Determination of the diversity of astroviruses in feces from cats in Florida

Abstract

Astroviruses are small, nonenveloped RNA viruses that have been linked to numerous diseases in a variety of species, including enteric disease in humans and cheetahs. Species Mamastrovirus 2, previously known as feline astrovirus, has been isolated from the feces of domestic cats and cheetahs. A total of 122 cat fecal samples from Alachua County, FL Animal Services and the Veterinary Community Outreach Program at the University of Florida were analyzed, and 35 contained astroviral RNA that was amplified and identified using consensus RT-PCR and sequence analysis. Using phylogenetic analysis, 19 of the astroviral sequences were identified as Mamastrovirus 2, making it the most prevalent astrovirus in this population. Three samples were identified as an astrovirus similar to viruses previously identified in foxes in The Netherlands and a cat in California, and one was similar to a bat astrovirus. One astroviral sequence was identified as an Avastrovirus. Although a causative relationship between mamastroviruses and enteric disease in cats has yet to be established, it is clear that mamastroviruses are prevalent, and an understanding of prevalence of astroviral types may help direct future test development.

Astroviruses are small, nonenveloped, single-stranded RNA viruses with a “star-like” ultrastructure. The family Astroviridae is divided into 2 genera, Avastrovirus in avian hosts and Mamastrovirus in mammalian hosts.¹⁷ Several strains of Astroviridae have been shown to be pathogenic in nature, suggesting they may be major causative agents of enteric disease. Human astroviruses are a predominant cause of viral gastroenteritis in neonates.⁸ Mamastroviruses have since been associated with diarrhea in diverse species including dogs and cheetahs.^3,24 The species Mamastrovirus 2 (MAstV-2), formerly known as feline astrovirus, has been isolated from cat feces, both diarrheic and non-diarrheic, but has yet to be statistically correlated to enteric disease in cats.^5,13,26

Enteric disease is an important health concern for many mammals within shelter environments. Animal shelters may be an ideal location for the transmission of infectious agents in general, given that animals in these facilities are subject to frequent population changes, high-stress situations, and close contact with other potentially infected animals. Given that pathogens are able to spread rapidly among individuals within the crowded environment of a shelter, the evolutionary pressures to keep the host alive dramatically decrease for the virus, allowing for greater virulence and risk of host fatality. This provides an excellent environment not only for the study of the pathogenesis and epidemiology of a virus, but also for rapid evolution of more pathogenic strains that could have significant impact on animal health. Astroviruses have high rates of mutation, making them capable of rapid evolution. Therefore, they are especially well-suited for proliferation and transformation within the shelter environment.

Although shown to be a viral cause of gastroenteritis in multiple species, isolating and characterizing astroviruses in the laboratory has remained a challenge for researchers. The intestinal villus blunting and inflammation commonly seen with other gastroenteric viruses is minimal or absent in astrovirus infections. Astroviruses increase gut epithelial permeability in the absence of cell death.¹² Therefore, histopathology remains an insensitive test for diagnosis of astroviral diarrhea. Similarly, identification through electron microscopy can be difficult, as astroviruses, caliciviruses, picornaviruses, and parvoviruses are all morphologically similar and can be misidentified.¹⁸ To date, the only reliable methodology for recognition of astroviral presence in feline fecal samples is consensus reverse-transcription (RT)-PCR and sequencing. Astroviral polymerase and capsid consensus RT-PCR protocols have been developed, validated, and applied to detect astroviruses in cheetahs, pinnipeds, and bats.^3,20,6

Knowledge of the diversity of feline astroviruses may enable design of more rapid and less expensive testing, such as quantitative PCR (qPCR). RT-qPCR has been used to detect and quantify the amount of astroviruses in fecal samples of humans, but has yet to be used for astroviruses in the domestic cat.^14,23 Appropriate primer design requires knowledge of where invariant regions lie in target regions of MAstV-2.

Fecal samples were collected from litter boxes at 2 facilities: Alachua County Animal Services (ACAS; Gainesville, FL) and the University of Florida, College of Veterinary Medicine, Veterinary Community Outreach Program (VCOP; Gainesville, FL). Although both facilities serve the local domestic cat population, VCOP provides outpatient care to many animals that are sponsored through rescue groups. This is important to note, because these individuals are often in foster care or are frequently housed in less-crowded facilities than their counterparts at ACAS. Thus, these 2 facilities may represent a difference in environmental stressors to the population. Samples were collected on a weekly basis from ACAS. Neither a systematic nor random sampling method was used in collection from VCOP; rather, samples were collected periodically from October 2012–November 2014, when outpatient animals happened to defecate. The samples were then frozen at −80°C until laboratory analysis was performed. Samples were identified by an ID number or animal name at each facility to reduce the possibility of repeat sampling from the same animal; however, the possibility of repeat sampling from an animal that was presented to both facilities cannot be ruled out.

Feces were aliquoted into 200-mg portions for RNA extraction, suspended in sterile saline, and centrifuged; resultant supernatant was then filtered. RNA was extracted from the filtrate (High Pure viral RNA kit, Roche, Indianapolis, IN). Degenerate primers based on conserved astroviral sequences were used in a nested RT-PCR reaction to amplify conserved regions of astroviral ORF1b (RNA-dependent RNA polymerase) using previously published protocols.^3,20 Round 1 of the nested RT-PCR protocol was performed using the primers Astr4380F (5’-GAYTGGRCNCGNTWYGATGGNA-CIAT-3’) and Astr4811R (5’-GGYTTNACCCA-CATNCCAAA-3’). Round 2 employed the primers Astr4574F (5’-GGNAAYCCMTCWGGICA-3’) and Astr4722R (5-ARNCKRTCATCNCCATA-3’).

Electrophoresis was performed on the PCR products using 1% agarose gels containing ethidium bromide, and bands of interest with an estimated length of 150 or 450 bp (nested vs. un-nested) were extracted (QIAquick, Qiagen, Valencia, CA). Direct sequencing was performed (BigDye terminator kit, Applied Biosystems, Carlsbad CA) with ABI automated sequencers (Applied Biosystems) at the University of Florida Interdisciplinary Center for Biotechnological Research (Gainesville, FL). All amplicons were sequenced in both directions and primer sequences edited out prior to phylogenetic analysis.

Predicted homologous polymerase nucleotide sequences from MAstV-2 were aligned using MAFFT.¹¹ Bayesian analysis was performed using Mr. Bayes 3.2.6 on the CIPRES (Cyberinfrastructure for Phylogenetic Research) server with gamma distributed rate variation and a proportion of invariant sites, and a general time reversible model.^16,21 Porcine astrovirus 1 (KJ533344) was used as an outgroup. Four chains were run, and statistical convergence was assessed by looking at the standard deviation of split frequencies as well as potential scale reduction factors of parameters. The first 10% of 1,000,000 iterations were discarded as a burn in, based on examination of trends of log probability generation. Two independent analyses were performed to prevent entrapment.

For broader examination of all of the viruses detected, homologous amino acid sequences were also aligned using MAFFT.¹⁶ Bayesian analysis was performed using Mr. Bayes on the CIPRES server with gamma distributed rate variation and a proportion of invariant sites and model jumping for amino acid substitution model selection.²¹ Avastrovirus 3 (AFW05431) was used as an outgroup for the alignment.

Maximum likelihood (ML) analyses were performed using RAxML on the CIPRES server as described previously.²² The same outgroups were chosen as were used in the Bayesian analyses. Amino acid substitution models were selected using ProtTest.¹ To test the strength of the tree topology, bootstrap analysis was used (1,000 re-samplings).

Of the 122 samples collected and analyzed, 35 provided sequences resembling MAstV-2 ORF1b (ADF57166). Using BLASTX (https://goo.gl/ct5JQ3), 30 of those sequences were classified as MAstV-2 with greater than 90% conserved sequence to the RNA-dependent RNA polymerase of cheetah astrovirus 1 (ACD13862), a MAstV-2 strain.² Three samples (Fc13022, Fc13023, and Fc13038) were closely related, and had some homology with a fox astrovirus (AGK45543). Two additional samples (Fc14009 and Fc13004), were astroviral, but unique. Sequences were deposited in GenBank (KT598665–KT598687).

Only sequences >200 bp were considered for phylogenetic analysis. Twelve astroviral sequences were subsequently excluded from the analysis given their length of sequence (11 MAstV-2 sequences and 1 “fox-like”). Although accepted sequences were sometimes as short as 200 bp, they showed no significant nucleotide differences over the available sequence.

Phylogenetic analysis found that 19 of the astroviral samples were included in the MAstV-2 clade with 97.4% posterior probability and ML bootstrap value of 50.4% (Fig 1). Furthermore, there was 100% posterior probability (87.7% bootstrap) that 2 of the samples, Fc13022 and Fc13023, were most closely related to the fox astrovirus (AGK45543), and to a previously identified feline astrovirus (YP009052461) with 99.1% posterior probability (96.1% ML bootstrap). One astrovirus sample, Fc14009, proved to be distinct. There was also 100% posterior probability (88.2% bootstrap) that one sample, Fc13004, was seated in the Avastroviridae clade, with 99.2% posterior probability (100% ML bootstrap) of clustering with avastrovirus 3. There was 100% posterior probability and 100% ML bootstrap support that a distinct canine astroviral clade, bovine clade, human Mamastrovirus 1 clade, bat clade, and sea lion/Mamastrovirus 10 clade occur (Fig 2).

Figure 1.

Bayesian phylogenetic analysis of the feline Alachua County astroviruses within the Mamastrovirus and Avastrovirus clades. Red circles denote the samples obtained from Alachua County Animal Services, and red triangles denote the samples obtained through the Veterinary Community Outreach Program at the University of Florida. The first numeric value indicates the Bayesian posterior probability, and the second indicates maximum likelihood bootstrap value.

Figure 2.

Bayesian phylogenetic analysis of the Florida feline astroviruses within the Mamastrovirus 2 clade. Symbols and numbering are consistent with Figure 1.

Within the MAstV-2 clade, a majority of the samples cluster in an individual clade with 65% posterior probability and 76.5% ML bootstrap value (Fig 1). Furthermore, there is 99% posterior probability (99.8% ML bootstrap) that 7 of these samples are in the same group, and 99% posterior probability (99.7% ML bootstrap) that 9 of the samples are in a separate but similar group. There is 71% posterior probability (82.1% ML bootstrap) that one of the samples, Fc13081, is most closely related to a previously identified MAstV-2 in diarrheic cat feces in a shelter in California.²⁶ Two samples, Fc13003 and Fc13007, form a separate group within MAstV-2 with 100% posterior probability (87% ML bootstrap). There also is a grouping of the previously identified cheetah astrovirus (EU650332) with 81% posterior probability (72.3% ML bootstrap) with MAstV-2 sequences from domestic cats in Hong Kong and Korea.^5,13

A relatively low number of fecal samples were collected from Alachua County animals, and yet almost 30% those samples were positive for astroviral RNA. This suggests that astroviruses are a common and potentially important infectious agent within the Alachua County domestic cat population. Of the astroviruses detected, MAstV-2 was identified most commonly. This is significant, because MAstV-2 has been directly linked to enteric disease in cheetahs and has been proposed to cause enteric disease in the domestic cat, as well.^3,19

Within the MAstV-2 clade, there appear to be 3 separate groupings of astroviruses present in Alachua County, FL (Fig 1). It is important to note that, even though ACAS samples represented <45% of the sampled population, 15 of the 16 most closely related MAstV-2 sequences were from samples collected at the ACAS site. This supports the idea that high-volume, high-stress shelter environments are more likely to promote the transmission and evolution of astroviral strains.

Although further epidemiologic data are needed on MAstV-2 in different shelter populations, MAstV-2 appears to be the most prevalent astrovirus in feline feces from Alachua County. However, other mamastroviruses were also identified. The second most common group of Mamastroviridae identified from feces in our study was first identified in foxes in The Netherlands, but has since been identified in 1 cat in California and 3 cats in our study. Astroviruses are very small viruses with high mutation and recombination rates, enabling rapid adaptation to new hosts.^4,7,15 Thus, low host fidelity may explain the presence of a fox astrovirus–like virus in cat feces. This is also supported by the presence of porcine and bat astroviruses scattered throughout the phylogenetic tree.^6,10

One of the samples, Fc13004, was determined to be an Avastrovirus most closely related to avastrovirus 3. Previous studies^9,25 have shown that viruses known to infect a food source are often identified in the feces of animals. Although direct infection of the cat cannot be ruled out, it seems plausible that the cat corresponding to sample Fc13004 ingested an avastrovirus 3–infected bird and passed the virus in its feces. This may also be the case for sample Fc14009, which contained a sequence similar to that of previously established bat astroviruses.

Although mamastroviruses have been identified from many species with enteric disease, including the cheetah and the domestic cat, more studies are needed to determine the causative relationship between astroviruses and enteric disease in the domestic cat. Standard RT-PCR with sequencing remains the most effective way to identify astroviral RNA in clinical samples, but the data obtained herein will assist in the rational design of quantitative RT-PCR assays targeting conserved areas of specific astroviral clades, which will allow for rapid, sensitive, specific, and less expensive quantification of the astroviral load in fecal samples of domestic cats. Such assays may be useful in understanding epidemiologic correlates of astrovirus infections, thus enhancing our knowledge of these viruses within feline populations.