We performed metagenomic sequencing for each animal using pooled RNA extracted from the cerebrum, cerebellum, brain stem, and spinal cord as previously described (6,7). We analyzed the sequences obtained using the MiSeq System (Illumina, San Diego, CA, USA) by using Kraken, an ultrafast and highly accurate program for assigning taxonomic labels by examining the k-mers within a read and querying a standard Kraken database with those k-mers (8). We assembled reads de novo using CLC Genomics Workbench (QIAGEN, Valencia, CA, USA) and identified the contigs by blastn (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The largest contig, encompassing ≈2,000 reads, encoded a near-complete astrovirus genome of 6,461 nt and was designated PoAstV3/USA/IA/7023/2017 (GenBank accession no. KY940545). This sequence originated from a sow sample. A near-complete astrovirus genomic sequence was also obtained from the piglet (contig length 5,935 bp; E = 0) and had 100% nucleotide identity to PoAstV3/USA/IA/7023/2017. We also identified porcine endogenous retrovirus (contig lengths 1,865 bp and 1,317 bp; E = 0) in sow samples. When using a minimum contig length of 500 nt, we identified rocilivirus (contig length 832 bp; 32 reads; E = 0; GenBank accession no. KU058672.1) in piglet samples.

Figure

Figure. Phylogenetic trees of capsid protein (A), RNA-dependent RNA polymerase protein (B), and whole-genome nucleotide (C) sequences of a PoAstV type 3 strain (PoAstV3/USA/IA/7023/2017, filled circle) from central nervous system tissues of...

Phylogenetic comparisons of the capsid protein sequence, RNA-dependent RNA polymerase protein sequence, and whole-genome nucleotide sequence placed PoAstV3/USA/IA/7023/2017 in the same cluster as other strains of PoAstV-3 (Figure, panels A–C). The isolate we identified was most closely related to PoAstV3/USA/US-MO123 (GenBank accession no. NC_019494.1; 94.1% amino acid identity; Technical Appendix Table 1), which was detected in a swine fecal sample (9). On the basis of these phylogenetic analyses, PoAstV3/USA/IA/7023/2017 is more closely related to neurotropic astroviruses from humans, minks, cows, and sheep (2‒5) than to PoAstV-1, PoAstV-2, PoAstV-4, and PoAstV-5.

We detected viral RNA by using a PoAstV-3 quantitative real-time PCR with previously fresh-frozen CNS tissues from animals with polioencephalomyelitis (Technical Appendix Table 2). We did not detect viral RNA in serum, feces, lung, liver, kidney, or spleen samples of animals with histologic lesions or any sample from animals without histologic lesions (9).

We describe the identification and genetic characterization of PoAstV-3 in CNS tissue from a piglet and sows with neurologic signs and histologic lesions compatible with a neurotropic virus similar to those described in neurotropic astrovirus cases in other species (2‒5). In humans, disease is primarily associated with immunocompromised patients. In cows, the virus is not commonly detected in feces, and the disease does not appear to be associated with immunocompromised animals (4). In this case, PoAstV-3 was not detected in feces of affected animals, and evidence of immunosuppression was lacking. The overall PCR prevalence of PoAstV-3 in feces of pigs in North America is reported to be low (1.2%) (10).

The PoAstV-3 we identified had 92.2% nucleotide sequence similarity to PoAstV-3 identified from a survey that evaluated feces samples from pigs (9).The significance of this finding is unclear. Investigations are needed to clarify the ecology and epidemiology of PoAstV-3 and the pathophysiology of neurotropic astroviruses. Studies have demonstrated the potential for recombination between porcine and human astroviruses, suggesting zoonotic potential (9,10).