Canine Influenza A Virus
Overview and Taxonomy of Canine Influenza A Virus
Canine influenza A virus (CIV) represents a paradigm-shifting emergence of influenza A virus (IAV) into a novel mammalian host species, the domestic dog (Canis lupus familiaris). Prior to the turn of the 21st century, dogs were not considered a natural reservoir for IAV; however, two distinct host-range shifts have fundamentally altered our understanding of IAV ecology and the potential role of companion animals in viral emergence and evolution [12, 24]. The taxonomic classification of CIV places it unequivocally within the family Orthomyxoviridae, genus Influenzavirus A, characterized by a negative-sense, single-stranded, segmented RNA genome. As with all IAVs, CIV is subtyped based on the antigenic properties of its two surface glycoproteins: hemagglutinin (HA) and neuraminidase (NA). To date, two antigenically and genetically distinct subtypes have established sustained circulation in dog populations: the equine-origin H3N8 CIV and the avian-origin H3N2 CIV [6, 24]. The emergence of these viruses, their evolutionary trajectories, and their ongoing adaptation to the canine host constitute a landmark event in influenza virology, providing an invaluable model for understanding the complex processes of interspecies transmission, host adaptation, and epidemic dynamics.
1. H3N8 Canine Influenza Virus: An Equine-to-Canine Host-Range Shift
The first recognized canine-specific influenza A virus, the H3N8 subtype, emerged in the United States in 2004, although molecular clock analyses place its initial transmission event from horses to dogs around 2002 [12]. This virus originated from the H3N8 equine influenza virus (EIV), specifically a reassortant virus of the circulating Florida-1 clade [12]. Unlike the avian-to-mammalian host-range shift seen with H3N2 CIV, the H3N8 CIV emergence represents a unique instance of a direct mammalian-to-mammalian host switch [12, 23]. Phylogenetic and phylodynamic investigations have demonstrated that once established in the canine population, H3N8 CIV evolved and diverged into multiple clades and sublineages, with evidence of both intra- and inter-lineage reassortment, underscoring the virus's ability to rapidly adapt to its new host [12]. The epidemiology of H3N8 CIV in the United States has been characterized by a patchy distribution, with sporadic outbreaks in the general dog population and sustained endemic transmission within high-density environments such as large animal shelters [23]. These shelter populations act as critical refugia, allowing the virus to persist despite the overall contact heterogeneity of the greater dog population [23]. The basic reproductive number (R₀) of H3N8 CIV within these high-density settings is estimated to be high, around 3.9, yet its effective reproductive number (Rₑ) in the broader, less-connected dog population hovers near the critical threshold of 1.0, a situation that poises the virus on the extinction-invasion boundary of the host contact network [23]. This dynamic has profoundly shaped the evolution of the virus, selecting for a transmission efficiency that is sufficient to persist in these hot spots but not high enough to enable efficient, sustained spread across the general dog population [1, 23].
2. Avian-Origin H3N2 Canine Influenza Virus: An Emerging Global Pathogen
The second and more globally expansive CIV subtype is the avian-origin H3N2 virus, which emerged in dogs in East Asia, specifically China and South Korea, around 2004–2005 [1, 3, 14]. Comprehensive genetic and phylogenetic analyses have unequivocally demonstrated that all eight gene segments of the initial H3N2 CIV isolates are of avian origin, closely related to H3N2 influenza A viruses circulating in the poultry reservoir [10, 14]. This cross-species transmission event is believed to have occurred through direct spillover from an avian reservoir, likely via contact at live bird markets or poultry-raising environments [14, 20]. Since its establishment, H3N2 CIV has evolved into a genetically and antigenically diverse virus, with phylogeographic studies dividing the virus into at least seven major clades that exhibit strong geographic clustering [14]. These clades have been further categorized through principal component analysis into three primary lineages: the Origin clade, the China clade, and the Korea/USA clade [15]. Notably, the H3N2 CIV has since spread beyond Asia. In early 2015, the virus was introduced into the United States, most likely via a single introduction of a South Korean strain into the Chicago, Illinois area [2]. Phylogenetic and epidemiological tracing revealed that this introduction seeded a major outbreak in the Midwest, which subsequently spread to other states, including Georgia and North Carolina [2, 8]. The virus was later detected in Canada in late 2017, where it caused multiple epidemiologic clusters in Ontario, driven by several independent introductions of South Korean/Chinese-origin viruses over a 10-month period [8, 11]. The evolution of H3N2 CIV has been characterized by an elevated rate of nonsynonymous substitutions compared to its avian ancestors, a clear signature of intense and ongoing selection pressure as the virus adapts to its new mammalian host [1, 6]. This adaptation has included the fixation of numerous amino acid substitutions across the genome, many of which are known to play critical roles in mammalian adaptation, such as HA-G146S, M1-V15I, NS1-E227K, PA-C241Y, PB2-K251R, and PB2-G590S [6]. The interplay between host population structure and viral evolution has been particularly well-documented for H3N2 CIV. In the United States, the virus exhibits recurrent epidemic burst-and-fade-out dynamics, relying on high-density metapopulations (e.g., animal shelters and boarding kennels) for sustained circulation and requiring frequent long-distance dispersal, likely through the movement of infected dogs, to initiate new outbreaks [1, 3]. This epidemiological pattern creates evolutionary cul-de-sacs, where the virus is not selected for the attributes necessary for sustained transmission in the general, contact-heterogeneous dog population [1].
3. Molecular Taxonomy and Key Adaptive Signatures
The taxonomic differentiation of CIV extends beyond subtype classification to include a suite of molecular markers that distinguish these viruses from their source reservoirs and provide insights into their ongoing host adaptation. One of the most striking and consistent evolutionary changes observed in both H3N8 and H3N2 CIV is the truncation of the PA-X protein [22]. In the equine H3N8 and avian H3N2 influenza viruses, the PA-X open reading frame (X-ORF) encodes a full-length 61-amino-acid protein. However, upon introduction into dogs, this X-ORF becomes truncated to encode only 41 amino acids in both CIV lineages [22]. This truncation is a convergent evolutionary event that enhances viral replication, pathogenicity, and transmissibility in dogs, likely by modulating host gene expression and inflammatory responses [22]. Another critical adaptive marker identified in H3N2 CIV is a two-amino-acid insertion at the distal end of the neuraminidase (NA) stalk, a feature that has become fixed in Chinese isolates since 2010 [17]. This stalk elongation enhances viral replication in mammalian cells, including Madin-Darby canine kidney (MDCK) cells and primary canine bronchiolar epithelial cells, and increases pathogenicity in a mouse model, representing a key step in the virus's adaptation to the canine host [17]. The polymerase basic protein 2 (PB2) segment has also been a major focus of evolutionary analysis. The I714S mutation, located within the nuclear localization signal (NLS) region of PB2, has been identified as a crucial determinant of mammalian adaptation in H3N2 CIV, influencing nuclear import efficiency and ribonucleoprotein (RNP) complex assembly [4]. Notably, while the canonical mammalian-adaptive markers E627K and D701N in PB2 enhance polymerase activity in H3N2 CIV, they do not necessarily alter virulence in mice or dogs, suggesting that alternative adaptive pathways are operative in the canine host [16]. Furthermore, at a whole-genome level, H3N2 CIV shows a progressive increase in the codon adaptation index (CAI) towards the dog host and a reduction in the abundance of CpG dinucleotide motifs, a hallmark of adaptation to mammalian cellular environments and a potential mechanism for evading host innate immune recognition [6, 15].
4. Genomic Diversity and the Potential for Reassortment
Dogs are now recognized as potential "mixing vessels" for influenza A viruses, a role previously attributed primarily to swine [12, 13, 18]. The canine respiratory tract expresses both avian-type (SAα2,3-Gal) and human-type (SAα2,6-Gal) sialic acid receptors, providing a permissive environment for co-infection with avian, human, and other mammalian IAVs [13, 19]. Surveillance and experimental studies have confirmed the occurrence of natural reassortment events involving H3N2 CIV. A landmark discovery was the isolation of novel H3N6 CIVs in China, which arose from a reassortment event between a circulating H3N2 CIV and an H5N6 avian influenza virus [5]. Additionally, a reassortant feline-origin H3N2 virus was isolated from a cat in China, which possessed an NS gene derived from a circulating human seasonal H3N2 virus [9]. More recently, a reverse-zoonotic reassortment event was identified in ducks on the Leizhou Peninsula, China, where an H3N2 avian influenza virus acquired HA and PB2 segments from a canine H3N2 virus [20]. Co-infection experiments in ex vivo canine tracheal explants have demonstrated that H3N2 CIV can readily reassort with human pandemic H1N1 and avian H9N2 viruses, generating viable reassortants with enhanced replication efficiency and a heightened proinflammatory cytokine response in human alveolar epithelial cells [7]. These findings underscore the significant zoonotic risk posed by the circulation of H3N2 CIV and highlight the urgent need for ongoing genomic surveillance of influenza viruses in dog populations as part of a comprehensive One Health strategy, as advocated by the World Health Organization (WHO) and the World Organisation for Animal Health (WOAH) [7, 19, 21].
Molecular Pathogenesis and Host Tropism of H3N2 Canine Influenza Virus
The avian-origin H3N2 canine influenza virus (CIV) represents a paradigm of cross-species transmission and subsequent mammalian adaptation. Since its emergence in dogs in Asia around 2005, this virus has not only established enzootic cycles but also revealed a sophisticated molecular pathogenesis that is intimately linked to a dynamic and, in many ways, restrictive, host tropism. Unlike the equine-origin H3N8 CIV, which involved a mammalian-to-mammalian host switch, the H3N2 lineage originated from a direct avian reservoir, necessitating a more profound genetic remodeling to overcome the interspecies barrier [6, 14]. The molecular mechanisms underpinning this adaptation involve a complex interplay of mutations in the viral polymerase complex, receptor binding properties, and strategies to subvert the canine innate immune system.
Molecular Determinants of Mammalian Adaptation: The Polymerase Complex
A primary barrier for an avian influenza virus (AIV) in a mammalian host is the inefficiency of its RNA-dependent RNA polymerase (RdRp) at the lower body temperature of the mammalian respiratory tract. The H3N2 CIV has overcome this through the accumulation of specific adaptive mutations. Critical residues in the PB2 subunit, such as PB2-627 and PB2-701, are well-characterized mammalian adaptation markers in other influenza viruses. However, for H3N2 CIV, the picture is more nuanced. While introducing E627K or D701N substitutions enhances polymerase activity in vitro, these changes do not significantly alter virulence in mice or beagles, suggesting that other residues are more consequential for this specific virus-host interaction [16]. Instead, a pivotal adaptation has been identified at position 714 in the nuclear localization signal (NLS) region of PB2. The PB2 I714S mutation, which emerged during the transition from AIV to CIV, dramatically enhances polymerase activity and pathogenicity in mammalian models. Mechanistically, the I714S substitution improves the nuclear import efficiency of the PB2 protein and facilitates robust interactions within the ribonucleoprotein (RNP) complex, specifically with PA and NP, leading to more efficient RNP assembly and viral replication [4]. Concurrently, the PB2-K251R and PB2-G590S substitutions, which became fixed around 2015, have been identified as playing imperative roles in facilitating transmission and spillover, likely by further tuning polymerase function in the canine environment [6].
Host Tropism and Receptor Specificity: A Shift Towards Mammalian Recognition
The most definitive determinant of host tropism for any influenza A virus is its ability to bind to specific sialic acid (SA) receptors on host cells. Avian influenza viruses exhibit a preference for SAα2,3-Gal linkages (avian-type receptors), which are abundant in the avian intestinal tract and, to a lesser extent, in the canine respiratory tract. Conversely, human influenza viruses preferentially bind SAα2,6-Gal linkages (human-type receptors). The H3N2 CIV has undergone a striking, yet incomplete, shift in receptor specificity. Early isolates retained a predominantly avian-like receptor binding profile, but as documented by the World Health Organization (WHO) and Food and Agriculture Organization (FAO) in their pandemic risk assessments, continuous circulation in dogs has driven a gradual acquisition of human-type receptor binding. CIVs now circulating in China and North America can recognize both avian-type (SAα2,3-Gal) and human-type (SAα2,6-Gal) receptors [19, 27, 28]. A key structural change facilitating this is the HA-G146S substitution, which alters the conformation of the receptor binding site (RBS) to accommodate the human-type receptor [6]. Furthermore, the stability of the hemagglutinin (HA) protein under acidic conditions, its acid stability, is crucial for successful viral entry via membrane fusion. H3N2 CIVs have been shown to acquire increased HA acid stability over time, a trait essential for surviving the acidic environment of the mammalian upper respiratory tract and for efficient transmission via respiratory droplets [19]. The truncation of the PA-X protein, which occurs universally in CIVs, further contributes to enhanced viral replication and transmission in dogs. A full-length PA-X suppresses host gene expression, including antiviral responses; its truncation in CIVs leads to a loss of this suppression, paradoxically increasing viral yields and pathogenicity by upregulating genes related to inflammatory responses [22].
Molecular Pathogenesis: Subversion of Canine Host Defenses
The pathogenesis of H3N2 CIV is not solely about the virus acquiring new functions; it is equally about evading the canine host's intrinsic and innate immune defenses. The host restriction factor canine interferon-inducible transmembrane (caIFITM) proteins, particularly caIFITM1 and caIFITM3, are potent inhibitors of CIV replication, blocking viral entry into the host cytoplasm [29]. Similarly, the viral tetherin protein acts as a mild restriction factor, preventing the release of budding virions from the cell surface, though CIV has evolved mechanisms to partially overcome this block [34]. On the host side, the interferon response is critical. The MDA5 signaling pathway, a key sensor of viral RNA, plays a significant role in canine defense. The CARD region of MDA5 activates the IFN-β promoter, leading to the expression of antiviral proteins and cytokines that curb viral replication [32]. The virus counteracts this by inducing a specific cellular microRNA (miRNA) response. For example, H3N2 CIV infection upregulates cfa-miR-143, which targets insulin-like growth factor binding protein 5 (Igfbp5), leading to activation of the p53-caspase3 apoptotic pathway and subsequent cell death, a potential mechanism for viral dissemination or immunopathology [33]. Conversely, the host utilizes miRNAs like cfa-miR-125b and cfa-miR-151 to directly target the viral NP and NS1 mRNAs, acting as negative regulators of viral replication [30]. The virus also manipulates the cellular lipidome; proteomic analysis of infected canine lungs reveals a suppression of apoptosis and cytoskeleton proteins coupled with induction of interferon-induced proteins, underscoring a complex battle for control of the cellular environment [36].
Reassortment Potential and the Expanding Host Tropism
A cornerstone of the public health threat posed by H3N2 CIV is its ability to reassort with other influenza A viruses. The canine respiratory tract is susceptible to a wide range of human, avian, and swine influenza subtypes, as demonstrated by studies on ex vivo canine tissue explants [7, 13]. This co-infection potential has been realized in nature. Novel reassortant H3N6 CIVs have been isolated in China, arising from a H3N2 CIV backbone acquiring an N6 neuraminidase from an H5N6 AIV [5]. Furthermore, a reassortant H3N2 virus was isolated from a cat in China, which carried an NS gene from a human seasonal H3N2 virus, providing direct evidence that felines, and potentially dogs, can serve as "mixing vessels" for influenza viruses of different host origins [9]. These reassortment events are not benign; they can confer enhanced mammalian adaptation, increased replication efficiency in human airway epithelial cells, and altered cytokine induction profiles, substantially elevating zoonotic risk [7, 28].
Despite this potential, the strict host tropism of H3N2 CIV is demonstrated by its inability to infect several other species. Experimental inoculation of domestic poultry (chickens, ducks, geese, pigeons, and quails) showed no evidence of infection, viral shedding, or seroconversion, confirming a unidirectional evolution towards mammalian tropism [25]. Similarly, the virus replicates inefficiently in swine [35] and fails to transmit from experimentally infected dogs to horses [26]. However, the picture is different for other mammals. The virus replicates efficiently and transmits via direct contact in guinea pigs and, critically, can achieve 100% respiratory droplet transmission in ferrets, the gold-standard model for human influenza transmission [19, 31]. This indicates that a relatively few additional adaptive changes could allow the virus to achieve sustained human-to-human transmission. The circulating H3N2 CIVs have already acquired a haemagglutination (HA) acid stability comparable to human seasonal viruses and demonstrate robust replication in human alveolar epithelial cells [19, 28]. The fact that human populations lack pre-existing immunity to these canine-origin H3N2 viruses, as highlighted by the WHO's pandemic preparedness frameworks, underscores a significant gap in herd immunity, making dogs a potential intermediate host for the adaptation of avian influenza viruses to the human population [19].
Epidemiology and Evolutionary Dynamics of Canine Influenza A Virus
The emergence and sustained circulation of canine influenza A virus (CIV) represents a singular event in the evolutionary history of influenza A viruses (IAVs), marking the first documented instances of a respiratory pathogen of this family becoming enzootic in a companion animal population following host-switching events from distinct reservoirs. Two antigenically and genetically distinct CIV subtypes have been recognized: the equine-origin H3N8 virus, which emerged in the United States around 1999–2004, and the avian-origin H3N2 virus, which first appeared in Asia around 2004–2005 and has since become the dominant globally circulating lineage [1, 3, 24]. The epidemiological patterns and evolutionary trajectories of these viruses, particularly the H3N2 subtype, provide a unique and instructive model for understanding the constraints and drivers of viral emergence in a new host species, with profound implications for both veterinary and public health.
Origins and Host-Switching Events
The H3N8 CIV lineage arose from a direct host-range shift of the equine influenza A virus (H3N8) from horses to dogs, a rare example of a mammalian-to-mammalian transmission event. Phylodynamic analyses indicate that this transfer occurred around 2002, involving a reassortant virus derived from the circulating Florida-1 clade of equine influenza virus [12]. Once established in the canine population, H3N8 CIV spread efficiently, evolved into multiple clades, and demonstrated intra- and inter-lineage reassortment, underscoring the capacity of dogs to serve as a permissive host for influenza virus evolution [12]. However, the global epidemiology of CIV shifted dramatically with the emergence of the avian-origin H3N2 subtype.
The H3N2 CIV originated from the direct transfer of an avian influenza A virus (H3N2) from an aquatic bird reservoir to dogs, likely occurring in Asia around 2002–2005 [3, 14]. This host-switch was unidirectional; experimental studies have demonstrated that H3N2 CIVs have lost the ability to infect domestic poultry, including ducks, chickens, geese, pigeons, and quails, indicating a definitive and irreversible adaptation to the mammalian host [25]. The earliest isolations of H3N2 CIV were reported in South Korea in 2007 and in southern China between 2006 and 2007, with phylogenetic analyses confirming that all eight genomic segments were of avian origin and closely related to contemporaneous H3N2 avian influenza viruses circulating in live poultry markets [10, 14, 37]. Following its initial establishment, the virus became enzootic throughout China and South Korea, and by 2015, it had been introduced into North America, triggering a major epizootic [1, 2].
Phylogeography and Transmission Dynamics
The global spread of H3N2 CIV is characterized by a complex pattern of regional endemicity punctuated by long-distance dispersal events and recurrent epidemic fade-out. In Asia, the virus became enzootic in China and South Korea, but genomic surveillance indicates that it died out in South Korea around 2017, leaving China as the primary global reservoir [3]. The introduction of H3N2 CIV into the United States in early 2015 was traced to a single incursion from South Korea, likely via the importation of infected dogs, with the initial outbreak centered in Chicago, Illinois [2]. This introduction seeded a large-scale epidemic that spread to multiple states, but subsequent analyses revealed that the virus did not establish a single, continuously circulating lineage. Instead, the epidemiology in North America is defined by recurrent epidemic burst–fade-out dynamics, driven by multiple independent introductions of virus from Asia [1, 3]. Phylogenetic studies of outbreaks in Ontario, Canada, in 2017–2018 confirmed that the Canadian epizootic was the result of at least four distinct introductions of South Korean/Chinese-origin H3N2 CIVs over a ten-month period, forming a distinct new clade alongside contemporaneous US and Chinese strains [8]. This pattern of repeated incursion and local extinction suggests that the virus faces significant barriers to sustained transmission in the general dog population.
The fundamental epidemiological unit for H3N2 CIV is the high-density host population, such as animal shelters, boarding kennels, and veterinary clinics. Within these facilities, the virus exhibits a high reproductive potential, with basic reproductive numbers (R0) estimated to be as high as 3.9 [23]. However, in the broader, sparsely connected dog population, the effective reproductive number (Re) hovers near 1.0, placing the virus on the extinction/invasion threshold of the host contact network [1, 23]. This dynamic is driven by widespread host contact heterogeneity; the virus spreads rapidly through dense, immunologically naive subpopulations, but after all susceptible individuals in a given facility become infected and immune, the local outbreak extinguishes [3]. Sustained regional circulation therefore requires the long-distance dispersal of virus to new, susceptible populations, a process that appears to be inefficient for H3N2 CIV. This reliance on metapopulation dynamics and repeated reintroduction from Asia explains why the virus has not become permanently enzootic in North America, despite causing large, multi-state outbreaks [1, 3, 23].
Evolutionary Dynamics and Host Adaptation
The evolutionary pressures acting on H3N2 CIV following its host-switch are distinct from those in its avian reservoir and have shaped the virus in profound ways. A comprehensive analysis of the virus’s evolution over 20 years in dogs reveals that it has evolved at a constant rate, but the nature of the selective forces has changed dramatically [3]. Compared to avian influenza viruses, H3N2 CIV consistently exhibits a higher ratio of nonsynonymous to synonymous substitutions (dN/dS) across all gene segments, indicative of a large-scale shift in selection pressures [1, 6]. This elevated dN/dS reflects a period of rapid adaptive evolution as the virus adjusted to the novel canine host environment. A total of 54 amino acid substitutions have become fixed in the H3N2 CIV population, with 11 of these sites also showing high prevalence in H3N8 CIV, providing compelling evidence for convergent evolution on different CIV lineages [6].
Several of these fixed substitutions are located in key viral proteins and are known to play imperative roles in mammalian adaptation. Notably, the HA-G146S, M1-V15I, NS1-E227K, PA-C241Y, PB2-K251R, and PB2-G590S substitutions have been reported to facilitate transmission and spillover of IAVs across species barriers [6]. Most of these adaptive changes became fixed around 2015, a period that coincided with the successful spread of H3N2 CIV from South Asia to North America [6]. The PB2 gene, in particular, has been a hotspot for adaptive evolution. The I714S mutation in the nuclear localization signal (NLS) region of PB2 has been shown to be a critical determinant of mammalian adaptation, enhancing nuclear import efficiency and the assembly of the viral ribonucleoprotein (RNP) complex, thereby increasing polymerase activity and pathogenicity in mammalian cells and mice [4]. Furthermore, while the classic mammalian-adaptive mutations PB2-E627K and PB2-D701N enhance polymerase activity in H3N2 CIV, they do not appear to alter virulence in mice or dogs, suggesting that alternative pathways for mammalian adaptation are operative in this lineage [16].
A particularly striking example of host-driven evolution is the truncation of the PA-X protein. The PA-X gene of avian H3N2 influenza viruses encodes a full-length 61-amino-acid X-ORF, but all H3N2 CIVs possess a truncated PA-X encoding only 41 amino acids [22]. This truncation occurred after the virus entered the canine host and is now conserved across all circulating strains. Functional studies have demonstrated that PA-X truncation increases virus yields in cell culture, enhances viral replication and pathogenicity in dogs, and improves transmission efficiency [22]. The mechanism appears to involve a reduced suppression of host gene expression and an upregulation of genes related to inflammatory responses, which paradoxically may enhance viral replication and shedding [22]. This adaptation highlights how the virus has fine-tuned its virulence and host interaction to optimize fitness in the canine respiratory tract.
Antigenic Evolution and Receptor Binding
The antigenic evolution of H3N2 CIV has been characterized by the emergence of distinct clades with significant implications for vaccine efficacy and zoonotic risk. A new antigenically and genetically distinct clade of H3N2 CIV emerged in China in 2016, possessing mutations associated with mammalian adaptation, and rapidly replaced previously circulating strains [39]. This clade, which includes the viruses that subsequently spread to North America, has continued to evolve, with strains isolated in Guangdong Province between 2018 and 2021 forming a distinct lineage that clusters with US and northern Chinese strains from 2017–2019 [38]. Critically, the hemagglutinin (HA) of contemporary H3N2 CIVs has undergone a functional shift in receptor-binding specificity. While early isolates maintained a strict preference for avian-type α2,3-linked sialic acid receptors, viruses isolated after 2015 have acquired the ability to bind to human-type α2,6-linked sialic acid receptors [19, 27, 28]. This change is a fundamental prerequisite for zoonotic transmission and pandemic potential. Structural analyses have confirmed that the HA of US H3N2 CIVs retains an avian-like receptor-binding profile, but the acquisition of human-type receptor binding in later isolates represents a significant step in mammalian adaptation [19, 27].
Reassortment and the Potential for Novel Genotypes
The canine respiratory tract is susceptible to a wide range of influenza A viruses, including human seasonal strains, avian viruses, and swine viruses, creating opportunities for co-infection and reassortment [7, 13]. Dogs have been shown to express both avian-type (α2,3) and human-type (α2,6) sialic acid receptors in the respiratory tract, fulfilling a key criterion for a “mixing vessel” host [13, 18]. Experimental co-infection of canine tracheal explants with H3N2 CIV and either human pandemic H1N1 or avian H9N2 viruses results in a high rate of reassortment, with the MP, NA, and NS segments most frequently exchanged [7]. The resulting reassortants are viable, replicate efficiently in human alveolar epithelial cells, and induce a more robust proinflammatory cytokine response than the parental strains, suggesting an enhanced zoonotic risk [7]. Field evidence supports the occurrence of natural reassortment events. A novel H3N6 CIV was isolated from dogs in Liaoning, China, in 2018–2019, which was generated by reassortment between a circulating H3N2 CIV and an H5N6 avian influenza virus [5]. Furthermore, a reassortant H3N2 virus isolated from a cat in Jiangsu, China, possessed an NS gene derived from a human H3N2 virus, while the remaining seven segments were of CIV origin [9]. These findings demonstrate that reassortment between CIVs and other IAVs is not merely a laboratory phenomenon but a real-world occurrence that could generate viruses with altered host range and pathogenicity.
Epidemiological Significance and Public Health Context
The World Organisation for Animal Health (WOAH) and the Food and Agriculture Organization (FAO) recognize the emergence of influenza in dogs as a significant development in the ecology of IAVs, given the close and frequent contact between dogs and humans. The Centers for Disease Control and Prevention (CDC) and the World Health Organization (WHO) have highlighted the potential for canine influenza viruses to act as a source of novel human infections. While no sustained human-to-human transmission of CIV has been documented, serological evidence indicates that human populations lack immunity to H3N2 CIVs, and preexisting immunity from seasonal human H3N2 viruses does not provide cross-protection [19]. The virus has demonstrated the ability to replicate in human airway epithelial cells and to transmit via respiratory droplets in ferrets, the gold-standard model for human influenza transmission [19]. The continued evolution of H3N2 CIV in dogs, including the acquisition of human-type receptor binding, increased HA acid stability, and enhanced replication in human cells, suggests that the virus is progressively adapting to the human host environment [19]. The epidemiological pattern of rapid spread through high-density dog populations, coupled with the potential for long-distance dispersal via human-mediated dog movement, creates a scenario where a fully human-adapted variant could emerge and spread rapidly before detection. The ongoing circulation of H3N2 CIV in China, combined with repeated introductions into North America and Europe, underscores the need for sustained, global surveillance of influenza viruses in companion animals, as mandated by the WOAH/FAO international standards for influenza surveillance.
Clinical Manifestations and Pathological Findings in Dogs
Canine influenza A virus (CIV), particularly the avian-origin H3N2 subtype, presents a spectrum of clinical disease in dogs that ranges from subclinical infection to severe, and occasionally fatal, respiratory compromise. The clinical manifestations are a direct consequence of the virus’s tropism for the respiratory epithelium, its capacity to induce a robust inflammatory response, and the frequent complication by secondary bacterial infections. Understanding the full pathological picture is critical for differential diagnosis, outbreak management, and assessing the zoonotic risk posed by this emerging pathogen.
Spectrum of Clinical Disease
The clinical course of H3N2 CIV infection in dogs is characterized by an acute onset of respiratory signs following an incubation period typically ranging from 2 to 4 days. The most consistent and prominent clinical sign is a persistent, productive cough that can last for 10 to 21 days or longer, often described as a “honking” cough that may be mistaken for kennel cough (tracheobronchitis) [24, 37]. This cough is frequently accompanied by serous to mucopurulent nasal discharge and ocular discharge [26, 31, 37]. Affected dogs exhibit varying degrees of lethargy, anorexia, and pyrexia. Experimental infections have consistently demonstrated elevated rectal temperatures, often peaking between 1- and 4-days post-inoculation (dpi) [26, 37, 41]. In a study comparing inoculation routes, intratracheal administration was found to induce more severe clinical signs, including higher and more sustained fevers, compared to intranasal inoculation, highlighting the role of lower respiratory tract involvement in systemic illness [41].
While the majority of infections are mild and self-limiting, a subset of dogs develops severe disease, particularly in high-density populations such as shelters and kennels. In these settings, morbidity rates can approach 100%, with a notable proportion of dogs progressing to pneumonia [11, 24]. The mortality rate directly attributable to CIV infection is generally low, estimated at 1-5% in outbreak situations, but can be significantly higher in cases complicated by secondary bacterial infections or in very young, geriatric, or immunocompromised animals [11]. A 2018 outbreak in Ontario, Canada, involving 104 dogs, reported a 2% mortality rate directly attributable to CIV, with death occurring in two cases [11]. The clinical severity is also modulated by the viral strain; for instance, the emergence of a novel antigenically distinct clade in China in 2016 was associated with increased pathogenicity in experimental settings [39].
Pathological Findings: Gross and Histopathological Lesions
The pathological hallmarks of H3N2 CIV infection are centered on the respiratory tract, with the severity of lesions correlating with the clinical presentation. Gross pathological examination of dogs that succumb to or are euthanized during severe infection reveals characteristic lung consolidations. These lesions are typically multifocal to coalescing, firm, dark red to purple, and are most commonly distributed in the cranioventral lung lobes, particularly the right middle, left cranial, and right cranial lobes [26, 41, 42]. This cranioventral distribution is a classic pattern for aspiration pneumonia and bronchopneumonia, reflecting the gravitational spread of infectious exudate. In a radiographic study of six naturally infected dogs, an unstructured interstitial to alveolar pulmonary pattern with a cranioventral distribution was the most common finding, and mild pleural effusion was noted in one case [42]. The consolidated lung tissue is often heavy and fails to collapse upon opening the thoracic cavity. In severe cases, the entire lung lobe may be affected, and purulent exudate may be present in the airways.
Histopathological examination provides a detailed view of the cytopathic and inflammatory damage. The hallmark lesion is a severe, necrotizing tracheobronchitis and bronchioalveolitis [37]. The tracheal and bronchial epithelium undergoes extensive necrosis and sloughing, with loss of cilia and cellular debris accumulating in the lumen. The alveolar spaces are filled with a mixture of proteinaceous edema fluid, fibrin, sloughed epithelial cells, and inflammatory cells, primarily neutrophils and macrophages, a picture consistent with acute interstitial pneumonia [37, 41]. The alveolar septa are thickened due to congestion, edema, and infiltration by mononuclear cells. In cases with secondary bacterial infection, the histopathology is compounded by suppurative inflammation, with large numbers of neutrophils and bacterial colonies visible within the airways and alveoli [26, 40]. A study by Yamanaka et al. (2012) isolated Streptococcus equi subsp. zooepidemicus from the lung consolidations of experimentally infected dogs, underscoring the importance of bacterial co-infection in exacerbating pathology [26].
Pathogenesis and Host Response
The severity of clinical disease and pathological lesions is not merely a function of direct viral cytopathology but is significantly driven by the host’s immune and inflammatory response. The virus initially infects and replicates within the epithelial cells of the upper and lower respiratory tract. Viral antigen can be detected in the trachea, bronchi, bronchioles, and alveoli [41]. The host response involves a complex interplay of innate immune factors. Upon infection, there is a rapid upregulation of interferon-inducible transmembrane proteins (caIFITMs), which act as potent host restriction factors to limit viral replication [29]. Similarly, the MDA5 pathway is activated, leading to the production of type I interferons and pro-inflammatory cytokines [32]. However, this antiviral response can become dysregulated.
Proteomic and transcriptomic analyses of infected canine lung tissue have revealed a significant upregulation of proteins and genes associated with apoptosis, the cytoskeleton, and innate immunity [36]. Specifically, the virus induces the expression of cfa-miR-143, which promotes apoptosis in infected cells via the p53-caspase3 pathway, a mechanism that may contribute to tissue damage [33]. Furthermore, the virus has evolved strategies to evade host defenses. The truncation of the PA-X protein, a common feature of both H3N8 and H3N2 CIVs, has been shown to enhance viral replication, pathogenicity, and transmission in dogs. This truncation leads to a more robust suppression of host gene expression and an exaggerated upregulation of genes related to inflammatory responses, thereby driving more severe immunopathology [22]. The resulting "cytokine storm," characterized by elevated levels of TNF-α, IL-6, and other chemokines, contributes to the alveolar damage, pulmonary edema, and systemic signs of illness [44, 45].
Radiographic and Clinical Pathology Correlates
Thoracic radiography is a valuable tool for assessing the extent and severity of pulmonary involvement. As described by Secrest and Sharma (2016), the predominant radiographic pattern is an unstructured interstitial to alveolar opacity, most frequently affecting the cranioventral lung lobes [42]. This pattern is consistent with the gross pathological findings of bronchopneumonia and distinguishes CIV from other causes of canine respiratory disease that may present with a more diffuse or caudodorsal distribution. The absence of intrathoracic lymphadenopathy or cranial mediastinal widening in the studied cases helps differentiate CIV from fungal diseases or neoplasia [42]. Clinically, these radiographic changes correlate with the severity of hypoxemia and respiratory distress observed in severely affected dogs.
Factors Influencing Clinical Outcome
Several factors modulate the clinical outcome of H3N2 CIV infection. Host population structure is paramount. The virus is highly adapted to spread in dense populations, such as animal shelters and boarding kennels, where the basic reproductive number (R0) can be as high as 3.9 [23]. In these settings, rapid transmission and high morbidity are the norm. Conversely, in the general dog population with lower contact rates, the virus struggles to maintain sustained transmission and often undergoes epidemic fade-out [1, 3]. Viral genetics also play a critical role. The emergence of new clades with enhanced receptor binding (e.g., recognition of human-like SAα2,6-Gal receptors), increased hemagglutination (HA) acid stability, and higher replication efficiency in human airway cells indicates ongoing adaptation that could alter clinical severity in dogs and pose a greater zoonotic risk [19]. Co-infections are a major determinant of disease severity. The canine infectious respiratory disease complex (CIRDC) frequently involves multiple pathogens. While CIV is a primary pathogen, its clinical impact is often amplified by concurrent infections with Mycoplasma cynos, Mycoplasma canis, Bordetella bronchiseptica, or Streptococcus equi subsp. zooepidemicus [26, 43]. The presence of these agents can transform a mild, self-limiting illness into a severe, life-threatening bronchopneumonia.
Diagnostic Approaches for Canine Influenza A Virus Detection
The accurate and timely detection of canine influenza A virus (CIV) is paramount for effective outbreak management, epidemiological surveillance, and the implementation of appropriate biosecurity measures. Given the virus's capacity for rapid dissemination within susceptible dog populations, particularly in high-density environments such as shelters, boarding kennels, and veterinary clinics, diagnostic approaches must balance sensitivity, specificity, and speed. The diagnostic armamentarium for CIV encompasses a spectrum of methodologies, ranging from traditional virological techniques to advanced molecular and serological assays, each with distinct applications and interpretative nuances. The selection of an appropriate diagnostic strategy is dictated by the clinical context, the stage of infection, the purpose of testing (e.g., individual diagnosis versus population surveillance), and the available laboratory infrastructure.
Molecular Detection: Reverse Transcription Polymerase Chain Reaction (RT-PCR) and Quantitative Variants
The cornerstone of contemporary CIV diagnosis is the detection of viral RNA through reverse transcription polymerase chain reaction (RT-PCR) and its quantitative derivative, real-time RT-PCR (RT-qPCR). These assays offer unparalleled sensitivity and specificity, enabling the detection of viral nucleic acids in clinical specimens before the onset of clinical signs and during the early stages of infection when viral shedding is at its peak. The vast majority of field surveillance studies and outbreak investigations rely on these platforms [2, 3, 5, 11, 25, 31, 36-38, 41-43, 46-48, 53].
The most common target for diagnostic RT-PCR assays is the highly conserved matrix (M) gene, which is shared across all influenza A virus subtypes [46, 48, 53]. This pan-influenza A approach allows for initial screening of suspect cases. Subsequently, subtyping assays, typically targeting the hemagglutinin (HA) and neuraminidase (NA) genes, are employed to differentiate between the two principal CIV subtypes, H3N8 and H3N2, and to distinguish them from other influenza A viruses that might sporadically infect canines [2, 38, 47]. The advent of multiplex RT-PCR panels has streamlined this process, allowing for the simultaneous detection of multiple respiratory pathogens, including canine parainfluenza virus, canine respiratory coronavirus, and Bordetella bronchiseptica, within the canine infectious respiratory disease complex (CIRDC) [43]. This syndromic approach is particularly valuable in clinical settings where co-infections are common and can complicate the clinical picture.
Sample Types and Collection Considerations. Nasal swabs, oropharyngeal swabs, and nasal washes constitute the primary specimens for RT-PCR-based diagnosis [25, 31, 37, 38, 41, 47, 53]. The anatomical site of viral replication is critical; H3N2 CIV replicates predominantly in the upper respiratory tract, making nasal and oropharyngeal swabs highly sensitive during the acute phase of illness [47]. In experimentally infected dogs, viral RNA can be detected as early as one day post-inoculation and shedding can persist for 7–10 days [31, 37, 41]. For optimal sensitivity, samples should be collected within the first 3–4 days of clinical signs. Furthermore, studies have demonstrated that different inoculation routes influence viral distribution within the respiratory tract, with intratracheal inoculation resulting in more robust replication in both upper and lower airways compared to intranasal inoculation, which has implications for the diagnostic yield of different sample types [41].
Interpretation of Results. A positive RT-PCR result confirms the presence of viral RNA but does not necessarily indicate the presence of infectious virus or active shedding, as nucleic acids can be detected for a period after the virus has been cleared. Conversely, a negative result does not definitively rule out infection, particularly if the sample was collected late in the disease course when viral load has declined, or if the sample was improperly handled. Despite these limitations, RT-qPCR remains the most reliable method for confirming acute CIV infection and is widely recommended by international health authorities, including the World Organisation for Animal Health (WOAH), for confirmatory testing.
Virus Isolation: The Historical Gold Standard
Virus isolation in embryonated chicken eggs or in Madin-Darby canine kidney (MDCK) cells remains a cornerstone of definitive CIV diagnosis and is essential for the procurement of live virus for downstream applications such as antigenic characterization, antiviral susceptibility testing, and vaccine development [2, 16, 22, 25, 31, 35, 37, 41, 48, 56].
The standard protocol involves inoculating clinical specimens (e.g., nasal swab eluates or tissue homogenates) into the allantoic cavity of 9–11-day-old specific-pathogen-free (SPF) embryonated chicken eggs [25, 37]. After 48–72 hours of incubation at 35°C, allantoic fluid is harvested and tested for hemagglutinating activity. While highly sensitive for many influenza A strains, this method is labor-intensive, time-consuming (taking several days), and requires specialized biocontainment facilities. It is critical to note that some CIV strains, particularly those with mammalian adaptive mutations, may not replicate efficiently in eggs, potentially leading to false-negative results.
Cell culture-based isolation using MDCK cells offers an alternative that may be more suitable for mammalian-adapted viruses [2, 16, 22, 41, 48]. Following inoculation, cell monolayers are monitored for the development of a cytopathic effect (CPE), characterized by cell rounding, detachment, and syncytia formation. The presence of virus is confirmed by hemadsorption or immunofluorescence using subtype-specific antibodies. The susceptibility of different MDCK cell clones to CIV can vary significantly, influencing the efficiency of virus isolation [50-52, 55]. High-yield clones, such as those identified for influenza vaccine production, may be more sensitive for isolating CIV field strains [51, 52, 55]. Despite its complexity, virus isolation is irreplaceable for obtaining the virus stocks needed for detailed genetic and phenotypic analysis, including the study of reassortment events, as documented in the emergence of novel H3N6 [5] and reassortant H3N2 viruses [9].
Serological Surveillance: Detecting Past Exposure and Population Immunity
Serological assays are indispensable for epidemiological surveillance, retrospective diagnosis, and assessing vaccine immunogenicity. They detect antibodies against CIV, which typically appear 7–10 days post-infection and can persist for months to years [31, 37, 47, 54].
The hemagglutination inhibition (HI) assay is the most widely used serological test for subtype-specific antibody detection [31, 37, 46, 47]. This assay measures the ability of antibodies in serum to block the hemagglutination of red blood cells by the virus. HI titers correlate with protective immunity and are used to monitor seroconversion following natural infection or vaccination [56]. However, the HI assay is laborious, requires a constant supply of standardized antigens and red blood cells, and is subject to inter-laboratory variability.
Enzyme-linked immunosorbent assays (ELISAs) offer a more high-throughput, standardized, and objective alternative [46, 54]. Commercial ELISAs that detect antibodies against the influenza A virus nucleoprotein (NP) are available and can identify exposure to any influenza A subtype, making them useful for initial screening. However, they do not provide subtype-specific information. Subsequent confirmatory HI testing is required to differentiate between H3N2 and H3N8 exposures.
Interpretation and Limitations. A fundamental limitation of serology is its inability to diagnose acute infections, as antibodies are not present during the early stage of illness. Paired serum samples, collected during the acute and convalescent phases (2–4 weeks apart), are necessary to demonstrate a four-fold or greater rise in antibody titer, which is diagnostic of recent infection. Serosurveys are critical for understanding the true prevalence of CIV in a population, as many infections are subclinical or mild. Studies in military working dogs in Korea [54] and coyotes in Illinois [49] have employed serological methods to demonstrate the absence of CIV exposure in these specific populations, providing valuable baseline data for ecological and epidemiological assessments.
Point-of-Care and Rapid Antigen Tests
Rapid antigen detection tests, typically lateral flow immunochromatographic assays, are available for veterinary use and offer the advantage of speed, with results obtainable in 15–30 minutes. These tests detect viral NP in nasal swabs or other respiratory secretions. However, their sensitivity is substantially lower than that of RT-PCR, particularly in samples with low viral loads or when testing is performed later in the course of disease [53]. A study in Bangladesh using a commercial rapid CIV antigen test kit reported zero positive results among 50 dogs, despite clinical suspicion, highlighting the limitations of this approach for diagnosis [53]. While useful as a preliminary screening tool in outbreak situations where immediate action is required, a negative rapid test should always be confirmed by a more sensitive molecular method. The CDC and WOAH generally do not recommend rapid antigen tests as stand-alone diagnostic tools for confirmatory purposes due to their lower sensitivity.
Genomic Sequencing and Phylogenetic Analysis
Beyond simple detection, genomic sequencing of CIV isolates is essential for molecular epidemiology, evolutionary tracking, and risk assessment. Whole-genome sequencing, now increasingly accessible through next-generation sequencing (NGS) platforms, provides a comprehensive view of the viral genome, enabling the identification of key mutations associated with mammalian adaptation, antigenic drift, and antiviral resistance [1-3, 5, 8, 10, 14, 15, 20, 22, 38, 39, 47].
Phylogenetic analysis of HA and NA gene sequences has been instrumental in tracing the global spread of H3N2 CIV from its Asian origin to North America [1-3] and in identifying multiple, distinct introductions of the virus into the United States and Canada [1, 8]. This analysis revealed that the Canadian outbreak in 2017–2018 was driven by multiple introductions of South Korean/Chinese-origin viruses [8] and that recurrent epidemic fade-out in the United States is followed by reintroduction from Asian lineages [1, 3]. Furthermore, sequencing has identified the emergence of antigenically distinct clades, such as the clade that emerged in China in 2016, which possessed mutations associated with increased mammalian adaptation and zoonotic potential [39].
Specific mutations in the polymerase basic protein 2 (PB2), such as I714S [4], and in other genes like HA, M1, and NS1 have been shown to influence nuclear import efficiency, RNP complex assembly, and host adaptation [6]. The detection of these markers through genomic surveillance provides critical insights into the ongoing evolution of CIV and its potential threat to public health.
Imaging and Pathological Diagnostics
While not a direct method for viral detection, thoracic radiography can support a clinical diagnosis of CIV and help differentiate it from other causes of respiratory disease. The radiographic hallmarks of CIV infection include an unstructured interstitial to alveolar pulmonary pattern, often with a cranioventral distribution, predominantly affecting the right middle, left cranial, and right cranial lung lobes [42]. Pleural effusion is an uncommon finding. These patterns, while not pathognomonic, can raise the index of suspicion for CIV in a dog presenting with acute respiratory signs, prompting definitive testing.
Histopathological examination of lung tissue from severely affected or deceased animals reveals characteristic lesions, including necrotizing tracheobronchitis, bronchioalveolitis, and the presence of viral antigen within respiratory epithelial cells, as confirmed by immunohistochemistry [37, 41]. Post-mortem diagnosis is critical for understanding the pathogenesis of emerging strains and for confirming the cause of death in fatal cases, as was observed in the 2% mortality rate during the 2017–2018 Ontario outbreak [11]. The combined use of virological, serological, pathological, and genomic methods provides a comprehensive diagnostic framework that is essential for the effective management and control of canine influenza.
Transmission Dynamics and Interspecies Barriers
The emergence and sustained circulation of H3N2 canine influenza virus (CIV) within the global dog population represents a singular event in influenza ecology: the successful host-switch of an avian influenza A virus (IAV) into a novel mammalian companion animal species, followed by its establishment as a enzootic pathogen. Understanding the transmission dynamics of H3N2 CIV requires a multi-faceted analysis of its intraspecies spread, its capacity (or lack thereof) to traverse species barriers to other mammals and birds, and the evolutionary and ecological forces that govern its circulation. The virus, which originated from an avian reservoir in Asia around 2004-2005 [3, 14], has since demonstrated a complex epidemiological pattern characterized by rapid, high-morbidity outbreaks in dense populations, recurrent epidemic fade-out, and a reliance on metapopulation dynamics for regional persistence. Critically, while H3N2 CIV has shown a remarkable ability to adapt to the canine host, its transmission to other species, including humans, poultry, swine, and horses, remains highly constrained, though not without significant risk, particularly regarding reassortment potential.
Intraspecies Transmission Dynamics in Dogs: The Role of Host Population Structure
The transmission of H3N2 CIV among dogs is not uniform but is profoundly shaped by the social and ecological structure of the host population. The virus exhibits a high basic reproductive number (R₀) within high-density environments such as animal shelters, boarding kennels, and veterinary clinics. Early modeling of H3N8 CIV dynamics, which shares a similar epidemiological profile, estimated a mean R₀ of 3.9 within such facilities, indicating a high potential for explosive outbreaks [23]. For H3N2 CIV, epidemiological models have consistently estimated an R₀ between 1.0 and 1.5 across most U.S. outbreaks, a value that is consistent with maintained but heterogeneous circulation [1]. This relatively low R₀ in the general dog population, contrasted with the high R₀ in dense subpopulations, is a defining feature of CIV transmission.
The virus’s reliance on these “hotspots” is driven by widespread host contact heterogeneity. The general dog population is characterized by a sparsely connected network of individuals, where opportunities for transmission are limited. In contrast, shelters and kennels represent highly connected, immunologically naive metapopulations that act as refugia, allowing the virus to persist locally before dying out [1, 23]. This dynamic leads to recurrent epidemic burst–fade-out cycles, where the virus rapidly sweeps through a susceptible shelter population, infecting nearly all individuals, and then fades out locally as herd immunity develops [1, 3]. Sustained regional circulation, therefore, requires the long-distance dispersal of the virus to initiate new outbreaks in other naive populations, a process that is often facilitated by the movement of infected dogs, particularly those recently imported from endemic regions [3, 8, 11].
This pattern is vividly illustrated by the introduction and subsequent spread of H3N2 CIV in North America. The virus was first detected in the United States in Chicago, Illinois, in March 2015, following a single introduction from South Korea [2]. Despite local control measures, the virus spread to several other states, including Georgia and North Carolina, but these secondary outbreaks typically ended within a few months [2]. A similar pattern was observed in Ontario, Canada, where the 2017-2018 outbreak was driven by multiple introductions of Asian-origin viruses over a 10-month period, leading to distinct epidemiological clusters that were contained through aggressive testing and isolation [8, 11]. These observations underscore that H3N2 CIV, while capable of causing significant morbidity within groups, has not yet evolved the transmission efficiency required to sustain itself in a general dog population with low contact rates, a phenomenon described as an “evolutionary cul-de-sac” [1]. The virus is therefore poised on the extinction/invasion threshold of the host contact network, and its continued presence in North America is dependent on repeated reintroduction from Asia, particularly from China, where the virus remains endemic after dying out in South Korea around 2017 [3].
Interspecies Barriers: A One Health Perspective on Spillover and Spillback
A critical aspect of H3N2 CIV’s ecology is its host range. While the virus has successfully adapted to dogs, it has not demonstrated a similar capacity to infect other species efficiently. This is a crucial finding from a One Health perspective, as it suggests that the immediate risk of a large-scale spillover event from dogs to other mammals, including humans, is currently low, but not negligible.
Transmission to Avian Species: Given its avian origin, a key question is whether H3N2 CIV can transmit back to birds. Experimental infections have definitively shown that domestic poultry, including chickens, ducks, geese, pigeons, and quails, are not susceptible to H3N2 CIV [25]. Oropharyngeal and cloacal swabs from experimentally infected birds were negative for virus, and no seroconversion or clinical signs were observed [25]. This indicates a unidirectional evolution of mammalian host tropism, where the virus has lost the ability to replicate in its original avian reservoir. This finding is further supported by field studies showing that H3N2 CIV replicates poorly in chickens and cannot be transmitted via contact or aerosol in this species [47]. The only exception is ducks, which can support low-level replication, but this is insufficient to facilitate onward transmission [47].
Transmission to Swine and Horses: The potential for H3N2 CIV to infect swine is of particular concern, as pigs are considered classic “mixing vessels” for influenza viruses due to the presence of both avian-type (α2,3-linked sialic acid) and human-type (α2,6-linked sialic acid) receptors in their respiratory tract [18]. However, experimental infection of pigs with the avian-origin H3N2 CIV that emerged in the United States in 2015 demonstrated that the virus does not replicate efficiently in swine [35]. Similarly, studies on the H3N8 CIV lineage have shown that infected dogs do not transmit the virus to horses, even when housed in close contact [26]. These findings suggest that the canine-adapted H3N2 virus has not acquired the necessary molecular determinants to productively infect these other mammalian hosts.
Transmission to Humans and Zoonotic Potential: The most significant public health question surrounding H3N2 CIV is its zoonotic potential. To date, there have been no confirmed reports of sustained human-to-human transmission or even sporadic human infections with H3N2 CIV. However, a growing body of evidence indicates that the virus is accumulating adaptive mutations that bring it closer to a state where human infection could become possible. This is a critical area of active research, as the World Health Organization (WHO) and the Food and Agriculture Organization of the United Nations (FAO) consider the emergence of novel influenza A viruses with pandemic potential a top priority for global health security.
The barriers to human infection are multi-layered. First, the hemagglutinin (HA) of early H3N2 CIV isolates demonstrated a strong preference for binding to avian-type (α2,3-linked) sialic acid receptors, which are abundant in the canine respiratory tract but scarce in the human upper airway [19, 27]. However, over the course of adaptation in dogs, H3N2 CIVs have undergone a critical shift in receptor-binding specificity. Contemporary isolates, particularly those from the novel clade that emerged in China around 2016, have acquired the ability to recognize both avian-type and human-type (α2,6-linked) sialic acid receptors [19, 28]. This dual receptor-binding capacity is a prerequisite for human infection and is a hallmark of pandemic influenza viruses.
Second, the virus has shown increasing replication efficiency in human airway epithelial cells. Studies have demonstrated that H3N2 CIVs isolated after 2015 replicate to titers comparable to seasonal human influenza viruses in human airway epithelial cells [19, 27]. Furthermore, the virus has acquired increased hemagglutinin acid stability, a trait that is essential for surviving the acidic environment of the human upper respiratory tract and facilitating transmission via respiratory droplets [19]. In a landmark study using a ferret model, the gold standard for assessing influenza transmission, contemporary H3N2 CIVs achieved a 100% transmission rate via respiratory droplets, a stark contrast to the inefficient transmission observed with earlier isolates [19]. This finding is alarming, as it demonstrates that the virus has the potential to become highly transmissible in a mammalian model that closely mimics human physiology.
Third, the human population is largely immunologically naive to H3N2 CIV. Serological surveys have shown that pre-existing immunity derived from seasonal human H3N2 influenza viruses does not provide cross-protection against H3N2 CIVs [19]. This means that if a canine-origin H3N2 virus were to acquire the ability to transmit efficiently among humans, it would encounter a completely susceptible population, a scenario that could precipitate a pandemic.
Molecular Determinants of Host Range and Transmission
The differences in transmission dynamics and interspecies barriers are underpinned by specific molecular changes in the viral genome. The host-switch from avian to canine hosts has been accompanied by a dramatic shift in selection pressures, with H3N2 CIV exhibiting a significantly higher ratio of non-synonymous to synonymous substitutions (dN/dS) compared to its avian ancestors, indicating rapid adaptive evolution [1, 6, 14]. Several key mutations have been identified that are critical for mammalian adaptation and transmission.
Polymerase Basic Protein 2 (PB2): The PB2 protein is a major determinant of host range. In avian influenza viruses, the presence of a glutamic acid (E) at position 627 is associated with cold-adapted replication, while a lysine (K) at this position is a well-known marker of mammalian adaptation. However, in H3N2 CIV, the E627K substitution alone does not enhance virulence in mice or dogs, suggesting that other residues are more critical for canine adaptation [16]. Instead, the I714S mutation in the nuclear localization signal (NLS) region of PB2 has been identified as a key factor. This mutation, which arose during the adaptation from avian to canine hosts, enhances nuclear import efficiency and facilitates the assembly of the viral ribonucleoprotein (RNP) complex, leading to increased polymerase activity and replication in mammalian cells [4]. Other adaptive mutations in PB2, such as K251R and G590S, have also become fixed in the H3N2 CIV population and are known to play roles in mammalian transmission [6].
Hemagglutinin (HA): The G146S substitution in the HA protein is a critical adaptation that contributes to the shift in receptor-binding preference from avian-type to human-type sialic acids [6]. This mutation, along with others in the receptor-binding site, has allowed H3N2 CIV to acquire the dual receptor-binding specificity that is a prerequisite for human infection [19, 28].
PA-X: The PA-X protein, a virulence factor encoded by a +1 frameshift in the PA gene, has undergone a unique truncation in both H3N8 and H3N2 CIVs. While avian and equine influenza viruses encode a full-length 61-amino-acid PA-X, CIVs encode a truncated 41-amino-acid version [22]. This truncation is not a loss of function but an adaptive change that enhances viral replication, pathogenicity, and transmission in dogs. The truncated PA-X appears to suppress host gene expression more effectively and upregulate inflammatory responses, contributing to the increased virulence of CIVs in their new host [22].
The Role of Reassortment in Expanding Host Range
Perhaps the greatest long-term risk posed by H3N2 CIV is its potential to act as a “mixing vessel” for the generation of novel reassortant viruses with altered host range. The canine respiratory tract is susceptible to a wide range of influenza A viruses, including human seasonal strains and avian viruses [7, 13]. Co-infection of a dog with a human influenza virus and a canine influenza virus could lead to the exchange of gene segments, potentially creating a reassortant virus that possesses the human-adapted surface proteins (HA and NA) for efficient human-to-human transmission, combined with the internal genes of a canine-adapted virus that confer high replication fitness in mammals.
This is not a theoretical risk. Experimental co-infection of canine tracheal explants with H3N2 CIV and human pandemic H1N1 or avian H9N2 viruses resulted in a high rate of reassortment, with the most frequent exchanges occurring in the matrix protein (MP), neuraminidase (NA), and non-structural (NS) segments [7]. The resulting reassortants were viable and able to replicate in human alveolar epithelial cells, with some showing enhanced replication efficiency and a more robust proinflammatory cytokine response than the parental strains [7]. Furthermore, a novel reassortant H3N6 CIV has been isolated from dogs in China, which likely arose from a reassortment event between an H3N2 CIV and an H5N6 avian influenza virus [5]. This reassortant showed increased mammalian adaptation ability compared to its H3N2 parent [5]. Similarly, a reassortant H3N2 virus isolated from a cat in China contained an NS gene from a human H3N2 virus, further demonstrating that companion animals can serve as intermediates for genetic exchange between human and animal influenza viruses [9]. These findings highlight the ongoing risk that dogs, due to their close contact with humans and their susceptibility to multiple IAV subtypes, could serve as a platform for the emergence of a pandemic influenza virus. The World Organisation for Animal Health (WOAH) and the FAO have emphasized the need for enhanced surveillance of influenza viruses in companion animals to monitor this risk.
Prevention, Control, and Public Health Implications
The management of canine influenza A virus (CIV), particularly the avian-origin H3N2 subtype, necessitates a multi-faceted strategy that integrates biosecurity, vaccination, antiviral therapeutics, and robust surveillance. The public health implications of this pathogen are profound, given the intimate cohabitation of humans and dogs, the virus’s demonstrated capacity for reassortment, and its ongoing adaptive evolution towards mammalian hosts. This section provides an exhaustive analysis of the current strategies for prevention and control, while critically evaluating the zoonotic risk and the broader implications for global health security.
Vaccination Strategies and Immunoprophylaxis
Vaccination remains the cornerstone of preventing clinical disease and reducing viral shedding in canine populations. The development of effective vaccines against H3N2 CIV has been a priority since its emergence in North America. A significant advancement is the development of an injectable RNA Particle (RP) vaccine. In a controlled challenge study, dogs vaccinated with an RP-CIV H3N2 vaccine demonstrated a statistically significant reduction in the duration and severity of clinical signs, a marked decrease in the duration and quantity of viral shedding, and a substantial reduction in lung consolidation and the incidence of suppurative pneumonia compared to placebo controls [40]. This vaccine was deemed safe, with lethargy being the most common adverse event, reported at a rate of only 1.6% [40]. This highlights the potential of next-generation vaccine platforms to provide robust protection while maintaining an excellent safety profile.
Beyond traditional injectable vaccines, novel delivery systems are being explored to improve compliance and efficacy, particularly in shelter and kennel environments. Insertion-responsive microneedles (IRMNs) represent a promising transcutaneous vaccination platform. In a guinea pig model, IRMNs coated with inactivated H3N2 CIV antigen elicited hemagglutination inhibition (HI) antibody titers that were two-fold higher than those induced by conventional intramuscular injection [56]. Furthermore, when challenged with a wild-type H3N2 virus, both the IRMN and intramuscular groups exhibited complete elimination of viral shedding by 8 days post-infection [56]. The IRMN system also demonstrated superior thermal stability, maintaining antigenic activity for three weeks at 50°C, a condition that completely inactivated the liquid formulation [56]. This thermostability is a critical advantage for field deployment and vaccination campaigns in resource-limited settings.
The immunological rationale for vaccination is further supported by studies on the canine innate immune system. The interferon-inducible transmembrane (IFITM) proteins are potent host restriction factors. Canine IFITM1, IFITM2a, IFITM2b, and IFITM3 have been shown to possess potent antiviral activity against influenza virus, with their expression upregulated upon interferon stimulation or viral infection [29]. Similarly, the MDA5 protein, a key sensor in the RIG-I-like receptor family, plays a crucial role in activating the interferon-β promoter via its CARD domain, thereby inhibiting CIV replication [32]. These findings underscore that effective vaccination should aim to prime these innate immune pathways, providing a first line of defense that can rapidly curtail viral replication upon exposure.
Biosecurity, Containment, and Antiviral Therapeutics
The epidemiological dynamics of H3N2 CIV are characterized by rapid spread within high-density populations, such as animal shelters and boarding kennels, followed by local extinction [1, 3]. This pattern is driven by the virus’s reliance on metapopulations of susceptible hosts and its inability to sustain transmission in the broader, more sparsely connected dog population without repeated reintroduction [1, 23]. Consequently, biosecurity measures are paramount. The successful containment of the 2017-2018 H3N2 outbreak in Ontario, Canada, provides a powerful case study. A combination of aggressive testing of suspected cases, comprehensive contact tracing and testing, and strict 28-day isolation of infected dogs effectively halted transmission in each of the five identified epidemiological clusters [11]. This demonstrates that even in an immunologically naive population, rapid diagnosis and stringent quarantine protocols can be highly effective.
Antiviral therapy serves as a critical adjunct to vaccination and biosecurity. While oseltamivir is commonly used, alternative agents are under investigation. Nitazoxanide (NTZ) and its active metabolite, tizoxanide (TIZ), have shown potent in vitro activity against CIV, with 50% inhibitory concentrations (IC50) ranging from 0.17 to 0.21 μM [57]. The mechanism of action involves the mild uncoupling of oxidative phosphorylation, leading to a reversible, dose-dependent decrease in cellular ATP levels, which is essential for viral replication [58]. This broad-spectrum antiviral activity makes NTZ a promising candidate for treating canine influenza, though further in vivo efficacy and toxicity studies are required [57].
Beyond synthetic drugs, natural compounds are being explored for their antiviral properties. Aloe vera extract and its constituents (quercetin, catechin hydrate, kaempferol) have been shown to inhibit influenza A virus replication by blocking autophagy and targeting the M2 protein [59]. Similarly, extracts from Vigna radiata (mung bean) interfere with multiple stages of the viral life cycle, including attachment, penetration, assembly, and release [60]. While these compounds are not yet standard of care, they represent a growing arsenal of potential therapeutic options.
Public Health Implications and Zoonotic Risk Assessment
The most pressing public health concern regarding H3N2 CIV is its zoonotic potential. Dogs are considered potential "mixing vessels" for influenza A viruses, as their respiratory tract expresses both avian-type (SAα2,3-Gal) and human-type (SAα2,6-Gal) sialic acid receptors [13, 18]. This dual receptor tropism creates a permissive environment for co-infection and reassortment. Indeed, co-infection studies in canine tracheal explants have demonstrated that H3N2 CIV can readily reassort with human pandemic H1N1 and avian H9N2 viruses, generating viable reassortants capable of replicating in human alveolar epithelial cells and inducing a heightened proinflammatory cytokine response [7]. This substantiates the risk that dogs could serve as an intermediate host for the emergence of novel influenza viruses with pandemic potential.
The adaptive evolution of H3N2 CIV in dogs has further narrowed the species barrier. Over a decade of circulation, the virus has acquired key mammalian-adaptive mutations. Notably, the PB2 I714S mutation, located in the nuclear localization signal (NLS) area, has been shown to enhance nuclear import efficiency and RNP complex assembly, thereby increasing polymerase activity and pathogenicity in mammalian cells and mice [4]. More alarmingly, systematic analyses have revealed that contemporary H3N2 CIVs have acquired the ability to recognize human-like SAα2,6-Gal receptors, demonstrate increased hemagglutination (HA) acid stability, and exhibit enhanced replication in human airway epithelial cells [19]. In a ferret model, which is the gold standard for influenza transmission studies, these viruses achieved a 100% transmission rate via respiratory droplets [19]. Critically, human populations lack pre-existing immunity to these CIVs, and immunity derived from seasonal human H3N2 viruses does not provide cross-protection [19].
The risk is not merely theoretical. The isolation of a novel reassortant H3N6 CIV from dogs in China, which possesses genes from both H3N2 CIV and H5N6 avian influenza virus, demonstrates the ongoing genetic plasticity of these viruses [5]. Furthermore, the detection of a reassortant H3N2 virus in a cat, containing an NS gene from a human H3N2 virus, indicates that the "mixing vessel" concept extends beyond dogs to other companion animals, such as cats [9]. The World Health Organization (WHO) and the World Organisation for Animal Health (WOAH) have emphasized the need for enhanced surveillance at the human-animal interface. The data strongly suggest that H3N2 CIV poses a credible and escalating zoonotic threat, warranting continuous, coordinated global surveillance and risk assessment to preempt a potential pandemic.
References
[1] Voorhees IEH, Dalziel B, Glaser A, Dubovi E, Murcia P, Newbury SP, et al.. Multiple Incursions and Recurrent Epidemic Fade-Out of H3N2 Canine Influenza A Virus in the United States. Journal of Virology. 2018. DOI: https://doi.org/10.1128/JVI.00323-18
[2] Voorhees IEH, Glaser A, Toohey-Kurth K, Newbury SP, Dalziel B, Dubovi E, et al.. Spread of Canine Influenza A(H3N2) Virus, United States. Emerging Infectious Diseases. 2017. DOI: https://doi.org/10.3201/eid2312.170246
[3] Wasik B, Damodaran L, Maltepes MA, Voorhees IEH, Leutenegger CM, Newbury SP, et al.. The evolution and epidemiology of H3N2 canine influenza virus after 20 years in dogs. bioRxiv. 2024. DOI: https://doi.org/10.1017/S0950268825000251
[4] Li X, Jia T, Wang K, Wang L, Zhou L, Li M, et al.. The PB2 I714S mutation influenced mammalian adaptation of the H3N2 canine influenza virus by interfering with nuclear import efficiency and RNP complex assembly. Emerging Microbes and Infections. 2024. DOI: https://doi.org/10.1080/22221751.2024.2387439
[5] Meng B, Li H, Feng C, Guo W, Feng Y, Zhu D, et al.. Emergence of a novel reassortant H3N6 canine influenza virus. Frontiers in Microbiology. 2023. DOI: https://doi.org/10.3389/fmicb.2023.1186869
[6] Guo F, Roy A, Wang R, Yang J, Zhang Z, Luo W, et al.. Host Adaptive Evolution of Avian-Origin H3N2 Canine Influenza Virus. Frontiers in Microbiology. 2021. DOI: https://doi.org/10.3389/fmicb.2021.655228
[7] Kam T, Bui CHT, Yeung H, Wong P, Chin AWH, Nicholls J, et al.. Viral characterization of the reassortants between canine influenza H3N2 and human pandemic (2009) H1N1 and avian H9N2 viruses in canine ex vivo tracheal explants. Virology Journal. 2025. DOI: https://doi.org/10.1186/s12985-025-02836-1
[8] Xu W, Weese J, Ojkić D, Lung O, Handel K, Berhane Y. Phylogenetic Inference of H3N2 Canine Influenza A Outbreak in Ontario, Canada in 2018. Scientific Reports. 2020. DOI: https://doi.org/10.1038/s41598-020-63278-z
[9] Zhao J, He W, Lu M, He H, Lai A. Emergence and Characterization of a Novel Reassortant Canine Influenza Virus Isolated from Cats. Pathogens. 2021. DOI: https://doi.org/10.3390/pathogens10101320
[10] Li S, Shi Z, Jiao P, Zhang G, Zhong Z, Tian W, et al.. Avian-origin H3N2 canine influenza A viruses in Southern China. Infection, Genetics and Evolution. 2010. DOI: https://doi.org/10.1016/j.meegid.2010.08.010
[11] Weese J, Anderson MEC, Berhane Y, Doyle KF, Leutenegger C, Chan RY, et al.. Emergence and Containment of Canine Influenza Virus A(H3N2), Ontario, Canada, 2017–2018. Emerging Infectious Diseases. 2019. DOI: https://doi.org/10.3201/eid2510.190196
[12] He W, Li G, Wang R, Shi W, Li K, Wang S, et al.. Host-range shift of H3N8 canine influenza virus: a phylodynamic analysis of its origin and adaptation from equine to canine host. Veterinary Research. 2019. DOI: https://doi.org/10.1186/s13567-019-0707-2
[13] Bui CHT, Yeung H, Ho J, Leung C, Hui KPY, Perera RA, et al.. Tropism of SARS-CoV-2, SARS-CoV, and Influenza Virus in Canine Tissue Explants. Journal of Infectious Diseases. 2021. DOI: https://doi.org/10.1093/infdis/jiab002
[14] He W, Li G, Zhu H, Shi W, Wang R, Zhang C, et al.. Emergence and adaptation of H3N2 canine influenza virus from avian influenza virus: An overlooked role of dogs in interspecies transmission. Transboundary and Emerging Diseases. 2019. DOI: https://doi.org/10.1111/tbed.13093
[15] Li G, Wang R, Zhang C, Wang S, He W, Zhang J, et al.. Genetic and evolutionary analysis of emerging H3N2 canine influenza virus. Emerging Microbes and Infections. 2018. DOI: https://doi.org/10.1038/s41426-018-0079-0
[16] Zhou P, Cao Z, Zeng W, Hao X, Zheng Q, Lin X, et al.. PB2 E627K or D701N substitution does not change the virulence of canine influenza virus H3N2 in mice and dogs.. Veterinary Microbiology. 2018. DOI: https://doi.org/10.1016/j.vetmic.2018.05.004
[17] Lin Y, Xie X, Zhao Y, Kalhoro D, Lu C, Liu Y. Enhanced replication of avian-origin H3N2 canine influenza virus in eggs, cell cultures and mice by a two-amino acid insertion in neuraminidase stalk. Veterinary Research. 2016. DOI: https://doi.org/10.1186/s13567-016-0337-x
[18] Abdelwhab EM, Mettenleiter T. Zoonotic Animal Influenza Virus and Potential Mixing Vessel Hosts. Viruses. 2023. DOI: https://doi.org/10.3390/v15040980
[19] Chen M, Lyu Y, Wu F, Zhang Y, Li H, Wang R, et al.. Increased public health threat of avian-origin H3N2 influenza virus caused by its evolution in dogs. eLife. 2023. DOI: https://doi.org/10.7554/eLife.83470
[20] Yao Q, Mai W, Lian Y, Zhang M, Yao Q, Huang C, et al.. Emergence and Evolution of Novel Canine-Avian Reassortant H3N2 Influenza A Viruses in Duck in Leizhou Peninsula, China. Frontiers in Microbiology. 2022. DOI: https://doi.org/10.3389/fmicb.2022.857800
[21] Borland S, Gracieux P, Jones M, Mallet F, Yugueros-Marcos J. Influenza A Virus Infection in Cats and Dogs: A Literature Review in the Light of the “One Health” Concept. Frontiers in Public Health. 2020. DOI: https://doi.org/10.3389/fpubh.2020.00083
[22] Liu L, Song S, Shen Y, Ma C, Wang T, Tong Q, et al.. Truncation of PA-X Contributes to Virulence and Transmission of H3N8 and H3N2 Canine Influenza Viruses in Dogs. Journal of Virology. 2020. DOI: https://doi.org/10.1128/JVI.00949-20
[23] Dalziel B, Huang K, Geoghegan J, Arinaminpathy N, Dubovi E, Grenfell B, et al.. Contact Heterogeneity, Rather Than Transmission Efficiency, Limits the Emergence and Spread of Canine Influenza Virus. PLoS Pathogens. 2014. DOI: https://doi.org/10.1371/journal.ppat.1004455
[24] Parrish C, Voorhees IEH. H3N8 and H3N2 Canine Influenza Viruses: Understanding These New Viruses in Dogs.. The Veterinary clinics of North America. Small animal practice. 2019. DOI: https://doi.org/10.1016/j.cvsm.2019.02.005
[25] Zhou P, Chen B, Hu X, Xiao X, Liu R, Li S. Domestic poultry are not susceptible to avian-origin H3N2 subtype canine influenza A virus.. Veterinary Microbiology. 2022. DOI: https://doi.org/10.1016/j.vetmic.2022.109501
[26] Yamanaka T, Nemoto M, Bannai H, Tsujimura K, Kondo T, Matsumura T, et al.. No evidence of horizontal infection in horses kept in close contact with dogs experimentally infected with canine influenza A virus (H3N8). Acta Veterinaria Scandinavica. 2012. DOI: https://doi.org/10.1186/1751-0147-54-25
[27] Pulit-Penaloza J, Simpson N, Yang H, Creager HM, Jones J, Carney P, et al.. Assessment of Molecular, Antigenic, and Pathological Features of Canine Influenza A(H3N2) Viruses That Emerged in the United States. Journal of Infectious Diseases. 2017. DOI: https://doi.org/10.1093/infdis/jiw620
[28] Wu M, Su R, Gu Y, Yu Y, Li S, Sun H, et al.. Molecular Characteristics, Antigenicity, Pathogenicity, and Zoonotic Potential of a H3N2 Canine Influenza Virus Currently Circulating in South China. Frontiers in Microbiology. 2021. DOI: https://doi.org/10.3389/fmicb.2021.628979
[29] Lu G, Ou J, Cai S, Lai Z, Zhong L, Yin X, et al.. Canine Interferon-Inducible Transmembrane Protein Is a Host Restriction Factor That Potently Inhibits Replication of Emerging Canine Influenza Virus. Frontiers in Immunology. 2021. DOI: https://doi.org/10.3389/fimmu.2021.710705
[30] Xie X, Pang M, Liang S, Lin Y, Zhao Y, Qiu D, et al.. Cellular microRNAs influence replication of H3N2 canine influenza virus in infected cells.. Veterinary Microbiology. 2021. DOI: https://doi.org/10.1016/j.vetmic.2021.109083
[31] Tangwangvivat R, Chaiyawong S, Nonthabenjawan N, Charoenkul K, Janethanakit T, Udom K, et al.. Transmission and pathogenicity of canine H3N2 influenza virus in dog and guinea pig models. Virology Journal. 2022. DOI: https://doi.org/10.1186/s12985-022-01888-x
[32] Fu C, Ye S, Liu Y, Li S. Role of CARD Region of MDA5 Gene in Canine Influenza Virus Infection. Viruses. 2020. DOI: https://doi.org/10.3390/v12030307
[33] Zhou P, Tu L, Lin X, Hao X, Zheng Q, Zeng W, et al.. cfa-miR-143 Promotes Apoptosis via the p53 Pathway in Canine Influenza Virus H3N2-Infected Cells. Viruses. 2017. DOI: https://doi.org/10.3390/v9120360
[34] Zheng Y, Hao X, Zheng Q, Lin X, Zhang X, Zeng W, et al.. Canine Influenza Virus is Mildly Restricted by Canine Tetherin Protein. Viruses. 2018. DOI: https://doi.org/10.3390/v10100565
[35] Abente EJ, Anderson T, Rajão D, Swenson S, Gauger P, Vincent A. The avian‐origin H3N2 canine influenza virus that recently emerged in the United States has limited replication in swine. Influenza and Other Respiratory Viruses. 2016. DOI: https://doi.org/10.1111/irv.12395
[36] Su S, Tian J, Hong M, Zhou P, Lu G, Zhu H, et al.. Global and quantitative proteomic analysis of dogs infected by avian-like H3N2 canine influenza virus. Frontiers in Microbiology. 2015. DOI: https://doi.org/10.3389/fmicb.2015.00228
[37] Song D, Lee C, Kang B, Jung K, Oh T, Kim H, et al.. Experimental Infection of Dogs with Avian-Origin Canine Influenza A Virus (H3N2). Emerging Infectious Diseases. 2009. DOI: https://doi.org/10.3201/eid1501.080755
[38] Ou J, Zheng F, Cheng J, Ye S, Ye C, Jia K, et al.. Isolation and Genetic Characterization of Emerging H3N2 Canine Influenza Virus in Guangdong Province, Southern China, 2018–2021. Frontiers in Veterinary Science. 2022. DOI: https://doi.org/10.3389/fvets.2022.810855
[39] Lyu Y, Song S, Zhou L, Bing G, Wang Q, Sun H, et al.. Canine Influenza Virus A(H3N2) Clade with Antigenic Variation, China, 2016–2017. Emerging Infectious Diseases. 2019. DOI: https://doi.org/10.3201/eid2501.171878
[40] Classe HM, Dant JC, Mogler M, Stachura K, Lafleur R, Xu Z, et al.. Efficacy and Safety in Dogs Following Administration of an Alphavirus RNA Particle Canine Influenza H3N2 Vaccine. Vaccines. 2024. DOI: https://doi.org/10.3390/vaccines12101138
[41] Luo J, Lu G, Ye S, Ou J, Fu C, Zhang X, et al.. Comparative pathogenesis of H3N2 canine influenza virus in beagle dogs challenged by intranasal and intratracheal inoculation.. Virus Research. 2018. DOI: https://doi.org/10.1016/j.virusres.2018.05.023
[42] Secrest S, Sharma A. THORACIC RADIOGRAPHIC CHARACTERISTICS OF CANINE INFLUENZA VIRUS IN SIX DOGS. Veterinary Radiology & Ultrasound. 2016. DOI: https://doi.org/10.1111/vru.12391
[43] Yondo A, Kalantari A, Fernandez-Marrero I, McKinney A, Naikare HK, Velayudhan B. Predominance of Canine Parainfluenza Virus and Mycoplasma in Canine Infectious Respiratory Disease Complex in Dogs. Pathogens. 2023. DOI: https://doi.org/10.3390/pathogens12111356
[44] Aghaei F, Moradi M, Karimi A. Punicalagin inhibits pro-inflammatory cytokines induced by influenza A virus. European Journal of Integrative Medicine. 2021. DOI: https://doi.org/10.1016/J.EUJIM.2021.101324
[45] Zhou H, Li R, Wang Y, Shen L, Cai L, Weng Y, et al.. Total alkaloids from Alstonia scholaris inhibit influenza a virus replication and lung immunopathology by regulating the innate immune response.. Phytomedicine. 2020. DOI: https://doi.org/10.1016/j.phymed.2020.153272
[46] Omoniwa D, Chinyere C, Agusi E, Mkpuma N, Oyetunde J, Igah O, et al.. Serological and molecular investigation of canine influenza virus in Plateau State, Nigeria. Sokoto Journal of Veterinary Sciences. 2022. DOI: https://doi.org/10.4314/sokjvs.v20i3.8
[47] Li Y, Zhang X, Liu Y, Feng Y, Wang T, Ge Y, et al.. Characterization of Canine Influenza Virus A (H3N2) Circulating in Dogs in China from 2016 to 2018. Viruses. 2021. DOI: https://doi.org/10.3390/v13112279
[48] Dao DT, Coleman KK, Bui VN, Bui AN, Tran L, Nguyen QD, et al.. High Prevalence of Highly Pathogenic Avian Influenza: A Virus in Vietnam's Live Bird Markets. Open Forum Infectious Diseases. 2024. DOI: https://doi.org/10.1093/ofid/ofae355
[49] Winter JM, Anchor C, Gehrt S, Landolfi JA, Allender M. Serosurvey for Canine Influenza in Coyotes (Canis latrans) in Illinois, USA, 2000–23. Journal of Wildlife Diseases. 2024. DOI: https://doi.org/10.7589/JWD-D-23-00196
[50] Yehia N, Rabie N, Adel A, mossad Z, Nagshabandi M, Alharbi M, et al.. Differential replication characteristic of reassortant avian influenza A viruses H5N8 clade 2.3.4.4b in Madin-Darby canine kidney cell. Poultry Science. 2023. DOI: https://doi.org/10.1016/j.psj.2023.102685
[51] Zinnecker T, Badri N, Araújo D, Thiele K, Reichl U, Genzel Y. From single‐cell cloning to high‐yield influenza virus production – implementing advanced technologies in vaccine process development. Engineering in Life Sciences. 2024. DOI: https://doi.org/10.1002/elsc.202300245
[52] Menard JA, Zinnecker T, Godbout E, Roberts J, Arulanandam R, Chen A, et al.. Lipidomic profiling of influenza A virus production in MDCK cells towards targeted clone selection. Scientific Reports. 2026. DOI: https://doi.org/10.1038/s41598-025-33499-1
[53] Sen S, Rahman S, Nag M, Rahman MM, Sarker R, Kabir S. Prevalence of canine parvo virus and canine influenza virus infection in dogs in Dhaka, Mymensingh, Feni and Chittagong districts of Bangladesh. Asian Journal of Medical and Biological Research. 2016. DOI: https://doi.org/10.3329/AJMBR.V2I1.27579
[54] Kim H, Yang D, Seo B, Cho I. Serosurvey of rabies virus, canine distemper virus, parvovirus, and influenza virus in military working dogs in Korea. Journal of Veterinary Medical Science. 2018. DOI: https://doi.org/10.1292/jvms.18-0012
[55] Yang Z, Yu S, Xu Y, Zhao Y, Li L, Sun J, et al.. The Screening and Mechanism of Influenza-Virus Sensitive MDCK Cell Lines for Influenza Vaccine Production. Diseases. 2024. DOI: https://doi.org/10.3390/diseases12010020
[56] Choi I, Kang A, Ahn M, Jun H, Baek S, Park J, et al.. Insertion‐responsive microneedles for rapid intradermal delivery of canine influenza vaccine. Journal of Controlled Release. 2018. DOI: https://doi.org/10.1016/j.jconrel.2018.08.017
[57] Ashton L, Callan RL, Rao S, Landolt G. In Vitro Susceptibility of Canine Influenza A (H3N8) Virus to Nitazoxanide and Tizoxanide. Veterinary Medicine International. 2010. DOI: https://doi.org/10.4061/2010/891010
[58] Rossignol J, Tijsma A, Baalen CVv. Mechanism of Antiviral Activity of Nitazoxanide against the Influenza Virus: Effect of Tizoxanide on AdenosineTriphosphate in Influenza-virus Infected Madin Darby Canine Kidney Cells. bioRxiv. 2021. DOI: https://doi.org/10.1101/2021.07.30.454324
[59] Choi J, Lee H, Kim Y, Hwang Y, Oh Y, Lee B, et al.. Aloe vera and its Components Inhibit Influenza A Virus-Induced Autophagy and Replication.. The American Journal of Chinese Medicine. 2019. DOI: https://doi.org/10.1142/S0192415X19500678
[60] Lo C, Pi C, Chen Y, Chen H. Vigna radiata (L.) R. Wilczek Extract Inhibits Influenza A Virus by Targeting Viral Attachment, Penetration, Assembly, and Release. Frontiers in Pharmacology. 2020. DOI: https://doi.org/10.3389/fphar.2020.584973