Circulation of pantropic canine coronavirus in autochthonous and imported dogs, Italy

Abstract Canine coronavirus (CCoV) strains with the ability to spread to internal organs, also known as pantropic CCoVs (pCCoVs), have been detected in domestic dogs and wild carnivores. Our study focused on the detection and molecular characterization of pCCoV strains circulating in Italy during the period 2014–2017 in autochthonous dogs, in dogs imported from eastern Europe or illegally imported from an unknown country. Samples from the gut and internal organs of 352 dogs were screened for CCoV; putative pCCoV strains, belonging to subtype CCoV‐IIa, were identified in the internal organs of 35 of the examined dogs. Fifteen pCCoV strains were subjected to sequence and phylogenetic analyses, showing that three strains (98960‐1/2016, 98960‐3/2016, 98960‐4/2016) did not cluster either with Italian or European CCoVs, being more closely related to alphacoronaviruses circulating in Asia with which they displayed a 94%–96% nucleotide identity in partial spike protein gene sequences. The pCCoV‐positive samples were also tested for other canine viruses, showing co‐infections mainly with canine parvovirus.

able to spread to extraintestinal tissues. This variant was associated with a fatal disease of dogs, characterized by leukopenia, gastroenteritis and severe lesions in the major organs Chen et al., 2019;Pinto et al., 2014), and subsequent studies have proved its impact on canine immune response (Marinaro et al., 2010).
Taking into account the scarce information existing about the actual circulation of pCCoV in the dog population, the aims of our study were the following: (a) to conduct an epidemiological survey for this virus in autochthonous and imported dogs in Italy during years 2014-2017; (b) to investigate the genetic relatedness of the detected pCCoV strains to extant coronaviruses; and (c) to evaluate the presence of co-infections in pCCoV positive and negative samples.

| Samples collection
During the period 2014-2017, 2,112 necropsy samples collected from different tissues (brain, heart, intestine, liver, spleen, lungs, kidney) of 352 dogs were submitted to molecular analysis to investigate possible viral causes of disease. The sampled animals included 141 clientowned, 151 stray and 60 imported dogs. Two hundred ninety-two of these dogs were from Italy, additional 56 animals had been imported by Hungary, while 4 dogs had been illegally imported from an unknown country. None of these dogs had undergone euthanasia, since their death was caused by illness or accident, but clinical signs occurring intravitam were not reported. At post-mortem examination, the analysed dogs showed catarrhal or hemorrhagic enteritis (n = 137), enlargement of the mesenteric lymph nodes (n = 79), pneumonia or other pulmonary lesions (n = 70), and meningeal and/or encephalic hyperaemia (n = 49). For 55 animals, post-mortem findings were not reported or were not fitting with those observed in pCCoV-infected dogs.

| Nucleic acid extraction
Samples collected for molecular investigations were homogenized with phosphate-buffered saline (PBS), and subsequently, RNA/ DNA extraction was performed using the automated extractor QIAsymphony (Qiagen) and the QIAsymphony DSP Virus/Pathogen Kit (Qiagen), following the manufacturer's instructions.

| CCoV genotyping and subtyping
The detected CCoV strains were characterized by means of two distinct real-time RT-PCR assays, specific for the genotypes CCoV-I and CCoV-II, targeting a fragment of the S gene (Decaro et al., 2013).
Samples that tested positive for CCoV-II were subjected to subtype-specific CCoV-IIa and CCoV-IIb gel-based RT-PCR assays targeting the S gene (Table 1) (Decaro et al., 2013). The PCR products were detected using the TapeStation 2,200 (Agilent Technologies) according to the manufacturer's protocol.

| Molecular characterization of pCCoV and CPV
Lung samples from the pCCoV-infected dogs were used for the molecular characterization of pCCoV. The spike protein gene (ORF2) of the putative pCCoV strains was sequenced and analysed using the protocol reported by Alfano et al. (2019). The sequences were analysed using BioEdit software package and the NCBI and EMBL analysis tools.
Samples that tested positive for CPV were further characterized by type-specific minor groove binder probe assays Decaro et al., 2007; and sequence analysis of partial VP2 gene (Buonavoglia et al., 2001).

| Sequence analysis and phylogeny
CCoV sequences were manually edited and analysed using the For the construction of phylogenetic trees, a multiple alignment of all target sequences was performed, using MAFTT Multiple Sequence Alignment software version 7 (Katoh & Standley, 2013) and Geneious software, using the neighbour-joining method, with the p-distance model, 1,000 bootstrap replicates, and, otherwise, the default parameters in Geneious (version 10.1.3).

| Nucleotide sequence accession number
The nucleotide sequences of the analysed pCCoV strains were deposited in GenBank under the following accession numbers:

| Data analysis
The comparison between positive and negative pCCoV dogs was carried out by examining the data with a chi-squared test, considering the statistically significant values p < .05, using the IBM SPSS Statistics 25 software. The confidence interval (CI 95%) was calculated using the prevalence parameter estimate with the Excel software.  (Table 2).

| Molecular characterization of putative pCCoV strains
Fifteen pCCoV strains were sequenced: 6 were from dogs of Italy and 9 from animals imported from other countries (6 from Hungary and 3 from an unknown country). The sequenced pCCoV strains presented neither the deletion in the genes of the accessory 3abc proteins nor the D125N mutation that had been suggested as potential markers for the pantropic behaviour (Decaro et al., , 2013. The phylogenetic tree, generated from partial ORF2 gene sequences, based on neighbour-joining method ( Figure 1  Abbreviations: J, juvenile (6-to 12-month-old); UKN, unknown; Y, young (0-to 6-month-old).

| D ISCUSS I ON
To date, pCCoV has been detected only sporadically in Italy and other countries Chen et al., 2019;Decaro et al., 2012Decaro et al., , 2013Ntafis et al., 2012;Pinto et al., 2014;Zicola et al., F I G U R E 1 Phylogenetic tree generated with the neighbour-joining method from partial spike protein gene (ORF2) sequences of the putative pantropic canine coronavirus strains and reference carnivore alphacoronaviruses  In fact, previous studies have demonstrated that this virus is frequently associated to subclinical infections and impairment of the lymphocyte counts, rather than to severe clinical signs and death of the infected dogs (Marinaro et al., 2010). Most of the pCCoVpositive animals also displayed post-mortem findings accounting for a systemic involvement, but at which extent those lesions were induced by pCCoV or by other co-pathogens, including the highly pathogenic CPV, CAdV-1 and CDV, infecting the same dogs could not be assessed.

TA B L E 3 Prevalence of viral co-pathogens in dogs with and without pCCoV infection
Remarkably, most of the pCCoV-infected dogs had been recently imported from Hungary, which may account for a wider circulation of this virus in eastern Europe. This finding is in contrast with those of a previous study aiming to assess the pCCoV circulation in Europe, which reported similar prevalences in Italy and Hungary, with detection rates of 8.69% and 9.33%, respectively (Decaro et al., 2013).
Currently, no test is available to differentiate the pantropic from the enteric CCoV strains, since no specific genetic markers have been identified in pCCoV so far. As a consequence, only the detection of a CCoV-IIa strain in extraintestinal tissues accounts for a possible pCCoV infection in dogs. This situation is similar to that observed in calicivirus infections in cats, where markers of pathogenicity have been not yet detected in highly virulent strains, so that diagnosis of systemic calicivirosis is obtained when the virus is detected in extrarespiratory tissues (Caringella et al., 2019).
Different from coronavirus infections in dogs, in cats potential genetic signatures were recently detected, which are able to discriminate between feline infectious peritonitis and feline enteric coronavirus strains (Felten et al., 2017).
According to previous observations (Decaro et al., 2013) (Wang, Ma, Lu, & Wen, 2006). These strains were from dogs illegally imported from an unknown country, which highlights the role of illegal trade of dogs in the introduction of pathogens into Italy (Decaro et al., 2007;Mira et al., 2018).
An additional finding of the present study is the high frequency of co-infections with pCCoV and other viruses. Enteric CCoV infections in dogs are very frequent Pratelli et al., 2003;Priestnall, Pratelli, Brownlie, & Erles, 2007). Co-infection by CPV and CCoV in dogs is known to enhance the severity of clinical signs Evermann, Abbott, & Han, 2005;Pratelli, 2006), with fatal outcomes being frequently reported in pups . A high frequency of co-infections with pCCoV and other pathogens has been previously reported (Alfano et al., 2019;Decaro et al., 2013;Ntafis et al., 2012;Pinto et al., 2014;Zicola et al., 2012). In the present study, we found a significant association of pCCoV with CPV and CAdV-2 infections. Pantropic CCoV is able to affect lymphocyte counts, thus causing a prolonged lymphopenia, so that it has been postulated that pCCoV infection may predispose to the increase in virulence of other pathogens by inducing immunosuppression in the infected dogs (Marinaro et al., 2010).
The large population of unvaccinated free-ranging dogs present in Italy (Corrain et al., 2007;Verardi, Lucchini, & Randi, 2006) considerably increases the density of susceptible hosts, and may thus importantly impact the spread and maintenance of canine pathogens in the environment. Therefore, it is strongly recommended to vaccinate not only private-owned dogs but also stray dogs whenever possible. In addition, the epidemiological risk related to the legal and illegal trade of carnivores from Asian countries must be taken into account, since this trade may represent a source of several emerging pathogens in domestic and wild canids (Mira et al., 2018(Mira et al., , 2019.

| CON CLUS ION
The present study demonstrates an increasing circulation of pCCoV in Italy, which reinforces the need for intensive and continuous surveillance on the importation and illegal trade in animals and the need for increased controls on both autochthonous and imported dogs.

ACK N OWLED G EM ENTS
We thank Dr Gianvito Lanave for his help in the deposit of the sequences in the GenBank database, Dr Lorena Cardillo, Dr Lucia Vangone and Dr Antonella De Angelis for their assistance with part of the experimental work. We are also grateful to Dr Loredana Baldi for her excellent support in the epidemiological and statistical investigation.

E TH I C A L A PPROVA L
Ethical statement is not applicable since sample collection was obtained from dead animals that were submitted to diagnostic investigations upon request of the owners or public authorities.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are openly available in the GenBank database at https://www.ncbi.nlm.nih.gov/nucle otide / under accession numbers MN086803-MN086817.