Pathology of Immunodeficient Mice With Naturally Occurring Murine Norovirus Infection

Abstract

Murine norovirus (MNV) was recently discovered in Rag2 ^−/−/Stat1 ^−/− mice in a U.S. medical research facility. Presently, little is known concerning the epidemiology and natural history of this virus. We studied the pathology of naturally occurring MNV infection in 28 immunodeficient mice of several different genotypes (Rag1 ^−/−/IFNγ R^−/−, OT1 Rag1 ^−/−/IFNγ R^−/−, OT2 Rag1 ^−/−/IFNγ R^−/−, Rag1 ^−/−/Stat1 ^−/−, and Rag2 ^−/−) that were maintained in two U.S. research facilities. The mice were selected for study because sentinel mice housed in their holding rooms had been identified as positive for MNV-specific antibodies during routine screening for infectious agents. Our data indicate that in certain lines of immunodeficient mice, MNV can establish a disseminated infection that is characteristically associated with inflammation in multiple tissues, including liver (hepatitis), lung (focal interstitial pneumonia) and the peritoneal and pleural cavities. In addition, MNV can establish an asymptomatic infection in the mesenteric lymph nodes of Rag2 ^−/− mice. Further studies are needed to determine whether MNV presents a confounding variable in immunological, toxicological and pathological studies in mice naturally infected with MNV.

[The supplemental data referenced in this paper is not printed in this issue of Toxicologic Pathology. It is available as a downloadable file in the online edition of Toxicologic Pathology, 34(6). In order to access the full article online, you must have either an individual subscription or a member subscription accessed through www.toxpath.org.]

Introduction

Noroviruses, members of the family Caliciviridae, have been associated with infection and disease in humans, pigs, cattle, and mice (Green et al., 2000; Karst et al., 2003). The first murine norovirus (strain MNV-1) was identified in immunodeficient laboratory mice (Karst et al., 2003) and was subsequently shown to infect murine macrophage and dendritic cells in tissue culture (Wobus et al., 2003, 2006). The virus may be widespread. It was recently reported that approximately 22% of 12,639 serum samples examined from mice in North American research facilities had antibodies that reacted with MNV-1 (Hsu et al., 2005). MNV-1 infection was shown to be lethal in certain immunodeficient mice with impaired innate immunity, and included mice with genotypes Rag2 ^−/−/Stat1 ^−/−, Stat1 ^−/−, Stat1 ^−/−/Pkr ^−/− and IFNαβγ R ^−/− (Karst et al., 2003). However, the virus was not associated with illness or death in experimentally infected wild-type mice, or in certain other immunodeficient mice such as Rag1 ^−/− or Rag2 ^−/− (Karst et al., 2003). Immunohistochemical studies showed that MNV-1 infection was associated with cells of macrophage and dendritic cell-like morphology in the liver and spleen tissues of experimentally challenged Stat1 ^−/− mice (Wobus et al., 2004). Furthermore, it was reported that MNV-1 could be detected by RT-PCR in spleen, mesenteric lymph nodes, and jejunum of outbred mice at five weeks following oral challenge (Hsu et al., 2005).

The detection of murine norovirus in laboratory mice has raised concerns for its effect on biomedical research. The purpose of this study was to analyze the pathology of MNV in laboratory mice that may have been infected by natural exposure to the virus, most likely by the introduction of the virus from an outside source into the animal facility. We euthanized immunodeficient mice of various genotypes at two U.S. medical research facilities and examined selected tissues for histopathological lesions and expression of MNV antigen. The MNV-positive immunodeficient mice (with the exception of Rag2 ^−/−) examined in this study had characteristic inflammatory lesions including hepatitis, focal interstitial pneumonia, peritonitis and pleuritis with evidence of macrophage infection. The implications of these findings for medical research are discussed.

Materials and Methods

Pathology

Various tissues were obtained from 14 MNV-positive 1–3 month-old male or female immunodeficient mice, on a C57BL/6 background, at facility 1 (Washington University School of Medicine, St. Louis, MO), including 6 Rag1 ^−/−/IFNγ R^−/− mice, 5 Rag1 ^−/−/Stat1 ^−/− mice, 2 OT1 Rag1 ^−/−/IFNγ R^−/− mice and 1 OT2 Rag1 ^−/−/IFNγ R^−/− mouse (Braaten et al., 2005; Sparks-Thissen et al., 2004, 2005) and at facility 2 (NIAID, NIH, Bethesda, MD) from fourteen MNV-positive 3–5-month-old male and female Rag2 ^−/− mice, also on a C57BL/6 background (Shinkai et al., 1992) (Table 1). The mice in both facilities were housed in microisolator cages with standard bedding and diets under animal study protocols approved by Institutional Animal Care and Use Committees of each institution. All mice in this study originated from animal rooms in which antibodies to MNV had been detected in sentinel mice. In facility 2, there was a high (up to 80%) prevalence of antibodies to MNV in sentinel mice (female Swiss-Webster) that were housed in the same animal room as the Rag2 ^−/− mice for several months, as detected by a microsphere-based serological multiplexed fluorescent immunoassay conducted at the University of Missouri Research Animal Diagnostic Laboratory (Hsu et al., 2005). Sentinel mice in the rooms at both facilities were negative for antibodies to other known mouse pathogens (Sendai virus, reovirus type 3, Theiler’s mouse encephalomyelitis virus, mouse hepatitis virus, ectromelia virus, epizootic diarrhea of infant mice, mouse cytomegalovirus, polyoma virus, pneumonia virus of mice, K virus, orphan parvovirus and mouse adenovirus). Helicobacter sp. DNA was not detected by PCR in the feces of sentinel mice in the room for 8 of the 14 mice at facility 1, but was detected in feces of sentinel mice in the room that housed 6 of the 14 mice. In facility 2, Helicobacter sp. was detected in sentinel mice housed in the same room and in the cecal contents of some of the mice in this study. Mice were euthanized with carbon dioxide. Necropsies were performed on all 28 mice at their respective facilities. Tissues were fixed in 10% neutral-buffered formalin, embedded in paraffin and sections were stained with hematoxylin and eosin. Steiner silver stain was performed on selected liver sections with lesions (1–3 liver sections per mouse) for detection of Helicobacter species.

Immunohistochemistry

Selected tissues were examined by immunohistochemistry (IHC). The primary MNV anti-viral antibodies used in the IHC included those raised against: (1) CsCl-purified MNV-1 particles, primarily recognizing MNV-1 capsid protein (VP1) (Wobus et al., 2004) in rabbits (serum diluted at 1:5000); (2) recombinant MNV-1 capsid protein (rVP1) in guinea pigs described below (serum diluted at 1:2000) or (3) recombinant MNV-1 proteinase-polymerase protein (ProPol) in guinea pigs (serum diluted at 1:250) (Sosnovtsev et al., 2006). Antigen retrieval from sections of fixed tissues for analysis of reactivity with the rabbit and guinea pig antibodies was performed in a food steamer (Sunbeam, Boca Raton, FL) (Ramos-Vara, 2005) with 0.01M, pH 6.0 molar citrate buffer (Biocare, Concord, CA) or Trilogy solution (Cell Marque, Hot Springs, AR), respectively. The Dako polymer rabbit ENVISION kit (Dako Corp, Santa Barbara, CA) and the Vector Guinea pig ABC kit (Vector Labs, Burlingame, CA) were used to visualize antigen-positive cells. Diaminobenzidine was used as the chromogen with hematoxylin as the counterstain. To generate a positive and negative control for immunohistochemistry, RAW 264.7 cells (ATCC, Manassas, VA) were infected with MNV-1 at a multiplicity of infection of 1 or mock-infected and collected after 24 or 48 hours. Cells were pelleted by centrifugation, fixed in formalin, embedded in paraffin, sectioned, and used as a positive and negative control for immunohistochemistry (Figures 1A–B, S1A–B). Other negative controls included the use of pre-immunization rabbit or guinea pig serum and the inclusion of tissues from uninfected mice. In addition, all reagents were tested in the absence of primary antibodies. Antigen staining was not observed in the negative controls.

Double stains for viral and mononuclear phagocyte antigens were performed on spleen, liver and mesenteric lymph nodes from selected mice. Antigen retrieval was performed utilizing a food steamer (model 5713, Oster) with DIVA buffer (BioCare, Walnut Creek, CA). To detect MNV viral antigen, we used the antibody to recombinant MNV-1 proteinase-polymerase protein (ProPol) as noted above at a 1:250 or 1:500 dilution with a HRP detection system and Vector Red Chromogen (PK-4007, Vector Labs. Burlingame, CA). To detect antigens specific for cells of the mononuclear phagocyte series, we used goat anti-mouse CD40 (T-20, sc-1731, Santa Cruz Biotechnology, Santa Cruz, California) at a 1:50 dilution with a HRP detection system and DAB/Nickel chromogen. In addition, a biotinylated rat anti-mouse F4/80 antibody (MCA497BB, Serotec, Oxford, UK) was used at a 1:50 dilution with an alkaline phosphatase detection system and Ferenge Blue Chromogen (AP506US, Biocare Medical, Walnut Creek, California).

RT-PCR Detection of MNV

Mesenteric lymph node or duodenal tissue was placed into 0.5 ml sterile PBS and the cells lysed with a Dounce homogenizer on ice. An aliquot (100 μl) was removed for the extraction of total RNA with the RNeasy kit (Qiagen, Valencia, CA). The detection of MNV-specific RNA was performed by RT-PCR with the One-step RT-PCR Kit (Invitrogen, Carlsbad, CA) and primer pair 5′-GTTCTCCTTCTATGGTGATGACG-3′ and 5′-GCTGG-CGGTCGATGCTGGCACG-3′ that would amplify nucleotides 4556–4788 of the MNV genome. DNA products were analyzed by electrophoresis in a 1% agarose gel and bands corresponding to the expected size were excised, purified, and subjected to sequence analysis with reagents in the BigDye Terminator Cycle Sequencing kit (Applied Biosystems, Foster City, CA) (using each of the RT-PCR primers). The identity of the product as MNV-specific was verified by a BLAST search against the GenBank database.

Fecal material was placed into 1 ml sterile PBS and homogenized using a Minibeadbeater (Biospec products). The fecal homogenate was clarified by centrifugation, passed through a 0.22 μM filter and inoculated onto RAW264.7 cells. After 4 days, the RAW264.7 cells were frozen at −80°C. The thawed virus was clarified by centrifugation and was used to inoculate RAW 264.7 cells. Three days later, total RNA was extracted from infected RAW264.7 cells using Trizol (Invitrogen) and the cDNA was synthesized using SuperScriptIII reverse transcriptase (Invitrogen) and an oligo d(T) primer. The detection of MNV-specific sequences was performed by PCR with primer pair 5′-GCCTCCGCTGCTACTGTAGG-3′ and 5′-CCCGGGAAGCCACAGTCC-3′ that would amplify nucleotides 2416–4929 of the MNV genome. DNA products were subjected to sequence analysis and the identity of the product as MNV-specific was confirmed as described previously.

Cloning, Expression and Purification of the Recombinant MNV Capsid Protein and Production of Capsid (VP1)-Specific Antiserum: Antiserum specific for recombinant (r) MNV-1 capsid protein (anti-rVP1) was prepared as follows. The ORF2 sequence (nt 5056–6681) of the virus genome was PCR-amplified from the full-length MNV-1 cDNA clone p20.3 (Sosnovtsev et al., 2006) as template and cloned into bacterial expression vector pET-28b (Novagen). Oligonucleotides 5′-tatattttaacgtctcacATGAGGATGAGTGATGGCGCAGCG-3′ and 5′-tattatttttactcgagTTGTTTGAGCATTCGGCCTGTTGC-3′ were employed to amplify the corresponding DNA fragment. The first oligonucleotide corresponded to the beginning of the ORF2 sequence (nt 5056 to 5079 of the MNV-1 genome, large case) and included a BsmBI site (underlined). The second oligonucleotide contained the sequence complementary to the end of the ORF2 (nt 6655 to 6678, large case) with engineered XhoI site (underlined). Purified PCR fragment was treated with BsmBI and XhoI and ligated into NcoI-XhoI-linearized pET-28b vector. The resulting plasmid (pCAPHis) contained the ORF2 sequence fused to the C-terminal His₆-tag and placed under control of the T7 RNA polymerase promoter downstream from the bacterial ribosome-binding site. The recombinant VP1 protein was expressed in Escherichia coli BL21 (DE3) cells (Novagen) transformed with the pCAPHis. The cells were grown in the presence of kanamycin (30 μg/ml) in LB medium at 37°C. When the A₆₀₀ of the culture reached 0.6, expression was induced by the addition of IPTG at a concentration of 1 mM. The recombinant VP1 protein was purified from the insoluble fraction of the induced cells by immobilized-metal affinity chromatography and used for immunization of guinea pigs as described previously (Sosnovtsev and Green, 2000; Sambrook and Russell, 2001).

Expression of the MNV-1 ProPol (proteinase-polymerase) protein and production of ProPol-specific antiserum was conducted in a similar manner (Sosnovtsev et al., 2006).

Results

We studied 28 mice at 2 research facilities for clinical signs, histopathology and immunohistochemistry associated with MNV infection. The designation, facility of origin, and strain of each mouse studied is shown in Table 1. To confirm MNV infection, virus was cultured from feces and identified by RT-PCR in 6/6 of the mice in facility 1 for which samples were available (Table 1). MNV-specific RNA was identified by RT-PCR in 13/13 of the mice in facility 2 for which samples were available (Table 1). Mice with no evidence for MNV infection by RT-PCR or immunohistochemistry were excluded from this study.

Clinical Data

Immunodeficient mice in this study that were housed at facility 1 and shown to be naturally infected with MNV exhibited clinical symptoms as they aged. Clinical symptoms included body weight loss, ruffled fur, and hunched backs that typically developed by 2–3-months of age or shortly before animals were necropsied. Clinical signs in the analyzed Rag1 ^−/−/IFNγ R^−/− , Rag1 ^−/−/Stat1 ^−/−, OT1Rag1 ^−/−/IFNγ R^−/− and OT2 Rag1 ^−/−/IFNγ R^−/− mice were associated with histopathological lesions (Table 1 and see next). The Rag2 ^−/− mice in this study that were housed at facility 2 showed no clinical symptoms.

Pathology

At necropsy of mice in facility 1, tissues appeared grossly normal, except for Rag1 ^−/− Stat1 ^−/− mice, which had splenomegaly and/or pale spots in the liver. All 14 mice at facility 1 had varying degrees of hepatitis (Figures 2–7, S2–3), characterized by mild-to-severe diffuse and focal inflammatory infiltrates, which appeared to be primarily composed of mononuclear cells and some neutrophils (Figure 3). Helical bacteria were not seen in the liver lesions with Steiner stain. Vasculitis with adhesion of leukocytes to hepatic (Figures 5 and S2–3) and pulmonary veins was sometimes observed. The most severely affected livers had loss of hepatocytes in periportal areas (Figures 2 and 5). Focal interstitial pneumonia (Figures S4–5) was seen in most mice but was usually mild. It was characterized by a focal, or multifocal interstitial infiltration of macrophage-like cells in alveoli and alveolar walls, with variable numbers of neutrophils present. No lesions of Pneumocystis murina were seen in the lungs of any of the mice studied. Peritonitis (Figure S6–7) often with pleuritis was observed in 8/14 mice. Cecal protozoa were often present, but not associated with inflammation of the gut.

At necropsy of mice in facility 2, no visible tissue abnormalities were observed, except for variably thickened ceca in most mice. Examination of the tissue showed that the mice often had typhlitis, which may be due to the Helicobacter sp. present in the colony (Ward et al., 1996) and protozoa present in the cecum. However, no other lesions were observed that could be associated with viral infection in any tissue.

Mesenteric lymph nodes of most Rag1 ^−/− or Rag2 ^−/− mice in both facilities often had many apoptotic cells within the paracortex with the usual lymphoid hypoplasia seen in immunodeficient mice. In addition, focal fibrosis was observed in the mesenteric nodes of 2 of the 6 Rag1 ^−/−/Stat1 ^−/− mice and 1 Rag1 ^−/−/IFNγ R^−/− mouse had a mild inflammatory lesion in the node.

Immunohistochemistry

To identify whether MNV antigen was present in lesions, immunohistochemistry was performed using antibodies directed against MNV-1 structural (VP1) or nonstructural (ProPol) proteins. Immunohistochemistry analysis for viral antigens in mice from facility 1 revealed the presence of cytoplasmic MNV antigens in inflammatory cells of the liver in all cases of hepatitis (Figures 4 and 6), the red and white pulp of the spleen (Figure 8), the lamina propria of the small intestine and intestinal lymphoid foci (Figures S8–9), lesions of the lung (Figure S5), in peritonitis and pleuritis (Figure S6), and in the mesenteric lymph nodes (Figures 11 and 12). The same lesions did not show reactivity with the pre-immunization serum (illustrated in Figures 7, S7, S10). The numbers of antigen-positive cells varied among the samples examined. All 3 MNV-specific antibodies stained cells in all tissues that were immunoreactive for viral antigens, although the degree of reactivity varied within the same tissue. The guinea pig anti-rVP1 antibody often showed the highest intensity of staining. In all mice examined from facility 1, cells with macrophage-like morphology in focal inflammatory hepatic lesions and Kupffer cells expressed MNV antigens. Double staining of the liver lesions with F4/80, a pan-macrophage marker, and antibodies specific for MNV ProPol, showed that the majority of cells that were positive for the presence of MNV ProPol, antigen were positive also for the F4/80 antigen and were located in focal inflammatory lesions (Figures 4, S11), along hepatic sinusoids or in macrophages adherent to the vascular endothelium (Figure 5).

In some cases, intravascular cells, suggestive of monocytes, were immunoreactive, especially in lung (Figure 9) and liver (Figure S2). A few cells in the Peyer’s patches (Figure S9) of some mice and a focal peritonitis (Figure S6) and pleuritis in three Rag1 ^−/−/Stat1 ^−/− mice showed antigen expression in inflammatory cells. The small intestine contained immunoreactive cells in the lamina propria (Figure S8) and the epithelium of two Rag1 ^−/−/Stat1 ^−/− mice (Figure 10). The MNV-positive epithelial cells were sometimes reminiscent of intestinal endocrine cells, but additional studies using cell lineage-specific markers will be needed to investigate the tropism of MNV in vivo.

At facility 1, viral antigens were found in mesenteric lymph nodes of 3/3 mice (2 Rag1 ^−/−/Stat1 ^−/−, one Rag1 ^−/−/IFNγ R ^−/−) and at facility 2, antigen was found in the mesenteric nodes of 7/9 Rag2 ^−/− mice for which nodes were available for study. There were usually only 3–10 immunoreactive cells detected in each node section of the Rag2 ^−/− mice. In these cases, the ProPol antibody yielded positive cells, suggesting active viral replication. Other lymph nodes of the same mice showed no reactivity with the 3 MNV-specific antibodies. The mesenteric lymph nodes of the Rag2 ^−/− mice were composed mostly of dendritic-like cells with no obvious lymphocytes, follicles or germinal cells present as observed previously in mice of this genotype (Shinkai et al., 1992; J. Ward, unpublished observations). Some of these dendritic-like cells appeared to be expressing MNV antigens (Figures 11 and 12). A higher number of MNV-positive cells was usually observed in the mesenteric lymph node from the Rag1 ^−/−/Stat1 ^−/− mice, and the infected cell type was examined for both F4/80 (macrophage) and CD40 (activated dendritic cell) markers. In the lymph node, F4/80+ cells were mostly in the medullary cords while the CD40-positive cells were in the paracortex (Figure S12). Double staining of the nodes with CD40 and MNV-specific antibodies showed that most CD40-positive cells were localized to the paracortex, and few cells expressing both CD40 and MNV antigens were found in the paracortex (Figure S12). Most of the virus-positive cells were in the medulla and were not expressing CD40. Controls that included the same tissues as above (Figures S3, S7, S10) or tissues from noninfected mice (not shown) did not show reactivity with pre-immunization serum.

Discussion

Genetically engineered immunodeficient mice play an important role in many areas of biomedical research. We studied the pathology of 28 MNV-infected immunodeficient mice in animal rooms of 2 U.S. research facilities in which evidence for the presence of MNV was first found in sentinel mice. Our data show that animal rooms with MNV-positive sentinel mice can indeed reflect the presence of the virus in immunodeficient mice housed in the same room. In facility 1, naturally occurring MNV infection was associated with progressive clinical disease, histopathological lesions, and MNV antigen expression in liver, lung and other tissues in various types of immunodeficient mice, while at facility 2, no MNV-associated clinical symptoms or lesions were observed in Rag2 ^−/− mice.

The first murine norovirus, strain MNV-1, was identified at facility 1 in Rag2 ^−/−/Stat1 ^−/− mice (Karst et al., 2003). The clinical disease in the facility 1 MNV-infected immunodeficient mice examined in our present study presented as varying degrees of hepatitis, peritonitis, and pneumonia. These lesions were similar to those previously observed in experimentally infected Rag2 ^−/−/Stat1 ^−/− and Stat1 ^−/− mice (Karst et al., 2003, Wobus et al., 2004; Ward and Wobus, unpublished). It has been shown previously that experimentally infected Rag2 ^−/−/Stat1 ^−/−, Stat1 ^−/−, Stat1 ^−/−/Pkr ^−/− and IFNαβγ R ^−/− mice were highly susceptible to MNV-1-induced lethality (Karst et al., 2003). The lesions observed in the current and previous studies most likely caused debilitating illness in these mice. However, the exact cause of death remains unknown. Our data suggest that mice with deficiencies in the innate immune system could become ill if infected with MNV by inadvertent exposure to the virus in an animal room.

Taken together, these data suggested a tropism of MNV for macrophages and dendritic cells in immunodeficient mice. Our study confirmed the tropism of MNV for macrophages in the liver as cells in hepatic lesions of MNV-infected mice stained with both MNV ProPol and F4/80. Whether MNV also infects dendritic cells in these mice remains to be investigated as MNV-positive cells in the mesenteric nodes usually did not double stain for MNV and CD40. However, MNV may not infect CD40-positive dendritic cells, since CD40 is a marker of activated dendritic cells and previous studies only demonstrated infections of unstimulated dendritic cells with MNV-1 (Wobus et al., 2004). Future studies are directed at verifying the infected cell types using other cell type specific markers and investigating other organs as well as identifying the cell type(s) infected by MNV in wild-type mice.

The identification of MNV antigen-positive cells of macrophage-like and dendritic cell-like morphology in liver, red pulp of the spleen, lamina propria of the intestine, and mesenteric lymph nodes from naturally infected mice, suggested that MNV replicates in cells of the mononuclear phagocyte lineage in certain strains of immunocompromised mice. This is similar to previous findings of infection in macrophage-like cells in the liver and spleen of Stat1 ^−/− mice experimentally infected with MNV-1 (Wobus et al., 2004). Taken together, these data suggest a tropism of MNV for macrophages and dendritic cells in immunodeficient mice. Although our study showed that many of the MNV-infected cells in the liver expressed F4/80, a macrophage antigen, most MNV-infected cells in the mesenteric lymph node did not express F4/80 or CD40.

Future studies are directed at verifying the infected cell types using other cell type specific markers and identifying the cell type(s) infected by MNV in wild-type mice. An interesting finding in this study was the discovery of MNV in the mesenteric lymph nodes of Rag2 ^−/− mice in facility 2. The presence of MNV in the mesenteric lymph nodes or duodenum of these mice was not associated with clinical disease, which was consistent with previous experimental studies. Rag1 ^−/− and Rag2 ^−/− mice challenged with MNV-1 did not develop disease, although MNV-1 RNA was detected in multiple organs (lung, liver, spleen, intestine, blood, and brain) and shed in feces at 12 weeks after infection (Karst et al., 2003). The histopathology was not examined in the previous challenge study, but in this study we did not observe MNV-associated lesions in the tissues examined from Rag2 ^−/− mice. The detection of MNV antigen and genome in mesenteric lymph nodes of clinically normal Rag2 ^−/− mice raises the possibility that MNV can cause a persistent infection in these mice. We have isolated MNV in cell culture from these tissues, and the characterization of these viruses is in progress to examine whether viral strain differences are involved in determining virulence and tissue and cell tropisms (data not shown).

Certain observations in our study may be relevant to studies of noroviruses associated with human disease. First, MNV may establish a persistent infection in certain immunodeficient mice, and prolonged human norovirus infection in immunocompromised patients has been documented (Nilsson et al., 2003; Gallimore et al., 2004; Rodriguez-Guillen et al., 2005). It is not known whether immunocompetent hosts can maintain a persistent norovirus infection, but it was noteworthy that following oral challenge of immunocompetent mice with MNV-1, viral genome was detected in mesenteric lymph nodes, jejunum and spleen in a portion of wild-type CD1 mice as long as 5 weeks after infection (Hsu et al., 2005, 2006). It should be noted that persistent infection of immunocompetent cats with a related virus, feline calicivirus (FCV), is well documented and the virus persists in the tonsils of infected animals (Wardley and Povey, 1977; Povey, 1986; Dick et al., 1989). A second finding of possible relevance to the human noroviruses was the identification of MNV-positive cells in the lamina propria of the small intestine and in dendritic-and macrophage-like cells of the mesenteric lymph nodes in some immunodeficient mice. The site of human norovirus replication is presumed to be the upper small intestine, but the susceptible cells that support replication have yet not been confirmed. It will be of interest to compare the cell and tissue tropisms of the murine and human noroviruses in future studies.

The finding that MNV-infected Rag2 ^−/− mice do not show clinical signs of disease and pathology while Rag1 ^−/−/Stat1 ^−/− or Rag1 ^−/−/IFNγ R^−/− mice show clinical signs of disease and pathology is most likely due to the essential role of innate immune responses in resistance to MNV. Previous work with experimental MNV-1 infection (Karst et al., 2003) demonstrated that mice with defects in the innate immune system, specifically Stat1 ^−/− and interferon receptors, succumb to MNV-1 infection. Genetic differences between strains of MNV may also partially account for differences in clinical disease and histopathology observed in Rag1 ^−/− and Rag2 ^−/− mice at facility 1 and Rag2 ^−/− mice at facility 2. Previous work with experimental MNV-1 infection (Karst et al., 2003) demonstrated that Rag1 ^−/− and Rag2 ^−/− mice have high levels of MNV-1 RNA in multiple tissues even 3-months postinoculation. In addition, strain differences between MNV-1 and newly isolated strains of MNV (Thackray, Green et al., unpublished) could also partially account for differences in virulence as regards clinical disease and histopathology.

The vigilance required in the housing and maintenance of immunodeficient mice is well known. The Stat1 ^−/− mice, in particular, are highly susceptible to viral disease (Durbin et al., 1996; Hogan et al., 2004). Our data showing MNV infection in the lymph nodes of asymptomatic immunodeficient mice indicate that the virus can spread undetected in a mouse colony. We conclude that researchers should be aware of the serological status of their sentinel mice, so that the associated disease and lesions of MNV infection will not present a confounding variable in the interpretation of pathology results in immunodeficient mice.