Human NoVs are a common cause of epidemic and sporadic acute gastroenteritis worldwide 
. The prototype Norwalk virus was discovered in 1972 
, however NoV research and the development of prevention strategies have since been hampered by the lack of robust cell culture system and animal model that closely mimics the clinical symptoms of NoV gastroenteritis in humans. Despite these difficulties, NoV VLP vaccines are in early phase clinical trials 
. Evaluation of these vaccines largely depends on challenge studies in human volunteers.
Development of an animal model for human NoV gastroenteritis might be based on infection of animals with human NoVs or the use of cultivable animal CVs as surrogates. Attempts to experimentally induce human NoV infection in NHP species such as common marmosets, cotton top tamarins, cynomolgus macaques, pigtail macaques, rhesus macaques and chimpanzees almost invariably resulted in asymptomatic infections 
. In one study however, a symptomatic infection characterized by diarrhea and vomiting of newborn pigtail macaques was induced following inoculation with the Toronto virus 
, but these results have not been confirmed by other studies. Moreover, detection of CVs or CV antibodies, including NoVs was reported in colonies with captive NHPs 
. In captive NHP colonies, CVs have been detected in stool samples of animals with symptoms of diarrhea. Recent data suggest an ~80% ReCV seroprevalence in rhesus macaques housed at three separate U.S. colonies [27, Sestak and Farkas, unpublished]. While ReCV seroprevalence rates are high in juvenile and adult rhesus monkeys, other NHP species exhibit different and/or lower seroprevalence rates. As suggested, these differences could be due to species-specific susceptibility/resistance factors such as HBGAs 
. Considering the high TV seroprevalence combined with high rates of juvenile diarrhea-associated morbidity in these colonies, it is suspected that ReCV infections play an important role in the etiology of NHP gastroenteritis. Here we examined for the first time whether experimental challenge of juvenile rhesus macaques with the tissue culture-adapted prototype ReCV (TV) induces clinical disease. Although no equivalent inoculum of tissue culture-adapted human NoV was available, two additional animals were challenged with human NoV stool suspensions to compare the clinical outcome between TV- and human NoV-inoculated animals. In order to increase the chance of infection, both animals were inoculated with a mixture of GII.2 and GII.4 NoVs. Mixed NoV infections have been described in oyster-associated human NoV outbreaks 
. Also, the existence of recombinant NoVs 
indicates that mixed NoV infections do frequently take place.
Two out of three TV-inoculated macaques developed symptoms of gastroenteritis (diarrhea, fever and virus shedding in stools but no vomiting or reduced appetite were noticed) while the third remained asymptomatic. All three animals shed virus for more than a week (8–10 PID) even though maximum virus shedding reached ~105
viral copies per gram of stool only in HC55 and HB61. Quantitative evaluation of virus shed in these two macaques combined with the daily output of stools () indicate that the amount of virus shed exceeded the amount of virus in the inoculum. Vomiting was not detected as a clinical sign of TV infection. This could be due to small number of animals in this study or other reasons. Studies with NoV-infected children suggest that ~25–30% of these children do not vomit 
. In addition, still unknown differences between ReCV and NoV pathogenesis might also be linked with the lack of vomiting in TV-infected macaques.
All three TV-inoculated macaques developed ≥4 fold TV-specific VN antibody responses within 7 days PI, indicating an early memory response, consistent with low levels of VN antibodies prior to inoculation. VN antibodies reached their peak by PID 7–9 in all of the three animals and persisted at high levels up to 4–5 weeks PI. After 5 weeks PI, VN antibody titers started to decline, possibly due to absence of re-exposure(s) that take place in colony animals 
. It is also possible that the binding (ELISA) antibodies would be detected for much longer period of time than functional (VN) antibodies. Finally, it is important to highlight that the low levels (1
80) of VN antibodies were still present at the end of the study in TV-inoculated macaques. Consistent with these results, declining NoV immunity has also been observed in human volunteers 
. Thus, questions still remain regarding the longevity, magnitude, class and role of CV antibodies in protection.
Interestingly, VN antibody responses could not be detected in any of the TV-infected animals against FT285 a GI.2 ReCV isolate, indicating that the two strains likely belong to different serotypes. TV and FT285 ORF2 and capsid proteins share 69% nt and 74% aa homology (Farkas and Sestak unpublished), which is comparable with differences between different human NoV genotypes. If humoral immunity plays a role in protection against enteric CV infections, our results are translational for NoV vaccine design since multiple genetic types have been described for both ReCVs and human NoVs. At present, ~10 cell culture-adapted ReCVs representing at least three genotypes are available in our laboratory. Consistent with previous reports, none of the two human NoVs used in this study caused clinical infection. Since all macaques secreted type-B HBGA in their saliva (not shown), and both NoV challenge strains were obtained from a type-B patient, it was predicted that both strains would replicate in macaques. However, only GII.2 virus was detected in macaque GI96, while both Gll.2 and Gll.4 viruses were shed by HB11 (). Although duration of shedding in HB11 was suggestive of virus replication, no antibody responses were detected against Minerva (GII.4) VLPs. Macaque GI96 had relatively high level (1
800) of ELISA antibodies at the beginning of the study. This could explain the complete absence of GII.4 virus shedding in GI96. GII.2 VLPs were not available to evaluate the GII.2-specific seroresponses in this study. Low viral loads in stools of Norwalk virus-challenged rhesus macaques were also reported by Rockx and colleagues 
. In their study, two animals stopped shedding virus by PID 2 while one animal shed up to PID 19 while it developed strong antibody responses. In the same study, marmosets and tamarins, but not cynomolgus macaques, shed the virus up to PID 4 but did not develop antibody responses. Taken together, these studies indicate that NHPs can be subclinically infected with human NoVs but the infection is very limited and its intensity varies in different species and individuals.
Histopathological examination of H&E-stained duodenum biopsies obtained at PID 3 demonstrated lymphocytic infiltration of the lamina propria in all five inoculated macaques while mild villous blunting was present in three TV-inoculated macaques. These signs were consistent to some degree with reported histopathology of NoV-infected humans 
. Although no ultrastructural examination of enterocytes was performed by transmission electron microscopy, confocal microscopy examination of intestinal brush border and epithelial layer revealed no significant damage of enterocytes indicating the possibility that virus might have translocated into lamina propria without replicating in enterocytes. This finding differs from what was reported with human NoV-inoculated gnotobiotic pigs or calves that showed presence of viral particles inside the enterocytes by PID 3 
. Notwithstanding, our results are consistent with reported findings of human NoV i.v.-inoculated chimpanzees which showed the presence of NoV antigens in intestinal lamina propria while exhibiting the signs of asymptomatic virus infection such as virus shedding in stools and seroconversion 
. In all of these studies, enterocyte damage was minimal or undetectable regardless of the detection of virus antigen raising questions about CV pathogenesis and their exact host cell tropism. In case of TV-inoculated macaques, TV antigen-containing cells with perinuclear fluorescence were found in intestinal lamina propria ( and ). In many of these cells TV-specific antibodies co-localized with calnexin, an endoplasmic reticulum marker, consistent with the formation of replication complexes reported for feline CV 
To determine the authenticity of TV-containing cells, several of the human-specific cell markers were utilized (). A few CD20+TV+ B lymphocytes were identified directly in duodenum lamina propria by confocal microscopy. Since suitability of intestinal pin-head-sized biopsies for quantitative cell phenotype analysis by confocal microscopy is limited, in vitro
experiment with TV-inoculated PBMCs obtained from eight healthy macaques was performed, utilizing quantitative multicolor flow cytometry, to determine if CD20+TV+ B cells could be identified in these cultures ( and ). The presence of CD20+TV+ B cells was demonstrated corroborating the confocal microscopy data. Most of the TV+ cells expressed HLA-DR antigen, suggesting a) involvement of TV+ lymphocytes in antigen presentation to other immune cells or b) possibility of activated lymphocytes being infected. Increased viral RNA load (p<0.05) that was detected in cultured PBMCs within 24 h following TV inoculation supported the flow cytometry data and was suggestive of virus replication. In study by Lay and colleagues, Norwalk virus failed to replicate in macrophages or dendritic cells derived from human peripheral blood 
. A more recent study by Chan and colleagues indicates however that human NoV binds to the lamina propria and Brunner's gland cells of the human duodenum 
. We have also incorporated CD123 marker to determine whether some B cells start to acquire dendritic cell (DC)-like features following their in vitro
exposure to virus. Earlier reports by Bjorck and colleagues (1998) showed that CD19+ B cells can differentiate into DCs following cytokine stimulation 
. Although commonly used as a DC marker, CD11c is also expressed by activated or leukemic B cells 
. Recently, a subpopulation of memory B cells was shown to express CD11c 
. In our study we observed increased CD11c expression by TV-infected B cells suggesting the activation of these cells following TV infection.
Antibodies used for immunofluorescent staining of duodenum biopsy tissues.
The fact that TV was identified in both B and T cells is to some degree different from findings reported for murine NoVs that replicate efficiently in macrophage and dendritic cell lines 
. Since Rhesus CD14+ cells can differentiate into dendritic cells in vitro
, it will be interesting to evaluate if these cells will become permissive to TV infection. In summary, our results are consistent with findings of human NoV -infected chimpanzees that remained asymptomatic but showed the presence of NoV antigen in intestinal lamina propria 
. Many features of CV infection in primates such as HBGA-associated genetic susceptibility to infection remain to be elucidated. In this proof-of-concept study, we demonstrated the potential of rhesus macaques using ReCV as a diarrheal disease model for human NoV. In future studies candidate animals will be selected by prescreening for the specific HBGAs and/or by using newborn SPF or animals completely naïve of previous ReCV exposure while using highly cytopathic ReCV strain for challenge.