Ferret Influenza A Virus

Overview and Taxonomy of Ferret Influenza A Virus

The ferret (Mustela putorius furo) occupies a singular and indispensable position in influenza A virus (IAV) research, serving as the preeminent small mammalian model for investigating viral pathogenesis, transmissibility, and pandemic potential. This primacy is not arbitrary but is rooted in a confluence of anatomical, physiological, and molecular characteristics that render the ferret uniquely susceptible to and representative of human IAV infection. The comparative anatomy of the ferret respiratory tract, its distribution of sialic acid (SA) receptors, and its clinical response to infection collectively establish a platform that recapitulates human disease with a fidelity unmatched by other conventional laboratory species such as mice or guinea pigs [1, 2]. A comprehensive understanding of the taxonomy and overview of IAVs studied within the ferret model is therefore foundational to interpreting the vast corpus of literature informing pandemic preparedness, antiviral development, and vaccine efficacy assessments.

The Ferret as a Model Organism: Biological and Historical Context

The suitability of the ferret for IAV research is fundamentally grounded in its expression of sialic acid receptors on the epithelial cells of its respiratory tract. Human-adapted IAVs preferentially bind to α2,6-linked SA receptors, which are abundant in the human upper respiratory tract (URT). Critically, ferrets, like humans, possess a high density of α2,6-linked SA receptors on ciliated and nonciliated cells of the nasal epithelium, trachea, and bronchi, making them permissive to infection with human IAV isolates without the need for prior viral adaptation [3, 4]. Conversely, the ferret URT also expresses α2,3-linked SA receptors, the preferential binding target of avian influenza viruses, allowing for the direct study of zoonotic strains with pandemic potential [3]. This dual receptor expression pattern is a defining feature that distinguishes the ferret from the mouse, which predominantly expresses α2,3-linked SA, thereby often requiring mouse-adapted virus strains for productive infection [1]. Furthermore, a seminal discovery revealed that ferrets, uniquely among common laboratory carnivores, lack a functional cytidine monophosphate-N-acetylneuraminic acid hydroxylase (CMAH) gene, resulting in the exclusive synthesis of N-acetylneuraminic acid (Neu5Ac) over N-glycolylneuraminic acid (Neu5Gc) [4]. This "humanized" sialic acid profile mirrors the human condition, where a fixed pseudogene renders the CMAH enzyme inactive, providing a profound molecular basis for the ferret's exceptional susceptibility to human IAV strains [4].

The clinical manifestations of IAV infection in ferrets closely parallel those observed in humans. Inoculated ferrets develop pronounced clinical signs, including fever, lethargy, sneezing, nasal discharge, and weight loss, which can be quantified to assess disease severity [1, 2, 5]. The onset of fever, a cardinal sign of systemic infection, is a reliable indicator of disease progression [6]. Importantly, the ferret model uniquely permits the simultaneous evaluation of both pathogenicity and transmissibility, two critical axes of pandemic risk assessment [7, 8]. Transmission studies can be configured to model direct contact, respiratory droplet, or fomite routes, providing mechanistic insights into how a virus may spread among humans [9, 10, 11]. The ability of a virus to transmit via the airborne route, defined as the spread of infectious particles over a distance without direct contact, is considered a hallmark of pandemic-capable strains, and the ferret model is the gold standard for assessing this trait [12, 13, 14, 8, 15].

Taxonomic Spectrum of Influenza A Viruses in the Ferret Model

The taxonomic diversity of IAVs evaluated in the ferret model is vast, encompassing viruses of human, swine, and avian origin across multiple hemagglutinin (HA) subtypes (H1-H9, H16) and neuraminidase (NA) subtypes [16, 5]. This breadth is a testament to the ferret model's utility in addressing fundamental questions of host range, adaptation, and interspecies transmission. The viruses can be broadly categorized by their host of origin and their associated risk profile.

Human Seasonal and Pandemic Viruses: The ferret model has been extensively employed to study the virological properties and transmission dynamics of contemporary seasonal IAVs, including the A(H1N1)pdm09 and A(H3N2) subtypes [17, 18, 19, 5, 20]. These viruses are naturally adapted to human hosts and typically cause mild to moderate respiratory disease in ferrets, mirroring their behavior in immunocompetent human populations [21]. The A(H1N1)pdm09 virus, which emerged in 2009, has been a particularly frequent subject of study, serving as a prototype for a fully pandemic-adapted virus that transmits efficiently via the airborne route [22, 6]. Its replication kinetics in ferrets, characterized by rapid and robust shedding from the URT, directly correlate with its high transmissibility [22, 13, 15]. Studies utilizing A(H1N1)pdm09 have also been instrumental in elucidating the mechanisms of antiviral drug resistance, revealing that mutations such as NA-H275Y (conferring oseltamivir resistance) and PA-I38T (conferring baloxavir resistance) can emerge and, importantly, retain the capacity for airborne transmission in ferrets despite in vitro fitness costs [23, 24, 25]. Similarly, A(H3N2) viruses, which undergo rapid antigenic drift, are routinely passaged in ferrets to generate post-infection antisera for the WHO global influenza surveillance system, a process that directly informs annual vaccine strain selection [26, 19].

Swine-Origin Influenza A Viruses: Pigs serve as a critical intermediate host for IAV evolution, capable of generating reassortant viruses with pandemic potential, as tragically demonstrated by the 2009 H1N1 pandemic. The ferret model has been pivotal in risk assessment of emerging swine-origin IAVs, particularly the H1N1 variant (H1N1v) viruses that sporadically infect humans following swine exposure [27]. Studies have demonstrated that while H1N1v viruses can replicate efficiently and transmit among co-housed ferrets, their ability to transmit via respiratory droplets is often less efficient than fully human-adapted viruses, suggesting that additional mammalian adaptation mutations are likely required for sustained human-to-human transmission [27]. The emergence of antigenically drifted H1N1v strains, such as A/Ohio/09/2015, which possesses a key HA substitution (G155E) that facilitates escape from pre-existing immunity elicited by prior A(H1N1)pdm09 infection, underscores the ongoing threat these viruses pose and the necessity for continuous surveillance in swine populations [27].

Avian-Origin Zoonotic Influenza A Viruses (Low-Pathogenicity): Avian IAVs constitute the vast reservoir of genetic diversity from which pandemic viruses emerge. The ferret model has been indispensable for characterizing the threat posed by zoonotic avian viruses that have caused human infection, most notably the low-pathogenicity avian influenza (LPAI) A(H7N9) virus lineage. Following its emergence in China in 2013, A(H7N9) viruses were rapidly assessed in ferrets, revealing a capacity for robust replication in the URT and, in some strains, efficient transmission among ferrets via respiratory droplets [28]. This alarming phenotype, combined with a high case fatality rate in humans, led to the classification of H7N9 as having the highest pandemic potential among avian influenza viruses at the time. The ferret model further demonstrated that prior immunity from seasonal A(H1N1)pdm09 infection could provide partial, but not complete, cross-protection against H7N9 challenge, a phenomenon attributed to the presence of group 1 HA stalk antibodies [29]. Other low-pathogenicity avian viruses, such as H9N2, have also been studied in ferrets, albeit to a lesser extent, to assess their potential for mammalian adaptation [30].

Highly Pathogenic Avian Influenza (HPAI) Viruses: In the current era, the most intensive focus of ferret-based research has been on HPAI H5Nx viruses, particularly those belonging to the A/goose/Guangdong/1/96 lineage, clade 2.3.4.4. This lineage has undergone extensive diversification through reassortment, giving rise to multiple HA-NA subtype combinations, including H5N1, H5N2, H5N6, and H5N8, all of which have been evaluated in ferrets [31, 32]. The taxonomic focus has sharpened dramatically on the clade 2.3.4.4b H5N1 sub-lineage since its panzootic spread across five continents, with unprecedented spillover into a wide array of mammalian species, including dairy cattle [29, 9, 10, 12, 33, 11]. Ferret studies have shown that the pathogenicity and transmissibility of these H5N1 viruses are highly variable and strain-dependent. Earlier isolates from the 2021-2022 epizootic, such as those from North American wild birds, were found to be highly pathogenic, causing severe systemic disease and death in inoculated ferrets, but lacked the ability to transmit via the airborne route [9, 33, 11]. However, a critical and alarming shift has been observed with isolates from the 2024 U.S. dairy cattle outbreak. A/Texas/37/2024 (TX/37), a virus isolated from a dairy farm worker, demonstrated a remarkable capacity for robust airborne transmission in ferrets, albeit at lower levels than human seasonal strains [11, 34]. This finding, substantiated by air-sampling studies that detected infectious TX/37 virus expelled by infected ferrets, represents a historic milestone, marking the first report of an HPAI H5N1 clade 2.3.4.4b virus exhibiting measurable airborne transmissibility in the mammalian ferret model [35, 11]. The virus retains an avian-like receptor binding preference for α2,3-linked SA, yet its ability to transmit via the air suggests that other mammalian adaptation traits, such as a stabilized HA, an enhanced polymerase activity mediated by mutations like PB2 T271A or E627K, and a capacity for systemic spread, can, in combination, partially overcome the receptor-binding barrier [12, 14, 32, 36]. The taxonomic and functional diversity of H5Nx viruses, as revealed through ferret studies, underscores the dynamic and unpredictable nature of the pandemic threat these viruses represent.

Molecular Pathogenesis of Ferret Influenza A Virus

The molecular pathogenesis of influenza A virus (IAV) in the ferret model is a multifaceted interplay between viral genetic determinants, host receptor landscapes, innate immune barriers, and the capacity for systemic dissemination. Ferrets (Mustela putorius furo) possess a unique constellation of molecular features that render them exquisitely susceptible to human-adapted and zoonotic IAV strains, making them the gold-standard model for recapitulating human influenza disease. The molecular underpinnings of this susceptibility, the viral strategies for replication and immune evasion, and the mechanistic basis for extrapulmonary spread and severe disease are detailed below, drawing exclusively from the corpus of available experimental evidence.

Molecular Determinants of Host Tropism and Entry

At the molecular level, the initial and perhaps most critical determinant of IAV pathogenesis in ferrets is the expression and distribution of sialic acid (SA) receptors. A seminal finding is that ferrets, uniquely among common laboratory mammals, exclusively synthesize N-acetylneuraminic acid (Neu5Ac) and do not produce N-glycolylneuraminic acid (Neu5Gc) [4]. This is due to an ancient, nine-exon deletion in the ferret CMAH gene, a molecular lesion shared with Pinnipedia and Musteloidia members of the Carnivora [4]. This exclusive expression of Neu5Ac mirrors the human condition, where the CMAH gene is also inactive. The absence of Neu5Gc on ferret cell surfaces means that human-adapted IAV strains, which have evolved to recognize Neu5Ac-containing receptors, encounter a permissive environment. Receptor-binding specificity is further refined by the linkage of SA to galactose. Differentiated primary ferret nasal epithelial cell (FNEC) cultures, which recapitulate the pseudostratified respiratory epithelium, display both α2,6-linked and α2,3-linked SA receptors on their apical surface, mirroring the distribution in the human upper respiratory tract [3]. The α2,6-linked SA receptors, the preferential target of human seasonal IAV, are abundant on ciliated cells, while α2,3-linked SA receptors, bound by avian influenza viruses, are found predominantly on nonciliated cells [3]. This differential cellular tropism has profound pathogenic consequences: a highly pathogenic avian influenza (HPAI) A(H5N1) virus primarily infects nonciliated cells, whereas a seasonal A(H1N1) virus infects both ciliated and nonciliated cells [3]. The molecular adaptation of IAV hemagglutinin (HA) to recognize human-like α2,6-linked SA is a hallmark of pandemic potential. For instance, avian-origin H3N2 canine influenza viruses (CIVs), after a decade of adaptation in dogs, acquired the ability to recognize human-like α2,6-Gal receptors, a molecular shift that correlated with 100% respiratory droplet transmission in ferrets [37]. Conversely, the 2024 A/Texas/37/2024 (H5N1) virus isolated from a dairy farm worker, despite maintaining an avian-like receptor-binding specificity, demonstrated heightened virulence and airborne transmission in ferrets [11]. This suggests that while receptor specificity is a critical barrier, other molecular factors can partially compensate to facilitate mammalian pathogenesis.

Molecular Regulation of Viral Replication and Polymerase Adaptation

Once inside the host cell, the viral ribonucleoprotein (vRNP) complex, comprising the polymerase basic 1 (PB1), PB2, polymerase acidic (PA), and nucleoprotein (NP), orchestrates replication and transcription. The efficiency of this complex is a central molecular determinant of pathogenesis. The PB2 subunit is a hotspot for host-range adaptation. Acquisition of mammalian-adapting mutations, such as PB2 E627K, is a well-documented facilitator of efficient replication in mammalian cells. A human clade 2.3.4.4 A/H5N6 virus possessing PB2 E627K exhibited high polymerase activity in vitro, correlating with high-titer replication in the ferret respiratory tract, though this alone did not confer airborne transmissibility [36]. The PB2 T271A substitution, identified in a 2023 H5N1 clade 2.3.4.4b isolate from a mink outbreak, was directly linked to pathogenic potential; reversing this mutation reduced mortality and airborne transmission in ferrets [12]. Similarly, the D701N substitution, a known mammalian adaptation marker, was positively selected in ferrets infected with a Dutch H5N6 virus, indicating that in vivo selection pressures can rapidly refine polymerase function for enhanced replication in the mammalian host [31].

The PA protein has emerged as a critical node in both replication fitness and antiviral resistance. The I38T substitution in PA, which confers reduced susceptibility to the antiviral baloxavir, impairs PA endonuclease activity but has variable consequences on overall viral fitness [24]. While such mutant viruses often show attenuated replication in vitro, they retain the capacity for contact and airborne transmission in ferrets, with the A/H1N1pdm09 I38T variant showing substantially lower between-host fitness compared to wild-type [24, 25]. This highlights a nuanced molecular pathogenesis where within-host fitness costs (reduced replication) do not always preclude between-host transmission. Furthermore, deep sequencing studies of swine H1N1 isolates revealed that minor variants with polymerase-enhancing mutations (e.g., PA-S321) could be selected and propagated in ferrets after airborne transmission, illustrating how quasispecies dynamics can rapidly alter the molecular composition of the infecting virus to favor mammalian adaptation [14].

Beyond the polymerase subunits, the HA protein's acid stability is a molecular trait that strongly correlates with transmissibility. The HA protein must undergo a pH-dependent conformational change to mediate membrane fusion. Swine H1N1 variants with HA-stabilizing mutations, such as HA1-S210, were selected within days in inoculated ferrets and were transmitted by both contact and airborne routes [14]. The study demonstrated that HA stabilization played a more prominent role than polymerase enhancement in promoting replication and transmission of these viruses, underscoring that the entry step, governed by HA stability, is a rate-limiting molecular bottleneck [14].

Innate Immune Antagonism and Host-Virus Dynamics

The host innate immune response, particularly the interferon (IFN) system, presents a formidable barrier that IAV must overcome. The ferret model has yielded detailed insights into this molecular arms race. Primary differentiated FNECs infected with IAV induce a robust and rapid type-I/II and type-III IFN response [18]. In contrast, influenza B virus (IBV) infection in the same system triggers a delayed and reduced IFN response, with diminished type-III IFN secretion and downregulation of thymic stromal lymphopoietin (TSLP), an IFN-induced gene that enhances adaptive immunity [18]. This differential molecular response in the ferret upper respiratory tract correlates with the observation that IAV elicits more robust antibody responses in ferrets than IBV, highlighting the critical role of early innate events in shaping downstream adaptive immunity.

At the molecular level, interferon-inducible transmembrane proteins (IFITMs) are potent restriction factors that inhibit the cytosolic entry of pH-dependent viruses, including IAV. The ferret IFITM locus has been characterized, revealing synteny with other mammals, and functional studies demonstrated that ferret IFITM1, -2, and -3 are transcriptionally upregulated in response to both IFN-α stimulation and direct IAV infection [38]. This establishes that the ferret possesses a functional IFITM-mediated intrinsic immune barrier that can limit viral entry. The virus counteracts these defenses. For instance, IAV infection itself upregulates the host microRNA miR-1290 through the ERK pathway [39]. This miRNA targets and reduces vimentin expression. Since vimentin binds the PB2 subunit of vRNP, its knockdown paradoxically increases vRNP nuclear retention and enhances viral polymerase activity [39]. This represents a host species-specific proviral mechanism, as miR-1290 is not present in chickens or mice, and its induction in ferrets (and humans) provides a molecular advantage to the virus.

The role of viral interference in shaping pathogenesis is also notable. Infection with one IAV can prevent or limit infection with another, unrelated respiratory virus (e.g., human respiratory syncytial virus) or a different IAV subtype [17, 40]. The interval between infections and the specific virus hierarchy are critical determinants, with ongoing shedding from a primary infection associated with interference against a secondary challenge [40]. This suggests that the molecular state of the infected ferret's respiratory epithelium, potentially mediated by IFN-induced antiviral states, directly influences susceptibility to subsequent viral infections.

Molecular Drivers of Disseminated and Systemic Pathogenesis

A hallmark of highly pathogenic IAV infections in ferrets is the capacity for systemic dissemination, particularly to the central nervous system (CNS). The molecular basis for neurotropism has been directly investigated. Using an HPAI A/Indonesia/5/2005 (H5N1) virus, three substitutions, PB1 E177G, PB1 A652T, and NP I119M, were identified in the CNS of a ferret that developed severe meningoencephalitis [41]. Individually or collectively, these mutations increased polymerase activity in vitro. In vivo, the virus bearing these CNS-associated mutations retained the capacity to infect the CNS but displayed reduced dispersion to other anatomical sites, suggesting a neurotropic specialization [41]. Critically, analysis of viral populations in the nasal turbinate and olfactory bulb revealed a lack of a genetic bottleneck at the entry point to the CNS, and virus populations bearing the CNS mutations showed signs of positive selection in the brainstem [41]. This demonstrates that selective evolutionary processes can operate within the CNS microenvironment, driving adaptation to this privileged site.

The capacity for systemic spread is further exemplified by the 2024 A/Texas/37/2024 (H5N1) virus, which, in ferrets, caused a severe and fatal infection characterized by viraemia and extrapulmonary spread to multiple organs [11]. Similarly, a Dutch H5N6 virus replicated and caused severe lesions not only in respiratory tissues but also in the brain, liver, pancreas, spleen, lymph nodes, and adrenal gland [31]. The molecular correlates of this pan-tropism are complex and multifactorial. The highly pathogenic phenotype often cannot be explained by a single known mammalian adaptation marker [31]. The virus's ability to cause thrombocytopenia is another element of systemic pathogenesis. In ferrets, the degree of platelet count reduction correlates with the pathogenicity of the infecting strain (H3N2 < H1N1pdm09 < H5N1) [42]. The molecular mechanism involves virus binding to sialic acids on platelets, with the viral neuraminidase removing sialic acids, triggering hepatic clearance of the platelets [42]. This provides a direct molecular link between viral replication and a systemic clinical sequela.

The role of the upper respiratory tract as the primary source of airborne transmission has been molecularly dissected. Using genetically tagged viruses, it was demonstrated that viruses transmitted via the air from ferret donors are consistently of the same genotype as those inoculated intranasally, confirming that the nasal respiratory epithelium is the exclusive source of expiratory aerosols [15]. Transmissible viruses replicate robustly in the nasal epithelium and are released into the air rapidly after infection, whereas poorly transmissible viruses show delayed or abrogated airborne shedding [22, 35]. This reinforces that the molecular efficiency of replication in the upper respiratory tract epithelium is the gatekeeper for onward transmission.

Molecular Determinants of Pathogenicity and Transmission

Pathogenicity in ferrets is not a monogenic trait but is governed by a complex polygenic network. The presence of mammalian adaptation markers in HA and PB2 correlates with more rapid growth in the upper respiratory tract early after infection [5]. However, the relationship between in vitro viral titers and in vivo outcomes is not straightforward. A meta-analysis of over 50 IAV found that viral titers in ferret nasal washes and nasal turbinates correlated positively with peak titers in human bronchial epithelial (Calu-3) cells, suggesting that intrinsic replicative capacity in human airway cells is a relevant correlate [16]. Yet, additional viral traits, such as HA stability and polymerase activity, are essential components of the transmissible phenotype.

The molecular pathogenesis of influenza in ferrets is also profoundly shaped by host immunological history. Immune imprinting, resulting from a first influenza infection in childhood, dictates the outcome of subsequent infections. In ferrets, imprinting with Group 1 H1N1 or H2N3 viruses conferred complete protection (100% survival) against a lethal H5N1 challenge, whereas H3N2-imprinted ferrets suffered severe disease with 40% mortality [43]. This is mechanistically linked to cross-reactive antibodies targeting the conserved HA stalk and N1 neuraminidase from the group 1 HA lineage [29]. The molecular basis of this protection lies in the redirection of the antibody response from the variable HA head to the conserved stalk domain, a strategy that is the foundation for universal influenza vaccine development [44, 45, 46]. The ferret model has thus become indispensable for dissecting the molecular pathways of immune-mediated protection and for understanding how antigenic drift, even in the neuraminidase protein, can reduce vaccine effectiveness [19, 47].

Epidemiology and Host Range of Ferret Influenza A Virus

The ferret (Mustela putorius furo) occupies a uniquely pivotal position in the ecology of influenza A viruses (IAV), serving simultaneously as a natural host, a sentinel species for zoonotic spillover events, and the gold-standard experimental model for assessing pandemic risk. Understanding the full spectrum of ferret IAV epidemiology requires dissecting their susceptibility across diverse viral subtypes, the mechanisms governing species-barrier crossings, the role of immune imprinting in shaping outbreak dynamics, and the increasingly critical interface between domestic ferrets and enzootic avian reservoirs. The recent expansion of highly pathogenic avian influenza (HPAI) H5N1 clade 2.3.4.4b into mammalian populations has thrust the ferret into an unprecedented position as both a victim of spillover and an indispensable tool for predicting human pandemic potential.

Natural History and Susceptibility as a Permissive Host

Ferrets are uniquely permissive to a remarkably broad array of IAV subtypes without prior adaptation, a phenomenon rooted in their evolutionary biology. Unlike mice, which require serial passage to become susceptible to human IAV isolates, ferrets express a sialic acid receptor profile that closely mirrors that of humans. Critically, ferrets exclusively synthesize N-acetylneuraminic acid (Neu5Ac) on their cell surfaces due to a nine-exon deletion in the cytidine monophosphate-N-acetylneuraminic acid hydroxylase (CMAH) gene [4]. This nine-exon deletion, shared with other Musteloidia and Pinnipedia members but absent in most other mammals, renders ferrets incapable of producing N-glycolylneuraminic acid (Neu5Gc). Humans share this exact biochemical deficiency, resulting in "naturally humanized" influenza virus receptors [4]. This molecular convergence means that ferret respiratory epithelium presents both α2,6-linked sialic acids (preferred by human-adapted viruses) and α2,3-linked sialic acids (preferred by avian viruses) on the apical surface of ciliated and nonciliated cells, respectively [3]. The differentiated primary ferret nasal epithelial cell (FNEC) culture model has demonstrated that seasonal H1N1 viruses infect both ciliated and nonciliated cells, while HPAI H5N1 viruses preferentially target nonciliated cells, reflecting the differential receptor tropism observed in human airway epithelium [3].

The epidemiological consequence of this receptor architecture is that ferrets are susceptible to natural infection with human seasonal IAV subtypes (H1N1, H3N2, and influenza B viruses), swine-origin variants, and a wide array of avian-origin viruses, including H5N1, H5N6, H5N8, H7N9, and H9N2 subtypes [1, 2]. Experimental inoculation of ferrets with over 125 contemporary IAV isolates spanning H1, H2, H3, H5, H7, and H9 subtypes has consistently demonstrated productive viral replication in the upper respiratory tract, with viral titers in nasal wash specimens, nasal turbinate tissue, and lung tissue all demonstrating positive inter-site correlations [5]. Importantly, the presence of mammalian adaptation markers in the hemagglutinin (HA) and polymerase basic 2 (PB2) proteins is associated with more rapid growth kinetics in the ferret upper respiratory tract early after infection, distinguishing viruses with pandemic potential from strictly avian-adapted strains [5].

Zoonotic Spillover Events and Emerging Outbreaks in Domestic Ferrets

Until recently, natural IAV infection in pet ferrets was considered rare, but the global expansion of HPAI H5N1 clade 2.3.4.4b has fundamentally altered this epidemiological landscape. In June 2023, concomitant with an unprecedented outbreak of HPAI H5N1 in domestic cats across Poland, the first documented natural infection of pet ferrets with A/H5N1 avian influenza was reported [10]. This outbreak involved five ferrets from a single household, including three clinically affected 9-week-old juveniles, their healthy mother, and another clinically normal adult. The juvenile ferrets presented with profound lethargy, pulmonary distress, and, in one fatal case, neurological symptoms culminating in death. Postmortem analysis of the succumbed ferret revealed A/H5N1 viral RNA in lungs, trachea, heart, brain, pancreas, liver, and intestine, demonstrating the capacity for systemic dissemination in naturally infected animals [10]. Critically, two of the five ferrets were shedding virus in throat swabs despite displaying no overt clinical signs, suggesting the possibility of asymptomatic viral shedding by ferrets [10]. This finding carries profound implications for zoonotic risk, as ferrets are kept in close proximity to humans as companion animals, and the detection of viral RNA in the gastrointestinal tract of the deceased animal underscores the potential for fecal-oral transmission routes within households.

The source of infection in this Polish outbreak was suspected to be contaminated poultry products, as the ferrets were fed a diet of fresh or frozen chicken, a common practice among ferret owners [10]. The World Organisation for Animal Health (WOAH) has subsequently emphasized that feeding raw or undercooked poultry to carnivorous companion animals represents a previously underappreciated transmission pathway from the avian reservoir to mammalian hosts. This outbreak highlights that domestic ferrets can serve as sentinel animals for HPAI activity in the environment, paralleling their traditional use in laboratory settings for risk assessment.

Experimental Host Range and the Spectrum of Viral Fitness

The experimental host range of ferret IAV extends far beyond naturally occurring infections, as ferrets are routinely employed to assess the pandemic potential of emerging zoonotic strains. Systematic evaluation of 14 diverse IAV strains encompassing human, swine, and avian origins has demonstrated that transmissible viruses display robust replication kinetics and rapid release of infectious virus into the air, while poorly transmissible or non-transmissible viruses exhibit significantly reduced or delayed replication with lower detection of airborne viral RNA at early time points post-inoculation [22]. This correlation between upper respiratory tract replication efficiency and airborne transmission has been mechanistically linked to the nasal respiratory epithelium as the primary anatomical site from which infectious virus is expelled into the air [15]. Using genetically tagged viruses, it has been definitively shown that viruses inoculated intranasally (infecting the upper respiratory tract) are transmitted to contact ferrets, whereas viruses inoculated intratracheally (restricted to the lower airways) are not [15]. This observation has profound epidemiological implications: any IAV strain that fails to achieve robust replication in the ferret nasal epithelium is unlikely to achieve sustained human-to-human transmission.

The evaluation of HPAI H5Nx viruses in the ferret model has revealed a spectrum of pathogenicity and transmissibility that informs global risk assessments by the World Health Organization (WHO) and the U.S. Centers for Disease Control and Prevention (CDC). Clade 2.3.4.4b H5N1 viruses isolated from human cases have demonstrated remarkable virulence in ferrets. The A/Chile/25945/2023 H5N1 virus, a novel reassortant containing four gene segments (PB1, PB2, NP, MP) from North American lineage, caused fatal disease characterized by high morbidity and extrapulmonary spread in inoculated ferrets, with transmission to naïve contacts in direct contact settings but no productive respiratory droplet transmission [9]. Similarly, the A/Texas/37/2024 H5N1 virus isolated from a dairy farm worker during the unprecedented U.S. dairy cattle outbreak caused severe and fatal infection in ferrets, characterized by viremia and extrapulmonary spread [11]. Critically, this virus was capable of efficient direct contact transmission, indirect transmission via fomites, and even airborne transmission, albeit at lower levels than seasonal H1N1 strains, despite maintaining an avian-like receptor-binding specificity [11]. Real-time air sampling of infected ferrets has confirmed that recent H5N1 strains, including the 2024 bovine-associated human isolate, are efficiently expelled into the air, whereas earlier 2005 zoonotic and 2024 bovine H5N1 viruses were not detected in air samples [35]. This temporal evolution suggests that ongoing circulation in mammalian hosts is selecting for variants with improved airborne shedding capacity.

Host Range Expansion Through Mammalian Adaptation

The epidemiological trajectory of HPAI H5N1 clade 2.3.4.4b is defined by its remarkable host range expansion, which has been systematically interrogated using ferret models. Spillover events into farmed mink in Spain, seals, and dairy cattle have raised urgent questions about mammalian adaptation. One isolate derived from mink was demonstrated to transmit by direct contact to 75% of exposed ferrets and by airborne transmission to 37.5% of contacts, the first report of a clade 2.3.4.4b H5N1 virus exhibiting both direct contact and airborne transmissibility in ferrets [12]. This isolate carried the PB2 T271A adaptive mutation, and reversion of this mutation reduced mortality and eliminated airborne transmission, directly linking a single amino acid change to a quantitative increase in pandemic potential [12]. Sequence analyses of additional mammalian isolates have identified the emergence of the mammalian adaptation substitution D701N in the PB2 protein in ferrets infected with a Dutch H5N6 virus, indicating that positive selection for enhanced replication in mammalian cells occurs rapidly upon infection [31].

The evolution of H5N1 viruses within the central nervous system of ferrets represents another dimension of host range expansion with critical clinical implications. In-depth analysis of viral populations in the olfactory bulb and brainstem of infected ferrets has demonstrated that neurotropic H5N1 strains can undergo positive selection within the brain, with substitutions in PB1 (E177G, A652T) and NP (I119M) identified as associated with severe meningoencephalitis [41]. These CNS-adapted mutations increased polymerase activity in vitro, and the lack of a genetic bottleneck in the olfactory bulb-brainstem axis suggests that neuroinvasion does not impose severe fitness constraints, allowing diverse viral quasispecies to access the CNS [41]. The WHO has noted that the increasing frequency of neurological complications in human H5N1 cases may reflect ongoing adaptation of these viruses to mammalian CNS tissue.

Immune Imprinting and Its Impact on Host Range Dynamics

The epidemiological landscape of ferret IAV infection is profoundly shaped by immune imprinting, the phenomenon whereby an individual's first influenza virus exposure determines the trajectory of subsequent immune responses. Using a ferret pre-immune model, investigators have demonstrated that ferrets imprinted with Group 1 hemagglutinin viruses (H1N1, H2N3) are completely protected against lethal H5N1 challenge, with 100% survival and minimal clinical symptoms, whereas H3N2-imprinted ferrets experience severe disease with 40% mortality [43]. This differential protection correlates with the presence of cross-reactive antibodies targeting the conserved Group 1 HA stalk domain and N1 neuraminidase [29, 43]. The immunological hierarchy observed in ferrets mirrors epidemiological patterns in humans, where historical imprinting with Group 1 viruses has been associated with reduced mortality during H5N1 outbreaks.

The implications of immune imprinting for ferret IAV epidemiology are twofold. First, the establishment of pre-immune ferret models is essential for accurately predicting vaccine effectiveness in human populations with diverse immune histories [48]. Second, the finding that ferrets consecutively infected with H1N1 followed by H3N2 still maintain protection against H5N1 challenge suggests that imprinting by early-life infection creates an immunological "memory hierarchy" that persists even after subsequent heterosubtypic exposures [43]. This has direct relevance for ferret populations in multi-pet households or breeding facilities, where sequential circulation of different IAV subtypes may shape herd immunity dynamics.

Viral Interference and Co-infection Dynamics within the Ferret Host Range

The epidemiology of IAV in ferrets is further complicated by viral interference, the phenomenon whereby infection with one virus temporarily prevents or limits infection with a second virus. Using the ferret model, it has been demonstrated that infection with influenza A virus can prevent or limit subsequent infection with human respiratory syncytial virus (hRSV), while prior hRSV infection paradoxically reduces morbidity attributed to subsequent IAV infection [17]. The interval between infections is a critical determinant, with viral interference observed only when the interval between primary infection and challenge is less than one week [40]. Moreover, different IAV subtypes exhibit an ordered hierarchy in their ability to block or delay heterologous infection, which may contribute to the dominance patterns observed in seasonal influenza epidemics [40]. This hierarchy has been systematically mapped: A(H1N1)pdm09 consistently blocks or delays A(H3N2) and influenza B infection more effectively than the reverse, suggesting that competitive dynamics within the ferret upper respiratory tract could influence which subtypes circulate in ferret populations during periods of co-circulation.

The ferret upper respiratory tract microbiome also plays a role in modulating susceptibility to IAV infection. Longitudinal metagenomic analysis of ferret and human URT microbiomes during IAV infection has revealed that both species exhibit a similar pattern of microbiome disturbance and resilience, with progression from a stable "healthy ecostate" to an "unhealthy ecostate" characterized by blooms of Pseudomonadales, followed by recovery coincident with viral clearance [49]. This parallel between ferret and human microbiome dynamics validates the ferret model for investigating bacterial co-infections, which are a major cause of influenza-associated mortality.

Swine-Origin IAV and the Human-Ferret-Swine Interface

Swine-origin IAVs represent a particularly important component of the ferret host range, as pigs are considered "mixing vessels" for the generation of pandemic strains. Ferrets exposed to swine H1N1 variant viruses (H1N1v) have demonstrated efficient replication in the respiratory tract and transmission among co-housed animals, though respiratory droplet transmission is often less efficient than that of pandemic H1N1 2009 viruses [27]. The antigenic diversity of swine-lineage H1 viruses is captured by ferret antisera, which is used by the WHO Global Influenza Surveillance and Response System (GISRS) to monitor antigenic drift. Serological investigation of zoonotic swine H1avN1 infections in humans has revealed cross-reactivity with ferret antisera raised against pandemic H1N1 2009, indicating antigenic relatedness that may inform vaccine strain selection [50].

The stabilization of hemagglutinin and enhancement of polymerase activity are critical traits that promote airborne transmission of swine IAV in ferrets. Deep sequencing of swine H1N1 isolates has revealed that minor variants with stabilized HA (HA1-S210) and enhanced polymerase (PA-S321) are selected within days of infection in ferrets, and these variants are preferentially transmitted by the airborne route [14]. This selection pressure demonstrates that ferrets can act as an "evolutionary sieve" that enriches for mammalian-adapted variants present at low frequency in swine populations, a finding that has prompted calls for enhanced surveillance of swine IAV using ferret transmission models.

Implications for Global Health Security

The expanding host range of HPAI H5N1 clade 2.3.4.4b, as elucidated through ferret studies, has prompted updated risk assessments from the WHO, CDC, WOAH, and the Food and Agriculture Organization (FAO). The detection of airborne transmissibility in a mink-derived isolate [12] and the demonstration of fomite and airborne transmission of the 2024 bovine-associated human isolate [11] represent a qualitative shift in the pandemic threat posed by these viruses. The CDC's Influenza Risk Assessment Tool (IRAT) now incorporates ferret transmission data as a core component, with recent assessments placing H5N1 clade 2.3.4.4b viruses in the "moderate to high" risk category for pandemic emergence.

The ability of ferrets to be infected with and transmit a diverse array of IAV subtypes, from seasonal H1N1pdm09 and H3N2 to HPAI H5Nx, H7N9, and swine-origin variants, makes them uniquely valuable for pandemic preparedness. However, this very susceptibility also renders domestic ferret populations vulnerable to spillover events from avian reservoirs, particularly given the dietary risks posed by raw poultry feeding. The 2023 Polish outbreak serves as a sentinel event, demonstrating that as HPAI H5N1 continues to circulate at unprecedented levels in wild bird populations, the interface with companion animals will require expanded surveillance and public health guidance to protect both animal and human health.

Diagnostics for Ferret Influenza A Virus Infection

The ferret model has become the gold standard for the study of influenza A virus (IAV) pathogenesis, transmission, and pandemic risk assessment. Its unique susceptibility to human-adapted and zoonotic IAV strains, driven by the exclusive expression of N-acetylneuraminic acid (Neu5Ac) on respiratory epithelial surfaces [4], necessitates a robust array of diagnostic approaches that mirror human clinical and laboratory methods. Diagnostics in this model serve multiple critical purposes: confirming infection, quantifying viral burden, characterizing viral genotypes and phenotypes, monitoring immune responses, and assessing transmissibility. The integration of molecular, virological, serological, and aerobiological techniques has evolved to capture the full spectrum of host–pathogen interactions, and these tools are increasingly standardized across laboratories to ensure comparability in global risk assessments [8, 51].

Sample Collection and Processing

The diagnostic workup in ferrets begins with meticulous sample collection, typically from the upper and lower respiratory tracts. Nasal washes – performed by instilling a sterile buffer (e.g., phosphate-buffered saline containing antibiotics) into the nares and collecting the effluent – are the most common ante-mortem specimen for quantifying viral shedding kinetics and infectious titers [22, 21]. These washes are collected serially over the course of infection, often at daily intervals, to monitor the rise and fall of viral load. The peak viral titer in nasal washes correlates positively with titers in nasal turbinate tissue and lung homogenates, though the strength of this association varies with the specific summary measure used (e.g., peak titer vs. area under the curve) [5]. In addition to washes, throat or nasal swabs are frequently employed, particularly in natural infection settings such as the 2023 outbreak of highly pathogenic avian influenza A(H5N1) in pet ferrets in Poland, where point-of-care antigen tests and RT-qPCR on throat swabs confirmed infection even in asymptomatic carriers [10]. For post-mortem diagnostics, systematic necropsy with collection of nasal turbinates, trachea, lung lobes, olfactory bulb, brain, and other extra-respiratory tissues (liver, pancreas, spleen, heart, intestine) is essential to evaluate viral tropism and extrapulmonary spread [9, 31, 2]. The selection of tissues should be guided by the virus subtype; for example, highly pathogenic H5N1 and H5N6 viruses routinely disseminate to the central nervous system and multiple viscera [9, 31, 32]. Samples are typically stored in viral transport medium at –80°C until processing.

Molecular Diagnostics: RT‑qPCR and Sequencing

Quantitative reverse transcription PCR (RT‑qPCR) targeting the matrix (M) gene or other conserved regions of the IAV genome is the primary molecular tool for detecting viral RNA in ferret specimens. This assay provides high sensitivity and allows for absolute quantification when coupled with standard curves of known copy numbers. In transmission studies, RT‑qPCR is used not only on nasal washes but also on aerosol filters to measure viral RNA emitted into the air, offering a non-invasive correlate of ferret-to-ferret transmissibility [22, 35]. The dynamics of viral RNA shedding in the air have been shown to differ significantly between transmissible and non-transmissible viruses; human seasonal and swine-origin strains exhibit rapid and robust emission, whereas most avian-origin H5N1 strains, including the 2024 bovine isolate A/Texas/37/2024, shed infectious virus into the air at lower levels [22, 35, 11]. This discrepancy underscores the value of RT‑qPCR as a screening tool for assessing pandemic potential.

Beyond detection, next-generation sequencing (NGS) of viral genomes from ferret specimens is increasingly employed to identify adaptive mutations acquired during mammalian passage. For instance, deep sequencing of nasal turbinate and olfactory bulb samples from H5N1-infected ferrets revealed positive selection of substitutions such as PB1 E177G and NP I119M in the brainstem, demonstrating the action of selective forces within the central nervous system [41]. Similarly, minor variants with enhanced polymerase activity (PA‑S321) or hemagglutinin stability (HA1‑S210) were selected during transmission of swine H1N1 viruses in ferrets, and these variants could be detected at low frequencies in the inoculum only through deep sequencing [14]. Such analyses are critical for risk assessment, as they identify viral populations with traits that may enhance human adaptation. The World Health Organization (WHO) and the World Organisation for Animal Health (WOAH) have recognized the importance of sequence-based surveillance for zoonotic influenza viruses, and ferret diagnostic data directly contribute to these global monitoring efforts.

Virus Isolation and Titration

Infectious virus titers are routinely measured by plaque assay or TCID50 on Madin–Darby canine kidney (MDCK) cells, using serial dilutions of nasal washes, tissue homogenates, or aerosol samples. This traditional virological method remains indispensable because RT‑qPCR does not distinguish between infectious and non-infectious particles; the detection of infectious virus in air samples, for example, is a more stringent indicator of transmission potential [35, 52]. The ferret model also allows for the isolation of virus from clinical specimens for antigenic characterization. Ferret antisera raised against reference viruses are used in hemagglutination inhibition (HI) and neuraminidase inhibition (NI) assays to monitor antigenic drift, a process that has been documented for both seasonal A(H1N1)pdm09 viruses and emerging H5Nx clades [26, 19]. However, it is notable that post-infection ferret antisera may not capture all antibody-specific epitopes recognized by humans or pigs, as demonstrated for certain immunodominant HA sites [53]; this limitation should be considered when interpreting surveillance data.

Serological Diagnostics

Serology in ferrets is performed primarily to assess immune responses post-infection or post-vaccination. The hemagglutination inhibition (HI) assay is the standard for measuring neutralizing antibodies against the HA head domain, with titers ≥40 typically considered protective in humans. In ferret studies, HI titers correlate inversely with disease severity after H5N1 challenge, and even undetectable HI titers can be associated with milder outcomes if non-neutralizing IgG antibodies and T-cell responses are present [54]. The role of antibodies against the neuraminidase (NA) protein is also increasingly recognized; NA inhibition assays using a substrate-based method (e.g., MUNANA) can detect antigenic drift in NA that may reduce vaccine effectiveness despite HA similarity [19]. Moreover, enzyme-linked immunosorbent assays (ELISAs) for ferret IgG, IgA, and interferon-gamma (IFN-γ) have been developed to capture humoral and cellular immune correlates [20, 54]. For instance, IFN-γ ELISpot assays on peripheral blood mononuclear cells or lung lymphocytes allow quantification of T-cell cross-reactivity across diverse IAV subtypes, including H1N1, H3N2, H5N1, and even influenza B [55, 56]. These serological diagnostics are essential for evaluating imprinting effects, where prior infection with a Group 1 HA virus (e.g., H1N1) confers superior protection against H5N1 compared with Group 2 imprinting (H3N2) [43, 48].

Antigen Detection and Point‑of‑Care Testing

In field and clinical veterinary settings, rapid antigen detection tests (RIDTs) targeting the influenza A nucleoprotein are available for ferrets. During the natural H5N1 outbreak in pet ferrets in Poland, such point-of-care tests performed on throat swabs from all five ferrets in a household correctly identified type A influenza antigens, even in the asymptomatic mother and another clinically normal adult [10]. While RIDTs are less sensitive than RT‑qPCR, they offer speed and ease of use, making them valuable for initial screening of suspected cases, especially when zoonotic transmission is a concern. The Centers for Disease Control and Prevention (CDC) and WOAH recommend confirmatory RT‑qPCR for all positive or inconclusive RIDT results during outbreak investigations.

Air Sampling Diagnostics

A recent advance in ferret diagnostics is the integration of continuous air sampling to measure the kinetics of infectious virus expulsion. Devices placed in aerobiology chambers or transmission hoods filter airborne particles, which are then eluted and assayed by both RT‑qPCR and MDCK titration [22, 35, 52]. This technique has revealed that the lack of airborne transmission for many avian H5N1 viruses is due to an absence of infectious virus shedding rather than a deficiency in mammalian adaptation mutations [35]. Furthermore, the use of size-selective particle separators (e.g., with a 50% cutoff at 5–8 µm aerodynamic diameter) allows differentiation between droplet and aerosol transmission routes [52]. Air sampling data are now considered a valuable adjunct to conventional transmission experiments, providing quantitative metrics that can be compared across studies and laboratories [22, 8].

Integration with Pandemic Risk Assessment

The diagnostic arsenal for ferret IAV infection directly underpins systematic risk assessment frameworks, such as the WHO’s Tool for Influenza Pandemic Risk Assessment (TIPRA). Key parameters including receptor-binding specificity, replication kinetics in nasal epithelium [15], and airborne transmissibility are derived from ferret diagnostic data. The observation that viral titers in human bronchial epithelial cells (Calu‑3) correlate with nasal wash and turbinate titers in ferrets [16] allows researchers to partially substitute in vitro diagnostics for in vivo experiments, reducing animal use while maintaining predictive power. Moreover, the routine inclusion of competitive mixture models – where wild-type and drug-resistant viruses are co-administered – enables assessment of within-host and between-host fitness through differential diagnostic readouts (e.g., relative titers in nasal washes and transmission frequencies) [25, 57]. These diagnostics collectively support evidence-based public health decisions regarding antiviral stockpiling, vaccine strain selection, and containment strategies for emerging zoonotic influenza viruses.

Transmission Dynamics and Pathogenicity in the Ferret Model

The ferret (Mustela putorius furo) stands as the preeminent small mammalian model for investigating influenza A virus (IAV) transmission dynamics and pathogenicity, owing to its remarkable physiological and anatomical parallels with human respiratory disease. Unlike murine models, ferrets exhibit clinical signs of influenza that closely mirror human illness, including fever, sneezing, nasal discharge, lethargy, and weight loss, and they are naturally susceptible to human, swine, and avian IAV strains without requiring prior adaptation [1, 2]. This susceptibility is underpinned by the distribution of both α2,6-linked and α2,3-linked sialic acid receptors throughout the ferret respiratory tract, which permits the attachment and entry of viruses with diverse host tropisms [3, 4]. Critically, ferrets exclusively synthesize N-acetylneuraminic acid (Neu5Ac) due to an ancient, nine-exon deletion in the CMAH gene, rendering them functionally similar to humans, who also lack N-glycolylneuraminic acid (Neu5Gc) [4]. This naturally humanized receptor landscape is a foundational determinant of ferret permissiveness to human-adapted IAV and contributes directly to the translational validity of data derived from this model.

The World Health Organization (WHO) and the U.S. Centers for Disease Control and Prevention (CDC) have incorporated ferret-based risk assessment into their pandemic preparedness frameworks, particularly through the WHO Tool for Influenza Pandemic Risk Assessment (TIPRA) and the CDC Influenza Risk Assessment Tool (IRAT). These frameworks rely heavily on ferret transmission experiments to gauge the human-to-human transmission potential of emerging zoonotic IAV [13, 51]. The ferret model thus serves as a critical bridge between in vitro assays and human epidemiological reality, providing actionable data for public health decision-making.

Pathogenicity Profiles: From Mild Respiratory Infection to Systemic Lethal Disease

Pathogenicity in the ferret model is assessed along a multidimensional continuum that includes clinical scoring (weight loss, temperature elevation, activity level), viral replication kinetics in the upper and lower respiratory tract, histopathological lesion severity, and the extent of extrapulmonary dissemination. Seasonal human IAV strains, such as A(H1N1)pdm09 and A(H3N2), typically induce mild to moderate, self-limiting upper respiratory tract infection in ferrets, characterized by peak nasal wash titers of 10⁴–10⁶ TCID₅₀/mL occurring between days 2 and 4 post-inoculation, followed by rapid viral clearance within 7–10 days [23, 5, 21]. Lower respiratory tract involvement is generally minimal in these infections, with virus recovery from lung tissue being sporadic and of low magnitude [5, 58].

In stark contrast, highly pathogenic avian influenza (HPAI) H5N1 and H5N6 viruses of the A/goose/Guangdong/1/96 lineage induce a profoundly different disease spectrum. Infection with clade 2.3.4.4b viruses, such as A/Chile/25945/2023 [9] and A/Texas/37/2024 [11], produces rapid, severe, and frequently fatal disease in ferrets. These viruses exhibit a remarkable capacity for systemic dissemination, with infectious virus detected not only in respiratory tissues, nasal turbinates, trachea, and lungs, but also in the brain, liver, spleen, pancreas, adrenal glands, and gastrointestinal tract [9, 31, 11]. The 2024 Texas isolate, derived from a dairy farm worker, demonstrated robust viremia and extrapulmonary spread, causing severe morbidity and mortality in all inoculated animals [11]. Similarly, the 2022 North American clade 2.3.4.4b virus evaluated by Pulit-Penaloza et al. induced systemic infection and was capable of transmitting to co-housed contacts, where it caused equally severe disease [33]. This capacity for systemic invasion is not uniformly distributed among HPAI viruses; the H5N6 virus A/black-headed gull/Netherlands/29/2017 caused severe disease with widespread extrapulmonary lesions, including the brain, liver, and pancreas, but did not transmit via the airborne route [31]. Sequence analysis of viruses recovered from infected ferrets frequently reveals positive selection for mammalian adaptation markers, most notably the PB2 D701N substitution, which emerged in nearly all ferrets infected with the Dutch H5N6 virus, indicating strong in vivo selection pressure for enhanced replication in mammalian hosts [31].

The neurological tropism of HPAI viruses is of particular concern. Siegers et al. demonstrated that A/Indonesia/5/2005 (H5N1) virus evolves within the central nervous system (CNS) of infected ferrets, with specific substitutions, PB1 E177G and A652T, and NP I119M, emerging in the brainstem and olfactory bulb [41]. These substitutions conferred increased polymerase activity in vitro and were associated with severe meningoencephalitis. Notably, the viral populations accessing the CNS via the olfactory route did not pass through a detectable genetic bottleneck, yet displayed signs of positive selection in the brainstem, suggesting that the CNS microenvironment actively selects for variants with enhanced neurotropism [41]. This finding has profound implications for understanding the pathogenesis of human H5N1 infections, where CNS complications are disproportionately reported compared to seasonal influenza.

Lower pathogenicity avian influenza viruses, such as H7N9 and H9N2, display an intermediate phenotype. The A/Anhui/1/2013 (H7N9) virus, for instance, replicates efficiently in both the upper and lower respiratory tract of ferrets, with peak shedding at 3–5 days post-inoculation, and can be recovered from the olfactory bulb, heart, and liver [28]. However, it does not typically cause the rapid systemic lethality observed with H5N1 viruses, and transmission via respiratory droplets is inefficient or absent [13, 59].

Transmission Dynamics: Direct Contact, Fomite, and Airborne Routes

The ferret model is uniquely suited to dissect the three principal modes of IAV transmission: direct contact, indirect contact (fomite), and airborne (respiratory droplet/aerosol) transmission. Each mode is evaluated using distinct experimental configurations. Direct contact is assessed by co-housing an inoculated donor ferret with one or more naïve sentinels, allowing unrestricted physical interaction. Airborne transmission is evaluated using either side-by-side cages with perforated separators that prevent direct contact but permit airflow, or specialized aerosol chambers that can be equipped with particle size separators to differentiate between droplet (>5 µm) and true aerosol (<5 µm) transmission [13, 52].

Seasonal human IAV strains, including A(H1N1)pdm09 and A(H3N2), transmit with high efficiency via the airborne route in ferrets. Donor animals shed infectious virus into the air at levels sufficient to infect sentinels placed in adjacent cages, with transmission typically occurring within 1–3 days of exposure [23, 13, 6]. Critically, Roberts et al. demonstrated that transmission of A(H1N1)pdm09 occurs before the onset of fever and is temporally correlated with peak viral titers in the nasal wash rather than with clinical signs such as sneezing or coughing [6]. This pre-symptomatic transmission has profound implications for pandemic containment strategies, as infected individuals may be contagious before they become clinically recognizable.

The dynamics of airborne shedding are now quantifiable through the use of air-sampling devices that continuously capture infectious virus expelled by infected ferrets. Tosheva et al. demonstrated that A(H1N1)pdm09 virus is efficiently shed into the air, with infectious particles detected at high levels. In contrast, earlier zoonotic A(H5N1) viruses from 2005 and a 2024 bovine isolate were not detected in air samples, despite robust replication in the nasal epithelium [35]. This absence of airborne shedding, rather than a lack of mammalian adaptation mutations per se, appears to be the primary barrier to airborne transmission for these strains. However, a 2022 European polecat A(H5N1) isolate and the 2024 dairy farm worker isolate (A/Texas/37/2024) both shed infectious virus into the air, albeit at lower levels than seasonal human viruses [35, 11]. The Texas isolate transmitted efficiently in a direct contact setting (100% of contacts infected) and via indirect fomite exposure, but airborne transmission was less efficient than that observed for human-adapted H1N1 strains [11]. These findings underscore a critical threshold: viruses must achieve a minimum level of airborne shedding to sustain human-to-human transmission.

Pulit-Penaloza et al. extended these observations by measuring both viral RNA in nasal washes and viral RNA emitted into the air for 14 diverse IAV strains, encompassing human, swine, and avian origins. They found that transmissible viruses, including human seasonal and swine-lineage strains, exhibited robust replication kinetics and rapid release of viral RNA into the air within the first 24–48 hours post-inoculation. Poorly or non-transmissible viruses, including many avian strains, showed significantly reduced or delayed replication and lower airborne viral RNA at early time points [22]. This study established that efficient ferret-to-ferret transmission via the air is directly associated with fast emission of virus-laden particles, making quantification of airborne viral RNA a powerful adjunct to traditional transmission assessments.

The H5N1 clade 2.3.4.4b virus isolated from mink (A/mink/Spain/2023) represents a notable exception to the general rule that avian H5N1 viruses lack airborne transmissibility. Restori et al. reported that this isolate transmitted via direct contact to 75% of exposed ferrets and, importantly, via the airborne route to 37.5% of contacts, the first documented instance of an H5N1 clade 2.3.4.4b virus exhibiting airborne transmissibility in ferrets [12]. Sequence analysis revealed that the virus carried the mammalian adaptation mutation PB2 T271A, and reversion of this substitution abrogated both mortality and airborne transmission. This finding carries enormous significance for pandemic risk assessment, as it indicates that a single amino acid change in a mammalian host can dramatically alter the transmission phenotype of a panzootic H5N1 virus.

The Upper Respiratory Tract as the Locus of Airborne Transmission

A seminal investigation by Richard et al. employed genetically tagged viruses to unequivocally demonstrate that airborne transmission of IAV in ferrets occurs from the upper respiratory tract. Ferrets were simultaneously inoculated intranasally (upper respiratory tract) and intratracheally (lower respiratory tract) with distinguishable but otherwise identical viruses. In every transmission event, the virus recovered from contact ferrets matched the genotype of the intranasally inoculated virus, establishing that the virus is expelled from the nasal respiratory epithelium rather than from the trachea or lower airways [15]. Moreover, viruses that are transmissible via the air preferentially infect ferret and human nasal respiratory epithelium, correlating with high replicative capacity in these tissues. This finding refocuses attention on the nasal epithelium as the critical anatomical site for both shedding and acquisition of airborne IAV, and it has direct implications for the design of vaccines and therapeutics intended to block transmission: these interventions must achieve robust protection at the nasal mucosa.

Viral Interference and Coinfection Dynamics

The ferret model has also illuminated the phenomenon of viral interference, whereby prior infection with one virus can limit or prevent infection with a second virus. Chan et al. showed that infection with influenza A virus can prevent or limit subsequent infection with human respiratory syncytial virus (hRSV), while hRSV infection reduces morbidity associated with subsequent influenza virus infection [17]. The interval between infections and the specific combination of viruses are critical determinants: viral interference is most pronounced when the interval between primary infection and secondary challenge is less than one week [40]. Furthermore, influenza viruses appear to adhere to a hierarchical order, where some viruses (e.g., A(H1N1)pdm09) can dominate over others (e.g., A(H3N2)) when introduced sequentially, a phenomenon that may contribute to the observed dominance of specific subtypes within a given season [40]. Coinfections are possible when the interval between infections is short (1–3 days), but ongoing shedding from the primary infection is associated with interference against the secondary challenge [40].

Immune Imprinting and Its Modification of Pathogenicity and Transmission

Immunological imprinting, the lasting influence of an individual's first influenza virus infection on subsequent immune responses, is a powerful modulator of disease outcome in the ferret model. Nuñez et al. established a ferret pre-immune model in which animals were first infected with seasonal H1N1, H2N3, or H3N2 viruses, allowed to recover for 84 days, and then challenged with HPAI H5N1 virus. Ferrets imprinted with Group 1 hemagglutinin viruses (H1N1, H2N3) were completely protected against lethal H5N1 challenge, exhibiting 100% survival with minimal clinical signs. In contrast, Group 2-imprinted ferrets (H3N2) experienced severe clinical disease, delayed progression, and 40% mortality [43]. Consecutive infections with H1N1 followed by H3N2 did not abrogate the protection conferred by the original H1N1 imprint, suggesting that the first infection establishes a dominant immunological memory that is not easily overwritten. Critically, vaccination of H3N2-imprinted ferrets with a broadly reactive H5 HA-based vaccine was able to redirect the immune response and rescue animals from the sublethal phenotype to complete protection [43].

This imprinting effect has profound implications for understanding the age-specific severity of zoonotic IAV infections in humans. Individuals born before 1968, who were primarily imprinted with Group 1 H1N1 or H2N2 viruses, may possess a degree of cross-protective immunity against Group 1 H5N1 viruses, whereas younger individuals imprinted with Group 2 H3N2 viruses may be at greater risk of severe disease. The ferret model thus provides a mechanistic basis for epidemiological observations of age-related susceptibility to H5N1.

Antiviral Resistance, Fitness, and Transmission

The ferret model is indispensable for evaluating the fitness and transmission potential of antiviral-resistant influenza viruses. Baloxavir marboxil, a cap-dependent endonuclease inhibitor targeting the PA protein, has proven highly effective against IAV, but resistance-associated substitutions at PA residue 38 (I38T/F/M) have been detected in treated patients. Stannard et al. and Jones et al. independently assessed the fitness of these resistant variants. I38T/F/M substitutions impair PA endonuclease activity and, in some backgrounds, reduce replication efficiency in vitro. However, despite these fitness costs, I38T- and I38M-containing A(H1N1)pdm09 and A(H3N2) viruses retained the ability to transmit between ferrets via both direct contact and airborne routes [23, 24]. Notably, the dual-resistant isolate (NA-H275Y plus PA-I38T) from an immunocompromised patient showed reduced replicative fitness in the ferret upper respiratory tract compared to wild-type virus, and was outcompeted in co-infection experiments, suggesting that widespread community transmission of this dual mutant is unlikely [23]. However, the stable retention of I38T/F/M during transmission, with minimal reversion to wild-type, indicates that once selected, these resistant variants could persist and spread within a population [24].

Lee et al. further demonstrated that baloxavir treatment of infected donor ferrets significantly reduces infectious viral shedding in the upper respiratory tract and decreases the frequency of onward transmission to sentinels, even when treatment is delayed until 48 hours post-infection [60]. In contrast, oseltamivir treatment did not substantially affect shedding or transmission under the same conditions. Combination therapy with baloxavir plus oseltamivir was shown to reduce the selection of viruses with reduced baloxavir susceptibility compared to baloxavir monotherapy, and importantly, prevented the de novo emergence of oselt

Immunology and Cross-Protection from Prior Influenza A Infection

The immune response to influenza A virus (IAV) in the ferret model is a profoundly complex interplay of innate, humoral, and cellular mechanisms, shaped decisively by the host’s history of prior infections. This immunological memory, often termed “immune imprinting” or “original antigenic sin,” dictates the trajectory of subsequent infections with heterologous or heterosubtypic strains, a phenomenon of paramount importance for pandemic risk assessment. The ferret, owing to its human-like sialic acid receptor distribution (Neu5Ac exclusively) and comparable clinical symptomatology, provides an unparalleled system for dissecting these intricate cross-protective pathways [1, 4]. Central to this understanding is the recognition that prior IAV exposure can either confer substantial protection, induce minimal impact, or, in some contexts, paradoxically skew the immune response, a concept that must be rigorously defined using the controlled experimental paradigms available in this model.

The Conceptual Framework of Immune Imprinting and Original Antigenic Sin in Ferrets

The initial encounter with influenza virus during childhood indelibly shapes the repertoire of B-cell and T-cell memory. In the ferret model, this imprinting has been systematically recapitulated by exposing naïve animals to seasonal IAV subtypes, allowing recovery, and subsequently challenging with antigenically distinct viruses. Nuñez et al. [43] provided a seminal demonstration of this phenomenon, showing that ferrets imprinted with Group 1 hemagglutinin (HA) viruses (specifically H1N1 or H2N3) exhibited 100% survival and minimal clinical signs upon lethal challenge with a highly pathogenic avian influenza (HPAI) H5N1 virus. In stark contrast, ferrets imprinted with a Group 2 HA virus (H3N2) suffered severe clinical disease and 40% mortality. This stark dichotomy underscores that the HA group of the imprinting strain is a primary determinant of cross-protection, likely due to the structural conservation of the HA stalk domain within a group. The study further demonstrated that sequential imprinting, first with H1N1 and then with H3N2, did not abrogate the protective benefit conferred by the initial Group 1 exposure, suggesting a hierarchical dominance of the first encountered virus in shaping the protective antibody landscape [43]. This aligns conceptually with the phenomenon observed in human populations where individuals born during H1N1-circulating eras have shown lower mortality during H5N1 outbreaks.

The mechanistic underpinnings of this imprinting effect are multifaceted. Skarlupka and Ross [48] have extensively characterized the ferret pre-immune model, emphasizing that the specificity and avidity of the antibody response post-challenge are heavily biased towards conserved epitopes of the original infecting strain. This can be a double-edged sword: while it can provide robust heterosubtypic protection against viruses sharing conserved stalk epitopes, it can also lead to a suboptimal response against a novel virus where the HA head is entirely unfamiliar. The work of Francis et al. [61] extends this to vaccination, demonstrating that H1N1-imprinted ferrets mounted a more robust and sustained antibody response to subsequent vaccination compared to naïve ferrets, with higher virus-specific IgG levels and greater virus neutralization activity. This indicates that imprinting not only affects the outcome of natural infection but also fundamentally alters the host’s capacity to respond to vaccine antigens, a critical consideration for universal vaccine development.

Humoral Correlates of Cross-Protection: The Role of Stalk and Neuraminidase Antibodies

While neutralizing antibodies directed against the immunodominant HA head are highly protective against homologous viruses, cross-protection against heterologous strains is largely mediated by antibodies targeting more conserved regions. The HA stalk domain has emerged as a key target. Sun et al. [29] provided direct evidence of this in a ferret model where prior infection with A(H1N1)pdm09 virus significantly reduced replication and transmission of a clade 2.3.4.4b HPAI H5N1 virus. This protection was specifically correlated with the presence of cross-reactive group 1 HA stalk antibodies and N1 neuraminidase (NA) antibodies in the pre-immune ferrets. Notably, the same prior infection offered less robust protection against a group 2 H7N9 virus, confirming the group-specific nature of stalk-based humoral immunity [29]. This study is a powerful example of how existing immunity from seasonal circulation can partially mitigate the threat posed by a novel zoonotic virus.

The contribution of NA-specific antibodies to cross-protection is increasingly recognized as a critical and often overlooked component. Rosu et al. [47] systematically dissected this in ferrets using vaccine formulations with matched and mismatched HA and NA components. They demonstrated that vaccination with an NA-matched but HA-mismatched vaccine (vacH1N2) provided substantial protection against H3N2 challenge, including reduced fever, weight loss, and elimination of lower respiratory tract virus replication, comparable to the protection afforded by a fully homologous vaccine. This highlights the ability of NA antibodies to mediate protection independently, potentially compensating for a drifted HA in seasonal vaccine mismatches. The study by Gao et al. [19] further underscores the importance of monitoring NA antigenic drift, showing that while ferret antisera against older H1N1pdm09 strains could still inhibit the NA of newer strains in vitro, they were less effective in passive transfer protection experiments in mice, indicating a functional drift that reduces in vivo efficacy. Thus, comprehensive assessment of cross-protection in the ferret model must evaluate both HA and NA antibody responses.

The development of broadly protective vaccines, such as those targeting the conserved HA stalk, has been extensively validated in the ferret model. Sequential immunization with chimeric hemagglutinin (cHA)-based vaccines, which feature exotic head domains and a conserved stalk, successfully redirects the immune response away from the variable head and towards the stalk [44, 45, 46]. Liu et al. [45] demonstrated that a live-attenuated influenza virus (LAIV) prime followed by an LAIV boost (LAIV-LAIV) regimen using cHA vaccines provided the most durable and robust protection against heterologous H1N1 and H6N1 challenge, correlating with strong stalk-specific antibody and CD4⁺ and CD8⁺ T-cell responses. This platform, moving towards clinical trials, exemplifies how mechanistic insights from the ferret model directly inform human vaccine design [44].

Cellular Cross-Protection: T-Cell Mediated Immunity and its Correlates

Beyond humoral immunity, cross-reactive T cells represent a second pillar of heterosubtypic protection, targeting highly conserved internal proteins such as nucleoprotein (NP), matrix protein 1 (M1), and polymerase subunits. Gooch et al. [55] established a direct correlation between heterosubtypic cross-protection and the presence of cross-reactive interferon-gamma (IFN-γ)-secreting T cells in the ferret model. Following sublethal H1N1 infection, ferrets challenged four weeks later with H3N2 virus showed significantly reduced clinical disease and shortened virus shedding duration, despite the absence of cross-reactive neutralizing antibodies. The protection was temporally correlated with the peak of circulating IFN-γ⁺ T cells at day 11 post-primary infection, and these cells were rapidly recalled upon challenge [55]. This study provides compelling evidence that cellular immunity alone can mitigate disease severity, independent of antibody.

Reber et al. [56] further characterized the breadth of T-cell cross-reactivity, showing that peripheral blood T cells from ferrets recovered from either H1N1pdm09 or H3N2 infection responded robustly to a wide panel of historical and contemporary IAV strains, as well as to influenza B viruses. The most dominant and cross-reactive responses targeted peptides from the NP protein, stimulating both CD4⁺ and CD8⁺ T-cell populations. This extensive cross-reactivity persisted in the spleen even after peripheral blood responses waned, suggesting a durable reservoir of memory T cells [56]. The differential cellular response in the lung is also noteworthy. Ryan et al. [20] found that while peripheral blood IFN-γ responses were comparable between H1N1 and H3N2 infections in ferrets, the IFN-γ response in lung lymphocytes was significantly different between the two subtypes. This suggests that the quality and anatomical localization of the T-cell response are subtype-specific and may influence the outcome of heterologous challenge.

The role of innate immune mechanisms in modulating cross-protection should not be neglected. Viral interference, the phenomenon whereby an initial viral infection temporarily blocks or limits a subsequent infection, has been systematically studied in ferrets by Laurie et al. [40]. They demonstrated that infection with one IAV subtype could prevent or delay infection with a second, antigenically unrelated virus, but only when the interval between infections was short (less than one week). This effect was virus-specific and hierarchical, with A(H1N1)pdm09 appearing to dominate over A(H3N2) and influenza B. This interference is likely mediated by the initial virus’s stimulation of a broad antiviral state, including type I and III interferons and interferon-stimulated genes (ISGs) like IFITMs, which can broadly restrict viral entry [17, 38]. The timing and viral identity are crucial; understanding this hierarchy informs epidemiological models of viral dominance during co-circulating seasons.

Determinants of Cross-Protection: Interval, Viral Hierarchy, and Antigenic Distance

The efficacy of cross-protection is not a binary phenomenon but is modulated by several critical variables. The interval between primary and secondary infection is a key determinant, as shown by Laurie et al. [40]. Protection via viral interference is short-lived (days), while adaptive immune memory (antibody and T-cell) takes weeks to mature and can last for months to years. The identity of the viruses involved also dictates a clear hierarchy. Pre-existing immunity to A(H1N1)pdm09, as demonstrated by Sun et al. [29], provides a selective advantage against group 1 H5N1 viruses but not group 2 H7N9 viruses. However, this protection is not absolute. Pulit-Penaloza et al. [27] showed that ferrets with pre-existing immunity to A/California/07/2009 (H1N1pdm09) were not fully protected against challenge with a antigenically drifted H1N1 variant virus (A/Ohio/09/2015). The single amino acid substitution (G155E) in the HA of the variant virus allowed it to escape neutralization by pre-existing antibodies, demonstrating that antigenic distance within the same subtype can overcome cross-protection.

The functional avidity and specificity of the recalled immune response are paramount. In the context of H5N1, the presence of stalk antibodies is protective, but the response can be suboptimal if the host’s immune system is focused on an irrelevant head domain. The work of Nachbagauer et al. [44] and Liu et al. [45] demonstrates that overcoming this requires deliberate antigen design to refocus the immune system. Furthermore, the adjuvant used can significantly alter the quality of the response. Wong et al. [54] found that while hemagglutination inhibition (HAI) titers were the best correlate of protection against H5N1 in ferrets, non-neutralizing antibodies and cellular responses also contributed, particularly in animals with low HAI titers. This nuanced understanding, derived from the ferret model, is critical for defining immune correlates for licensure of vaccines against pandemic-potential viruses where clinical efficacy data in humans may be unavailable. The robustness of these findings is underscored by multi-laboratory validation exercises, such as the one conducted by Belser et al. [51], which confirmed the reproducibility and concordance of transmission and pathogenicity data in the ferret model, establishing it as a reliable tool for assessing cross-protective immunity.

In Vitro and In Vivo Correlates for Risk Assessment of Ferret Influenza A Virus

The ferret model remains the preeminent small mammalian system for evaluating the pandemic potential of emerging and re-emerging influenza A viruses (IAV), providing a physiologically relevant platform that recapitulates critical features of human infection, including clinical symptomology, respiratory tract receptor distribution, and transmission dynamics [1, 7]. However, the utility of this model is fundamentally enhanced when integrated with robust in vitro methodologies, as the systematic examination of how viral titer measurements obtained in cell culture align with results from in vivo experimentation provides a rigorous foundation for risk assessment [16]. The correlation between these two experimental paradigms is not merely a matter of convenience but represents a critical, evidence-based framework for contextualizing the threat posed by novel IAV, including those of high-pathogenicity avian influenza (HPAI) H5N1, swine-origin variant viruses, and human seasonal strains. The World Health Organization (WHO) has developed a Tool for Influenza Pandemic Risk Assessment (TIPRA) that explicitly incorporates parameters such as receptor-binding specificity, replication in human airway epithelial cells, and transmission in animal models, underscoring the necessity of understanding in vitro–in vivo correlates for accurate risk stratification [13, 8].

Viral Replication Kinetics as Foundational Correlates

The most direct and extensively characterized correlate between in vitro and in vivo systems is the measurement of viral replication kinetics. A landmark meta-analysis by Creager and colleagues systemically compared viral titers from over 50 human and zoonotic IAV tested concurrently in human bronchial epithelial (Calu-3) cells and in ferrets, demonstrating a statistically significant positive correlation between peak viral titers in Calu-3 cells and those recovered from ferret nasal wash specimens and nasal turbinate tissue [16]. This finding was further validated by Kieran and colleagues, who compiled viral titer data from over 1,000 ferrets inoculated with 125 contemporary IAV, revealing that viral titers in nasal turbinate, lung tissue, and nasal washes were positively correlated with one another, with the strength of these associations influenced by the specific nasal wash summary measure employed and the intrinsic properties of the virus [5]. These analyses indicate that the replicative capacity of a virus in a standardized in vitro system, particularly a human respiratory epithelial cell line, serves as a robust, albeit not absolute, predictor of its growth kinetics in the ferret upper respiratory tract.

More nuanced insights emerge from the application of mathematical modeling to raw in vitro viral titers. By estimating generalized replication kinetic parameters (such as growth rate, peak titer, and decay rate), researchers have identified commonalities in infection progression between in vitro and in vivo systems that are not apparent from simple endpoint titers [16]. For instance, IAV that possess mammalian host adaptation markers in the hemagglutinin (HA) and polymerase basic 2 (PB2) proteins, such as the PB2 E627K or D701N substitutions, exhibit more rapid growth in the ferret upper respiratory tract early after infection, a phenotype that is mirrored in their replication dynamics in Calu-3 cells [5]. These findings are particularly relevant for risk assessment, as they allow for the identification of viruses that, even if they do not achieve exceptionally high peak titers, may demonstrate a kinetic advantage that facilitates early and robust shedding, a critical determinant of transmissibility.

Transmission Phenotypes and Airborne Shedding Correlates

The assessment of transmissibility in ferrets is arguably the most consequential component of pandemic risk assessment, and it is here that in vitro–in vivo correlates have proven both powerful and revealing. While direct contact transmission is relatively permissive, airborne transmission via respiratory droplets or aerosols is the gold standard for gauging human-to-human spread potential [8]. Critically, robust viral replication in the upper respiratory tract, specifically within the nasal respiratory epithelium, has been identified as a primary driver of airborne transmission [15]. This observation is directly supported by in vitro studies using primary differentiated ferret nasal epithelial cells (FNECs), which recapitulate the cellular composition, mucociliary clearance system, and sialic acid receptor distribution of the nasal epithelium [3]. Differentiated FNEC cultures demonstrate that both α2,6-linked (human-type) and α2,3-linked (avian-type) sialic acid receptors are present on the apical surface, with distinct cellular tropisms: human-adapted IAV primarily infect ciliated cells, whereas avian IAV such as HPAI H5N1 predominantly infect nonciliated cells [3]. Furthermore, the replication of an HPAI H5N1 virus in FNECs was significantly attenuated at 33°C (the temperature of the upper respiratory tract) compared to 37°C, while a seasonal H1N1 virus replicated efficiently at both temperatures [3]. This temperature-dependent restriction of avian IAV in nasal epithelial cells provides a mechanistic correlate for their generally poor airborne transmissibility in ferrets and humans.

The relationship between in vitro replication and in vivo transmission is further refined by direct measurements of viral RNA shedding into the air. Pulit-Penaloza and colleagues demonstrated that efficiently transmissible viruses, including human seasonal and swine-origin strains, exhibit robust replication and rapid release of viral RNA into the air of an aerobiology chamber, whereas poorly transmissible or non-transmissible viruses show significantly reduced or delayed airborne shedding at early time points post-inoculation [22]. The kinetics of this shedding are critical: airborne transmission correlates not with the overall magnitude of shedding but with the speed and efficiency of virus emission during the first 24–48 hours post-infection [22, 6]. This temporal window is precisely when the virus is replicating most actively in the nasal epithelium, reinforcing the in vitro–in vivo link. Air-sampling studies using HPAI H5N1 viruses have further corroborated these findings, showing that earlier zoonotic H5N1 strains (e.g., from 2005) shed no detectable infectious virus into the air, while more recent isolates from mammals, including a 2024 virus from a dairy farm worker (A/Texas/37/2024), were efficiently expelled into the air, albeit at lower levels than human seasonal H1N1 [35]. This incremental acquisition of airborne shedding capacity, detectable only through sophisticated in vivo aerosol monitoring, represents a critical risk correlate that would be missed by in vitro assays alone.

Molecular Determinants and Phenotypic Correlates of Pathogenicity

Beyond replication kinetics, specific molecular determinants identified through in vitro assays serve as powerful correlates of in vivo pathogenicity. The presence of mammalian adaptation markers in the HA, PB2, and other gene segments is routinely used to triage viruses for ferret studies. For example, the PB2 T271A substitution, which enhances polymerase activity in mammalian cells, was found to be critical for the airborne transmissibility and mortality of a mink-derived H5N1 clade 2.3.4.4b virus in ferrets; reversion of this mutation reduced both mortality and airborne transmission [12]. Similarly, the acquisition of the PB2 D701N substitution was positively selected in ferrets infected with an H5N6 virus from the Netherlands, correlating with increased virus replication but not necessarily with airborne transmission [31]. The HA acid stability, a property that can be measured in vitro by exposing viruses to low pH and assessing their ability to retain receptor-binding activity, is another key correlate. Hu and colleagues demonstrated that HA stabilization played a more prominent role than polymerase enhancement in promoting the replication and transmission of swine H1N1 isolates in ferrets, with HA-stabilizing variants selected within days of inoculation [14]. These observations underscore that while in vitro assays can identify individual molecular traits, the interplay of multiple traits, often in a virus- and subtype-specific manner, determines the in vivo phenotype.

The capacity for extrapulmonary spread and systemic infection, a hallmark of highly pathogenic IAV, is also informed by in vitro data. HPAI H5N1 and H5N6 viruses that cause severe, fatal disease in ferrets are characterized by their ability to replicate not only in respiratory tissues but also in brain, liver, pancreas, spleen, and adrenal glands [9, 31]. Siegers and colleagues identified specific substitutions in the PB1 (E177G, A652T) and NP (I119M) genes of an H5N1 virus that emerged in the central nervous system (CNS) of infected ferrets, demonstrating increased polymerase activity in vitro [41]. These CNS-associated mutations were selected in the brainstem under positive selection, indicating that viral evolution within a ferret can produce variants with enhanced neurotropism that are detectable through in vitro polymerase assays [41]. This capacity for CNS invasion is not merely an academic concern; infection with an H5N1 virus isolated from a human case in Chile (A/Chile/25945/2023) led to high morbidity and extrapulmonary spread in ferrets, including infection of the CNS, a finding that directly informed the risk assessment of this novel reassortant virus by the WHO and CDC [9].

Immune Correlates and the Role of Pre-Existing Immunity

The concept of immune imprinting, whereby an individual's first influenza virus infection shapes their lifelong immune responses to subsequent exposures, is a critical variable in risk assessment that can be modeled in the ferret. Sun and colleagues demonstrated that ferrets previously infected with an A(H1N1)pdm09 virus exhibited significantly reduced replication and transmission of a subsequent HPAI H5N1 clade 2.3.4.4b virus, an effect mediated by cross-reactive group 1 HA stalk antibodies and N1 neuraminidase antibodies [29]. This in vivo protection had a clear in vitro correlate: sera from H1N1pdm09-immunized ferrets showed cross-reactive binding to the H5 HA, providing a mechanistic explanation for the reduced pathogenesis [29]. Nuñez and colleagues further defined this phenomenon, showing that ferrets imprinted with group 1 HA viruses (H1N1, H2N3) were completely protected against lethal H5N1 challenge (100% survival), while those imprinted with group 2 HA viruses (H3N2) experienced 40% mortality [43]. These studies established that the in vitro measurement of cross-reactive antibody titers, specifically against the HA stalk, is a robust predictor of in vivo protection against avian IAV, a finding with direct implications for vaccine strain selection and pandemic preparedness.

The cellular arm of the immune response also provides important correlates. Gooch and colleagues showed that heterosubtypic protection in ferrets, induced by prior low-dose H1N1 infection against H3N2 challenge, correlated not with cross-reactive neutralizing antibodies (which were absent) but rather with the presence of cross-reactive interferon-gamma (IFN-γ) secreting T cells in the circulation [55]. This T-cell response, which peaked at 11 days post-infection and was strongly recalled upon challenge, provided a reduction in virus shedding duration and clinical disease, even though it did not prevent infection [55]. The ability to measure ferret IFN-γ responses using enzyme-linked immunospot (ELISpot) and ELISA assays [20] has thus become a valuable in vitro correlate for assessing vaccine-induced cellular immunity and the potential for cross-protection against antigenically distinct IAV. Furthermore, the differential interferon responses observed between IAV and influenza B virus (IBV) in primary ferret nasal epithelial cells, with IBV eliciting delayed and reduced type-I/II/III interferon responses and downregulation of thymic stromal lymphopoietin (TSLP), provide a mechanistic correlate for the weaker antibody responses typically observed after IBV infection in ferrets [18].

Refining Risk Assessment Through Standardized Protocols and Quantitative Metrics

The robustness of in vitro–in vivo correlates depends critically on the standardization of experimental protocols. An international exercise involving 11 laboratories, coordinated by Belser and colleagues, demonstrated that while overall transmission outcomes for two H1N1 viruses were concordant across sites, specific parameters, including donor-to-contact airflow directionality, could significantly influence transmission rates [51]. This study underscored that aggregating data from diverse laboratories is feasible, but only when key variables such as inoculation dose (e.g., 10^6 pfu versus 10^2 pfu), route (intranasal versus intratracheal), and housing conditions are controlled [21, 51]. Low-dose challenge models, for instance, have been shown to result in a more authentic disease course that parallels human infection, with prolonged virus shedding and greater sensitivity to antiviral interventions, such as oseltamivir [21]. These refinements to the ferret model enhance the translational value of in vitro data, as they allow for more precise comparisons between the replicative fitness of viruses in cell culture and their behavior in an animal model that more closely mimics natural infection.

Quantitative metrics derived from ferret studies, such as the area under the curve (AUC) for nasal wash titers, the time to peak titer, and the duration of shedding, have been systematically correlated with in vitro measurements. Kieran and colleagues used correlation coefficients and mediation analyses to support the interconnectedness of viral titer measurements across different respiratory tract sites, identifying summary metrics most closely linked with virological and phenotypic outcomes [5]. These analyses have practical implications: for instance, peak viral titer in Calu-3 cells was found to be a better predictor of peak titer in ferret nasal washes than were other molecular markers, such as the presence of a multi-basic cleavage site in HA [16]. However, the same study cautioned that additional phenotypic and molecular determinants of virulence and transmissibility varied in their association with in vitro measurements, indicating that no single in vitro assay can replace the holistic assessment provided by the ferret model [16]. The integration of mathematical modeling to estimate replication kinetic parameters from raw in vitro titers provides a promising avenue for generating more generalizable correlates, potentially allowing for the identification of viruses that, while exhibiting moderate peak titers, possess the dynamic growth characteristics necessary for efficient airborne transmission [16, 5].

The assessment of antiviral resistance also relies on in vitro–in vivo correlations. Baloxavir marboxil, a novel antiviral targeting the PA endonuclease, has been associated with the emergence of resistant viruses carrying I38T/F/M substitutions. While these substitutions impair PA endonuclease activity and reduce viral replication in cell culture (particularly in the case of I38T), they do not necessarily abolish transmissibility in ferrets [24, 25]. Jones and colleagues demonstrated that influenza A viruses with I38T/F/M substitutions could still transmit between ferrets via contact and airborne routes, indicating a moderate fitness cost that may not prevent community spread [24]. Competitive mixture studies in ferrets further revealed that I38T viruses have lower within-host fitness compared to wild-type, but this fitness cost is greater for H1N1pdm09 viruses than for H3N2 viruses, suggesting subtype-specific risk profiles [25]. These studies highlight the importance of interpreting in vitro resistance data through the lens of in vivo fitness, as the presence of a resistance mutation does not automatically equate to a loss of transmissibility or pandemic potential.

References

[1] Belser J, Katz J, Tumpey T. The ferret as a model organism to study influenza A virus infection. Disease Models & Mechanisms. 2011. DOI: https://doi.org/10.1242/dmm.007823

[2] Belser J, Eckert A, Huynh T, Gary J, Ritter J, Tumpey T, et al.. A Guide for the Use of the Ferret Model for Influenza Virus Infection. American Journal of Pathology. 2019. DOI: https://doi.org/10.1016/j.ajpath.2019.09.017

[3] Zeng H, Goldsmith C, Kumar A, Belser J, Sun X, Pappas C, et al.. Tropism and Infectivity of a Seasonal A(H1N1) and a Highly Pathogenic Avian A(H5N1) Influenza Virus in Primary Differentiated Ferret Nasal Epithelial Cell Cultures. Journal of Virology. 2019. DOI: https://doi.org/10.1128/JVI.00080-19

[4] Ng PSK, Böhm R, Hartley-Tassell L, Steen J, Wang H, Lukowski S, et al.. Ferrets exclusively synthesize Neu5Ac and express naturally humanized influenza A virus receptors. Nature Communications. 2014. DOI: https://doi.org/10.1038/ncomms6750

[5] Kieran T, Sun X, Maines T, Beauchemin CAA, Belser J. Exploring associations between viral titer measurements and disease outcomes in ferrets inoculated with 125 contemporary influenza A viruses. Journal of Virology. 2024. DOI: https://doi.org/10.1128/jvi.01661-23

[6] Roberts K, Shelton H, Stilwell P, Barclay W. Transmission of a 2009 H1N1 Pandemic Influenza Virus Occurs before Fever Is Detected, in the Ferret Model. PLoS ONE. 2012. DOI: https://doi.org/10.1371/journal.pone.0043303

[7] Belser J, Pulit-Penaloza J, Maines T. Ferreting Out Influenza Virus Pathogenicity and Transmissibility: Past and Future Risk Assessments in the Ferret Model.. Cold Spring Harbor Perspectives in Medicine. 2019. DOI: https://doi.org/10.1101/cshperspect.a038323

[8] Belser J, Barclay W, Barr I, Fouchier R, Matsuyama R, Nishiura H, et al.. Ferrets as Models for Influenza Virus Transmission Studies and Pandemic Risk Assessments. Emerging Infectious Diseases. 2018. DOI: https://doi.org/10.3201/eid2406.172114

[9] Pulit-Penaloza J, Brock N, Belser J, Sun X, Pappas C, Kieran T, et al.. Highly pathogenic avian influenza A(H5N1) virus of clade 2.3.4.4b isolated from a human case in Chile causes fatal disease and transmits between co-housed ferrets. Emerging Microbes and Infections. 2024. DOI: https://doi.org/10.1080/22221751.2024.2332667

[10] Golke A, Jańczak D, Szaluś-Jordanow O, Dzieciątkowski T, Sapierzyński R, Moroz-Fik A, et al.. Natural Infection with Highly Pathogenic Avian Influenza A/H5N1 Virus in Pet Ferrets. Viruses. 2024. DOI: https://doi.org/10.3390/v16060931

[11] Pulit-Penaloza J, Belser J, Brock N, Kieran T, Sun X, Pappas C, et al.. Transmission of a human isolate of clade 2.3.4.4b A(H5N1) virus in ferrets. Nature. 2024. DOI: https://doi.org/10.1038/s41586-024-08246-7

[12] Restori KH, Septer KM, Field C, Patel DR, VanInsberghe D, Raghunathan V, et al.. Risk assessment of a highly pathogenic H5N1 influenza virus from mink. Nature Communications. 2024. DOI: https://doi.org/10.1038/s41467-024-48475-y

[13] Gudymo AS, Onkhonova G, Danilenko A, Susloparov I, Danilchenko N, Kosenko M, et al.. Quantitative Assessment of Airborne Transmission of Human and Animal Influenza Viruses in the Ferret Model. Atmosphere. 2023. DOI: https://doi.org/10.3390/atmos14030471

[14] Hu M, Jones JC, Banoth B, Ojha C, Crumpton J, Kercher L, et al.. Swine H1N1 Influenza Virus Variants with Enhanced Polymerase Activity and HA Stability Promote Airborne Transmission in Ferrets. Journal of Virology. 2022. DOI: https://doi.org/10.1128/jvi.00100-22

[15] Richard M, Brand JVDvd, Bestebroer T, Lexmond P, Meulder Dd, Fouchier R, et al.. Influenza A viruses are transmitted via the air from the nasal respiratory epithelium of ferrets. Nature Communications. 2020. DOI: https://doi.org/10.1038/s41467-020-14626-0

[16] Creager HM, Kieran T, Zeng H, Sun X, Pulit-Penaloza J, Holmes KE, et al.. Utility of Human In Vitro Data in Risk Assessments of Influenza A Virus Using the Ferret Model. Journal of Virology. 2023. DOI: https://doi.org/10.1128/jvi.01536-22

[17] Chan K, Carolan L, Korenkov D, Druce J, McCaw J, Reading P, et al.. Investigating Viral Interference Between Influenza A Virus and Human Respiratory Syncytial Virus in a Ferret Model of Infection. Journal of Infectious Diseases. 2018. DOI: https://doi.org/10.1093/infdis/jiy184

[18] Rowe T, Davis WG, Wentworth DE, Ross T. Differential interferon responses to influenza A and B viruses in primary ferret respiratory epithelial cells. Journal of Virology. 2024. DOI: https://doi.org/10.1128/jvi.01494-23

[19] Gao J, Couzens LK, Burke D, Wan H, Wilson P, Memoli MJ, et al.. Antigenic Drift of the Influenza A(H1N1)pdm09 Virus Neuraminidase Results in Reduced Effectiveness of A/California/7/2009 (H1N1pdm09)-Specific Antibodies. mBio. 2019. DOI: https://doi.org/10.1128/mBio.00307-19

[20] Ryan K, Slack GS, Marriott A, Kane J, Whittaker CJ, Silman N, et al.. Cellular immune response to human influenza viruses differs between H1N1 and H3N2 subtypes in the ferret lung. PLoS ONE. 2018. DOI: https://doi.org/10.1371/journal.pone.0202675

[21] Marriott A, Dove B, Whittaker CJ, Bruce C, Ryan K, Bean T, et al.. Low Dose Influenza Virus Challenge in the Ferret Leads to Increased Virus Shedding and Greater Sensitivity to Oseltamivir. PLoS ONE. 2014. DOI: https://doi.org/10.1371/journal.pone.0094090

[22] Pulit-Penaloza J, Brock N, Belser J, Sun X, Pappas C, Tumpey T, et al.. Kinetics and magnitude of viral RNA shedding as indicators for Influenza A virus transmissibility in ferrets. Communications Biology. 2023. DOI: https://doi.org/10.1038/s42003-023-04459-0

[23] Stannard HL, Mifsud EJ, Wildum S, Brown SK, Koszalka P, Shishido T, et al.. Assessing the fitness of a dual-antiviral drug resistant human influenza virus in the ferret model. Communications Biology. 2022. DOI: https://doi.org/10.1038/s42003-022-04005-4

[24] Jones JC, Pascua P, Fabrizio TP, Marathe BM, Seiler P, Barman S, et al.. Influenza A and B viruses with reduced baloxavir susceptibility display attenuated in vitro fitness but retain ferret transmissibility. Proceedings of the National Academy of Sciences of the United States of America. 2020. DOI: https://doi.org/10.1073/pnas.1916825117

[25] Lee L, Zhou J, Koszalka P, Frise R, Farrukee R, Baba K, et al.. Evaluating the fitness of PA/I38T-substituted influenza A viruses with reduced baloxavir susceptibility in a competitive mixtures ferret model. bioRxiv. 2020. DOI: https://doi.org/10.1371/journal.ppat.1009527

[26] Kirsebom F, Thompson C, Talts T, Kele B, Whitaker HJ, Andrews NJ, et al.. Early influenza virus characterisation and vaccine effectiveness in England in autumn 2025, a period dominated by influenza A(H3N2) subclade K. Euro surveillance : bulletin Europeen sur les maladies transmissibles = European communicable disease bulletin. 2025. DOI: https://doi.org/10.2807/1560-7917.ES.2025.30.46.2500854

[27] Pulit-Penaloza J, Jones J, Sun X, Jang Y, Thor S, Belser J, et al.. Antigenically Diverse Swine Origin H1N1 Variant Influenza Viruses Exhibit Differential Ferret Pathogenesis and Transmission Phenotypes. Journal of Virology. 2018. DOI: https://doi.org/10.1128/JVI.00095-18

[28] Xu L, Bao L, Deng W, Zhu H, Chen T, Lv Q, et al.. The mouse and ferret models for studying the novel avian-origin human influenza A (H7N9) virus. Virology Journal. 2013. DOI: https://doi.org/10.1186/1743-422X-10-253

[29] Sun X, Belser J, Li Z, Brock N, Pulit-Penaloza J, Kieran T, et al.. Effect of Prior Influenza A(H1N1)pdm09 Virus Infection on Pathogenesis and Transmission of Human Influenza A(H5N1) Clade 2.3.4.4b Virus in Ferret Model. Emerging Infectious Diseases. 2025. DOI: https://doi.org/10.3201/eid3103.241489

[30] Li R, Yuan B, Xia X, Zhang S, Du Q, Yang C, et al.. Tree shrew as a new animal model to study the pathogenesis of avian influenza (H9N2) virus infection. Emerging Microbes and Infections. 2018. DOI: https://doi.org/10.1038/s41426-018-0167-1

[31] Herfst S, Begeman L, Spronken M, Poen M, Eggink D, Meulder Dd, et al.. A Dutch highly pathogenic H5N6 avian influenza virus showed remarkable tropism for extra-respiratory organs and caused severe disease but was not transmissible via air in the ferret model. Msphere. 2023. DOI: https://doi.org/10.1128/msphere.00200-23

[32] Pulit-Penaloza J, Brock N, Pappas C, Sun X, Belser J, Zeng H, et al.. Characterization of highly pathogenic avian influenza H5Nx viruses in the ferret model. Scientific Reports. 2020. DOI: https://doi.org/10.1038/s41598-020-69535-5

[33] Pulit-Penaloza J, Belser J, Brock N, Thakur PB, Tumpey T, Maines T. Pathogenesis and Transmissibility of North American Highly Pathogenic Avian Influenza A(H5N1) Virus in Ferrets. Emerging Infectious Diseases. 2022. DOI: https://doi.org/10.3201/eid2809.220879

[34] Watanabe T, Iwatsuki-Horimoto K, Kiso M, Nakajima N, Takahashi K, Lopes TJdS, et al.. Experimental infection of Cynomolgus Macaques with highly pathogenic H5N1 influenza virus through the aerosol route. Scientific Reports. 2018. DOI: https://doi.org/10.1038/s41598-018-23022-0

[35] Tosheva II, Filaire F, Rijnink WF, Meulder Dd, Kekem Bv, Bestebroer T, et al.. Influenza A(H5N1) shedding in air corresponds to transmissibility in mammals. Nature Microbiology. 2024. DOI: https://doi.org/10.1038/s41564-024-01885-6

[36] Herfst S, Mok CK, Mok CK, Brand J, Vliet Svd, Rosu M, et al.. Human Clade 2.3.4.4 A/H5N6 Influenza Virus Lacks Mammalian Adaptation Markers and Does Not Transmit via the Airborne Route between Ferrets. Msphere. 2018. DOI: https://doi.org/10.1128/mSphere.00405-17

[37] Chen M, Lyu Y, Wu F, Zhang Y, Li H, Wang R, et al.. Increased public health threat of avian-origin H3N2 influenza virus caused by its evolution in dogs. eLife. 2023. DOI: https://doi.org/10.7554/eLife.83470

[38] Horman W, Kedzierska K, Rootes CL, Bean A, Nguyen THO, Layton D. Ferret Interferon (IFN)-Inducible Transmembrane Proteins Are Upregulated by both IFN-α and Influenza Virus Infection. Journal of Virology. 2021. DOI: https://doi.org/10.1128/JVI.00111-21

[39] Huang S, Huang C, Chen C, Chen T, Lin C, Lin Y, et al.. Novel Role for miR-1290 in Host Species Specificity of Influenza A Virus. Molecular Therapy: Nucleic Acids. 2019. DOI: https://doi.org/10.1016/j.omtn.2019.04.028

[40] Laurie K, Guarnaccia T, Carolan L, Yan A, Aban M, Petrie S, et al.. Interval Between Infections and Viral Hierarchy Are Determinants of Viral Interference Following Influenza Virus Infection in a Ferret Model. Journal of Infectious Diseases. 2015. DOI: https://doi.org/10.1093/infdis/jiv260

[41] Siegers JY, Ferreri L, Eggink D, Kroeze EVV, Velthuis AJTt, Bildt MVDvd, et al.. Evolution of highly pathogenic H5N1 influenza A virus in the central nervous system of ferrets. PLoS Pathogens. 2023. DOI: https://doi.org/10.1371/journal.ppat.1011214

[42] Jansen AG, Spaan T, Low H, Iorio DD, Brand J, Tieke M, et al.. Influenza-induced thrombocytopenia is dependent on the subtype and sialoglycan receptor and increases with virus pathogenicity.. Blood Advances. 2020. DOI: https://doi.org/10.1182/bloodadvances.2020001640

[43] Nuñez IA, Jang H, Huang Y, Kelvin A, Ross TM. Influenza virus immune imprinting dictates the clinical outcomes in ferrets challenged with highly pathogenic avian influenza virus H5N1. Frontiers in Veterinary Science. 2023. DOI: https://doi.org/10.3389/fvets.2023.1286758

[44] Nachbagauer R, Liu W, Choi A, Wohlbold TJ, Atlas T, Rajendran M, et al.. A universal influenza virus vaccine candidate confers protection against pandemic H1N1 infection in preclinical ferret studies. npj Vaccines. 2017. DOI: https://doi.org/10.1038/s41541-017-0026-4

[45] Liu W, Nachbagauer R, Stadlbauer D, Strohmeier S, Solórzano A, Berlanda-Scorza F, et al.. Chimeric Hemagglutinin-Based Live-Attenuated Vaccines Confer Durable Protective Immunity against Influenza A Viruses in a Preclinical Ferret Model. Vaccines. 2021. DOI: https://doi.org/10.3390/vaccines9010040

[46] Liu W, Nachbagauer R, Stadlbauer D, Solórzano A, Berlanda-Scorza F, García‐Sastre A, et al.. Sequential Immunization With Live-Attenuated Chimeric Hemagglutinin-Based Vaccines Confers Heterosubtypic Immunity Against Influenza A Viruses in a Preclinical Ferret Model. Frontiers in Immunology. 2019. DOI: https://doi.org/10.3389/fimmu.2019.00756

[47] Rosu M, Kok A, Bestebroer T, Meulder Dd, Verveer EP, Pronk MR, et al.. Contribution of Neuraminidase to the Efficacy of Seasonal Split Influenza Vaccines in the Ferret Model. Journal of Virology. 2022. DOI: https://doi.org/10.1128/jvi.01959-21

[48] Skarlupka AS, Ross T. Immune Imprinting in the Influenza Ferret Model. Vaccines. 2020. DOI: https://doi.org/10.3390/vaccines8020173

[49] Kaul D, Rathnasinghe R, Ferres M, Tan G, Barrera A, Pickett B, et al.. Microbiome disturbance and resilience dynamics of the upper respiratory tract during influenza A virus infection. Nature Communications. 2020. DOI: https://doi.org/10.1038/s41467-020-16429-9

[50] Dürrwald R, Wedde M, Biere B, Oh D, Heßler-Klee M, Geidel C, et al.. Zoonotic infection with swine A/H1avN1 influenza virus in a child, Germany, June 2020. Euro surveillance : bulletin Europeen sur les maladies transmissibles = European communicable disease bulletin. 2020. DOI: https://doi.org/10.2807/1560-7917.ES.2020.25.42.2001638

[51] Belser J, Lau E, Barclay W, Barr I, Chen H, Fouchier R, et al.. Robustness of the Ferret Model for Influenza Risk Assessment Studies: a Cross-Laboratory Exercise. bioRxiv. 2022. DOI: https://doi.org/10.1128/mbio.01174-22

[52] Turgeon N, Hamelin M, Verreault D, Lévesque A, Rhéaume C, Carbonneau J, et al.. Design and Validation with Influenza A Virus of an Aerosol Transmission Chamber for Ferrets. International Journal of Environmental Research and Public Health. 2019. DOI: https://doi.org/10.3390/ijerph16040609

[53] Holzer B, Rijal P, Mcnee A, Paudyal B, Clark B, Manjegowda T, et al.. Protective porcine influenza virus-specific monoclonal antibodies recognize similar haemagglutinin epitopes as humans. bioRxiv. 2020. DOI: https://doi.org/10.1371/journal.ppat.1009330

[54] Wong S, Duan S, Debeauchamp J, Zanin M, Kercher L, Sonnberg S, et al.. The immune correlates of protection for an avian influenza H5N1 vaccine in the ferret model using oil-in-water adjuvants. Scientific Reports. 2017. DOI: https://doi.org/10.1038/srep44727

[55] Gooch K, Marriott A, Ryan K, Yeates P, Slack GS, Brown PJ, et al.. Heterosubtypic cross-protection correlates with cross-reactive interferon-gamma-secreting lymphocytes in the ferret model of influenza. Scientific Reports. 2019. DOI: https://doi.org/10.1038/s41598-019-38885-0

[56] Reber A, Music N, Kim J, Gansebom S, Chen J, York I. Extensive T cell cross-reactivity between diverse seasonal influenza strains in the ferret model. Scientific Reports. 2018. DOI: https://doi.org/10.1038/s41598-018-24394-z

[57] Koszalka P, George AM, Dhanasekaran V, Hurt A, Subbarao K. Effect of Baloxavir and Oseltamivir in Combination on Infection with Influenza Viruses with PA/I38T or PA/E23K Substitutions in the Ferret Model. mBio. 2022. DOI: https://doi.org/10.1128/mbio.01056-22

[58] Toots M, Yoon J, Hart M, Natchus M, Painter G, Plemper RK. Quantitative Efficacy Paradigms of the Influenza Clinical Drug Candidate EIDD-2801 in the Ferret Model. Translational research : the journal of laboratory and clinical medicine. 2019. DOI: https://doi.org/10.1016/j.trsl.2019.12.002

[59] Liu YV, Massare M, Pearce MB, Sun X, Belser J, Maines T, et al.. Recombinant virus-like particles elicit protective immunity against avian influenza A(H7N9) virus infection in ferrets. Vaccine. 2015. DOI: https://doi.org/10.1016/j.vaccine.2015.03.009

[60] Lee L, Zhou J, Frise R, Goldhill DH, Koszalka P, Mifsud EJ, et al.. Baloxavir treatment of ferrets infected with influenza A(H1N1)pdm09 virus reduces onward transmission. PLoS Pathogens. 2020. DOI: https://doi.org/10.1371/journal.ppat.1008395

[61] Francis ME, McNeil M, Dawe N, Foley M, King ML, Ross T, et al.. Historical H1N1 Influenza Virus Imprinting Increases Vaccine Protection by Influencing the Activity and Sustained Production of Antibodies Elicited at Vaccination in Ferrets. Vaccines. 2019. DOI: https://doi.org/10.3390/vaccines7040133

Disclaimer: This article is for educational and informational purposes only. It is not intended to substitute for professional veterinary advice, diagnosis, treatment, or regulatory guidance. Always consult a licensed veterinarian or qualified specialist regarding animal health, disease diagnosis, and therapeutic decisions.