Community Research and Development Information Service - CORDIS


ANTIGONE Report Summary

Project ID: 278976
Funded under: FP7-HEALTH
Country: Netherlands

Final Report Summary - ANTIGONE (ANTIcipating the Global Onset of Novel Epidemics)

Executive Summary:
In recent years, an increased number of zoonotic viruses and bacteria have crossed the species barrier to humans and caused or threatened to cause human pandemics with high morbidity and mortality. Because of our inability to predict the emergence of these pathogens, it is difficult to take preventive measures. For any zoonotic pathogen to emerge from its original animal reservoir(s) and develop into a pathogen with human pandemic potential, it must successfully cross various barriers. These barriers may be divided into interspecies, intrahuman barriers and interhuman barriers. It is poorly understood which factors allow zoonotic pathogens to successfully cross these barriers. In order to identify the key factors that render zoonotic pathogens prone to cross the species barrier and gain efficient transmissibility among humans, ANTIGONE had the following main objectives:
•To identify and understand the key factors that render zoonotic viruses and bacteria with human pandemic potential prone to cross the species barriers, adapt to the human host and further to gain human-to-human transmissibility;
•To translate our increased understanding of key factors in the chain of emergence to risk assessment, and options for prevention and intervention of human pandemics emerging from zoonotic pathogens;
•And to develop and implement a One Health training programme, combining human and veterinary medical expertise with those from other relevant disciplines, in order to equip the future generation of scientists with the necessary knowledge to deal with emerging zoonotic infectious diseases.
In order to reach these objectives, we had a two-pronged approach. First, we performed primary research studies to fill important gaps in our understanding of how zoonotic pathogens can gain pandemic potential. Integral to these activities was a cross-disciplinary training programme for young scientists (Young ANTIGONE) and the development of risk assessment models. Second, we organized expert workshops where experts discussed key issues. In research on key factors that allowed pathogens to cross from different animal reservoirs to humans, ANTIGONE partners played an important role in our understanding of the newly emerged MERS-CoV, such as identification of its receptor, and establishing the central role of dromedaries in human exposure to MERS-CoV. Regarding the intrahuman barrier, a new method—antibody landscape—was developed to study the role of antibody-mediated herd immunity on antigenic variation among pathogen strains, and indicated that pre-emptive vaccine updates may improve influenza vaccine efficacy in previously exposed individuals. The research on the interhuman barrier has been in acceleration in part as a result of research on HPAIV H5N1, which showed that several recently circulating isolates of HPAIV H5N1 are only a few mutations away from becoming efficiently transmissible among mammals, and highlighted critical areas of research to mitigate this risk. The level of knowledge gained for efficient human-to-human transmissibility of influenza viruses, is valuable for similar studies in other respiratory pathogens. Important research was performed on the relationship between climate change and emergence of tick-borne diseases, indicating that the combination of climate, host abundance and social factors may explain the upsurge of epidemics transmitted by ticks to humans. For risk assessment of introduction of bat-borne zoonotic pathogens in Europe, a generic model framework was developed. A special focus in ANTIGONE is on the E. coli O104:H4 outbreak. Research on rapid, sensitive, and specific detection of O104:H4 and the other major EHEC serotypes that cause haemolytic uremic syndrome, resulted in a system of real-time multiplex PCR-based detection assays that could be used as surveillance tool. ANTIGONE has contributed to a better understanding of the emergence and transmission of pathogens with pandemic potential. This will help to predict the zoonotic and pandemic potential of pathogens, and improve our abilities of early detection and effective response against outbreaks of emerging zoonotic pathogens. In addition, ANTIGONE’s combined training in human medicine, veterinary medicine, and other critical disciplines—“One Health”—has prepared a new generation of scientists to deal with tomorrow’s global health problems. We thus expect ANTIGONE to not only create a step-change in scientific basis for, but critically also in the implementation of, broadly effective preparedness plans, prevention strategies, and intervention strategies.

Project Context and Objectives:
In order to identify the key factors that render zoonotic pathogens prone to cross the species barrier and gain efficient transmissibility among humans, ANTIGONE has the following main objectives for the entire five-year duration, clustered into five main activity areas: expert workshops, primary research studies, risk assessment and modeling, Young ANTIGONE, and management:
• To identify and understand the key factors that render zoonotic viruses and bacteria with human pandemic potential prone to cross the species barriers, adapt to the human host and further to gain human-to-human transmissibility.
• To translate our increased understanding of key factors in the chain of emergence to risk assessment, and options for prevention and intervention of human pandemics emerging from zoonotic pathogens.
• To develop and implement a One Health training programme, combining human and veterinary medical expertise with those from other relevant disciplines, in order to equip the future generation of scientists with the necessary knowledge to deal with emerging zoonotic infectious diseases.
Figure 1. ANTIGONE framework and main acitvity areas.
In order to reach these objectives, we have a two-pronged approach. First, we will perform primary research studies to fill important gaps in our understanding of how zoonotic pathogens can gain pandemic potential. These studies will focus on selected viruses and bacteria, including Severe Acute Respiratory Syndrome coronavirus (SARS-CoV), Crimean-Congo haemorrhagic fever virus (CCHFV), Nipah virus (NiV), Ebola virus (EBOV), Escherichia coli (E. coli), Mycobacterium bovis (M. bovis), Borrelia burgdorferi (B. burgdorferi), Coxiella burnetii (C. burnetii) and Streptococcus suis (S. suis). Integral to these activities will be a cross-disciplinary training programme for young scientists (Young ANTIGONE) and a web-based pathogen information sharing platform. Second, we will organize expert workshops where experts from the human and veterinary fields and other relevant disciplines, from within and outside ANTIGONE, will discuss key issues in infectivity, pathogenicity, and transmissibility of zoonotic pathogens and determine general criteria to assess the risk of these pathogens to gain human pandemic potential.

Project Results:

The overall objective of work packages 1 to 4 was to identify key factors that allow pathogens to cross the interspecies barriers. This was studied for non-vectorborne pathogens, vectorborne pathogens, and Escherichia coli.

Interspecies barriers for non-vectorborne zoonotic pathogens
Non-vectorborne pathogens studied included Nipah virus, Lagos bat virus, Duvenhage virus, zoonotic influenza viruses, MERS Co-V, and Ebola virus.

Exposure of European bats to Nipah virus
In order to determine if European bats host henipaviruses, we obtained available bat serum samples from laboratories across Europe and we proactively blood-sampled bats in Spain and Cyprus, since these European countries border Africa. Also, a species of bat, the Egyptian fruit bat (Rousettus aegyptiacus), that is known to carry henipavirus in Africa is known to live in Cyprus. This species was also tested from the Middle East. We used the Luminex platform – a rapid-throughput binding assay – to detect antibodies that bind to either the F or G proteins of Nipah virus, Hendra virus or Cedar virus.
Archived samples comprised 340 Daubenton’s bat (Myotis daubentonii) serum samples from the United Kingdom and 70 serotine bats (Eptesicus sp.) from Spain. Proactive sampling resulted in the collection of 130 sera from Schreiber’s bats (Miniopterus schreiberseii) from Spain and of 35 Egyptian fruit bat (Rousettus aegyptiacus) and 10 bentwinged bat (Miniopterus sp.) serum samples from Cyprus. In addition, sera from 18 Egyptian fruit bats from the United Arab Emirates were tested. None of the samples gave higher than expected background readings for mean fluorescence intensity (MFI), indicating that all tested negative for each of the analytes used. There was, therefore, no evidence of exposure of the European bats tested to Nipah virus or to related henipaviruses.

Factors determining risk of human exposure to Lagos bat virus
Lagos bat virus (LBV) is a zoonotic pathogen carried by the straw-coloured fruit bat (Eidolon helvum), amongst other African bats. This phylogroup 2 lyssavirus causes rabies in terrestrial mammals. Vaccination against classical rabies virus (phylogroup 1) does not protect against LBV. To understand risk factors for public health, an understanding of viral infection dynamics in reservoir hosts is required. Three LBV isolates obtained from wild E. helvum were available, including an isolate obtained from Ghana during the course of the Antigone project. For this project, we bred E. helvum bats in captivity to provide experimental animals of known (negative) infection history for Lagos bat virus. Also, a biosecure facility for the carrying out of lyssavirus infection studies in Accra, Ghana was constructed specifically for this research.
Prior to investigating viral infection dynamics in E. helvum, we conducted a pilot study to determine if each of the available LBV isolates would cause disease should it reach the brain of this host. Each of three groups of captive-bred E. helvum were experimentally inoculated intracerebrally with one of the LBV isolates. Equal titres of virus were inoculated across all isolates using a standard site and rate of inoculation. A fourth group was mock-infected using virus-free tissue culture fluid. While mock-infected animals remained healthy throughout, each LBV isolate produced a distinct disease phenotype (weakness - Nigeria strain; hyperactivity - Senegal strain; aggression - Ghana strain), each with a clinical duration of < 24 hours. The incubation period and the pattern of viral spread to other tissues, including salivary glands, also appeared to be isolate-specific.
In a second experiment, we determined the lowest viral dose of Ghana strain (deemed preferable because it had the least cell passages and resulted in an obvious, distinct disease phenotype) to cause 100% productive infection. Each of five doses of Ghana strain was inoculated intramuscularly into four E. helvum. The end point was 90 days after last death due to rabies to allow enough time for viral incubation and possible recovery, as might occur in the wild. A blood sample (for serology) and a saliva swab (for virus detection) were taken from each animal immediately prior to inoculation, then twice per week for the first three weeks post-inoculation (PI). From 22-42 days PI, these samples were taken from each surviving bat once/week, followed by once every other week from 6 weeks PI onwards. A saliva swab was taken daily from each bat that became ill.
The number of bats that developed clinical signs showed no clear correlation to dose. There was no clear correlation between seroconversion and development of disease. The virus was neurotropic in all diseased bats. Only one surviving bat (inoculated with a 10-1 viral dose) was seropositive at the end of the experiment.

Factors determining excretion of Lagos bat virus
In our experimental animal model, of the three LBV strains tested, the Ghana strain was shown to be most suitable based on efficient inter-neuronal transmission and ability to spread to peripheral organs. Lyssavirus antigen was not expressed in the salivary glands of any of the inoculated bats, but was present in the taste buds of 4/6 bats examined that died with rabies, suggesting possible viral excretion via taste buds. All saliva swabs examined (from all bats which developed rabies and from one bat that did not) were negative for LBV using qPCR. Thus, the tongue may be a relatively more important site of virus excretion in bats, than the salivary gland. These advances improve our ability to assess the risk of bat lyssaviruses for zoonotic transmission

Characterization of Duvenhage virus
The main objective was to identify key factors in infection dynamics of Duvenhage virus, a member of the lyssavirus genus of the family Rhabdoviridae. We intended to study the tropism and replication efficiency of Duvenhage virus for different neuronal cell types and cells of the salivary glands in vitro and in vivo and determine the efficiency of virus transport from the brain to the salivary glands in animal models in vivo.
We characterized Duvenhage virus (DUVV) infections in vitro (mouse and human neuroblastoma cell lines) and in vivo (BALB/c mouse model of infection). For this purpose a close to primary clinical isolate (DUVV-NL07; p2 on human neuroblastoma cells) was used. The reduction of viral titers in N2a cells over time could be explained by cell loss observed in these cells after two weeks in culture. Inoculation of BV-2 cells or human astrocytes with DUVV-NL07 at low moi (moi=0.1) did not result in productive infection as no infectious virus or viral RNA could be detected at different time points post inoculation.
DUVV-NL07 has also been characterized in vivo in 8-week old BALB/c mice. A high dose of virus (10^6 TCID50) administered intramuscularly was able to induce lethal encephalitis in 50% of inoculated animals. Confocal microscopy has confirmed that DUVV-NL07 is exclusively neurotropic (similar to other members of the lyssavirus genus) since only cells expressing neuron-specific markers are co-expressing DUVV-NL07 antigen in the brains of DUVV-infected mice. Full-length genome sequence of DUVV-NL07 was obtained using standard Sanger sequencing, and next generation sequencing. Exact leader and trailer secuences were obtained using the 3’ and 5’end RACE method. Next, we aimed to establish in our laboratory the reverse genetics system for wild-type lyssaviruses in order to study tropism and pathogenesis of DUVV in vivo. As both DUVV and SHBRV are wild-type viruses, standard existing protocols of reverse genetics that apply to lab-adapted viruses had to be extensively optimized to be applicable for these wild-type viruses. At the end of the project recombinant virus (rDUVV) was rescued from BSR-T7 cells using homologous helper plasmids, albeit at low titers (only 103 pfu/ml on N2A cells).

Cross-species infections of selected zoonotic influenza viruses using the ferret and swine as a model for influenza in humans.
Candidate contemporary swine H3N2 viruses were sourced and approved via the FP7 ESNIP3 (KBBE2010-13-05). Chronological panels of human sera were selected to determine cross reactivity between human H3N2 antibodies (natural exposure and following vaccination) and swine H3N2 viruses (non-TRIG). These showed that over a period of approximately 40 years human viruses of this subtype have diverged away from the common ancestors while swine viruses have remained more static, contemporary viruses will be compared to historical isolates e.g. 2011, 1992, 1984, 1973 (PC-like).
Contemporary European swine H3N2 A/swine/Italy/55925/2010 was selected as the candidate virus strain and all ethical approval for all animal experimentations (WP7 and 11) were received and forwarded to the commission (19/6/12). Pilot studies were undertaken to establish the susceptibility in swine and ferrets to a contemporary European swine H3N2 isolate - A/swine/Italy/55925/2011 in order to generate base line data for WP7 swine to ferret (human) and WP11 onward human-to-human transmission using a ferret model with naïve and vaccinated (with seasonal human vaccine 2011/12) animals.
A comparison of pathogenesis was determined following aerosol inoculation in the swine primary host (n=5) and the ferret (n=6) novel host at 6 and 10 weeks old respectively. Pigs and ferrets were innoculated intranasally with A/swine/Italy/55925/2011 amplified in embryonated fowls eggs at a final dose of 105 EID50. Both species were monitored for clinical signs and sampled daily (including nasal swabs/nasal washes). Shedding profiles were determined by Matrix gene RRT-PCR.
There was no clinical disease in the pigs, and mild clinical signs were observed in the ferrets. Both species were successfully infected as indicated by the detection of viral RNA from 1 day until dpi 6 in the pigs and dpi 9 in the ferrets. A double peak of shedding was detected at dpi 2 and 5 in the ferrets, whereas one peak of shedding was detected in the pigs between dpi 2-4. Post-mortems at dpi 5 in both species demonstrated gross lung lesions.
In addition, the transmission of swH3N2 from ferret-to-ferret was also assessed, where 3 naïve ferrets were placed directly in contact with 3 infected ferrets from dpi 2 onwards. All contact animals became infected from 1-2 days post contact (dpc) with a double peak of virus shedding at 4 and 7 dpc. Both species are highly susceptible to swH3N2, and transmission was confirmed in the ferrets.
In a second experiment, pigs were infected (n=8) with swH3N2 A/swine/Italy/55295/2011 (MDCK2/EP2), via intranasal-MAD route with a total of 4x10^6 EID50 at day one the pigs were co-housed with naïve ferrets (n=8) to determine if pig to ferret (model for human) transmission was possible via the airborne route. The two species were separated by a barrier with a 10cm gap. Close proximity was encouraged through animal husbandry and environmental enrichment.
Successful airborne transmission from infected pigs to co-housed ferrets was demonstrated in 3/4 pairs (6/8 animals) with proximity to the pigs a significant factor in the infection of the ferrets. The level and duration of virus shedding in the ferrets (dpi 1-16) was much greater than that in the pigs (dpi 1-10) (area under the curve was 42.4 and 24.2 respectively) almost 2-fold different. No clinical signs were detected in either species. Sero-conversion in the ferrets was 100%.

Discovery of hepadednaviruses in bats
We developed multiple novel pathogen detection techniques to be able to identify unknown viruses. In addition, biobanks (>10,000 arthropod, small mammal, livestock and human samples), and cell culture systems (>200 cell lines comprising >20 species) were established in order to be able to isolate and characterize novel viruses.
The focus of the virus ecology studies was on small mammals (bats and rodents) and livestock. We published high impact articles dealing with novel viruses identified in small mammals. For the first time, bat-specific ancestral Hepatitis E viruses were identified (Drexler et al., JVI, 2012). The zoonotic potential of the bat-associated Hepatitis E viruses is yet to be explored. Novel bat hepadnaviruses were identified that were able to use the human Hepatitis B receptor but were not neutralized by sera of Hepatitis B-vaccinated individuals suggesting a spillover potential and at the same time a lack of human population immunity (Drexler et al., 2013, PNAS). In addition, novel rodent hepaciviruses were identified. In light of the absence of a suitable Hepatitis C animal model these viruses may pave the avenue for the establishment of a rodent hepacivirus animal model (Drexler et al., 2013, PLoS Path).

Key role in studies on emergence of MERS
We played a major role in the MERS outbreak response. A prerequisite for all MERS-Coronavirus-related ecological and epidemiological studies was the development of state-of-the art diagnostic tools. This included viral nucleic acid as well as antibody detection methods. The first diagnostic algorithm for MERS-Coronavirus detection was already published early after the first MERS cases occurred (e.g. Buchholz et al., 2012 Eurosurveillance). The UBMC-developed real time RT-PCR became a standard assay for laboratories around the world and was included in the WHO diagnostic recommendations. More than 200 laboratories in >50 countries have received material and support from UBMC for MERS-Coronavirus diagnostics. We wrote a review summarizing the pitfall and challenges of serodiagnostics for emerging coronaviruses to raise awareness in the scientific community (Meyer et al. 2014, Virus Research). The translation of the in-house diagnostic tools into commercial kits has been facilitated (Altona Diagnostics, TIB MOLBIOL, EUROIMMUN). The commercially available kits enabled laboratories around the world to perform MERS diagnostics independently. Antigone partners became expert advisors for WHO diagnostic recommendations. We organized the first WHO diagnostic workshop in Dubai to raise awareness and improve diagnostic skills in MERS endemic areas. More than 40 international laboratory experts from 20 different countries attended the course. As part of the MERS rapid response, samples were used that had been accumulated in frame of the ANTIGONE project. The samples were the basis to explore the putative reservoir spectrum of this newly emerged coronavirus. By adapting previously established coronavirus detection assays Antigone partners forerunners in resolving the ecology and epidemiology of MERS-Coronavirus.
We were one of the first to show a link between dromedary camel exposure and MERS-Coronavirus infection (Memish et al., 2014 EID). In the so far largest cross-sectional study on more than 10,000 individuals in Saudi Arabia, we showed that MERS-Coronavirus may be far more widespread than previously thought and may cause subclinical infections (Müller et al., 2015, Lancet Infectious Diseases). In addition, this study showed that the infection risk for dromedary-exposed individuals was indeed elevated.
In addition, we identified MERS-Coronavirus ancestor viruses in bats (Annan et al., 2013 EID, Ithete et al., 2013 EID, Corman et al., 2014, Journal of Virology) and provided first evidence for dromedary camels to be a putative source of infection (Reusken/Haagmans/Müller et al. 2013, Lancet Infectious Diseases). Multiple follow-up studies showed a wide temporal and geographical distribution of MERS-Coronavirus in dromedaries in Africa and The Middle East (Corman et al., 2014, EID; Meyer et al., 2014, EID; Reusken et al., 2013, Eurosurveillance). We organized and tested samples that dated back to the year 1983, providing first empirical evidence that MERS-Coronavirus has been circulating in African dromedaries for at least 30 years prior to emergence (Müller et al. 2014, EID). Finally, were the first to show that young dromedary camels may be an important source of infection (Wernery et al., 2014, EID).

Sociological aspects of the emergence of MERS
We aimed to define modes of collaboration between anthropologists, ecologists and microbiologists on the human/animal/environment interface in the emergence of new pathogens. This meant defining common questions and identifying complementarities between disciplinary methods. This included writing the interdisciplinary article “Crossing the Interspecies Barrier: Opening the Door to Zoonotic Pathogens » at the Dahlem workshop of the Antigone project in September 2013, and organizing a conference at the Collège de France in June 2013 entitled « Zoonoses and emerging infectious diseases. When biologists meet anthropologists » in which two members of the Antigone project met with biologists and anthropologists from France and the rest of the world.
The next step was to define a fieldwork common to biologists and anthropologist. After building trusting relationships with the public health authorities in Qatar under the supervision of Elmoubasher Abu Bakr Abd Farag, we sent to Doha an anthropologist specialized in camels domestication in Africa, Sarah Cabalion. She made two missions in Qatar in December 2014 and in May 2016 to study contacts between camel workers and camels and their potential role in the emergence of MERS-CoV.
Two main targets were defined : the camel race course of Shahanyia, 30 kilometers north-west of Doha, and the camel market of Doha. It is estimated to be around 700 000 camels in Qatar, with 30 000 racing camels. As the grazing system in the desert has been prohibited ten years ago, camels are concentrated in markets and farms. While the life standards of Qatari citizens have raised significantly in this period, the consumption of camels for meat and for races has also increased, as well as imports of human workforce and animal resources from abroad.
While it has not been possible for Sarah Cabalion to study migrant workers at the market and slaughterhouse of Doha, she has been able to study relations between owners, workers and camels in the Shahaniya race course and farms. Her main finding is that Qatari citizens are very attached to their camels, who represent for them nomadic life in the desert, which explains why the dream of rich Qataris is to have a farm in Shahaniya where there can spend their weekends. While owners keep at a distance from camels (they give them orders from their cars when they run, and they give them a few strokes when they come to the farm), workers are in daily contact with camels, which they feed and provide with medicine. Consequently, when veterinarians and public health officers came from Doha to check the health of camels, owners and workers didn’t believe that camels were responsible for the transmission of MERS-CoV and rather invoked the role of birds.
These qualitative findings, based on interviews and long-term observations, confirm the quantitative studies based on polls made by the public health administration of Qatar. The results of this collective research have been quite contrasted : while Mongolian breeders had a high level of awareness of zoonotic pathogens, because the government has communicated heavily about the threat of zoonoses for the Mongolian economy (where one third of resources come from animal breeding), elephant workers in Laos denied them in a way that can be compared to camel workers in Qatar. Australian Aboriginals have another perception: for them, bats are totemic animals with which they have always been in relation, which explain that they can eat them as food and medicine, horses and camels are imported species with which they have developed distanced relations.
The perception of camels in Australia and the Arab Peninsula has consequently been compared by our group in contrast with the perception of bats, even if the role of bats in the emergence of MERS-CoV has not been identified. This study might be useful to study animal workers in the Doha market in the future, since a third of camels imported for meat come from Australia.
The remaining task is to compare camel breeding in the Arab Peninsula and in Sudan, as most camels raised for the racecourse come from Sudan. One of the main hypotheses for the emergence of MERS-CoV, is that the weaning of camel calves in Sudan has been shortened to increase importation, which may have increased the period of exposure of young calves to viruses. This hypothesis should be tested by observing relations between mothers and calves in Sudanese farms and by tracing the roads of camel exportation between East Africa and the Arab Peninsula.

Factors determining risk of human exposure to filoviruses
Bats host an array of zoonotic pathogens, but there is little research on how bats and humans interact, how people perceive bats and their accompanying disease risk, or who is most at risk. We, therefore, investigated epidemiological, anthropological and behavioural factors that determine risk of human exposure to key zoonotic pathogens.
The straw-coloured fruit bat (Eidolon helvum) is a large, abundant species of bat in sub-Saharan Africa. It is widely hunted and eaten, including in Ghana, and it is known to carry potentially zoonotic pathogens. This raises concerns, as hunting and butchering are common sources of zoonotic transmission. Interviewing 577 Ghanaians, we identified the characteristics of people involved in the bat-bushmeat trade and explored their perceptions of risk. Our methods and results are described in Kamins et al. Ecohealth 12,104-120;2015. This work builds on our study of the bat bushmeat commodity chain that had been published prior to the West African Ebola crisis (Kamins et al. Biol Cons 144,3000-3008;2011).
Bat-bushmeat actors comprised a complex array of groups. Bat hunting, selling and consumption are widely distributed across regional and ethnic lines, with local hotspots, while butchering is predominantly done by women and hunters. Bats were easier for less active men to hunt than other species. Interviewees held little belief of disease risk from bats, saw no ecological value in fruit bats and associated eating bats with specific tribes. Generally, disease risk was considered greatest by those who did not eat bats and lowest by those who hunted or sold bats. It is unclear whether a lack of familiarity with bats leads to a higher perception of risk, or if this decreases the likelihood of involvement with bat bushmeat.

Ebolavirus reservoir infection dynamics
We studied longitudinal antibody responses to ebolaviruses in three sympatric fruit bat species (Epomophorus gambianus, E. helvum and Rousettus aegyptiacus) in four locations across Ghana. A Luminex immunoassay was used to detect Ebola virus (EBOV) antibodies. A statistical analysis was performed separately for each bat species to determine the cutoff value between seronegative and seropositive samples.
Statistically significant differences between the seroprevalences of the three species were demonstrated (p <0.001). For each species, there was no statistically significant difference between roosts. Independent variables, such as age, sex & weight, showed no significant effect on EBOV seroprevalence. There was, however, a consistently significant “year effect” for E. gambianus, with differences between 2013 (higher seroprevalence) and 2015 – 2016 (lower seroprevalence). Eidolon helvum did not show an overall “year effect” (ANOVA p >0.05) but seroprevalence was higher in 2013 than 2015 (p <0.05). Only E. gambianus presented a consistently significant effect of season, with higher seroprevalence in June-July (season of synchronous pregnancy).

Filovirus spillover dynamics
We used our Luminex platform to identify the presence of exposure to EBOV in people and domestic animals. Each sample was tested in duplicate. Raw MFI values were used and only when values of ≥ 1,000 were obtained for both replicates were these results interpreted as being seropositive for the purposes of this report.
In one study, rural villagers were blood-sampled one year apart. Of 297 people in year 1, sera from 25 (8%) gave MFI values ≥ 1,000. In year 2, 199 people were blood-sampled; 180 of whom were repeat samples from year 1. Of these, 158 remained seronegative and ten had “seroconverted”. Thirteen of the initial “seropositive” people were unavailable for re-sampling; of the 12 who were resampled, eight stayed “seropositive” while four became “seronegative”.
Additionally, people and their domestic animals in Ghana were tested for evidence of exposure to Ebola virus in four rural villages with differing degrees of bat hunting, as well as in the capital city of Accra. Bat hunting is most prolific in the Volta region and here we found an apparent human seroprevalence against EBOV of 19.3%, the highest of any location sampled. The next highest seroprevalence in people (13.2%) was in Ve-Golokwati, which is a village where people live amongst a large colony of bats (Epomophorus gambianus), in which there was a 7% seroprevalence to EBOV. There was also evidence of seropositivity to EBOV (or to a related virus) in domesticated animals, especially in pigs. Pigs are known to be amplifier hosts for a type of Ebolavirus in the Philippines, so our results are consistent with the known ecology of Ebolaviruses. We are now working with external reference laboratories to validate these results for future publication and risk assessment.
In order to identify the virus(es) stimulating the production of antibodies that react with our Luminex EBOV serology platform, we harvested tissues from non-human species in Ghana in which high seroprevalences were identified; i.e. E. gambianus, R. aegyptiacus (see deliverable 2.7) and the domestic pig, Sus scrofa domesticus. These tissues are now being examined for the presence of filoviruses using molecular methods and virus isolation. Additional funding (£600,000, 2017-2019) has been recently obtained from the UK Medical Research Council to enable the continuation of this work programme.
Furthermore, in a collaborative study with the Albert Einstein College of Medicine Ney York (Kartik Chandran) UBMC identified NPC1 as a genetic determinant of filovirus susceptibility in bats, suggesting that some NPC1 variations reflect host adaptations to reduce filovirus replication and virulence (Ng et al., 2015, eLIFE).

Does Bartonella spillover from bats to other species, including humans, occur?
Bartonella is a type of bacterium that can cause severe disease in people. It is generally known to be carried by domestic cats, but we identified widespread Bartonella infection in bats in sub-Saharan Africa. We, therefore, examined samples collected for our other studies for evidence of zoonotic transmission of bat-borne Bartonella in Ghana. We tested serum samples from 335 volunteers who had close contact with bats and also sampled 70 domestic animals that lived underneath bat colonies (5 cats, 23 chickens, 7 cows, 6 dogs, 21 goats, 8 sheep) in Accra. Using both serological and culture assays, we found no evidence of spillover of bat-associated infection (Mannerings et al, Emerg Infect Dis 22,922-924;2016).

Reservoir-pathogen interactions for Q fever
The obligate intracellular bacterial pathogen Coxiella burnetii is the causative agent of Q fever in humans. The most important infection source is ruminants which may shed the bacteria in high numbers via placental products, feces and milk. After internalization into host cells via phagocytosis, C. burnetii replicates especially in monocytes and other immune cells in parasitophoric vacuoles. We used epithelial cells of the site of entry into the ruminant host (airway / lung epithelia) and exit from the host (epithelia from placenta, mamma, intestine) and studied their permissiveness for C. burnetii invasion and replication quantitatively (monitoring replication kinetics) and qualitatively (electron microscopy of parasitophorus vacuoles). Dynamics of host cell responses with particular emphasis on (intracellular) microbicidal effector molecules and (extracellular) immune mediators were the main scope and were monitored at the proteome and transcriptome level. Biomarkers for prediction of epidemic potential of C. burnetii strains were deduced from comparisons between different bacterial isolates.
To analyze the interactions between C. burnetii and the host at the entry (lung) and the exit (placenta, intestine, udder) site, several epithelial cell lines have been collected and characterized. Maintenance of epithelial cell specific characteristics was confirmed by detection of cytokeratin and zonula occludens-2 protein. While all cells expressed tight junction proteins, only caruncular cell line BCEC12 fully developed into a polarized monolayer when grown on filter inserts as deduced from an increase in transepithelial resistance. To study permissiveness for bacterial invasion and replication, cell lines were inoculated with C. burnetii apathogen strain “Nine Mile phase II clone 4 (NMII)” and the pathogen “Nine Mile phase I RSA 493 (NMI)”. Formation of parasitophorous vacuoles varied relative to LPS-phase type between epithelia and between infection conditions. Bovine epithelial cell lines exhibited different permissiveness to C. burnetii while maintaining cell viability with udder cells (PS) allowing for the highest invasion rates. For infection of host cells, especially monocytes and macrophages, C. burnetii utilizes specific cell surface receptors, leukocyte response integrin (LRI, aVß3) and complement receptor 3 (CR3, aMß2). The expression of these two surface markers on epithelial cells was investigated. While the complement receptor was expressed on all cell lines selected (F3 – no invasion; FKDR - slight invasion rate; PS - high invasion rate) only the udder cell PS line expressed LRI, providing a first insight into the molecular basis for the strikingly different permissiveness of bovine epithelial cells for C. burnetii infection and replication.
Response of bovine epithelial cells to inoculation with C. burnetii (suspensions of both strains and of heat-inactivated bacteria) was also analyzed by quantifying cytokine expression (IL-1beta, IL-6, IL-8, IL-10, GM-CSF, MCP1, INF-gamma, TNF-alpha, TGF-beta, iNOS). The infected cells of all bovine epithelial lines did not differ in their cytokine expression profiles from uninfected control cells. Only small variations were seen in the profile of lung cells (Bel-26) with a down-regulation of IL-6 and a strong up-regulation of TNF-α during infection with NMII. The cell line PS, which in general had the highest infection rate, totally failed to mount an immune response to infection. C. burnetii replicates in acidic parasitophorous vacuoles inside eukaryotic cells, which require a possibility to manipulate the host metabolism for bacterial survival. To identify how the host is affected by C. burnetii infection we investigated the proteome of bovine udder epithelial cells before and after infection with C. burnetii strains NMI and NMII. We applied two-dimensional gel electrophoresis using fluorescent dyes (DIGE) and gel-free LC-MS/MS analysis with TMTsixplex labeling of whole cell lysate proteins. Overall, in both infections, we found 22 regulated bovine proteins, with most of them upregulated. Both strains led to the regulation of a wide spectrum of proteins involved in cell (energy) metabolism, cell compartmentalization and signaling, protein processing and regulation, stress and immune response. The pI distribution of the regulated bacterial proteins was quite noticeable. Most of them had a basic or acidic pI. This may support the survival of Coxiellae inside the acidic compartment inside the host cells.
C. burnetii was chosen as a suitable model bacterium to unveil factors determining the pandemic potential at sites of entry into and exit from the animal host. Initial contact of C. burnetii with the respiratory epithelium likely shapes the subsequent host response, paves the way for persistent infection, and determines the number of shedders. Intensity of shedding, by contrast, relies on the interaction of C. burnetii with placental, intestinal and mammary epithelia, i.e., the sites of exit. The primary research studies conducted within ANTIGONE have contributed to the definition of key factors determining intensity of shedding of C. burnetii from animal reservoir species at the individual and the population level. Our findings from in vitro studies clearly indicate that the tissue tropism of the agent and thereby the preferred route of excretion of this zoonotic pathogen is determined by properties of epithelial cells in different organs. These properties become effective at the cellular level rather than being shaped by multi-cellular interactions at the tissue level. C. burnetii strains from human and animal sources differ in their ability to interact with the respective epithelial cells. These findings pave the way to the definition of biomarkers for virulence and pandemic potential of C. burnetii.
In 2014 we identified vaginal shedding of C. burnetii in deer hinds, confirming red deer as a source of Q fever zoonotic infection, as confirmed by a human case (farm veterinarian; Gonzalez-Barrio et al. 2014). In 2015 we evidenced that host and environmental factors modulate the exposure of free-ranging and farmed red deer to Coxiella burnetii. Although red deer may maintain the circulation of C. burnetii without third species, the most frequent scenario probably includes other wild and domestic host species. These findings, taken together with previous evidence of C. burnetii shedding by naturally infected red deer, point at this wild ungulate as a true reservoir for C. burnetii and an important node in the life cycle of C. burnetii, at least in the Iberian Peninsula. We also studied the role of European rabbits as a reservoir for C. burnetii. High individual and population seroprevalences observed in wild and farmed rabbits, evidence of systemic infections, and vaginal shedding support the reservoir role of the European rabbit (Gonzalez-Barrio et al. 2015). Finally, in wild boar C. burnetii shedding was detected in nasal, rectal and genital swabs (Gonzalez-Barrio et al. 2015).

Shedding of Mycobacterium by host species
One of our goals was identifying key factors determining intensity of shedding of Mycobacterium bovis, the main causative agent of animal tuberculosis, from animal reservoir species at the individual and the population level. Research within ANTIGONE was performed both at the individual level, partly through challenge experiments, and at the population level, through cross-sectional and longitudinal studies using different animal hosts as models. An experimental infection in the goat model, using M. bovis, M. caprae and M. tuberculosis strains, revealed that there are Mycobacterium species-related differences in the clinical picture in this model host, essentially indicating a lower pathogenicity (and hence lower shedding) in animal hosts infected with human strains of M. tuberculosis as compared to animal strains of M. bovis and M. caprae (Bezos et al. 2015). Second, we reported the detection of MTC shedding in 80% of naturally MTC-infected wild boar and red deer. For the first time we provided evidence for the existence of a proportion of super-shedders within the naturally infected population of these host species. These super-shedders are responsible for a disproportionately large amount of MTC excretion from infected wild ungulates. MTC DNA concentrations greater than the minimum infective doses for cattle, red deer or wild boar are present in excretion routes from both species. These results have implications for the design of control programs in multi-host pathogen systems where these species are maintenance hosts for bTB (Santos et al. 2015).

Interspecies transmission of Streptococcus suis
Our specific aim was to study the transmission dynamics, host and pathogen factors that modulate infectivity and cell-tropism, and the human immune response to the model pathogen Streptococcus suis. Whole genome sequencing of invasive isolates from humans and pigs identified an emerging zoonotic clone of S. suis in the Netherlands, likely the result of mixing imported pig breeds in the late 19th century (PLoS One. 2012;7(5):e33854; Sci Rep. 2016 Jul 6;6:28984). This clone contained a novel type 1 restriction-modification system, which has not been described before in S. suis, and a pathogenicity island which showed similarities with a pathogenicity island observed in outbreak strains in China (Pathogens. 2016 Nov 18;5(4)). In vitro and in vivo studies of host-pathogen interaction in the intestinal tract showed a correlation of serotype and genotype with zoonotic potential and confirmed the intestinal tract as a potential site of entry for S. suis (Gut Microbes. 2016;7(2):154-62. J Infect Dis. 2015 Jul 1;212(1):95-105). The S. suis adhesin SadP was shown to contribute to intestinal infection by binding to the Gb3 receptor on the host intestinal cell surface (submitted for publication). Serological studies in pigs and humans indicated very poor immunogenicity of the S. suis serotype 2 capsule in both hosts, potentially contributing to zoonotic potential (manuscript in preparation).

Interspecies barriers for vector-borne zoonotic pathogens
Vector-borne pathogens studied included bunyavirus, Crimean-Congo haemorrhagic fever virus, Anaplasma phagocytophilum, and Borrelia burgdorferi.

A new clade of bunyaviruses
We identified a new clade of bunyaviruses in insects revealing a yet unknown diversity of arthropod-borne viruses that we may have underestimated (Marklewitz et al., 2013, Journal of Virology) The vast diversity of arthropod-borne viruses will help to understand the origin and evolution of viruses in general.

A new host for Crimean-Congo haemorrhagic fever virus
We provided evidence for widespread infection of African bats with Crimean-Congo hemorrhagic fever-like viruses. Tick-infested migratory bats may thus be a yet undiscovered host that may contribute to the distribution of distinct CCHFV lineages into distant geographic areas (Müller et al., 2016, Scientific Reports).

Host-tick-pathogen interactions for Anaplasma phagocytophilum
Our main objective was the identification of key environmental and genetic factors that allow vector-borne pathogens to cross the interspecies barriers. The model used in our experiments was based on the emerging tick-borne pathogen in the United States, Europe and Asia, Anaplasma phagocytophilum, which causes human and animal granulocytic anaplasmosis and tick-borne fever in ruminants. The characterization of host-tick-pathogen molecular interactions is essential to understand vector competence and disease transmission, and to develop novel control strategies for vector infestations and pathogen infection and transmission. In our studies, the interactions between A. phagocytophilum and the tick vector were characterized using latest “omics” technologies. The results revealed complex responses by both ticks and pathogens that were necessary for maintenance of tick fitness while ensuring vector capacity. At the tick-pathogen interface, A. phagocytophilum induces an antifreeze glycoprotein and heat shock proteins to increase tick survival, feeding fitness and vector capacity. Mechanisms have evolved which enable A. phagocytophilum to subvert host immune response, favoring pathogen infection, multiplication and transmission. Common mechanisms were demonstrated for pathogen infection of vertebrate and tick host cells, and include but are not limited to, remodeling of the cytoskeleton, inhibition of cell apoptosis, manipulation of the immune response, reprogramming of several metabolic processes and control of host cell epigenetics. The results also showed that the evolution of the tick-host-pathogen molecular interactions resulted in conflict and cooperation between them with mutual beneficial effects for ticks, hosts and pathogens.
Changes in environmental traits were identified that impact the prevalence of pathogens in ticks affecting the (a) tick mortality, activity rates, abundance and seasonality, (b) host phenology, changing the contact rates between reservoir host(s) and ticks, and (c) changes in the host community, producing different rates of circulating pathogens. We found that A. phagocytophilum genotypic diversity follows a bioclimatic gradient. Evidence suggests that human and ruminant infecting strains may be maintained by different reservoirs, therefore introducing uncertainty in assessing the environmental processes affecting the transmission cycles of this pathogen.
In summary, genetic and environmental factors were identified that affect pathogen infection and transmission by ticks. These results illustrate evolutionary mechanisms by which A. phagocytophilum activates bacterial biological processes and manipulates tick protective responses and other biological processes in order to facilitate infection, while ticks respond to limit pathogen infection and preserve feeding fitness and vector capacity for survival of both ticks and pathogens.

Ecological and Epidemiological risk factors for acquisition of Crimean-Congo hemorrhagic fever virus (CCHFV) infection in humans
CCHFV (genus Nairovirus, family Bunyaviridae) causes severe disease in humans with fatality rate up to 30%. The disease is characterized by high fever, headache, malaise, while in severe cases, hemorrhagic manifestations are present. Endemic foci of the disease are present in many areas of Africa, Asia, and Europe. Humans become infected with CCHFV through tick bites, mainly from Hyalomma spp. ticks, or by direct contact with blood or tissues from viremic patients or animals. CCHFV is characterized by a great genetic variability and CCHFV strains can be grouped into 7 genetic lineages. Among them, lineage Europe 1 contains pathogenic CCHFV strains from Balkans, Russia and Turkey, while lineage Europe 2 contains the CCHFV strain AP92, isolated from Rhipicephalus bursa ticks collected in 1975 from goats in Vergina village in northern Greece.
Greece is a country in the south part of the Balkan Peninsula. Although sporadic cases or outbreaks of CCHF are often observed in other Balkan countries such as Bulgaria, Albania, and Kosovo, only one, fatal, CCHF case has been reported in Greece. The causative strain clusters with other pathogenic Balkan CCHFV strains, differing by >20% at the nucleotide level from the CCHFV strain AP92. Previous studies showed that CCHFV seroprevalence in Greece ranges from 0% up to 14%. In the absence of human cases, it has been suggested that a non- or low-pathogenic strain, like AP92, is circulating in the country causing asymptomatic infections and high seroprevalence.
In the frame of ANTIGONE project, studies were performed aiming to understand better the unique CCHF epidemiology in Greece (high seroprevalence and absence of CCHF cases). Focused studies were performed in areas with high seroprevalence, followed by a spatial cluster analysis to assess the geographical variations in CCHFV seropositivity and to identify the risk factors.
It was shown that the overall CCHFV seroprevalence is 3.8%, with significant rate difference between the eastern and western part of the country. Apart the risk factors described in previous studies (age, sex, tick bite, agropastoral activities), the altitude, the land cover type and the transitional woodland/shrub land per person, as well as the number of livestock per person, and specifically the number of goats, sheep and cattle per person, were shown to affect significantly the seroprevalence. It was suggested that tick studies were needed to identify the circulating strains and unravel the mystery of CCHF epidemiology in Greece.
More than 2,000 ticks were collected from various geographic areas of the country. The species with the widest distribution were Rhipicephalus sanguineus s.l. (64.8%) and R. bursa (25.9%), while the population of H. marginatum was low (0.2%). It has to be mentioned that in CCHF endemic areas, the population of H. marginatum ticks is high (>60% of the ticks). An initial study was performed in ticks collected from ruminants in Kastoria region (northeastern Greece), and a novel AP92-like strain (lineage Europe 2) was detected in R. bursa ticks, differing from AP92 by 9.7% at the nucleotide level. During the last year of the project, an AP92-like strain was isolated in cell culture and fully sequenced using next generation sequencing. This is the second isolate of CCHFV Europe 2 lineage. Additional studies on the whole tick collection showed that several Riphicephalus spp. ticks were infected by CCHFV lineage Europe 1 (the lineage with the pathogenic strains). Rhipicephalus spp. ticks are three-host ticks, while Hyalomma marginatum (main vector of CCHFV) is a two-host tick, which may play a role in tick competence. The conclusion of these studies is that both CCHFV lineages are present in Greece: Europe 2 (AP92-like strains) and Europe 1 (pathogenic strains), which however, was detected in Riphicephalus spp. ticks (and not in Hyalomma marginatum ticks, which were present in very low numbers), suggesting the CCHF incidence is related not only with the circulating CCHFV strain, but also with the species and the population of the ticks.
A number of collected ticks were tested for additional tick-borne viral and bacterial pathogens. Several Rickettsia and two Anaplasma species (A. phagocytophylum and A. platys) were detected. Among Ricketsiae, some species were detected for the first time in Greece (R. africae, R. monacensis). In addition, a variety of Coxiella-like endosymbionts were detected in all tick genera tested; their exact role remains to be elucidated.
Regarding viral microorganisms, two novel phleboviruses were detected in Greek ticks: 1) Antigone virus, detected in 6 pools of R. sanguineus ticks collected from sheep in the mainland of Greece, and 2) Lesvos virus, detected in two pools of Haemaphysalis parva ticks collected from sheep in a Greek island (Lesvos Island). The new viruses differ highly each other and from all other known phleboviruses; however, they cluster together with all other tick-borne phleboviruses, including the recently identified severe syndrome with thrombocytopenia syndrome virus (SFTSV) and Heartland virus (HRTV), which cause severe disease in humans. Further studies are needed to show whether Antigone and Lesvos viruses are of interest for the public or animal health.
The Dahlem meeting on human-to-human (H2H) transmission included pathogens transmitted by arthropods (mainly mosquitoes). Among pathogens, most are viruses (such as Dengue, Zika, Chikungunya viruses), but bacteria (such as Rickettsia prowazeckii) and parasites (like Plasmodium spp.) are also included. The H2H transmission is the result of complex interactions between pathogens, vectors and humans, which are strongly influenced by human interventions and demographic and environmental conditions

Tick-host-pathogen interactions for Borrelia
During the transmission events molecules from all three organisms, host-tick-Borrelia, interact in two overlapping and complex micro worlds, - the feeding lesion in the vertebrate host and in vector midgut. Switching it antigenic profile and interaction with vector and host proteins enables survival of Lyme disease spirochete in both worlds.
A better understanding of three-way interactions between spirochete, the tick and the host is necessary to address the question “What makes the Lyme borreliosis system so successful?” Diverse interaction that occurs between spirochete, the tick vector and the host makes Borrelia an elusive pathogen.¬¬¬¬¬¬¬¬¬
Since all Borrelia species are host-propagated bacteria that shuttle between a vertebrate and tick, the spirochetes developed strategies to sense and survive in these diverse environments. This is achieved by altering the level of gene expression in response to changes in temperature, pH, salts, nutrient content and other host and vector dependent factors. To differentially synthesize proteins during infection cycle, Borrelia must possess regulatory networks to sense its environment. Quorum sensing is a mechanism by which bacteria monitors the surrounding by producing and responding to signaling molecules known as autoinducers. The complex life cycle and mechanism of pathogenesis means that B. burgdorferi precisely senses its environment and regulates protein expression accordingly. An individual bacterium must interact with different tissue types during its cycle, derive nutrition from warm-blooded hosts as well as from vector tick, and avoid clearance by hosts and vectors immune systems. Through quorum sensing, a whole population of Lyme disease spirochetes can synchronize production of proteins needed for infection and survival. There are additional benefits in having the whole population of Borrelia to coordinate certain functions. For example, simultaneous transmission of large numbers of spirochetes from a tick to a warm-blooded host might increase the odds that at least some bacteria survive the host’s immune system and establish infection.
As soon as the spirochetes contact complement in the host blood meal, innate immunity comes into play. At the beginning of infection, before any pathogen-specific antibodies are made, the invading spirochetes are attacked by the non-specific alternative complement pathway. Borrelia expresses several complement regulator-acquiring surface proteins that can bind regulatory proteins of the alternative complement pathway. Coated with mentioned host regulatory proteins, spirochetes avoid recognition and eradication by complement.
The interaction between spirochetes with complement has a major impact on spirochete transmission dynamic. Complement mediated selection operates directly in the midgut of feeding tick, so spirochetes are destroyed directly in tick, prior to transmission to the host. Vertebrate host range for Borrelia is determined by spirochetes sensitivity to complement of particular animal species; the ability to bind factor H and FHL-1 appears to depend on the Borrelia genotype. Vertebrate hosts may be infected concurrently with different genospecies of B. burgdorferi sensu lato, even with those for which they are apparently not transmission competent. A possible explanation for this apparent paradox brings us back to the beginning of this discussion - paradox may be related to the fact that Borrelia genes are differentially expressed during the life cycle. Studies to identify complement receptors on the spirochetes and to unveil the protective mechanisms of spirochetes against complement represent one of the fascinating area in complete understanding of three way interactions among spirochetes, the tick and the host. For clearer definition of spirochete molecules associated with the pathogenesis of Lyme borreliosis, the better understanding of the adaptive mechanism of this spirochete is required.
After the host adaptive immune response becomes effective, spirochetes disappear from the bloodstream but are disseminating to host internal organs, synovial fluid, collagenous tissue of the skin and joints. They simply need to survive until they are transmitted back to competent ticks and they find protection from host immune system in host tissues. The genome of B. burgdorferi does not contain any known virulence genes. At this time the successful persistence of spirochete within host depends on evading the host’s immune system, rather than exploiting the host tissues for reproduction or growth. The extracellular matrix provides an immune-privileged environment for spirochetes.
B. burgdorferi encounters the host immune system initially upon deposition into the skin and subsequently in the bloodstream en route to target tissues. Cellular components of peripheral blood mononuclear cells (PBMCs) are found in the inflammatory infiltrate of erythema migrans (EM) lesions and thus comprise a population of innate immune cells that mediate the initial interactions between host and pathogen. PBMCs constitute a readily accessible source of host immune cells. To examine the global transcriptional response of host to B. burgdorferi infection, an ex vivo co-culture model of infection using PBMCs was widely employed. It is believed that innate immune responses to Borrelia are primarily triggered by the spirochete's lipoproteins signaling through cell surface TLR1/2.
Although Borrelia evolves its own mechanism to fight host immune defense, the transmission via tick saliva provides spirochetes with certain adaptive advances.
Host-tick interactions that influence Borrelia populations include the processes whereby ticks (a) contact host and (b) successfully obtaining replete blood meal. The ixodid tick engorges for 4-8 days in order to feed to repletion. While it gives the spirochete a lot of time to migrate from vector to host or from host to vector, the feeding side is still a battleground, not pleasant to any participant. With a growing knowledge of tick transcriptome, it is becomes clear that Borrelia has a distinct advantage in entering the host via tick spit, as it employs the tick salivary proteins for self-protection.
Advances in modern molecular, genomic and proteomic tools have increased our ability to recognize and characterize the events that are important in enhancing of Borrelia transmission either directly or indirectly. The driving goal that motivates much current researches that are narrowed at vector-host-pathogen triangle is a quest for anti-pathogen or anti-tick vaccines. The goal that our team completed was by using the modern approach for the analysis of the interaction between the participants of vector-host-pathogen triangle to identify the factors involved in pathogen transmission and survival and, markers of pathogenicity and resistance for development of control measures.

Interspecies barriers for Escherichia coli
One of our goals was studying the prevalence and distribution of virulence factors of human enteropathogenic E. coli, of the O104:H4 serotype, and of antibiotic resistance genes in animal and environmental reservoirs, as well as in humans. We described a new set of quantitative real-time PCR assays for the direct detection and quantification of nine virulence-associated genes (VAGs) characteristic of the most important human E. coli pathotypes and four serotype-related genes (wzxO104,fliCH4, rbfO157, fliCH7) that can be used as a novel and efficient surveillance tool for detection of pathogenic E. coli strains. This tool was developed as a collaboration between P6 UCLM and P9 FLI. Using this tool, we were able to simultaneously detect all genes characteristic of the hybrid O104:H4 strain of the 2011 German outbreak (stx2/aggR/wzxO104/fliCH4) in fecal pools from one German abattoir located near the outbreak epicenter (Cabal et al. 2015). Concomitant detection by this direct approach of VAGs from a novel human pathogenic E. coli strain in cattle samples implies that the E. coli gene pool in these animals can be implicated in de novo formation of such highly-virulent strains.
Another specific objective was to study the prevalence of virulence factors of human diarrheagenic Escherichia coli in healthy animal and environmental reservoirs, including E. coli of the O104:H4 serotype. We assessed the opportunities for E. coli carrying both the aggR and stx genes to emerge by studying fecal samples from 204 'backyard' chicken farms, 204 farmers and 306 age- and gender-matched individuals not exposed to chicken farming in southern Vietnam. Only 0.5 % of chicken farm samples contained the stx or the aggR gene, whilst 2.8% of human faecal samples contained stx genes and 6.8 % contained the aggR gene. E. coli isolates that simultaneously contained the four genetic markers associated with E. coli O104:H4 epidemic strain were not found. In southern Vietnam, humans are a more likely reservoir of aggR and stx gene carrying E. coli than chicken but opportunities for transmission of isolates and/or genes between human and animal reservoirs resulting in the emergence of highly virulent E. coli strains still exist, given the poor hygiene conditions in 'backyard' farms in Vietnam (BMC Microbiol. 2016 Sep 9;16:208.)
We also studied the acquisition of diarrheagenic E. coli in 479 Dutch international travelers. Significant differences in acquisition of the aggR and eae genes were found between geographical regions with the highest acquisition observed in Africa. Stx genes were acquired by 6 travellers of whom three acquired the combination of aggR and stx1 genes. This particular combination of genes however was not detected in single isolates (manuscript in preparation).

The overall objective of work packages 5 to 9 was to identify key factors that allow pathogens to cross the intrahuman barriers. This involved genetic and antigenic variation of pathogens, factors allowing pathogens to infect and adapt to humans, and immune factors associated with disease.

Role of genetic and antigenic variation of zoonotic pathogens at intrahuman barriers
Model pathogens studied included zoonotic influenza viruses, hantaviruses, and MERS-CoV.

Adaptation of zoonotic influenza viruses and MERS-CoV to replication in humans
The main objectives were:
• To identify the key processes of adaptation of zoonotic influenza viruses to replication in humans. Specific objectives were:
o to determine the zoonotic potential of avian influenza viruses of the subtypes H7N9, H5N8, and H10N7, which emerged during the course of the Antigone project.
o to characterize the mechanism by which human and avian influenza viruses damaged the pulmonary alveolar wall, resulting in acute respiratory distress syndrome.
o to understand how human and avian influenza viruses could use the olfactory nerve to spread directly from the upper respiratory tract to the brain.
o to elucidate the damage that human and avian influenza viruses caused to extra-respiratory tissues.
• To identify and characterize the cause of Middle East Respiratory Syndrome (MERS), a disease that emerged during the course of the Antigone project. Specific objectives were:
o identify the viral agent causing MERS
o identify the receptor for MERS-CoV in the human respiratory tract
o develop a laboratory animal model for MERS in humans
• To improve a mathematical model that combines elements of within-host pathogenesis and between-host transmission for influenza in humans. Specific objectives were:
o To identify conditions that allow within-host emergence of a transmissible mutant influenza virus after evolution in a person.
o To compare the risk of emergence of such a virus between immunocompetent and immunocompromised hosts.

The main results were as follows:
• H7N9 virus attached moderately or abundantly to both upper and lower human respiratory tract, suggesting that it has the potential both to transmit efficiently among humans and to cause severe pneumonia.
• The pattern of attachment of H7N9 virus in the respiratory tracts of macaques, mice, and to a lesser extent pigs and guinea pigs resembled that in humans more closely than the attachment pattern in ferrets. This information contributes to our knowledge of the different animal models for influenza.
• The 2014 H5N8 virus replicated poorly in ferrets, had low virulence and lacked the ability to transmit via the airborne route in the ferret model, suggesting that the public health threat of the HPAI H5N8 strains is low.
• Large-scale geographical spread of H5N8 virus in 2014-2015 most probably occurred via long-distance flights of infected migratory wild birds, first in spring 2014 from South Korea or other unsampled locations in the region to northern breeding grounds and then in autumn 2014 from these breeding grounds along migration routes to wintering sites in North America and Europe.
• H10N7 virus infection in ferrets caused moderate inflammation in the lower respiratory tract, indicating that humans may potentially be infected by this virus and develop disease.
• An integrated literature review showed that the pathogenesis of influenza-virus-induced acute respiratory distress syndrome is centred on the alveolar epithelium.
• By use of and in vitro co-culture model, we showed that IAV damages the pulmonary epithelial-endothelial barrier by disruption of tight junctions amongst epithelial cells, and specifically due to loss of tight junction protein claudin-4.
• Influenza virus and other respiratory viruses can use the olfactory nerve to spread directly from the nasal cavity to the central nervous system (CNS), where they may cause clinical disease ranging from mild nervous dysfunction to severe meningoencephalitis and neurodegenerative disease.
• We showed the first case of influenza virus entry into the CNS via the olfactory route in an immune-compromised infant.
• Studies in our ferret model showed that vaccination is more effective than prophylactic oseltamivir in preventing CNS invasion by H5N1 virus via the olfactory nerve.
• In ferrets, both H5N1 and H1N1 virus infections increased the expression of pro-inflammatory cytokines in the respiratory tract and the central nervous system. These data demonstrate that extra-respiratory tissues contribute to the systemic pro-inflammatory cytokine response during influenza virus infections and that this is particularly pronounced during H5N1 virus infections.
• A novel coronavirus, later classified as MERS-CoV, was identified as the cause of MERS in humans.
• Dipeptidyl peptidase 4 (DPP4; also known as CD26) was identified as a functional receptor for MERS-CoV in the human respiratory tract.
• A rabbit model for MERS in humans was developed; this model may be important to test intervention strategies to block MERS-CoV replication in vivo.
• Mathematical modelling showed that evolution of a transmissible mutant of a zoonotic influenza virus, resulting in within-host replication to titres allowing transmission, could occur only rarely in immunocompetent individuals. In contrast, immunocompromised individuals could represent the elusive index cases for the generation of a transmissible variant of zoonotic influenza virus in the human population.

Genetic diversity of highly pathogenic avian influenza A/H5N1 during human infection and identification of novel human adaptive pathways
Our objective was to study the genetic diversity of highly pathogenic avian influenza A/H5N1 during human infection and identification of novel human adaptive pathways. We aimed to identify evolutionary pathways involved in adaptation of avian influenza viruses to humans. We have used next generation whole genome quasi species sequence analyses from clinical specimens obtained H5N1-infected patients, followed by phenotypic characterization of mutations potentially relevant for human adaptation using recombinant viral proteins and viruses. We emphasize that we investigated naturally occurring virus variants found in clinical specimens during actual human infection, as opposed to experimentally selecting or creating pathogenic or transmissible viruses in so-called gain-of-function experiments (manuscript submitted for publication).

Identification of the receptor for MERS-CoV
The above mention cell culture collection enabled us to perform in vitro experiments that provided essential information on the putative host range of the newly emerged MERS-Coronavirus. Already 2 months after the first human case was announced on PROMED, we showed that MERS-Coronavirus did not use the same entry receptor as SARS-Coronavirus (Müller et al. 2012, mBio). Instead, and contrary to SARS-Coronavirus, MERS-Coronavirus had a broad replicative capacity in a wide range of mammalian cell cultures including bats and livestock (Müller et al. 2012, mBio; Eckerle et al., 2014, EID). The rapid identification of the MERS-Coronavirus receptor, dipeptidyl peptidase 4, by Antigone partners, already 6 months after the first MERS case helped to explain the observed the severe outcome MERS and the putative broad host range. Similar to the SARS-Coronavirus receptor, angiotensin converting enzyme 2, dipeptidyl peptidase 4 is primarily expressed in the lower repiratory tract on type 2 pneumocytes and alveolar macrophages. Infections in the lower respiratory tract are commonly associated with a more severe disease outcome than upper respiratory tract infections. The broad host range may be explained by the fact that DPP4 is more conserved among different mammalian species than the SARS-Coronavirus receptor. Importantly, we were the first to identify a putative virulence marker of MERS-Coronavirus. The MERS-Coronavirus specific accessory protein 4a was shown to block the intracellular immune defense mechanisms (Niemeyer et al., 2013, Journal of Virology).

Adaptation of hantaviruses to replication in humans
Our ultimate goal is to elucidate the molecular bases behind this gain in pathogenicity. Hantaviruses are deeply adapted to their rodent hosts and provoke a Hantavirus Cardio-Pulmonary Syndrome (HCPS) in the New World and a Haemorrhagic Fever with Renal Syndrome (HFRS) in the Old World. We are comparing the Puumala virus (PUUV), responsible for HFRS in Europe with the two non-pathogenic Tula virus (TULV, Old World), and Prospect Hill virus (PHV, New World). Myodes glareolus and Microtus arvalis from the Arvicolinea sub-family are hosting PUUV and TULV, respectively.
As Hantaviruses have been poorly explored previously, many tools had to be built (some still pending) to generate the essential prerequisites claimed in the program. We evaluated the entry capacity of Hantaviruses in cell lines derived from different organs and hosts. Among primate/human cells, only Vero E6 (kidney), HuH7 (liver) and THP1 (differentiated into macrophages) were susceptible to PUUV, TULV and PHV while HepG2 (liver), A549 (lung), HEK (kidney), Caco2 (intestine) or NK and neutrophil primary human cells were poorly/not susceptible. Interestingly, kidney MyglaSWRecB epithelial cells and lung alveolar MyglaAECB6 cells from M. glareolus tissues (the PUUV reservoir) were infected by PUUV but also by PHV. The differential susceptibility to infection does not correlate with the surface expression of the classical (co)receptors (β3-, β1-integrins, DAF, gC1qR) previously proposed to differentiate between pathogenic and non pathogenic hantaviruses.
We observed clear morphological differences in the way viruses replicate in the different cell types and the Y-2H approach was used to search for host factors involved in replication or assembly that may differentiate human from rodent hosts. Screening a cDNA library from Mus Musculus and the Human orfeome we did not find any cellular partner for the N and NS proteins of TULV, PUUV and Hantaan virus (HNTV). In contrast, the cytosolic tail of the Gn glycoprotein (GnCT) of TULV and PUUV interacted specifically with the rodent pericentrin (PCNT), a major scaffold protein of the microtubule organising centre. Immunofluorescence analysis of Vero E6 cells infected with PUUV showed that the Gc glycoprotein (no Ab is available for Gn) is localized in the Golgi and surrounds both PCNT and the microtubule organising centre. This observation is important since hantaviruses, like other members of the Bunyaviridae family, do not encode a matrix protein and their GnCT is thought to play a key role in particle assembly. To further understand the function of the GnCT-PCNT complex in the viral cycle, the respective domains responsible for the interaction are dissected, specific Gn antibodies are prepared and cells stably deleted of PCNT expression are generated through shRNAs transduced by lentivrus vectors.
In parallel, we used proteomic antibody arrays to evaluate the impact of PUUV, TULV and PHV infection on the regulation of cellular signalling pathways. Highly susceptible cells (i.e. Vero E6 and HuH7) at day 7 post-infection showed a differential activation or inhibition of factors involved in angiogenesis, pro-inflammation and recruitment of cells of the innate immune system such as macrophages or PMN. Theses results correlates with the observation that pro-inflammatory factors are elevated in the serum of patient with HFRS.

Potential for zoonotic transmission of African henipavirus and Nipah virus
The second part of our work within the consortium concerned the molecular characterisation and evaluation of the potential for zoonotic transmission for a recently identified African relative of the Henipaviruses. The genus Henipavirus comprises two known virus species causing fatal encephalitis in humans. These viruses, Hendra virus (HeV) in Australia and Nipah virus (NiV) in Asia, have been sporadi¬cally acquired from bats of the Pteropus genus by humans, swine and horses. Although Henipaviruses are traditionally restricted to flying foxes of the genus Pteropus of South Asia, a paper published by Partner 4 (UBMC) in 2012 had identified distinct viral clades in phylogenetic relation to Henipaviruses in six bat species sampled in five African countries (Drexler et al., Nat. Comm. 2012). Full genome sequencing of a representative virus (GH-M74a) confirmed formal classification in the genus Henipavi¬rus (GenBank: HQ660129.1). It was therefore of vital interest to continue research into the detection of related viruses in Africa and to determine if these viruses present cause for public health concerns and potential for zoonotic transmission. As the direct isolation and amplification of replicating virus from bat samples has to date proven unsuccessful for the African strains, we aimed to amplify and to study viral gene products in order to study the molecular characteristics of this newly identified Henipavirus, using this complete genomic sequence as a starting point. With this aim, experiments were undertaken to study the viral surface F and G entry proteins in terms of the phenomena of viral entry into a host cell as well as receptor usage for these African Henipaviruses. Expression plasmids were produced and tested for the viral surface proteins G and F that when expressed together in cell lines expressing the appropriate viral receptor are able to induce cell-to-cell fusion and promote the formation of large multi-nucleated cells known as syncytia. In these experiments we were able to show, for example, that co-expression of GH-M74a G and F proteins in various mammalian or simian cell lines fails to result in syncytia formation whereas expression of the same proteins in cell lines of bat origin results in active cell fusion (Lawrence et al., 2014). Overall, similar data from our group and others would appear to suggest that differences in expression and cleavage of the surface F protein for the GH-M74a lineage in certain cell types for certain species vs. bat cell lines may represent a species barrier to spillover of these particular viruses into a new host. At least one additional publication in this area in collaboration with Partner 4 (UBMC) is expected to be submitted shortly. This new paper will feature an additional >20 African henipavirus-related F gene sequences mentioned above, as well as our continuing work at addressing the functional differences observed for these surface proteins. Our current data suggest an impaired trafficking and/or maturation of the surface protein in certain simian or human cell lines.
Another major contribution concerns our understanding of the molecular mechanisms involved in NiV replication. More specifically, our ongoing work with an external partner concerns the molecular structure/function relationship between NiV replicative proteins and is based on crystal structure data and molecular reverse genetic modelling of the interactions involved in viral replication. A publication detailing our joint work in this area and providing a structure/function model was published in Nature Structural and Molecular Biology (Yabukarski et al., 2014). In this publication we identified the molecular interaction between the viral N and P proteins to be a potent target for antiviral therapy and showed that a peptide mimicking part of the P protein can effectively block viral replication in vitro.

Adaptation of Mycobacterium bovis to replication in humans
One goal was identifying the pathogen- and host-specific factors correlated with differences in pathogenicity and transmissibility between different M. bovis strains. Significant progress in this aspect was derived from studies on the wild boar model, a good proxy for human tuberculosis since the clinical picture includes frequent lung involvement. Comparative genomics analyses have provided new insights into the evolution and adaptation of the MTBC to survive inside the host and facilitate host to host spreading. We sequenced the genome of three M. bovis (MB1, MB3, MB4) and one M. caprae (MB2) field isolates with different lesion score, prevalence and host distribution phenotypes. Genome sequence information was used for whole-genome and protein-targeted comparative genomics analysis with the aim of finding correlates with phenotypic variation with potential implications for tuberculosis risk assessment and control. The comparison of the MB1 and MB4 isolates revealed differences in genome sequence, including gene families that are important for bacterial infection and transmission, thus highlighting differences with functional implications between isolates otherwise classified with the same spoligotype. Strategic protein-targeted analysis using the ESX or type VII secretion system, proteins linking stress response with lipid metabolism, host T cell epitopes of mycobacteria, antigens and peptidoglycan assembly protein identified new genetic markers and candidate vaccine antigens that warrant further study to develop tools to evaluate risks for TB disease caused by M. bovis and for TB control in humans and animals (de la Fuente et al. 2015).

Human immune response to Duvenhage virus
Our main objective was to determine the differences in immune response explaining the different pathogenetic mechanisms of Duvenhage virus in the human host. In order to achieve our objective we first set up an animal model of DUVV infection. The optimal route of inoculation (subcutaneous; s.c. or intramuscular; i.m.) as well as age-dependent susceptibility was determined in BALB/c mice. We found that DUVV-NL07 was neuroinvasive for both 3-week and 8-week old mice inoculated with 106 TCID50 or 104 TCID50 either i.m or s.c. The earliest time point on which virus was detected in the brain was day 7 post inoculation and clinical signs (including ruffled hair, hunched position and muscle weakness) were apparent from 10 days post inoculation onward. There was no strong association between development of paralysis and route of inoculation, age or level of viral RNA in the brain. Virus was not detected in the brain samples of animals inoculated with 102 TCID50. The highest dose administered (106 TCID50/mouse) resulted in 50% mortality.
Intracranial inoculations with DUVV-NL07 of inbred (BALB/c and C57/Bl6) and outbred (Swiss albino) mouse strains were also performed to study the transcriptomic profile of DUVV-NL07 and identify molecular pathways involved in pathogenesis of disease. Differentially expressed mRNA transcripts were identified using LIMMA version 2.12.0. Dysregulated proteins were identified with MS. Gene/protein enrichment analysis was done using different publicly available software. A number of differentially expressed genes/proteins after DUVV-NL07 infection compared to mock-inoculated control mice were identified. The majority of these genes are associated with innate immunity and cell death mechanisms. Of interest is the activation of caspase 1 (CASP-1) and the downstream upregulation of pro-inflammatory signals that may lead to cell death. Validation of these findings with qRT-PCR and/or IHC was performed in samples from different mouse strains.

Factors in Hantaviruses that mediate evasion of human immune responses
In absence of a cytopathic effect in infected cells, the role of pro-inflammatory response in the physiopathology of the HFRS induced by PUUV is presumed, leading to an alteration of the endothelial cell barrier functions and subsequent capillary leakage. Among innate immune cells, neutrophils (PMN) interact with the vascular endothelium and increased permeability through adhesion processes and chemokine secretion (CXCL1, 2, 3 and 8).
We studied the impact of PUUV, TULV and PHV infection on PMN and NK cells isolated from blood of healthy donors. If both cell types were poorly infected, we observed an increase of survival of PMN infected with the pathogenic PUUV while no effect occurred in presence of the non pathogenic TULV or PHV. This PUUV induced survival correlates with a delayed apoptosis of the PMNs. This delay is independent of caspases related to the intrinsic apoptosis pathway but rather involves death receptors of the extrinsic pathway, namely TRAILR1. Altogether, the cumulated results of WP7 and WP9 are in favour of the role of PMN in hantavirus pathogenesis through the secretion of inflammatory mediators provoking endothelium damage.

Pathogenicity mechanisms of Ebola virus
Two viral pathogens that are able to sporadically spillover into the human population and that would appear to intrinsically be able to replicate in humans resulting in human-to-human transmission are the Filoviruses and the Henipaviruses of the Mononegaviridae order. In this respect, the aims of our research team were to study, for these two viral families, the molecular characteristics that allow host adaptation and immune evasion leading to cross species transmission. Indeed both pathogens are considered as serious threats to global health and belong to a short list of viruses that can be seen as “loaded guns”, intrinsically primed to cause human pandemics, and as such are classified as biosafety level 4 (BSL-4) pathogens. How emerging viruses cause such severe and lethal diseases is still poorly understood, and both of these viruses are well-suited model systems for studying, at the level of molecular interactions, how zoonotic infection results in human pathology. The investigation of the mechanisms and strategy of expression of viral proteins, innate immune evasion and pathogenicity provides valuable information about the evolution and adaptation of these viruses to a new host environment, as well as concerning the potential of zoonotic viruses in general to cross the species barrier into the human population.
Our aims concerned the filoviruses and in particular, in understanding the mechanisms of pathogenicity that make certain members of this particular family of viruses so deadly. For example, we had previously demonstrated that these viruses are easily adaptable to new animal hosts in laboratory conditions, e.g. adaptation of these viruses to rodents requires only a few consecutive passages. Analysis of the viral genomes of guinea-pig-adapted viruses revealed that changes in VP24 of Ebola virus (EBOV) are both necessary and sufficient to adapt EBOV for replication. It appeared however that the adaptation was not related to the IFN signalling antagonist function of VP24. This suggests the existence of other important host specific barriers that are overcome by mutations in VP24. In this respect, we and others have recently shown that both EBOV and the related Marburg virus (MARV) are able to modulate a cellular mechanism directed at maintaining cellular redox balance, a pathway that was not previously thought to be specifically targeted during viral infection. Thus, we have shown that synthesis of VP24 affects the expression of a panel of cytoprotective genes regulated by the nuclear transcription factor Nrf2. Our studies demonstrate that the VP24 protein from MARV is able to activate NRF2 transcriptional activity by binding negative regulator of NRF2 (Keap1) and thus prevent degradation of NRF2 (Page et al., 2014). As a consequence, NRF2 increases cellular resistance to oxidative and toxic damage. Such an effect may cause a short-term increase in the survival of MARV infected cells and help to promote replication of this virus. On the contrary EBOV VP24 dramatically downregulates the ability of cells responding to oxidative damage induced by the massive inflammatory reactions followed the course of infection (Nemirov et al., in preparation). Such data again indicates that changes induced by this pathogen to cellular oxidative stress pathways are involved in the pathogenesis seen with EVD.
One of our major contributions to our understanding of EBOV pathogenicity mechanisms globally concerns the key roles played by the viral glycoprotein in host pathogenesis mechanisms. Our work with this virus clearly suggests that the phenomena of GP editing and release from infected cells as a soluble form via shedding are important pathogenicity mechanisms that not only allow the virus to replicate successfully in a broad range of host cell types and organisms but that are also very likely mechanisms that are highly involved in adaptation of the virus to a new host species. Indeed, it is becoming increasingly apparent that the unique expression strategy used by ebolavirus to produce its different GP gene products is at the heart of successful virus replication in host cells. Three papers detailing the mechanisms involved in the cellular expression of GP via RNA editing phenomena and the effects of modification of this strategy and in modifying the ratios of surface glycoprotein expression vs. its secreted and shedding forms have been recently published by our group in the Journal of Infectious Diseases. Importantly, these works show that even slight modification of editing or of the expression of GP products can dramatically alter virus cytotoxicity in cellula and virulence in animal model of infection (Dolnik et al., 2015; Volchkova et al., 2015a; Volchkova et al., 2015b). It is also clear that high pathogenicity is linked to high virus replication rates and an inadequate or inappropriate host response to massive virus replication, and in this respect one of our salient contributions to this field was the demonstration that the soluble shed form of the viral glycoprotein, termed shed GP, could play a particular role during EBOV infection. In effect we have shown that this protein is able to bind and activate non-infected dendritic cells and macrophages, inducing the secretion of pro- and anti-inflammatory cytokines and that this activation is likely mediated through cellular toll-like receptor 4 (TLR4) (Escudero-Pérez et al., 2014). Furthermore, we have also highlighted the ability of shed GP to affect endothelial cell function both directly and indirectly, demonstrating the interplay between shed GP, systemic cytokine release and increased vascular permeability (Escudero-Pérez et al., 2014). These activities of shed GP could be at the heart of the excessive and dysregulated inflammatory host reactions to infection but may also contribute to the disseminated intravascular coagulation syndrome that disrupts blood flow in the disease victim leading to organ failure and ultimately death in cases of EBOV disease. The continuing study of the mechanisms involved for this action and the role that shed GP may play in activating other cellular targets and/or in the phenomena of coagulation impairment is ongoing. In this respect, materials and fixed tissue samples have been sent to Partner 1 for immunohistological and pathological analysis.

Development of an efficient vaccine therapy against Ebolavirus infection
The unprecedented scale of the 2014-15 ebolavirus outbreak in West Africa led us to instigate a collaboration with an external partner addressing this development of efficient vaccine therapy against ebolavirus infection. In this study, hyper immune anti-EboV GP polyclonal IgG were obtained from pigs that are knock-out for two major xenoantigens in humans. Following purification from serum, hyper-immune polyclonal IgGs were obtained, exhibiting virus neutralizing ability. In Guinea pig experiments, compared to control animals treated with IgGs from non-immunized double KO pigs, the anti-EBOV IgG-treated animals exhibited a significantly prolonged survival and a decreased virus load in blood on day 3. The data obtained indicated that these IgGs, lacking epitopes that are highly immunogenic in humans, have a protective effect upon EBOV infection (Reynard et al., 2016). In a similar vein, through funding provided by the consortium, we were able to continue work to characterise a collection of previous produced monoclonal antibodies against the viral glycoprotein and to test the neutralisation capacity of these molecules. The characterisation of the mode of action of these antibodies also allowed us to further dissect some of the intricacies of the entry process and has offered detail with which to development new treatment strategies. Two papers describing this work were published in the Journal of Infectious Diseases (Reynard & Volchkov, 2016a; Reynard & Volchkov, 2016b).


Pathogenicity factors of Escherichia coli O104:H4
Escherichia coli O104:H4, that caused the worldwide largest outbreak of hemolytic uremic syndrome in 2011, is a highly virulent hybrid of enterohemorrhagic (EHEC) and enteroaggregative (EAEC) E. coli [herein after referred to as EHEC O104:H4]. EHEC, a subpopulation of Shiga toxin (Stx)-producing E. coli (STEC), are zoonotic pathogens, basically adapted to cattle, whereas the highly human adapted EAEC are rarely detected in animal hosts. The emergence of novel pathogens and potentially asymptomatic carriers pose a major public health threat for widespread epidemics in susceptible populations.
Partner 9 FLI determined pattern and magnitude of fecal shedding and site of colonization in a bovine infection model to assess if EHEC O104:H4 can utilize ruminants as reservoir. In two independent experimental infection trials (I - short duration (till 4 days), II - long duration (till 28 dpi)), calves (in each trial five calves/strain) were inoculated with 1010 CFU of EHEC O104:H4, of EHEC O157:H7 (positive control) or of apathogenic E. coli O43:H28 (negative control) and necropsied. As deduced from the fecal excretion pattern, EHEC O104:H4 took an intermediate position compared to EHEC O157:H7 and the apathogenic O43:H28 being detectable day 24 p.i. compared to day 28 p.i. and day 12 p.i., respectively. Analyses of sequential fecal samples demonstrated that increasing percentages of colonies of EHEC O104:H4 were formed by bacteria having lost the pAA plasmid, but the ESBL plasmid remained stable. On day 4 p.i., EHEC O104:H4 was re-isolated from intestinal content and mucosal tissue samples, but (in contrast to EHEC O157:H7) no detection in intestinal samples on day 28 p.i. was possible. By immunoperoxidase staining of adherent anti-O157 and anti-O104 positive bacteria in intestinal tissue samples, single EHEC O104:H4 bacteria could be identified in 1 of 5 calves at the recto-anal junction only 4 d.p.i., while micro-colonies of EHEC O157:H7 were visible in 3 of 5 calves 4 d.p.i.. No induction of a humoral (measured by serum antibodies against Stx2) immune response took place within trial II. These results are the first evidence that cattle can carry EHEC O104:H4 at least transiently, although EHEC O104:H4 appears to be less well adapted to bovines than classical EHEC strains.
An additional study aimed at appraising the relative level of adaptation of the hybrid strain EHEC O104:H4 to humans and cattle by in vitro cell culture models. Adherence of EHEC O104:H4 to intestinal (jejunal and colonic) epithelial cells (IEC) of human and bovine origin was compared to that of E. coli strains representative of different pathovars and apathogenic E. coli. Strain-specific host gene expression profiles of selected cytokines and host-induced expression of bacterial virulence genes were assessed by quantitative RT-PCR. Amounts of Stx released upon host cell contact were quantified by ELISA. The outbreak strain EHEC O104:H4 is capable of attachment to IEC of human and bovine origin. It adheres in intermediate quantities as compared to other E. coli strains, but resembled other strains by preferentially binding to less differentiated small intestinal cells. Relative amounts of Stx produced and released by classical EHEC strains correlated well with the level of adhesion but strain EHEC O104:H4 released strikingly less toxin in bovine and human cell cultures. Host cell response was strain-dependent with a strong pro-inflammatory effect, however, independent from the level of adhesion. Overall EHEC O104:H4’s reaction profile most closely resembled that of other human-adapted EAEC strains, in particular strain 55989.
Although EHEC O104:H4 is able to colonize cattle under experimental conditions, the genetic and phenotypic attributes of this strain primarily qualifies it to colonize a human reservoir. Overall, our findings from in vitro and in vivo studies form a sound basis for risk analyses anticipating the likelihood of zoonotic infections with bacteria belonging to this particular subset of the EHEC pathovar (manuscript in preparation).


The overall objective of work packages 10 and 11 was to identify key factors that allow pathogens to cross the interspecies barriers. Model pathogens included influenza virus, human metapneumovirus, MERS-CoV, Yersinia, and Escherichia coli 0104:H4.

Hamster model for transmissibility of metapneumoviruses
We have worked on two main objectives: 1) to evaluate the guinea pig model to study transmissibility of human and avian metapneumoviruses, and 2) to evaluate the potential of aerosol transmission of zoonotic influenza viruses.
To study airborne transmissibility beyond influenza viruses, we used the human metapneumovirus (HMPV) as a model pathogen to design an airborne transmission model for paramyxoviruses. Although HMPV was confirmed to replicate very well in guinea pigs, transmission of virus from inoculated to naïve animals either via contact or via aerosols was not observed. As an alternative model system, we decided to test hamsters as these animals were also found to be very permissive for HMPV. Our first experiments indicated that one strain of HMPV was indeed transmissible by aerosols (3 out of 4 animals via direct contact, and 2 out of 4 animals via the airborne route), while another strain of HMPV was not. The avian counterpart of HMPV, AMPV was not transmitted by aerosol. This means that the guinea pig model can be replaced by the hamster model.

Airborne transmission of influenza viruses in the ferret model
Genetic changes and associated phenotypic changes that were necessary for airborne transmission of zoonotic influenza A/H5N1 viruses in the ferret model were identified (Herfst, Science, 2012, Linster, Cell, 2014). When the A/H7N9 virus emerged in China in the spring of 2013, we studied the airborne transmissibility of this newly emerging virus in order to determine its impact on public health. With the publication of this research (Richard, Nature, 2013), we showed that natural A/H7N9 viruses isolated from human cases were transmissible between ferrets, although they displayed a limited airborne transmissibility. Substitutions in A/H7N9 viruses that could potentially confer transmissibility to A/H7N9, based on the critical requirements for the airborne transmission of A/H5N1 virus, were also identified (Schrauwen, JVI, 2016). Moreover, in the light of the accumulated knowledge on airborne transmissibility of A/H5N1 viruses, we worked on the characterization and risk assessment of newly emerged viruses: the Dutch A/H5N8 (Richard, PLoS One,2015), A/H10N7 viruses isolated in 2014 from seals and the zoonotic H5N6 virus from China.
In other studies, infected (Inf) co-housed ferrets from WP7/D7.2/M24/m12 were then housed in a separate room inside an airborne only cage system with naïve (N) ferrets on one side or ferrets prior exposed (Vacc) to human seasonal influenza immunisation (including seasonal human H3N2) on the other side using a (2N+2Inf+2vacc) x 4 model. The ferret cage design was modified following consultation with P1 EMC
Overall, the studies that were conducted under Antigone allowed us to gain fundamental knowledge on specific phenotypic properties required for airborne transmission of zoonotic influenza viruses, especially A/H5N1 viruses. The extrapolation of this knowledge to other zoonotic viruses, such as A/H7N9 and A/H5N6 has helped us to perform targeted risk assessment and improved pandemic preparedness.

Sustained human-to-human transmission of MERS-CoV
The currently largest case-contact study in a household setting revealed that secondary infections are rather uncommon and that the R0-value may be below 0.5 which means that sustained human-to-human transmission chains are not very likely (Drosten et al., 2014, New England Journal of Medicine). However, in an intensified outbreak investigation in Jeddah (Drosten et al., 2015, Clinical Infectious Diseases) UBMC found out that nosocomial transmission chains may, in fact, be the main reason for locally limited MERS outbreaks.

Ability of EBOV to transmit among humans
Our research revealed that specific amino acid substitutions in the EBOV GP increased tropism for human cells, while reducing tropism for bat cells. The observed increased infectivity may have enhanced the ability of EBOV to transmit among humans and contributed to the wide geographic distribution of some viral lineages (Urbanowicz et al., 2016, Cell).

Transmissibility of Yersinia under harsh host tissue conditions
The aim of this study was to identify factors that enable pathogenic and potentially pathogenic Yersinia to highly replicate under harsh host tissue specific conditions. We chose hypoxia i.e. the condition when oxygen demand exceeds oxygen supply as a harsh environmental condition. Cellular responses to hypoxia are central to the pathophysiology of inflammation and major diseases. As a consequence to hypoxia the microenvironment of the hypoxic tissue is altered by reprogramming metabolic pathways.
The human pathogens Yersinia (Y.) pseudotuberculosis and Y. enterocolitica cause various gut-associated syndromes, also known as yersiniosis, whereas Y. pestis is the responsible agent for pneumonic, bubonic and septicaemic plague. Common to all three agents is their infection strategy based on a plethora of virulence factors enabeling them to enter, attach, invade and ultemately colonise the host. To date the influence of an hypoxic environment as found e.g. in the mammalian gastrointestinal tract or the bubo on the internalisation of Gram negative bacteria such as members of the Yersinia pseudotuberculosis group (i.e. Y. similis, Y. pseudotuberculosis and Y. pestis) is still not entirely clear. In order to develop more effective therapeutic strategies for the treatment of Yersinia infections and their sequelae the complex interrelationships between the host and the agent and the surrounding microenvironment have to be determined.
Two cell lines were selected for infection assays. Both cell lines were infected with a pathogenic Y. pseudotuberculosis strain, and a putative non pathogenic Y. similis strain under normoxic or hypoxic conditions, respectively. Infection rates differed between the different bacterial species and environmental conditions. Both cell lines proved to be suitable surrogates. For Y. pseudotuberculosis and Y. similis infections stable and reproducible hypoxic conditions could be maintained. Infection experiments with an attenuated Y. pestis are currently ongoing.
In order to complement the work intensive plating efficiency tests a flow cytometry based method was successfully developed to assess invasion frequency. In order to increase invasion frequency of apathogenic yersiniae a transfection agent based method was successfully developed. The effective infection has been additionally confirmed by MALDI TOF analysis.
The influence of the hypoxic microenvironment on the transcriptome of different Yersinia species has been analyzed by RNA-Seq and the sum of differentially expressed genes calculated. Because the transcriptome includes all mRNA transcripts in the cell, the transcriptome reflects the genes that are being actively expressed at any given time in the bacteria and the cells. The affected pathways in the host and in the bacteria are currently being analyzed in silico and evaluated for their application as putative therapeutic targets.

Transmissibility of Escherichia coli 0104:H4
The primary goal of our research activities was to identify key factors of the EHEC O104:H4 pathogenicity that render this pathogen prone to cross the species barrier and gain efficient transmissibility among humans. Furthermore, we aimed at the generation of data that promote the establishment of preventive, therapeutic and diagnostic measures to counteract EHEC infections and outbreaks.
So far, it is not recommended to treat EHEC infections with antibiotics due to the induction of the Shiga toxin-encoding bacteriophage and thus increased expression of Shiga toxin. To identify antimicrobials that are suitable for treatment of patients infected with EHEC O104:H4 and of shedders, antimicrobials to which the outbreak strain is susceptible, were tested for their ability to induce stx2-phages and to increase Shiga toxin production in vitro. This will be the initial step in the investigation of the effects of antibiotics on the outbreak strain regarding their potential benefit or harm to patients. We demonstrated that several antibiotics decrease or do not influence baseline levels of stx2-harboring phage induction, stx2 transcription, and Stx2 production in the EHEC O104:H4 outbreak strain and might therefore be potentially therapeutically useful in indicated cases.
The selective induction of the loss of the Stx-encoding bacteriophage results in a significant attenuation of EHEC strains and would represent an important step towards improved treatment options for EHEC infection by losing the major EHEC virulence factor. We therefore aimed at the identification of growth conditions, under which the lysogenic stx2-bacteriophage could be induced in the majority of bacterial cells of the culture. We have identified conditions (incl. low pH, mitomycin C treatment, growth in the presence of subinhibitory concentrations of chloramphenicol and ciprofloxacin) which promote loss of stx-harbouring bacteriophages in different highly virulent EHEC clones (O26 or O157). Growth under these conditions led to the loss of the stx-phage in up to 90 % of the bacterial population. In contrast, these conditions did not result in even a single colony of stx2-negative O104:H4.
The O104:H4 outbreak strain produces extended spectrum beta-lactamases and is resistant to several antibiotics. Lytic bacteriophages that specifically target EHEC strains may thus be an alternative treatment option of EHEC infections. We isolated lytic bacteriophages that are able to infect E. coli O104:H4 from water samples. These bacteriophages were not specific for E. coli O104:H4, but were also able to infect and lyse other E. coli variants. Nevertheless, these phages may be useful for the eradication of EHEC O104:H4 from, e.g. food samples.
We focused on the role of genetic processes of evolutionary change in EHEC O104:H4 on their ability to adapt to human hosts. We studied genomic differences that might have led to the increased virulence of EHEC O104:H4 and enabled crossing of the intrahuman barrier, thereby resulting in high numbers of patients with severe disease. We also investigated the microevolution of E. coli O104:H4 during the outbreak and within patients. Specifically, using next generation sequencing (NGS) technology we determined in 49 German EHEC O104:H4 isolates genomic variations. Analyzed strains included isolates representing different host pathogen interaction scenarios (single isolates per patient, follow up isolates of the same patient, household isolates of familiar communities). Overall, genome sequencing resulted in 28 single nucleotide polymorphisms (SNPs), the majority (21) were nonsynonymous (ns)SNPs, however no significant positive selection was detected. In total, the SNPs differentiated 15 genotypes with varying genotypic changes within follow up isolates derived from the same patient demonstrating the impact of host-pathogen interaction. This was accompanied by shifts over time regarding the genotypes due to potential microevolutionary events during the outbreak selecting different adaptive genotypes of different isolates.
Comparative genomic analysis of the E. coli O104:H4 outbreak strain provided insights into specific genomic features that provided valuable information for the development of rapid molecular tests for the differentiation of EHEC. We developed improved detection and typing approaches for all clinically relevant EHEC variants. The real-time PCR protocols developed represent a rapid and reliable method for HUSEC identification and Stx subtyping in diagnostic laboratories.
We demonstrated that OMVs represent a powerful tool of the E. coli O104:H4 outbreak strain for the delivery of its virulence factors into intestinal epithelial cells and causing cell injury and inflammatory responses. This finding has also implications for the use of OMVs as potential vaccine candidates to prevent disease caused by E. coli O104:H4.
We studied horizontal gene transfer events that may have resulted in the highly virulent 2011 EHEC O104:H4 isolate and compared host range and transfer frequencies of the ESBL plasmid pO104_90. pO104_90 was readily transferable by conjugation under standard laboratory conditions, including from and into other clinically relevant Gram-negative pathogens. Co-transfer of chromosomal markers by chromosomally integrated pO104_90 0 could not be observed.
In contrast to the closely related enteroaggregative E. coli O104:H4 strain 55989 or the historic O104:H4 outbreak strain HUSEC041 which express AAF/III fimbriae, the EHEC O104:H4 outbreak strain from 2011 expresses the AAF/I allelic variant of this fimbrial type. To study putative differential functions of the five known AAF fimbrial variants, we systematically compared AAF-mediated phenotypes, e.g. autoaggregation, biofilm formation, and adherence to epithelial cells. Our analyses revealed both, phenotypic differences as well as similarities, among the five AAF subtypes. We showed that there are no major differences in phenotype for AAF/I, AAF/III and AAF/V with regard to autoaggregation and adherence to epithelial cells, but variations in biofilm production could be detected. We did not observe a superior survival capacity of EHEC O104:H4 relative to closely related E. coli strains on dry fenugreek seeds. Also the ability to express AAF/III fimbriae did not affect the persistence of EHEC O104:H4 on dry fenugreek seeds. For none of the strains tested cultivatable bacterial cells were found without enrichment on contaminated seeds after more than 24 weeks of storage. Our findings suggest that contamination of fenugreek seeds with E. coli O104:H4 previous to the distribution from the importer may be less likely than previously assumed.


The overall objective of work packages 12 and 13 were to translate our increased understanding of key factors in the chain of emergence to risk assessment, prevention and intervention, including using the data from primary studies to parametrize epidemiological and ecological risk assessment models and dynamics models.

Results in this area include:

1) A review paper discussing the potential for introduction of bat borne viruses into the EU, highlighting the most likely routes of entry for Nipah Virus (NiV) and Marburg Virus (MARV).
2) Development of a generic quantitative risk assessment model to assess the relative risk of introduction of bat-borne viruses to the EU. The deterministic framework combines multiple routes by which bat-borne viruses are likely to be imported into the EU: human travel, legal trade (e.g. foodstuffs and products of animal origin), live animal movements and illegal importation of bushmeat and highlights the relative risk of introduction of these viruses into the EU, which Member States are of highest relative risk and from which route.
3) A paper describing the generic model and demonstrating application using NiV as a case study was produced. An additional paper is in preparation to compare results across a number of viruses: NiV, MARV, Ebola virus, Hendra Virus and Middle East Respiratory Syndrome Coronavirus.
4) A quantitative assessment for risk of bushmeat as a vector for introduction of Ebola virus into the United Kingdom, from the Democratic Republic of Congo. Expanding on the bushmeat route considered in the generic model framework, this assessment focusses on the risk pathway associated with this route and further explores the potential of virus introduction via illegal bushmeat entering the UK. The model is stochastic in nature, predicting the number of EBOV-positive bushmeat portions entering the UK per year, based on published reports on sample surveys of bushmeat entering the EU.
5) A review paper discussing the importance of the infectious dose of Ebola virus in bushmeat within risk assessment, when the transmission history of the virus is unknown. The infectivity of EBOV to humans may depend not only on the identity of the donor species itself but also on the heritage of the virus i.e. the wildlife species-passage history reflecting differences in the number of passages within a given species, and hence the degree of adaptation of the EBOV. The use of the infectious dose within a risk assessment framework would entail an unknown element of uncertainty.

Epidemiological and ecological risk assessment and dynamic models of the risk of vector-borne pathogens into human populations.
The objective was to provide vector competence data to inform epidemiological and ecological risk assessment and dynamic models of the risk of vector-borne pathogens into human populations. Our studies applied the methodology of ecological networks to quantify >14,000 interactions among ticks, vertebrates, and pathogens in the western Palearctic. The results showed that natural foci of ticks, pathogens, and vertebrate reservoirs display complex relationships that are key to the circulation of pathogens and infection dynamics through the landscape. These natural networks are highly structured, modular, coherent, and nested to some degree. These results indicate that ticks and vertebrates interact along the shared environmental gradient, while pathogens are linked to groups of phylogenetically close reservoirs. As a result of these studies, interactions between tick species, their hosts, the pathogens they carry and transmit, and the geographic distribution of species in the Western Palearctic were determined based on evidence published during 1970–2014. These relationships were linked to remotely sensed features of temperature and vegetation and used to extract the network of interactions among the organisms. The resulting datasets focused on niche overlap among ticks and hosts, species interactions, and the fraction of the environmental niche in which tick-borne pathogens may circulate as a result of interactions and overlapping environmental traits. The resulting datasets provide a valuable resource for researchers and agencies interested in tick-borne pathogens, as they conciliate the abiotic and biotic sides of their niche, allowing exploration of the importance of each host species acting as a vertebrate reservoir in the circulation of tick-transmitted pathogens in the environment.

Potential Impact:
Antigone has had impact in different areas, including increased knowledge, often leading to better surveillance, prevention and/or intervention; archives of pathogens, cells, tissues, and reagents that facilitate research on emerging infectious diseases; new tests to improve diagnostic and surveillance; better tissue and animal models for infectious diseases; more interdisciplinary research; and education of a new generation of scientists with a One Health approach.

Increased knowledge, with translation potential

The rapid and ground-breaking studies on the ecology, epidemiology and biology of MERS-Coronavirus shortly after emergence have provided invaluable and important information to the public, clinicians, researchers and national and international public health institutions (ECDC, WHO).
The detection of DPP4 as receptor for MERS-CoV may provide therapeutic opportunities to combat hCoV-EMC infection, through manipulation of DPP4 levels or development of inhibitors that target the binding interface between the S1 domain and receptor in vivo. In addition, future studies should address the development of effective vaccines, including those that elicit antibodies that prevent the binding of hCoV-EMC to DPP4.
Although camels are suspected to be the primary source of MERS-CoV infection for humans, exact routes of transmission have yet to be determined. By doing so, preventive measures can be taken to reduce the zoonotic transmission of MERS-CoV. Further serosurveillance studies in livestock should continue to monitor the spread of MERS-CoV in the Middle East, Africa and Asia.

One of our major contributions in terms of scientific knowledge is to our understanding of EBOV pathogenicity and globally concerns the key roles played by the viral GP glycoprotein in host pathogenesis mechanisms. Our work with this virus clearly suggests that viral strategies used for the expression of this viral glycoprotein on the surface of infected cells and the release of a related, soluble form of GP via shedding, are important pathogenicity mechanisms that not only allow the virus to replicate successfully in a broad range of host cell types and organisms but are also very likely mechanisms that are highly involved in adaptation of the virus to a new host species infection (Escudero-Pérez et al., 2014; Dolnik et al., 2015; Volchkova et al., 2015a; Volchkova et al., 2015b) . Understanding these mechanisms will clearly help us to understand the high pathogenicity associated with this virus and in the future potentially in managing its disease. In terms of the molecular strategies exploited by the virus to regulate surface GP expression levels, our discovery of the existence of a non-stop mRNA mechanism (Volchkova et al., 2016a) clearly has implications for filovirus research and virology but also from a wider perspective in terms of mammalian cell biology and research into the cellular mechanisms involved. Likewise, our studies demonstrating that the VP24 protein from Marburg virus is able to activate NRF2 transcriptional activity and increase cellular resistance to oxidative and toxic damage and the discovery that this pathway is specifically targeted during viral infection (Page et al., 2014), opens new avenues for research in terms of viral replication strategies in general and for therapeutic discovery.
Our ongoing, collaborative research into the finding and characterisation of filoviruses in West Africa (Ghana) has potentially important impacts in terms of our understanding of the ecology of these viruses and their host species and in terms of our ability to predict and even to prevent spillover events, public health threats and the drivers of such threats. Ultimately, the detection and molecular characterisation of additional non-, weakly- or highly-pathogenic filoviruses circulating in sub-Saharan Africa will contribute greatly to our understanding of the molecular pathology mechanisms of this family of viruses. Our work provides an important first insight into infection dynamics of Ebola virus in reservoir species over an extended period (2012-2106); an understanding of these infection dynamics is essential to gauge the overall risk of zoonotic exposure, as demonstrated in our ground-breaking study “A Unified Framework for the Infection Dynamics of Zoonotic Spillover and Spread” of this neglected area (Lo Iacono et al, PLOS Negl Trop Dis 10(9), e0004957;2016).
The results of our risk assessment can potentially help to aid both national and international surveillance activities by highlighting both the most likely routes of introduction of bat borne viruses, such as Ebola virus and Nipah virus, and where in the EU introduction is most likely.

Influenza virus
• Role of genetic and antigenic variation. There have been three major impacts of our research in this area. First, the risk assessment of nature evolving an airborne mammal-to-mammal transmissible avian influenza virus. This is a key question for influenza pandemic preparedness. We performed an analyses of all available influenza A/H5N1 virus surveillance data, and found that there are many viruses that have recently circulated that might require only three additional mutations to become mammal-to-mammal transmissible. There were also a small number of viruses that might require only two mutations. We then developed a mathematical model of within-host virus evolution to study factors that could increase and decrease the probability of the remaining mutations evolving after the virus has infected a mammalian host. Our work indicates that it is possible for a mammal-to-mammal transmissible A/H5N1 virus to evolve in nature, and highlights critical areas of research that are needed to further refine the risk analysis, understand the evolutionary processes, and potentially mitigate the risk. This work was one of the key pieces of information instrumental in the outcomes of meetings at the WHO, the US government's NSABB, and the Dutch government's decisions on the original EMC and Madison findings. The work was published in Science (Russell et al, Science 2012) Second, a key question in the evolution of influenza viruses is how they change antigenically, and any receptor binding chance they might have when they change. This is important both for seasonal and pandemic influenza virus evolution, seasonal influenza virus vaccination, and pandemic preparedness. Our paper in Science (Koel et al, 2013), in which we show that antigenic change of seasonal influenza viruses is caused by amino acid substitutions on the periphery of the receptor binding site—substitutions that may well reduce the intrinsic fitness of these viruses by distorting the receptor binding site and thus be the primary governing factor in the antigenic evolution of human seasonal influenza viruses. This publication is key, as it is this process—of escaping from herd immunity that may well change the receptor binding properties of the virus that could accidentily cause the virus to gain related receptor binding affinity in a new host and thus change host specificity. This paper is in the process of changing the field’s perspective on how influenza viruses evolve. Third, we have developed a breakthrough method for the analysis of human serology data. This work was published in Science (Fonville et al, 2014), describes a new high-resolution technique to describe immunological pressures, and show how these can be used to design better seasonal and pandemic influenza vaccines. The above work, when combined, leads to to a new generation influenza virus vaccine that will be tested phase II clinical trials that start in 2019.
• Genetic diversity of highly pathogenic avian influenza A/H5N1 during human infection. The study on avian influenza is unique in that it provides detailed insights into the diversity of H5N1 quasi species and evolutionary patterns of HPAI H5N1 during actual human infection and the first steps are made towards better insight into possible molecular mechanism of human adaptive mutations. These findings are highly relevant for an informed assessment of the pandemic risk of this virus.
• Susceptibility to swine influenza viruses. Data suggest that naïve animals (human children) would be susceptible to European swine H3N2. Those that had been previously exposed to human seasonal H3N2, infected or immunized, are at less risk from this hazard. However, escape mutants may be a possibility, this factor requires further investigation. This work has fed into APHA’s (formerly AHVLA) role in providing consultancy for the pig expert group UK, OIE network of expertise on animal influenza – swine influenza (which has linkage to OFFLU), academia through presentation of the data at international conferences.
• Olfactory route of influenza virus entry. Demonstration that the olfactory nerve is a relatively neglected route for influenza virus and many other viruses to enter the CNS implies that those investigators studying experimental animal models and those involved in human clinical cases need to take the olfactory nerve into account in sampling and analysis. The evidence that seasonal H3N2 influenza virus may use the olfactory route to invade the CNS emphasizes the need to examine the olfactory tract for evidence of virus replication in any autopsy case involving influenza. Based on the demonstration in a ferret model that vaccination is more effective than prophylactic oseltamivir in preventing CNS invasion by HPAI H5N1 virus via the olfactory nerve, and considering the fact that many other viruses can enter the CNS via the olfactory nerve in animal models, it is important to include the olfactory tract in future pathogenesis, antiviral, and vaccine studies involving influenza viruses and other viruses that replicate within the nasal cavity, especially viruses with a known neurotropic potential.
• New model for influenza-induced ARDS. Our new model for influenza-virus-induced ARDS has implications for pathogenesis research, both by identification of topics of particular interest and by exposure of gaps in our knowledge. The model described in this review also provides a rationale to choose therapeutic targets to minimise the tissue-damaging response against influenza virus infection of pulmonary alveoli. Our detailed co-culture study shows that the alveolar epithelium both forms the main barrier against fluid leakage to the alveolar lumen and is the key target of IAV upon infection via the respiratory route. This implies that improving and restoring the integrity of the alveolar epithelium may play an important role in reducing the severity of IAV-induced oedema and ARDS.
• Role of wild birds in global spread of HPAI strains. The potential role of wild birds in the circumpolar circulation of influenza viruses, such as the currently spreading H5N8 virus, does point to the need to increase our knowledge about the connectedness at the vast circumpolar (sub)arctic breeding areas between migratory waterfowl populations originating from different wintering areas. Surveillance of waterfowl at the crossroads of migratory flyways to wintering areas in Europe, Asia, and North America would inform epidemiological risk analysis and provide early warning of specific HPAI threats to poultry, and potentially human, health.
• Role of immunocompromised individuals for pandemic threat of influenza. Immunocompromised individuals could represent the elusive index cases for the generation of a transmissible variant of zoonotic influenza virus in the human population. This possibility could have serious public health implications for isolation and monitoring of patients with zoonotic influenza in countries where such measures are possible. Although these measures can be difficult to implement, they would also be essential in areas with poor medical infrastructure and high or increasing numbers of immunocompromised individuals.
• Required properties for airborne transmission of influenza virus. Over the past years, many mutations have been described in the literature which were the result of adaptation of avian influenza viruses to the human host. With our studies done under Antigone, we have identified the genetic changes that are required for airborne transmission between mammals. Furthermore, we have identified the phenotypic changes associated with these genetic changes as well as other functionally equivalent substitutions. Ongoing surveillance studies and epidemiological investigations may now be improved by including virus phenotyping assays to screen if the required properties for airborne-transmissibility are present. Our new knowledge of phenotypic viral properties for airborne transmission is thus not only fascinating from the basic scientific perspective, it is also crucial to predict risks, enhance surveillance, and perhaps even to adapt pharmaceutical and non-pharmaceutical intervention strategies to prevent outbreaks or pandemics.

Crimean-Congo hemorrhagic fever virus and other tick-borne pathogens
Most of our work was on Crimean-Congo hemorrhagic fever (CCHF), one of the major emerging disease threats in Europe. Although CCHF is an old disease (known since 40’s), the knowledge about its epidemiology and pathogenesis is still limited. As an example, it is still unknown whether ticks of different species, besides Hyalomma, are competent, that means capable to orally acquire, maintain and transmit the virus to a next vertebrate host. It is also unknown whether there are pathogenicity differences among the genetic lineages, and which parts of the genome are associated with increased pathogenicity. In the frame of ANTIGONE project we tried to unravel the mystery on CCHF epidemiology in Greece, where the seroprevalence is high, suggesting that CCHFV strains are circulating in the country, however cases of the disease are not observed. We found that strains of lineage Europe 2 (AP92 and novel AP92-like strains) are often detected in the country, while Europe 1 is also present, but it is detected in Rhipicephalus spp. ticks, which seems to be not competent vectors. Furthermore, it was found that the population of H. marginatum ticks, the known virus vectors, is very low and sufficient to produce outbreaks. Since CCHFV causes a severe disease in humans, often in an outbreak form, knowledge about its epidemiology is essential to assess properly the risk of infection. The results of the study showed that in the absence of Hyalomma ticks, the likelihood of a human case is very low. Therefore, the risk models have to be based on biotic and abiotic factors that may facilitate the spread and establishment of the competent vector. During the project we were able to isolate a CCHFV strain of lineage Europe 2, which strenthen our capacity to study better the genetic differences among lineages and indentify the parts of the genome which are associated with pathogenicity. At societal level, the analysis of the data by public health authorities provides the tools for proper risk assessment to minimize CCHFV epidemic potential and to manage disease outbreaks effectively.
We also worked on additional tick-borne pathogens, and we detected several rickettsial species, some of them for the first time in Greece. This knowledge is very important to understand better the distribution and epidemiology of the Rickettsia species in various ecosystems and environments. In addition, two novel tick-borne phleboviruses were detected in Greece (Antigone virus and Lesvos virus), which differ highly from all currently known phleboviruses. Tick-borne phleboviruses were totally neglected until recently, when it was found that two tick-borne phleboviruses (SFTSV and HRTV) cause severe and potentially fatal disease in humans. The discovery of new tick-borne phleboviruses provides knowledge on the epidemiology of this group of viruses and enables the design of diagnostic tools.

Most of the emerging viruses are originating in zoonotic reservoirs. The vast majority is non pathogenic in its specific reservoir probably due to a long term co-adaptation that have allowed the selection of efficient countermeasures by the host. On the other hand, these viruses become pathogenic when jumping across the human barrier to human. Our goal is to dissect and understand the molecular basis explaining this difference of pathogenicity. Beside its pure fundamental interest, this research may help to develop antiviral strategies against the pathogenic viruses. We are studying the Hantavirus genus that is combining a certain number of handicaps. It is the last genus among negative strand RNA viruses for which no reverse genetics have been established so far, what seriously impairs its molecular analysis. In addition, Hantaviruses circulating in Europe that constitute our principal interest, grow very poorly in cell culture. Finally, no satisfactory animal model exists since rodents are the reservoir hosts of Hantaviruses. Despite these difficulties, the ANTIGONE project has permitted preliminary results that may explain the pathogenic outcome of HFRS in Europe and thus allow us to search for countermeasures (ultimately drugs) against them.
The discovery of a new Hantaviruses circulating among Rattus rattus in the Indian Ocean (Mayotte Island) and related to the lineages circulating in Asia may also have a potential impact for Public Health in this region of the world. Indeed, Hantaviruses from Asian lineages are usually more pathogenic in human (Hantaan virus and Seoul virus are known to reach 5-10% lethality). Due to the promiscuity between human and rodents in this region of the world, it is plausible that several HFRS with no related aetiology so far, would be linked to these Hantaviruses. We have engaged a sero-surveillance in this part of the world (Madagascar) and preliminary results indicate about 2% of seropositive people in the normal population, leaving open the option that a greater percentage will be observed in patients. This observation may be enlarged to the whole Africa, unexplored for Hantaviruses so far. In addition, it is of note that a greater number of Seoul or Seoul-like viruses are isolated in Europe, particularly in pet shops selling Rattus spp.

Duvenhage virus
ANTIGONE is a consortium dedicated to contribute to a “better understanding of the emergence and transmission of pathogens with pandemic potential and improve preparedness planning, in particular modelling and prediction, but also development of appropriate intervention measures.” The work that P1 has performed with DUVV has significantly added to our knowledge of lyssavirus infection in general and DUVV in particular. Unfortunately, we were not able to study the role of different DUVV genes in the pathogenesis of DUVV. However, we were able to identify using genomics and proteomics, important pathogenic mechanisms that lyssaviruses use to induce disease. This knowledge, not only added to our better understanding of the virus-host interactions and the (innate) response of the host to these important human pathogens, it also paved the way to identify more rational and tailored intervention strategies for these viruses. The results obtained from the studies in WP9 significantly contributed to the design of novel and innovative treatment options for rabies.

African Henipaviruses
Our research has helped to advance our understanding of the species barriers that such viruses lmust overcome to replicate in a new host species in terms of viral entry and replication mechanisms. Understanding these mechanisms will lead to clues concerning not only the adaptation of such viruses but also again for our ability to predict the potential for zoonotic spillover based on the phenotypic analysis of genetic data.

Nipah virus
Building on our important publication outlining a structure/function model for virus replication, we have shown that the molecular interaction between the viral N and P proteins, as part of the viral replication machinery, is a potent target for antiviral therapy (Yabukarski et al., 2014). Indeed in this paper, we have demonstrated that a peptide mimicking part of the P protein can effectively block viral replication upon transfection in vitro. Building on this publication, we have continued in this area, using diverse peptides mimicking part of the viral P protein to block viral replication in other related human and animal paramyxoviruses. The development of synthetic peptides is envisaged in this respect, with the aim of testing their efficiency in cell or animal models of viral infection, including for NiV. A European patent application was deposited for this work in July 2014 and extended internationally through Inserm in 2015. This patent is currently under negotiation for take-over by a start-up biotech company set up indirectly by our external partners aimed at exploiting the therapeutic potential of our discovery. The therapeutic scope of such peptides is vast, potentially impacting on a range of both human and animal diseases with important socio-economic implications.

Escherichia coli
• Cattle as potential reservoirs. Between May and July 2011, germany and parts of Europe experienced the that-far largest outbreak of henolytic uremic syndrome (HUS). With an EHEC strain of serotype O104:H4, a novel type of EHEC emerged, that far only known to colonized the intestinal tract of humans. Previous EHEC outbreaks were primarily caused by consumption of contaminated food or direct contact to cattle. Primary research studies conducted within ANTIGONE proved for the first time that the strain is capable of utilizing cattle as a reservoir host by colonization of the intestinal tract for a certain period of time. As a consequence, cattle shall be regarded as potential reservoirs for humans of this emerging pathogen. Diagnostic methods implemented for monitoring of zoonoses have to be adapted to allow for a reliable and rapid detection of O104:H4 and related strains with similar virulence potential.
• Micro-evolution during outbreak. The emergence of so-called hybrid E. coli strains as a result of horizontal gene transfer events combining multiple virulence factors of different pathotypes complicates the epidemiology of pathogenic E. coli strains. The unusual combination of virulence factors as well as the presence of antimicrobial resistance traits, including extended-spectrum beta-lactamases (ESBLs), highlights the potential risk posed by the O104:H4 outbreak strain and can also further complicate the treatment. Accordingly, our results regarding the roles of stx2-harboring bacteriophages, outer membrane vesicles, ESBL-encoding plasmid pO104_90, and the AAF/I-encoding plasmid pO104_80 increase our knowledge regarding the risk of interaction between E. coli isolates or virulence factors, leading to new hybrid strains. These results also deepen our understanding of events involved in the emergence of the outbreak strain and on its survival on plants. The pathogenic potential of the putative E. coli O104:H4 progenitor strain increased upon acquisition of virulence and resistance genes located on mobile DNA elements such as bacteriophages, and plasmids. It can still not be ruled out that mobile genetic elements may also be associated with and thus be spread by outer membrane vesicles. Our data on the microevolution of E. coli O104:H4 during the outbreak and within patients shed new light on genomic regions of the outbreak strain that may be under selection in vivo.
• Development of intervention strategies. Our results provide first steps in the development of strategies to attenuate or counteract the O104:H4 outbreak strain and other EHEC in general by (i) selectively inducing the loss of the stx-harboring bacteriophages or (ii) the use of lytic bacteriophages targeting EHEC variants incl. the outbreak strain. The comparison of the impact of different antibiotics on the induction of stx-harboring phages may also pave the way for the development of antibiotic treatment options during EHEC infection avoiding strong Shiga toxin-dependent symptoms or sequelae.
• Epidemiology of virulence factors. The work on enteric E. coli has provided insight in the epidemiology of virulence factors and the potential for outbreak strains to emerge. In addition, ANTIGONE partners 3, 6, 8 and 9 who collaborated in WP4, together with researchers from the University of Oxford, UK and the Robert Koch Institute, Germany, have successfully built the HECTOR consortium, which will study host restriction of E. coli using whole genome sequencing, animal experiments and mathematical modeling, funded by the EU through ERA-net and the JPI-AMR program (C. Schultsz project leader), between 2017 and 2020.

Streptococcus suis
The work on S. suis has contributed to improved knowledge of infection routes and host-pathogen interaction. This work has had substantial impact as it has enhanced knowledge of the role the gastro-intestinal tract may play as an entry site of infection in both pigs and humans. This enhanced knowledge has significantly contributed to research targeted at the design of potential interventions to reduce S. suis colonization and infection in pigs and hence in humans. In addition, the work has provided ample opportunities for future research into the zoonotic potential of S. suis particularly through the study of epigenetic regulation through restriction-modification system as well as horizontal gene transfer of yet unknown virulence factors.

Anaplasma phagocytophilum
Traditional tick control methods, based primarily on chemical acaricides, have proven not to be sustainable because of the selection of acaricide resistant ticks. Tick vaccines appear to be a promising and effective alternative for control of tick infestations and pathogen transmission. The results of our studies with the application of omics technologies to the characterization of tick-host-pathogen interactions resulted in effective screening platforms and algorithms for the discovery of new tick protective antigens. Vaccinomics and reverse vaccinology approaches could be used to identify and fully characterize candidate protective antigens and validate vaccine formulations. New candidate protective antigens will most likely be identified by focusing on abundant proteins with relevant biological function in tick feeding, reproduction, development, immune response, subversion of host immunity and pathogen infection and transmission. Consequently, tick protective antigens will be discovered with multiple impacts when used in a vaccine including reductions in (a) tick infestations and fertility, (b) tick pathogen infection, (c) tick vector capacity for pathogen transmission and (d) tick response to pathogen infection. These new vaccines will likely combine tick antigens associated with different protective mechanisms alone or in combination with pathogen-derived antigens to have an effect on reducing tick infestations while affecting pathogen infection and transmission to ultimately result in the control of tick-borne diseases. Finally, the most economical integrated tick control strategies will be those combining tick vaccines with other control methods while reducing acaricide applications to reduce risks for humans, animals and the environment. Additionally, the datasets of niche overlap among ticks and hosts, species interactions, and the fraction of the environmental niche in which tick-borne pathogens may circulate as a result of interactions and overlapping environmental traits, provide a valuable resource for researchers and agencies interested in tick-borne pathogens, as they conciliate the abiotic and biotic sides of their niche, allowing exploration of the importance of each host species acting as a vertebrate reservoir in the circulation of tick-transmitted pathogens in the environmental niche.

The analysis of host-tick-Borrelia interactions is essential to understand the molecular mechanisms that affect transmission of pathogen by vector, development of Borrelia infection, or triggers the host reaction to tick bite and infection. Genomic, transcriptomic, proteomic, lipidomic, and metabolomic analyses of the parasite, vector and the host yield a global view on molecular and biochemical processes in both spatial and temporal compartments. This will lead us to the identification of molecular markers for rapid infection diagnosis or drug design or to the identification of novel antigens as vaccine candidates.
Borrelia lives in two distinct and different milieus. B. burgdorferi genes are differentially expressed as the spirochete traverses the enzootic cycle between its tick vector and mammalian host. Passing back and forth between vector and host spirochete regulates expression of mammalian phase-specific or tick phase-specific genes. Transcriptional changes in Borrelia are regulated by specific signals that include temperature, pH, nutrients, and as yet undefined vector- and host-specific factors. The nutrients themselves, especially carbon sources, and/or their metabolic by-products seem to provide regulatory and chemotactic signals that guide the spirochete as it moves between hosts and vectors, crossing interspecies barriers. Despite its small genome, the Lyme disease spirochete possesses deceptively complex machinery for tightly regulating gene and protein expression, with an unusual combination of components identified to date. Discovery that spirochete lipoproteins are major inflammatory agonists led to understanding of the mechanisms by which the mammalian host senses the presence of live spirochetes and mobilizes cellular and humoral defences to combat the intruder. By contrast, little is understood about the processes that occur late in infection and the mechanisms that enable the bacterium to persist in the face of the robust cellular and humoral immune responses that it elicits.
Co-evolution of the spirochete with its arthropod vector and mammalian hosts has enabled the bacterium to take advantage of host physiological processes with compensatory reductions in its own biosynthetic machinery. Borrelia has developed ingenious and parsimonious strategies to obtain and utilize the nutrient sources available within the ”feast (host associated) and famine (vector associated)” periods of its life cycle.
Analysis of the factors that contribute to spirochete persistence and are involved in modulation of pathogen-sensing pathways and host responses will lead to a better understanding of the determinants of vector and host specificity for B. burgdorferi. Manipulation with these determinants might interrupt the cycle of spirochete transmission. Although eliminating Lyme disease spirochetes from nature is unrealistic, diminishing their threat to humans would seem to be an achievable goal.

Our Bartonella studies, conducted in collaboration with a WHO international reference center, suggest a radically different ecology for Bartonella species in West Africa compared to other global regions such as South East Asia. The reasons for the differences are not entirely clear and could include the relative lack of contact with pet cats and dogs in West Africa, in contrast to parts of SE Asia.

Animal and tissue models.
• Bat-Lagos bat virus. Our results indicate that intramuscular inoculation of a 10-2 dilution of Ghana LBV strain into E. helvum comprises an appropriate model to study LBV pathogenesis in this species. No evidence exists for common human exposure to LBV, but our work serves to provide a unique non-rabies lyssavirus model in a natural host for subsequent study. The results regarding the presence of the virus in taste buds have particularly important implications for the study of virus excretion, not least given the relative rarity of the virus in the salivary glands of infected, diseased bats. We hope to extend these studies into further natural transmissions studies of LBV in their natural Eidolon helvum host.
• Hamster-metapneumovirus. The newly developed hamster model is the first model that is available to study airborne transmission of other respiratory viruses than influenza viruses. This model can be used to study the requirements for airborne transmission of paramyxoviruses, and results from these experiments will help to identify commonalities in phenotypes required for transmission of members of both virus families.
• Cell culture/Yersinia model. The work carried out in this study established a cell culture/pathogen model which can be used to investigate highly potent pathogens, adapted parasites and commensals. Examples include Burkholderia mallei/B. pseudomallei or Brucella suis bv1 (highly pathogenic for humans) and B. suis bv2 (pathogenic for swine only). This model could also be applied to pathogens causing granulomatous infections such as Mycobacteria/Tbc. Its applicability to viral or fungal agents has not yet been determined and ought to be investigated. This model could possibly also be used to identify excessive and undesired reactions of the body to numerous agents and their control via pharmaceuticals or for testing new substances.

Diagnostic assays
• MERS-CoV. The identification of MERS-CoV as the cause of MERS allowed development of rapid and specific diagnostic tests.These allowed epidemiologic investigations to identify the original source of the virus and potential transmission events between the infected patient and direct contacts. It also allowed the development of serologic assays for surveillance studies. We developed and validated MERS-Coronavirus detection assays and collaborated with numerous companies (Altona Diagnostics, TIB MOLBIOL, EUROIMMUN) to translate the in-house tests into commercial products. The availability of commercial kits allows countries to address local epidemics independently.
• Escherichia coli. Our comparative genomic analysis of the E. coli O104:H4 outbreak strain and other EHEC variants provided valuable information for the development of rapid molecular tests for the differentiation of EHEC. Our improved typing approaches for all clinically relevant EHEC variants by real-time multiplex PCR represent a rapid and reliable method for HUSEC identification and Stx subtyping in diagnostic laboratories. Regarding pathogenic E. coli, the application of the new set of qPCRs in surveillance studies could be an efficient early-warning tool for the emergence of zoonotic strains.

We established an invaluable biobank and cell culture collection that will be highly beneficial for future research projects. In the past years, hundreds of new viruses were identified in the accumulated samples that will help to understand the virus diversity and evolution in animals. In order to understand and identify the main drivers for zoonotic spillover events it is crucial to know the existing virus diversity in order to i) understand the virus evolution ii) to identify genetic markers that will help to predict future epidemics/pandemics. The established cell culture collection has already led to numerous collaborations and high impact publications in mBio, Nature, Cell and eLIFE. To date more than 35 institution in >20 countries have received such cell cultures. The cell cultures were highly valuable to establish in vitro assays that allowed to investigate fundamental traits of known and novel pathogens. This has been highlighted during the ANTIGONE project, e.g., to explore the receptor and host range of MERS-Coronavirus and Ebola virus (Müller et al., 2012, mBio; Raj et al., 2013 Nature; Urbanowicz et al., 2016, Cell and Ng et al., 2015 eLIFE).

Model framework
The generic entry assessment model framework developed as part of WP13 involved a number of novel methodologies which are already being utilised in other EU projects, SPARE (funded through Animal Health and Welfare ERA-Net) and COMPARE (which received funding from the Horizon 2020 research and innovation programme under grant agreement No. 63476), both of which involve development of generic quantitative entry assessments for pathogens into the EU.

Improved communication policy
Our work in Qatar might improve the communication policy of the public health authorities, that can be inspired from other models we have studied in Mongolia, Laos and Australia. The problem for public health authorities is how to combine the local attachment of humans to animals considered as emblematic species (such as camels in the Arab Peninsula, elephants in Laos and bats in Australia) with the protection of public health faced with the threat of emerging pathogens. A communication policy needs to be created showing that distances with animals are not incompatible with traditional forms of attachment, in so far as it involves animal workers in the definition of adequate relations of proximity and distance with animals. The relations between tradition and modernity must also be involved in communication policies, because in each of our settings, the emergence of zoonotic pathogens is linked to the transformations of traditional practices by the globalization of exchanges.

Improvement of interdisciplinary research
Through Antigone, there has been improved interdisciplinary research involving human and veterinary medicine, virology, bacteriology and parasitology. The involvement of anthropology was novel and should be encouraged in future. The impact of this study is also to enhance further collaborations between biologists, anthropologists and ecologists on the role of the human/animal/environment interface in the emergence of zoonotic pathogens. It has been an experimental venue to combine quantitative and qualitative analysis, interviews and long-term observation and participation, in defining significant fieldwork sites for the study of zoonotic risks. Since this question will increasingly motivate public health agencies and insurance companies, as witnessed by the support of Axa to our research team, this method can be proposed for other consortiums bringing interdisciplinary competences on the study of zoonoses.

The ANTIGONE project has greatly increased the interaction among established scientists but has also contributed to the education of young scientists. The four One Health course attracted over 100 young scientists from approximately 20 countries, and exposed them to reknowned experts in the area of emerging infectious diseases, and emphasized to them the value of a One Health approach to research in this area.

List of Websites:

Related information

Reported by



Life Sciences
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top