Influenza A virus in dairy cattle: infection biology and potential mammary gland-targeted vaccines

Abstract

Influenza, a major “One Health” threat, has gained heightened attention following recent reports of highly pathogenic avian influenza in dairy cattle and cow-to-human transmission in the USA. This review explores general aspects of influenza A virus (IAV) biology, its interactions with mammalian hosts, and discusses the key considerations for developing vaccines to prevent or curtail IAV infection in the bovine mammary gland and its spread through milk.

Introduction

Influenza A is a worldwide spread and highly contagious viral disease that ranks among the major public and veterinary health issues. Recent reports of infections by highly pathogenic avian influenza viruses (HPAIVs) in wild and domestic mammals worldwide^1,2 point to an elevated risk of adaption of these viruses to mammalian hosts, which may ultimately lead to the emergence of a new pandemic strain. This threat has recently reached an unprecedented alarming level after the reports of sustained HPAIV (H5N1 clade 2.3.4.4b) infections in dairy cattle and cases of cow-to-human transmission in the USA^3–6.

In this review, we provide a brief overview of the biology and natural history of influenza A virus (IAV) infections and describe the key mechanisms involved in the interactions between IAV and its mammalian hosts. We also highlight the main recent observations on HPAIV infections in dairy cows and discuss the major aspects to be considered for the development of candidate vaccines able to prevent productive IAV replication in the mammary gland and its spread through milk.

Key features of Influenza A virus

IAV is the type species of the genus Alphainfluenzavirus and belongs to the Orthomyxoviridae family. Because of their rich diversity, IAVs are commonly classified as viral subtypes according to combinations of distinct serotypes of the two surface glycoproteins, haemagglutinin (HA or H, when referring to the serotype) and neuraminidase (NA or N), that decorate the viral envelope. Sixteen serotypes of HA and 9 serotypes of NA have been identified, mostly in avian influenza viruses (AIVs), and almost all of their possible combinations (HxNy) have been detected in field isolates. The viral genome, totalling about 13 kilobases, comprises 8 single-stranded RNAs of negative polarity. Both in the virion and in the infected cell, each of these genomic RNAs are associated with multiple copies of the viral nucleoprotein NP⁷, while the RNA termini are associated with the three subunits (PA, PB1, PB2) of the viral RNA-dependent RNA polymerase. Following HA binding to sialic acids (SAs) linked to cellular glycoproteins, the virus is internalized. After fusion of the viral envelope with the endosome membrane, its genetic material, in the form of eight ribonucleoproteins, is directed to the nucleus, where both viral mRNAs transcription and viral genome replication occur.

At the end of the viral cycle, 8 newly synthesized ribonucleoproteins (RNPs) associate with each other, along with the cellular membrane, leading to the budding of enveloped progeny virions. During this step, the segmented nature of the viral genome allows genomic reassortments, leading to the creation of viruses with new antigenic and phenotypic properties. Whenever a cell is infected by two distinct IAVs, their individual RNPs can be mixed in theoretically up to 256 possible viral genome combinations, including the two parental ones. Reassortments can be observed between two human IAVs (e.g. H1N1 and H3N2) infecting a single person. However, these events are far more common in wild aquatic and shore birds, which collectively constitute the main natural reservoir for avian influenza viruses (AIVs). The co-infection of a wild aquatic bird by two or more AIVs is not uncommon^8,9, providing multiple opportunities for reassortment events.

Expansion and diversification of H5N1 HPAIVs: impacts and trends

H5N1 viruses belong to a long lineage of H5 AIVs that originated in China in the 1990s. The HA of all the current panzootic avian H5Nx viruses can be traced back to the progenitor virus A/goose/Guangdong/1996 (A/gs/Gd). This HA is the principal virulence determinant of these panzootic viruses, because of its particularly long polybasic cleavage site (PCS): RERRRKKR | GLFGA, which resulted from the insertion of untemplated purines by the “stuttering” viral polymerase (Fig. 1). The membrane fusion that occurs at the beginning of the viral cycle is mediated by the HA, once it is activated through cleavage into HA1 and HA2 by extracellular proteases acting at the N-terminus of the GLFGA peptide sequence. Due to the particularly long PCS of the HA of A/gs/Gd, this cleavage can be readily catalysed by ubiquitous proteases (such as furin), a prerequisite for the systemic dissemination of HPAIVs within their avian hosts. The HA of A/gs/Gd emerged from the Eurasian pool of low-pathogenicity avian H5 viruses¹⁰, most likely from H5N2 and H5N3 viruses that circulated in Asia at that time (Fig. 1). While the reassortment events between avian influenza viruses before 1996 were only poorly or not at all documented, the acquisition of this polybasic cleavage site was clearly the decisive factor in the surge of the current H5 HPAI viruses.

Fig. 1. Alignment of HA nucleotide sequence (nucleotides 1008-1070 of the coding sequence).

The peptide sequence corresponding to the cleavage site is shown on the right (blue rectangle highlights the stretch of basic residues upstream of the cleavage site). Alignment was performed using the “emma” and “shoawalign” tools in EMBOSS and manually edited. A few representative sequences from the Eurasian set of H5 sequences (up to year 1996) were retrieved from the Influenza virus resource. Sequences 1 and 9, corresponding to previous HPAIVs in Europe, were included for comparison.

The major HPAI epizootic caused by A/gs/Gd and its descendants in Guangdong and Hong Kong in 1996–97 resulted in 18 human infections, of which 6 were lethal. This first epizootic was stopped, thanks to the culling of millions of poultry. However, “stamping-out” strategies failed to eliminate the pool of Asian H5 viruses. H5 HPAIVs expanded and diversified in the 1997–2002 period, infecting several species of wild and domestic birds¹¹. From 2002 onwards, HPAIV belonging to the “genotype Z” caused several outbreaks in China and South East Asia. After these initial events at a regional scale, several important waves of H5 HPAI spread to other continents in the next two decades. In 2005–2008, Qinghai-lineage H5N1 reached Europe and Africa^11,12, while a major wave of H5N8 viruses harbouring clade 2.3.4.4 H5 spread worldwide in 2014–2015¹³. This was followed in 2020 by the ongoing global wave of H5N1 viruses harbouring clade 2.3.4.4b H5^14,15 and since 2022, more and more avian and mammalian species have been found to be infected by the currently circulating H5N1 viruses^16,17.

IAV infections in mammals: natural history and host response

Infections by IAVs have been detected in a wide array of avian and mammalian species. Wild migratory birds, especially waterfowl, are considered as the natural viral reservoir and play a role in the spread of IAV infections across all continents. Occasionally, these avian-origin IAVs, commonly referred to as AIVs, cross over from their natural reservoir to domestic birds, notably farmed chickens, turkeys, geese and ducks. In these hosts, some low-pathogenicity AIVs (LPAIVs) of the H5 or H7 types can evolve into HPAIVs through the above-mentioned insertion of untemplated purines. On the other hand, AIVs (LPAIVs or HPAIVs) in domestic birds can spill over to mammals including humans, causing sporadic or sustained infections. IAVs that are currently circulating in swine and in horses (H1N1, H1N2 and H3N2 strains in swine; H3N8 in horses) result from intricate combinations of spill-over and reassortment events that occurred in the preceding decades^18–20. In humans, all seasonal influenza viruses that have been or that are currently circulating have likewise resulted from combinations of spill-over infections and virus reassortments. Such evolutionary events, often involving swine as an intermediate host, have caused the four influenza viruses pandemics of the 20th and 21st centuries: Spanish flu (H1N1) in 1918, Asian flu (H2N2) in 1957, Hong Kong flu (H3N2) in 1968 and the 2009 swine flu pandemic (H1N1/2009)²¹. Constantly circulating in the human population, human H3N2 and H1N1/2009 viruses are nowadays the cause of seasonal influenza epidemics (together with influenza B viruses of the B/Victoria and B/Yamagata lineages) which result in 290.000 to 645.000 deaths annually worldwide²². Sporadic infections by AIV without sustained intra-species transmission have been detected in numerous mammalian species including humans, pigs and various domestic or wild carnivores and marine mammals²³. Sporadic spill-over infections leading to sustained intra-species transmission have so far only been reported in cats²⁴, seals²⁵, farmed fur animals (minks, foxes and racoon dogs)^26,27 and, only recently, in cattle³.

Adaptation of AIVs to mammalian hosts requires amino acid substitutions in various viral proteins, with modifications in the HA protein and the viral polymerase playing a key role in this process. Certain substitutions in the HA protein can allow it to switch its receptor binding specificity, from avian-type SA-α2,3 (preferentially bound by AIVs) to mammalian-type SA-α2,6 (preferentially bound by human/mammalian IAVs)^28,29. Beside several substitutions in HA that are required to allow airborne person-to-person transmission, human adaptation also involves substitution E627K in the PB2 subunit of the viral polymerase to enhance viral replication in mammalian cells^16,29. While most H5 AIVs that have been isolated so far from milk or cattle have not acquired these well-known adaptation markers, they carry several other substitutions that may increase their transmissibility and virulence in mammals³⁰.

Upon entry into the host cell via clathrin-mediated or clathrin- and caveolin-independent endocytic pathways³¹, IAV is recognized by at least 3 types of pattern recognition receptors (PRRs) from the innate immune system: toll-like receptors (TLRs) that bind dsRNA (TLR3) or ssRNA (TLR7 and TLR8), the retinoic acid-inducible gene I (RIG-I, recognizing 5’-triphosphate RNA) and the NOD-like receptor NLRP3 (a component of inflammasome)^32,33. This recognition leads to the activation of the nuclear factor kappa B (NF-κB) and interferon-regulatory factors (IRFs), which trigger the expression of pro-inflammatory cytokines (IL-1β, IL-18, IL-6, IL-8, TNF-α among others) and type I and III interferons (IFN-α/β, IFN-λ). Activation of the IFNα receptor by type I IFNs triggers the JAK-STAT signalling cascade and the expression of interferon stimulated gene (ISGs) with potent antiviral activities^32–34. Cytokine secretion will lead to the early recruitment of immune effector cells, such as NK cells, monocytes and neutrophils, with a later arrival and activation of lymphocytes and plasma cells³⁵.

Immune responses against IAV infection are a double-edged sword. While a strong activation of innate immunity is needed to clear infection, a dysregulated response, especially in the case of human infections with certain AIVs, can initiate a so-called “cytokine storm”. This excessive cytokine secretion syndrome leads to severe and often fatal disease outcomes³⁶. H5N1 HPAIV infections in humans often lead to pneumonia, acute respiratory distress syndrome (ARDS) and multi-organ failure, with a case fatality rate of approximately 50%. Patients show important lung inflammation, leukocyte infiltration, decreased peripheral lymphocyte count and high cytokine levels^37,38.

Experimental infections of human volunteers with a seasonal human influenza virus (A/Texas/36/91 H1N1) demonstrated that the severity and magnitude of symptoms and viral replication correlate with the levels of IL-6 and IFN-α³⁹. A similar correlation between the cytokine response and the severity of disease was observed using the ferret model of infection with seasonal influenza H1N1 and H3N2. While mild disease was associated with elevated levels of IFN-α and IFN-γ, severe disease was associated to elevated IL-6 levels⁴⁰. In this model, infection with a H5N1 HPAIV produced a pathology similar to that observed following human H5N1 infections, including weight loss, lymphopenia, severe respiratory symptoms and death^41,42. The key role of IFN type I signalling in the host defence to IAV has been confirmed in murine models based on animals deficient for IFNα receptor (Ifnar−/−) and STAT (Stat−/−). Ifnar−/− mice expressed decreased levels of interferon-stimulated genes (ISGs), while Stat−/− mice showed deficient viral clearance⁴³. In addition, IFN-β deficient mice showed decreased survival rates and enhanced viral replication when infected with the SC35M strain (H7N7)⁴⁴. Type I IFNs and dendritic cells help naïve CD8⁺ T cells to differentiate into cytotoxic T lymphocytes (CTLs) that target infected cells. Memory CD8⁺ T cells respond quickly to a second influenza infection but their efficacy decreases with time. Importantly, unlike neutralizing antibodies that are specific to a single HA serotype⁴⁵, memory CD8 + T cells can potentially cross-react to different influenza subtypes since they recognize epitopes that are conserved in a variety of IAVs⁴⁶.

HPAIV outbreaks in bovine populations: general considerations and the unexpected role of the mammary gland

Unlike other livestock hosts at the human-animal interface (such as poultry and swine), bovine species have previously not been considered as susceptible targets for IAV infection, much less as potential reservoirs for continuous viral evolution⁴⁷. Experimental infections of calves with HPAIV (H5N1 strain A/cat/Germany/R606/2006) via the intranasal route failed to induce clinical symptoms and resulted in low levels of viral shedding⁴⁸. In addition, avian-type IAV receptors have previously been considered as absent from the upper and lower respiratory tract in cattle⁴⁹. These observations, combined with data from multiple field studies⁴⁷, suggested that the bovine species are less susceptible to infections with AIVs compared to other mammalian species expressing the avian-type SA-α2,3 receptor in the respiratory tract such as pigs and humans^50,51. However, this understanding has been challenged by recent outbreaks of H5N1 clade 2.3.4.4b HPAIVs in multiple dairy cattle herds in the USA^3,5,52. Two recent studies exploring the diversity of SA in bovine tissues have uncovered that both avian and mammalian-type IAV receptors are widely expressed in the respiratory tract and mammary gland of dairy cows and beef calves^53,54. These observations demonstrate that the specific SA present in bovine tissues enable avian or mammalian-origin IAV to infect this species.

Since the initial detection in commercial dairy herds in Texas, USA, on March 2024, 924 confirmed cases of HPAI H5N1 clade 2.3.4.4b virus infection have been confirmed across 16 states as of January 11, 2025⁵². These infections are clinically characterized by nonspecific systemic symptoms including lethargy, reduced feed intake and rumen motility, as well as a sharp drop in milk production^3,55. Although reports from the 1990-2000s have associated the presence IAVs or IAV-specific antibodies to a loss of milk production in bovine^56–58, the recent cases of HPAIV infections are distinguished by a prominent involvement of the mammary gland in viral replication and spread within and between affected herds^3,55. Infected animals show thick yellow milk with flecks and/or clots, often bacterial culture negative, indicating the occurrence of viral mastitis. Post-mortem analyses of affected cows confirmed the presence of neutrophilic mastitis concurrent with the presence of viral antigens in mammary alveolar epithelial cells and intraluminal cells³.

Mammary epithelial cells and leukocytes (macrophages, lymphocytes and neutrophils) present in numbers in the mammary gland act in concert by cytokine signalling to implement local innate and adaptive immune responses^59,60 (Fig. 2). Upon IAV infection, these cells probably initiate immune mechanisms by the recognition of viral RNA and surface proteins. The expression of TLR7, a key endosomal receptor for the detection of viral RNA⁶¹, has not been demonstrated in the bovine mammary gland. However, the TLR1/TLR2 complex as well as the RIG-I receptor are reported to be expressed in the mammary gland^62–65 and have been associated to viral detection⁶¹. Recognition of herpesviral motifs by the TLR1/TLR2 complex has been demonstrated in other species and the HA from measles virus was shown to activate murine and human antigen-presenting cells via the TLR2 receptor⁶⁶. Therefore, the contribution of these receptors to viral recognition in the mammary gland shall not be overlooked and remains to be explored.

Fig. 2. Anatomy of the bovine mammary gland and mediators of local immune response mechanisms.

For details refer to⁶⁰.

Virus-related mastitis has been mostly associated to diseases such as bovine viral leukemia⁶⁷, foot and mouth disease⁶⁸, bovine viral diarrhea and bovine herpesvirus-1 infections⁶⁹, which predispose affected animals to secondary infections due to general immune suppression. Some evidence indicates that changes in the mammary gland and its secretion can be attributed to bacterial infections secondary to tissue damage caused by viruses^70,71. Viruses can also infect the udder as a consequence of prolonged bacterial mastitis⁷² and, in some cases, multiply within the mammary gland without any clinical manisfestation^70,73. Although viruses are not considered major udder pathogens^71,74, intramammary inoculation experiments showed that the bovine mammary gland can act as a replication site for multiple viruses including bovine enterovirus and other viruses causing Newcastle disease, fowl plague, mumps, canine distemper, poliomyelitis, vesicular stomatitis, foot-and-mouth disease, infectious bovine rhinotracheitis, pseudo-cowpox, African swine fever, parainfluenza and human influenza^75–77. These pioneer observations showed that large amounts of virus progeny can be produced in the bovine mammary gland and pointed to a potential role of milk in the transmission of viral diseases. Indeed, natural infections by bovine herpesviruses can result in intermittent shedding of the virus in milk by persistently affected animals⁷⁴. Viral dissemination in milk has also been reported in cows infected with parainfluenza 3 virus⁷³, foot and mouth disease virus⁷⁸, bovine leukemia virus⁷⁹ and recently, IAV³.

A genomic and epidemiologic investigation attributes the source of the multi-state cattle outbreak in the USA to a single transmission event from avian species⁸⁰. Infections among cattle were reported to be horizontally transmitted by the displacement of lactating cows showing or not clinical disease^4,55 but the precise mode of transmission and the initial site of virus replication remains unclear. Caserta et al. ⁵⁵ hypothesized that HPAIVs could disseminate to organs such as the mammary gland via a short and low-level viremia upon oral or respiratory infection. The authors also proposed that infection could also occur through the teat orifice and cisternae, leading to viremia and virus dissemination to other tissues. To address this issue, an experimental infection study has been carried out to establish the role of the respiratory and intramammary routes in the onset of disease⁸¹. Viral replication in infected mammary gland was shown to be significantly greater than anticipated, reinforcing the notion that the udder and milk are primary sources of viral spread. Although experimental respiratory infection led to the detection of viable virus in only 2 out of 4 experimental animals, this route could be relevant for viral transmission in the context of animal facilities holding hundreds of animals⁸¹.

The establishment of endemic infections by HPAIVs in cattle could have significant implications from a One Health perspective. Evidence demonstrating multidirectional interspecies transmission of this pathogen suggests that, in this scenario, cattle and other animal species co-located on agricultural premises could act as mixing vessels facilitating the emergence of novel IAVs with enhanced zoonotic potential^3,80,81. HPAIV dissemination in milk also represents a risk to individuals with occupational exposure to infected animals. Two cases of cow-to-human spread involving dairy farm workers have been reported. Both patients had mild illness with conjunctivitis and fully recovered⁶. As pasteurization inactivates influenza viruses^82,83, the commercial milk supply is considered to be safe for consumption⁸⁴. Nevertheless, raw milk and raw milk-based products from affected farms represent a risk for human consumers and other animals.

Evaluating vaccine strategies to prevent HPAIV infections: focus on the bovine mammary gland

Efficient vaccines against bovine mastitis are still lacking⁸⁵. However, the extensive trial-and-error efforts over the last decades to set up improved solutions have provided valuable insights into mammary gland immunity. This accumulated knowledge could contribute significantly to the design of candidate vaccines aimed at preventing viral proliferation in the bovine mammary gland and its subsequent dispersal via milk.

Vaccines against human seasonal influenza and swine, equine, or canine influenza are successfully used on a global scale since decades. Still, most of the current vaccines only protect against clinical disease and their effectiveness is often limited due to viral antigenic drift and resulting virus-vaccine mismatches⁸⁶. Most of the licensed IAV vaccines (both in animals and humans) are inactivated vaccines that are mainly produced in embryonated chicken eggs. There are three types of inactivated influenza vaccines (IIV): whole virus vaccines (virus inactivation by formaldehyde or β-propiolactone treatment), split virus vaccines (virus disruption by physicochemical treatment) and subunit vaccines (preparations of the HA and NA viral surface proteins). These adjuvanted or non-adjuvanted vaccine formulations are predominantly applied via the intramuscular route. Live attenuated influenza vaccines (LAIV) are based on attenuated/non-virulent (often cold-adapted) viruses that were generated by serial passaging or reverse genetics. They are typically applied via the mucosal/intranasal route⁸⁷.

Neutralizing antibodies directed against the HA globular head domain can provide sterilizing antiviral immunity and represent a well-established correlate of protection^88,89. Antibodies directed against NA are likewise important as they can help limiting viral spread within the infected host⁹⁰. IIV usually trigger strong IgG responses targeting both viral surface proteins, but their longevity is limited. Moreover, viral antigenic drift leading to virus-vaccine mismatches necessitates constant IIV reformulations that come with variable vaccine efficacies⁹¹. LAIV, mimicking natural viral infection, trigger less potent IgG responses than IIV, but they also induce robust mucosal (IgA secretion) and cellular (T-cell) immune responses, thus providing a more long lasting and potentially cross-protective (against antigenically distinct viruses) immunity^87,92. These observations indicate that vaccine formulations capable of inducing not only a neutralizing IgG response, but also mucosal and cellular immune responses in the mammary gland might represent ideal candidates to limit symptoms and viral spread during HPAIV infections in cattle.

Polyreactive antibodies produced without prior antigenic stimulation, defined as natural antibodies (NAbs), are found in bovine serum and milk. These antibodies, primarily of the IgM isotype, bind to microbial molecules with low affinity and have been associated with improved defence against bacterial mastitis⁹³. However, high titres of HPAIV in the mammary gland of infected cows^3,55 suggests that local NAbs are insufficient to block viral replication. Bovine milk contains low levels of immunoglobulins (Igs), with a predominance of the IgG1 isotype. This particularity significantly influences bovine mammary immunity, as IgG1 is not opsonic for neutrophils in this species. To increase IgG2, IgA and IgM levels, and consequently enhance humoral and phagocytic responses in the mammary gland, local stimulation through natural infections or vaccination is required⁹⁴. Baker et al. observed that the detection of virus-neutralizing antibodies in mammary quarters experimentally infected with HPAIV coincided with negative virus isolation⁸¹. This finding indicates that the induction of local influenza-specific antibodies might contribute to viral clearance from the udder, but further research is necessary to clarify the nature of involved Igs, the duration of this response, and importantly, if it can be modulated through vaccination.

Different immunisation approaches have enabled the production of antibodies targeting antigens from mastitis-causing bacteria in serum, but achieving protective antibody responses at sufficient levels in the mammary gland remains a challenging task. Prime and booster subcutaneous (SC) injections of an oil-in-water adjuvanted subunit vaccine in the mammary gland suspensory ligament have been shown to induce higher vaccine-specific IgG1 levels in milk, but not IgG2 or IgA levels, compared to two intramuscular (IM) administrations in the neck or an intramammary (IMM) prime followed by an SC booster in the mammary gland ligament⁹⁵. An increase in vaccine-specific IgG1, but not IgG2, in milk has also been reported following prime and booster SC injections in the neck of a vaccine composed of recombinant antigens combined with Emulsigen^®-D⁹⁶. Another study demonstrated that neither IM/IM nor IM/IMM prime and boost immunisations with killed Escherichia coli emulsified in oil adjuvant (Montanide ISA 61VG®) or E. coli culture supernatant improve milk bactericidal and opsonic activity⁹⁷, probably due to insufficient IgG2 levels, a limiting factor for bacterial clearance by neutrophils. Vaccination-induced IgG1 in the mammary gland could contribute to viral elimination by increasing the capacities of macrophages to act as phagocytic and antigen presenting cells. This Ig isotype represents the majority of IgG antibodies targeting IAV and is highly effective at direct virus inhibition⁹¹. Nevertheless, IgA are considered to present the highest affinity for influenza HA⁹¹ and this isotype is present at low levels in bovine milk⁹⁴. Immunisations with S. aureus α-toxoid combined with different adjuvants showed that alum supplemented with both saponin and oil result in higher IgG1, IgG2 and IgA titres in milk when compared to alum–oil or alum–saponin⁹⁸.

Evidence suggests that adaptive cell-mediated immunity (CMI) also confers host resistance to influenza. A study carried out in a natural infection setting uncovered that a higher frequency of cross-reactive cytotoxic T-cells (CTL; CD8⁺IFN-γ⁺IL-2⁻) is associated with reduced illness severity and absence of viral shedding in healthy human individuals infected with H1N1pdm09⁹⁹. In another report exploring pre-infection cellular components related with asymptomatic and symptomatic influenza illness in vaccinated and unvaccinated adult individuals, it has been revealed that protection from infection is associated with increased frequency of polyfunctional CD4⁺ and CD8⁺ T cells, circulating T follicular helper, myeloid dendritic cells, Th17 cells and innate effector CD16-expressing cytotoxic and cytokine-producing NK cells¹⁰⁰.

Pre-clinical evaluation of virus-vectored candidate vaccines indicate that the presence of CTL⁹¹ and CD4⁺⁹² T-cells recognizing highly conserved influenza proteins generates heterosubtypic immunity across IAV strains. T-cell-inducing vaccines could therefore counterbalance IAV escape mechanisms to circumvent humoral immunity. Of note, the direct activity of antigen-specific T-cells, rather than antibodies, has been reported as a major mechanism for controlling and preventing infections in the mammary gland^97,101. These observations suggest that the capacity to efficiently trigger local CMI mechanisms should be incorporated into the requirements of a candidate vaccine designed to prevent HPAIV infections in the bovine mammary gland.

A recent study demonstrated that mucosal immunization via the nasal route outperforms IM immunization in conferring cross-protection against influenza in mice¹⁰². The authors also highlighted that influenza cross-protection can be bolstered by heterologous sequential immunizations combining different administration routes and these findings have been attributed to improved IFN-γ-mediated effector functions and T-cells avidity¹⁰². Interestingly, local immunization has also been reported to modify favourably the course of infection in the mammary gland by limiting inflammation and accelerating bacterial clearance. Heterologous immunisation against E. coli in an IM/IMM prime/booster regime resulted in earlier and higher IFN-γ production in milk upon IMM infectious challenge⁹⁷. Further investigation suggested that intramammary immunizations are pivotal for the establishment of resident memory T-cells in the bovine mammary tissue¹⁰³. In this sense, an immunization strategy taking advantage of the intramammary route could promote the elimination of HPAIV at the site of entry in case of intramammary transmission and/or prevent productive viral replication in the mammary gland and dispersal in milk. Nevertheless, the development of an intramammary vaccine candidate against HPAIV necessitates careful consideration in the design of formulations, including the selection of appropriate adjuvants, to avoid compromising the role of the mammary gland as a milk-producing organ.

Intranasal immunisation with an oil-in-water adjuvanted subunit vaccine failed to elicit measurable antibody responses in dairy cattle and this outcome has been attributed to the inability of recombinant proteins to stimulate bovine immune responses at the nasal mucosa¹⁰⁴. In contrast, intranasal vaccination with live formulations has been reported to provide benefits in preventing bovine respiratory diseases that include achieving full efficacy in the presence of maternally derived antibodies and stimulating the common mucosal immune system (CMIS)¹⁰⁵. In this context, induction of mucosal immunity at the respiratory tract by vaccination could play a role in reducing airborne transmission of HPAIV. However, multiple observations suggest that the mammary gland of dairy ruminants does not belong to the CMIS¹⁰¹. This particularity represents a limitation for strategies based on oral or intranasal immunisations that simultaneously target the mammary gland. These and other factors involved in designing a vaccine candidate targeting the mammary gland are outlined in Fig. 3.

Fig. 3. Key considerations for designing vaccines against HPAIV infections in the bovine mammary gland.

APC antigen presenting cells, HA haemagglunitin.

Conclusions

Controlling HPAIV infections in cattle represents a major “One Health” emergency. Improvements in biosecurity and hygiene measures, coupled to a “cull plus vaccination” strategy, have proven to be efficient in controlling successive H5 HPAI epizootics in domestic birds¹⁰⁶. For this approach to be applied to outbreaks in cattle, the development of efficient vaccines will be crucial. Although the role of the mammary gland in the pathophysiology of HPAIV infections in the bovine species needs to be better understood, the observed viral tropism for mammary epithelial cells strongly suggests that a successful vaccine candidate must induce robust anti-viral protective responses in the udder. The limited advances in the implementation of efficient vaccines for mastitis in cattle over the past decades⁸⁵ indicate that this task represents a significant challenge and collaborative efforts among teams from the fields of virology, vaccinology and bovine immunology will be essential to achieve this objective.

Author contributions

Writing—original draft preparation: R.P.M., D.M., S.T., I.C.P. and P.G.; Writing—review & editing: R.P.M., S.T., D.M. and I.C.P.; Funding acquisition: R.P.M.

Competing interests

The authors declare no competing interests.