Biology and Cellular Tropism of a Unique Astrovirus Strain: Murine Astrovirus 2

Abstract

Biology and tropism of MuAstV2Murine astrovirus 2 (MuAstV2) is a novel murine astrovirus recently identified in laboratory and wild mice. MuAstV2 readily transmits between immunocompetent mice yet fails to transmit to highly immunocompromised mouse strains—a unique characteristic when contrasted with other murine viruses including other astroviruses. We characterized the viral shedding kinetics and tissue tropism of MuAstV2 in immunocompetent C57BL/6NCrl mice and evaluated the apparent resistance of highly immunocompromised NOD-Prkdc^em26Cd52Il2rg^em26Cd22/NjuCrl mice to MuAstV2 after oral inoculation. Temporal patterns of viral shedding were determined by serially measuring fecal viral RNA. Tissue tropism and viral load were characterized and quantified by using in-situ hybridization (ISH) targeting viral RNA. Cellular tropism was characterized by evaluating fluorescent colocalization of viral ISH with various immunohistochemical markers. We found a rapid increase of fecal viral RNA in B6 mice, which peaked at 5 d after inoculation (dpi) followed by cessation of shedding by 168 dpi. The small intestine had the highest percentage of hybridization (3.09% of tissue area) of all tissues in which hybridization occurred at 5 dpi. The thymus displayed the next highest degree of hybridization (2.3%) at 7 dpi, indicating extraintestinal viral spread. MuAstV2 RNA hybridization was found to colocalize with only 3 of the markers evaluated: CD3 (T cells), Iba1 (macrophages), and cytokeratin (enterocytes). A higher percentage of CD3 cells and Iba1 cells hybridized with MuAstV2 as compared with cytokeratin at 2 dpi (CD3, 59%; Iba1, 46%; cytokeratin, 6%) and 35 dpi (CD3, 14%; Iba1, 55%; cytokeratin, 3%). Neither fecal viral RNA nor viral hybridization was noted in NCG mice at the time points examined. In addition, mice of mixed genetic background were inoculated, and only those with a functioning Il2rg gene shed MuAstV2. Results from this study suggest that infection of, or interaction with, the immune system is required for infection by or replication of MuAstV2.

Astroviruses are nonenveloped, positive-sense, single-stranded RNA viruses with a star-like appearance—from which the name derives—when examined by transmission electron microscopy. First identified in 1975, astroviruses are commonly associated with gastrointestinal illness in children.^1,23 They demonstrate considerable diversity, and unique strains have been identified in numerous species through advances in molecular diagnostics.^{2,4-6,11-13,18,19,21,25,27,29,30,32,35-40} This broad distribution likely resulted from cross-species transmission and subsequent adaptation to the novel host.¹¹ Clinical presentation varies among species, although most infections are asymptomatic or limited to mild gastrointestinal illness.^4,9,11 Extraintestinal disease resulting in fatal encephalitis has been described in several species (including cows, mink, and immunocompromised people).^{3,19,22,26,35}

Astroviral infection of mice was first described in 1985, when an unknown astrovirus was identified by electron microscopy in the feces of nude mice.¹⁷ Since then, astroviruses have been detected in many wild and laboratory mouse populations.^12,29,30,34 Despite their prevalence, studies have been limited and their effects on host biology remains largely unknown. Murine astrovirus (MuAstV) was identified through molecular sequencing in 2012 and has since been discovered to be enzootic in numerous research and production mouse colonies.^12,29,34 Whether the strain described in 1985 was MuAstV is unknown. Immunocompetent and immunodeficient mouse strains are both susceptible to MuAstV infection, although no clinical disease and only minimal pathology are observed.^7,47 Similar to astroviruses infecting other species, MuAstV infection is frequently localized to the gastrointestinal tract.⁴⁷ A recent study demonstrated MuAstV replication in goblet cells and altered mucus production within the gastrointestinal tract, highlighting the potential effect of the virus on select research studies despite the lack of clinical disease and pathology.⁸

Our group previously reported the detection of a novel murine astrovirus, murine astrovirus 2 (MuAstV2), in a laboratory mouse colony.³¹ MuAstV2 is genetically distinct from MuAstV but is closely related to a strain recently reported in wild mice.^31,43 The MuAstV2 strain identified in the laboratory mouse colony shares 89.2% nucleotide identity to a strain detected in wild mice in New York City but less than 50% nucleotide identity to MuAstV, the strain commonly isolated from laboratory mice. In addition, MuAstV2 was found to share as much as 80.8% nucleotide similarity to an astrovirus strain isolated from urban brown rats (Rattus norvegicus) in Hong Kong.^5,31 Antibodies to MuAstV2 were inadvertently detected in laboratory colony mice when a serologic immunoassay for mouse thymic virus prepared from a murine T-cell line tested positive. Further analysis showed that the mice were negative for mouse thymic virus and that the T-cell line was contaminated with a novel astrovirus strain similar to MuAstV2, resulting in the positive test. The observation that MuAstV2 did not appear to infect highly immunocompromised mice via natural exposure or experimental inoculation was highly unusual.³¹ This finding is distinct from other murine viruses, including MuAstV, given that infection of immunocompromised mice leads to persistent infection and chronic virus shedding.^12,15,16,47

We sought to further understand the biology of MuAstV2 by evaluating viral shedding kinetics and tissue tropism in immunocompetent mice and to further characterize the presumptive resistance to infection observed in highly immunocompromised mice. Temporal patterns of viral shedding were determined by serially measuring fecal viral RNA after oral inoculation. Tissue and cell tropism were characterized using in-situ hybridization (ISH) and immunohistochemistry during the course of infection. We hypothesized that MuAstV2 initially infects the gastrointestinal tract, as occurs with other astroviruses, but speculated that components of the immune system were required to support infection or replication or both. Furthermore, we sought to characterize the extraintestinal spread of MuAstV2.

Materials and Methods

Mice.

C57BL/6NCrl (female, 8 wk old) and NOD Prkdc^em26Cd52Il2rg^em26Cd22/NjuCrl (female, 8 wk old) mice obtained from Charles River Laboratories (Wilmington, MA) were used. Mice were housed by strain in polysulfone shoebox cages in a ventilated caging system (Maxi-Miser, Thoren Caging Systems, Hazelton, PA) on autoclaved aspen-chip bedding (PWI Industries Canada, Quebec, Canada) with cotton nesting material (0.5 in²; Nestlets, Ancare, Bellmore, NY) and were given flash-autoclaved, γ-irradiated feed (LabDiet 5058, PMI, St Louis, MO) and acidified water (pH 2.5 to 2.8) ad libitum. The animal room was maintained on a 12:12-h light:dark cycle, 10 to 15 air changes hourly (100% outside air), temperature at 22 ± 1 °C, and relative humidity at 30% to 70%. Mice were obtained from rooms free of and were considered seronegative for: ectromelia virus, hantavirus, lactate dehydrogenase-elevating virus, minute virus of mice, mouse adenovirus, mouse cytomegalovirus, mouse hepatitis virus, mouse parvovirus, K virus, mouse rotavirus, mouse thymic virus, murine norovirus, pneumonia virus of mice, mouse polyoma virus, reovirus, Sendai virus, Theiler murine encephalomyelitis virus, Citrobacter rodentium, Clostridium piliforme, Corynebacterium kutscheri, Filobacter rodentium, Helicobacter spp., Mycoplasma pulmonis, Salmonella spp., and endo- and ectoparasites. Prior to inoculation, all mice were confirmed to be free of MuAstV and MuAstV2 via RT-PCR. All procedures were performed at Weill Cornell Medicine after approval by the IACUC, in accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals, 8th edition²⁸.

MuAstV2 inoculation.

A viral stock was created from small and large intestinal contents of colony mice positive for MuAstV2. Intestinal contents (approximately 2.5 mL) were homogenized with sterile PBS (approximately 11.5 mL total), clarified by centrifugation (4696 xg rpm for 20 min), and passed through a sterile 0.22-µm filter (Nalgene 726-2520; Thermo Fisher Scientific, Waltham, MA). The viral stock (filtrate) was maintained at –80 °C until use and was confirmed via RT-PCR to be positive for MuAstV2 and negative for MuAstV. In addition, the viral stock was confirmed to be PCR-negative for: lymphocytic choriomeningitis virus, ectromelia virus, mouse adenovirus, mouse cytomegalovirus, mouse hepatitis virus, minute virus of mice, mouse parvovirus, K virus, mouse rotavirus, mouse thymic virus, pneumonia virus of mice, mouse polyoma virus, reovirus, Sendai virus, Theiler murine encephalomyelitis virus, Citrobacter rodentium, Clostridium piliforme, Filobacter rodentium, Klebsiella oxytoca, Klebsiella pneumoniae, Mycoplasma pulmonis, Pasteurella pneumotropica, Salmonella spp., and endo- and ectoparasites. Prior to mouse inoculation, the viral stock was thawed and diluted 1:4 with sterile PBS.

Viral shedding kinetics.

To characterize viral shedding, 10 B6 and 10 NCG mice were inoculated via oral gavage with 200 µL of the diluted viral stock solution containing an estimated 174,000 viral copies. However, 6 of the 10 NCG mice originally used were later found to have genetic contamination (see genotype confirmation below). Therefore, another 5 genetically verified NCG mice were added to the study. An additional 2 control mice of each strain were inoculated with 200 µL of PBS and housed together by strain in separate cages. Mice were monitored daily for clinical disease.

Mice of the same strain were housed in groups of 5 mice per cage. One fecal pellet was collected at 1, 2, 3, 5, 7, 14, 21, 28, 35, 42, 56, 84, 114, 140, and 168 d postinoculation (dpi) from each B6 mouse; at 1, 2, 3, 5, 7, 14, and 21 dpi from each of the first 10 NCG mice; and at 3, 7, 14, 21, and 28 dpi from each of the 5 mice in the second NCG group. For fecal collection, each mouse was placed in a clean cage for fecal collection and returned to its’ home cage immediately thereafter. Fecal pellets were stored at –80 °C until analysis by using MuAstV2 qRT-PCR on individual mice to characterize viral shedding through 114 dpi. Mice that tested positive at 114 dpi continued to be sampled through 168 dpi. Mice were euthanized with carbon dioxide and submitted for gross necropsy at the final time point.

Genotype confirmation.

After inoculation of the initial group of NCG mice (n = 10) to determine viral shedding kinetics, MuAstV2 RNA was detected for as long as 21 dpi in feces of some mice in this group; these mice underwent further analysis. Hematology was performed by collecting a maximum of 100 µL of whole blood from the retroorbital venous sinus under isoflurane anesthesia. Blood was collected into EDTA tubes and the lymphocyte count determined using an automated analyzer (IDEXX Procyte DX). Because 4 of the 10 mice had WBC counts greater than expected for the strain, the vendor subsequently genotyped each of the aforementioned mice via PCR analysis at both the Prkdc and IL2rg loci and concurrently via single-nucleotide polymorphism analysis to confirm the background strain, by using nucleic acid extracted from an approximately 3-mm section of the mouse’s distal tail.

Pathology and tissue tropism.

To characterize pathology and tissue tropism, 15 B6 and 6 NCG mice were inoculated with 200 µL of the diluted viral stock solution via oral gavage. Mice were housed according to strain with as many as 5 mice per cage; housing was re-adjusted to 5 mice per cage whenever mice were euthanized. As controls, an additional 5 B6 mice were inoculated with 200 µL of PBS and were housed together in a separate cage. Four B6 mice (3 inoculated, 1 control) were euthanized at 3, 5, 7, 14, 28, and 168 dpi and submitted for gross pathology, histopathology, and ISH. Mice from the viral shedding kinetics experiment were used for the 168-dpi time point. Three NCG mice (all inoculated with MuAstV2) were euthanized each at 3 and 7 dpi and processed similarly. One fecal pellet was collected from each mouse at euthanasia for MuAstV2 qRT-PCR analysis to determine viral shedding.

RT-PCR analysis.

Viral copy numbers were estimated by using a RT-PCR assay as previously described.³¹ Quantitative analysis was performed to evaluate viral shedding by determining the fecal PCR copy number in each fecal pellet. A minimum positive threshold of 100 copies per reaction was used, according to the positive control template used for the assay.

Postmortem evaluation.

After CO₂ euthanasia, mice underwent a complete necropsy. Whole blood (maximum, 1 mL) was obtained from each mouse via cardiocentesis immediately after euthanasia and allowed to clot in a serum collector (BD Microtainer, BD Bioscience, San Jose, CA). The clot was removed and fixed in 10% neutral buffered formalin for a minimum of 72 h prior to embedding and staining. After gross examination, all organs were fixed in 10% neutral buffered formalin. After fixation, the skull, spinal column, sternum, femur and tibia were decalcified in a 14% EDTA solution and maintained at 4 °C for at least 2 wk; the solution was refreshed every 48 to 72 h. Tissues were then processed in ethanol and xylene and embedded in paraffin in a Leica ASP6025 tissue processor (Leica Biosystems, Wetzlar, Germany). Paraffin blocks were sectioned at 5 µm, stained with hematoxylin and eosin, and examined by a board-certified veterinary pathologist (AOM). The following tissues were processed and examined: heart, thymus, lungs, liver, gallbladder, kidneys, pancreas, stomach, duodenum, jejunum, ileum, cecum, colon, lymph nodes (mandibular, mesenteric), salivary glands, skin (trunk and head), urinary bladder, uterus, cervix, vagina, ovaries, oviducts, adrenal glands, spleen, thyroid gland, esophagus, trachea, spinal cord, vertebrae, sternum, femur, tibia, stifle joint, skeletal muscle, nerves, skull, nasal cavity, oral cavity, teeth, ears, eyes, pituitary gland, and brain. Small and large intestines were collected and processed in a Swiss roll pattern, allowing evaluation along the entire length of the intestines.

ISH.

A pilot study was performed to optimize tissue selection for RNA ISH. Briefly, 5 B6 mice were inoculated as described earlier and euthanized, and the complete set of tissues and organs described earlier were evaluated by using ISH at 2, 3, 5, 7, and 14 dpi (1 mouse per time point). Viral RNA was detected only in stomach, small intestine, large intestine, spleen, thymus, mesenteric lymph node, mandibular lymph node, and bone marrow at one or more time points. Therefore, ISH was limited to these tissues for viral detection at 3, 5, 7, 14, 28, and 168 dpi in B6 mice and at 3 and 7 dpi in NCG mice. In addition, clotted blood was evaluated at 3, 5, 7, 14, and 28 dpi to detect viremia. An automated RNA ISH assay (RNAScope 2.5 LS Assay-Red, Advanced Cell Diagnostics, Newark, CA) was used to detect the presence of viral RNA. Slides were prepared from the paraffin-embedded blocks, containing formalin-fixed, sectioned at 5 µm. Slides were processed according to the manufacturer’s instructions on a Leica BOND (Leica Biosystems), and viral RNA was detected by using custom probes targeting nucleotides 66 through 991 (Advanced Cell Diagnostics) of the MuAstV2 genome. Assay controls were performed for each run against a housekeeping gene (mouse cyclophilin B) as a positive control, and a bacterial specific gene (4-hydroxy-tetrahydrodipicolinate reductase [dapB]) gene as a negative control. Positive RNA hybridization was identified as discrete, punctate chromogenic red dots under bright field microscopy. Slides positive for hybridization were scanned with a 20· objective on a Panoramic P250 whole-slide scanner (3DHistech, Hungary). Quantitative digital image analysis was performed by using the HALO v3.0.311.255 Area Quantification v1.0 module (Indica Labs, Albuquerque, NM). Region of interest and thresholding values were validated by a board-certified veterinary pathologist (AOM). The intestinal luminal contents and all tissue artifacts were excluded from the analysis. The parameters used for all samples for thresholding, determined from the mouse cyclophilin B positive control, were ‘probe stain color’ (R = 0.335, G = 1.929, B = 1.084) and ‘probe stain minimum optical density’ (02725, 0.3938, 0.5). The output value was defined as the percentage total area of tissue hybridized to MuAstV2 probe.

Fluorescent colocalization assays.

A colocalization assay combining both fluorescent immunohistochemistry and chromogenic ISH was performed on paraffin-embedded blocks containing formalin-fixed small intestines of MuAstV2 ISH-positive B6 mice at 2 and 35 dpi. Tissue sections (thickness, 5 µm) were hybridized with ISH as described earlier. When applied to tissue sections and excited by using light with wavelength of 542 to 582 nm, the RNAScope ISH red chromogen probe can be detected in the fluorescent spectrum without the addition of a fluorophore. After application of RNA-ISH, a combined IF–immunohistochemistry (IHC) assay was conducted with the following antibodies (single IHC marker per slide): antiB220 (dilution, 1:100; catalog no. 550286, BD Bioscience, San Jose, CA), antiCD3 (1:100; ab135372, Abcam, Cambridge, MA), antiCD4 (1:100; 14-9766-82, eBioscience, San Diego, CA), antiCD8a (1:250; 14-0808, eBioscience), anticytokeratin (1:100; Z0622, Dako, Santa Clara, CA), antiIba1 (1:500; ab5076, Abcam), and antisynaptophysin (1:200; 611880, BD Bioscience); these primary antibodies were followed by an appropriate fluorescent-labeled secondary antibody (1:500; A11055, A21202, A21206, or A21208, Alexa Fluor 488, Invitrogen, Waltham, MA) applied to each ISH slide. All IHC–ISH marker combinations were screened for colocalization by using widefield fluorescence microscopy of samples collected at 35 dpi. Immunohistochemistry was performed only with the antibodies that resulted in a combination of immunopositive cells and viral RNA hybridization (i.e., anticytokeratin, antiIba1, antiCD3) were evaluated on samples collected at 2 dpi. Colocalization was characterized by RNA probe hybridization within the boundary of the cell membrane highlighted by the IF-IHC antibodies. Slides from immunopositive IHC antibodies (i.e., anticytokeratin; antiIba1, antiCD3) from both time points were then imaged with a Zeiss LSM880 confocal microscope (Carl Zeiss, Oberkochen, Germany) to confirm colocalization, and results were reviewed by a board-certified pathologist (AOM).

The proportion of colocalized cells as a proportion of total cells within a single field and the amount of intracellular viral RNA per cell were quantified for each immunopositive IHC marker at 2 and 35 dpi. Small intestinal villi containing virus (either between or within hybridized cells) were identified, and only cells within these villi were used for quantification. All cells expressing the IHC marker within villi were evaluated for colocalization with viral RNA. The distinct red dots within each cell of 100 total cells were counted for each IHC marker per time point. These values were used to estimate the percentage of colocalized cells and the amount of intracellular viral RNA after exposure to the virus. Dot size and intensity were not quantified.

Results

Viral shedding kinetics.

Among B6 mice, viral RNA was detected in the feces of several mice as early as 1 dpi and in all mice by 2 dpi. Fecal viral RNA increased rapidly, consistently peaking at 5 dpi in all mice before declining through the remaining time points. Four of 9 mice (2 mice per cage) were PCR-negative at 56 dpi, and copy numbers were low in mice that remained positive. At 84 dpi, 2 additional mice were PCR-negative, whereas 3 mice that had tested negative at 56 dpi tested positive, yielding low viral RNA copy numbers in all samples. Viral RNA was present in 2 mice at 140 dpi, and all mice were negative by 168 dpi (Table 1).

Table 1.

Fecal qRT-PCR results from B6 mice after inoculation

	1 dpi	2 dpi	3 dpi	5 dpi	7 dpi
No. of positive mice/total no.	1/10	10/10	10/10	10/10	10/10
% positive	10	100	100	100	100
Copy numbera	201 ± 0	1960 ± 965	108,697 ± 65012	2,965,630 ± 2,487,172	395,126 ± 214,784
	14 dpi	21 dpi	28 dpi	35 dpi	42 dpi
No. of positive mice/total no.	10/10	10/10	10/10	10/10	9/9b
% positive	100	100	100	100	100
Copy number	218,849 ± 257,902	10,748 ± 9185	3795 ± 3944	2302 ± 2801	897 ± 843
	56 dpi	84 dpi	114 dpi	140 dpi	168 dpi
No. of positive mice/total no.	5/9	6/9	2/9	2/9	0/9
% positive	56	67	22	22	0
Copy number	398 ± 592	404 ± 223	201 ± 0	201 ± 0	0

^aAverage copy number per fecal pellet (1 fecal pellet per sample; mean ± 1 SD). Calculated from positive samples only.

^bOne animal was euthanized prior to day 42.

For NCG mice, Table 2 provides the genotype at both the Prkdc and IL2rg loci, peripheral blood lymphocyte count, and fecal MuAstV2 RNA copy number at 3, 5, 7, 14, and 21 dpi from each of the initial 10 MuAstV2-inoculated ‘NCG’ mice. MuAstV2 RNA was detected in the feces of 4 of the 10 of the ‘NCG’ mice at each of the 5 time points examined, peaking at day 7, whereas 6 of the 10 were negative. Lymphocyte counts were at least 7.5 times higher in mice shedding viral RNA than in those that were not; this feature is not consistent with the NCG phenotype (Table 2). Single-nucleotide polymorphism and PCR analysis was used to confirm the NCG background and homozygous deletions of both the Prkdc and IL2rg genes in 4 of the 10 mice. MuAstV2 RNA was not detected in the feces at any time point of these 4 mice. Single-nucleotide polymorphism analysis confirmed that the 6 negative mice were of mixed genetic background. Among these, of 2 mice from which MuAstV2 could not be detected, 1 had homozygous deletions of both the Prkdc and IL2rg genes, whereas the other had a homozygous deletion of the IL2rg gene but was heterozygous at the Prkdc locus. Conversely, all 4 mice from which viral RNA was detected were heterozygous or wildtype for the IL2rg and Prkdc genes (Table 2). An additional 5 NCG mice (genotype confirmed) were subsequently inoculated with MuAstV2, and fecal viral RNA was not detected at any time point through 28 dpi.

Table 2.

Fecal MuAstV2 RNA copy number, lymphocyte count, and genotype of the initial 10 ‘NCG’ mice used in the virus infectivity experiment

Mouse	Genotype		Lymphocyte count (·103/µL)	qRT-PCR copy number at indicated dpi
	Prkdc-26-del52	IL2RG-26-del22		3	5	7	14	21
1a	Hom	Hom	0.42	0	0	0	0	0
2	WT	WT	8.85	26561	107227	1747528	215443	53367
3	Hom	Hom	0.76	0	0	0	0	0
4	Het	Hom	0.28	0	0	0	0	0
5	Het	Het	6.77	6579	107227	1747528	26561	26561
6	WT	Het	7.49	13219	215443	432876	6579	1630
7a	Hom	Hom	0.77	0	0	0	0	0
8a	Hom	Hom	0.53	0	0	0	0	0
9	Het	Het	7.41	26561	1747528	1747528	26561	3275
10a	Hom	Hom	0.90	0	0	0	0	0

Het, heterozygous; Hom, homozygous.

^aNCG genotype confirmed.

Pathology.

No gross or microscopic lesions were observed in the NCG mice, and lesions in B6 mice were attributed to those typically observed in the strain. One B6 mouse was euthanized at 35 dpi due to a distended abdomen. At necropsy, the mouse had mucometra secondary to an imperforate vagina.

ISH.

Table 3 summarizes the tissue tropism and temporal distribution of hybridization in B6 mice. Chromogen was detected in the small and large intestines, mesenteric and mandibular lymph nodes, spleen, thymus, bone marrow, and circulating leukocytes from mice for at least 1 of the 6 time points evaluated. Quantification of the percentage total tissue area with chromogen detection revealed that the small intestine exhibited the greatest proportion of and change in hybridization, increasing between 3 dpi (0.36%) and 5 dpi (3.09%), subsequently decreasing at 7 dpi (0.43%), and continuing to decline through the remaining time points, thus mirroring the pattern of fecal viral shedding. The mesenteric lymph nodes and large intestine showed a similar increase in hybridization through 5 dpi, albeit with a lower proportion of hybridized tissue (mesenteric lymph node, 1.23%; large intestine, 0.17%; Figure 1). Viral RNA persisted at an increased level for a longer duration in mesenteric lymph node in contrast to the other tissues. The thymus, spleen, and mandibular lymph node showed a similar pattern of hybridization, with a 48-h delay as compared with the peak fecal viral shedding and hybridization in the intestines. This finding was consistent in all 3 tissues, peaking at 7 dpi (thymus, 2.30%; spleen, 0.29%; mandibular lymph node, 0.21%). Of these lymphoid tissues, the thymus exhibited the greatest proportion of hybridization, peaking at 7 dpi, and was second only to the small intestine in the amount of hybridization. In contrast, the bone marrow exhibited the lowest level of hybridization (0.02%), which was detected at 5 and 7 dpi only (Figure 2).

Table 3.

Tissues positive for MuAstV2 viral RNA hybridization at 3 to 168 dpi in B6 mice

	no. of positive mice/total no. at indicated dpi
	3	5	7	14	28	168
Small intestine	3/3	3/3	3/3	3/3	3/3	2/3
Large intestine	0/3	3/3	3/3	3/3	3/3	3/3
Mesenteric LN	3/3	3/3	3/3	2/2a	3/3	NC
Mandibular LN	0/3	3/3	3/3	3/3	0/3	0/3
Spleen	1/3	3/3	3/3	2/3	1/3	2/3
Thymus	0/3	2/3	3/3	3/3	0/3	1/3
Bone marrow	0/3	1/3	3/3	0/3	0/3	NC
Blood clotb	1/1	0/1	1/1	0/1	0/1	NC

LN, lymph nodes; NC, not collected.

^aNot collected from 1 mouse.

^bSamples pooled per time point.

Figure 1.

Average percentage total area of MuAstV2 RNA hybridization and fecal viral copy number at 3, 5, 7, 14, 28, and 168 dpi in intestine-associated tissues of B6 mice (n = 3 per time point; mesenteric lymph node, n = 2 at 14 dpi). Background hybridization indicated by black dotted line. Error bars indicate the standard deviation.

Figure 2.

Average percentage total area of MuAstV2 RNA hybridization and fecal viral copy number at 3, 5, 7, 14, 28, and 168 dpi in hematopoietic and lymphoid-associated tissue of B6 mice (n = 3 per time point). The black dotted line indicates background hybridization. Error bars represent the standard deviation. *, standard deviation = 0.64%.

Figures 3, 4, and 5 are photomicrographs from select tissues with MuAstV2 RNA hybridization at 3, 5, and 7 dpi. The pattern of hybridization ranged from single punctate to coalescing clusters of red dots, varying between time point and tissue type (Figure 3). At 3 dpi, multifocal regions with dense clusters of hybridization were seen randomly throughout all sections of the small intestine, occurring largely within the lamina propria, and with regions of nonhybridized tissue between hybridization clusters. By 5 dpi, intense hybridization was present diffusely throughout the small intestine, affecting most of the villi and predominately localized within the lamina propria. The regions of nonhybridized tissue present at 3 dpi were no longer present. By 7 dpi, hybridization remained throughout all sections of the small and large intestines, albeit at a lower intensity and hybridizing single or small groups of cells in contrast to the intense clusters noted previously (Figure 3). The most intense hybridization was associated with cells located throughout the lamina propria, whereas lower hybridization intensity was present in a small subset of enterocytes at all time points. Within these enterocytes, hybridization was localized to the apical portion of the cell, perhaps reflecting tropism or uptake of viral RNA (Figure 4). Within the thymus, the intensity of hybridization appeared to be greater in the thymic medulla as compared with the cortex at 7 dpi (Figure 3). No specific pattern was identified in the remaining tissues or time points. In addition, hybridization was detected in circulating leukocytes at 3 and 7 dpi (Figure 5). These leukocytes were cytologically suggestive of lymphocytes. The circulating leukocyte data were not assessed quantitatively. Chromogen was not detected under bright-field microscopy in NCG mice at either time point (Figure 6).

Figure 3.

Sections of small intestine, large intestine, mesenteric lymph node, thymus, and spleen from MuAstV2-infected B6 mice at 3, 5, and 7 dpi evaluated for MuAstV2 RNA-ISH counterstained with hematoxylin. Red chromogen dots indicate positive RNA hybridization. Small intestine: strong multifocal hybridization present in the small intestinal lamina propria at 3 dpi; diffuse strong hybridization in the lamina propria at 5 dpi; and diffuse moderate hybridization in the lamina propria at 7 dpi. Large intestine: No hybridization at 3 dpi; scant, multifocal hybridization in the mucosa/submucosa at 5 and 7 dpi. Mesenteric lymph node: scant, multifocal hybridization throughout at 3 dpi; strong, multifocal hybridization throughout at 5 and 7 dpi. Thymus: no hybridization at 3 dpi; scant, multifocal hybridization at 5 dpi; and strong, multifocal hybridization within the thymic cortex and medulla at 7 dpi. Spleen: no hybridization at 3 dpi; scant, multifocal hybridization mostly within the red pulp at 5 dpi; multifocal, strong immunoreactivity within the red and white pulp at 7 dpi.

Figure 4.

Section of small intestine from MuAstV2-infected B6 mice evaluated 5 dpi for MuAstV2 RNA-ISH counterstained with hematoxylin. Strong multifocal to coalescing areas of hybridization within the mononuclear cells in the lamina propria as well as within a subset of enterocytes.

Figure 5.

Section through a formalin-fixed, paraffin-embedded blood coagulum obtained from clotted centrifuged peripheral blood from a B6 mouse infected with MuAstV2 and evaluated at 5 dpi for MuAstV2 RNA-ISH counterstained with hematoxylin. Multiple mononuclear cells consistent with circulating lymphocytes (1.5 to 2 erythrocytes in diameter) show punctate cytoplasmic positive MuAstV2 hybridization.

Figure 6.

Section of small intestine from MuAstV2-infected NCG mice evaluated at 3 dpi for MuAstV2 RNA-ISH counterstained with hematoxylin. No hybridization was observed.

Fluorescent colocalization assays.

MuAstV2 RNA hybridization was found to colocalize with only 3 of the 7 markers evaluated at 35 dpi: CD3 (T-cells), Iba1 (macrophages), and cytokeratin (enterocytes). Subsequently, colocalization was also identified in samples evaluated for these 3 markers at 2 dpi. Representative images from 2 or 35 dpi are shown in Figure 7. Hybridization was identified as a variable number of red dots randomly distributed within the cytoplasm. This pattern was consistent between markers and between time points. There was no colocalization of ISH signal and CD4, CD8, B220, or synaptophysin at 35 dpi. The percentage of hybridized cells and the average number of dots per cell at each time point are shown in Table 3.

Figure 7.

Representative sections of small intestine from B6 mice infected with MuAstV2 and evaluated with MuAstV2 RNA-ISH (red fluorescence), IF-IHC (green fluorophore), and DAPI (blue fluorescence). White bar, 10 µm. (A) AntiCD3 IF-IHC at 2 dpi shows multiple lymphocytes within the lamina propria, with MuAstV2 hybridization and colocalization. (B) High magnification of CD3-immunopositive lymphocyte at 2 dpi shows MuAstV2 within the cytoplasm. (C) Antipancytokeratin IF-IHC at 35 dpi shows an enterocyte (white arrow) with MuAstV2 hybridization and colocalization. (D) High magnification of an enterocyte at 35 dpi shows MuAstV2 hybridization within the apical aspect of the cytoplasm. (E) AntiIba1 IF-IHC at 2 dpi shows multiple histiocytes within the lamina propria with MuAstV2 hybridization and colocalization. (F) High magnification of a single histiocyte at 35 dpi shows colocalization of RNA within the cytoplasm.

A higher percentage of T cells and macrophages hybridized with MuAstV2 as compared with enterocytes at both time points (2 and 35 dpi) (Table 4). The sample mean percentage of hybridized cells and number of dots per cell at 2 and 35 dpi was 59% and 14% (3.63 and 2.36 dots per cell) for CD3; 6% and 3% (2.00 and 1.67 dots per cell) for cytokeratin; 46% and 55% (2.37 and 2.73 dots per cell) for Iba1, respectively.

Table 4.

Percentage of positive cells in the small intestine at 2 and 35 dpi that costained for MuAstV2 RNA and each cell marker

	2 dpi		35 dpi
	% positive cellsa	Average no. of dots per cellb	% positive cells	Average no. of dots per cell
CD3	59	3.63±2.38	14	2.36±1.65
CK	6	2.00±1.10	3	1.67±1.15
Iba1	46	2.37±1.79	55	2.73±1.77
CD4	NEc	—	Negc	—
CD8a	NE	—	Negc	—
B220	NE	—	Negc	—
Synaptophysin	NE	—	Negc	—

NE, not evaluated; Neg, negative.

^aPer 100 cells.

^bmean ± 1 SD.

^cNot evaluated with confocal microscopy.

Discussion

This study revealed that B6 mice were susceptible to MuAstV2 infection after oral inoculation. The pattern of infection resembled that detected after experimental inoculation of B6 mice with MuAstV and is highly suggestive of the oral–fecal pattern of transmission observed with other astrovirus strains and enteric viruses.^8,9,24,44 Fecal viral RNA was detected at 84 dpi in several mice that had tested negative at the previous time point. The significance of this finding and whether it is artifactual or related to the biology of MuAstV2 are unknown. The low copy numbers detected at both time points suggest that the sensitivity of the PCR assay or variation in fecal sample size may account for the finding. It may also reflect intermittent shedding during viral clearance, or because the cells supporting viral replication remain unknown, it also could reflect variability in viral trafficking between the target cell or organ and the gastrointestinal tract lumen. Although we found no evidence of viral shedding in the feces after 140 dpi, viral RNA remained detectable in both the gastrointestinal tract and lymphoid tissues at 168 dpi, the last time point evaluated. Further studies are necessary to determine whether MuAstV2 persists and can be reactivated after stress or immunosuppression, or whether the persistent hybridization reflects noninfectious remnants of viral nucleic acid.

ISH was used to characterize the location and quantity of MuAstV2 RNA at the tissue and cellular levels throughout the course of infection. MuAstV2 RNA was detected in intestinal, lymphoid, and hematopoietic tissue of B6 mice with no microscopic pathology. The quantity of MuAstV2 RNA detected in the small intestine and mesenteric lymph nodes correlated to the fecal viral copy number at the same time points, with a delay in peak viral load in the large intestine and distant lymphoid organs. This pattern reflects the experimental route of inoculation, which we believe is its natural fecal–oral mode of transmission. The pattern and degree of hybridization in the small intestine was surprising, because we expected that enterocytes would be the initial and primary target of viral replication. Enterocytes either displayed only a few small, discrete areas of hybridization within the apical cytoplasm or were devoid of viral RNA, whereas the lamina propria contained many cells with multiple to coalescing larger, more intense areas of hybridization. This finding was consistent at all time points and suggests the gastrointestinal epithelium may not be the primary site of infection for MuAstV2; this idea was further supported by the immunohistochemical staining. Although we observed colocalization of viral RNA in cells bound by cytokeratin (enterocytes) at 2 and 35 dpi, the lamina propria had considerably greater colocalization to CD3 and Iba1 (T cells and macrophages, respectively). This cellular infection pattern is distinct from that of MuAstV (which was recently shown to infect goblet cells) and other astroviruses.^8,10,45 The persistence of viral RNA in the lamina propria and continued shedding in the feces suggests gastrointestinal tissue contains—at least transiently—cells that support viral replication or shedding. The thymus displayed the highest percentage of extraintestinal tissue hybridization and the second highest percent of tissue hybridization of any of the tissues and time points evaluated. We believe this pattern suggests that the thymus and other lymphoid tissues may be sites of continuing viral replication of MuAstV2; however, further work is required to confirm this hypothesis, given that the hybridization in these tissues could reflect trafficking of immune cells that were infected in the gastrointestinal tract.

To further characterize cellular tropism of MuAstV2, we performed immunohistochemistry on representative tissues from early (2 dpi) and late (35 dpi) infection. Cell markers for immune cells and intestinal epithelium were used in light of the observed pattern of hybridization and known astrovirus biology. Colocalization of viral RNA to CD3-, Iba1-, and cytokeratin-positive cells was observed at 2 and 35 dpi, with the highest degrees of colocalization to CD3 and Iba1. Although these markers are present on multiple cell types, we believe that the colocalized cells were enterocytes (cytokeratin-positive), T cells (CD3-positive), and macrophages (Iba1-positive), given their morphology, location, and staining pattern. The mean percentage of hybridized cells and in the amount of hybridization per cell in CD3-positive cells was lower at 35 dpi (14%, 2.36 dots per cell) than was noted at 2 dpi (59%, 3.63 dots per cell) but slightly higher in Iba1-positive cells at the same time points (46%, 2.37 dots per cell at 2 dpi; 55%, 2.73 dots per cell at 35 dpi). Colocalization was most observed in cells expressing CD3 or Iba1, that is, T cells and macrophages, respectively, suggesting that either or both cell types are the principal target for MuAstV2. The observed colocalization to CD3 and Iba1 is, at least in part, due to viral clearance resulting from the normal immune response; however, these findings in conjunction with the observed tissue hybridization and strain susceptibility differences (inability to infect NCG mice) support T cells or macrophages as MuAstV2 targets. To our knowledge, this tropism is unique to MuAstV2, as it has not been reported for other strains of astrovirus. Colocalization to CD4 and CD8 was not observed in the samples evaluated, despite colocalization to CD3-positive cells. This outcome may reflect the susceptibility of the CD3⁺ CD4^– CD8^– (i.e., double-negative) T-cell subset to MuAstV2 infection. Alternatively, this finding could reflect a lower percentage of hybridized cells or number of hybridizations per cell at 35 dpi, or the methodology used may have obscured colocalization of MuAstV2 to the CD4 or CD8 cell markers.

The percentage of hybridized cells and amount of hybridization was low for cytokeratin at both time points (6%, 2.00 dots per cell; 3%, 1.67 dots per cell) as compared with the other cell markers, providing further evidence that enterocytes are likely not the primary cell type supporting MuAstV2 replication. We speculate that the viral RNA in enterocytes may indicate low levels of viral replication or may not reflect actively replicating virus but rather virus uptake from the lumen via endo- or pinocytosis. A definitive conclusion cannot be drawn, because the methodology used cannot distinguish free viral RNA from actively replicating virus.

Consistent with the aforementioned viral tropism, we found that NCG mice were resistant to infection under these experimental conditions. The lack of fecal viral shedding and the absence of viral RNA detection with ISH in NCG mouse tissues is consistent with earlier observations.³¹ The lack of susceptibility to astroviruses is highly unusual for immunocompromised mice and implies as does the viral hybridization data, that some component of the immune system is required for MuAstV2 infection or replication. This requirement is in stark contrast to other astrovirus strains, which cause sustained infection or more severe clinical disease in immunocompromised species.^{3,10,22,35,47} The fact that we unintentionally inoculated genetically contaminated NCG mice shed greater light on MuAstV2 pathogenesis and cellular tropism because mice that lacked a functional Il2rg did not shed virus after experimental inoculation Il2rg is required for the function of multiple cytokine receptors, including those for IL2, IL4, IL7, IL9, IL15, and IL21.²⁰ These cytokines have many functions associated with immune cell development and function, predominantly in T cells (IL2, IL4, IL7, IL9, IL15, IL21), B cells (IL2, IL4, IL7, IL21), NK cells (IL2, IL15), dendritic cells (IL15, IL21), macrophages (IL4), and mast cells (IL9).^{14,20,33,40-42,46} Possible roles of the common γ chain include viral translocation from the gastrointestinal lumen, proliferation or activation of the target cell or cells, and cellular uptake or release of the virus.^{14,20,33,40-42} Although we cannot conclude specifically why Il2rg is required for MuAstV2 infection, we know based on our observation of hybridization in T cells and macrophages and Il2rg’s impact on the cytokine receptors of these cells, highlighted earlier in the paragraph, that lack of Il2rg affects at least some of the cells observed to contain the most viral RNA as determined by hybridization. Given that many of these cytokines are responsible for cell proliferation and activation, the loss of Il2rg may limit either the number of target cells available for MuAstV2 infection or proliferation after infection. Another possibility is that MuAstV2 translocation from the intestinal lumen was inhibited due to the limited number of susceptible cells within the intestinal epithelium of Il2rg deficient mice. Alternatively, Il2rg may mediate viral entry into or exit from the cell through either direct interaction of the virus with the cellular receptor or indirectly through an alternate pathway regulated by receptor activation by one or more of the previously identified cytokines. To our knowledge, this scenario has not been described previously. Further studies are required to characterize the nature of this interaction.

This study further highlights the highly unique biology of the MuAstV2 astroviral strain, which is an enteric virus whose cellular tropism and replication require interaction with select elements of the host’s immune system. Limited conclusions can be drawn with regard to target cell types. Additional studies will be necessary in various genetically engineered and mutant mouse strains that have specific immune system alterations or by using additional cell identification techniques to elucidate the biology of MuAstV2 more fully. Although the low prevalence of MuAstV2 in laboratory mouse colonies indicates the virus’s rarity makes it a low biosecurity risk, its unique cellular tropism warrants further study.