Canine Influenza H3N2: Dog Flu Reference

Overview and Taxonomy of Canine Influenza H3N2: Origins and Classification

Taxonomic Classification and Virological Context

Canine influenza virus (CIV) H3N2 is a member of the genus Alphainfluenzavirus within the family Orthomyxoviridae, classified as Influenza A virus (IAV) based on the antigenic properties of its nucleoprotein (NP) and matrix (M) proteins [19, 22]. The influenza A virus classification system further subtypes viruses according to the antigenic characteristics of the two major surface glycoproteins: hemagglutinin (HA) and neuraminidase (NA). The H3N2 subtype designation indicates that this virus possesses the H3 hemagglutinin and the N2 neuraminidase, a combination that has demonstrated remarkable capacity for interspecies transmission and sustained circulation in mammalian hosts [6, 15, 22]. Unlike the H3N8 CIV, which originated from an equine influenza virus spillover event into greyhounds in the United States around 1999 [17], the H3N2 CIV represents a direct host-switching event from an avian reservoir, making it a unique model for studying the evolutionary dynamics of influenza A virus adaptation to a novel mammalian host [8, 10, 15].

The taxonomic distinction between H3N2 CIV and other influenza A viruses is critical for understanding its pathogenic potential and epidemiological behavior. As a type A influenza virus, H3N2 CIV possesses a segmented, single-stranded, negative-sense RNA genome comprising eight gene segments that encode for at least 11 proteins [19]. This segmented genome architecture is of paramount importance because it facilitates genetic reassortment, the exchange of entire gene segments between different influenza A viruses co-infecting the same host cell [2, 20]. The capacity for reassortment has been experimentally demonstrated in dogs naturally co-infected with H3N2 CIV and pandemic H1N1 (pH1N1) virus, yielding 23 distinct reassortant viruses. Notably, these reassortants exhibited a dominance of the M gene from pH1N1 and the HA gene from canine H3N2, with some reassortants displaying heightened pathogenicity in mammalian models [2]. This genetic plasticity underscores the taxonomic significance of H3N2 CIV as a potential platform for the emergence of novel influenza viruses with pandemic potential, a concern recognized by global health authorities including the World Health Organization (WHO) and the World Organisation for Animal Health (WOAH).

Avian Origins and Initial Emergence

The H3N2 canine influenza virus originated from an avian influenza A virus reservoir, with the host-switching event estimated to have occurred around 2004 in Asia [8, 10, 15]. The first documented isolation of H3N2 CIV occurred in South Korea in 2007 from dogs exhibiting severe respiratory syndrome, with the virus subsequently identified in China during the same period [2, 6, 16]. Phylogenetic analyses have consistently demonstrated that the ancestral virus was an avian-origin H3N2 influenza A virus, likely derived from wild waterfowl or domestic poultry populations [10, 12, 15]. The initial spillover event was not an isolated occurrence but rather represented a successful host-switch that enabled sustained dog-to-dog transmission, a critical distinction from dead-end spillover infections that do not establish onward transmission chains [8, 16].

Experimental infections conducted shortly after the initial isolation confirmed that the avian-origin H3N2 virus could replicate efficiently in dogs, cause clinical disease, and transmit horizontally through direct contact. Song et al. (2009) demonstrated that susceptible dogs co-housed with experimentally infected animals developed elevated rectal temperatures, virus shedding, seroconversion, and severe necrotizing tracheobronchitis and bronchioalveolitis [16]. This experimental confirmation of transmissibility was foundational for establishing H3N2 CIV as a bona fide canine respiratory pathogen rather than an incidental spillover event. The virus's ability to establish sustained transmission cycles in dogs represents a significant evolutionary leap, as it required the virus to overcome multiple host barriers, including differences in receptor binding specificity, intracellular replication efficiency, and innate immune evasion mechanisms [10, 12].

Phylogenetic Structure and Clade Classification

Comprehensive phylogenetic analyses of H3N2 CIV genomes have revealed a structured evolutionary history characterized by distinct clades and genotypes that reflect both geographic segregation and temporal evolution. Li et al. (2018) identified three major evolutionary clades through timescaled phylogenetic tree reconstruction and principal component analysis: the Origin clade, the China clade, and the Korea/USA clade [12]. This tripartite structure reflects the initial emergence of the virus in Asia, its subsequent differentiation into geographically distinct lineages, and the eventual transcontinental spread to North America. The Korea/USA clade is particularly significant because it encompasses the viruses that were introduced into the United States in 2015, demonstrating the direct phylogenetic linkage between Asian and North American H3N2 CIV populations [12, 14].

More recent genotyping efforts have refined our understanding of H3N2 CIV diversity. Liu et al. (2024) identified 15 distinct genotypes among all available H3N2 CIV sequences, with genotype 15 becoming predominant among dogs in China since approximately 2017, indicating the establishment of a stable virus lineage [5]. Ge et al. (2024) further characterized a novel clade, designated clade 5.1, which emerged after 2019 and has become prevalent in China, demonstrating continued viral evolution and adaptation [4]. This clade classification is not merely a taxonomic exercise; it has direct implications for vaccine strain selection, diagnostic assay design, and risk assessment for zoonotic potential. The emergence of new clades and genotypes signals ongoing adaptation to the canine host and potentially altered biological properties, including changes in receptor binding affinity, antigenicity, and pathogenicity [4, 5, 10].

Genotypic Diversity and Reassortment Events

The genotypic landscape of H3N2 CIV is characterized by considerable complexity, driven by both gradual accumulation of point mutations and punctuated reassortment events. The segmented genome of influenza A viruses allows for the generation of novel genotypes through reassortment when a host is co-infected with two or more distinct influenza viruses [2, 20]. In the case of H3N2 CIV, multiple reassortment events have been documented, involving gene segments from human seasonal H3N2 viruses, pandemic H1N1 (2009) viruses, and other avian influenza viruses [2, 3]. Wen et al. (2025) reported the isolation of a novel canine/human reassortant H3N2 virus from golden monkeys at a zoo in China, in which the PB1 gene segment originated from a human H3N2 strain while the remaining seven segments were derived from early local canine H3N2 virus [3]. This reassortant virus exhibited enhanced virulence in mice, attributed to a unique cytotoxic motif (68I, 69L, and 70V) in the PB1-F2 protein [3]. Such findings highlight the potential for reassortment to generate viruses with altered pathogenicity and expanded host range.

The genotypic diversity of H3N2 CIV is further compounded by the introduction of gene segments from the pandemic H1N1 (2009) virus. Na et al. (2015) demonstrated that in naturally co-infected dogs, the M gene of pH1N1 and the HA gene of canine H3N2 were predominant in the resulting reassortants, suggesting a selective advantage for certain gene segment combinations [2]. Some of these reassortants displayed high pathogenicity in mice, raising concerns about the emergence of novel influenza viruses with pandemic potential from canine populations [2]. The genotypic complexity of H3N2 CIV necessitates continuous molecular surveillance, as recommended by the WHO and WOAH, to monitor for the emergence of reassortant viruses that may pose threats to both animal and human health.

Geographic Distribution and Lineage Dynamics

The geographic distribution of H3N2 CIV has expanded significantly since its initial emergence in Asia, with the virus now circulating endemically in China, having died out in South Korea around 2017, and causing recurrent outbreaks in the United States following its introduction in 2015 [8, 14]. The epidemiological dynamics of H3N2 CIV in North America are particularly instructive for understanding how an emerging respiratory pathogen spreads within a structured host population. Wasik et al. (2024) demonstrated that H3N2 CIV transmission in the United States is dominated by fast-moving outbreaks within dense populations such as animal shelters and kennels, with the virus dying out locally after susceptible animals become infected and immune [8]. Sustained circulation at the continental scale requires repeated reintroduction from Asia, with 2–3 introductions into North America documented in the past three years alone [8].

Voorhees et al. (2018) provided a detailed analysis of the recurrent epidemic burst–fade-out dynamics of H3N2 CIV in the United States, demonstrating that the virus's basic reproductive number (R₀) ranges between 1 and 1.5 across nearly all outbreaks [14]. This relatively low R₀, combined with widespread host contact heterogeneity, means that the virus cannot sustain transmission through a general dog population and instead relies on metapopulations of high host density for continued circulation [14]. The epidemiological model suggests that H3N2 CIV may have reached an evolutionary cul-de-sac in the United States, where there is limited opportunity for selection of more transmissible phenotypes [14]. This pattern of recurrent fade-out and reintroduction contrasts with the endemic situation in China, where the virus has established sustained circulation in the dog population [8, 9].

Molecular Markers of Host Adaptation

The transition of H3N2 CIV from an avian host to sustained circulation in dogs has been accompanied by the accumulation of specific amino acid substitutions that reflect adaptation to the mammalian host environment. Guo et al. (2021) conducted a comprehensive analysis of host adaptive evolution and identified 54 amino acid substitutions that have become fixed in the H3N2 CIV population, with 11 of these sites also showing high prevalence in H3N8 CIV, indicating convergent evolution across different canine influenza lineages [10]. Among the most significant adaptive substitutions are HA-G146S, M1-V15I, NS1-E227K, PA-C241Y, PB2-K251R, and PB2-G590S, all of which have been previously implicated in facilitating transmission and spillover of influenza A viruses across species barriers [10]. Notably, most of these substitutions became fixed around 2015, coinciding with the spread of H3N2 CIV from Asia to North America [10].

The hemagglutinin (HA) protein has been a particular focus of adaptation studies, as it mediates receptor binding and is a major determinant of host range. Liu et al. (2024) identified a newly emerged adaptive mutation, HA-V223I, which is predominantly found in human and swine H3N2 viruses and appears to enhance mammalian adaptation [5]. This substitution decreases the virus's affinity for human-type α-2,6 sialic acid receptors while enhancing thermal stability, suggesting a trade-off between receptor binding specificity and virion stability [5]. Peng et al. (2025) further characterized the receptor-binding properties of contemporary H3N2 CIV isolates, demonstrating strong affinity for avian-type α-2,3 receptors and limited binding to human-type α-2,6 receptors, indicating that while the virus has adapted to dogs, it has not fully acquired the receptor specificity required for efficient human-to-human transmission [1]. However, the presence of both avian and human-type sialic acid receptors in the canine respiratory tract creates opportunities for further adaptation and reassortment [7, 18].

Antigenic Classification and Implications for Vaccination

The antigenic properties of H3N2 CIV have evolved significantly since the virus's emergence, with implications for vaccine efficacy and serological surveillance. Antigenic analysis using hemagglutination inhibition (HI) assays has revealed that contemporary H3N2 CIV strains are antigenically distinct from both human seasonal H3N2 viruses and avian H3N2 viruses [1, 5]. Peng et al. (2025) demonstrated that human antiserum showed only partial cross-reactivity with canine H3N2 virus, while no reactivity was detected with avian strains, indicating that the canine virus has acquired unique antigenic characteristics through its evolution in dogs [1]. These antigenic changes are driven by amino acid substitutions in antigenic sites of the HA protein, as well as altered glycosylation patterns that can mask or expose epitopes [1, 21].

The antigenic divergence of H3N2 CIV from human influenza viruses has important implications for public health. Liu et al. (2024) demonstrated that current human H3N2 vaccines do not confer protection against H3N2 CIVs, meaning that human populations remain immunologically naive to canine-origin H3N2 viruses [5]. This lack of cross-protection, combined with the close contact between humans and dogs, creates a scenario in which H3N2 CIV could potentially emerge as a zoonotic pathogen [5, 20]. The Centers for Disease Control and Prevention (CDC) and WHO have recognized the importance of monitoring influenza viruses at the human-animal interface, and the antigenic characterization of H3N2 CIV is a critical component of pandemic preparedness efforts. The development of effective canine influenza vaccines, including virus-like particle (VLP) vaccines, live recombinant adenovirus-vectored vaccines, and inactivated vaccines, relies on the selection of antigenically appropriate strains that match circulating viruses [11, 13].

Molecular Pathogenesis of Canine H3N2: Reassortment, Receptor Binding, and Host Adaptation

The emergence and sustained circulation of the avian-origin H3N2 canine influenza virus (CIV) within dog populations represents a paradigm of rapid viral evolution and host adaptation. Unlike the equine-origin H3N8 CIV, the H3N2 subtype has demonstrated a remarkable capacity for genetic plasticity, driven by its segmented genome, which facilitates reassortment, and by intense selective pressures operating at the interface of receptor specificity and host cellular machinery. The molecular pathogenesis of H3N2 CIV is therefore a multifaceted narrative of reassortment events that reshuffle gene constellations, alterations in receptor binding that determine tissue tropism and host range, and a suite of adaptive mutations that refine viral fitness within the canine host and, critically, modulate the potential for spillover to other mammals, including humans [6, 15, 22].

Reassortment: The Engine of Genetic Novelty

The segmented nature of the influenza A virus genome makes reassortment a potent driver of evolutionary change, and H3N2 CIV has been a prolific participant in this genetic exchange. One of the most consequential natural reassortment events documented involved the co-infection of a dog with avian-origin H3N2 CIV and the pandemic H1N1 (pH1N1) virus, which is itself a reassortant of avian, swine, and human lineages [2]. Analysis of 23 reassortant viruses arising from this single co-infection revealed a clear pattern of viral dominance: the M (matrix) gene from pH1N1 and the HA gene from canine H3N2 were preferentially retained in the majority of progeny viruses [2]. This selective pressure for the pH1N1 M segment is particularly significant, as the M gene encodes the M1 matrix protein and the M2 ion channel, both critical for viral assembly, budding, and uncoating. The acquisition of the pH1N1 M segment by CIV H3N2 was not merely a passive genomic shuffle; it conferred a tangible phenotypic advantage. Intriguingly, some of these reassortants demonstrated increased pathogenicity in a mouse model compared to the parental CIV H3N2 strain, underscoring the capacity of reassortment to generate viruses with heightened virulence and altered host range without prior adaptation [2].

Further evidence of the power of reassortment to bridge species barriers came from a fatal outbreak in zoo-housed golden monkeys in China in 2022 [3]. A novel H3N2 reassortant, designated A/golden monkey/Jiangsu/1/2022 (Gm-1), was isolated from the lungs of deceased animals. Genomic sequencing revealed that while seven of its eight gene segments originated from an early local canine H3N2 virus (A/canine/Jiangsu/06/2010, JS06), the PB1 segment was derived from a human influenza A virus, sharing over 97% nucleotide identity [3]. This precise reassortment event, swapping a single polymerase gene, was directly linked to increased virulence. Reverse genetics experiments, in which the human-origin PB1 was placed into the canine background, demonstrated that the resulting virus (rGm-1) caused significantly more severe lung pathology and elevated levels of proinflammatory cytokines in infected mice compared to the parental canine virus (rJS06) [3]. Mechanistically, this enhanced virulence was attributed to a unique cytotoxic motif (68I, 69L, 70V) within the PB1-F2 protein of the human-origin PB1 segment, a motif absent in the canine strain [3]. This finding highlights how the acquisition of even a single gene segment from a human-adapted virus can dramatically amplify the pathogenic potential of a canine virus, representing a clear public health concern for both non-human primates and, by extension, humans.

The structural consequences of reassortment extend to viral morphology itself. Studies comparing the Chinese mutant JS10, which possesses a two-amino-acid insertion in the NA stalk, and the Korean strain KR07, demonstrated that swapping the M segment from a pandemic H1N1 strain (CA04) into the KR07 background alone (creating KR07M) was sufficient to induce a morphological change towards more filamentous particles and a higher efficiency of cellular uptake [26]. This demonstrates that reassortment can influence fundamental viral properties like shape and entry kinetics, which are intimately linked to pathogenesis and transmission. The ongoing circulation of H3N2 CIV in enzootic areas like China [8-10] provides a continuous, dynamic genetic pool from which such potentially dangerous reassortants can emerge, particularly as dogs are known to be susceptible to infection with human seasonal H1N1 and H3N2 viruses, setting the stage for future co-infection and reassortment events [18, 25, 27].

Receptor Binding: The Molecular Gatekeeper of Host Tropism

The initial and critical barrier to cross-species transmission of influenza A virus is the specificity of the viral hemagglutinin (HA) for sialic acid (SA) receptors on the surface of host cells. Avian influenza viruses generally bind preferentially to α-2,3-linked SA receptors, which are abundant in the avian intestinal tract. In contrast, human-adapted influenza viruses have a strong affinity for α-2,6-linked SA receptors, which predominate in the human upper respiratory tract. The canine respiratory tract harbors both α-2,3- and α-2,6-linked SA receptors [7, 18], a dual-receptor landscape that theoretically provides a "mixing vessel" environment permissive to both avian and mammalian viruses. The molecular pathogenesis of H3N2 CIV is therefore critically defined by its evolving receptor-binding properties.

As a virus of recent avian origin, early H3N2 CIV isolates retain a strong binding affinity for avian-type α-2,3 receptors. However, more recent isolates have demonstrated a widening of their receptor specificity, showing an ability to bind, albeit with varying degrees, to human-type α-2,6 receptors as well [1, 5]. This dual-receptor binding capacity is a hallmark of pandemic potential. Research by Liu et al. (2024) systematically characterized four H3N2 CIVs isolated from dogs in China between 2018 and 2020 [5]. Receptor-binding analysis confirmed that these contemporary viruses could bind to both avian and human-type receptors. Crucially, attachment assays demonstrated that the viruses could bind to human tracheal tissues, directly confirming their ability to interact with the relevant anatomical substrate for human infection [5].

The evolutionary trajectory of this receptor specificity is fine-tuned by specific amino acid substitutions in the HA1 subunit. Guo et al. (2021) identified the HA-G146S substitution as one of several key adaptive changes that became fixed in the CIV population around 2015, a period that coincided with the spread of the virus from Asia to North America [10]. This substitution is known to alter receptor binding properties. Perhaps more compelling is the discovery of the HA-V223I substitution. While initially identified as a mammalian-adaptive mutation predominantly found in human and swine H3N2 viruses, when introduced into the canine H3N2 background, the V223I substitution paradoxically decreased the virus's binding affinity for human-type α-2,6 receptors [5]. However, this same substitution significantly enhanced the thermal stability of the virus [5]. This finding suggests a trade-off: the adaptive landscape for CIV in dogs may prioritize viral stability and transmission efficiency within the canine host, even at the cost of reducing its inherent affinity for human receptors. This provides a molecular explanation for why, despite sustained circulation and ample opportunity for human exposure, H3N2 CIV has not yet caused a widespread human pandemic. The virus's "zoonotic potential" remains high, but its immediate human-transmission potential appears to be actively modulated by ongoing host-adaptive epistasis. A recently isolated strain from Jiangsu, China, in 2025, further confirmed this pattern, showing strong affinity for avian-type α-2,3 receptors but only limited binding to human-type α-2,6 receptors, alongside altered glycosylation patterns that contributed to a distinct antigenic profile compared to human and avian H3N2 viruses [1].

Host Adaptation: A Genomic Signature of Establishment and Refinement

The transition of H3N2 CIV from an avian reservoir to a new mammalian host, the dog, has imposed unique selective forces that have left a distinct genomic signature. A comprehensive analysis of the evolution of H3N2 CIV over its 20 years in dogs revealed that the virus has evolved at a constant rate, but with a significantly higher ratio of nonsynonymous (amino-acid-changing) to synonymous (silent) substitutions (dN/dS) compared to its avian ancestors [10, 14]. This elevated dN/dS ratio is a hallmark of positive selection and rapid adaptive evolution, as the virus responds to the pressures of replicating and transmitting within a novel host environment [10].

Guo et al. (2021) systematically catalogued 54 amino acid substitutions that have become fixed in the global H3N2 CIV population [10]. Many of these are in known functional domains. Among the most significant are those in the polymerase genes (PB2 and PA), which are critical for replication efficiency. The presence of mammalian-adaptive PB2 mutations like PB2-I292T, PB2-G590S, and PB2-S714I has been documented in contemporary CIVs [5, 10]. The PB2-G590S mutation, in particular, is a well-characterized marker of mammalian adaptation that, along with PB2-E627K (which is notably absent in most canine H3N2 strains despite its prevalence in other mammalian-adapted viruses), enhances polymerase activity in mammalian cells. Intriguingly, experimental introduction of PB2-E627K or PB2-D701N into a CIV H3N2 backbone increased polymerase activity in vitro but did not alter virulence in mice or dogs [24]. This suggests that the path to enhanced virulence in canines involves a complex polygenic interaction, rather than reliance on a single, powerful mutation like E627K, which is common in human H1N1 and H5N1 viruses. This nuanced adaptation highlights the concept of "host-specific" adaptive solutions.

Convergent evolution is also strikingly apparent. Eleven of the 54 fixed substitutions identified in H3N2 CIV were also found at high prevalence in the independently evolved H3N8 CIV lineage [10]. This convergence at sites within HA (G146S), M1 (V15I), NS1 (E227K), PA (C241Y), and PB2 (K251R, G590S) strongly implies that these are universal solutions to the challenges of replicating and transmitting in a canine host, regardless of the viral subtype [10]. This pattern suggests a limited number of optimal adaptive pathways for influenza A virus adaptation to dogs, which is a critical insight for risk assessment and vaccine design.

Beyond protein-coding changes, host adaptation has also involved genetic "fine-tuning" at the level of codon usage and dinucleotide composition. The H3N2 CIV genome has shown a steady increase in its codon adaptation index (CAI) towards the codon usage bias of the dog, indicating a refinement of translation efficiency for optimal replication in canine cells [10, 12]. Moreover, an analysis of relative dinucleotide abundance has revealed a significant reduction in CpG motifs in the CIV genome compared to the ancestral avian virus [10]. This depletion of CpG dinucleotides is a known strategy for evading the host's innate immune system, particularly the zinc-finger antiviral protein (ZAP), which targets RNA sequences with high CpG content. This is a sophisticated example of molecular pathogenesis operating at the level of host-mediated RNA degradation. The downstream consequences of these adaptations are dramatic; phosphoproteomic analysis of H3N2-infected dog lungs revealed that infection causes profound changes in host protein phosphorylation, affecting over 1,200 proteins and 3,000 modification sites, thereby hijacking critical signaling networks involved in cellular metabolism, immune response, and cytoskeletal organization [23]. This molecular hijacking by a now well-adapted canine virus underscores the complex interplay between viral genetic determinants and host cellular pathways that defines its pathogenesis.

Epidemiology and Global Distribution of Canine Influenza H3N2

The epidemiological landscape of canine influenza A virus (CIV) subtype H3N2 represents a compelling narrative of cross-species viral emergence, sustained enzootic circulation, and complex metapopulation dynamics. Unlike the H3N8 CIV, which arose from an equine reservoir and was first recognized in racing greyhounds in the United States around 1999 [17], the H3N2 CIV originated from an avian influenza A virus reservoir, making its initial jump into domestic dogs in Asia approximately two decades ago [8, 10, 15]. This host-switching event, which likely occurred around 2004-2005, has since given rise to a pathogen that now circulates enzootically in specific geographic regions while exhibiting a pattern of recurrent epidemic fade-out and reintroduction in others [8, 14]. The global distribution of H3N2 CIV is not uniform; rather, it is characterized by distinct epidemiological phases, regional endemicity, and a dynamic interplay between viral evolution, host population structure, and anthropogenic factors such as international animal movement.

Emergence and Early Enzootic Establishment in Asia

The first recognized isolation of avian-origin H3N2 CIV occurred in South Korea in 2007, with retrospective analyses tracing the virus's initial emergence in dogs to approximately 2005 in both China and Korea [2, 8, 12]. Experimental infections conducted shortly after this discovery confirmed that the virus was highly transmissible among dogs, with direct contact leading to rapid spread, severe respiratory pathology, and seroconversion in naive animals [16]. This early work established that the virus was not merely a spillover event but had acquired the capacity for sustained dog-to-dog transmission, a prerequisite for enzootic establishment. Following its initial detection, H3N2 CIV spread rapidly through canine populations in South Korea and China, becoming endemic in these regions [7, 10, 30]. Phylogenetic analyses of early isolates revealed the formation of distinct evolutionary clades, with principal component analysis identifying three major lineages: an ancestral "Origin" clade, a "China" clade, and a "Korea/USA" clade [12]. This genetic diversification occurred relatively quickly after the host switch, driven by a combination of mutation pressure, natural selection, and alterations in dinucleotide abundance that facilitated adaptation to the canine host [10, 12]. Critically, the virus experienced significantly greater selection pressure in dogs than in its avian reservoir, evidenced by extremely high global non-synonymous to synonymous substitution ratios (dN/dS) across all gene segments [10]. This adaptive evolution led to the fixation of at least 54 amino acid substitutions in the circulating CIV population, including several known mammalian adaptive markers such as HA-G146S, PB2-G590S, and PB2-K251R, many of which became fixed around 2015 [10]. The establishment of these adaptive mutations likely facilitated the virus's subsequent geographic expansion.

The North American Incursion and Recurrent Epidemic Dynamics

A pivotal event in the global epidemiology of H3N2 CIV occurred in early 2015, when the virus was first detected in the Chicago, Illinois area of the United States, having been introduced from South Korea [8, 14]. This incursion triggered a large-scale outbreak that spread across multiple states, representing the first known introduction of H3N2 CIV into North America [9, 14, 29]. The epidemiological behavior of the virus in the United States has been markedly different from its enzootic circulation in Asia. Comprehensive genomic and surveillance analyses have revealed that H3N2 CIV circulation in the U.S. is characterized by recurrent epidemic burst–fade-out dynamics [14]. The virus spreads rapidly through dense, high-contact populations such as animal shelters and boarding kennels, where the basic reproductive number (R₀) has been estimated to range between 1.0 and 1.5 [14]. However, after exhausting the pool of susceptible individuals in these localized populations, the virus tends to die out regionally, requiring reintroduction from external sources to spark new outbreaks [8, 14]. This pattern is driven by the host population structure of dogs in the U.S., which is highly heterogeneous; the general pet dog population has relatively low contact rates compared to the concentrated populations in shelters, making it difficult for the virus to sustain transmission chains without repeated introductions [14]. Indeed, the virus has been introduced into North America from Asia on multiple occasions, with at least two or three separate introductions occurring in the three years preceding 2024 [8]. These introductions have seeded outbreaks in different regions, but sustained circulation across the entire continent has not been achieved, and the virus has repeatedly undergone epidemic extinction in the U.S. [8, 14]. The 2018 outbreak in New York City, which prompted a rapid public health response involving a web-based reporting platform and interactive dashboard for veterinarians, exemplifies this pattern of localized, high-intensity outbreaks that require active surveillance to monitor [32].

Shifting Endemicity: The Current Status in Asia

While the United States experiences intermittent outbreaks, the epidemiological picture in Asia has evolved significantly. South Korea, which was one of the initial epicenters of H3N2 CIV, appears to have experienced viral fade-out in its dog population around 2017 [8]. The reasons for this extinction are not entirely clear but may relate to the development of herd immunity following widespread exposure, the implementation of control measures, or a combination of factors. In contrast, China has emerged as the primary enzootic reservoir for H3N2 CIV, with the virus continuing to circulate persistently in dog populations across multiple provinces [4, 8-10]. Phylogenetic analyses indicate that the viral lineages currently circulating in China have been the source of the most recent introductions into North America, underscoring China's role as a global reservoir for this pathogen [8]. Within China, the virus has continued to evolve, with new genotypes emerging and becoming dominant. A comprehensive analysis of all available H3N2 CIV sequences identified 15 distinct genotypes, with genotype 15 becoming predominant in dogs since approximately 2017, indicating the establishment of a stable and successful viral lineage [5]. Furthermore, a new clade, designated 5.1, emerged after 2019 and has become prevalent in China, demonstrating ongoing adaptation and diversification [4]. Surveillance efforts in Guangdong Province, southern China, between 2018 and 2021, isolated six emerging H3N2 CIV strains that formed a novel phylogenetic group closely related to strains from the United States and northern China, confirming the interconnectedness of these global viral populations [9, 29]. The continuous circulation in China provides a persistent source of viral genetic diversity, from which new variants can emerge and potentially spread internationally.

Host Range Expansion and Interspecies Transmission Events

A critical dimension of the epidemiology of H3N2 CIV is its expanding host range and demonstrated capacity for interspecies transmission, which raises significant concerns from a One Health perspective [20]. While dogs are the primary reservoir, the virus has been shown to infect other mammalian species, both naturally and experimentally. Serological evidence has documented H3N2 CIV infection in horses with a history of dog exposure in China, with 2.2% of racehorses tested showing positive antibodies, and a significant correlation between seropositivity and contact with dogs [7]. This finding is particularly noteworthy given that both SA-α-2,3-Gal and SA-α-2,6-Gal receptors are present in the respiratory tracts of horses, providing a potential anatomical basis for infection [7]. More alarmingly, a novel reassortant H3N2 virus, containing a human-origin PB1 gene segment, was isolated from zoo-housed golden monkeys (Rhinopithecus roxellana) in Jiangsu, China, in 2022, resulting in fatal respiratory disease [3]. This event represents the first confirmed influenza A virus infection in this endangered primate species and demonstrates the potential for CIV to acquire genetic elements from human influenza viruses, thereby enhancing its virulence in new hosts [3]. The human-origin PB1 segment conferred a significant virulence advantage, leading to more severe lung pathology and elevated proinflammatory cytokine levels in experimentally infected mice [3]. Experimental studies have also demonstrated that H3N2 CIV can infect and transmit among guinea pigs, a common mammalian model, with both direct contact and aerosol exposure leading to infection, fever, and viral shedding [28]. Furthermore, cats have been shown to be highly susceptible to H3N2 CIV infection, exhibiting sustained nasal viral shedding and viral replication in the respiratory tract, even in the absence of overt clinical signs [31]. This susceptibility positions cats as a potential bridging host in the influenza transmission network [31]. The detection of influenza A virus genetic material in 9.01% of nasopharyngeal swabs from dogs in Kazakhstan, with serological evidence of H3N2-specific antibodies in 9.44% of sampled animals, further underscores the widespread and ongoing nature of CIV circulation, even in regions where it has not been previously well-documented [25]. The ability of dogs to be co-infected with multiple influenza A virus subtypes, including human pandemic H1N1 and avian strains, creates a potential "mixing vessel" scenario where genetic reassortment can generate novel viruses with unpredictable pathogenicity and host range [2, 26]. Indeed, naturally occurring co-infection has already been documented, leading to the generation of 23 distinct reassortant viruses, some of which exhibited increased pathogenicity in mice [2]. The presence of both avian-type (α-2,3) and human-type (α-2,6) sialic acid receptors in the canine respiratory tract provides the molecular foundation for this dual susceptibility [5, 18].

Clinical Manifestations and Pathology in Dogs and Other Hosts

Clinical Spectrum in Domestic Dogs

Canine influenza H3N2 infection in dogs presents a remarkably broad spectrum of clinical disease, ranging from subclinical infection, where animals harbor and shed the virus without overt signs, to severe, life-threatening pneumonia [20, 22]. The disease is fundamentally an acute respiratory syndrome, with hallmark clinical signs emerging approximately 1–3 days post-exposure. The most frequently documented constellation of signs includes a persistent, productive cough, serous to mucopurulent nasal discharge, sneezing, ocular discharge, lethargy, and anorexia [16, 38, 39]. Fever, defined as a rectal temperature exceeding 39.5°C, is a prominent feature in symptomatic animals, with a geometric mean temperature of 39.86°C ± 0.49 reported in experimentally infected contact dogs [39]. Critically, the severity of clinical signs is directly correlated with viral shedding; dogs exhibiting fever shed significantly higher viral titers (mean 2.99 log EID50/mL) compared to afebrile animals, underscoring the role of pyrexia as a marker of high viral burden and transmission potential [39].

The upper respiratory tract is the primary site of viral replication. Initial replication occurs in the epithelium of the nasal cavity, trachea, and bronchi, leading to necrotizing tracheobronchitis [16]. The virus demonstrates a strong tropism for the lower respiratory tract, and in severe cases, infection progresses to bronchoalveolitis, interstitial pneumonia, and even fatal hemorrhagic pneumonia [3, 4]. The severity of the clinical picture is profoundly influenced by the challenge dose, host immune status, and the presence of concurrent infections. In a dose-dependent model using beagles, experimental inoculation with 10^6 EID50 induced the most pronounced clinical signs, including overt dyspnea, marked pyrexia, and profound lethargy, alongside the highest viral titers and most severe pulmonary pathological changes [4]. Conversely, lower inoculum doses (10^4 and 10^5 EID50) tended to produce mild, self-limiting disease characterized by transient coughing and mild serous nasal discharge, with minimal to no significant lung pathology [4]. Furthermore, the route of inoculation significantly modulates disease expression; experimental data reveals that intratracheal challenge consistently induces more severe clinical signs, higher fevers, and more extensive lung consolidation compared to intranasal administration, suggesting that direct delivery of virus to the lower airways bypasses key innate immune defenses of the upper tract [36].

A crucial factor in clinical outcome is the phenomenon of co-infection. Canine infectious respiratory disease (CIRD) is a polymicrobial syndrome, and pathogenic synergism between viruses (e.g., canine parainfluenza virus, canine adenovirus) and bacteria (e.g., Bordetella bronchiseptica, Mycoplasma cynos, Streptococcus equi subsp. zooepidemicus) dramatically exacerbates respiratory compromise [38]. In such complex infections, the initial viral insult impairs mucociliary clearance and suppresses local immune responses, facilitating secondary bacterial invasion, which can lead to suppurative bronchopneumonia and a protracted, more severe disease course [6, 38]. In immunocompromised dogs, such as those receiving immunosuppressive glucocorticoid therapy (e.g., prednisolone at 3.0 mg/kg), the clinical course is markedly different. These animals experience a prolonged duration of viral shedding (13 days vs. 8 days in immunocompetent controls) and a higher peak viral titer (5.5 log EID50 vs. 4.6 log EID50), highlighting the critical role of host immune competence in controlling viral replication and limiting transmission [34].

Gross and Histopathological Findings in Dogs

The pathological signature of H3N2 CIV infection in dogs is a severe, acute, necrotizing inflammation of the respiratory tract. On gross examination at necropsy, the lungs exhibit multifocal to coalescing areas of consolidation, characterized by a dark red, congested, and firm texture, predominantly affecting the cranioventral and middle lung lobes [16, 28]. These consolidated areas are sharply demarcated from adjacent normal, aerated lung parenchyma. Cut surfaces often exude a frothy, serosanguinous fluid, indicative of pulmonary edema. The trachea and bronchi contain variable amounts of mucopurulent exudate, and the mucosa is erythematous and edematous [16].

Histopathologically, the most consistent and severe lesions are found in the bronchi and bronchioles. There is extensive necrosis and sloughing of the epithelial lining, with infiltration of neutrophils, lymphocytes, and plasma cells into the lamina propria and submucosa [16, 28]. The bronchiolar lumina are often filled with a mixture of necrotic debris, inflammatory cells, and fibrin. In the alveolar spaces, the septa are thickened due to congestion, edema, and infiltration of mononuclear cells and neutrophils, resulting in a pattern of acute interstitial pneumonia [4, 36].

Immunofluorescence staining of lung tissue from infected beagles confirms that the virus is predominantly localized within the epithelial cells of the bronchi, bronchioles, and type II pneumocytes in the alveoli, validating the lung as the primary target organ for viral replication [4]. More advanced molecular analyses, such as phosphoproteomic profiling of infected dog lungs, have revealed that H3N2 CIV infection triggers dramatic changes in host protein phosphorylation, with 1,235 differentially expressed phosphorylated proteins identified at 5 days post-infection [23]. These alterations implicate key signaling pathways involved in inflammation, apoptosis, and cytoskeletal rearrangement, providing a molecular basis for the profound tissue damage observed [23]. Similarly, comparative microRNA expression analysis demonstrates that CIV infection alters the expression of numerous microRNAs involved in regulating the innate immune response and metabolic pathways, such as glycerolipid and glycerophospholipid metabolism. These changes are postulated to contribute to the clinical signs of anorexia, weight loss, and the suppression of antiviral immunity [33].

Hematological and Immunological Disturbances

Beyond the respiratory tract, H3N2 CIV infection can induce significant hematological derangements. Experimental infection of beagles has demonstrated the development of transient, self-resolving immune thrombocytopenia (ITP), evidenced by decreased platelet counts and increased platelet-associated immunoglobulin G (PAIgG) levels [35]. In one study, a proportion of infected dogs became both thrombocytopenic and positive for PAIg by flow cytometry, a finding that temporally correlated with the acute phase of infection [35]. The mechanism is hypothesized to involve viral-induced immune dysregulation, leading to the formation of anti-platelet autoantibodies, a phenomenon also observed in human influenza infections. This finding is clinically relevant, as it highlights that thrombocytopenia, while typically mild and transient, can be a feature of the acute phase of H3N2 CIV and may complicate diagnostic and therapeutic decision-making in affected dogs [35].

Interspecies Transmission and Pathology in Alternative Hosts

The H3N2 CIV is not strictly confined to dogs. Its capacity to spill over into other mammalian species is a significant concern from both a veterinary and a One Health perspective.

Cats

Domestic cats are highly susceptible to H3N2 CIV. Experimental inoculation studies demonstrate that while cats may not consistently exhibit overt clinical signs, they are highly permissive for infection. Infected cats show seroconversion by 7 days post-inoculation and sustain nasal viral shedding for approximately one week, with viral replication confirmed in the lungs, trachea, and nasal turbinates by immunohistochemistry and virus isolation [31]. Critically, comparative studies suggest that cats are more susceptible than dogs to the H3 subtype of influenza viruses, including H3N2 CIV, H3N8, and H3N2 swine influenza viruses. Dogs in these comparative studies, while seroconverting, did not shed virus or show evidence of viral replication in tissues, whereas cats did [31]. This increased susceptibility positions cats as potentially important incidental hosts and sentinel animals for influenza A virus circulation in domestic environments.

Guinea Pigs and Ferrets

The guinea pig model has been instrumental in studying interspecies transmission. Intranasal inoculation of guinea pigs with Thai CIV-H3N2 results in fever (evident at 1–2 dpi), active viral shedding in respiratory secretions, and seroconversion. Significantly, transmission occurs to both direct contact and aerosol-exposed guinea pigs, confirming the virus's capacity for airborne spread in a non-canine host [28]. Pathological changes in guinea pig lungs are milder than in dogs, primarily characterized by mild histopathological changes such as peribronchiolar lymphocytic infiltration [28]. This model highlights that the virus can establish productive infection and onward transmission in select mammalian species without requiring severe pathology. Ferrets, a gold-standard model for influenza pathogenesis and transmission, have also been shown to be susceptible to mutated H3N2 CIV variants, particularly those from China and Korea, which exhibited increased virulence and signs of severe respiratory disease in these models [26].

Golden Monkeys (Non-Human Primates)

A particularly alarming demonstration of the zoonotic potential of H3N2 CIV comes from a 2022 outbreak in zoo-housed golden monkeys (Rhinopithecus roxellana). This unprecedented event, caused by a novel canine/human reassortant H3N2 virus (Gm-1), resulted in classic flu-like symptoms in seven animals, with two succumbing to fatal respiratory distress [3]. Histopathological analysis of the deceased monkeys lungs revealed severe pulmonary consolidation and inflammation consistent with acute influenza pneumonia. Immunostaining confirmed widespread viral antigen in the lung parenchyma. Notably, the human-origin PB1 segment of this reassortant virus was shown, through reverse genetics studies in mice, to confer a significant virulence advantage. When mice were infected with rGm-1 (containing the human PB1), they developed more severe lung pathology and elevated levels of proinflammatory cytokines compared to those infected with the canine-origin parent virus [3]. The presence of a unique cytotoxic motif (68I, 69L, 70V) in the PB1-F2 protein of the reassortant is hypothesized to drive this increased pathogenicity [3]. This case is a stark reminder that CIV can not only infect but also cause fatal disease in non-human primates, emphasizing the potential for the virus to adapt to species genetically close to humans.

Rodent Models (Mice)

Mice are widely used to study CIV pathogenesis and to dissect the role of specific viral genes. Standard canine H3N2 isolates generally cause mild to moderate disease in mice, with weight loss and lung lesions, but are often non-lethal [13]. However, the generation of reassortant viruses, a naturally occurring event, can profoundly alter this pathogenicity. For instance, some reassortant viruses between canine H3N2 and human pandemic H1N1 (2009), created through natural co-infection, display significantly enhanced virulence in mice compared to the parental canine virus [2]. These reassortants, which often contain the M gene from pH1N1 and the HA gene from canine H3N2, induce more severe weight loss, higher lung viral titers, and more extensive pulmonary histopathology [2]. This demonstrates that genetic reassortment, even within the same host, can produce variants with unpredictable and heightened virulence in a new host. Furthermore, specific mutations, such as PB2-E627K or D701N, while enhancing polymerase activity in vitro, do not inherently alter virulence in mice or dogs, suggesting that other genetic or host factors are rate-limiting for pathogenesis in vivo [24].

Horses and Other Species

Serological evidence indicates that horses in close contact with dogs, such as those in shared riding club environments, can become infected with CIV H3N2, with a seroprevalence of 2.2% detected in a Chinese study [7]. The presence of both SA-α-2,3-Gal and SA-α-2,6-Gal receptors in the equine respiratory tract, similar to dogs, provides the anatomical substrate for this cross-species transmission [7]. While clinical disease in these seropositive horses was not reported, the finding underscores the potential for CIV to circulate within a wider range of domestic animals. Conversely, surveillance in free-ranging canids (coyotes, foxes) in the United States has failed to detect antibodies to CIV H3N2, suggesting that the virus is currently not established in those wild populations despite their interface with domestic dogs [37].

Diagnostic Approaches for Canine H3N2: Molecular, Serological, and Virological Methods

The accurate and timely diagnosis of canine influenza A virus (CIV) subtype H3N2 is a cornerstone of effective outbreak management, epidemiological surveillance, and fundamental research into viral pathogenesis and evolution. Given the virus's capacity for rapid transmission within dense dog populations, such as those found in shelters, kennels, and boarding facilities, and its documented potential for interspecies transmission, a multi-faceted diagnostic strategy is essential [8, 14, 20]. This strategy must integrate molecular techniques for direct pathogen detection, serological assays for retrospective analysis and immune status assessment, and classical virological methods for virus isolation and characterization. The selection of an appropriate diagnostic modality is dictated by the clinical context, the stage of infection, the specific objectives of the investigation (e.g., acute diagnosis vs. seroprevalence survey), and the available laboratory infrastructure. This section provides an exhaustive analysis of the molecular, serological, and virological methods employed in the diagnosis and study of H3N2 CIV, drawing upon the latest research to detail their mechanisms, applications, limitations, and interpretive nuances.

Molecular Diagnostics: The Cornerstone of Active Surveillance and Acute Diagnosis

Molecular methods, particularly reverse transcription-polymerase chain reaction (RT-PCR), have become the gold standard for the direct detection of H3N2 CIV RNA in clinical specimens. These techniques offer unparalleled sensitivity and specificity, enabling the identification of viral genetic material even in samples with low viral loads or during the early stages of infection before a robust antibody response has developed. The primary targets for these assays are highly conserved regions of the influenza A virus genome, most commonly the matrix (M) gene, which is shared across all influenza A subtypes [38, 41, 43]. A positive result for the influenza A M gene is then followed by subtype-specific assays targeting the hemagglutinin (HA) and neuraminidase (NA) genes to confirm the H3N2 subtype [25, 38].

Real-Time RT-PCR (qRT-PCR) is the most widely deployed molecular tool. Its quantitative nature allows for the estimation of viral RNA copy number, which can be correlated with viral shedding levels and, consequently, infectiousness. Studies have demonstrated that viral shedding in experimentally infected dogs can be detected as early as 1 day post-inoculation (dpi) and can persist for several days, with peak titers often coinciding with the onset of fever and clinical signs [28, 39]. The duration and magnitude of shedding can be influenced by host factors; for instance, immunocompromised dogs treated with glucocorticoids have been shown to exhibit prolonged and higher-magnitude viral shedding compared to immunocompetent controls, a critical consideration for clinical management and biosecurity in veterinary settings [34]. The high sensitivity of qRT-PCR makes it the method of choice for screening nasal, pharyngeal, or oropharyngeal swabs from dogs presenting with acute respiratory illness, particularly during outbreaks [9, 32]. The World Health Organization (WHO) and the World Organisation for Animal Health (WOAH) recognize RT-PCR as a primary diagnostic tool for influenza surveillance in both human and animal populations.

The development of multiplex RT-PCR assays has further enhanced diagnostic efficiency. These assays can simultaneously detect and differentiate between multiple respiratory pathogens that contribute to the canine infectious respiratory disease (CIRD) complex. For example, a single reaction can be designed to target influenza A (including H3N2 and H3N8), canine distemper virus, canine adenovirus, canine parainfluenza virus, and bacterial agents like Bordetella bronchiseptica and Mycoplasma cynos [38]. This syndromic approach is invaluable for differential diagnosis, as the clinical signs of CIV infection, cough, nasal discharge, fever, and lethargy, are indistinguishable from those caused by other CIRD agents. Understanding the etiological agent is crucial for implementing appropriate treatment and control measures, as the use of antibiotics, for instance, would be ineffective against a primary viral infection.

The utility of molecular diagnostics extends beyond clinical diagnosis to include genomic surveillance and evolutionary studies. Full-genome sequencing of RT-PCR-positive samples allows researchers to track the emergence of new viral clades, identify mutations associated with host adaptation or antigenic drift, and monitor for reassortment events. For instance, phylogenetic analyses of H3N2 CIV sequences have revealed the emergence of distinct clades (e.g., clade 5.1 in China) and documented multiple independent introductions of the virus from Asia into North America, where it exhibits recurrent epidemic fade-out and reintroduction dynamics [4, 8, 14]. Furthermore, sequencing has identified specific amino acid substitutions, such as HA-V223I, which reduces binding affinity to human-type receptors but enhances thermal stability, and PB2-E627K, which enhances polymerase activity in mammalian cells, highlighting the ongoing adaptation of this avian-origin virus to its canine host [5, 24]. The use of synthetic analytical reference materials (ARMs) , which contain complete sequences of key diagnostic targets (HA, NA, M, NP, NS), has been validated as a safe and reliable BSL-1 alternative to live virus for the development and quality control of these molecular assays, ensuring consistent performance across laboratories [40].

Serological Diagnostics: Uncovering Historical Exposure and Population Immunity

Serological assays detect antibodies produced by the host's immune system in response to H3N2 CIV infection or vaccination. These methods are indispensable for seroprevalence studies, vaccine efficacy trials, and retrospective investigations of outbreaks, particularly when the acute phase of infection has passed and viral shedding has ceased. The most commonly employed serological techniques are the Hemagglutination Inhibition (HI) assay and the Enzyme-Linked Immunosorbent Assay (ELISA).

The Hemagglutination Inhibition (HI) assay is considered the gold standard for subtype-specific serology. It is based on the principle that the influenza virus's HA protein can agglutinate red blood cells (RBCs). Specific antibodies in a serum sample will bind to the HA protein, preventing this agglutination. The HI titer is the highest dilution of serum that completely inhibits hemagglutination. HI assays are highly specific for the HA subtype and are used extensively to measure antibody responses following natural infection or vaccination [7, 11, 13, 28]. For H3N2 CIV, HI titers ≥ 1:20 or ≥ 1:40 are often used as a threshold for seropositivity, indicating prior exposure [7, 27]. This assay has been instrumental in documenting the seroprevalence of H3N2 CIV in various populations, including pet dogs in Ohio (2.4%) [27], horses with dog exposure in China (2.2%) [7], and dogs in Kazakhstan (9.44%) [25]. The HI assay is also critical for antigenic characterization, revealing, for example, that human H3N2 vaccines do not confer protection against circulating H3N2 CIVs, underscoring the distinct antigenic nature of the canine virus [5]. However, the HI assay has limitations, including the need for species-specific RBCs, the removal of non-specific inhibitors from sera, and its inability to detect antibodies against the neuraminidase (NA) protein.

The Enzyme-Linked Immunosorbent Assay (ELISA) offers a more high-throughput and often less technically demanding alternative. Most commercial and in-house ELISAs for influenza A target the nucleoprotein (NP), which is highly conserved across all subtypes. A positive NP-ELISA result indicates past or present infection with influenza A, but it does not provide subtype information [27, 41]. Subtype-specific ELISAs, using recombinant HA or NA proteins, have been developed but are less standardized than HI assays. While ELISAs are excellent for large-scale screening, studies have shown that they can lack sensitivity compared to the HI assay. For example, one study found that a commercial NP-ELISA failed to detect antibodies in many HI-positive canine sera, highlighting the need for validated, species-specific serological tests for dogs [27]. The microneutralization (MN) assay is another functional serological test that measures the ability of antibodies to neutralize viral infectivity in cell culture. It is often more sensitive than the HI assay and can detect antibodies against both HA and NA, providing a more comprehensive measure of the functional antibody response [7, 44].

The interpretation of serological results requires careful consideration of the infection timeline. Antibodies typically become detectable by HI or MN assays around 7-10 days post-infection, peaking at 2-3 weeks [28, 31]. Therefore, a single negative serological result does not rule out acute infection. A four-fold or greater rise in antibody titer between paired acute and convalescent sera (collected 2-4 weeks apart) is considered definitive evidence of recent infection. Serology is also a powerful tool for investigating cross-species transmission. For instance, serological surveys have provided the first evidence of H3N2 CIV infection in horses [7] and have demonstrated that cats are more susceptible than dogs to H3 subtype influenza viruses, showing sustained seroconversion and viral shedding [31].

Virological Methods: Virus Isolation and Characterization

Virus isolation remains the definitive method for confirming an active H3N2 CIV infection and is essential for obtaining live virus for detailed biological characterization, such as antigenic analysis, receptor-binding studies, and pathogenicity testing. The process involves inoculating clinical specimens (e.g., nasal swab eluates, lung tissue homogenates) into a permissive culture system.

The embryonated chicken egg has been the traditional gold standard for influenza virus isolation. The virus is inoculated into the allantoic cavity of 9- to 11-day-old specific-pathogen-free (SPF) embryonated chicken eggs. After 2-3 days of incubation, the allantoic fluid is harvested and tested for the presence of virus, typically by a hemagglutination assay using chicken or guinea pig RBCs. This method has been used successfully to isolate numerous H3N2 CIV strains from field samples [1, 9, 16]. However, the egg-based system has drawbacks, including the potential for the virus to acquire adaptive mutations (e.g., in the HA gene) that alter its antigenicity and receptor-binding properties, a phenomenon well-documented for human influenza viruses [21].

Madin-Darby Canine Kidney (MDCK) cells are the most widely used mammalian cell line for the isolation of canine influenza viruses. MDCK cells are highly permissive to influenza A virus infection and support robust viral replication, often producing higher yields and more genetically authentic virus than eggs [4, 42]. The use of a qualified MDCK cell line (qMDCK) has been shown to achieve high initial isolation efficiencies (over 70%) from clinical specimens, with the resulting isolates maintaining antigenicity equivalent to circulating viruses [42]. For H3N2 CIV isolation, MDCK cells are inoculated with the clinical sample and monitored daily for the development of a cytopathic effect (CPE), characterized by cell rounding, detachment, and syncytia formation [4]. The presence of virus is confirmed by hemagglutination assay, immunofluorescence (IF) staining using subtype-specific antibodies, or RT-PCR [4, 36]. The choice between egg and cell culture can influence the characteristics of the isolated virus, particularly its glycosylation patterns. MDCK-derived HAs tend to have more high-molecular-weight glycans and a higher proportion of partially occupied glycosylation sites compared to egg-derived HAs, which can affect receptor binding and immunogenicity [21].

Once isolated, the virus can be subjected to a battery of characterization assays. Hemagglutination Inhibition (HI) with a panel of reference antisera is used for antigenic typing and to detect antigenic drift [1]. Receptor-binding assays, such as the solid-phase binding assay using sialylglycopolymers, determine the virus's preference for α-2,3 (avian-type) or α-2,6 (human-type) sialic acid receptors. H3N2 CIVs have been shown to bind strongly to avian-type receptors, with some strains also exhibiting limited binding to human-type receptors, a property that is modulated by specific HA mutations like V223I [1, 5]. Transmission electron microscopy (TEM) can be used to visualize viral morphology and size, revealing pleomorphic particles that are typically spherical or filamentous [4, 26]. In vivo pathogenicity studies in animal models (e.g., beagles, mice, guinea pigs, ferrets) are the ultimate test of virulence, assessing clinical signs, viral replication in tissues, and histopathological changes [24, 26, 28, 36]. These studies have demonstrated that intratracheal inoculation is more virulent than intranasal inoculation in dogs [36] and that reassortant viruses carrying a human-origin PB1 gene can cause more severe pathology in mice [3]. The isolation and characterization of H3N2 CIV are therefore not merely diagnostic steps but are fundamental to understanding the virus's biology, its potential threat to animal and public health, and its ongoing evolution.

Antigenic Evolution and Glycosylation Patterns of H3N2 Canine Influenza Virus

The H3N2 canine influenza virus (CIV) represents a paradigm of host-switch adaptation, evolving from an avian influenza A virus (IAV) reservoir into a fully established mammalian pathogen over the past two decades. Its sustained circulation within dog populations, now spanning over 20 years, has imposed unique selective pressures that have driven profound alterations in its antigenic landscape and glycosylation architecture [8, 10, 12]. These changes are not merely incidental; they are the molecular fingerprints of a virus navigating a novel host environment, shaping its transmissibility, immune evasion capacity, and zoonotic potential. Understanding the intricate interplay between antigenic drift and glycosylation pattern shifts is critical for predicting viral emergence, designing effective vaccines, and informing surveillance strategies recommended by the World Organisation for Animal Health (WOAH) and the Centers for Disease Control and Prevention (CDC).

Molecular Drivers of Antigenic Diversification in the Hemagglutinin Protein

The hemagglutinin (HA) glycoprotein is the primary target of neutralizing antibodies and thus the principal locus of antigenic evolution. Since its avian-to-canine transfer, H3N2 CIV HA has accumulated a substantial number of amino acid substitutions, many of which map to antigenic sites A, B, C, and D, homologous to those defined in human H3N2 viruses. Guo et al. [10] conducted a comprehensive codon-based frequency diagram analysis and identified 54 fixed amino acid substitutions across the H3N2 CIV genome during its circulation in dogs, with a disproportionately high number occurring in HA. Critically, six of these substitutions, including HA-G146S, have been experimentally implicated in facilitating interspecies transmission and spillover events. The G146S substitution, located near the receptor-binding site (RBS), has been shown to alter receptor-binding specificity and enhance viral replication in mammalian cells, representing a key step in canine adaptation [5, 10].

More recent surveillance has revealed the emergence of novel adaptive mutations that continue to refine antigenic properties. Liu et al. [5] identified the HA-V223I substitution in H3N2 CIV isolates from 2018–2020, a mutation predominantly found in human and swine H3N2 viruses. Functional characterization demonstrated that V223I reduces the virus’s binding affinity for human-type α-2,6 sialic acid receptors while paradoxically enhancing HA thermal stability. This finding has profound implications for antigenic evolution: by reducing human-type receptor affinity, the virus may be selectively shaped to maintain canine-specific tropism, yet the increased thermostability could enhance environmental persistence and transmission efficiency among dogs. Furthermore, the presence of this substitution in mammalian-adapted lineages suggests ongoing selective pressure toward improved fitness in the canine host, potentially at the expense of broader zoonotic receptor binding.

Antigenic cartography studies using hemagglutination inhibition (HI) assays have consistently demonstrated that H3N2 CIV strains are antigenically distinct from contemporary human H3N2 viruses. Peng et al. [1] showed that human H3N2 antisera exhibited only partial cross-reactivity with canine isolates, while no reactivity was detected with avian H3N2 antisera. This antigenic divergence is driven by a constellation of amino acid changes at key epitopes, including residues 145, 155, 156, 158, 159, 189, 193, and 225 (H3 numbering), which collectively remodel the antibody-binding surface. Critically, the five amino acid substitutions and altered glycosylation patterns identified by Peng et al. [1] conferred distinct antigenic properties that effectively separate canine H3N2 from both human and avian lineages, a finding corroborated by the work of Ou et al. [9] and Li et al. [12], who demonstrated the emergence of novel phylogenetic clades within the canine H3N2 population.

Glycosylation as a Determinant of Antigenic Masking and Receptor Binding

N-linked glycosylation of the HA globular head is a powerful mechanism for modulating antigenicity, as glycans can sterically shield antibody epitopes from recognition. The H3N2 CIV HA possesses a variable number of potential N-glycosylation sequons (PNGS), and the gain or loss of these sites has been a hallmark of its antigenic evolution. Comparative analysis between H3N2 CIV and its avian progenitor reveals that the canine virus has acquired several lineage-specific glycosylation sites while losing others, a process that directly impacts immune evasion [1, 21]. Li et al. [21] demonstrated that in H3N2 HA, the occupancy of PNGS can vary substantially depending on the production host, chicken embryo versus Madin-Darby canine kidney (MDCK) cells, indicating that glycan processing is influenced by the host cellular machinery. For the H3N2 and B strains, MDCK-derived HAs contained more sites being partially occupied (<95%) than embryo-derived HAs, and MDCK-derived HAs harbored more glycans of higher molecular weight [21]. This host-dependent glycosylation has direct implications for antigenic characterization: viruses passaged in different substrates may exhibit altered antigenic profiles due to differential glycan masking, potentially confounding vaccine strain selection and serological surveillance.

The strategic placement of glycans near the receptor-binding domain profoundly influences receptor-binding specificity. Peng et al. [1] demonstrated that H3N2 CIV isolates from 2025 retained a strong affinity for avian-type α-2,3 sialic acid receptors but exhibited only limited binding to human-type α-2,6 receptors. This receptor-binding profile is consistent with the virus's avian origin and suggests that, despite years of circulation in dogs, the HA has not fully shifted its binding preference. However, the acquisition of mammalian-adaptive substitutions such as HA-G146S and HA-N188D has modulated this binding, allowing the virus to engage both receptor types to varying degrees [5, 10]. The balance between α-2,3 and α-2,6 binding is mediated, in part, by glycosylation patterns around the RBS. For instance, the presence or absence of a glycan at residue 158 (H3 numbering) has been shown to influence the accessibility of the receptor-binding pocket, with glycan loss often correlating with increased binding to human-type receptors in human H3N2 viruses. In canine H3N2, the glycosylation status at this and adjacent positions remains dynamic, and ongoing surveillance is needed to monitor shifts that could enhance zoonotic potential.

Phylogenetic Clade Dynamics and Antigenic Signature Evolution

The antigenic evolution of H3N2 CIV is inextricably linked to its phylogenetic trajectory. Timescaled phylogenetic analyses have delineated at least three major clades: the ancestral origin clade, a China clade, and a Korea/USA clade [12]. More recent work by Ge et al. [4] identified the emergence of a 5.1 clade after 2019, which has become predominant among circulating strains in China. This clade is characterized by a unique constellation of antigenic site mutations and glycosylation pattern shifts, resulting in a distinct antigenic signature that may not be fully recognized by antibodies elicited by older vaccine strains. The continuous emergence of new genotypes, with Liu et al. [5] identifying 15 distinct genotypes among all available H3N2 CIV sequences, genotype 15 having become dominant since approximately 2017, underscores the rapid pace of antigenic diversification.

The global dissemination of H3N2 CIV has followed a pattern of repeated introductions and local extinction. Wasik et al. [8] and Voorhees et al. [14] demonstrated that virus lineages circulating in China have seeded multiple introductions into North America, with 2 or 3 introductions occurring in the past 3 years alone. Each introduction event represents a genetic bottleneck that may select for viruses with specific antigenic and glycosylation profiles. The epidemic burst-fade-out dynamics observed in the United States, where the virus spreads rapidly through high-density populations (kennels, shelters) before dying out locally, may exert selection for variants that are optimally adapted to these transmission contexts [8, 14]. Importantly, epidemiological models indicate a basic reproductive number (R₀) between 1 and 1.5 across U.S. outbreaks, suggesting that while the virus is sufficiently transmissible to cause outbreaks, it lacks the viral fitness to sustain circulation in a general dog population with widespread contact heterogeneity [14]. This ecological constraint may limit the opportunity for antigenic drift to accumulate, but it also means that each reintroduction carries the potential for new antigenic variants to arise.

Convergent Evolution and the Influence of Reassortment on Antigenic Properties

A striking feature of H3N2 CIV evolution is the evidence for convergent evolution with the H3N8 CIV lineage. Guo et al. [10] identified 11 amino acid substitutions that are highly prevalent in both H3N2 and H3N8 CIV, suggesting that common selective pressures, specific to the canine host, drive the fixation of advantageous mutations independently in both lineages. These convergent changes include mutations in HA, M1, NS1, PA, and PB2, many of which have been shown to play roles in host adaptation, immune evasion, and viral replication. The existence of convergent evolution strongly implies that there is a finite set of adaptive solutions for influenza viruses to thrive in dogs, and that antigenic evolution is constrained by these functional requirements.

Reassortment has further shaped the antigenic landscape of H3N2 CIV. Natural co-infection of a dog with pandemic H1N1 (2009) and H3N2 CIV generated 23 distinct reassortant viruses, with the M gene from pH1N1 and the HA gene from H3N2 CIV predominating [2]. While these reassortants retained the antigenic properties of the H3N2 HA, the acquisition of heterologous internal genes, particularly the M segment, can alter viral morphology, cell entry efficiency, and pathogenicity. Na et al. [26] demonstrated that swapping the M segment of pH1N1 into an H3N2 CIV background (creating KR07M) resulted in morphological changes and higher cell uptake efficiency. More concerning, some reassortants exhibited increased pathogenicity in mice compared to wild-type H3N2 CIV [2]. The emergence of a novel canine/human reassortant virus, A/golden monkey/Jiangsu/1/2022, which acquired a human-origin PB1 gene segment while retaining seven segments from canine H3N2, highlights the ongoing potential for reassortment to generate viruses with altered virulence and antigenic properties [3]. Although the HA remained canine in origin, the enhanced pathogenicity conferred by the human PB1 segment, associated with a unique cytotoxic motif (68I, 69L, 70V) in the PB1-F2 protein, could select for variants with increased fitness in mammalian hosts, indirectly influencing the evolutionary trajectory of the HA antigenic domain.

Implications for Public Health Surveillance and Vaccine Design

The antigenic and glycosylation evolution of H3N2 CIV carries direct implications for public health. The CDC and WOAH have emphasized the need for enhanced surveillance of influenza viruses in companion animals, as dogs may serve as intermediate hosts for the emergence of novel pandemic strains. Liu et al. [5] demonstrated that current human H3N2 vaccines do not confer protection against H3N2 CIVs, a finding echoed by the antigenic analyses of Peng et al. [1] confirming poor cross-reactivity. This antigenic gap means that pre-existing human immunity is unlikely to provide a barrier against canine-to-human spillover, should the virus acquire the ability to efficiently bind human-type receptors. The receptor-binding analyses showing limited α-2,6 binding provide some reassurance, but the plasticity of the HA glycoprotein and the dynamic nature of glycosylation patterns mean that the emergence of fully human-adapted variants remains a distinct possibility under sustained selective pressure.

Furthermore, the presence of mammalian-adaptive substitutions (PB2-E627K, PB2-D701N, PB2-G590S, PB2-S714I, PB1-D154G, NP-R293K) in circulating H3N2 CIV strains [5, 10, 24] indicates that the virus is already equipped with molecular signatures associated with increased replication and pathogenicity in mammals. The fact that H3N2 CIV has been shown to infect guinea pigs, ferrets, and cats experimentally, and has serological evidence of infection in horses with dog exposure [7, 28, 31], underscores its cross-species capability. Antigenic monitoring, therefore, must be coupled with glycosylation profiling to identify variants that may escape herd immunity or alter receptor specificity. The development of cell-based vaccine platforms, such as those using MDCK cells, must account for host-dependent glycosylation patterns to ensure that vaccine antigens faithfully represent circulating strains [21, 42]. As the H3N2 CIV continues its evolutionary trajectory, marked by constant rates of evolution, acquisition of adaptive substitutions, and dynamic glycosylation shifts, sustained, globally coordinated surveillance remains the cornerstone of pandemic preparedness.

Interspecies Transmission and Zoonotic Potential: Human and Non-Human Primates

The H3N2 canine influenza virus (CIV), having successfully crossed the avian-mammalian species barrier to establish a stable lineage in dogs, presents a complex and evolving landscape of interspecies transmission risks. The close and often intimate proximity of dogs to humans, coupled with the virus’s demonstrated ability to infect a range of mammalian species, positions H3N2 CIV as a significant pathogen of concern within the “One Health” framework [6, 20]. This section provides an exhaustive analysis of the documented and potential transmission of H3N2 CIV to humans and non-human primates, examining the molecular determinants, epidemiological evidence, and pathobiological consequences of such spillover events.

Molecular Determinants of Host Range and Receptor Specificity

The fundamental barrier to influenza A virus (IAV) cross-species transmission is the specificity of the viral hemagglutinin (HA) protein for sialic acid (SA) receptors on host cells. Avian influenza viruses typically exhibit a preference for SA linked to galactose by an α-2,3 linkage (SAα-2,3-Gal), which are abundant in the avian intestinal tract. In contrast, human influenza viruses preferentially bind SAα-2,6-Gal receptors, which predominate in the human upper respiratory tract. The H3N2 CIV, being of avian origin, initially displayed a strong affinity for avian-type SAα-2,3-Gal receptors [1]. However, sustained circulation in a mammalian host (dogs) has driven significant adaptive evolution in the HA gene.

Molecular characterization of contemporary H3N2 CIV strains has revealed the acquisition of several mammalian-adaptive amino acid substitutions. Notably, substitutions such as HA-G146S, HA-N188D, and the newly emerged HA-V223I have been identified in circulating strains [5, 10]. The HA-V223I substitution is particularly intriguing as it is predominantly found in human and swine H3N2 viruses, suggesting convergent evolution towards mammalian adaptation [5]. Receptor-binding assays have demonstrated that while many H3N2 CIVs retain the ability to bind both avian and human-type receptors, the HA-V223I mutation paradoxically reduces the virus’s affinity for human-type SAα-2,6-Gal receptors while enhancing its thermal stability [5]. This suggests a complex trade-off between receptor binding, viral stability, and host adaptation. Despite this reduced affinity, attachment analyses have confirmed that H3N2 CIV can still bind to human tracheal tissues, albeit with lower efficiency when carrying the V223I mutation [5]. This residual binding capability, combined with the presence of both SAα-2,3-Gal and SAα-2,6-Gal receptors in the canine respiratory tract, creates a permissive environment for the virus to act as a potential “mixing vessel” for reassortment events [7, 18].

Further molecular adaptations extend beyond the HA gene. A comprehensive analysis of H3N2 CIV evolution identified 54 fixed amino acid substitutions, including several known to facilitate IAV spillover, such as PB2-K251R, PB2-G590S, PB1-D154G, and NP-R293K [5, 10]. The PB2 gene is a well-established host range determinant. While the canonical mammalian-adaptive mutations PB2-E627K and PB2-D701N have been shown to enhance polymerase activity in H3N2 CIV, they did not increase virulence in mice or dogs, indicating that the pathobiology of this virus is governed by a more complex polygenic architecture [24]. However, the presence of these and other adaptive markers in the circulating gene pool underscores the ongoing mammalian adaptation of the virus.

Zoonotic Potential and Evidence for Human Infection

Despite the theoretical risk and the presence of molecular markers for mammalian adaptation, there are currently no confirmed, documented cases of natural H3N2 CIV infection in humans [20, 25]. This absence of evidence is a critical point, but it does not equate to evidence of absence. Several factors may contribute to this apparent lack of zoonotic transmission. First, the receptor-binding preference of contemporary H3N2 CIV strains, particularly those with the HA-V223I mutation, shows a reduced affinity for human-type receptors, which may constitute a significant barrier to efficient infection of the human upper respiratory tract [5]. Second, antigenic analysis has demonstrated that current human seasonal influenza vaccines do not confer protection against H3N2 CIV, indicating that the human population is immunologically naïve to this virus [5]. This immunological naivety could paradoxically increase the risk of a severe outcome should a fully human-adapted variant emerge, but it also means there is no pre-existing cross-reactive immunity that might facilitate detection through serosurveys.

Serological evidence, however, provides a more nuanced picture. Hemagglutination inhibition (HI) assays using human antiserum have shown partial cross-reactivity with canine H3N2 viruses, suggesting that some humans may have antibodies that can recognize the canine virus, likely due to prior exposure to human seasonal H3N2 strains [1]. Conversely, studies have detected antibodies against human H1N1 and H3N2 viruses in dog populations, indicating that reverse zoonosis (human-to-dog transmission) is a frequent event [25, 27]. This bidirectional flow of influenza viruses between humans and dogs is a critical concern. The co-circulation of human and canine IAVs in the same host (the dog) creates a high-risk scenario for genetic reassortment. This was dramatically illustrated by the natural co-infection of a dog with H3N2 CIV and pandemic H1N1 (pH1N1) virus, which led to the generation of 23 distinct reassortant viruses [2]. Some of these reassortants, which carried the M gene from pH1N1 and the HA gene from H3N2 CIV, demonstrated increased pathogenicity in mice, highlighting the potential for novel, more virulent strains to emerge from the canine host [2].

The risk is further amplified by the global epidemiology of the virus. H3N2 CIV is now endemic in China, with periodic introductions into North America and other regions [8, 14]. The virus spreads rapidly in high-density dog populations such as shelters and kennels, creating a large viral biomass in close contact with humans [8, 32]. The World Health Organization (WHO), the World Organisation for Animal Health (WOAH), and the Centers for Disease Control and Prevention (CDC) all recognize the pandemic potential of novel influenza A viruses, and the sustained circulation of an avian-origin virus in a mammalian companion animal is a scenario that warrants continuous, high-level surveillance [40, 45].

Interspecies Transmission to Non-Human Primates: A Critical Sentinel Event

While human infection remains unconfirmed, a landmark study has provided definitive evidence of H3N2 CIV transmission to non-human primates (NHPs). In June 2022, an outbreak of severe respiratory disease occurred in a zoo-housed population of golden monkeys (Rhinopithecus roxellana) in Jiangsu Province, China [3]. Seven animals developed flu-like symptoms, and two succumbed to fatal respiratory distress. Histopathological and immunohistochemical analyses confirmed a diagnosis of influenza A virus pneumonia, and a novel H3N2 reassortant virus, designated A/golden monkey/Jiangsu/1/2022 (Gm-1), was isolated from the lungs of the deceased animals [3].

This event is of profound significance for several reasons. First, it represents the first confirmed report of IAV infection in golden monkeys, an endangered species, highlighting the threat that CIV poses to wildlife conservation [3]. Second, and more critically for public health, genomic sequencing revealed that Gm-1 was a canine/human reassortant virus. The PB1 gene of Gm-1 showed greater than 97% identity with human seasonal H3N2 strains, while the remaining seven genomic segments originated from early local canine H3N2 viruses [3]. This demonstrates unequivocally that reassortment between canine and human influenza viruses can occur in nature and that the resulting progeny virus can be highly pathogenic in a new mammalian host.

To understand the biological implications of this reassortment, researchers used reverse genetics to reconstruct the event. They generated two viruses: rGm-1 (carrying the human-origin PB1) and rJS06 (a control with a canine-origin PB1). In a mouse model, infection with rGm-1 resulted in significantly more severe lung pathology and elevated levels of proinflammatory cytokines compared to rJS06 [3]. This enhanced virulence was attributed to a unique cytotoxic motif (68I, 69L, and 70V) in the PB1-F2 protein of Gm-1, which was absent in the canine virus [3]. The PB1-F2 protein is known to modulate the host immune response and contribute to viral pathogenicity. This finding provides a direct mechanistic link between the acquisition of a human viral gene segment and an increase in virulence in a mammalian host.

The implications of this NHP outbreak are far-reaching. It demonstrates that the canine host can serve as a bridge for the transmission of human influenza genes into a novel viral context, creating a pathogen with enhanced virulence for primates. This event validates the long-held concern that dogs could be a source of novel pandemic influenza strains. The fact that the spillover occurred in a zoo setting, where animals are under human care, underscores the need for enhanced biosecurity measures to protect both captive wildlife and the humans who work with them. The CDC and WOAH have long emphasized the importance of surveillance at the human-animal interface, and this case serves as a stark reminder of the unpredictable consequences of viral mixing.

Experimental Models of Interspecies Transmission

Beyond natural events, experimental studies have been crucial in defining the host range of H3N2 CIV. Guinea pigs (Cavia porcellus) have been established as a valuable mammalian model for studying influenza transmission. Intranasal inoculation of guinea pigs with a Thai H3N2 CIV isolate resulted in successful infection, characterized by fever, viral shedding in the respiratory tract, seroconversion, and mild histopathological lung changes [28]. Crucially, the virus was transmitted from inoculated guinea pigs to both direct-contact and aerosol-exposed sentinel animals, demonstrating that H3N2 CIV can be transmitted via the airborne route in a mammalian model [28]. This finding is critical as it suggests that if the virus were to fully adapt to humans, it could potentially spread via respiratory droplets, similar to seasonal influenza.

Studies in ferrets, which are considered the gold standard model for human influenza transmission, have also shown that some H3N2 CIV variants, particularly those with mutations or reassortments, can exhibit increased virulence and transmissibility [26]. The ability of H3N2 CIV to infect and transmit in these models, combined with the documented presence of both avian and human-type SA receptors in the canine respiratory tract, reinforces the concept of the dog as a potential intermediate host that can facilitate the adaptation of avian influenza viruses for eventual transmission to humans [18, 20].

Conclusion of Analysis

The interspecies transmission and zoonotic potential of H3N2 CIV represent a dynamic and significant threat to public health and wildlife conservation. While the virus has not yet caused a documented human pandemic, the molecular machinery for mammalian adaptation is actively evolving within the canine host. The acquisition of human-like receptor binding, the presence of multiple mammalian-adaptive mutations, and the demonstrated ability to reassort with human influenza viruses, as proven by the fatal outbreak in golden monkeys, collectively paint a picture of a virus on the cusp of a potential host-range expansion. The absence of human cases likely reflects a combination of incomplete receptor adaptation and a lack of exposure opportunity, rather than an inherent inability to infect humans. The continuous, intensive surveillance of H3N2 CIV in dog populations, as advocated by global health authorities, is not merely a veterinary concern but a critical component of pandemic preparedness. The golden monkey outbreak serves as a powerful sentinel event, a clear warning that the next pandemic could emerge not from a bird or a pig, but from man’s best friend.

Prevention and Control Strategies: Vaccination and Biosecurity

The management of H3N2 canine influenza virus (CIV) in domestic dog populations presents a unique and formidable challenge to veterinary practitioners, public health authorities, and animal welfare organizations. Unlike many well-characterized respiratory pathogens of dogs, H3N2 CIV is a relatively recent host-switched pathogen, having emerged from an avian reservoir around 2004 and subsequently establishing sustained circulation in canine populations across Asia and North America [8, 10, 15]. This evolutionary history has profound implications for prevention and control, as the virus retains molecular signatures of its avian origin while accumulating host-adaptive mutations that enhance its fitness in mammals [10, 12]. The cornerstone of any effective control program rests upon two interdependent pillars: vaccination strategies designed to induce robust, durable immunity, and comprehensive biosecurity protocols aimed at interrupting transmission chains. These measures must be implemented with an understanding of the virus’s unique epidemiological behavior, characterized by rapid, explosive outbreaks in high-density populations such as shelters and kennels, followed by local extinction and reliance upon long-distance viral dispersal for perpetuation [8, 14].

Vaccination Strategies: Current Platforms and Future Directions

The development of effective vaccines against H3N2 CIV has been a priority since the virus’s recognition as an emerging pathogen, and several platforms have demonstrated protective efficacy in experimental settings. Inactivated whole-virus vaccines represent the most traditional approach, and studies have confirmed that a monovalent inactivated H3N2 CIV preparation can elicit robust humoral immune responses and provide protection against homologous challenge in mice [13]. The immunogenicity of such vaccines can be significantly enhanced through the use of appropriate adjuvants; for instance, the addition of Montanide ISA 25 adjuvant to a virus-like particle (VLP) vaccine containing hemagglutinin (HA) and M1 proteins resulted in markedly higher hemagglutination inhibition (HI) titers compared to unadjuvanted formulations [11]. The VLP platform itself offers distinct advantages, as these non-infectious particles preserve the native conformation of surface glycoproteins while lacking viral genetic material, thereby eliminating any risk of reversion to virulence [11]. In a critical proof-of-concept study, a single dose of adjuvanted H3 HA VLP vaccine protected specific-pathogen-free beagles against wild-type virus challenge, with vaccinated animals showing reduced viral shedding and attenuated clinical signs [11].

Beyond traditional inactivated and VLP vaccines, recombinant vector-based approaches have shown considerable promise. A live attenuated canine adenovirus type 2 recombinant expressing H3N2 HA (rCAV2-HA) was demonstrated to induce both humoral and cell-mediated immune responses in mice, and provided protective efficacy equivalent to that of an inactivated vaccine upon challenge [13]. This platform is particularly attractive because it leverages a well-characterized vaccine vector already licensed for use in dogs, potentially allowing for combination with routine core vaccinations. In contrast, a DNA vaccine based on the pVAX1 plasmid expressing HA (pVAX1-HA) elicited immune responses but conferred only limited protection, suggesting that protein-based or viral-vectored approaches are currently superior for this pathogen [13]. It must be emphasized, however, that the antigenic landscape of H3N2 CIV is not static. Phylogenetic analyses have revealed that the virus has evolved into multiple clades and genotypes, with the currently predominant genotype 15 having established a stable lineage in dogs since approximately 2017 [5]. The HA gene, in particular, has accumulated amino acid substitutions, including the mammalian-adaptive mutation V223I and the glycosylation-altering residue G146S, that may affect vaccine antigenicity [5, 10]. Critically, antigenic analyses have demonstrated that contemporary human H3N2 influenza vaccines do not confer protection against H3N2 CIV, underscoring the necessity for species-specific vaccine formulations that are updated in accordance with circulating canine strains [5].

The route and method of vaccine delivery also warrant careful consideration. Conventional intramuscular injection, while effective, can be associated with fear and pain responses in canine patients, potentially compromising owner compliance. A novel insertion-responsive microneedle (IRMN) system has been developed that allows for patchless administration of H3N2 vaccine to the ear without requiring hair removal [46]. This system demonstrated 95% delivery efficiency and elicited antibody responses equivalent to intramuscular injection, with markedly improved behavioral compliance from the animals [46]. Such innovations may be particularly valuable for mass vaccination campaigns in shelter environments, where minimizing stress and handling time is paramount. Additionally, the use of qualified Madin-Darby canine kidney (MDCK) cell lines for vaccine virus isolation and propagation is gaining traction, as these systems can yield antigenically appropriate viruses while avoiding the egg-adaptive amino acid substitutions that sometimes compromise vaccine efficacy [42]. The glycosylation patterns of hemagglutinin differ between MDCK cell-derived and egg-derived vaccines, with MDCK-derived HAs containing more partially occupied glycosylation sites and higher-molecular-weight glycans, which may influence immunogenicity [21]. Ongoing surveillance of circulating H3N2 CIV strains is therefore essential to ensure that vaccine antigens remain antigenically matched to field viruses, particularly given the capacity of this virus for rapid evolutionary change [1, 9].

Biosecurity Measures: Interrupting Transmission Chains

Biosecurity represents the second, and arguably more immediately actionable, pillar of H3N2 CIV control. The epidemiology of this virus is characterized by swift spread within high-density populations, with a basic reproductive number (R₀) consistently estimated between 1 and 1.5 across United States outbreaks [14]. This moderate transmissibility, combined with the rapid development of immunity in infected animals, leads to a pattern of epidemic burst–fade-out dynamics: the virus spreads explosively through susceptible populations in shelters and kennels, then dies out locally once the pool of naïve hosts is exhausted [8, 14]. Sustained circulation therefore requires metapopulation dynamics, specifically, the movement of infected animals to new, susceptible populations before local extinction occurs [8, 14]. This has profound implications for biosecurity: interventions that reduce the probability of viral introduction into naïve populations, or that slow the rate of spread within affected populations, can tip the balance from sustained circulation toward extinction.

In kennel and shelter settings, strict isolation of newly arrived animals is the single most important biosecurity measure. Experimental transmission studies have consistently demonstrated that direct contact is sufficient for efficient virus spread, with contact-exposed dogs developing clinical signs within 1–2 days of exposure and shedding virus at titers sufficient to infect subsequent contacts [16, 28, 39]. The duration of viral shedding is influenced by host immune status; dogs receiving immunosuppressive doses of prednisolone shed virus for a significantly longer period (13 days versus 8 days) and at higher peak titers compared to immunocompetent controls [34]. This finding carries direct clinical relevance: veterinary practitioners should exercise caution in prescribing glucocorticoids to dogs with suspected or confirmed influenza, as such therapy may prolong the infectious period and amplify environmental contamination. Furthermore, the virus can be shed before the onset of clinical signs, making quarantine based solely on symptom observation insufficient [39]. A minimum isolation period of 14 days for exposed animals, combined with negative RT-PCR testing before introduction into the general population, is recommended.

Environmental decontamination is another critical component. H3N2 CIV, like other influenza A viruses, is enveloped and relatively susceptible to inactivation by appropriate disinfectants, including quaternary ammonium compounds, accelerated hydrogen peroxide, and sodium hypochlorite solutions. Given that the virus can survive on fomites for hours to days under appropriate conditions, rigorous cleaning and disinfection of kennels, food bowls, and handling equipment between occupants is essential. The role of human fomites in transmission should not be underestimated; shelter and veterinary personnel should implement strict hand hygiene protocols and consider the use of dedicated outerwear or disposable gowns when handling potentially infected animals. In outbreak situations, cohorting of confirmed cases, suspected cases, and unexposed animals into separate air handling zones can reduce the risk of aerosol transmission, although the relative contribution of aerosol versus droplet versus contact transmission in natural settings remains incompletely defined [28].

Surveillance and diagnostic capacity form the intelligence arm of any control program. The H3N2 CIV outbreak in New York City in 2018 demonstrated the value of rapid, centralized data capture and visualization; the New York City Department of Health and Mental Hygiene partnered with the local veterinary medical association to deploy a web-based reporting platform that collected case information and produced an interactive dashboard within two business days [32]. Such systems allow for real-time situational awareness and enable timely deployment of control measures. At the diagnostic level, rapid influenza diagnostic tests (RIDTs) developed for human use have been evaluated for detection of H3N2 CIV, but their sensitivity is poor at low viral loads, one study reported only 36.1% sensitivity compared to real-time RT-PCR [47]. Nucleic acid amplification tests therefore remain the gold standard for confirmation, though RIDTs may have utility as screening tools in high-prevalence settings where viral loads are likely to be elevated [47]. The development of synthetic analytical reference materials for influenza diagnostics, including those containing H3N2 sequences, provides a safe and standardized means of validating molecular assays without requiring live virus, facilitating broader access to quality-assured testing [40].

Finally, it is essential to consider the One Health dimensions of H3N2 CIV control. Dogs are not only maintenance hosts for this virus but also potential bridging hosts for reassortment events with human and swine influenza viruses. The co-circulation of H3N2 CIV with pandemic H1N1 (2009) viruses in naturally co-infected dogs has already generated reassortants in which the M gene of pandemic origin and the HA gene of canine origin predominate, and some of these reassortants exhibited heightened pathogenicity in mammalian models [2]. The presence of both avian-type (α-2,3) and human-type (α-2,6) sialic acid receptors in the canine respiratory tract provides a permissive environment for dual infection and reassortment [7, 18]. Indeed, serological evidence has demonstrated that horses with dog exposure can seroconvert to H3N2 CIV, and human-origin PB1 segments have been found in canine/human reassortant viruses that caused fatal infections in zoo-housed golden monkeys [3, 7]. The World Organisation for Animal Health (WOAH) and the World Health Organization (WHO) both recognize the zoonotic and pandemic potential of influenza A viruses with novel gene constellations, and the circulation of H3N2 CIV in a species that shares intimate domestic environments with humans demands continued vigilance. Biosecurity measures at the human-animal interface, including hand washing after handling dogs, avoiding contact with sick animals, and ensuring that canine influenza vaccine is used in high-risk populations, are prudent public health interventions that align with the broader goals of pandemic preparedness.

References

[1] Peng J, Miao X, Zhang X, Li Z, Wang Y, Liu G, et al.. Molecular Characterization of an H3N2 Canine Influenza Virus Isolated from a Dog in Jiangsu, China, in 2025. Veterinary Sciences. 2025. DOI: https://doi.org/10.3390/vetsci13010032

[2] Na W, Lyoo K, Song E, Hong M, Yeom M, Moon H, et al.. Viral dominance of reassortants between canine influenza H3N2 and pandemic (2009) H1N1 viruses from a naturally co-infected dog. Virology Journal. 2015. DOI: https://doi.org/10.1186/s12985-015-0343-z

[3] Wen X, Chen R, Lv Y, Geng J, Guo J, Sun Y, et al.. Fatal infection of a novel canine/human reassortant H3N2 influenza A virus in the zoo-housed golden monkeys.. Veterinary Microbiology. 2025. DOI: https://doi.org/10.1016/j.vetmic.2025.110792

[4] Ge F, Shen H, Yang D, Yang X, Li X, Wang J, et al.. The biological characteristics and infection dynamics of a novel H3N2 canine influenza virus genotype in beagles. Virology Journal. 2024. DOI: https://doi.org/10.1186/s12985-024-02422-x

[5] Liu L, Wang F, Wu Y, Mi W, Zhang Y, Chen L, et al.. The V223I substitution in hemagglutinin reduces the binding affinity to human-type receptors while enhancing the thermal stability of the H3N2 canine influenza virus. Frontiers in Microbiology. 2024. DOI: https://doi.org/10.3389/fmicb.2024.1442163

[6] Rell F, Apada AMS, Wahyuda AAPJ, Yusuf B, Suharto RH, Jamaluddin AW, et al.. LITERATUR REVIEW : INTRODUCTION OF THE CANINE INFLUENZA VIRUS. Buletin Veteriner Udayana. 2025. DOI: https://doi.org/10.24843/bulvet.2025.v17.i04.p22

[7] Zhou P, Luo A, Xiao X, Hu X, Shen D, Li J, et al.. Serological evidence of H3N2 canine influenza virus infection among horses with dog exposure. Transboundary and Emerging Diseases. 2019. DOI: https://doi.org/10.1111/tbed.13104

[8] 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

[9] 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

[10] 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

[11] Lee D, Bae S, Park J, Kwon J, Yuk S, Song J, et al.. Virus-like particle vaccine protects against H3N2 canine influenza virus in dog. Vaccine. 2013. DOI: https://doi.org/10.1016/j.vaccine.2013.05.023

[12] 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

[13] Tu L, Zhou P, Li L, Li X, Hu R, Jia K, et al.. Evaluation of protective efficacy of three novel H3N2 canine influenza vaccines. OncoTarget. 2017. DOI: https://doi.org/10.18632/oncotarget.21104

[14] 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

[15] Wasik B, Voorhees IEH, Parrish C. Canine and Feline Influenza.. Cold Spring Harbor Perspectives in Medicine. 2019. DOI: https://doi.org/10.1101/cshperspect.a038562

[16] 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

[17] Dubovi EJ, Njaa BL. Canine influenza.. The Veterinary clinics of North America. Small animal practice. 2008. DOI: https://doi.org/10.1016/j.cvsm.2008.03.004

[18] 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

[19] Yong J, Lu S, Lu C, Huang R. The Development History, Structural Composition, and Functions of Influenza Viruses and the Progress of Influenza Virus Inhibitors in Clinics and Clinical Trials.. Mini-Reviews in Medical Chemistry. 2024. DOI: https://doi.org/10.2174/0113895575316416240724043949

[20] 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

[21] Li J, Liu S, Gao Y, Tian S, Yang Y, Ma N. Comparison of N-linked glycosylation on hemagglutinins derived from chicken embryos and MDCK cells: a case of the production of a trivalent seasonal influenza vaccine. Applied Microbiology and Biotechnology. 2021. DOI: https://doi.org/10.1007/s00253-021-11247-5

[22] Klivleyeva N, Glebova T, Shamenova M, Saktaganov N. Influenza A viruses circulating in dogs: A review of the scientific literature. Open Veterinary Journal. 2022. DOI: https://doi.org/10.5455/OVJ.2022.v12.i5.12

[23] Liu Y, Fu C, Ye S, Liang Y, Qi Z, Yao C, et al.. Phosphoproteomics to Characterize Host Response During H3N2 Canine Influenza Virus Infection of Dog Lung. Frontiers in Veterinary Science. 2020. DOI: https://doi.org/10.3389/fvets.2020.585071

[24] 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

[25] Glebova TI, Klivleyeva N, Saktaganov N, Shamenova M, Lukmanova G, Baimukhametova A, et al.. Circulation of influenza viruses in the dog population in Kazakhstan (2023–2024). Open Veterinary Journal. 2024. DOI: https://doi.org/10.5455/OVJ.2024.v14.i8.17

[26] Na W, Xie X, Yeom M, Kang A, Kim H, Lim J, et al.. Morphological features and pathogenicity of mutated canine influenza viruses from China and South Korea.. Transboundary and Emerging Diseases. 2020. DOI: https://doi.org/10.1111/tbed.13494

[27] Jang H, Jackson YK, Daniels J, Ali A, Kang K, Elaish M, et al.. Seroprevalence of three influenza A viruses (H1N1, H3N2, and H3N8) in pet dogs presented to a veterinary hospital in Ohio. Journal of Veterinary Sciences. 2017. DOI: https://doi.org/10.4142/jvs.2017.18.S1.291

[28] 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

[29] Ou J, Lu G, Jia K, Li S. Genetic characterization of a novel group of H3N2 canine influenza viruses isolated from Guangdong Province in southern China in 2018. . 2020. DOI: https://doi.org/10.22541/au.159118775.54777062

[30] Lee E, Kim E, Kim B, Song J, Cho I, Shin Y. Molecular analyses of H3N2 canine influenza viruses isolated from Korea during 2013–2014. Virus genes. 2016. DOI: https://doi.org/10.1007/s11262-015-1274-x

[31] Deng J, Ma C, Yu J, Chen B, Li S, Zhou P. Cats are more susceptible to the prevalent H3 subtype influenza viruses than dogs. Virulence. 2025. DOI: https://doi.org/10.1080/21505594.2025.2605799

[32] Whittemore K, Corrado RE, Paladini M, Davidson A, Huang CC, Daskalakis D, et al.. Data capture and visualization for a canine influenza outbreak , New York City, 2018. Online Journal of Public Health Informatics. 2019. DOI: https://doi.org/10.5210/OJPHI.V11I1.9670

[33] Zheng Y, Fu X, Wang L, Zhang W, Zhou P, Zhang X, et al.. Comparative analysis of MicroRNA expression in dog lungs infected with the H3N2 and H5N1 canine influenza viruses.. Microbial Pathogenesis. 2018. DOI: https://doi.org/10.1016/j.micpath.2018.05.015

[34] Hong M, Kang B, Na W, An D, Moon H, Kim D, et al.. Prolonged shedding of the canine influenza H3N2 virus in nasal swabs of experimentally immunocompromised dogs. Clinical and experimental vaccine research. 2013. DOI: https://doi.org/10.7774/cevr.2013.2.1.66

[35] Williams M, Shropshire S, Chornarm N, Brewer M, Hawley J, Khorsand M, et al.. Effects of canine influenza infection and DA2PP vaccination on the development of platelet-associated immunoglobulins and platelet counts in dogs.. Veterinary clinical pathology. 2024. DOI: https://doi.org/10.1111/vcp.13333

[36] 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

[37] DiGeronimo PM, Why KV, Glass H, Dubovi E, Latney L. Serosurvey for Influenza Virus Subtypes H3N8 and H3N2 Antibodies in Free-Ranging Canids in Pennsylvania, USA. Journal of Wildlife Diseases. 2019. DOI: https://doi.org/10.7589/2018-03-078

[38] Maboni G, Seguel M, Lorton A, Berghaus R, Sanchez S. Canine infectious respiratory disease: New insights into the etiology and epidemiology of associated pathogens. PLoS ONE. 2019. DOI: https://doi.org/10.1371/journal.pone.0215817

[39] Song D, Moon H, Jung K, Yeom M, Kim H, Han S, et al.. Association between nasal shedding and fever that influenza A (H3N2) induces in dogs. Virology Journal. 2011. DOI: https://doi.org/10.1186/1743-422X-8-1

[40] Papazisi L, Asbury H, Bose J, Gujjari P, Castro L, Corriea T, et al.. P-1810. Development of Avian and Human Influenza Analytical Reference Materials for Diagnostics and Surveillance. Open Forum Infectious Diseases. 2026. DOI: https://doi.org/10.1093/ofid/ofaf695.1979

[41] 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

[42] Harada Y, Takahashi H, Fujimoto T, Horikoshi F, Chida S, Tanaka K, et al.. Evaluation of a qualified MDCK cell line for virus isolation to develop cell-based influenza vaccine viruses with appropriate antigenicity.. Vaccine. 2024. DOI: https://doi.org/10.1016/j.vaccine.2024.126242

[43] Plata-Hipólito CB, Cedillo-Rosales S, Obregón-Macías N, Hernández-Luna C, Rodríguez-Padilla C, Tamez-Guerra R, et al.. Genetic and serologic surveillance of canine (CIV) and equine (EIV) influenza virus in Nuevo León State, México. PeerJ. 2019. DOI: https://doi.org/10.7717/peerj.8239

[44] Flannery B, Chung J, Holiday C, Jefferson SN, Gaglani M, Murthy K, et al.. Influenza Antibody Levels Associated with Laboratory-Confirmed Influenza in a Test-Negative Study Design, US Flu VE Network, November 2018–May 2019. medRxiv. 2026. DOI: https://doi.org/10.64898/2026.03.24.26349239

[45] Grohskopf L, Blanton L, Ferdinands J, Chung J, Broder K, Talbot H, et al.. Prevention and Control of Seasonal Influenza with Vaccines: Recommendations of the Advisory Committee on Immunization Practices , United States, 2022–23 Influenza Season. MMWR Recommendations and Reports. 2022. DOI: https://doi.org/10.15585/mmwr.rr7101a1

[46] . Patchless administration of canine influenza vaccine on dog’s ear using insertion-responsive microneedles (IRMN) without removal of hair and its in vivo efficacy evaluation. European journal of pharmaceutics and biopharmaceutics. 2020. DOI: https://doi.org/10.1016/j.ejpb.2020.06.006

[47] Saipinta D, Chuammitri P, Ohkura S, Luangwaranan B, Tadee P. Human rapid influenza diagnostic test: efficacy and significance for using in animal. Veterinary Integrative Sciences. 2023. DOI: https://doi.org/10.12982/vis.2023.021