IN SITU HYBRIDIZATION AND VIRUS CHARACTERIZATION OF SKUNK ADENOVIRUS IN NORTH AMERICAN WILDLIFE REVEALS MULTISYSTEMIC INFECTIONS IN A BROAD RANGE OF HOSTS

Health surveillance of wildlife enables rapid recognition of new and emerging zoonotic diseases, which is a key tenet of the One Health movement (Mackenzie and Jeggo 2019). North American porcupines (Erethizon dorsatum; hereafter referred to as porcupines), raccoons (Procyon lotor), and striped skunks (Mephitis mephitis; hereafter referred to as skunks) are common across North America and are frequently admitted to wildlife rehabilitation centers due to frequent encounters with humans and domestic animals. Wildlife admitted to rehabilitation centers may be carrying infectious agents or develop disease while in temporary captivity. Consequently, situations in which stressed or immune-compromised individuals from multiple species are brought together in close proximity represent a unique scenario for novel diseases to emerge (Lempp et al. 2017).

Adenoviruses include a large group of double-stranded, nonenveloped, hexagonal and icosahedral DNA viruses documented infecting all major groups of vertebrate species and coevolved with hosts over time (Maclachlan and Dubovi 2017). Mastadenoviruses (genera infecting mammals) are thought to have arisen by a host-switching event in atadenoviruses, from a lineage that coevolved within squamatid reptiles but presumably jumped to an ancient ruminant during evolution (Wellehan et al. 2004). Adenoviruses may cause significant disease in domestic and wildlife species involving one or more of the respiratory, cardiovascular, hepatic, or gastrointestinal systems. Adenovirus infection in free-ranging wildlife is common; significant adenovirus-related disease is less so but includes severe adenoviral hemorrhagic disease of mule deer (Odocoileus hemionus) caused by Odocoileus adenovirus 1, fox encephalitis of red foxes (Vulpes vulpes) caused by canine adenovirus 1 (CAdV-1), infectious canine hepatitis of black (Ursus americanus) and brown (Ursus arctos) bears also caused by CAdV-1, and and enteritis of red squirrels (Sciurus vulgaris) caused by squirrel mastadenovirus A (Duff et al. 2007; Walker et al. 2016; Knowles et al. 2018; Woods et al. 2018). With the exception of CAdV-1, which infects a variety of species including members of family Canidae (dogs, foxes, coyotes, wolves), Ursidae (bears), and Mephitidae (skunks), adenoviruses are relatively species specific and usually do not cause significant infection or disease outside the evolved hosts (Borkenhagen et al. 2019). Phylogenetic analysis suggests that skunk adenovirus-1 (SkAdV-1) is most closely related to CAdV-1 and CAdV-2 and bat adenovirus 2 and 3 and likely shares a common ancestor (Kozak et al. 2015).

The SkAdV-1 was originally reported in 2015 from a striped skunk found dead in Ontario, Canada, and submitted for routine rabies surveillance, with negative results (Kozak et al. 2015). Necropsy revealed lymphoid depletion, interstitial pneumonia, and diffuse necrosis throughout the liver with large intranuclear inclusion bodies (INIBs). Virus isolation and viral genome sequencing revealed a new species of adenovirus. Since its original detection in skunks, SkAdV-1 has been detected in several other mammalian species and families, including Erinaceidae (African pygmy hedgehog; Atelerix albiventris), Erethizontidae (North American porcupine; Erethizon dorsatum), and most recently Canidae (gray fox, Urocyon cinereoargenteus; Madarame et al. 2016; Balik et al. 2020; Needle et al. 2020). For cases that received a necropsy, the most commonly described lesions included necrotizing bronchointerstitial pneumonia and tracheitis with INIBs (Kozak et al. 2015; Needle et al. 2019, 2020). To date in porcupines, SkAdV-1 has only been detected from intranasal swabs in individuals that subsequently recovered from respiratory disease and were released into the wild (Balik et al. 2020). We report on disease caused by SkAdV-1 in multiple wildlife species, including porcupines, skunks, and raccoons, the latter of which have not yet been reported as being infected with SkAdV-1. We describe the pathology and characterize tissue distribution via in situ hybridization (ISH) and compare complete SkAdV-1 genomes isolated in cell culture from three wildlife cases (porcupine, skunk, and raccoon).

A total of 24 diagnostic wildlife cases involving porcupines (n=16), skunks (n=3), and raccoons (n=5) were submitted to multiple rehabilitation centers and laboratories from a wide geographic range (Fig. 1). Additionally, paraffin-embedded tissues from a gray fox (Needle et al. 2020) were included for ISH, and nasal swabs from a rehabilitated porcupine (case 6, which developed respiratory signs but recovered and was released) were submitted for SkAdV-1 PCR.

All diagnostic cases received a complete necropsy performed by a veterinary pathologist. Representative tissue samples were submitted for histologic examination and stained with H&E by using standard procedures.

We performed RNAScope® 2.5 HD-Red ISH for SkAdV-1 on 4-µm tissue sections mounted on charged slides by using probes designed in collaboration with Advanced Cell Diagnostics Inc. (Newark, California, USA). Ten proprietary RNAScope ZZ probes were designed to cover the 5,629–6,229 base pair (bp) region of the DNA polymerase gene (KP238322), and a red chromogen was used. The probes were designed to exclude hybridization with the CAdV-1 reference genome. We performed ISH according to manufacturer's instructions (Advanced Cell Diagnostics) with a 15-min antigen retrieval step. Tissue from a North American porcupine (case 23) with SkAdV-1 infection confirmed by PCR, virus isolation, and sequencing of the DNA polymerase gene, as previously described, was used as a positive control (Balik et al. 2020). Lung tissue from a North American porcupine, a striped skunk, a raccoon, and a gray fox that tested negative for SkAdV-1 by PCR and with normal lung histology were used as negative controls. The designed probe was confirmed to lack cross reactivity with CAdV-1 by performing ISH on a case of CAdV-1–associated encephalitis in a red fox (confirmed via PCR).

In 17 of the cases, DNA was extracted from nasal swabs or frozen tissues. In brief, 25 mg of lung tissue was incubated overnight in lysis buffer (100 mM NaCl, 500 mM Tris HCl, 10% sodium dodecyl sulfate) with proteinase K. Swabs were swirled in 250 µL of lysis buffer for 30 s to remove the material from the swabs, and then another 250 µL of lysis buffer was added with proteinase K. The samples were incubated overnight in a 56 C water bath, followed by extraction of the nucleic acids by using the phenol-chloroform method, precipitated with absolute ethanol, and sodium acetate. Pellets were resuspended in 50 µL Trisethylenediaminetetraacetic acid buffer. Samples were quantified on the nanodrop to ensure there was adequate DNA concentration (50–1,000 ng). Degenerate primers (forward 5′ TYM GVG GVG GBM GVT GYT AYC C 3′; reverse 5′ GTR GCR AAN SWS CCR TAS AGG GCR TT3′) were used to amplify the partial sequence of the adenoviruse DNA polymerase gene. The PCR reaction was performed in a 50-µL volume, which contained 100 ng of DNA, 3 mM MgCl₂, 0.4 µM concentration of each primer, 0.25 mM dNTPs, and 1.25 U Acc-start Taq. The following conditions were used: initial denaturation 94 C for 3 min, followed by 40 cycles at 94 C for 30 s, 50 C for 30 s, and 72 C for 30 s. There was a final extension step at 72 C for 10 min. The original amplified DNA from the tested sample, along with a positive control was purified by using a commercial Qiagen kit, and submitted for sequencing at Macrogen (Seoul, South Korea). Sequence analysis of PCR fragments was performed by using the online tool Staden Package programs Pregap and Gap (SourceForge 2021).

We processed lung tissue and a nasal swab from porcupine cases 4 and 6, respectively, and lung tissue from raccoon cases 8 and 10 for virus isolation. Frozen lung samples and nasal swabs in viral transport media (VTM) were received from submitting agencies and stored at –80 C. Lung samples were rapidly thawed and then ground with sterile silica sand in a sterile mortar and pestle with added media (Hanks balanced salt solution plus penicillin 200 IU/mL, streptomycin 200 µg/mL, and gentamicin 50 µg/mL). They were further diluted to give approximately a 10% weight-to-volume ratio of tissue suspension and clarified by low-speed centrifugation for 10 min (2,060 × G). For swabs, VTM from the swab sample ampoules was inoculated directly onto cells. In each case, 0.5 mL of the supernatant fluid or the VTM was used to inoculate 25-cm² tissue culture flasks containing 80% confluent cell cultures of African green monkey (Chlorocebus sp.) kidney cells (Vero C1008) from the American Type Culture Collection (ATCC-CRL-1586). Flasks were incubated at 37 C for 1 h, and the inoculum was removed, and 10 mL of Dulbecco modified Eagle medium/Ham F-12, 1:1, with added HEPES, containing 2% fetal calf serum, combined in each flask. Mock-infected flasks inoculated with sterile medium served as negative controls. Flasks were incubated at 37 C and examined daily for cytopathic effects (CPE). Cells were passaged weekly (1:5) for 4 wk, at which time flasks not showing visible CPE were discarded. Media from flasks showing CPE were passed through a 0.45-µm filter, diluted (1/100), and passaged onto fresh cells. Only after CPE could be demonstrated in these flasks were the presumptive virus isolates included for further study. Flasks showing extensive CPE were frozen and held at –80 C and submitted for next-generation sequencing.

We extracted DNA from cultured viruses by using Quick-DNA/RNA™ Viral Kit (Zymo Research, Irvine, California, USA), as described in the company's protocol. We used 50 µL of nuclease-free water (Corning, Tewksbury, Massachusetts, USA) for elution. We quantified DNA by using Qubit™ dsDNA HS Assay Kit and a Qubit™ (ThermoFisher Scientific, Chelmsford, Massachusetts, USA). Libraries were done by using Nextera XT DNA Library Preparation Kit (Illumina, San Diego, California, USA). Briefly, 0.3 ng/mL of double-stranded DNA was used to start the libraries. Fragmentation and tagmentation was performed as suggested by the company's protocol apart from using 14 PCR cycles instead of 12. Libraries were then purified by using AxyPrep Mag™ PCR Clean-Up Kits (Axygen, Corning, New York, USA), as described in the Nextera XT protocol. The libraries quality was assessed by using Agilent High Sensitivity DNA Kit in a bioanalyzer (Agilent, Santa Clara, California, USA). Libraries were normalized by using LNB1 beads (Nextera XT protocol). Libraries were sequenced in a v3 600-cycle cartridge by using a MiSeq instrument, and PhiX was used at around 4% as a control for the sequencing runs (Illumina).

Reads were trimmed for adaptors and quality by the MiSeq software during FastQ generation. Using CLC Genomic Workbench software (version 12.0.3, Qiagen, Redwood City, California, USA), reads from each sample were mapped by using Map Reads to Reference application with default settings against a list of adenovirus full-length genomes obtained from GenBank. Then, all reads were mapped against the closest adenovirus full-length genome in the list (KP238322). Consensus sequence was extracted from alignments. Reads were trimmed again for quality and adaptors in CLC Genomic Workbench software, before, de novo analysis by using the application “De novo Assemble Metagenome,” with 5000 minimum contig length, and the scaffolding setting option was performed. Contigs were used to confirm the adenovirus full-length genome obtained by resequencing. Sequence analysis was done by using LaserGene software (DNAStar Inc., Madison, WI, USA). Whole-genome sequences were assembled by using LaserGene's SeqMan NGen module. Sequence comparisons were done with the SeqMan Pro module.

A total of 24 wildlife cases with a history of respiratory disease were found to have lesions consistent with SkAdV-1 infection (Table 1). Of these, 19 cases were confirmed to be positive for SkAdV-1 infection, and five cases were considered either probable or suspect based on clinical histories, characteristic lesions, INIBs, or immune positivity for adenovirus antigen within affected tissues. Cases were submitted for necropsy between June 2015 and October 2020, mostly between August 2018 and October 2020. The cases can be grouped into three categories: wildlife that developed SkAdV-1 disease during rehabilitation (n=21), wildlife that presented to rehabilitation centers with skunk adenoviral disease (n=1; case 22), and free-ranging wildlife that were found dead (n=2; cases 23 and 24). Submissions occurred throughout the year but primarily between the months of August and December and included 16 porcupines, five raccoons, and three skunks. Cases occurred in Nova Scotia and Ontario in Canada and in Maine, New Hampshire, and New York states in the US (Fig. 1). In cases of mortality, wildlife either died or were euthanized due to the severity of respiratory disease. The length of time between onset of disease and mortality (when known, n=18) ranged between 2 d and 82 d (mean 14 d and median 6 d). Recorded clinical signs invariably included cough, oculonasal discharge, and dyspnea, which were variably treated with antimicrobial medications, nebulization, and oxygen therapy.

The most common gross and microscopic lesions were consistent between species (Table 1). Grossly, the most common lesion was dark red, firm to rubbery lungs (n=20), frequently accompanied by prominent hemorrhagic pleural effusion (n=10) and pulmonary edema (n=12; Fig. 2C). Many cases had oculonasal discharge (n=11; Fig. 2A, B), and porcupines often (n=10) exhibited overt inflammatory exudate within the airways, including nasal passages, trachea, and bronchi (Fig. 2D). Less common lesions included poor body condition (n=7), fibrinous pleuritis (n=2), and peritonitis (n=1), mild to moderate enlargement of parenchymal organs (liver, spleen, kidneys, and lymph nodes; n=4), and skin crusting consistent with sarcoptic mange (n=4; porcupines only).

Microscopic lesions were similar between all species. The most common lesion was moderate to marked bronchopneumonia (n=22; Fig. 3A) that commonly (n=12) was accompanied by interstitial lesions, including prominent hyaline membranes (Fig. 3C) or type 2 pneumocyte hyperplasia or both. Disease progression in most cases was considered to be acute to subacute, with occasional cases (cases 13 and 17) demonstrating extensive type 2 pneumocyte hyperplasia, intra-alveolar fibrosis, and marked lymphoplasmacytic inflammation. Most cases (n=18) exhibited extensive bronchiolar epithelial necrosis with mixed inflammatory cell infiltrates. Many cases (if examined) also had evidence of mild to marked rhinotracheitis (n=16) that exhibited similar chronicity, inflammation, and epithelial changes, as described in the lungs (Fig. 3B). Almost all cases (n=23) exhibited large INIBs (Fig. 3D). Inclusions primarily occurred within sloughed necrotic epithelial cells throughout the respiratory tract (n=18). Microscopic lesions occurred less commonly outside the respiratory tract and included necrotizing tubulonephritis with INIBs (n=6), necrotizing hepatitis (n=1; case 5), and encephalitis (n=1; case 4). In three cases, INIBs were only observed in extrarespiratory tissues.

We found that PCR products (Table 1) from porcupines (n=11), skunks (n=3), and raccoons (n=3) had 100% sequence identity with SkAdV-1 PB1 (GenBank accession no. KP238322.1) over 525 bp and 100% sequence identity to Callithrix pygmaea adenovirus (GenBank no. HM245776.1) over 460 bp. The genomes sequenced from virus isolates from one raccoon (case 10; GenBank no. MZ073341) and one porcupine (case 4; GenBank no. MZ073342) were compared with the original SkAdV-1 genome (Kozak et al. 2015; GenBank no. KP238322; Table 2). Genome sizes of the virus isolates were 31,848 bp, with an average coverage for the porcupine samples of 1,600× and 744× for the raccoon sample. We identified 19 point mutations from each virus genome, with the majority being either within intergenic regions or silent. Four variable sites resulting in amino acid substitutions were located at nucleotide (nt) positions 23102, 27956, 27992, and 30288, and corresponded to a 100-kd protein (nt 23102), E4 ORFC (nt 30288), and to the fiber protein (nt 27956 and nt 27992). The same GTC to GCC mutation (substitution of Val to Ala) in the 100-kd protein occurred in both porcupine and raccoon. The TCC to TTC mutation (substitution of Ser for Phe) in the E4 ORFC protein occurred only in the porcupine. In the fiber protein, GCC to ACC mutation at 27956 (substitution of Ala to Thr) occurred in porcupines, while ACT to TCT mutation at 27992 (substitution of Thr to Ser) was found in the raccoon.

We observed visible CPE in both porcupine and raccoon samples after 1 d postinfection, appearing as focal areas of rounded swollen cells that rapidly progressed until the entire cell sheet detached. The mock-infected flasks remained normal throughout.

We consistently detected SkAdV-1 nucleic acid in nuclei of either intact or sloughed respiratory epithelium, either in the nasal turbinates, trachea, or lungs, and apparent interspecies variability was not found (Fig. 4, Supplementary Material Fig. S1, and Table 3). In two of the porcupines, hybridization was not detected in the lungs and was localized to the necrotic debris of nasal turbinates (case 14) and in the trachea and kidney (case 15). In one porcupine (case 20), mesothelial cells also had multifocal positive nuclear hybridization. In the trachea (three of three evaluated animals), the mucosal epithelium and submucosal glandular epithelium had multifocal nuclear hybridization (Fig. 4B). We found SkAdV-1 hybridization less commonly in other organs and tissues, including in the liver (three of six evaluated animals; infrequent hepatocytes, biliary epithelium, and sinusoidal endothelium), kidney (two of five evaluated animals; rare renal tubular epithelium), brain (one of one evaluated animals; rare cells), and lymph node (one of two evaluated animals; rare macrophages; Fig. 4C, D).

Based on lesion severity, SkAdV-1 was considered to be either the primary cause of death or significantly contributing to morbidity and mortality in the majority of cases (n=23). One suspect porcupine case (case 7) recovered from respiratory disease during rehabilitation and later succumbed to Baylisascaris procyonis cerebral larval migrans. Subsequent necropsy demonstrated only mild pneumonia with scattered INIBs consistent with adenovirus infection. Aerobic bacterial culture was conducted on lung tissue (n=13), and in four cases (two porcupines, one skunk, one raccoon), isolated bacteria were considered significant contributors to disease. These included B. bronchiseptica, Klebsiella pneumoniae, P. multocida, and S. zooepidemicus. We performed PCR for canine distemper virus on tissues from five raccoons and two skunks; all were negative (Supplementary Material Table S1). A variety of other ancillary diagnostics were performed on select cases, and all results were considered negative (Table S1).

With advanced medical care and monitoring, disease surveillance in wildlife rehabilitation centers provides a unique situation for the recognition of new and emerging diseases affecting wildlife species. In almost all cases we examined with disease caused by SkAdV-1, lesions were similar, whatever the species. The most significant and common lesion that we found (92% of cases) was severe necrotizing bronchopneumonia or bronchointerstitial pneumonia with INIBs within sloughed airway epithelial cells, which is consistent with reports from a striped skunk and from African pygmy hedgehogs, (Kozak et al. 2015; Needle et al. 2019).

Several lesions reported here indicate that SkAdV-1 can cause multisystemic infections, including necrotizing tubulonephritis, hepatitis, and encephalitis. The presence of tubulonephritis with inclusion bodies in renal epithelium in six cases (25%) suggests that similar to CAdV-1, SkAdV-1 may be shed via the urine (Evermann et al. 2006). Further research is needed to explore this mechanism of virus transmission, which could be a useful screening tool for wildlife rehabilitation centers. The ISH results confirmed epithelial cell tropism, with occasional infection of histiocytes, mesothelial cells, and endothelium.

Disease caused by SkAdV-1 has not previously been described in raccoons, nor has the pathology of SkAdV-1 infection in porcupines been detailed. Review of porcupine necropsy reports archived in the Canadian Wildlife Health Cooperative's database (>30 consecutive years of wildlife health surveillance data) did not reveal historical evidence of infectious pneumonia in porcupines, with the most common cause of death being linked to cerebral larva migrans of B. procyonis. Documented SkAdV-1 disease in porcupines, raccoons, and skunks has implications for both free-ranging and captive wildlife. Four cases of free-ranging wildlife have been previously reported with SkAdV-1 disease, including the first skunk case report, a gray fox, and two porcupines that recovered and were subsequently released (Kozak et al. 2015; Balik et al. 2020; Needle et al. 2020). Of the cases we describe, three animals (cases 22–24), including two skunks and one porcupine, developed disease while free ranging. Determining the reservoir species, transmission routes, and duration of shedding for this virus will be crucial for management of SkAdV-1 outbreaks in wildlife rehabilitation centers to help identify possible carriers and mitigate disease spread. Most of the free-ranging cases with SkAdV-1 have been in skunks (n=3) that may support this species as the primary host or carrier for SkAdV-1. However, this sample size is very small, and there have been more total instances of this virus reported in other species (porcupines, hedgehogs, raccoons, and gray foxes) than in skunks; therefore, targeted SkAdV-1 surveillance of free-ranging wildlife populations will be required before a carrier host can be confirmed.

The impact SkAdV-1 may have on free-ranging wildlife populations is currently unknown. Cases of SkAdV disease are known in free-ranging wildlife from both ours and previous reports (Balik et al. 2020; Needle et al. 2020). Additionally, it has been confirmed both in this article (case 6) and in another recent publication that porcupines can recover from SkAdV-1 infection to be subsequently released (Balik et al. 2020). In addition to the diagnostic cases included in this article, clinical histories were available for 11 rehabilitated porcupines that developed similar respiratory disease concurrently with confirmed cases of SkAdV-1 but recovered and were released; these may have been undocumented cases of SkAdV-1 infection. It is possible that recovered individuals may shed SkAdV-1 into the environment for a period postrecovery (Evermann et al. 2006); this has been documented for other adenoviruses with which infected individuals may harbor and shed the virus for months postinfection (Greene 2012; Huh et al. 2019).

The predilection for SkAdV-1 to infect multiple host species is unique among adenoviruses, which typically demonstrate high host fidelity. The adenovirus fiber protein is located on the viral capsid and is important for host cell infectivity and immune protection, inducing the synthesis of specific neutralizing antibodies in an infected host (Cook and Radke 2017; Greber and Flatt 2019). Taken together, the single nt polymorphisms we found result in three unique amino acid sequences for the fiber protein of SkAdV-1 from each species, which may suggest a possible role for genetic variation affecting host cell entry in SkAdV-1 promiscuity. Given that this finding is based on only three virus samples, involvement of confounding factors, including random sequencing error or genetic drift associated with cell culture, cannot be ruled out entirely. Further research with larger sample sizes across multiple populations will be required to determine if single nt polymorphisms are conserved within host species and what role they may play in host cell infectivity.

Since 2015, SkAdV-1–associated disease has now been confirmed in multiple free-ranging and rehabilitated wildlife species (porcupines, raccoons, gray fox, and skunks) in Canada and the eastern US. Further research is required to better understand SkAdV-1 disease pathogenesis and transmission and to identify carrier hosts for this virus in free-ranging populations.

We acknowledge the following people and organizations for their hard work and collaboration without which this study would not have been possible: Hope for Wildlife Society; Darlene Jones (Canadian Wildlife Health Cooperative [CWHC] wildlife technician); and Jordi Segers (CWHC, National White-Nose Syndrome, scientific program coordinator).

Supplementary material for this article is online at http://dx.doi.org/10.7589/JWD-D-21-00099.