Epizootic mortalities in American Crows (Corvus brachyrhynchos) during the winter months, referred to as winter mortality of crows, have been recorded in North America for almost two decades. The most common postmortem findings include necrotizing enteritis, colitis, and fibrinous splenic necrosis. These findings are proposed to be due to infection with a Reovirus sp. Our objectives were to characterize the pathology and seasonality of the epizootics in New York State (NYS), confirm the causative role of an Orthoreovirus sp., and determine its phylogeny. On the basis of our proposed case definition for reovirosis, we examined case data collected by the NYS Wildlife Health Program for 16 yr. A total of 558 cases of reovirosis were recorded between 2001 and 2017. Reovirosis had a clear seasonal presentation: cases occurred almost exclusively in winter months (71% in December–January). Detailed data from a 2-yr period (2016–17) demonstrated that reovirosis caused up to 70% of all recorded crow deaths during epizootic months. Crows with positive orthoreovirus isolation from the spleen or intestine were 32 times more likely to die with characteristic histologic lesions of enteritis or enterocolitis and splenic necrosis than crows with negative isolation results. An in situ hybridization probe specific to virus isolated from NYS crow reovirosis cases demonstrated a direct association between viral presence and characteristic histologic lesions. Sigma C (capsid protein) sequences of isolates from NYS crows showed high homology with Tvärminne avian virus, recently proposed as a novel Corvus orthoreovirus clade, and only distantly related to the avian orthoreovirus clade. Our study indicated that a novel orthoreovirus was the cause of winter mortality (or reovirosis) of American Crows and placed the NYS isolates in the newly proposed genus of Corvid orthoreovirus.
Since the turn of the century, large numbers of American Crows (Corvus brachyrhynchos), sometimes reaching the hundreds, have been found dead during cold winter months in the northeastern US. Winter mortality of crows, as such events have come to be known, have also been reported elsewhere in North America, from Ontario (Campbell et al. 2004, 2008) to California (Affolter et al. 2007; Meteyer et al. 2009). Epizootics tend to occur in consecutive or nonconsecutive years at a time of year when crows form large roosts and are, therefore, in constant direct contact with each other (Meteyer et al. 2009). The most common pathologic findings include necrotizing typhlocolitis (Affolter et al. 2007), necrotizing enteritis, fibrinous splenic necrosis, and occasional necrotizing hepatitis (Campbell et al. 2004, 2008; Meteyer et al. 2009). On the basis of the electron microscopy findings and viral isolation from tissues of affected crows, an orthoreovirus was postulated as the primary etiologic agent (Affolter et al. 2007; Meteyer et al. 2009). Immunohistochemical staining of characteristic lesions with antiavian reovirus antibodies supported those findings (Dobson et al. 2009). Conclusive evidence, however, such as fulfilment of Koch's postulates, is thus far lacking. The reported presence of reoviruses in the respiratory and digestive tract of healthy, or at least subclinical, domestic and wild birds (Hollmén and Docherty 2007), including an avian orthoreovirus (ARV) detected in a crow of an unspecified species in Japan (GenBank accession no. LC164026), has raised the question of whether reoviruses are the primary etiology or an incidental finding in epizootics of winter mortality recorded in American Crows.
The family Reoviridae consists of nonenveloped double-stranded RNA viruses of worldwide distribution. The genus Orthoreovirus, belonging to the subfamily Spinareovirinae, includes five species, typically aggregated in three clades: one clade consists of all mammalian orthoreoviruses, another clade includes ARV and Nelson Bay (isolated from a flying fox, Pteropus poliocephalus) orthoreoviruses, and a third clade contains baboon and reptilian orthoreoviruses. Recently discovered psittacine and Steller sea lion orthoreovirus strains (PsRV and SSRV, respectively) seem to be most closely related to the ARV and Nelson Bay species (Benavente and Martínez-Costas 2007; Palacios et al. 2011; Bányai et al. 2014). Disease associated with ARV infection has been typically restricted to domestic bird species, usually those grown in commercial operations, and manifested most commonly as arthritis and tenosynovitis, with less frequent myocarditis, pancreatitis, hepatitis, and enteritis (Tanyi et al. 1994; van Loon et al. 2001; Shivaprasad et al. 2009; Davis et al. 2012). Necrotizing hepatitis, splenitis, and typhlocolitis or enteritis are more frequently reported from wild or nongalliform species, such as psittacines (van den Brand et al. 2007), geese (Palya et al. 2003), Pekin ducks (Anas platyrhynchos; Liu et al. 2011), and corvids, including magpies (Pica pica; Lawson et al. 2015), and American Crows (Affolter et al. 2007; Dobson et al. 2009; Meteyer et al. 2009; Nemeth et al. 2016). Electron microscopy, virus isolation, and immunohistochemical staining of viruses associated with the earliest reported epizootics of American Crow mortalities in the US (Affolter et al. 2007; Dobson et al. 2009; Meteyer et al. 2009), as well as electron microscopy and sequencing of the virus associated with a single case of necrotizing hepatitis and splenitis in a magpie in the UK (Lawson et al. 2015), placed the virus within the ARV species. An isolate from a Hooded Crow (Corvus corone cornix) in Finland, which died after exhibiting abnormal behavior consistent with neurologic disease, suggested that corvids might be affected by a diverging orthoreovirus species. The isolate from this Finnish crow, Tvärminne avian virus (TVAV), could not be placed within the ARV clade on the basis of the sequence of its sigma C protein, an outer-capsid type-specific protein that mediates cell attachment (Benavente and Martínez-Costas 2007; Huhtamu et al. 2007). Subsequent full genome sequencing of the TVAV placed the isolate closer to SSRV and PsRV than to ARV (Dandár et al. 2014). Thus, TVAV has been proposed as a distinct virus species provisionally named Corvid orthoreovirus (Dandár et al. 2014).
Epizootics of winter mortality have been observed in American Crows in New York State (NYS) since at least 2001, when surveillance for West Nile virus increased crow diagnostic submissions. Postmortem examination typically revealed lesions similar to those reported in epizootics elsewhere in North America (i.e., necrotizing enteritis, typhlocolitis, and fibrous splenic necrosis; hepatitis was only occasionally observed). Virus isolation of several cases through the years, starting with an American Crow found dead in December 2001, resulted in the identification of a syncytial-forming virus consistent with an Orthoreovirus sp. During the winter of 2016–17 high numbers of dead crows were examined by wildlife pathologists of the NYS Wildlife Health Program (WHP). The NYS WHP conducts general and targeted surveillance of mortality in free-ranging wild species, with special interest in mortalities involving multiple individuals. Given the characteristic lesions observed and the time of year, an epizootic of winter mortality of American Crows, henceforth referred to as reovirosis, was suspected. To further investigate these mortalities and their link to orthoreovirus infection, all crow mortality cases recorded by the NYS WHP in 2016–17 were used to characterize the pathology and seasonality of winter mortality of crows, confirm the causative role of orthoreoviruses in the development of lesions, and determine the phylogeny of the orthoreovirus involved in NYS crow epizootic mortalities.
MATERIALS AND METHODS
Postmortem examination and reovirosis case definition
Detailed data on all crows examined during two consecutive years (2016 and 2017) were compiled and analyzed to investigate whether carcasses with characteristic lesions of winter mortality (i.e., necrotizing enteritis, typhlocolitis, and splenic necrosis) were confined to the cold season and whether they were associated with the presence of an orthoreovirus. Information collected included date found, age (established on the basis of plumage, color of the oral mucosa, feather status, presence or absence of bursa of Fabricius, and, when available, banding records), sex, gross lesions, histologic lesions, virus isolation (from spleen or intestine), and cause of death diagnosis. Of all crows submitted to the NYS WHS program during 2016–17 (n=247), only crows that underwent at least a gross postmortem examination (n=194) were included in the analysis. Most of the 194 necropsied crows also underwent other tests, such as histopathology and virus isolation. Additionally, historical records for crows examined both at the Animal Health Diagnostic Center and through the NYS WHP from 2001 to 2015 were added to the 2016–17 data to determine seasonality. Records in the historical data ranged from single tissue samples submitted for virus isolation (usually spleen or intestine) to full postmortem examination with ancillary testing. Although records sometimes consisted of multiple crows found dead at the same time and in the same location, all data analysis was performed at the individual level.
Per our case definition, reovirosis was considered confirmed as the cause of death if the individual crow had histopathologic evidence of necrosis in the spleen or intestine and either positive orthoreovirus isolation from fresh tissue (spleen or intestine) or positive in situ hybridization (ISH) staining in the lesions. Probable reovirosis corresponded to the presence of either gross evidence of spleen or intestinal necrosis without histopathologic confirmation but with positive orthoreovirus isolation or histopathologic evidence of necrosis in spleen or intestine without orthoreovirus isolation or ISH. Suspect reovirosis corresponded to crows with gross necrosis in spleen or intestine but with no confirmatory testing (no virus isolation, histopathology, or ISH). For the purpose of our study, negative crows were those that lacked lesions suggestive of reovirosis and were confirmed to have died of a cause other than reovirosis during 2016–17. To determine the odds of developing disease in the presence or absence of the virus, spleen samples from a randomly selected subset of the negative crows were processed for virus isolation.
Virus isolation, amplification, sequencing, and development of an ISH probe
Virus isolation was performed on splenic tissue samples homogenized in minimum essential medium (MEM) with Earle's salts (MEM-E; catalog no. 10370, Gibco, Grand Island, New York, USA), 0.5% bovine serum albumin, penicillin (200 U/mL), streptomycin (200 μg/mL), and fungizone (2.5 μg/mL), at a ratio of one part of tissue to nine parts of medium. Following tissue disruption, samples were subjected to low-speed centrifugation to remove large particulate matter. Approximately 1 mL of supernatant was used to inoculate Vero cells (ATCC® CCL-81™, American Type Culture Collection, Manassas, Virginia, USA) in each T-25 flasks. The inoculum remained on the cells for 1.5–2 h at 37 C with periodic swirling. Following removal of the inoculum and one rinse with 3 mL of phosphate-buffered saline, 6 mL of MEM-E with 10% irradiated fetal bovine serum was added to each flask. Cultures were monitored daily for evidence of cytopathology. At 5- to 7-d intervals, monolayers were subcultured with one new monolayer established to continue the isolation procedure. Cultures showing cytopathology consistent with a virus infection were processed for fluorescent antibody testing with a polyclonal antibody to chicken-origin ARV (catalog no. ab33986, Abcam, Cambridge, Massachusetts, USA), PCR assays, or negative stain electron microscopy.
Amplification and sequencing of orthoreovirus sigma C sequences from our crow isolates used primers selected on the basis of alignments of TVAV S1 with other ARV sequences (Table 1). Total nucleic acid was purified from culture lysates by using a commercial kit (catalog no. 52906, Qiagen, Valencia, California, USA). Amplicons were generated by using SuperScript III One-Step RT-PCR (catalog no. 12574, Invitrogen, Carlsbad, California, USA) with thermocycling conditions of 48 C for 30 min, 95 C for 5 min, and 40 cycles of 95 C for 30 s, 59 C for 30 s, and 72 C for 60 s. Sequencing was performed by the Cornell University Biotechnology Resource Center (Ithaca, New York, USA) and was edited and analyzed by using Lasergene 15.0 (Molecular Biology Suite, DNASTAR, Madison, Wisconsin, USA). Complete sigma C sequences for three isolates from the 2016–17 outbreak and archived NYS isolates from 2001 and 2008 were accessioned in GenBank (accession nos. MK396070–MK396074).
A partial sigma C consensus sequence of orthoreoviruses isolated from the spleens of four crows diagnosed with confirmed reovirosis during the 2016–17 epizootic was generated with primers 658F and 1642R (Table 1). The consensus sequence was submitted to Advanced Cell Diagnostics (Newark, California, USA). Validation of the ISH probe (RNAscope® Probe-V-CbRV, catalog no. 501641, Advanced Cell Diagnostics) was performed on uninfected and Corvid orthoreovirus–infected Vero cells suspended in an aqueous gel (HistoGel™, Thermo Fisher Scientific Inc., Winsford, Cheshire, UK) prior to being embedded in paraffin and processed as routine histologic samples. Tissues from three crows with characteristic necrotizing enteritis and multifocal splenic necrosis from the 2016–17 epizootic, including one spleen that yielded one of the isolates for the consensus sequencing, were stained with the ISH probe. Protocol (see Supplementary File 1) generally followed manufacturer's specifications for manual staining and included both uninfected and orthoreovirus-infected Vero cells as negative and positive controls.
Orthoreovirus phylogenetics: sigma C gene sequence
Evolutionary analysis was conducted in MEGA7 (Kumar et al. 2016). Sequences were aligned by using ClustalW (Thompson et al. 2002). Phylogenetic relationships were estimated by using the neighbor-joining method (Saitou and Nei 1987). Evolutionary distances were calculated by using the maximum composite likelihood method (Felsenstein 1985). Bootstrap test support percentages were determined by using 500 replicates (Tamura et al. 2004).
Reovirosis mortalities in American Crows
Of the total 247 American Crow carcasses that were submitted to the NYS WHP for cause of death determination during 2016 and 2017, 194 underwent a necropsy. Of the 194 necropsied, reovirosis was considered probable or suspect in 33 (17%) cases on the basis of gross or histopathologic evidence of necrotizing enteritis or splenic necrosis, as per our case definition. In 21 of those cases, cytopathology with syncytial cell formation characteristic of orthoreovirus was observed in 86% (18/21), confirming reovirosis in most probable and suspect cases (Table 2).
The odds of developing lesions associated with orthoreovirus infection were calculated on the basis of a subset of the total 194 crows necropsied during 2 yr (2016–17). Forty-six crows (21 orthoreovirus suspect and 25 negative crows) were tested for orthoreovirus infection via virus isolation from splenic tissue (Table 2). Eleven of the 25 negative crows were found in December or January; the others (n=13) were found in February (n=8), March (n=3), June (n=1), August (n=1), and September (n=1). Compared with the 86% positive virus isolation in suspect reovirosis cases, 21 (84%) of spleens from 25 crows with no evidence of necrotizing enteritis or splenic necrosis yielded no virus growth, while the other four (16%) were positive for orthoreovirus growth. The four crows with no lesions but positive virus isolation were found in December (n=1) or January (n=3). Based solely on these data, the odds of an American Crow developing necrotizing enteritis and splenic necrosis were approximately 32 times greater when the virus was isolated from the tissue than when the virus was absent (Fisher's exact test, P<0.001).
Of all reovirosis cases (suspect, probable, and confirmed) in the combined data (2001–17), 70% (391/558) were recorded in the months of December (30%, 167/588) and January (40%, 224/558). In addition, 17% (94/558) were recorded in February, 8% (43/558) in March, with 0–2% recorded in any of the other months (Table 3). Mortalities associated with reovirosis in historical data (2001–15) followed the seasonal distribution, but the total number of cases varied greatly by year (see Fig. 1a and Supplementary Material Table S1). Mortalities due to reovirosis (probable or confirmed) represented 17% (33/194) of the total number of crows necropsied as part of the NYS WHP during a 2-yr period and 70% (23/33) of those examined during the months of December 2016 and January 2017 alone (Table 2 and Fig. 1b). The only two crows with suggestive lesions found outside of the winter months (December–January) could not be confirmed as reovirosis cases. One, found in March 2017, had gross evidence of necrotizing enteritis, but no histopathology or virus isolation was performed. The other, found in May 2017, had both gross and histopathologic necrotizing enteritis but was negative for virus isolation from intestinal tissue. From the 2016 and 2017 data, the odds of developing reovirosis (confirmed or probable) were 79 times greater during the months of December and January than between February and November (Fisher's exact test, P<0.001).
In situ hybridization
The ISH with a probe designed to target a consensus sequence of the sigma C gene from orthoreovirus isolated from American Crows was developed. Validation was performed on Vero cells: the probe specifically and intensely stained cells infected with an isolate of orthoreovirus from one of the infected crows (Fig. 2). When the ISH was performed on the lesions in the intestine and spleen of a subset of three crows with confirmed reovirosis on the basis of positive virus isolation, strong staining was noted in direct association with the lesions in all three cases (Fig. 3). Infected and uninfected Vero cells were used as positive and negative controls.
Phylogenetic analysis of Orthoreovirus involved in mortalities of American Crows
Sequencing of the 981-base pair sigma C open reading frame in viruses isolated from five NYS American Crows showed 97–99% identity between NYS isolates and 91–93% identity between NYS isolates and TVAV (Huhtamo et al. 2007). Amino acid sequences were 98–100% identical between the NYS isolates and 93–94% identical to TVAV (Huhtamo et al. 2007). The close phylogenetic relationship between the NYS and the Finnish isolate as compared with other avian origin orthoreoviruses is depicted in Figure 4. This virus (TVAV) is postulated to be a new species of Orthoreovirus, tentatively named Corvid orthoreovirus (Dandár et al. 2014).
Winter mortality has been observed in American Crows in NYS since at least 2001. Our intensive investigation of an epizootic mortality event during the winter of 2016–17 confirmed that an orthoreovirus was the etiologic agent. Gross and histologic lesions were mostly consistent with previous reports associated with putative or confirmed reovirus infection in nongalliform species, including corvids (Campbell et al. 2004; Lawson et al. 2015). The characteristic necrotizing enteritis, typhlocolitis, and splenic necrosis in the NYS crows closely resembled the lesions reported in crows from Ontario, where reovirus was originally proposed as the likely cause of epizootic winter mortalities (Campbell et al. 2004). Similar to the single case of a magpie in the UK reported to have died from a reovirus infection, the NYS crows had severe splenic necrosis (Lawson et al. 2015). Unlike the UK corvid case, however, necrotizing stomatitis and hepatitis were not a feature of the disease in NYS crows, while necrotizing enteritis or typhlocolitis were common in the crows but absent in the magpie (Lawson et al. 2015). The intestines of the UK magpie were not examined histologically, however, because of advanced autolysis (Lawson et al. 2015), so enteritis or enterocolitis could have been present. The difference in presentation could be due to the idiosyncratic difference in host species, or perhaps more likely, it could be the result of infection with a different orthoreovirus species.
Our records showed that reovirosis caused the death of approximately one-fifth of all crows examined by the NYS WHP in a 2-yr period of time. Mortalities were concentrated in the winter months, when reovirosis became the leading cause of death. This is exemplified by the 2016–17 epizootic, when 70% of all recorded deaths were attributed to reovirosis. The seasonal presentation of the epizootic in NYS crows was demonstrated both by historical data and the presentation of the 2016–17 epizootic. Although data from the 2016–17 outbreak seemed to support seasonality more than the historical records from 2001–15, note that the date of death for some of the crows in the latter period could actually correspond to the date of necropsy or sample submission, which sometimes happened months afterwards. Even if the dates in the historical data are completely accurate, it is obvious that the vast majority of cases occurred during December and January. Seasonality was similar to other crow mortalities putatively associated with reovirosis, which are reported almost invariably in the coldest months in eastern North America (Campbell et al. 2004, 2008; Meteyer et al. 2009). Elucidating the basis for such seasonality is beyond the scope of this work, but impaired immunity during harsh weather conditions and increased fecal-oral transmission due to overcrowding during winter roosting are potential factors (Hollmén and Docherty 2007; Meteyer et al. 2009). The number of reovirosis cases in the historical data varied greatly by year, suggesting a possible trend in the presentation of the epizootics. This, however, could have been due simply to submission bias as, unfortunately, we could not examine the historical data to compare the percentage of reovirosis versus other causes of mortalities. Additionally, apparent trends could be due to our analysis of mortalities at the individual rather than the incident level.
On the basis of a consensus sequence obtained from orthoreovirus isolates from NYS crows, a specific ISH probe was developed and validated. Granted the specificity of the probe was not tested on other orthoreovirus clades, such as ARV, with which it could potentially cross react. However, this is unlikely as the consensus sequence used for the probe is different from other orthoreovirus clades. More importantly, the purpose of the probe was not to distinguish between clades but to demonstrate the presence of virus in histopathologic sections. By applying the probe to a subset of tissues from crows that died during the 2016–17 epizootic, we were able to directly associate this novel orthoreovirus with the presence of characteristic lesions. Although previous researchers suggested an association between orthoreoviruses and corvid epizootics characterized by enterocolitis and splenic necrosis (Campbell et al. 2004; Affolter et al. 2007; Meteyer et al. 2009), conclusive evidence was lacking. After the original epizootics were recognized at the turn of the century, an orthoreovirus was proposed as the most likely cause of death, on the basis of virus isolation of archived tissue samples from 2003 (Meteyer et al. 2009). That study based its findings on virus isolation alone, however, and recognized the presence of other potential pathogens, thus leaving some questions regarding the causal association of reovirus infection, lesion development, and mortality. Work done in Canada suggested that a different virus, perhaps a rotavirus or a paramyxovirus, could instead be responsible (Campbell 2008). Scepticism regarding the direct causal association of orthoreoviruses and mortality in crows was based on the reported presence of reoviruses in domestic birds that never developed clinical disease (Hollmén and Docherty 2007) and the occasional finding of ARV sequences in corvids (e.g., GenBank accession no. LC164026), which raised the possibility of subclinical (incidental) infections. Using an ISH probe developed specifically to target viruses isolated from NYS crows, we were able to demonstrate that an orthoreovirus was directly associated with the characteristic lesions of winter mortality, specifically necrotic lesions in the intestine and spleen, thus confirming it as the etiologic agent. We calculated that presence of the virus made crows 32 times more likely to develop lesions, further supporting the etiologic link between the virus and the disease. The four crows with no lesions but positive virus isolation from tissues could have been individuals innately less susceptible to developing disease. Alternatively, considering they were all found during the winter months of December and January, they could have been at an early stage of infection before lesions could be detected, particularly as they were not examined histologically. Also, those four crows were examined as part of incidents that involved confirmed reovirosis cases, and sample collection was performed without changing instruments, potentially yielding false-positive results. Thus, the odds of developing disease in association with presence of this novel orthoreovirus may be even higher than we observed.
Sigma C sequences of isolates from NYS crows with reovirosis show high homology with TVAV (Dandár et al. 2014). Sequencing of the sigma C gene, encoding a capsid protein involved in viral cell attachment, originally suggested that the TVAV isolate from a Finnish Hooded Crow does not fit within the ARV clade, where most other avian isolates cluster (Huhtamo et al. 2007). Further sequencing of the genome from the Finnish isolate support its location outside the ARV clade and suggest the existence of a novel Orthoreovirus species, temporarily named Corvus orthoreovirus (Dandár et al. 2014). Sequence analysis of a subset of the NYS isolates involved in epizootic mortalities revealed high homology to the TVAV, indicating they are closely related to the novel Corvus orthoreovirus and not within the ARV clade. The Hooded Crow source of TVAV suffered from incoordination, abnormal posture, cramps and paralysis, suggesting damage to its nervous system (Huhtamo et al. 2007). Unfortunately, no gross or histopathologic examinations are reported, so it is unknown whether the clinical signs represented agonal changes or were related to viral lesions in the nervous system, and whether there were lesions in the intestine or spleen of this crow that may have resembled those found in reovirus-infected American Crows. Dandar et al. (2014) speculated that the Hooded Crow could have been an incidental host, on the basis of the similarities between TVAV and Stellar sea lion orthoreovirus, or that TVAV has low host fidelity. We found the virus to be capable of infecting and causing disease in corvids, seemingly targeting them as its primary host. Interestingly, the sigma C protein sequence of the virus isolated from the UK magpie (Lawson et al. 2015) is not closely related to the TVAV virus but clusters among poultry isolates, unlike the isolates from our NYS crows. This could explain the difference in lesions between our crows and the UK magpie and Japanese crow (GenBank accession no. LC164026) and suggests that the difference is due to infection with different viruses rather than being host specific.
Our study indicates that a novel Orthoreovirus is the cause of winter mortality, or reovirosis, of American Crows by demonstrating its direct presence in characteristic lesions and the high odds of developing disease if infected. It further establishes its close relation to Tvärminne orthoreovirus, placing it in the newly proposed genus of Corvid orthoreovirus.
We thank Erica Sloma and Andrew Miller for help with in situ hybridization staining and Raphaël Vanderstichel for the production of the rose diagrams. We thank the pathology residents, laboratory and necropsy technicians, and other support staff at various institutions for their invaluable contributions. Our work was supported by Federal Aid in Wildlife Restoration grant W-178 and the New York State Department of Environmental Conservation.
Supplementary material for this article is online at http://dx.doi.org/10.7589/2019-01-015.