NOVEL POXVIRUS INFECTION IN NORTHERN AND SOUTHERN SEA OTTERS (ENHYDRA LUTRIS KENYONI AND ENHYDRA LUTRIS NEIRIS), ALASKA AND CALIFORNIA, USA

The sea otter (Enhydra lutris) historically ranged contiguously across northern Pacific coastlines from Mexico to Japan but was hunted to near extinction during the 19th century (Kenyon 1969). Following the ban of commercial harvest in the early 20th century, remnant populations expanded but remained fragmented, with the northern sea otter (NSO; E. l. kenyoni) in Alaska, Washington, and British Columbia; the southern sea otter (SSO; E. l. nereis) in California; and the Asian sea otter (E. l. lutris) in Russia and Japan (Reidman and Estes 1990). This trend was reversed, and populations of NSO in the Aleutians and western Alaska declined 95.5% from 1965 to 2003, with a 62% decline from 2000 to 2003 (Estes et al. 2005). Reasons for decline appear to be related to natural predation with an unknown role played by other factors such as disease and contaminants (US Fish and Wildlife Service 2010). Sea otter populations in California have failed to reach projected carrying capacity, and recent SSO declines appear to be associated with increased mortality rates rather than declining birth rates, with infectious disease implicated as a likely cause (Estes et al. 2003).

Because they lack adequate body fat, sea otters rely on their fur to maintain a waterproof thermal barrier that allows them to exist in their cold water environment (Yeates et al. 2007). Sea otters spend 5–35% of the day cleaning and grooming their fur (Davis et al. 1988; Reidman and Estes 1990). Dermatologic disease may significantly impair thermoregulation.

The Poxviridae are large, DNA viruses that replicate in the host cytoplasm. The subfamily Chordopoxvirinae infects diverse vertebrates. Variola virus, in the genus Orthopoxvirus, is the cause of smallpox in humans, which killed hundreds of millions before vaccination was implemented worldwide in the last century. There is evidence that variola has served as a major selective factor on humans (Galvani and Slatkin 2003). While there have been recent studies of the evolution of the genus Orthopoxvirus (Babkin and Babkina 2012), the diversity, ecology, and evolution of Poxviridae is poorly understood.

Data on poxviruses of the Carnivora is limited. Distinct viruses in the genus Orthopoxvirus infect raccoons (Procyon lotor) and skunks (Mephitis mephitis; Emerson et al. 2009). Raccoonpox can also infect domestic cats (Felis catus; Yager et al. 2006). Cowpox, an Orthopoxvirus, is also known to infect diverse Carnivora (Meyer et al. 1999) and man (Kurth et al. 2009). Orf virus, a Parapoxvirus of sheep, has been identified in a domestic cat (Hamblet 1993). Distinct members of the genus Parapoxvirus have been partially characterized from pinnipeds (Bracht et al. 2006), and transmission of “sealpox” lesions to human caregivers in rehabilitation facilities has been documented (Hicks and Worthy 1987). A poxvirus that does not cluster within recognized poxviral genera has been reported from Steller sea lions (Eumetopias jubatus; Bracht et al. 2006). We report a novel poxvirus of sea otters that appears to be distinct from previously characterized poxviruses to a degree consistent with a novel genus.

A female NSO pup, approximately 4 mo old, was found alone in the surf near Anchor Point in Cook Inlet, Alaska (59°78′N, 152°77′W) in early August 2009. Eight weeks after admission, a 15-mm-diameter pink area of superficially ulcerated skin developed on the caudal abdomen (Fig. 1a). Two weeks later, a 6-mm punch biopsy of the skin lesion was obtained and preserved in 10% neutral buffered formalin. The tissue was embedded in paraffin blocks, sectioned, and stained with H&E. Thin section as well as negative staining electron microscopy (EM) was performed on a portion of the formalin-fixed punch biopsy.

The skin lesion appeared to be resolving, but due to persistent poorly controlled seizures, the otter was declared nonreleasable and inappropriate for captive maintenance and was euthanized 15 wk after admission. Gross postmortem examination confirmed a healing skull fracture with an associated focal frontal lobe defect. Multiple tissues were submitted for histopathology and processed as above.

A second clinical case of poxvirus infection was identified in a SSO in California in September 2011. A female pup, weighing 1.8 kg and estimated to be 3 days old, was recovered from a beach in Cayucos, California (35°26′5.11″N, 120°53′19.81″W). The animal was transported to the Monterey Bay Aquarium for care. Twelve days after admission, a 2×3 mm, slightly raised, ulcerated, pink to red lesion with a broad base attachment was noted on the right upper lip just dorsal to the mucocutaneous junction. Within 6 days, a second lesion similar in appearance was observed asymmetrically located on the left upper lip (Fig. 1b). A wedge biopsy of the original lesion, which had increased in size to 6×7 mm, was obtained, submitted for histopathologic evaluation, and processed for light microscopy as above. A second set of biopsies was collected and submitted for poxviral PCR and sequencing 8 days later. Over the subsequent 2 mo, the SSO pup facial lesions healed uneventfully. Repeated examinations failed to identify any additional lesions. Approximately 88 days after initial admission, the pup was transferred to a permanent captive home.

In vitro isolation of virus culture was attempted using two primary cell lines that have previously allowed successful isolation and propagation of poxviruses from seals. Primary harp seal (Pagophilus groenlandicus) kidney cells (PSK) and primary beluga (Delphinapterus leucas) kidney cells (BWK) lines were established and propagated using standard methodology (Chan and Hsuing 1994). Previously frozen lesion material from the NSO was ground with sterile silica sand and added to media (Dulbecco's modified Eagle's medium/F-12 plus penicillin 200 IU/mL, streptomycin 200 µg/mL, and gentamicin 50 µg/mL; 10% w/v), centrifuged for 10 min. (2,060 × G), and inoculated onto BWK and PSK cells grown in tissue culture flasks. After adsorption for 1 hr at 37 C, inoculum was removed, and 10 mL of medium containing 2% fetal calf serum was added to each flask. Mock-infected flasks inoculated with sterile medium served as negative controls. Flasks were returned to 37 C and examined daily for cytopathologic effect (CPE). Medium was changed weekly on all flasks.

We extracted DNA from NSO tissue–inoculated PSK cells after passage seven and from a wedge biopsy of the SSO with a commercial kit (QIAamp® DNeasy kit, Qiagen Inc., Valencia, California, USA). PCR amplification of a partial sequence of the poxvirus DNA–dependent DNA polymerase gene was performed using consensus primers designed to amplify all Chordopoxvirinae (Table 1). Gaps were closed using sea otter poxvirus–specific primers designed from sequences obtained using consensus primers (Table 1).

The reaction mixture for the PCR amplifications consisted of 0.1 µL Platinum Taq DNA Polymerase (Invitrogen, Carlsbad, California, USA), 2 µL 10× PCR buffer, 0.8 µL 50 mM MgCl₂, 0.4 µL 10 mM dNTPs, 1 µL of 20 µM forward and reverse primers, and 3 µL DNA template. For the second round, 3 µL of the round 1 product was used as template. The PCR conditions used for all reactions included an initial denaturation of 5 min at 95 C, followed by 45 cycles of denaturation at 95 C for 1 min, annealing at 45 C for 1 min, and extension at 72 C for 1 min, followed by a final elongation step at 72 C for 10 min. After electrophoresis, bands of interest were cut from the gel and extracted using the Qiaquick gel extraction kit (Qiagen). Direct sequencing was performed using the Big-Dye Terminator Kit (Applied Biosystems, Foster City, California, USA) and read on ABI automated sequencers. All amplicons were sequenced at least twice in each direction, and primer sequences were edited out before constructing contiguous sequences.

Predicted homologous 730–760 amino acid polymerase sequences were obtained from GenBank, with the RefSeq database selected whenever possible, and were aligned using MAFFT (Katoh and Toh 2008). Bayesian analyses of the alignment were performed using MrBayes 3.1 (Ronquist and Huelsenbeck 2003) on the CIPRES server (Miller et al. 2010), with gamma distributed rate variation, a proportion of invariant sites, and amino acid substitution model jumping (Ronquist and Huelsenbeck 2003). Statistical convergence was assessed by the standard deviation of split frequencies as well as potential scale reduction factors of parameters and was stopped when statistical convergence was reached. Two independent Bayesian analyses were run to avoid entrapment on local optima. Maximum likelihood (ML) analyses of each alignment were performed using RAxML on the CIPRES server (Stamatakis et al. 2008), with gamma distributed rate variation and a proportion of invariant sites. The amino acid substitution model with the highest posterior probability in the Bayesian analysis was selected. Amsacta moorei entomopoxvirus (GenBank accession NP_064832), a member of the subfamily Entomopoxvirinae, was used as the out-group. Bootstrap analysis was used to test the strength of the tree topology, with 1,000 subsets (Felsenstein 1985).

Histologically, the skin biopsy of the NSO had marked epidermal hyperplasia with rete pegs and nests extending into the deep dermis. Epithelial cells had ballooning degeneration and necrosis, and intracytoplasmic eosinophilic inclusions typical of Bollinger bodies were detected in epithelium of the epidermis (Fig. 1c, d). Poxvirus particles were seen by negative stain EM and by transmission EM in scrapings from the epidermal surface and in intact epithelium of the epidermis, respectively (Fig. 1e, f).

Histologic examination of NSO tissues collected at postmortem examination confirmed that the pox lesion on the ventral abdomen was almost completely resolved, but the otter had diffuse pronounced hyperplastic, papillomatous gingivitis in the oropharynx, with small eosinophilic intracytoplasmic inclusion bodies suggestive of poxvirus and systemic lymphohistiocytic inflammation of unknown etiology in multiple tissues, including brain, skeletal muscle, nerves, lungs, and lymph nodes.

The lip biopsy of the SSO had a focus of well-differentiated epidermal cells proliferating as small and large epithelial nests extending into the dermis, similar to the lesion in the NSO. Central zones of dyskeratosis and degeneration with progression to individual cell necrosis were noted. Midzonal to inner areas of these lobules contained numerous prominent intracytoplasmic eosinophilic inclusion bodies. Cells toward the periphery of the lobules often exhibited prominent cytoplasmic vacuolation and cellular ballooning.

After 15 and 21 days postinfection for the BWK and PSK cells, respectively, swelling, rounding up, and detachment of cells was seen, with progressive CPE until the entire cell monolayer had disintegrated. The negative control flasks showed no CPE. The CPE could be replicated upon dilution and inoculation with this tissue culture fluid using freshly seeded flasks of both cell lines over at least four passages.

The PCRs resulted in a 2,233 base pair (bp) segment of the polymerase gene with a 30.5% GC content. Sequences from both animals were identical. Sequence data were submitted to GenBank (accessions KF425534 and KF425535). This virus is hereafter referred to as sea otter poxvirus (SOPV).

Bayesian analysis found that the JTT model (Jones et al. 1992) of amino acid substitution was most probable. Bootstrap values from the ML analysis are shown to the right of the Bayesian posterior probabilities on the phylogenetic tree (Fig. 2). The phylogenetic analyses found that SOPV clustered within the subfamily Chordopoxvirinae but did not cluster within known genera.

This is the first report of a poxvirus from a sea otter. The virus was isolated and grew well in two primary cell lines: one from a phocid (harp seal) and one from a cetacean (beluga) origin, demonstrating that SOPV can replicate in cells from different host species. This suggests that SOPV, like other poxviruses, may have relatively low host fidelity for a DNA virus.

Identical viruses were found from two animals from geographically and genetically distinct populations. As a species, sea otters have undergone a relatively recent population bottleneck, with 99% lost to the fur trade from 1741 to 1911, resulting in low gene flow (Larson et al. 2012). Examination of 320-bp mitochondrial d-loop sequences of sea otters found transitions at three sites between subspecies (Larson et al. 2002). One possible explanation is that this virus had a historical relationship with both populations antecedent to the bottleneck. Alternatively, transmission of the virus may have occurred between the populations. Northern and southern sea otters do not appear to cross between their ranges, suggesting that transmission could involve virus survival in seawater, use of a different host that spans their ranges, or mechanical movement of virus. Future investigations should prioritize species that are sympatric with both NSO and SSO populations. Although no differences were detected in the 2,233 bp of the DNA-dependent DNA polymerase sequence, this is a highly conserved gene. Examination of additional less conserved genes and additional isolates may help resolve the question of whether these viruses are truly identical.

The two cases presented here occurred in late summer/early fall; further cases are needed to determine whether there are seasonal differences or if this is coincidental. Both cases were identified in immature stranded animals, and lesions appeared during potentially stressful rehabilitation. Outbreaks of pox lesions have been previously reported in harbor seals (Phoca vitulina; Wilson et al. 1972; Muller et al. 2003) and California sea lions (Zalophus californianus; Nollens et al. 2006) undergoing rehabilitation, and it has been proposed that immune suppression during stress may cause animals to be more susceptible to poxviral disease.

Poxviral genera differ significantly in their base compositions (Li et al. 2010). Parapoxvirus, Molluscipoxvirus, and Crocodylidpoxvirus have GC-rich genomes; all others are AT-rich. These viruses with GC-rich genomes are not monophyletic; our analyses, and those of others, find that Avipoxvirus is basal to Parapoxvirus and Molluscipoxvirus, and Parapoxvirus is more closely related to the remaining AT-rich genera than it is to Molluscipoxvirus or Crocodylidpoxvirus. SOPV is supported as being outside the clade of non-Avipoxvirus AT-rich poxviruses. The resolution of our analyses beyond this is weak; support values outside of genera between the Chordopoxvirinae and the clade of non-Avipoxvirus AT-rich poxviruses were all <70% Bayesian posterior probabilities and <50% ML bootstrap values. Taking into consideration the AT bias in some clades of the Poxviridae, we chose to analyze amino acid alignments rather than nucleotides because of concerns regarding nonlineage factors on viral nucleotide composition bias outweighing the true phylogenetic signal. In other virus families, AT richness has been associated with host jumps (Wellehan et al. 2004; Poss et al. 2006). The DNA viruses that replicate in the cytoplasm, such as poxviruses, are associated with increased ability to jump hosts (Pulliam and Dushoff 2009). Considering that the Poxviridae contain the etiology of one of the most significant diseases in human history (smallpox), the evolution of the family as a whole is surprisingly poorly understood. The zoonotic potential of SOPV is unclear. There is precedent for zoonotic carnivore poxvirus infections, as previously discussed (Kurth et al. 2009). Although no pox infections have been reported in humans exposed to sea otters, it would be prudent for those working with otters to wear protective clothes and gloves when handling these animals either in the wild or in rehabilitation settings. Similar precautions have been advised for workers coming into contact with seal species that may be infected with seal parapoxviruses (Roess et al. 2011).

In conclusion, SOPV is the first poxvirus infection reported in a mustelid. The virus is phylogenetically distinct from other poxviruses at a level consistent with a novel genus. Identical viral sequences were obtained from animals of two geographically distinct sea otter subspecies. Cutaneous disease may be more significant in a species highly dependent on pelage for thermoregulation. Further studies are needed to determine the source of this virus, its mode of transmission, zoonotic potential, and biological significance.

The northern sea otter pup in this report was collected under US Fish and Wildlife Letter of Authorization LOA837414, and the southern sea otter pup was collected under Federal Fish and Wildlife Permit MA032027-1 and Letter of Authorization LOA032027-1. Laboratory procedures for the genetic sequencing were approved by the University of Florida IACUC 201106085. Consensus poxviral primer design was funded by Office of Naval Research grant N00014-09-1-0252 to J.F.X.W. We thank Phoenix Central Laboratory, Mukilteo, Washington, for slide preparation and data retrieval, Christie Buie of Northwest ZooPath for photo editing and image layout, Verena Gill, US Fish and Wildlife Service for assistance in sample shipment, and the sea otter rehabilitation staff at the Alaska SeaLife Center and the Monterey Bay Aquarium for assistance in sample collection and husbandry of the sea otter pups. We thank Michael Hammill of the Department of Fisheries and Oceans Canada for collecting the harp seal from which the PSK cells were derived and we thank the Hunters and Trappers Association at Tuktoyaktuk, Northwest Territories, for collecting the beluga from which the BWK cells were derived.