Astrovirus Induces Diarrhea in the Absence of Inflammation and Cell Death

Abstract

Astroviruses are a leading cause of infantile viral gastroenteritis worldwide. Very little is known about the mechanisms of astrovirus-induced diarrhea. One reason for this is the lack of a small-animal model. Recently, we isolated a novel strain of astrovirus (TAstV-2) from turkeys with the emerging infectious disease poult enteritis mortality syndrome. In the present studies, we demonstrate that TAstV-2 causes growth depression, decreased thymus size, and enteric infection in infected turkeys. Infectious TAstV-2 can be recovered from multiple tissues, including the blood, suggesting that there is a viremic stage during infection. In spite of the severe diarrhea, histopathologic changes in the intestine were mild and there was a surprising lack of inflammation. This may be due to the increased activation of the potent immunosuppressive cytokine transforming growth factor beta during astrovirus infection. These studies suggest that the turkey will be a useful small-animal model with which to study astrovirus pathogenesis and immunity.

Astroviruses are small, round, nonenveloped viruses, typically 28 to 30 nm in diameter (33). The name astrovirus comes from astron, Greek for star, describing the characteristic five- or six-pointed star-like surface projections detected by negative stained electron microscopy. Astroviruses were first observed in 1975 in association with outbreaks of gastroenteritis in infants (2, 31). Since then, astroviruses have been implicated as one of the leading causes of infantile viral gastroenteritis worldwide (15, 16). A longitudinal study in rural Mexico found that astrovirus was the most common cause of infantile gastroenteritis, suggesting that the burden of astrovirus disease in developing countries may be especially high (32). Astrovirus is an endemic cause of diarrhea in infants but is also capable of causing outbreaks in day care centers, hospitals, and other institutions (14, 17, 39-41). Astrovirus-induced gastroenteritis has also been reported in association with food-related illnesses in the United States (35, 36) and is an important cause of gastroenteritis in immunocompromised individuals (5, 9, 10, 18, 27, 66, 70).

Astroviridae family members have been described to cause diarrhea and enteritis in several mammalian and avian hosts (13, 19, 31, 34, 58, 64, 67). Unfortunately, very little is known about astrovirus pathogenesis or the host factors involved in viral clearance and disease resolution. Of the nonhuman astroviruses, only the bovine, ovine, and turkey astroviruses have been studied experimentally (47, 48, 57, 58, 62, 63, 67-69). Although limited, these studies have generated some insight into the potential mechanisms of astrovirus infection and disease. In gnotobiotic lambs, astrovirus infected the mature enterocytes of the small intestine and infected cells were sloughed and replaced by cuboidal epithelial cells (57). The resulting slight villous atrophy was transient, lasting less than 5 days, and resembled a mild rotavirus infection. Unlike gnotobiotic lambs, which exhibited diarrhea, gnotobiotic calves infected with bovine astrovirus alone showed no illness (69). However, a unique type of infection was observed. Astrovirus appeared to target the M cells overlying jejunal and ileal Peyer's patches and mononuclear inflammatory cells and eosinophils were observed atop the infected dome epithelial cells (69). In turkeys, astrovirus infection can be accompanied by a moderate increase in mortality (24, 34, 46, 47). Limited studies demonstrated that astrovirus infection causes only mild histopathology while inducing severe osmotic diarrhea (62, 63).

Although important data were gleaned from these animal experiments, significant questions remain unanswered. The gap in our understanding in astrovirus biology is partly due to the fact that none of the previous systems were fully developed into an animal model. Prior studies did not address the kinetics of astrovirus replication, the cells supporting infection, the effect of infection on cell death, or the host response to astrovirus infection. Recently, we isolated and characterized a genetically and immunologically distinct turkey astrovirus strain, turkey astrovirus type 2/North Carolina/034/1999 (TAstV-2), associated with an emerging disease in turkeys (4, 22-24, 53, 54). In the present studies we describe an in ovo system to culture virus and report the pathogenesis of TAstV-2 in infected turkey poults and embryos, including clinical disease and viral localization. These are the first studies, to our knowledge, to evaluate viremia, apoptosis, and the immunomodulatory cytokine transforming growth factor beta (TGF-β) during astrovirus infection. Results from these studies suggest that the turkey will be an important animal model for understanding the mechanism of astrovirus pathogenesis.

MATERIALS AND METHODS

TAstV-2 propagation.

TAstV-2 was isolated and propagated as described (24, 54). Briefly, the thymus or intestines from infected turkey poults were homogenized, filtered (0.2-μm pore size), and inoculated into the yolk sac of 20-day-old specific-pathogen-free turkey embryos (from a closed flock of Small Beltsville White turkeys housed at Southeast Poultry Research Laboratory). Viral replication in embryo intestines was monitored by in situ hybridization at 1, 3, and 5 days postinoculation (dpi). Virus was harvested at 5 dpi. Intestines were removed, homogenized, 0.2 μm filtered, and centrifuged at 150 × g for 10 min. Additionally, embryo intestinal fluid was collected separately, filtered (0.2-μm pore size), and centrifuged at 500 × g for 10 min.

RNA isolation and RT-PCR.

Total RNA was isolated from purified virus, embryo intestines, or tissues excised from experimentally inoculated or control turkeys with Trizol following manufacturer instructions (Invitrogen, Carlsbad, Calif.). Reverse transcription (RT)-PCR was performed as previously described (23).

TAstV-2 quantitation.

Viral load was assessed by developing a TAstV-2-specific competitive quantitative RT-PCR system. Briefly, total RNA, isolated from 100 μl of infectious material, was analyzed by one-step RT-PCR (Qiagen, Valencia, Calif.) in the presence of a competitor RNA. The competitor RNA was generated by modifying a plasmid (pTAstVpol18) which contains nucleotides 2863 to 5296 of the TAstV-2 genome. pTAstVpol 18 was digested with ScaI following the manufacturer's instructions (Invitrogen), and then two 30-bp randomly generated oligonucleotides were ligated to the cut plasmid to generate a construct with the TAstV-2 pol gene with 60 bp of additional sequence (pTAstVpolC). This new construct was then digested with SstI and NotI following the manufacturer's instructions (Invitrogen) and ligated into the corresponding sites in pGEM T-Easy vector (Promega, Madison, Wis.). This final construct, pTAstVpolCQ, was then used to generate positive-sense competitor RNA with the RNA polymerase SP6 (Roche Molecular, Indianapolis, Ind.). Competitor RNA was purified, and copy numbers were quantitated by spectrophotometry as described (50). TAstV-2 polymerase gene-specific primers, flanking the modified region in pTAstVpolCQ, were designed: CQ-RT-PCR Fwd (CCATGATATGCTACGGGGAT) and CQ-RT-PCR Rev (GACTCAACATCTGGTAGCCT). Sample RNA was added at a uniform concentration to each tube of a serial log dilution of competitor RNA and amplified under the following conditions: 50°C for 30 min, 95°C for 15 min, 30 cycles of 94°C for 30 s, 55°C for 30 s, 72°C for 30 s, and a final 72°C extension for 1 min, with the Qiagen One-Step RT-PCR kit (Qiagen, Valencia, Calif.) in a total reaction volume of 25 μl. Products were then separated by electrophoresis in an agarose gel and the amplification products were visualized with ethidium bromide. The copy numbers of viral RNA in the sample were calculated with Kodak Imaging software densitometry and plotting against the standard curve of the competitor as previously described (12).

Animals.

Two-day-old unvaccinated British United Turkeys of America poults (male and female) were obtained from a commercial hatchery. Control and infected poults were housed in separate biosafety level 2 containment facilities in individual Horsfall units with HEPA filtered inlet and exhaust air valves. Birds were fed routine turkey starter from the University of Georgia and given free access to clean water. After a brief acclimation period, 5-day-old poults were weighed (day 0) and randomly assigned to either a control group or a group infected with astrovirus (n = 60 per group). Poults were orally inoculated with ≈10⁶ genomic units of astrovirus in a 200-μl total volume or phosphate-buffered saline (PBS) alone. Birds were monitored daily for signs of clinical disease and weighed at 0, 3, 5, 9, and 12 dpi. At 1, 2, 3, 4, 5, 7, 9, and 12 dpi, five random poults per group were euthanized by cervical dislocation, and the small intestine, bursa, spleen, pancreas, thymus, liver, kidney, bone marrow, skeletal muscle (breast), feces, and blood were collected. All tissues were stored at −70°C or placed in 10% phosphate-buffered formalin. Blood was collected in syringes containing heparin, incubated overnight at 4°C and then separated into red cell, lymphocyte, and plasma fractions with Histopaque 1077 (Sigma Chemicals, St. Louis, Mo.). The bursa, spleen, and thymus from each group were weighed to the nearest milligram prior to processing.

To perform RT-PCR analysis and virus isolation studies, the individual tissues at each time point were pooled, homogenized, and aliquoted for RNA isolation with Trizol or inoculation into 20-day-old turkey embryos. The animal experiments were repeated five times with different groups of poults with similar results. All animal experiments were approved by the USDA Animal Care and Use Committee and complied with all federal guidelines.

In situ hybridization.

The TAstV-2-specific riboprobe was generated as described (4). Briefly, TAstV-2 plasmid p25.5 containing a 1.5-kb segment of the extreme 3′ end of the TAstV-2 genome (24) was digested with BamHI and transcribed with T7 RNA polymerase and digoxigenin-labeled UTP (Roche Molecular), creating an antisense riboprobe of approximately 1.6 kb in length. Digoxigenin incorporation was verified by dot blot. In situ hybridization was performed according to previously described techniques (8). Briefly, tissue sections were deparaffinized with Citrisolv (Fisher Scientific, Norcross, Ga.), digested with 35 μg of proteinase K/ml for 15 min at 37°C, and hybridized overnight at 42°C with approximately 35 ng of digoxigenin-labeled riboprobe per slide in 5× (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate), SSC 50% formamide, 5% modified milk protein (Roche Molecular), 1% N-lauroylsarcosine, and 0.02% sodium dodecyl sulfate (SDS). The following day, slides were washed in increasingly stringent solutions: 2× SSC with 1% SDS for 30 min at 50°C, 1× SSC with 0.1% SDS for 30 min at 50°, 1× SSC for 15 min three times at room temperature, and 0.1× SSC for 15 min at room temperature.

After the posthybridization washes, sections were incubated with antidigoxigenin antibody conjugated to alkaline phosphatase (Roche Molecular) for 2 h at 37°C and developed with nitroblue tetrazolium and bromcresylindolyl phosphate for 1 to 3 h. Sections were counterstained lightly with hematoxylin and coverslipped with Permount for a permanent record. Each group of slides was processed with a positive control tissue consisting of a section of positive embryo intestine and negative control sections from uninfected poults.

Histopathology.

Tissues from control and infected poults were fixed in 10% phosphate-buffered formalin overnight, then processed, embedded, sectioned (0.3 μm), stained with hematoxylin and eosin, and examined by light microscopy.

Detection of TAstV-2 antigen by immunofluorescence.

The distribution of TAstV-2 was monitored with a rabbit polyclonal antibody generated to a peptide sequence in the TAstV-2 capsid protein (K₆₇₆ to R₆₉₁) (ResGen, Carlsbad, Calif.), accession no. AAF18464. Briefly, tissue sections from turkeys sacrificed at 1, 2, 3, 4, 5, 7, 9, and 12 dpi were processed as described above, deparaffinized with Citrisolv, antigenic sites exposed by microwaving the tissues for 5 min in a citrate buffer, then incubated with primary antibody diluted 1:500 in phosphate-buffered saline containing 0.1% Tween 20 (PBST) overnight at 4°C. After incubation in primary antibody, the slides were washed in PBST, incubated with a biotinylated goat anti-rabbit antibody (Vector Laboratories, Burlingame, Calif.) for 30 min at room temperature, washed in PBST, then incubated with an Alexa488-streptavidin-labeled antibody (Molecular Probes, Eugene, Oreg.) diluted 1:200 in PBST for 1 h at room temperature. Slides were mounted in PBS-glycerol and fluorescence was examined on a motorized Zeiss Axioplan IIi equipped with a rear-mounted excitation filter wheel, a triple pass (DAPI/FITC/Texas Red) emission cube, and a Zeiss AxioCam B&W charge-coupled device camera. Fluorescence images were pseudocolored and merged with OpenLabs 3.0 software (Improvision Inc., Lexington, Mass.).

Colocalization of TAstV-2 antigen and apoptosis.

To determine if TAstV-2 induced cell death, intestinal sections from control or TAstV-2-infected turkey poults were deparaffinized and antigenic sites were exposed as described above, then incubated with terminal deoxynucleotide transferase labeled with tetramethylrhodamine red fluorescence (in situ end labeling TUNEL analysis; Roche Molecular) for 1 h at 37°C following the manufacturer's instructions. Immediately following TUNEL staining, the sections were washed three times with PBST and stained for TAstV-2 as described above.

NRK soft agar assay.

TGF-β activity was assessed by determining the colony-forming activity of normal rat kidney cells (NRK-49, CRL-1570; American Type Culture Collection, Manassas, Va.) in the presence of epidermal growth factor in soft agar, as described previously (51). Briefly, 5% Noble agar (Difco) was diluted 10-fold in 5% calf serum (Fisher Scientific) in Dulbecco's modified Eagle's medium (DMEM), and 0.5 ml of this 0.5% agar dilution was added per well to a 24-well tissue culture plate and allowed to solidify. Then 100 μl of serum from PBS-inoculated controls or TAstV-2-inoculated birds taken at 1, 3, 5, and 12 dpi containing epidermal growth factor (1 ng; EMD Biosciences, San Diego, Calif.) was combined with 0.6 ml of 0.5% agar and 0.2 ml (2 × 10³ cells) of an NRK suspension in 5% calf serum in DMEM, and 0.5 ml of this 0.3% agar sample solution was added to the cooled base layer. The samples were incubated for 3 to 5 days at 37°C in 5% CO₂, and the total number of colonies greater than 62 μm was quantified with an inverted microscope. Colony formation is indicative of the presence of activated TGF-β (51). All conditions were performed in triplicate.

Mink lung PAI luciferase assay.

Mink lung fibroblasts (Mv1Lu) stably transfected with the TGF-Β responsive plasminogen activator inhibitor (PAI) luciferase reporter (Mv1Lu-PAI) were a kind gift from Daniel Rifkin (New York University) and used to assay for TGF-β activity as described (1). Briefly, Mv1Lu-PAI cells were plated at 1.6 × 10⁴ cells per well in a 96-well tissue culture plate and allowed to attach for 7 h. After the attachment period, cells were cultured for 16 to 18 h at 37°C in 5% CO₂ in 100 μl of DMEM/well with 0.1% bovine serum albumin (Invitrogen) containing recombinant active TGF-β1 (6.25 pM; R&D Systems, Minneapolis, Minn.), 321.25 μg of protein from homogenized embryo intestinal filtrate from PBS- or TAstV-2-infected embryos at 5 dpi, or embryo intestinal filtrate preincubated (40 min at room temperature) with a polyclonal anti-TGF-β neutralizing antibody (Clone 1D11; R&D Systems). Cells were then lysed and luciferase activity was assessed with the Promega luciferase system (Promega) and the Turner luminometer (Turner Biosystems, Sunnyvale, Calif.).

Statistics.

Data comparing body weights and lymphoid organ weights were analyzed by one-way analysis of variance and pairwise multiple comparison with the Student-Newman-Keuls method (SigmaStat; Jandel Scientific, San Rafael, Calif.). The significance level was defined at P < 0.05.

RESULTS

Propagation of TAstV-2 in embryos.

Attempts to propagate TAstV-2 in cell culture with primary turkey embryo fibroblasts, turkey embryo kidney cells, chicken embryo fibroblasts, chicken embryo kidney cells, African green monkey kidney cells (Vero), mink lung epithelial cells (Mv1Lu), Madin-Darby canine and bovine kidney cells, a human colorectal adenocarcinoma cell line (Caco-2), and an ileocecal colorectal adenocarcinoma cell line (HCT-8) were unsuccessful. Therefore, specific-pathogen-free turkey embryos at 20 embryonic days of age were inoculated with a tissue filtrate prepared from healthy or TAstV-2-infected turkey poults and incubated for 1, 3, or 5 days at 39°C. Intestines were removed and tested for TAstV-2 RNA and replication by RT-PCR and in situ hybridization, respectively. RT-PCR analysis on embryo intestines was positive for TAstV-2 at days 1 through 5 postinoculation. In situ hybridization showed extensive viral replication within 1 dpi. TAstV-2 replication increased until 3 dpi (Fig. 1A) and then began to decrease by 5 dpi. No TAstV-2 in situ staining was detected in the control embryos (Fig. 1B). Interestingly, at 5 dpi, TAstV-2-infected embryo intestines were enlarged, thin-walled, and distended. An immense accumulation of intestinal fluid was also observed in the intestines of TAstV-2-infected embryos (Fig. 1C) but not the controls (Fig. 1D). These results demonstrate that turkey embryos support TAstV-2 replication and are a valuable source for in vitro propagation.

FIG. 1.

Propagation of TAstV-2 in embryos. Twenty-day-old specific-pathogen-free turkey embryos were inoculated with a tissue filtrate from healthy or TAstV-2 inoculated turkey poults and incubated for 5 days at 37°C. (A) TAstV-2 replication was detected in inoculated embryo intestines via in situ hybridization. (B) No TAstV-2 in situ hybridization staining was detected in PBS-inoculated embryos. (C) At 5 dpi, TAstV-2-infected embryo intestines were enlarged, thin-walled, fluid-filled and distended compared to controls (D).

Clinical signs and gross lesions of TAstV-2-induced disease.

Inoculation of naïve poults with 10⁶ genomic units of TAstV-2 resulted in 100% of the infected birds developing diarrhea within 24 h of challenge that continued throughout the course of the 12-day experiment (Fig. 2A). Diarrhea was watery, yellow, frothy, and mucus-filled but did not contain undigested food or blood. Control animals had no diarrhea. In addition to the diarrhea, infected birds exhibited statistically significant growth depression compared to uninfected controls (Fig. 2A, P < 0.05). At 5 dpi, there was a ≈27% difference in the growth, and a 38% difference occurred by 12 dpi (Fig. 2A). The TAstV-2-infected birds remained smaller throughout experiments extended to 28 dpi (data not shown).

FIG. 2.

TAstV-2 infection results in growth depression and decreased thymic size. Five-day-old turkey poults were orally inoculated with 10⁶ TAstV-2 particles or with PBS (0.2 ml) and (A) 20 to 45 random poults per group were weighed at days 0, 3, 5, and 12 postinfection. The second y-axis exhibits the percentage of poults exhibiting diarrhea at the same days of infection. (B) The thymi from five random poults per group were weighed on days 3, 5, 9, and 12 postinfection and the average weight per group determined. Stars represent statistically significant difference in weight as determined by one-way analysis of variance. P < 0.05.

Upon necropsy, the intestines of infected poults were distended, dilated, and gas-filled. The intestines appeared to be three to five times the size of those of the noninfected controls. In addition to the macroscopic changes seen in the intestines, we noted that the bursa and thymus, and to a lesser extent the spleens, of the infected animals appeared reduced in size. To examine this further, these organs were removed, weighed, and compared to those of the mock-infected poults. Birds infected with TAstV-2 had a statistically significant decrease in the size of the thymus beginning 3 dpi and continuing through 9 dpi (Fig. 2B, P < 0.05). Calculating the differences as a ratio of organ weight to body weight, we found, at 3 dpi, that the thymus of the TAstV-2-infected group was 36% smaller than that of the control group and 52% smaller at 9 dpi. However, by 12 dpi, there was no difference in the relative thymus size, suggesting that these changes were transient. There were no statistically significant differences in the sizes of the bursa or spleen compared to controls.

Histopathologic lesions.

To investigate the histologic changes resulting from TAstV-2 infection, tissues were examined by routine hematoxylin and eosin staining and light microscopy. In spite of the severe diarrhea, the intestinal lesions were mild. By 2 dpi, there were scattered single degenerating villous epithelial cells, predominantly in the basal portions of the villi (Fig. 3A). These degenerating cells were present through 9 dpi. Crypt hyperplasia was very mild at 3 dpi and continued through 12 dpi. By 5 dpi there was a minimal amount of mononuclear inflammatory infiltrate in the lamina propria that resolved by 12 dpi. Because of the gross changes seen in the thymus, we also examined extraintestinal tissues: bursa, spleen, pancreas, thymus, liver, kidney, bone marrow, skeletal muscle, and blood. No remarkable histologic changes were noted in any of these tissues. No lesions were seen in any of the control tissues (Fig. 3B). These findings demonstrate that TAstV-2 infection resulted in severe diarrhea, growth suppression, and reduction in thymic mass in the absence of widespread inflammation or cellular damage.

FIG. 3.

TAstV-2 infection results in minor histopathologic changes. Histologic lesions in the duodenum of (A) TAstV-2-infected or (B) control poults collected at 2 dpi and stained with hematoxylin and eosin. (A) In TAstV-2-infected poults, histopathologic lesions are limited to scattered single degenerating villous epithelial cells predominantly in the basal portions of the villi (arrows). Original magnification shown in figure. Inset image is a higher magnification of the affected area.

Localization of TAstV-2.

We originally isolated TAstV-2 from the thymus, suggesting that TAstV-2 is present outside the intestines (53). However, no studies to date have examined the distribution of astrovirus during infection. Therefore, we examined the distribution of TAstV-2 at different times postinfection by RT-PCR, isolation of infectious virus, immunofluorescence, and in situ hybridization (Table 1). Not surprisingly, infectious virus could be isolated from the feces and intestines at all time points in the experiment from day 2 onward; however, the levels of virus in the feces at 1 dpi were below the level of detection by RT-PCR. TAstV-2 RNA was also detected by RT-PCR in the thymus, bursa, spleen, liver, kidney, pancreas, skeletal muscle, bone marrow, and in the plasma fraction of infected birds, generally at 3 and 5 dpi, and the thymus and spleen were still positive at 7 dpi (Table 1).

TABLE 1.

Localization of TAstV-2

Sample	TAstV-2 found (dpi)
	RT-PCRa	Virus isolationb	IFAc	In situ hybridizationd
Feces	2-9	1-12	NDe	ND
Intestines	1-12	1-12	1-12	1-9
Bursa	5	5	3-5	Negf
Thymus	3-7	3-7	3-5	Neg
Spleen	3-7	3-7	3-5	Neg
Kidney	3	3-12	5	Neg
Liver	3-5	3	5	Neg
Skeletal muscle	5	5	ND	Neg
Marrow	3-5	3-5	3	Neg
Pancreas	3	3	3-5	Neg
Plasma	3-5	3-5	ND	ND

^aTissues from five random control or infected animals were collected at different days postinfection and pooled at each time point, and RNA was isolated for TAstV-2-specific RT-PCR.

^bTissues collected as described above, homogenized, filtered, diluted 1:100 and inoculated into 20-day-old embryonated turkey eggs. Five days postinoculation, embryos were monitored for clinical signs of infection and intestines were isolated for RT-PCR.

^cImmunofluorescence with a polyclonal antibody produced to a peptide sequence in the TAstV-2 capsid protein (K₆₇₆-R₆₉₁).

^dIn situ hybridization with negative-sense riboprobe specific for the 3′ end of the TAstV-2 genome.

^eND, not determined.

^fNeg, negative at all times tested.

Infectious virus could be isolated from all of the samples generally between 3 and 7 dpi. The presence of TAstV-2 outside the intestines was also detected by immunofluorescence. Mild, limited TAstV-2 capsid staining was detected in all tissues examined, most consistently between 3 and 5 dpi (Fig. 4A, C, and E). No staining was observed in control tissues (Fig. 4B, D, and F). Although there was infectious virus and viral antigen staining in extraintestinal tissues, in situ hybridization data suggested that astrovirus replication was limited to the intestines (Fig. 5A). No replicating virus was detected in representative extraintestinal tissues (thymus, bursa, and spleen) (Fig. 5B, C, D). In situ staining of the TAstV-2 genome in the intestines was generally found in the deep edges of the villi and not in the crypts (Fig. 5A). A similar staining pattern for TAstV-2 capsid protein was observed, with antigen detected in the cytoplasmic portion of specific enterocytes at the mid-region of the villi (Fig. 4E).

FIG. 4.

TAstV-2 is present in extraintestinal tissues. Photomicrograph of the distribution of specific immunofluorescence staining against astrovirus capsid antigen of (A) thymus, (C) bursa, and (E) intestine of 10-day-old TAstV-2-infected poults at 5 dpi and (B) thymus, (D) bursa, and (F) intestine of 10-day-old PBS mock-infected control birds. Original magnification shown in the figure.

FIG. 5.

TAstV-2 replication is limited to the intestines. Photomicrograph of the distribution of specific in situ hybridization staining for TAstV-2 in the duodenum, thymus, bursa and spleen of 10-day-old TAstV-2-infected poults at 5 dpi. (A) Arrows denote in situ positive cells; inset image is a higher magnification of TAstV-2 in situ positive cells. No staining detected in nonintestinal tissues or control tissues. Original magnification shown in figure.

TAstV-2 infection does not increase cell death.

The lack of histologic lesions in the intestines of TAstV-2-infected animals was surprising given the levels of viral replication and diarrhea. To determine if TAstV-2-infected cells undergo cell death, intestinal sections from control and infected poults were double-labeled for TAstV-2 capsid protein and cell death by TUNEL analysis. Not surprisingly, there was a great deal of TUNEL staining in both control and TAstV-2-infected intestines (Fig. 6A to D). In contrast, astrovirus staining was found only in the cytoplasm of enterocytes of infected (Fig. 6F) but not control (Fig. 6E) intestines. Double-labeling the tissues resulted in no overlap of TUNEL-positive cells with TAstV-2-infected cells, suggesting that astrovirus replication does not result in an increase in cell death (Fig. 6G and H). Identical results were observed in TAstV-2-infected embryos (data not shown). These experiments suggest that TAstV-2 does not increase cell death, which supports the histopathology observations (Fig. 3).

FIG. 6.

TAstV-2 infection does not increase apoptosis. Five-day-old turkey poults were orally inoculated with 0.2 ml of PBS or 10⁶ TAstV-2 particles, and formalin-fixed intestines from 10-day-old PBS mock-infected control birds (A, C, E, and G) or 10-day-old TAstV-2-infected poults at 5 dpi (B, D, F, and H) were sequentially stained for cell death with TUNEL and conjugated with rhodamine (A to D) and then with anti-TAstV-2 peptide antibody, followed by biotin, then Alexa488-conjugated avidin (E and F). Panels A to D represent TUNEL alone, panels E and F represent staining for TAstV-2, and panels G and H are merged differential interference contrast, rhodamine, and Alexa488 images. Panels A and B, original magnification shown on figure; white bar on panels C to H represents 20 μm.

Increased activation of TGF-β during TAstV-2 infection.

Given the severity of the clinical disease, the length of diarrhea and virus shedding (>9 dpi), one would expect that an inflammatory response would be initiated. One possible explanation for the apparent lack of inflammation would be the increased expression of an immunosuppressive cytokine during TAstV-2 infection. One of the most potent immunosuppressive cytokines is TGF-β (26). To determine if TGF-β-activity increased during TAstV-2 infection, serum from inoculated turkeys was collected and tested with the NRK colony-forming soft agar assay, a highly specific and sensitive biological assay for TGF-β activity (51). We observed substantial increases in serum TGF-β bioactivity after infection with TAstV-2 (Fig. 7A). TGF-β activity was elevated within 1 dpi and remained increased even at day 12 compared with the control turkeys. A neutralizing antibody against TGF-β inhibited the increased colony formation observed in the infected serum samples, suggesting that the in vivo activity is that of TGF-β (Fig. 7A). Intestinal filtrates were also tested but were toxic in the NRK assay.

FIG. 7.

TAstV-2 increases TGF-β activity. (A) Five-day-old turkey poults were orally inoculated with 0.2 ml of PBS or 10⁶ TAstV-2 genomic units, blood was collected from euthanized poults at 1, 3, 5, and 12 dpi, and sera were isolated. Aliquots of sera (100 μl) were tested for TGF-β activity by the NRK colony-forming soft agar assay. Epidermal growth factor was the negative control for the soft agar assay. Five poults per condition were tested, and the results are expressed as the means of triplicate determinations; error bars indicate standard deviations, while stars represent statistical significance (P < 0.05). To demonstrate specificity, the day 1 sample was preincubated with 2.5 μg of anti-TGF-β clone 1D11 (R&D Systems) for 40 min at room temperature. These results are representative of at least three experiments. (B) Twenty-day-old turkey embryos were infected with TastV-2 (10⁹ genome units) or PBS, and then 5 days postinfection, intestines were removed and homogenized and equal protein concentrations from infected and uninfected embryos were brought to a final volume of 100 μl in DMEM containing 0.1% bovine serum albumin. Duplicate samples were preincubated with 2.5 μg of anti-TGF-β clone 1D11 for 40 min at room temperature to show specificity. Recombinant TGF-β₁ (6.25 pM) was used as a positive control. MV1Lu-PAI cells were incubated with samples for 16 h before luciferase levels were determined. Results are reported as relative light units, and the results are expressed as the means of triplicate determinations; error bars indicate standard deviations, while stars represent statistical significance (P < 0.05).

To determine if TGF-β activity also increases in inoculated embryos, 20-day-old turkey embryos were inoculated with 10⁹ genome units of TAstV-2 or PBS and incubated for 5 days, at which time the intestines were removed, homogenized, and filtered (0.22-μm pore size). These tissues homogenates were then assayed for active TGF-β with the Mv1Lu-PAI (PAI, plasminogen activator inhibitor) luciferase assay. Intestines from TAstV-2-inoculated embryos had 11-fold more active TGF-β than did an equal amount of total protein from PBS-inoculated embryos (Fig. 7B). The increase in luciferase activity was inhibited with a TGF-β neutralizing antibody, demonstrating that active TGF-β in the intestinal homogenates was driving the luciferase expression off the PAI promoter (Fig. 7B). Finally, the intestinal fluid isolated from TAstV-2-infected specific-pathogen-free embryos contained elevated levels of active TGF-β (unpublished data).

DISCUSSION

Very little is known about the pathogenesis of astrovirus infection. Studies in vivo in humans and animals are limited (37, 46, 58, 62, 63, 69). In these studies we described the development of an in ovo method to propagate high titers of infectious virus and a small-animal model that will be useful to further understand astrovirus pathogenesis and the host response to infection. The present studies are the first, to our knowledge, to examine the pathogenesis of astrovirus infection including the kinetics of astrovirus replication, the location of the virus and its ability to localize to extraintestinal sites, and, most surprisingly, the induction of diarrhea in the absence of either cellular damage or an increased inflammatory response.

In vitro all of the human astrovirus (HAstV) strains were adapted to replicate in cell lines (7, 25, 60). To date, we have been unable to propagate TAstV-2 in primary turkey or chicken cells or the cell lines that support HAstV replication. Fortunately, we were successful in propagating TAstV-2 in turkey embryos. Inoculation of TAstV-2 in the yolk sac of 20-day-old turkey embryos resulted in productive viral replication, accompanied by an accumulation of fluid in the intestines of infected embryos. Routine testing of this fluid indicates that it typically contains 10¹¹ viral genomic units/ml as determined by competitive quantitative RT-PCR. Limiting dilutions in embryos followed by immunofluorescent staining for the viral capsid protein suggested that the fluid contained at least 10⁹ infectious viral particles/ml (data not shown).

Although previous animal studies yielded some information about astrovirus pathogenesis, none were fully developed into animal models. Therefore, we set out to determine if we could use the turkey poult model to understand viral pathogenesis. TAstV-2 was highly infectious and extremely stable in the environment (54); therefore, control birds had to be housed in separate rooms to avoid cross contamination. Additionally, placing naïve poults in contact with infected birds or in cages that previously housed TAstV-2-infected birds resulted in immediate infection and diarrhea. Similar to the case with mammalian astroviruses, younger animals are more susceptible to TAstV-2 infection. Infecting older naïve birds with TAstV-2 induced diarrhea; however, the duration of viral replication and the clinical signs were reduced in older animals (data not shown). Infecting naïve poults with TAstV-2 resulted in diarrhea in 100% of the birds within 24 h postinfection. Infected poults had a reduced growth rate and remained significantly smaller than controls throughout the experiment. In addition to the growth depression, infected poults also had significantly reduced thymus weights, although this difference had resolved by the end of the experiment. The mechanism for the reduced growth rate and undersized thymus is not understood; however, both are likely directly related to the diarrhea. Infected birds likely suffer some nutritional deficiencies. Infected birds consumed the same amount of feed as the age-matched controls but did not gain weight at the same rate. In additional studies, birds given nutritional additives did not have as severe weight loss or changes to the thymus.

TAstV-2 RNA and infectious virus were detected in every tissue examined, including the blood. To confirm that TAstV-2 RNA and infectious virus present in nonintestinal tissues were independent of contaminating blood, tissues were washed extensively in PBS or incubated overnight in large volumes of formalin followed by a second 48-h incubation in PBS prior to processing. Thus, it is unlikely that the TAstV-2 is due to contaminating blood. Additionally, we confirmed the presence of TAstV-2 in nonintestinal organs by immunofluorescent staining for the capsid protein. The distribution of viral antigen and RNA throughout nonintestinal organs peaked at 5 dpi and then waned. By 12 dpi, only the intestine contained virus (Table 1). There was limited capsid staining in lymphoid areas of the thymus and bursa and in the kidney epithelia. However, most of the TAstV-2 capsid staining in the extraintestinal tissues was associated with vasculature.

Previously it was unknown if astroviruses induced viremia. In this study, TAstV-2 RNA and low titers of virus were detected in plasma samples from infected poults. Many viruses induce viremia, during which the viruses circulate in the blood, serum, or white blood cells and are spread to target organs to initiate infection (38). The mechanism by which TAstV-2 enters the bloodstream and spreads to extraintestinal organs is unknown. Studies with astrovirus in lambs and calves suggested a possible role for macrophages, Peyer's patches, and M cells in infected animals (57, 68). We demonstrated that macrophages isolated from the spleens of TAstV-2-infected poults did not contain infectious virus (M. D. Koci, L. A. Kelley, D. Larsen, and S. Schultz-Cherry, submitted for publication). Unfortunately, at the current time, markers for turkey-specific APCs are not available. Collectively, these results suggest that viremia occurs following TAstV-2 infection and that the TAstV-2-positive sera contain infectious virus.

Although extraintestinal tissues contained TAstV-2 antigen and RNA, only the intestine appeared to support viral replication, as determined by in situ hybridization. Limited replication was observed in the cecal tonsils and distal small intestine within 1 dpi. By 3 dpi, replication was pronounced in the cells of the mid-villus of the cecal tonsils and distal small intestine (duodenum) with expansion to the epithelium of the large intestine and small intestine. By 9 dpi, only minimal viral replication was observed (4).

Many enteric pathogens induce diarrhea by destroying enterocytes in the villous epithelium, ultimately leading to cell death and villous atrophy (29). This does not appear to be the case with TAstV-2. In spite of the diarrhea, there were only minimal to mild histologic changes in the intestines during TAstV-2 infection. The lack of substantial histologic changes noted in the intestines was supported by TUNEL analysis. TUNEL staining demonstrated that cell death was not increased during infection, either in general or specifically in TAstV-2-infected cells. Similar results were obtained with the apoptosis-specific antibody against caspase 3 (data not shown).

Another common mechanism to induce diarrhea is through an increased inflammatory response. However, there is no increase in inflammatory cell infiltrates in response to TAstV-2 infection. This could be due to a number of factors including the lack of cell death and/or to the upregulation of an anti-inflammatory cytokine by TAstV-2 infection. One such factor is TGF-β. TGF-β is a potent immunomodulatory factor that is important in intestinal homeostasis. In the intestine, TGF-β is produced by enterocytes and localizes primarily at the villus tip in the jejunum (61) and occasionally lymphocytes in the lamina propria are immunopositive (3). Once activated, TGF-β mediates epithelial restitution (11, 45, 61), plays a major role in the development of regulatory T cells (56), potentially leading to generation of a TH3-type response eliciting oral tolerance (65), modulates the severity of inflammatory diseases (6, 30), activates neutrotrophic factors (21, 42, 59), and preserves the epithelial barrier function (43, 44).

There is no information in the literature on the role of activated TGF-Β during viral gastroenteritis. However, based on other models of intestinal injury, we hypothesize that the increased TGF-β-Β may be important in preserving or maintaining the epithelial barrier (3, 11, 20, 49). Our work demonstrated that systemic TGF-β activity was increased during TAstV-2 infection. Increased TGF-β activity occurred within 1 dpi and remained elevated throughout the course of the experiment. This pattern was similar to that seen with the activation of TGF-β by influenza virus (52). Similar experiments with infected embryos demonstrated that TAstV-2-infected intestines contain 11-fold more active TGF-β than do control intestines. This is a significant increase in TGF-β activity. Additionally, the fluid that accumulated in infected embryo intestines following 5 days of incubation contained high levels of active TGF-β (data not shown). These results are unique to TAstV-2. Embryos or poults infected with other enteric pathogens that induce diarrhea, including turkey coronavirus virus and reovirus, failed to activate TGF-β-Β (data not shown). Studies are under way to determine the mechanism of activation and role of TGF-β in astrovirus pathogenesis.

Currently, we do not know how TAstV-2 induces diarrhea. However, it is through a mechanism that does not involve either an inflammatory response or extensive cellular damage. Studies are under way to determine the mechanism of astrovirus-induced diarrhea, including the possibility that astrovirus, like rotavirus and simian immunodeficiency virus, encodes a viral toxin.

In summary, these studies developed a procedure to cultivate large quantities of infectious astrovirus and established the kinetics and distribution of astrovirus replication with an avian astrovirus in a young turkey model. We also showed, for the first time, that an enteric astrovirus induced viremia and extraintestinal distribution of virus. Finally, we demonstrated that astrovirus caused diarrhea without inducing cell damage or cell death and began exploring the cellular response to infection. At this time we cannot compare our results to those observed in a mammalian model since those studies have not been performed. It will be interesting to determine if viremia is a general feature of astrovirus infection. The presence of astrovirus RNA in serum may aid in the rapid diagnosis of infection. Additionally, experiments examining the induction of diarrhea by heterologous strains of astrovirus in the turkey model are necessary to fully explore the potential use of turkey poults or embryos as an animal model for astrovirus infection.

TAstV-2 is phylogenetically distinct from the mammalian astroviruses; however, given the similarities in disease and age distribution, this difference is likely due to the evolutionary distance between mammals and birds (28). Further studies are warranted to determine if the findings in turkeys are generalizable to mammals. Astroviruses are one of the leading causes of viral gastroenteritis worldwide. This animal model will be useful in increasing our knowledge of the mechanisms involved in inducing astrovirus diarrhea and in defining important features of the host response to infection and will possibly lead to improved therapeutics.