Restriction of Viral Replication, Rather than T Cell Immunopathology, Drives Lethality in Murine Norovirus CR6-Infected STAT1-Deficient Mice

ABSTRACT

Recent evidence indicates that viral components of the microbiota can contribute to intestinal homeostasis and protection from local inflammatory or infectious insults. However, host-derived mechanisms that regulate the virome remain largely unknown. In this study, we used colonization with the model commensal murine norovirus (MNV; strain CR6) to interrogate host-directed mechanisms of viral regulation, and we show that STAT1 is a central coordinator of both viral replication and antiviral T cell responses. In addition to restricting CR6 replication to the intestinal tract, we show that STAT1 regulates antiviral CD4⁺ and CD8⁺ T cell responses and prevents systemic viral-induced tissue damage and disease. Despite altered T cell responses that resemble those that mediate lethal immunopathology in systemic viral infections in STAT1-deficient mice, depletion of adaptive immune cells and their associated effector functions had no effect on CR6-induced disease. However, therapeutic administration of an antiviral compound limited viral replication, preventing virus-induced tissue damage and death without impacting the generation of inflammatory antiviral T cell responses. Collectively, our data show that STAT1 restricts MNV CR6 replication within the intestinal mucosa and that uncontrolled viral replication mediates disease rather than the concomitant development of dysregulated antiviral T cell responses in STAT1-deficient mice.

IMPORTANCE The intestinal microbiota is a collection of bacteria, archaea, fungi, and viruses that colonize the mammalian gut. Coevolution of the host and microbiota has required development of immunological tolerance to prevent ongoing inflammatory responses against intestinal microbes. Breakdown of tolerance to bacterial components of the microbiota can contribute to immune activation and inflammatory disease. However, the mechanisms that are necessary to maintain tolerance to viral components of the microbiome, and the consequences of loss of tolerance, are less well understood. Here, we show that STAT1 is integral for preventing escape of a commensal-like virus, murine norovirus CR6 (MNV CR6), from the gut and that in the absence of STAT1, mice succumb to infection-induced disease. In contrast to the case with other systemic viral infections, mortality of STAT1-deficient mice is not driven by immune-mediated pathology. Our data demonstrate the importance of host-mediated geographical restriction of commensal-like viruses.

INTRODUCTION

The mammalian intestinal immune system developed with colonizing species, including viruses, archaea, bacteria, fungi, protists, and helminths. Maintenance of this relationship is carefully regulated by host- and microbe-directed tolerance mechanisms which coevolved to limit the generation of inflammatory responses against the colonizing microbes (1). Various host-encoded factors regulate tolerance toward bacterial components of the microbiota, and in their absence, mice exhibit enhanced susceptibility to intestinal inflammation and/or systemic translocation of microbes or their by-products (2–5). In contrast, less is known about how the host regulates viral colonization.

Murine norovirus (MNV) CR6 demonstrates features of a mutualistic host-microbe interaction and is an attractive model to interrogate host-encoded mechanisms that maintain this relationship (6). In addition to providing supportive signals in immunosufficient germ-free (GF) mice (7–9), this strain can establish persistent infection in colonic tuft cells with continued shedding for at least 2 months in the absence of any detectable clinical or histopathological signs in wild-type (WT) mice (10, 11). In this context, CR6 persists via antagonization of the type III interferon (IFN) antiviral response (12). Notably, no clinical signs of disease have been reported for Rag1^−/−, Rag2^−/−, Ifnar^−/−, or Ifnlr1^−/− mice colonized with CR6 (13–15). In contrast, Stat1^−/− mice develop virus-induced disease following infection with CR6, as well as other persistent strains, MNV-O7 and MNV4 (16–19). In particular, elevated viral loads in the extraintestinal tissues of CR6-infected Stat1^−/− mice, through either parenteral infection or dissemination-enhancing mutation, result in severe disease and eventual death, in contrast to the case with WT mice (19, 20). These data suggest that signal transducer and activator of transcription 1 (STAT1)-dependent mechanisms control MNV CR6 replication and are necessary to maintain a commensal-like relationship between virus and host.

STAT1 signaling is critical for resistance to multiple viral pathogens. This includes MNV CW3, a strain that is cleared via coordinated adaptive immune responses by day 8 in WT mice (21, 22). In contrast, CW3 induces rapid mortality Stat1^−/− mice, indicating that STAT1 is essential for innate antiviral immunity in this setting (23). STAT1 has also been assigned a role in supporting adaptive antiviral immunity: in the context of vaccinia virus infection, antiviral CD8⁺ T cell expansion and survival are impaired in STAT1-deficient mice (24, 25). In contrast, STAT1 and type I interferon receptor (IFNAR) prevent immunopathology following infection with lymphocytic choriomeningitis virus (LCMV) Armstrong, influenza virus, and respiratory syncytial virus (26–30). These data indicate that STAT1-dependent protection from virus-induced disease is heavily context dependent. Importantly, the mechanisms underlying CR6-induced lethality in STAT1-deficient mice remain incompletely defined, and the influence of STAT1-dependent signaling on antiviral T cell function or immunopathology has not been addressed.

In this study, we examined the role of STAT1 in regulating persistent CR6 infection. As opposed to the rapid weight loss and mortality associated with CW3 infection (100% mortality by day 5), in our facility, CR6-infected STAT1-deficient mice displayed weight loss and mortality over a course of 28 days. At 7 to 10 days postinfection (p.i.), approximately 25% of STAT1-deficient animals displayed CR6-induced disease characterized by weight loss, splenomegaly, and spleen and liver histopathology. This was accompanied by systemic viral spread not seen in STAT1-sufficient littermate controls. Infection by the parenteral route resulted in 100% penetrance of CR6-induced disease, and this correlated with substantially higher viral loads in both innate and adaptive immune cells. Notably, the time frame of CR6-induced disease onset correlated with hyperaccumulation of antigen-specific CD8⁺ T cells, a Th17-skewed CD4⁺ T cell response, neutrophilia, and elevated levels of circulating cytokines, a phenotype commonly associated with immunopathology. These data, in combination with infection-induced dysregulation of colon-associated microbial communities in Stat1^−/− mice, suggested that STAT1-dependent regulation of CR6 may be mediated in a microbiota or immunoregulatory manner. Here, we demonstrate that neither antibiotic-mediated microbiota depletion nor attenuation of adaptive immune responses was sufficient to rescue mice from virus-induced disease. In contrast, therapeutic administration of a viral polymerase inhibitor, 2′-C-methylcytidine (2CMC), limited viral replication and prevented virus-induced disease in STAT1-deficient mice without restoring the profile of antiviral immune responses seen in WT mice. Thus, in contrast to STAT1’s role in limiting immune-mediated damage to host tissue seen in the context of other viral pathogens, STAT1 protects the host by restraining CR6 viral replication. These data effectively uncouple the effects of a dysregulated T cell response from virus-induced damage to the host and demonstrate that STAT1-mediated restriction of CR6 replication is necessary to prevent systemic spread and subsequent lethality.

RESULTS

STAT1 limits MNV CR6 dissemination and protects against virus-induced disease.

To limit the impact of environmental influences on our analysis of STAT1 function in the context of CR6 exposure, we bred Stat1^+/− × Stat1^+/− to generate littermate Stat1^+/+, Stat1^+/−, and Stat1^−/− (Stat1^WT, Stat1^Het, and Stat1^KO) mice. We took advantage of the clear requirement for STAT1 in response to CW3 infection to test whether Stat1^Het mice displayed any signs of haploinsufficiency. Consistent with previous observations (23), Stat1^KO mice demonstrated rapid weight loss and 100% mortality by day 5 post-CW3 infection (Fig. 1A and B). In contrast, 100% of Stat1^WT and Stat1^Het mice survived through day 8 post-CW3 infection with no signs of weight loss, mortality, or other clinical symptoms (Fig. 1A and B). In addition, there were no differences in viral clearance (Fig. 1C) or the frequency or number of MNV-specific CD8⁺ T cells (Fig. 1D to F) between Stat1^WT and Stat1^Het littermates, indicating that a single copy of the Stat1 gene is sufficient for antiviral immunity in the context of MNV CW3 infection. Consistent with this, Stat1^WT and Stat1^Het bone marrow-derived dendritic cells (BMDCs) expressed similar levels of Stat1 mRNA (Fig. 1G), and both responded to MNV infection by upregulating gene and protein expression of STAT1 (Fig. 1G to I). We thus modified our breeding strategy, using Stat1^Het × Stat1^KO mice to generate STAT1-sufficient (Stat1^Het) and STAT1-deficient (Stat1^KO) littermates for the remaining experiments.

FIG 1.

Stat1^Het phenocopy Stat1^WT mice. (A to F) Stat1^Het and Stat1^WT mice were infected with 10⁴ PFU of MNV CW3 or PBS p.o. and euthanized on day 8. Data are representative of those from two independent experiments with 3 or 4 mice per group. (A) Weight change as a percentage difference from initial. (B) Survival curve of CW3-infected mice reaching humane endpoint, Mantel-Cox test. (C) CW3 genome copies in ileal and splenic tissue in naive and infected Stat1^WT and Stat1^Het mice were quantified by reverse transcription-PCR (RT-PCR). (D to F) Frequency (D) and total number (E) of antigen-specific MNV-CW3 Kb-P1^519Y tetramer⁺ CD8⁺ T cells isolated from the spleen and small intestinal IEL compartment (F). (G to I) BMDCs derived from Stat1^WT, Stat1^Het, and Stat1^KO mice were pulsed for 1 h with MNV CW3 (multiplicity of infection [MOI] = 0.5), washed, and analyzed at the indicated time. Data are compiled from 3 independent experiments with triplicate samples from individual donors. LoD, limit of detection; RQ, relative quantity of Stat1 gene expression, normalized to hprt in Stat1^WT cells. (H) Flow cytometric detection of STAT1 in BMDCs at 6 h pi. (I) Kinetics of MNV CW3 infection-induced STAT1 expression in BMDCs.

Using this breeding strategy, we assessed the role of STAT1 in the context of a commensal-like persistent intestinal virus. Consistent with previous reports of CR6 infection in WT mice (10), Stat1^Het mice had no detectable clinical symptoms through 28 days of CR6 exposure despite a high colonic viral burden (Fig. 2A to C), indicating infectious tolerance (1). In contrast, 75% of Stat1^KO mice demonstrated clinical signs of disease, including piloerection, reduced mobility, sunken eyes, and a tiptoe gait, accompanied by >10% loss of body weight; ultimately, 70% reached a humane endpoint of >20% loss of body weight due to virus-induced disease and were euthanized or naturally succumbed to infection (Fig. 2A and B). At the time of euthanasia (day 28 p.i. for Stat1^Het mice), viral loads in the colon and spleen were measured. Consistent with previous reports (14) and in contrast to the case with Stat1^Het mice, CR6 was readily detectable in splenic tissue of Stat1^KO mice (Fig. 2D), suggesting that STAT1 plays a critical role in maintaining the intestinal restriction of CR6 infection. Moreover, tissue and systemic viral burdens correlated with virus-induced mortality in Stat1^KO mice (Fig. 2C and D). Notably, a fraction of Stat1^KO mice survived the infection without losing >10% of body weight or displaying other signs of virus-induced disease. At 28 days post-CR6 exposure, viral burdens were below the limit of detection in both colon and spleen of these mice (Fig. 2C and D). CR6-specific IgG was detected in the serum of these mice, confirming that they had been infected (Fig. 2E).

FIG 2.

STAT1 limits MNV CR6 dissemination and protects against virus-induced disease. Stat1^Het and Stat1^KO mice received 10⁴ PFU of MNV CR6 or PBS p.o. and were euthanized on day 28 p.i. (A to E), day 8 p.i. (F to P), or at the humane endpoint. (A) Weight change as a percentage difference from initial, at day 28 p.i. or at humane endpoint. (B) Survival curve of CR6-infected mice reaching humane endpoint, Mantel-Cox test. (C and D) Correlation between MNV CR6 genome copies in colon (C) and spleen (D) and time to humane endpoint in Stat1^Het and Stat1^KO mice, Spearman correlation coefficient test. Data in panels A to D are compiled from two independent experiments with 4 naive Stat1^Het mice, 2 naive Stat1^KO mice, 8 Stat1^Het mice, and 16 Stat1^KO mice. (E) Detection of MNV-specific anti-VP1 IgG in the blood of Stat1^KO mice at endpoint or day 28 p.i. Clearance is defined colonic viral load below LoD at endpoint. Data are compiled from two independent experiments with 2 naive mice, 5 mice not cleared, and 3 cleared mice. (F) Weight change as a percentage difference from initial at day 8 p.i. (G) Survival curve of mice reaching humane endpoint for 8-day CR6-infected mice, Mantel-Cox test. (H and I) MNV CR6 genome copies in colon (H) and spleen (I) were quantified by RT-PCR. Data in panels F to I) are compiled from six independent experiments with 11 naive Stat1^Het mice, 10 naive Stat1^KO mice, 22 Stat1^Het mice, and 23 Stat1^KO mice. (J) MNV CR6 genome copies in Stat1^KO spleen at day 8 p.i. and comparison with weight change as percentage difference from initial, Spearman correlation coefficient test. (K) MNV CR6 genome copies in the brain were quantified by RT-PCR. Data are from one independent experiment. (L and M) Stat1^Het and Stat1^KO mouse spleen weight at day 8 p.i. (L) and comparison with weight change as percentage difference from initial (M), Spearman correlation coefficient test. (N to P) Representative hematoxylin and eosin (H&E)-stained spleen (N), liver (O), and colon (P) sections at day 8 p.i. Scale bar = 200 μm. Images are representative of three independent experiments. Two-tailed Mann-Whitney U test comparing Stat1^Het and Stat1^KO mice was performed unless otherwise noted.

In order to examine mechanisms that could contribute to development of virus-induced disease in these mice, we focused on day 8 post-CR6 exposure, a time point at which the majority of Stat1^KO mice had survived the infection, including a portion (∼25%) that had lost >10% of their original body weight (Fig. 2F and G). At this time point, Stat1^KO mice had higher colonic viral burdens than their littermate controls (Fig. 2H), and viral dissemination to the spleen that correlated with weight loss (Fig. 2I and J). In contrast to previous studies utilizing CW3 infection of Stat1^KO mice, viral burdens were not detectable in the brain of CR6-infected Stat1^KO mice at this time point (Fig. 2K). At day 8 postinfection, a portion of Stat1^KO mice exhibited splenomegaly, and this sign of disease correlated with weight loss (Fig. 2L and M). In Stat1^KO mice that demonstrated >10% weight loss, we observed evidence of virus-induced disease characterized by acute multifocal severe necrosis in the spleen and liver (Fig. 2N and O). Notably, colonic architecture appeared similar between naive and CR6-infected Stat1^KO mice, with no consistent emergence of infection-induced epithelial damage/sloughing or crypt organization (Fig. 2P), indicating that infection-induced disease is related to a loss of control of systemic viral replication.

To further test the hypothesis that systemic viral spread contributes to CR6-induced morbidity and mortality, we infected mice intravenously (i.v.) and monitored clinical symptoms and weight loss. Notably, bypassing the intestinal phase of infection resulted in 100% penetrance of CR6-induced disease in Stat1^KO mice culminating in weight loss requiring euthanasia (Fig. 3A and B) that was accompanied by splenomegaly and disrupted liver architecture (data not shown). We used this system to determine the cellular susceptibility to CR6 infection in Stat1^KO mice. At day 5 postinfection (i.v.), we sorted multiple immune cell populations from Stat1^Het and Stat1^KO littermates and quantified viral genome copies in dendritic cells, macrophages, T cells, and B cells (Fig. 3C). Consistent with previous reports of broad tropism of MNV CW3 following systemic infection (20), these data demonstrate that CR6 can infect a wide range of immune cell types in Stat1^KO mice; however, in Stat1^Het mice, viral replication is restricted. Collectively, these data indicate that systemic viral spread contributes to CR6-induced disease and that STAT1 restricts the ability of CR6 to infect systemic tissues.

FIG 3.

Intravenous infection increases incidence of MNV CR6-induced disease in Stat1^KO mice and leads to infection of multiple immune cell populations. Stat1^Het and Stat1^KO mice received 2 × 10⁵ PFU of MNV CR6 or PBS i.v. and were euthanized at humane endpoint (A and B) or day 5 p.i. (C). (A and B) Weight change (body weight [BW]) as a percentage difference from initial (A) and representation of the percentage of mice with <10% weight loss (B). Data are compiled from 3 independent experiments with 3 Stat1^Het naive mice, 2 Stat1^Het mice, and 11 Stat1^KO mice. One-way analysis of variance (ANOVA) (A) and Mantel-Cox test (B) were used. (C) MNV CR6 genome copies in immune populations sorted from spleens day 5 p.i. were quantified by RT-PCR. Data are representative of those from 3 independent experiments. Two-tailed unpaired t test was used.

MNV CR6-induced disease is independent of antibiotic-mediated microbiome disruption in Stat1^KO mice.

The incidence of severe CR6-induced disease in our Stat1^KO mice is substantially higher than that previously reported by other groups (19, 31), suggesting that environmental conditions may influence disease severity. These data indicated a potential host-virus-microbiota interaction in the development of CR6-induced disease in Stat1^KO mice. Consistent with this hypothesis, we observed a significant difference in the colon-associated microbial populations between Stat1^KO and Stat1^Het mice (Fig. 4A) and neutrophilia in the spleen and colon (Fig. 4C) of Stat1^KO mice following infection, although neither the α- or β-diversity of fecal microbial communities was significantly different between Stat1^Het or Stat1^KO mice following infection (Fig. 4B). These data are consistent with a possible model where virus-host gene interactions alter colon-associated microbial communities which subsequently gain access to the circulation and/or tissues and contribute to the manifestation of clinical symptoms in some Stat1^KO mice.

FIG 4.

MNV CR6-induced disease is independent of antibiotic-mediated microbiome disruption in Stat1^KO mice. Stat1^Het and Stat1^KO mice received 10⁴ PFU of MNV CR6 or PBS p.o. and were analyzed on day 8 p.i. (A) NMDS plot of 16S sequencing data from feces or colonic tissue at day 1 or day 8 p.i. One independent experiment with repeated sampling from 4 CR6-infected Stat1^Het and Stat1^KO mice was performed. Multiple-sample analysis of molecular variance (AMOVA) test was used. (B) Chao1 index of 16S sequencing data from feces or colonic tissue at day 1 or day 8 p.i. One independent experiment with repeated sampling from 4 CR6-infected Stat1^Het and Stat1^KO mice was performed. (C) Flow cytometric detection of Ly6G⁺ CD11b⁺ neutrophils in the spleen (top) and colonic LPLs (bottom). Gated on CD45⁺ B220⁻ CD11c⁻ MHC-II⁻ cells, numbers in flow plots represent frequencies of gated population in total live CD45.2⁺ population ± SD. On the right are frequencies of neutrophils of total CD45⁺ cells in spleen and colon. Data are representative of those from two independent experiments with 3 to 6 mice per group. (D to F) Stat1^KO mice were provided water with Splenda (control) or a cocktail of antibiotics for 2 weeks prior to and throughout MNV CR6 infection. (D) 16S copies in the spleen (left) and liver (right) quantified by RT-PCR. (E) Weight change as a percentage difference from initial body weight. (F) Frequency of CD45⁺ B220⁻ CD11c⁻ MHC-II⁻ Ly6G⁺ CD11b⁺ neutrophils in the spleen in Stat1^KO mice at day 8 p.i. Data are compiled from 2 independent experiments with 2 naive mice and 10 or 11 mice per infected group. Two-tailed Mann-Whitney U test comparing Stat1^Het and Stat1^KO mice was performed unless otherwise noted.

To assess whether virus-induced disease in the context of STAT1-deficiency was related to bacterial translocation, we pretreated Stat1^KO mice with a broad-spectrum cocktail of antibiotics (ABX) that has been shown to reduce the intestinal bacterial biomass (32). However, despite ABX-mediated prevention of microbial translocation (Fig. 4D), the frequency of Stat1^KO mice demonstrating signs of CR6-induced disease or neutrophilia accumulation between control and ABX-treated groups did not significantly differ (Fig. 4E and F). Collectively, these data demonstrate that despite correlations between the microbiota and genetic susceptibility to an enteric commensal-like virus, microbiota depletion via broad-spectrum ABX treatment is insufficient to prevent disease. Notably, we were unable to detect signs of colonization with eukaryotic members of the microbiota, such as fungi or protists, in our colony or animal facility. Together, these data suggest that while microbial translocation may be associated with disease, it is likely a secondary effect.

STAT1 restrains antiviral CD8⁺ T cell responses, but CD8α⁺ T cells do not mediate MNV CR6-induced disease.

In WT mice, CR6 establishes a persistent replicative niche within intestinal tuft cells and evades detection by antigen-specific CD8⁺ T cells (11, 21). While previous work has shown that STAT1 and upstream IFN signaling are essential to restrict CR6 to this immunoprivileged niche (14, 19), it is unknown how STAT1 regulates the antigen-specific CD8⁺ T cell response in either acute or persistent MNV infection. We hypothesized that the loss of STAT1 may result in an altered cytotoxic CD8⁺ T cell response that contributes to disease. At day 8 after infection with CR6 orally, we quantified and phenotyped MNV-specific CD8⁺ T cells isolated from the colonic intraepithelial lymphocyte (IEL) and lamina propria lymphocyte (LPL) compartments and the spleen of Stat1^Het and Stat1^KO littermates. Similar to the case with WT mice (21), there was a small but distinguishable population of MNV-specific CD8⁺ T cells in the colonic IEL and LPL compartments and spleen of Stat1^Het mice (Fig. 5A and B). In the spleen, the frequencies of MNV-specific CD8⁺ T cells were similar between Stat1^Het and Stat1^KO mice, but Stat1^KO mice demonstrated an increased frequency of colonic IEL and LPL MNV-specific CD8⁺ T cells (Fig. 5A and B). Further, stimulation of splenocytes from CR6-infected Stat1^KO mice with immunodominant MNV-derived peptides elicited greater effector cytokine production (IFN-γ and tumor necrosis factor alpha [TNF-α]) but similar levels of production of the cytotoxic mediator granzyme B to that of Stat1^Het mice (Fig. 5C). Together, these data indicate that the loss of STAT1 leads to the generation of a significantly more inflammatory CR6-specific CD8⁺ T cell response. Notably, CD8⁺ T cells isolated from CR6-infected Stat1^Het and Stat1^KO mice both failed to upregulate CD107a and MIP1α (Fig. 5C), supporting previous studies showing that these cells appear to have suboptimal function in the context of CR6 infection (21).

FIG 5.

STAT1 restrains antiviral CD8⁺ T cell responses, but CD8⁺ T cells do not mediate MNV CR6-induced disease. (A to E) Stat1^Het and Stat1^KO mice received 10⁴ PFU of MNV CR6 or PBS p.o. and were analyzed on day 8 p.i. (A and B) Flow cytometric detection of antigen-specific MNV-CR6 Kb-P1^519F tetramer⁺ CD8⁺ T cells. Frequencies ± SEM of CR6 Kb-P1^519F tetramer⁺ cells within the CD8⁺ lymphocyte gate of colonic IELs are represented on flow plots (A). (B) Quantification in colonic IELs, LPLs, and splenocytes. Data in panels A and B are compiled from three independent experiments with 4 to 6 mice per naive group and 12 Stat1^Het and 8 Stat1^KO mice. (C) Total cell counts of IFN-γ⁺, TNF-α⁺, granzyme B⁺, CD107a⁺, and MIP1α⁺ CD8⁺ splenocytes following stimulation with MNV P1^519F peptide. Data are representative of those from three independent experiments with 2 or 3 mice per naive group and 3 or 4 mice per infected group. (D and E) Concentrations of IFN-γ (D) and TNF-α (E) in serum at endpoint, measured by cytometric bead array. Data are compiled from three experiments with 3 or 4 mice per naive group and 8 to 10 mice per infected group. (G to I). Stat1^KO mice received 2 × 10⁵ PFU of MNV CR6 or PBS i.v. and were treated with isotype or anti-CD8 mAbs on days 0, 3, and 6 and analyzed on day 8 post-CR6 infection. (F) Frequency of CD8⁺ cells as a percentage of CD45⁺ splenocytes (left) and colonic IELs (right). (G) Weight change in naive and CR6-infected Stat1^KO mice treated with anti-CD8 mAb or isotype control at day 8 p.i. (H) Flow cytometric detection of B220⁻ CD11c⁻ MHC-II⁻ Ly6G⁺ CD11b⁺ neutrophils in the spleen, represented as a percentage of CD45⁺ cells. (I) MNV CR6 genome copies in colon (left) and spleen (right) were quantified by RT-PCR. Data in panels G to I are from one independent experiment with 3 naive mice and 7 or 8 mice per infected group. Mann-Whitney U test was used.

Previous studies have shown that failure to regulate the magnitude of a CD8⁺ T cell response can contribute to immunopathology (33–35). Notably, the expanded populations of IFN-γ- and TNF-α-producing MNV-specific CD8⁺ T cells in CR6-infected Stat1^KO mice were associated with elevated levels of circulating IFN-γ and TNF-α (Fig. 5D and E). To determine whether this inflammatory population of virus-specific CD8⁺ T cells could contribute to immunopathology, we depleted CD8α⁺ cells (including T cells [36]) following i.v. CR6 infection and assessed clinical scores and viral loads in Stat1^KO mice. Despite efficient depletion of CD8α⁺ cells (Fig. 5F), there was no significant difference in virus-induced weight loss between isotype-treated and anti-CD8α monoclonal antibody (mAb)-treated Stat1^KO mice (Fig. 5G). However, there was a loss of neutrophil accumulation in the spleens of anti-CD8 mAb-treated Stat1^KO mice (Fig. 5H), providing indirect evidence that CR6-induced disease is unrelated to neutrophilia. Notably, depletion of CD8α⁺ cells had no impact on either intestinal or splenic viral burdens (Fig. 5I), indicating that loss of antiviral effector cells does not exacerbate impaired control of viral replication.

STAT1 restricts systemic and mucosal Th17 differentiation.

We also characterized the antiviral CD4⁺ T cell response in Stat1^Het and Stat1^KO mice via tetramer pulldown of MNV-specific CD4⁺ T cells (37, 38). There was no significant difference in the numbers of circulating MNV capsid-specific (tetramer IA^b-P1⁴⁹⁶) CD4⁺ T cells between CR6-infected Stat1^Het and Stat1^KO animals (Fig. 6A and B). Although MNV-specific Stat1^Het and Stat1^KO CD4⁺ T cells each expressed T-bet, the master regulator of Th1 differentiation involved in intracellular pathogen immunity, Stat1^KO CD4⁺ T cells also expressed RORγt (Fig. 6C), the transcription factor that specifies Th17 differentiation and is normally associated with protection from extracellular bacterial or fungal pathogens. Consistent with this, polyclonal stimulation of splenocytes or colonic lymphocytes resulted in expression of IFN-γ in Stat1^Het mice and robust expression of both IFN-γ and interleukin 17A (IL-17A) in Stat1^KO mice (Fig. 6D). These findings were further accompanied by increased il17a gene expression in the spleen and colon and increased circulating levels of IL-17A and IL-22 (Fig. 6E to G and data not shown). Upregulation of IL-22-responsive genes Reg3b and Reg3g, which encode antimicrobial peptides in the intestine, may be suggestive of a compensatory mechanism by which Stat1^KO mice protect the mucosal barrier (Fig. 6H). Together, these data suggest that STAT1 plays an essential role in coordinating CD4⁺ T helper cell differentiation in response to enteric viral infection.

FIG 6.

STAT1 restricts systemic and mucosal Th17 differentiation. Stat1^Het and Stat1^KO mice received 10⁴ PFU of MNV CR6 or PBS p.o. and were analyzed on day 8 p.i. (A) Lymphocytes from spleen and peripheral and mesenteric lymph nodes from naive and CR6-infected Stat1^Het and Stat1^KO mice were pooled and enriched for MHC-II IAb-P1⁴⁹⁶ tetramer⁺ cells (37). Numbers in flow plots represent frequencies ± SEM of MHC-II IAb-P1⁴⁹⁶ tetramer⁺ cells within the CD4⁺ lymphocyte gate. (B) Quantification of total IAb-P1⁴⁹⁶ tetramer⁺ CD4⁺ T cells. (C) Representative Tbet and RORγt expression in IAb-P1⁴⁹⁶ tetramer⁺ CD4⁺ T cells. Data in panels A to C are representative of those from two independent experiments with 4 mice per group. (D) Splenocytes and colonic IELs were stimulated with PMA/ionomycin in the presence of BFA/monensin for 5 h ex vivo and assayed for production of IFN-γ and IL-17A. Numbers in flow plots represent frequencies ± SEM of cells within the CD4⁺ lymphocyte gate. Data are representative of those from three (spleen) or two (colon) independent experiments. On the right are frequencies of IL-17A cells of total CD4⁺ cells in spleen (top) and colon (bottom). Data are pooled from three (spleen) or two (colon) experiments. (E) Expression of il17a was measured in spleen by RT-PCR and are expressed as fold of Stat1^Het values normalized to expression of hprt. RQ, relative quantity. Data are pooled from 3 independent experiments with 4 to 15 mice per group. (F and G) Concentrations of IL-17A (F) and IL-22 (G) in serum at endpoint, measured by cytometric bead array. (H) Expression of Reg3b (left) and Reg3g (right) was measured in colon by RT-PCR and is expressed as fold change in expression from naive Stat1^KO controls, normalized to expression of hprt. Two-tailed Mann-Whitney U test comparing Stat1^Het and Stat1^KO mice was performed unless otherwise noted.

Inappropriate Th17 skewing in Stat1^KO mice is CD4⁺ T cell intrinsic but dispensable for disease.

STAT1 plays a key role in allowing CD4⁺ T cells to receive signals for Th1 differentiation during T cell activation (39). Thus, the observed Th17 differentiation of STAT1-deficient T cells could be due to cell-intrinsic factors. To determine whether STAT1 plays a cell-intrinsic role in CD4⁺ Th1 cell differentiation, enriched Stat1^Het and Stat1^KO CD4⁺ T cells were stimulated in vitro under neutral (anti-CD3/anti-CD28) or Th1-polarizing (anti-CD3/anti-CD28 plus anti-IL-4, IL-2, and IL-12) conditions for 4 days. Under these conditions, there was no impairment in Stat1^KO CD4 T cell proliferation (Fig. 7A). Under neutral conditions, Stat1^Het CD4⁺ T cells produced detectable quantities of Th1-associated cytokines TNF-α and IFN-γ, and this production increased under Th1-polarizing conditions (Fig. 7B); however, under either condition Stat1^Het CD4⁺ T cells failed to produce detectable levels of Th17-associated cytokines IL-17A, IL-17-F, and IL-22 (Fig. 7C). In contrast, Stat1^KO CD4⁺ T cells produced a greater quantity of TNF-α under neutral conditions than did Stat1^Het CD4⁺ T cells and similar amounts under Th1-polarizing conditions, whereas IFN-γ production was reduced in both conditions (Fig. 7B). Strikingly, Stat1^KO CD4⁺ T cells produced Th17-associated cytokines when stimulated under neutral conditions, and although reduced, these cytokines were still produced under Th1-polarizing conditions (Fig. 7C). These data are consistent with the Th17 signatures observed in CR6-infected Stat1^KO mice and suggest that in the absence of STAT1, CD4⁺ T cells are intrinsically polarized toward a Th17-like phenotype.

FIG 7.

Inappropriate Th17 skewing in Stat1^KO mice is CD4⁺ T cell intrinsic but dispensable for disease. (A to C) Naïve CD4⁺ T cells were isolated from the spleens of Stat1^Het and Stat1^KO mice and stimulated in vitro for 4 days with anti-CD3 and anti-CD28 mAbs under neutral (∅) or Th1-skewing conditions. (A) CFSE intensity of CD4⁺ T cells stimulated under neutral conditions. (B and C) Concentrations of Th1-associated cytokines TNF-α and IFN-γ (B) and Th17-associated cytokines IL-17A, IL-17F, and IL-22 (C) in culture supernatant. (D to F) Stat1^KO mice received 10⁴ PFU of MNV CR6 or PBS p.o. and were treated with anti-CD4 mAb or isotype control i.p. on days 0, 3, and 6 p.i. and analyzed on day 8 p.i. (D) Frequency of CD4⁺ cells as a percentage of CD45⁺ splenocytes (left) and colonic IELs (right). (E) Expression of il17a in the colon of Stat1^KO mice treated with anti-CD4 mAb or isotype control, expressed as fold of Stat1^Het values normalized to expression of hprt. (F) Weight change in naive and CR6-infected Stat1^KO mice treated with isotype or anti-CD4 mAb. Data are representative (E) or compiled from (F) those of two independent experiments with 4 or 5 mice per infected group. (G and H) Stat1^KO mice received 10⁴ PFU of MNV CR6 or PBS p.o. and were treated with isotype or anti-IL-17A mAbs on days −1, 2, and 5 and analyzed on day 8 post-CR6 infection. (G) Flow cytometric detection of IL-17A-expressing splenocytes of MNV CR6-infected Stat1^KO mice at day 8 p.i. (H) Weight change in naive and CR6-infected Stat1^KO mice treated with anti-IL-17A mAb or isotype control at day 8 p.i. Data are compiled from two independent experiments with 3 naive mice and 8 to 12 mice per infected group. (I and J) Stat1^Het and Stat1^KO mice received 100 μg of anti-CD20 i.p. The following day, mice received 2 × 10⁵ PFU of MNV CR6 or PBS i.v. Mice also received 400 μg of anti-CD4 and anti-CD8α i.p. at 0, 3, and 6 days postinfection. (I) Flow cytometric detection of CD45.2⁺ TCRβ⁻ B220⁺ CD19⁺ B cells in the spleen of mice which received combination mAb treatment. (J) Weight change in naive and CR6-infected Stat1^KO mice which received combination mAb treatment. Data are compiled from two independent experiments with 2 naive mice and 7 to 9 mice per infected group. Two-tailed unpaired Welch’s t test (B and C) or two-tailed Mann-Whitney U test (D to F and H to J) was performed.

CR6-infected Stat1^KO mice demonstrated an expansion of Th17 CD4⁺ T cells and increased gene expression and serum IL-17A compared to STAT1-sufficient controls (Fig. 6D to F). In both viral and parasitic infections, IL-17A is pathogenic and is associated with morbidity and mortality (40, 41). Therefore, we used antibody-mediated CD4⁺ cell depletion to test whether the Th17-polarized response was causing immunopathology. At day 8 after infection with CR6 orally, Stat1^KO mice treated with anti-CD4 had an absence of CD4⁺ cells in the spleen and colon (Fig. 7D), with reduced colonic il17a gene expression (Fig. 7E), confirming the efficacy of the mAb treatment. However, there was no detectable effect on CR6-induced disease due to CD4⁺ cell depletion in Stat1^KO mice (Fig. 7F). Although the majority of IL-17A-expressing cells were CD4⁺ T cells (Fig. 7G), we postulated that other IL-17A-producing cells might contribute to disease. To test this, we treated Stat1^KO mice with either isotype or neutralizing anti-IL-17A mAb. Again, no difference in clinical manifestation was seen between isotype- and mAb-treated CR6-infected Stat1^KO animals (Fig. 7H). These data are consistent with previous observations that STAT1 has a CD4⁺ T cell intrinsic role in regulating Th1 differentiation (39) and that an inappropriate Th17 response contributes to lung injury in response to influenza virus infection (41). In the context of the viral pathogen LCMV Armstrong, the dysregulated, proinflammatory CD4⁺ T cell response generated in Stat1^KO mice mediates infection-induced mortality (29). However, in the context of this normally asymptomatic, commensal-like intestinal virus, these data indicate that the Th17-polarized CD4⁺ T cell response caused by STAT1 deficiency does not mediate lethal immunopathology.

Despite the failure of antibody-mediated depletion of CD4- or CD8α-expressing cells to rescue Stat1^KO mice from CR6-induced disease, these experiments did not rule out the possibility of each of these populations being sufficient to mediate disease alone. To more holistically investigate whether adaptive immune cells mediate CR6-induced disease, we used a combination antibody depletion strategy to deplete CD4-, CD8α-, and CD20-expressing cells simultaneously in Stat1^KO mice i.v. infected with CR6. Similar to anti-CD4 and anti-CD8α treatments, anti-CD20 mAb treatment efficiently depleted target B cells (Fig. 7I). Despite this, the combination mAb treatment did not rescue Stat1^KO mice from disease (Fig. 7J). Together, these data demonstrate that adaptive immune cells do not mediate CR6-induced disease in Stat1^KO mice.

Clinical symptoms of virus-induced disease and immune dysregulation can be uncoupled by antiviral treatment of Stat1^KO mice.

In the absence of T cell-mediated immunopathology or microbial translocation causing disease, we investigated whether direct inhibition of viral replication was sufficient to prevent disease in CR6-infected Stat1^KO mice. To this end, we treated mice with 2′-C-methylcytidine (2CMC), a viral polymerase inhibitor previously shown to limit MNV replication (42), starting concurrently with CR6 infection (Fig. 8A). Consistent with previous reports, this prophylactic treatment regimen resulted in viral burdens at or below the limit of detection in both colon and spleen (Fig. 8B) and prevented virus-induced weight loss in Stat1^KO mice (Fig. 8C).

FIG 8.

Clinical symptoms of viral-induced disease and immune dysregulation can be uncoupled by antiviral treatment of Stat1^KO mice. (A) Experimental scheme for prophylactic treatment regimen used for panels B and C. Data are compiled from two independent experiments with 6 naive mice and 7 or 8 mice per infected group. (B) CR6 genome copies in colon (left) and spleen (right) were quantified by RT-PCR. (C) Weight change as a percentage difference from initial. (D) Experimental scheme for therapeutic treatment regimen used for panels E to I. Data are compiled from two independent experiments with 4 naive mice and 7 or 8 mice per infected group. (E) CR6 genome copies in colon and spleen were quantified by RT-PCR. (F) Weight change as a percentage difference from initial. (G) Flow cytometric detection of splenic antigen-specific MNV-CR6 Kb-P1^519F tetramer⁺ CD8⁺ T cells isolated from colonic IELs. Numbers in flow plots represent frequencies ± SEM of CR6 Kb-P1^519F tetramer⁺ cells within the CD8⁺ lymphocyte gate. (H) Quantification of MNV CR6-specific Kb-P1^519F tetramer⁺ cells as a frequency of CD45⁺ CD8⁺ IELs. (I) Expression of il17a was measured in spleen by RT-PCR and is expressed as fold of Stat1^Het values normalized to expression of hprt. Data are representative of those from two independent experiments. Two-tailed Mann-Whitney U test was performed unless otherwise noted.

We then used a modified treatment regimen and started treatment with 2CMC at day 3 postinfection (Fig. 8D). We postulated that this therapeutic regimen would allow early stages of the infection to proceed and activate innate and adaptive immune responses, allowing us to interrogate whether these immune responses contribute to disease while limiting the confounding factor of viral replication. This modified approach of therapeutic antiviral treatment was associated with significantly reduced viral burdens in Stat1^KO mice; in the colon, burdens of 2CMC-treated mice were at or near the limit of detection, and 2CMC treatment resulted in a significant reduction in splenic viral burden compared to that in control-treated littermates (Fig. 8E). Notably, through day 8 p.i., therapeutic 2CMC treatment prevented CR6-induced weight loss in Stat1^KO mice (Fig. 8F). Consistent with the lack of clinical signs, therapeutic 2CMC treatment prevented CR6-induced splenomegaly and liver and spleen necrosis associated with CR6 infection of Stat1^KO mice (data not shown). Notably, this occurred despite detectable viral loads still being observed in the spleens of Stat1^KO mice, suggesting the existence of a threshold of viral burden beyond which disease occurs. Finally, we addressed whether 2CMC treatment also impacted the inflammatory antiviral CD4⁺ Th17 and CD8⁺ T cell responses that are associated with CR6-induced disease in Stat1^KO mice. Although we observed a significant reduction in viral burdens, the frequency of MNV-specific CD8⁺ T cells and splenic il17a gene expression were equivalent between vehicle- and 2CMC-treated Stat1^KO mice (Fig. 8G to I). Thus, despite complete protection from CR6-induced disease and significantly improved viral control, the signatures of an inflammatory, nonprotective immune response were still present in 2CMC-treated Stat1^KO mice. Collectively, our data uncouple virus-induced disease from the systemic inflammatory immune responses which occur in the absence of STAT1, suggesting that neither immunopathology nor secondary bacterial translocation is directly involved in CR6-induced disease. Instead, STAT1-dependent pathways maintain the commensal-like relationship between host and CR6 by restricting CR6 replication within mucosal tissues, thereby limiting damage to systemic organs caused by unchecked viral replication.

DISCUSSION

Recent work demonstrating the ability of the virome to both positively and negatively influence our health has underscored the importance of elucidating how these relationships are regulated by the host (7, 43). The mammalian virome includes viruses that have evolved distinct strategies for long-term persistence, such as latency, genomic integration, and/or immune evasion. In both mice and humans, noroviruses can establish long-term persistence without latency. Experimental infection of volunteers has demonstrated that human norovirus (huNoV) can establish persistent infection for weeks or months after resolution of clinical symptoms, and despite high viral loads, some patients never experience clinical symptoms (44). These results, together with the demonstration that MNV can confer host-protective effects in mice (7, 8), suggest that norovirus infection fits the profile of mutualism; however, the host mechanisms facilitating this relationship are not well understood (8). Here, we describe profound immune dysregulation following MNV CR6 infection of STAT1-deficient mice. However, unlike other pathogenic viral infections in which immune-mediated pathology contributed to morbidity and mortality, we demonstrate that controlling viral replication via treatment with a viral polymerase inhibitor is sufficient to abrogate disease, even in the presence of inflammatory immune responses.

In WT mice, CR6 infection is limited to the colonic epithelium, where it infects tuft cells expressing the viral entry receptor, CD300lf (11). However, a recent report demonstrated that the closely related MNV strain CW3 infects multiple immune subsets in the spleen, implying that CD300lf is also broadly expressed in systemic tissues (20). This is consistent with data from us and others demonstrating that in CR6-infected Stat1^KO mice, high viral burdens can be found in multiple peripheral tissues, including the spleen and liver. Furthermore, a previous study has shown that a conditional knockout of Stat1 in hematopoietic lineages is sufficient to permit viral dissemination in mice orally infected with CR6 (19), and conditional knockout of Stat1 in CD11c-expressing cells is sufficient to permit systemic persistence of CW3 (45). These data suggest a model of viral dissemination in which the absence of STAT1 renders multiple immune cell populations unable to control intrinsic CR6 replication, resulting in acquisition of infection in gut-associated lymphoid tissues and systemic dissemination. Consistent with this model, we observed that in the absence of STAT1, high viral genome copy numbers were detectable in T cells, B cells, macrophages, and dendritic cells following parenteral CR6 infection; in contrast, viral genomes in these populations were at or below the limit of detection in Stat1^Het littermates. In this context, systemic dissemination of CR6 and a subsequent failure to control this infection in tissues, including the liver and spleen, could lead to severe tissue damage and underlie the observed mortality in Stat1^KO mice.

Consistent with a previous report (46), CR6 infection had no effect on luminal or colon-associated microbial communities in STAT1-sufficient animals. However, maintenance of colonic-tissue-associated community structure was STAT1 dependent, which we considered indicative of potential bacterial translocation that could contribute to CR6-induced pathology. However, depletion of intestinal bacterial biomass prevented bacterial translocation but had no effect on CR6-induced disease in Stat1^KO mice. While these results do not eliminate the possibility that microbiota constituents (bacterial or otherwise) contribute to the observed disease, they suggest that bacterial translocation is secondary to virus-induced morbidity.

Notably, we observed a significantly higher rate of severe disease following oral CR6 infection in our Stat1^KO mice than in previous reports (19, 31); at day 21 p.i., approximately 50% of our mice had reached humane endpoint, in contrast to previous reports showing approximately 20% of mice succumbing at this time point. These data could suggest the presence of an environmental factor that contributes to protection from disease that is lacking in our facility. However, the nature of this difference remains unclear. The presence of a particular strain of astrovirus (STL5) has been demonstrated to confer resistance to intestinal CR6 infection via upregulation of IFN-λ, but this study did not investigate astrovirus-dependent effects on systemic infection (13). Studies to identify the nature of the environmental factor(s) (e.g., bacterial and nonbacterial microbiota constituents or diet) that contributes to differences in CR6-induced disease are ongoing.

Previous reports have also documented disease in B6/Stat1^KO mice upon infection with strains of MNV other than CR6 (31). Strains MNV-O7 and MNV4 form persistent infections similar to MNV CR6 in WT mice, but neither causes mortality in Stat1^KO mice (16, 17). Furthermore, MNV4 induces TNF-α- and IFN-γ-independent colonic inflammation in Stat1^KO mice, while in our facility colonic architecture was maintained in CR6-infected Stat1^KO mice. These data indicate that strain-specific differences may provide useful insights into viral determinants of disease. Notably, CR6 infection of Stat1^KO mice significantly upregulated expression of epithelium-protective factors such as IL-22 and the antimicrobial peptide-encoding genes Reg3b and Reg3g, which may contribute to maintenance of intestinal homeostasis. Consistent with this hypothesis, CR6 colonization fortifies the intestinal barrier of WT GF mice via IL-22-mediated STAT3 activation and upregulation of Reg3β (8).

STAT1 transduces signals from type I, II, and III interferons, and previous studies have investigated the individual roles of these signals in the context of CR6. Despite only weakly inducing type I interferon, CR6 infection of Ifnar^−/− mice leads to systemic viral dissemination. In contrast, loss of type III interferon signaling leads to increased viral shedding, but not dissemination, while loss of type II interferon signaling has no documented effect on CR6 infection (14). Finally, combined deficiency of type I and II IFN signaling results in increased viral shedding (42). Notably, the loss of any one of these signaling pathways in isolation does not lead to increased morbidity and mortality; however, we hypothesize that in the case of Stat1^KO mice, the combined loss of type I and type III signaling leads to enhanced susceptibility of systemic tissues to infection, resulting in a loss of control of systemic viral replication and eventual morbidity and mortality.

Our data demonstrate the novel finding that 2CMC remains efficacious against CR6 when administered therapeutically in a severely immunocompromised model. Given the risks posed by noroviruses to immunocompromised humans, this is further evidence for the potential usefulness of targeted antiviral treatment in the clinic. Additionally, STAT1 suppression is an off-target effect of the purine analog fludarabine, a chemotherapeutic agent used in the treatment of leukemia and lymphoma (47). Patients treated with fludarabine are at an elevated risk of opportunistic viral infection, and our data support the hypothesis that fludarabine-mediated suppression of STAT1 may contribute to this risk of infection. Together, our data highlight the importance of STAT1 in regulating ongoing—and often subclinical—chronic viral infections.

MATERIALS AND METHODS

Mouse strains and housing conditions.

Mice were housed at the University of British Columbia (UBC) specific-pathogen-free Centre for Disease Modeling. MNV-free C57BL/6 Stat₁^+/− mice (48) [The Jackson Laboratory; strain B6.129S(Cg)-Stat1^tm1Dlv/J (012606)] were bred as either Stat1^+/− × Stat1^+/− or Stat1^+/− × Stat1^−/−. To control for microbiome and other environmental effects (49), resulting Stat1^WT, Stat1^Het, and Stat1^KO littermates were used. If an individual experiment required multiple litters, age-matched STAT1-sufficient and -deficient mice from each litter were included. Mixed genotypes were cohoused before and throughout the experiment. Replicate experiments were performed on litters from multiple dams. Experimental mice were aged matched within each experiment, and were 6 to 12 weeks of age for all experiments.

All experiments were performed according to guidelines from UBC Animal Care Committee and Biosafety Committee-approved protocols. Mice were housed in ventilated Ehret cages prepared with BetaChip bedding and had ad libitum access to irradiated PicoLab diet 5053 and reverse osmosis/chlorinated (2 to 3 ppm)-purified water. Housing rooms were maintained on a 14/10-h light/dark cycle with temperature and humidity ranges of 20 to 22°C and 40 to 70%, respectively. Sentinel mice housed in experimental rooms were maintained on dirty bedding and nesting material and were tested on a quarterly basis for presence of mites (Myobia musculi, Myocoptes musculinis and Radfordia affinis, pinworm (Aspiculuris tetaptera and Syphacia obvelata), fungi (Encephalitozoon cuniculi), bacteria (Helicobacter spp., Clostridium piliforme, Mycoplasma pulmonis, and CAR bacillus), and viruses (Ectromelia virus, epizootic diarrhea of infant mice [EDIM]/rotavirus, murine hepatitis virus [MHV], MNV, metapneumovirus [MPV], minute virus of mice [MVM], LCMV, mouse adenovirus 1/2 [MAV1/2], murine cytomegalovirus [MCMV], polyomavirus, pneumonia virus of mice [PVM], reovirus type 3 [REO.3], Sendai virus, and Theiler’s murine encephalomyelitis virus [TMEV]).

Infection, monoclonal antibody, antibiotic, and antiviral treatments.

Mice were infected by oral gavage (per os [p.o.]) or intravenously (i.v.) by tail vein injection with 10⁴ (p.o.) or 2 × 10⁵ (i.v.) PFU of MNV CR6 diluted in sterile phosphate-buffered saline (PBS). Naive mice were sham infected with sterile PBS (Sigma). Mice were monitored every 24 to 48 h and euthanized at the humane endpoint (loss of ≥20% of initial body weight). In rare cases, true infection-induced mortality was seen in mice that appeared ill but had not yet reached humane endpoint (despite provision of supportive care, mice were found dead at the next health check).

To deplete CD4⁺ or CD8α⁺ cells, 400 μg of either anti-CD4 (clone GK1.5; UBC AbLab) or anti-CD8α (clone 53.67; UBC AbLab) antibody was administered intraperitoneally (i.p.) at days 0, 3, and 6 post-MNV CR6 infection. To deplete CD20⁺ cells, 100 μg of anti-CD20 (clone MB20-11; Bio X Cell) was administered i.p. 1 day prior to infection. Cellular depletion of CD4⁺ and CD8⁺ cells in spleens and intraepithelial lymphocyte compartments was monitored by flow cytometric detection with anti-CD4 BV650 (clone RM4-5; BD Biosciences) or anti-CD8β (clone H35-17.2; Thermo Scientific). To neutralize IL-17A, 500 μg of anti-IL-17A (clone 17F3; Bio X Cell) was administered i.p. at days −1, 2, and 5 of MNV CR6 infection. For all mAb treatment experiments, antibodies were prepared in sterile PBS (Sigma) and IgG1 isotype control antibodies (clone MOPC-21; Bio X Cell) were administered with the same dosing regimen and schedule as the treatment. For antiviral treatment, 1 mg of 2CMC was administered i.p. daily beginning as specified. Control mice received a vehicle control of PBS following an identical dosing regimen.

For antibiotic administration, water supplemented with 0.25 mg/mL of vancomycin, 0.5 mg/mL of neomycin, 0.5 mg/mL of gentamicin, 0.5 mg/mL of ampicillin, 0.25 mg/mL of metronidazole, and 4 mg/mL of artificial sweetener (Splenda) was provided ad libitum (32). Control mice received sterile drinking water supplemented with 4 mg/mL of Splenda.

Lymphocyte recovery and ex vivo lymphocyte stimulation.

Single cell suspensions of splenic lymphocytes were prepared by mashing tissue and passing cells through a 70-μm nylon mesh filter followed by ammonium-chloride-potassium (ACK) lysis to remove red blood cells. Intestinal lymphocytes were harvested from the cecum and colon as previously described (21). For peptide-specific cytokine responses, 2 × 10⁶ lymphocytes were stimulated with 0.4 μg/mL of major histocompatibility complex (MHC) class I-restricted peptides (Kb-P1^519Y) (21) for 5 h at 37°C in the presence of 10 μg/mL of brefeldin A (BFA; Sigma) and GolgiStop (BD Biosciences). For polyclonal cytokine responses, lymphocytes were stimulated with 0.1 μg/mL of phorbol myristate acetate (PMA) and 1 μg/mL of ionomycin (Sigma) in the presence of BFA and GolgiStop for 5 h at 37°C. To detect MNV-specific CD4⁺ T cells, cells from spleen and peripheral and mesenteric lymph nodes were pooled, passed through a 70-μm nylon mesh filter to create a single cell suspension, labeled with tetramers specific for MNV CW3 capsid P1⁴⁹⁶, and subjected to magnetic bead enrichment, as described previously (37, 38).

Tissue histology.

Samples were flushed with PBS, fixed in 4% paraformaldehyde, and embedded in paraffin. Sections (5 μm) were cut and used for hematoxylin and eosin staining (Histochemistry Service Laboratory at the University of British Columbia). Histology images were acquired with Zeiss Axio Observer 7 and an AxioCam 105 microscope camera. Qualitative assessment of tissue pathology was performed unblinded by an experienced veterinary pathologist comparing to samples from naive littermate controls.

Bacterial DNA extraction, amplification, and iTag sequencing.

DNA was extracted from liver, spleen, colon tissue, and fecal pellets using DNeasy PowerSoil kit (Qiagen) as per the manufacturer’s instructions. Resulting DNA was stored at −20°C. Liver and spleen 16S ribosomal DNA (rDNA) amplification was measured via SYBR incorporation detected on a Quant Studio 3 real-time PCR system (Applied Biosystems). Quantification was conducted using the PicoGreen assay (Invitrogen) for double-stranded DNA (dsDNA) (50) measured on a TECAN M200 plate reader. Bacterial and archaeal small subunit (SSU) rRNA gene fragments from the extracted genomic DNA were amplified using primers 515F and 806R. Sample preparation for amplicon sequencing was performed as described previously (51, 52). The amplicon library was analyzed on an Agilent Bioanalyzer using a high-sensitivity dsDNA assay to determine approximate library fragment size and to verify library integrity. Pooled library concentration was determined using the KAPA library quantification kit for Illumina. Library pools were diluted to 4 nM and denatured into single strands using fresh 0.2 N NaOH as recommended by Illumina. The final library was loaded at a concentration of 8 pM, with an additional PhiX spike-in of 5 to 20%. Sequencing was conducted through the UBC Sequencing and Bioinformatics Consortium (https://sequencing.ubc.ca/).

Bioinformatic analysis.

Sequences were processed using Mothur (53). In brief, sequences were removed from the analysis if they contained ambiguous characters, had homopolymers longer than 8 bp, and did not align to a reference alignment of the correct sequencing region. Unique sequences and their frequency in each sample were identified and then a preclustering algorithm was used to further denoise sequences within each sample (54). Unique sequences were identified and aligned against a SILVA alignment (55) (available at https://www.arb-silva.de/aligner/). Sequences were chimera checked using UCHIME (56), and reads were then clustered into 97% operational taxonomic units (OTUs) using OptiClust (57). OTUs were classified using the SILVA reference taxonomy database (release 132; available at https://mothur.org/wiki/silva_reference_files/#release-132). All data were visualized in R and Excel. For α- and β-diversity measures, all samples were subsampled to the lowest coverage depth and standard indices were calculated in Mothur. Community structure was investigated using the Yue and Clayton similarity estimator, nonmetric multidimensional scaling (NMDS) was used to compare bacterial community structures across all samples, and a multiple-sample analysis of molecular variance (AMOVA) was used to test the significance of differences between microbial communities.

Statistical analysis.

Statistical analyses were performed using GraphPad Prism (GraphPad Software; version 6.0), with two-tailed nonparametric Mann-Whitney U test, two-tailed unpaired Welch’s t test, Mantel-Cox or Spearman correlation coefficient, as appropriate (specified in figure legends). Values in flow cytometry plots are reported as means ± standard deviations. Bars on graphs indicate medians. Statistically significant differences are shown in figures as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; and ns, not significant.

Data availability.

Raw sequencing data are available under NCBI BioProject accession number PRJNA798435.