Novel recombinant H5-based vaccine provides effective protection against H5N1 influenza virus in cats

Abstract

The emergence and broad circulation of highly pathogenic avian influenza (HPAI) H5N1 virus in wild birds and its spillover into dairy cows with sustained transmission in this species pose a major risk to felines, which are highly susceptible and often succumb to the infection. Here, we developed a novel recombinant hemagglutinin H5-based vaccine and evaluated its safety, immunogenicity, and protective efficacy against HPAI H5N1 virus in domestic cats. Immunization of cats with H5-vaccine candidate elicited robust levels of neutralizing antibodies against H5N1 virus and protection against disease, mortality, and infection upon H5N1 virus challenge. The vaccine-elicited immunity significantly reduced virus shedding and viremia, limiting systemic spread and disease severity in immunized animals. Importantly, beyond protecting susceptible felids, vaccinating cats against the H5N1 virus could also reduce the risk of human exposure - underscoring the One Health impact of implementing such a vaccination strategy in feline populations.

Introduction

The highly pathogenic avian influenza (HPAI) H5N1 viruses from clade 2.3.4.4b comprise a large group of Influenza A viruses (IAVs) with a broad host range, increasing genetic diversity, and high evolutionary rate^1–5. The emergence and spread of HPAI H5N1 viruses in aquatic and terrestrial mammals pose a major threat to animal and human health. Since its first detection in the US in January 2022⁶, HPAI H5N1 clade 2.3.4.4b viruses have been detected in more than 53 wild and domestic animal species across North America⁷. These viruses have caused multiple outbreaks in red foxes, seals, and sporadic cases in wild mesocarnivores, including black bears and raccoons^8–10. Between March and December 2024, the HPAI H5N1 reassortant genotype B3.13 emerged as the main genotype infecting cattle and domestic cats, with documented evidence of transmission from cows to cows, and cows to cats, raccoons, pigeons, and humans^11,12. As a result of reassortment in wild bird reservoirs, a newly emerging H5N1 virus genotype D1.1 was identified in November-December 2024 and quickly became the predominant virus circulating in wild birds, leading to spillover to multiple species, including humans^13,14, dairy cows¹⁵, and domestic cats¹⁶.

HPAI H5N1 has caused multiple outbreaks in felids, including leopards and domestic cats in Southeast Asia¹⁷, Europe⁴, and more recently in the US^11,12. Notably, as of July 2025, there have been 200 confirmed cases of HPAI H5N1 infections in felids in North America¹⁶. Most of the infections in domestic cats have led to severe disease, often resulting in neurological disease and death of affected animals^18,19. Infections have been linked to the consumption of H5N1-contaminated raw milk on dairy farms^11,12, or raw milk bought at the marketplace²⁰, ingestion of affected wild birds²¹, or consumption of raw H5N1-contaminated pet food²². Although cat-to-cat or cat-to-human transmission has not been reported, the close contact of this companion animal with humans underscores the public health implications of HPAI H5N1 virus infections in domestic cats.

To mitigate the impact of HPAI H5N1 virus in domestic cats, we developed a subunit recombinant hemagglutinin (HA) H5 protein-based vaccine candidate (on a USDA-licensed platform)^23,24 and evaluated its safety, immunogenicity, and protective efficacy in cats against HPAI H5N1 challenge infection. The HA is the immunodominant structural glycoprotein of Influenza A viruses, which binds to the sialic acid (SA) receptors, enabling virus entry into host cells^25,26. The full-length HA glycoprotein (524 amino acids (aa)) is divided into a head (HA1) (1-328 aa) and stem (HA2) (329-524 aa) regions, with the globular head (116-261 aa), which contains the receptor binding domain, being the main target of neutralizing antibodies (NAbs)²⁵. Upon immunization of cats, we evaluated the cross-reactivity of NAb responses against the newly emergent genotype D1.1. Our results demonstrate that H5-based vaccine-elicited immunity resulted in reduced levels of virus replication and shedding as well as decreased viremia, limiting systemic spread and preventing clinical disease and mortality in immunized animals.

Results

The recombinant H5 vaccine elicited robust antibody responses in cats

We investigated the safety and immunogenicity of a baculovirus expressed recombinant H5 vaccine candidate in domestic cats. The full-length HA protein of HPAI H5N1 clade 2.3.4.4b genotype A2 virus (H5; A/common tern/Maine/W22-480B/2022; GenBank accession no. OP377550), expressed using the BestBac 2.0 baculovirus expression system, was formulated with aluminum hydroxide adjuvant (Fig. 1a, b) and used in immunization-challenge studies in cats. For this, 16 HPAI H5N1 virus seronegative research cats were allocated into two experimental groups as follows: G1: sham-immunized, H5N1 virus challenged (n = 8) and G2: H5-vaccine immunized, H5N1 virus challenged (n = 8) (Fig. 2a). Animals in G1 (sham-immunized) received a sham immunization consisting of supernatant extracts from Sf9 cells infected with an empty baculovirus and formulated in aluminum hydroxide adjuvant, while animals in G2 (H5-vaccine) received the recombinant H5 protein vaccine formulation. Both groups were immunized on days 0 and boosted on day 21 via the subcutaneous (SC) route (Fig. 2a). Following immunizations, animals were monitored for 48 h, and none of the animals developed adverse reactions at the site of injection. Additionally, no fever was observed in any of the immunized animals within 48 h after immunization.

Fig. 1. Recombinant H5-based subunit vaccine against HPAI H5N1 virus.

a Schematics depicting steps of H5-based subunit vaccine production using baculovirus in Spodoptera frugiperda insect cells (Sf9). A pBacPAK8 cloning vector was used to prepare a plasmid vector containing the H5 sequence of HPAI H5N1 virus. The plasmid vector contains flanking sequences homologous to the linear BestBac 2.0 Baculovirus vector for homologous recombination exchanging the H5 insert for the polyhedrin gene of the Baculovirus. The recombinant Baculovirus encoding the H5 gene under control of the polyhedrin promoter was harvested and infected cultures were centrifuged to remove the cells and the supernatant. The supernatant was processed through 0.2-micron sterile disposable filter. The premaster culture was titrated to determine the final concentration and Sf9 cells were scaled up to production quantities in suspension. Cells were infected with the recombinant baculovirus, harvested, and processed. Fluids were inactivated with Beta-propiolactone (BPL) at a final concentration of 0.2-0.3%. The final vaccine formulation was prepared from the inactivated virus fluids in aluminum hydroxide adjuvant. b Expression of H5 protein in Sf9 cells as shown by Western blot. Lanes 1-3 included negative controls as: 1 empty baculovirus, 2 non-target antigen, 3 Sf9 cell lysate. 4 Standard Marker, 5-10; Passage 0 to Passage 5 show stable expression of the recombinant H5 protein. The recombinant H5 protein was probed with anti-Influenza H5N1 rabbit polyclonal antibody (primary), and goat anti-rabbit IgG (H + L) HRP polyclonal (secondary) antibodies. The predicted molecular weight of the H5 protein is 65.9 kD. While the lower band is at the predicted size of approximately 65.9 kD, the upper band is believed to be the glycosylated form of H5 (8 out of 48 sites in HA are predicted to be glycosylated).

Fig. 2. Immunization of domestic cats with recombinant H5-vaccine results in humoral responses.

a Experimental design of immunization-challenge study (sham [G1] and H5-vaccine [G2] group) in cats immunized (day 0), boosted (day 21) with H5 HA-based subunit vaccine, and then challenged with HPAI H5N1 TX2/24 isolate (2 ×10^4.5 PFU/cat) at day 35 post immunization (pi). Serum titers (log2) of virus-neutralizing antibodies (b) and hemagglutination-inhibition antibodies (c) at days 0, 14, 21 and 35 pi. Each dot in the bar graph indicates an individual cat from each group (mean with SEM, n = 8 cats/group). Statistical significance was determined by use of Two-way ANOVA with column means comparison (****P < 0.0001).

Antibody responses elicited by immunization with the recombinant H5-vaccine were evaluated by virus neutralization (VN), and hemagglutination inhibition (HAI) assays. Notably, NAbs were detected in six out of eight (6/8) cats on day 14 post-immunization (pi) with NAb titers ranging from 3–5 log₂ and increasing to 6–8 log₂ on day 21 pi (p < 0.0001) (Fig. 2b). On day 35 pi – 14 days post booster vaccination – NAb titers increased to 8–12 log₂ and were significantly higher (p < 0.0001) when compared to NAb titers detected in serum collected on days 14 and 21 pi, prior to the booster immunization. Of note, one cat (cat #5, N243131) from the H5-vaccine group consistently presented lower NAb titers of 0, 6, 8 log₂ on days 14, 21, and 35 pi, respectively, when compared to the other animals in the group (Fig. 2b). HAI antibody titers elicited by immunization were similar to the NAb titers (Fig. 2c), confirming a robust humoral immune response following prime-boost immunization with the H5-vaccine in cats.

Immunization with the recombinant H5 vaccine provided clinical protection against H5N1

To assess the protective efficacy of the recombinant H5 vaccine developed here, all animals in G1 and G2 were challenged oronasally on day 35 pi (1 mL intranasally [IN] and 1 mL orally [PO]) with a virus suspension containing 10^4.5 PFU.mL⁻¹ of H5N1 virus isolate A/Cattle/Texas/063224-24-1/2024 (TX2/24, GISAID accession number: EPI_ISL_19155861) clade 2.3.4.4b genotype B3.13. Following the challenge, we monitored the clinical course of H5N1 virus infection daily for 14 days post-challenge (pc) and recorded body temperature, body weight, and clinical signs. We developed a clinical score index (CSI) in collaboration with Cornell’s Center for Animal Resources and Education (CARE) veterinarians and used it to document observed clinical signs. Each animal was scored daily as follows: hunched posture (3 points), ruffled fur/piloerection (3 points), decrease in food and water intake (2 points), body weight loss >15% (10 points), and neurological signs (10 points). The sum of individual daily scores was calculated and graphed as total daily CSI (Fig. 3a). Animals from sham-immunized G1 (8/8) started showing clinical signs of HPAI H5N1 infection on day 1 pc. These clinical signs included depression, pyrexia (up to 41 °C), reduced feed intake, ruffled fur, and hunched back posture. Additionally, 3/8 of the animals presented a decrease in food and water intake, labored breathing, and difficulty in moving. All animals showed signs of paresis and overall muscle weakness. None of the H5-vaccine immunized animals showed any clinical signs of disease, and remained active without changes in their feed and water intake (Fig. 3a).

Fig. 3. Clinical outcome after H5N1 virus challenge infection.

a Daily clinical score index (CSI) of cats from sham [G1] and H5-vaccine [G2] groups challenged with HPAI H5N1 TX2/24 isolate (2 × 10^4.5 PFU/cat) (n = 8 cats in the sham, n = 8 cats in H5-vaccine groups). b Change in daily body temperature (°C). The dotted line is the threshold for fever. c Change (%) in daily body weight of cats from sham [G1] and H5-vaccine [G2] groups. d Survival (%) of cats from sham [G1] and H5-vaccine [G2] groups. Each dot in the bar graph indicates an individual cat from each group (mean with SEM, n = 8 cats/group). Statistical significance was determined by use of Two-way ANOVA with column means comparison (****P < 0.0001).

To assess the clinical protection conferred by the H5-vaccine, changes from baseline (day 0 pc) rectal body temperature were monitored (Fig. 3b). Animals in the sham-immunized G1 presented increased body temperatures from day 1 pc onwards, which persisted until the animals reached humane euthanasia points or until the end of the 14-day experiment. On day 1 pc, all animals from the sham-immunized G1 developed fever with a marked increase in body temperature (0.89–2.9 °C above baseline, up to 41 °C), with animals maintaining high body temperature (>1 °C above baseline) until days 5 to 9 pc. Notably, most animals in the H5-vaccine G2 (7/8) did not present increased body temperatures throughout the 14-day experiment (Fig. 3b). Only one animal in the H5-vaccine G2 (cat #5) – which had the lowest NAb titers post-immunization – presented a transient increase in body temperature.

The animals from the sham-immunized G1 started losing body weight on day 1 pc onwards until they reached humane euthanasia endpoints (≥20% body weight loss) or until day 9 pc for the animal (cat #3) that survived the H5N1 virus challenge (Fig. 3c). By day 3 pc, seven out of eight (7/8) cats had lost 5–10% of body weight, while one animal (cat #3) lost 2.4%. Importantly, this cat consistently lost less body weight compared to the rest of the animals from the sham-immunized group. By day 5 pc, seven out of eight (7/8) animals had lost 10% of their body weight. By day 7 pc, one out of eight (1/8) animals lost 20% of its body weight, reaching the pre-established humane endpoint for euthanasia. On days 8 and 9 pc, four additional animals reached 20% body weight loss and were euthanized. Importantly, cat #3 started to gain weight from day 9 pc onwards, however, it did not reach the baseline body weight until the end of the experiment on day 14 pc. None of the vaccinated animals lost body weight, rather they steadily gained body weight (1.5–11%) throughout the 14-day experiment (Fig. 3c). The survival rate (determined based on days of endpoint euthanasia) for the animals in the sham-immunized group was 12.5%, whereas 100% of the animals in the H5-vaccine group survived the H5N1 virus challenge (Fig. 3d). Together, these results demonstrate that the recombinant H5-vaccine provided near complete clinical protection against HPAI H5N1 virus challenge.

Recombinant H5-vaccine reduced virus shedding

Next, we assessed virus shedding in nasal (nasal swabs, NS) and oral (oropharyngeal swabs, OPS) secretions and feces (rectal swabs, RS). To this end, total RNA was extracted from NS, OPS, and RS swabs and subjected to an Influenza A type M gene-specific RT-qPCR assay (Fig. 4a, c, e). The viral RNA levels detected in nasal secretions from sham-immunized animals were significantly higher (p < 0.0001) on days 1, 3, 5, 7 pc with the peak viral RNA load (6.93 log₁₀ copy number.mL⁻¹) observed on day 5 pc (Fig. 4a). Similarly, viral RNA loads in oral secretions were significantly lower in H5-vaccine G2 animals than in sham-immunized G1 animals on days 1, 5, and 7 pc (p < 0.0001), with the peak viral RNA load in one of the infected cats (6.62 log_10.copy number.mL⁻¹) being detected on day 5 pc (Fig. 4c). Additionally, viral RNA in feces from H5-vaccine G2 animals was significantly lower (p < 0.0001) between days 3–7 pc with a peak RNA load (5.87 log₁₀ copy number.mL⁻¹) being detected on day 7 pc in sham-immunized G1 animals (Fig. 4e).

Fig. 4. Immunization with H5-vaccine resulted in reduced virus shedding.

Viral RNA copy number/mL (log₁₀) in (a) nasal swabs, (c) oral swabs, and (e) rectal swabs. Infectious virus titers were determined by endpoint titrations in Cal-I cells and represented as TCID₅₀.mL⁻¹ (log₁₀) in (b) nasal swab (d) oral swabs, and (f) rectal swabs. Each dot in the bar graph indicates an individual cat from each group (mean with SEM, n = 8 cats/group). Statistical significance was determined by the use of Two-way ANOVA with column means comparison. ****P < 0.0001, **P < 0.01.

We also assessed infectious virus shedding dynamics in the RT-qPCR-positive NS, OPS, and RS samples. Consistent with the findings of viral RNA, infectious virus loads in nasal and oral secretions, and feces were significantly (p < 0.0001) lower between days 1-7 pc in H5-vaccine G2 animals when compared to animals in the sham-immunized G1 (Fig. 4b, d, f). Notably, infectious virus was only detected in one OPS sample collected from one H5-vaccine group cat on day 3 pc (cat #5; 2.8 log₁₀ TCID_50.mL⁻¹; which had lower NAb titers). This cat did not further shed virus in any swab samples until the end of the 14-day experiment (Fig. 4d). These results demonstrate that immunization of cats with the recombinant H5-vaccine results in robust protection against H5N1 virus, limiting viral replication in the nasal, oral, and gastrointestinal tract of animals and markedly reducing virus shedding through nasal and oral secretions and feces (Fig. 4f).

Recombinant H5 vaccine reduces viremia in immunized animals

To assess the efficacy of the vaccine to prevent viremia, we collected EDTA whole blood and serum samples on days 1, 3, 5, 7, 10, and 14 pc. High levels of viral RNA loads were detected in whole blood samples by RT-qPCR in all animals (8/8) from the sham-immunized G1, while significantly lower (p < 0.0001) levels of viral RNA were detected in H5-vaccine G2 animals (Fig. 5a). Notably, virus titrations in whole blood showed high-level viremia (peaked as 2.8 log₁₀ TCID₅₀.mL⁻¹ at day 3 and 5 pc) in sham-immunized G1 animals, while in the H5-vaccine G2, only cat #5 presented a short-duration viremia (day 1 pc, 1.8 log₁₀ TCID₅₀.mL⁻¹) (Fig. 5b). Similarly, viral RNA levels in serum were significantly lower (p < 0.0001) in H5-vaccinated cats between days 1–7 pc, when compared to sham-immunized animals (Fig. 5c). Infectious virus was detected in only one animal (1.8 log₁₀ TCID₅₀.mL⁻¹ at day 3 pc) in the sham-immunized G1 on day 3 pc, with no infectious virus being detected in cell-free serum in the other animals (Fig. 5d). These results demonstrate that immunization of cats with the recombinant H5-vaccine markedly reduced H5N1 virus viremia in immunized cats.

Fig. 5. Immunization with H5-vaccine prevented viremia post-HPAI H5N1 challenge-infection.

Viral RNA copy number/ml (log₁₀) in (a) whole blood, and (c) serum. Infectious virus titers were determined by end point titrations in Cal-I cells and represented as TCID₅₀.mL⁻¹ (log₁₀) in (b) whole blood and (d) serum. Each dot in the bar graph indicates an individual cat from each group (mean with SEM, n = 8 cats/group). Statistical significance was determined by use of Two-way ANOVA with column means comparison. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05, ns > 0.05 (mean with SEM, n = 8 cats/group).

IgA and IgG responses in nasal and oral secretions and serum

Given the limited H5N1 virus replication observed in the respiratory tract and pharynx of H5-vaccine G2 cats, we assessed H5-specific mucosal IgG and IgA antibody levels in nasal and oral secretions using a bead-based serological assay using the Luminex platform (Fig. 6). IgG levels in nasal and oral secretions were significantly higher (p < 0.0001) in H5-vaccinated animals on day 35 post-immunization when compared to sham-immunized animals (Fig. 6a, c). These results were consistent with the higher (p < 0.0001) levels of IgG detected in serum from H5-vaccine G2 cats (Fig. 6e). The IgA levels did not increase in immunized G2 nor in sham G1 animals (Fig. 6b, d, f). These results demonstrate that parenteral immunization with the H5-vaccine candidate elicited robust levels of mucosal IgG in cats.

Fig. 6. Immunization with recombinant H5-vaccine elicited higher levels of IgG in nasal, oral secretions, and serum.

Shown are IgG levels (mean fluorescence intensity, MFI) innasal secretions (a), oral secretions (c) and serum (e). The IgA levels in nasal secretions (b), oral secretions (d) and serum (f). Each dot in the bar graph indicates an individual cat from each group (mean with SEM, n = 8 cats/group). Statistical significance was determined by the use of Two-way ANOVA with column means comparison. ****P < 0.0001, ns > 0.05.

Neutralizing antibodies elicited by the H5-vaccine in cats cross-neutralize the genotype D1.1 virus

We assessed the cross-reactivity of the H5-vaccine elicited NAb to the newly emerged H5N1 virus genotype D1.1. To this end, we used HPAI H5N1 virus isolate A/Canada Goose/03149/2025 genotype D1.1 to perform a virus neutralization assay with the serum collected post-immunization and post-challenge infection (Fig. 7c, d). Neutralizing antibodies elicited by the H5-vaccine efficiently neutralized the H5N1 D1.1 virus, with the levels of NAb comparable to those observed against HPAI H5N1 genotype B3.13 TX2/24 strain (Fig. 7a, b). Additionally, cats from the sham-immunized group presented robust serum antibody responses post-challenge infection, developing high NAb titers (4–7 log₂) by day 7 pc, which cross-reacted with both H5N1 B3.13- and D1.1 viruses (Fig. 7b and d). Notably, the H5-vaccine G2 animals only presented a slight increase in antibody titers at day 14 pc. One animal from H5-vaccine G2 (cat #5) showed a 3-fold increase in NAb titer (log₂) on day 14 pc as compared to NAb titers on day 1 pc (Fig. 7e). Comparison of the group geometric mean NAb titers on days 1, 7, and 14 pc, revealed that H5-vaccine G2 animals had a modest 1 to 1.5 (log₂)-fold increase in the NAb geometric mean titer (GMT), while sham-immunized G1 animals presented a 6 to 9 (log₂)-fold increase in NAb levels (Fig. 7f).

Fig. 7. Immunization with recombinant H5-vaccine elicited cross-reactive antibodies against HPAI H5N1 clade 2.3.4.4b.

Shown are the virus-neutralizing (100%) antibody titers (log₂) against HPAI H5N1 clade 2.3.4.4b genotype B3.13 on 0-day post-challenge (35 days post-immunization) (a) and 1–14 days post challenge (b). Shown are the virus-neutralizing (100%) antibody titers (log₂) against HPAI H5N1 clade 2.3.4.4b genotype D1.1 on (0 day post challenge (35 day post vaccination) (c) 1–14 days post challenge (d)). e Shown are the virus-neutralizing (100%) antibody titers (log₂) of individual animals against HPAI H5N1 clade 2.3.4.4b genotype B3.13 on 1, 7, and 14 days post challenge. f Shown is the geometric mean titer of sham-immunized and H5-vaccine groups against HPAI H5N1 clade 2.3.4.4b genotype B3.13. Each dot in the bar graph indicates an individual cat from each group (mean with SEM, n = 8 cats/group). From the sham-immunized challenged group, 7/8 animals were euthanized between 7 and 10 days post challenge. There was only a single cat in the sham group on 10 and 14 days post challenge. Statistical significance was determined by the use of Two-way ANOVA with column means comparison. ****P < 0.0001, ns > 0.05.

To identify the potential association of immune parameters elicited by immunization with the H5-vaccine with clinical parameters, we performed Pearson multiple correlation analysis among serum neutralizing antibodies (NAbs) detected on day 35 post-immunization and IgG levels detected in nasal and oral secretions and serum. Pearson r values are shown with color gradient and numbers (Fig. 8). Interestingly, we observed a strong correlation between levels of IgG in: nasal vs oral secretions (r = 0.80), oral secretions vs serum, (r = 0.80); a medium correlation in: nasal secretions vs serum (r = 0.62), serum vs serum NAb (r = 0.56), and a weak correlation in: nasal secretions vs serum NAbs (r = 0.31) and oral secretions vs and serum NAbs (r = 0.27).

Fig. 8. Multiple correlations matrix of IgG and clinical parameters.

The mean fluorescence intensity (MFI) from Luminex assays (NS IgG, OPS IgG and serum IgG) and serum titers of virus neutralization antibodies (serum NAbs) were transformed into log₁₀. Statistical significance was determined by use of multiple correlation analysis comparing clinical parameters from day 35 post vaccination (n = 8 cats/group).

Discussion

Currently, multiple vaccine platforms are being developed and tested against the HPAI H5N1 virus, ranging from the traditional whole inactivated virus-based vaccines to the most recent lipid nanoparticles (LNP) mRNA-based vaccine platform, for use in avian, human, and other mammalian species^27–30. The ideal vaccine should be safe to be administered, induce robust immune responses, reduce virus replication and shedding, and provide protection against disease and mortality³¹. In the present study, we developed a baculovirus-expressed recombinant H5-vaccine and investigated its safety and protective efficacy against HPAI H5N1 virus in domestic cats. We have demonstrated that immunization with the recombinant H5-vaccine candidate induced robust immune responses and provided nearly complete clinical protection against an HPAI H5N1 virus genotype B3.13 virus challenge, when compared to sham-immunized animals, in which HPAI H5N1 virus resulted in a fatal outcome.

The prime-boost vaccination strategy (days 0 and 21) used in our study resulted in high NAb and HAI antibody titers ranging from 8–12 log₂ and 9–10 log₂ on day 35 pi, respectively. One animal in the H5-vaccine group (cat #5) presented lower titers of NAb throughout the study and had a NAb titer of 8 log₂ on day 35 pi, the day of H5N1 challenge, which was 3-4-fold lower than the NAb titers detected in the other H5-immunized animals. The antibody responses elicited by immunization with the H5-vaccine resulted in protection against clinical disease following H5N1 virus challenge, as evidenced by the lack of clinical signs, fever, and body weight loss in 7/8 H5-immunized animals. Only cat #5 presented mild and transient clinical signs, including a slight increase in body temperature and mild depression on day 3 pc, which resolved by day 4 pc. Importantly, all animals (100%) in the H5-vaccine group survived, while only one cat (12.5%) in the sham immunized group did not reach endpoint euthanasia following challenge infection with H5N1 virus. These results demonstrate strong protection against disease and mortality induced by the H5N1 virus in felids, a species that is highly susceptible to infection with the virus.

One of the most important outcomes of immunization with the H5-vaccine in cats in our study was the significant reduction in viral RNA in nasal and oral secretions and feces. Notably, most of the animals in the H5-vaccine G2 shed no infectious virus at any of the time points sampled in our study. The exception to this was cat #5, which presented detectable infectious virus shedding in oral secretions on day 3 pc. These findings indicate that immunization of cats with the recombinant H5 vaccine developed here reduces virus shedding from infected animals, thus reducing the risk of transmission and spread of H5N1 virus between cats³² and from cats to other species, including humans.

Although several details of H5N1 virus pathogenesis in domestic cats remain unknown, infection in this species results in systemic spread and replication³² with the virus ultimately invading and replicating in the brain^11,19,32. Our results in sham-immunized animals demonstrate that the H5N1 B3.13 virus established viremia in infected animals, with both viral RNA and infectious virus being detected consistently in whole blood and low levels of viral RNA detected in serum. Importantly, animals in the H5-vaccine G2 presented significantly lower levels of viral RNA in blood, with nearly no detection of infectious virus. The only animal with detectable infectious virus in whole blood samples was cat #5 on day 1 pc. While we have not collected tissues from immunized animals over the course of infection, the reduced levels of viremia detected in H5-vaccine G2 animals could contribute to the differences in clinical manifestation and mortality observed between H5- and sham-immunized animals. The immune responses elicited by immunization limited viremia and, consequently, hampered systemic spread and replication of the H5N1 virus in target organs in immunized animals. This, however, needs to be confirmed experimentally.

The detection of high levels of H5-specific IgG antibodies in nasal and oral secretions suggests the development of mucosal immunity in the H5-vaccine group. These antibodies likely played a role in limiting H5N1 virus replication in the nasal and/or oral mucosa and consequently decreased virus shedding through nasal and salivary secretions. Interestingly, no IgA responses were detected in the nasal or salivary secretions of immunized animals. Our findings are similar to those described following cat infection with feline immunodeficiency virus (FIV), in which mucosal secretion IgG levels increased while IgA levels remained constant following FIV infection³³. Although IgG is the predominant immunoglobulin in blood³⁴, its presence at mucosal surfaces can occur through transudation from the blood. While not directly eliciting IgA responses, the H5 vaccine here induced robust systemic neutralizing/IgG responses, which translated into high IgG antibody levels in nasal and salivary secretions. These responses were likely critical to neutralize the virus at earlier stages of viral replication in the nasal and oral mucosa.

Importantly, NAbs elicited by immunization with the recombinant H5 vaccine cross-reacted with two contemporary H5N1 virus isolates belonging to genotypes B3.13 and D1.1, which share 98.94 and 98.24% amino acid identity with the vaccine virus HA (Supplementary Table 2). After challenge with the B3.13 TX2/24 virus, nearly full protection against clinical disease and mortality, and significant reduction in virus replication, virus shedding, and viremia were observed in H5 vaccinated animals. Importantly, the robust clinical and virological protection was supported by post-challenge serological data. Virus-neutralizing antibody data showed that sham-immunized animals developed an anamnestic NAb response with titers increasing 6–8-fold post-challenge infection. In contrast, H5-vaccinated animals presented only a modest booster response, with 1–2 fold increase in Nab titers being observed post-challenge. Together, these results demonstrate that the immunity elicited by the H5-vaccine in our study resulted in robust protection in immunized animals.

It is important to note that in spite of the robust NAb responses elicited by the H5-vaccine in cats in our study, cellular immune responses were not investigated, and thus, the contribution of T cells to the observed post-challenge protection is unknown. Another limitation is the fact that paired tissues from the H5-vaccinated and sham-immunized animals were not collected during the study. Therefore, direct assessment of the H5-vaccine on limiting the virus tropism and distribution, tissue replication, and damage was not determined. Another aspect of the H5-vaccine developed here that needs further investigation is the duration of the immunity³³. This would be important to determine appropriate intervals for revaccination of felids in the field. In summary, we demonstrate the safety and protective efficacy of a recombinant H5-vaccine against HPAI H5N1 virus genotype B3.13 in domestic cats. In addition to decreasing disease severity and mortality in immunized cats, the H5-vaccine developed here dramatically reduced infectious virus shedding, suggesting that it would be an efficient tool to combat H5N1 virus transmission and spread in felines³².

Methods

Ethics statement

All animals were handled in accordance with the Animal Welfare Act (AWA). The study procedures were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at Cornell University (Protocol No. 2024-0177).

Cells and viruses

Bovine uterine epithelial cells (Cal-1; developed in-house in virology laboratory at Animal Health Diagnostic Center (AHDC)) were propagated in minimum essential media (MEM, Corning, In, Corning, NY) supplemented with 10% fetal bovine serum (FBS), Anti-Anti (Gibco,15240062), and gentamycin (50 µg.mL⁻¹). Cells were maintained at low passage, and only cells with fewer than 48 passages were used in these studies. The HPAI H5N1 virus isolate TX2/24 (A/Cattle/Texas/063224-24-1/2024, EPI_ISL_19155861)¹⁶ belonging to genotype B3.13, was isolated from milk of infected dairy cattle and used for challenge infection and VN assays in this study. The HPAI H5N1 A/Canada_goose/03149/2025 genotype D1.1 (EPI_ISL_20276734)¹⁶ isolated from a Canada goose was used for the VN assay. The original virus stocks were propagated in 9–10-day-old embryonated chicken eggs and titrated in Cal-1 cells via TCID₅₀ and plaque assay.

Recombinant H5 HA vaccine

The wild-type H5 HA full-length coding sequence from a highly pathogenic avian influenza A virus A/common_tern/Maine/W22-480B/2022 (GenBank Accession Number OP377550) clade 2.3.4.4b genotype A2 was codon optimized for expression in Spodoptera frugiperda insect cells (Sf9). We engineered a pBacPAK8 cloning vector (Clontech Laboratories, Inc., Mountain View, CA) containing the H5 HA sequence with flanking sequences homologous to the polyhedrin gene of linear BestBac 2.0 Baculovirus backbone (Expression Systems, Davis, CA). Through homologous recombination, the resulting Baculovirus expressing the H5 HA gene (under the polyhedrin promoter) was generated and harvested. The Sf9 cells and virus were grown in culture media obtained from Expression Systems (ESF AF 99-300) formulated without animal origin ingredients. At the final harvest step, infected cultures were centrifuged to remove the cells, and the supernatant was processed through a 0.2-micron sterile disposable filter. Finally, Sf9 cells were scaled up to large production quantities utilizing sterile disposable plastic vessels, infected (0.00001-10 MOI) with the supernatant previously mentioned, and grown at 25–31 °C for 2–6 days. Harvest fluids were concentrated on volume 4–6x via centrifugation, disrupted via sonication, and inactivated with Beta-propiolactone (BPL) at a final concentration of 0.2–0.3%. The final pH of the inactivated antigen was adjusted to 7.0–7.5 and stored at 2–8 °C until formulation. The final formulation of vaccine antigen (determined via a proprietary assay) was prepared with 20% aluminum hydroxide adjuvant as per approved USDA outlines for similar vaccines^23,24. After vaccine formulation, antigen content was determined using Medgene Lab’s proprietary quantification assay and expressed as 7.7 antigen units (AU)/mL. The vaccine sterility and safety were confirmed as per 9CFR regulations.

Animal housing

A total of 16 domestic cats (Felis catus) (8 males and 8 females) at 12 months of age were obtained from Marshal Bioresources (New York, USA) and housed in two rooms with eight cats each co-housed per room for 6 weeks (immunization phase) in the animal biosafety level 1 (ABSL-1) facility at East Campus Research Facility (ECRF) at Cornell University. During the challenge phase, animals were moved into Horsfall HEPA filtered isolators and housed individually in the Animal Biosafety Level 3 (ABSL-3) facility at East Campus Research Facility (ECRF) at Cornell University. Food and water were provided ad libitum.

Animal immunization-challenge study

After acclimation for 72 h in ABSL-1, nasal swabs and whole blood samples were collected from all animals for HPAI H5N1 RT-qPCR and serology testing, respectively. The nasal swab samples were tested using an influenza A M-specific RT-qPCR assay and the Path-ID Multiplex One-Step RT-PCR kit (ThermoFisher, described below) for HPAI H5N1 viral nucleic acid, and the serum samples were used for a virus neutralizing assay using a bovine HPAI H5N1 TX2/24 virus expressing a miniGFP reporter gene, as previously described³⁵. All animals tested negative for H5N1 viral RNA in nasal swabs and virus-neutralizing antibodies (NAbs) in the serum.

Animals were then randomly allocated into two experimental groups as follows: group 1 (G1, n = 8) sham-immunized, H5N1-virus-challenged; and group 2 (G2, n = 8) H5-vaccine immunized, H5N1-virus-challenged. Prior to immunization on day 0, basal rectal body temperature was taken from all animals and recorded. Immunization was performed by subcutaneous (SC) injection on the dorsocentral area of the neck with 1 mL of sham vaccine (G1) or the recombinant adjuvanted H5-vaccine candidate (G2). Animals were monitored for body temperatures and by physical examination at the injection site for 48 h post-immunization (pi). On day 21 pi, a booster dose was administered to all animals as described for the prime immunization, and animals were monitored for 48 h as described above. Blood samples were collected on days 0, 14, 21, and 35 pi, while nasal, oral, and rectal swabs were collected on days 0 and 35 pi. All sampling was conducted with animals under isoflurane anesthesia. Rectal body temperature was taken on days 0- 3 to monitor animal response to the priming dose, on days 21- 23 to assess for the boosting dose, and on day 35 pi to establish baseline prior to challenge.

On day 35 post immunization (day 0 post challenge (pc)), animals were moved into the ABSL-3 laboratory for challenge with HPAI H5N1 TX2/24 virus. All animals from G1 (n = 8, sham) and G2 (n = 8, H5-vaccine) were inoculated oronasally (1 mL intranasally; 0.5 mL per nostril; and 1 mL orally), with a virus suspension of bovine HPAI H5N1 TX2/24 virus (A/Cattle/Texas/06322424-1/2024) containing 10^4.5 PFU.mL⁻¹. Following inoculation, animals were monitored daily for clinical signs compatible with influenza A H5N1 virus infection. A clinical score index (CSI) was used to score the daily clinical signs from day 1 to 14 pc. The CSI was calculated based on daily observations on individual animals and included the following five clinical signs (Supplementary Table 1) and respective scoring system (based on points): hunched posture (3 points), ruffled fur (3 points), decreased eating or drinking as evidenced by less than ¼ of the food and water provide being consumed (2 points), >15% body weight loss (10 points), neurological signs (e.g. hind-limb paralysis, tremors, etc.) (10 points)³⁶. Each animal was scored daily, and if the CSI of any animal increased to a score ≥15 the frequency of monitoring was increased. The presence or absence of these signs was scored, with the presence receiving the full score for each clinical sign and the absence receiving a score of 0. This ensured objectivity of the CSI and increased the stringency of the clinical evaluation. Rectal body temperature and body weight were measured and recorded daily throughout the challenge phase of the experiment, and body weight loss ≥20% was considered the humane endpoint for euthanasia and necropsy. Nasal (NS), oropharyngeal (OPS), and rectal (RS) swabs were collected on days 1, 3, 5, 7, 10, 14 pc, in 1 mL virus transportation media (VTM) and stored at −80 °C until further use. Blood samples were collected in EDTA-coated tubes and were stored in both −80 °C as well as in a fridge at 4 °C. Serum samples were collected in serum separation tubes on days 1, 3, 5, 7, 10, and 14 pc. All animals were euthanized in accordance with the American Veterinary Medical Association (AVMA) Guidelines for the Euthanasia of Animals. Briefly, animals were sedated with intramuscular dexmedetomidine (Dexdomitor®, 0.01–0.04 mg/kg), followed by induction of deep anesthesia using inhaled isoflurane (3–5%). Depth of anesthesia was confirmed by the absence of the pedal withdrawal reflex. Once deep anesthesia and unconsciousness were verified, blood was collected via the intracardiac route until respiratory arrest occurred. Animals were then monitored for a minimum of three minutes to confirm death prior to the initiation of necropsy.

Real-time RT-qPCR

Total RNA was extracted from 200 µl of NS, OPS, and RS by using the IndiMag Pathogen kit (Indical Bioscience SP947855P196, Germany) in an automatic KingFisher flex (ThermoFisher). The RT-qPCR was performed on extracted RNA using the Path-ID Multiplex One-Step RT-PCR kits (ThermoFisher). The primers and probes targeting the M gene of IAV were synthesized (IDT DNA); M + 25: AGA TGA GTC TTC TAA CCG AGG TCG, M-124: TGC AAA AAC ATC TTC AAG TCT CTG, M + 64 Probe: FAM-TCA GGC CCC CTC AAA GCC GA-BHQ1 and used with the following thermal conditions: 15 min at 48 °C, 10 min at 95 °C, and 40 cycles of 15 s at 95 °C for denaturation and 60 s at 60 °C. All samples with Ct values below 40 were considered positive. Samples in which fluorescence did not cross the threshold within 40 cycles were considered negative. A standard curve using 10-fold serial dilution of HPAI H5N1 TX2/24 virus stock (2 × 10⁷log TCID₅₀ .mL⁻¹) RNA was used to determine the viral RNA copy number (log₁₀.mL⁻¹) using the relative quantification method, and values were plotted with GraphPad Prism Prime Software version 10.1.2 (GraphPad, La Jolla, CA, USA). Negative samples were attributed a copy number of zero.

Virus titrations

The RT-qPCR positive samples were processed for endpoint virus titrations under BSL-3 conditions in the research suite at the Animal Health and Diagnostic Center (AHDC), Cornell University. Briefly, Cal-1 cells were seeded in 96-well plates at 2.5 x 10³ per well, and 24 h post-seeding, the plates were moved to BSL3. A 10-fold serial dilution from 10⁻¹ to 10⁻⁶ of the test samples were prepared with plain MEM, and quadruplicate wells were inoculated with 50 μl of each dilution. After 1 h of incubation at 37 °C, 50 μl of complete growth media, MEM supplemented with 5% FBS, was added to each well of the 96-well plate. After 48 h of incubation at 37 °C, cells were fixed with 3.7% formalin for 30 minutes at RT, and an immunofluorescence assay (IFA) was performed using the mouse anti-NP mAb (HB65) mouse monoclonal antibody, followed by goat anti-mouse IgG (H&L) conjugated with Alexa Fluor 594 (ImmunoReagents Laboratories) secondary antibodies. The virus titers from endpoint dilutions were calculated using Spearman and Karber’s method and expressed as log₁₀ (TCID_50. mL⁻¹). The limit of detection of our titration assays was 10^1.05 log10 TCID₅₀.

Virus neutralization assay

The HPAI H5N1 genotypes B3.13 and D1.1 were used to determine virus NAb responses in serum samples from sham (G1) and H5-vaccine (G2) animals. Briefly, 75 μl MEM was added to the first column (A1 to H1) of a 96-well plate, and 50 µl MEM was added to the remaining wells of the 96-well plate. Heat-inactivated serum (25 µl) was added to the first well, mixed 5 times with a pipette, and 50 μl transferred to the second well. A two-fold serial dilution was prepared similarly until the last well on the plate (1:8 to 1:16384). A virus suspension containing 200 TCID_50.mL⁻¹ (0.001 MOI) in 50 µl (MEM) was added to all wells, and the virus-serum mixture was incubated at 37 °C for 1 h. A known positive serum sample was used as a positive control, and the virus inoculum was back-titrated by inoculating serial dilutions from 10⁻¹ to 10⁻³ in Cal-1 cells. Cal-1 cells (approximately 2.5 x 10³ per well) were added directly to 96-well plates with virus-serum. Following 48 h of incubation at 37 °C, cells were fixed with 3.7% formaldehyde, and IFA was performed as described above. All samples with 100% neutralization at the initial serum dilution of > 3 log₂ (1:8) were considered positive.

Hemagglutination inhibition assay

We assessed HAI antibody responses by using the hemagglutination inhibition (HAI) assay. Briefly, serum samples were heat-inactivated at 56 °C for 30 minutes prior to performing standard HAI protocol. Serum samples were prepared by adsorbing with 10% turkey red blood cells (RBCs), 0.16 M potassium periodate, and 3% glycerol, and subsequently tested by using a standard protocol using turkey RBCs³³. Two-fold serial dilutions (1:4 to 1:8192) of treated sera (25 µl) were prepared in 96-well U-bottom microtitration plates. Four hemagglutination units (HAU) of the HPAI H5N1 virus TX2/2024 isolate (per 25 µl) were added to each well containing the serum dilutions (1:4) and were incubated for 30 minutes at RT. Subsequently, 25 µl of 0.5% RBCs were added to each well and incubated for 45 minutes at room temperature. The HAI titer was determined as the highest dilution of the serum that completely inhibited hemagglutination, as indicated by RBC button formation at the bottom of the well. The assay included virus and cell control and was performed in duplicates. The HAI titers were log-transformed before statistical analysis. According to the World Organization for Animal Health (OIE) guideline manual³³, the samples with an HAI titer of > 3 log₂ were considered seropositive.

Recombinant H5 HA1 protein expression

To facilitate expression and purification, the globular domain of the HA protein, HA1 (aa 26-354), was selected for recombinant protein expression and purification. A previously described secretion system for mammalian expression using equine IL-4 as a fused secretion factor was employed³⁷. A pCAGGS-N-IL4 vector was used to clone HA1, using the restriction enzymes XmaI and NheI, and a 6xHis tag was added by primer to allow protein purification.

Ten 150 cm dishes were seeded with 6 x 10⁶ HEK-293T cells and were incubated at 37 °C for 24 h. Just prior to transfection, the cell culture media were removed, and the cells were rinsed twice with 1x PBS to remove residual FBS. 25 mL of warmed plain DMEM (Gibco 11965092) was added to the plates. Transfection complexes were prepared in plain DMEM with 10 μg DNA per plate and 20 μl (1 mg.mL⁻¹) PEI prime (SeroChem P100-1G). Transfected plates were incubated at 37 °C for 48 hours (h). Cell media containing the secreted protein was collected and cleared at 500 x g for 20 minutes at 4 °C. Cleared protein suspensions were concentrated to 5 mL using centrifugal concentrator units with a 10 kDa molecular weight cut off (MWCO) (Millipore UFC9010). The concentrated suspension containing the secreted protein was then dialyzed three times for 8 h at 4 °C with stirring using 500 mL standard nickel column purification buffer (10 mM HEPES, 10 mM imidazole, 500 mM NaCl, 5% glycerol, 5 mM β-mercaptoethanol). The protein was then purified using HisTrap HP 5 mL columns (Cytiva 17524802) and the AKTA Pure system (Cytiva). Bound proteins were eluted using an imidazole buffer gradient (30 – 480 mM imidazole). Peak fractions were collected and analyzed by 10% SDS-PAGE (Biorad 1610173) run at 110 V for 80 minutes and SimplyBlue™ SafeStain Coomassie staining (Invitrogen LC6060). Fractions containing the protein were pooled, concentrated as described earlier, and then dialyzed 3 times for 8 h at 4 °C with stirring using 500 mL sterile PBS. Protein concentration was determined using the Qubit Protein Assay (ThermoFisher Q33212), and aliquots were frozen at −80 °C. The purified recombinant HA1/H5 protein was used in the Luminex assays as described below.

IgA and IgG Luminex

Isotype-specific antibody assays based on the Luminex platform were performed to assess H5 HA-specific IgA and IgG antibody responses in nasal secretions, oral secretions, and serum samples collected at days 0 and 35 pi. Serum samples were diluted 1:1000 in phosphate buffer saline (PBS) with bovine serum albumin (BSA; 1%) and sodium azide (PBN, prepared in-house), and undiluted nasal and oral samples (swabs dipped in 1 mL of VTM) were plated at 50 μl per well into a pre-washed 96-well plate with a hydrophilic PVDF membrane (Millipore Sigma MSGVN2250). Controls, including blanks and low, medium, and high antibody concentration samples, were added in duplicate starting at wells A1 and B1. After loading the samples and controls, 50 μl of pre-mixed beads (b35, Luminex Corp, USA) pre-coupled with monoclonal anti-equine IL-4 antibody (clone 25)³⁸ and H5 HA1-IL4 fusion protein, were dispensed into each well, and plates were incubated on a shaker (600 rpm) for 30 minutes at room temperature (RT). Following the shaking incubation, all beads were washed by adding 300 μl per well of PBS-TBN (PBN + 0.02% TWEEN-20), followed by aspiration from the well bottom three times. The process was automated by an Agilent BioTek 50 TS Microplate Washer. After the washes, 50 μl of a biotinylated Goat anti-Feline IgA (H) antibody (Novus Biologicals NB7262B), and Goat anti-Cat IgG (H + L) (Jackson ImmunoResearch Laboratories 102-065-003) antibody diluted 1:1000 in PBN, was added to each well and incubated for 30 minutes at RT. After incubation, all wells were washed. Following the wash, 50 μl of streptavidin-phycoerythrin (SAV-PE, Invitrogen™S866), diluted 1:100 in PBN, was added and incubated for 30 minutes at RT. Plates were washed a final time, and the beads were resuspended in 100 μl of PBN per well and shaken for 15 minutes on a shaker (600 rpm). The plate was read on Luminex™ 200™ (Luminex Corp., http://www.luminexcorp.com), only wells with bead counts ≥100 were considered valid for analysis, and data were reported as mean fluorescence intensity (MFI) values.

Statistical analysis

Statistical analysis was performed by two-way analysis of variance (ANOVA) followed by multiple comparisons, and mean and standard error of mean values were presented. Multiple correlation analysis was performed on the log₁₀-transformed data, and correlation coefficient (r) values are presented GraphPad Prism Prime (version 15.1) was used to plot the data. Figures 1 and 2a were prepared using BioRender.com.

Author contributions

Conceptualization: D.G.D.; Methodology: S.L.B, R.R., M.N., P.B.S.O., A.N.D., S.G., A.Y., B.S.; Software: S.L.B., P.B.S.O.; Validation: S.L.B., P.B.S.O., A.N.D., S.G., A.Y; Formal analysis: S.L.B., P.B.S.O., D.G.D.; Investigation: S.L.B., R.R., M.N., P.B.S.O., D.G.D; Resources: D.G.D., A.Y., S.G.; Data curation: S.L.B., P.B.S.O.; Writing-original draft: S.L.B., D.G.D.; Writing-Review & Editing: S.L.B., R.R., M.N., P.B.S.O., A.N.D., S.G., A.Y.; Visualization: S.L.B., D.G.D; Supervision: D.G.D.; Project administration: D.G.D.; Funding acquisition: D.G.D.

Data availability

All data supporting the findings of this study are available within the manuscript and its supplementary information files.

Competing interests

The work was funded by VST LLC dba Medgene. A.Y., S.G,. and BS are employed by Medgene Labs. All other authors declare no conflict of interest. A patent application has been filed by VST LLC dba Medgene (Applicant), with Alan John Young listed as the inventor. The applications include U.S. Serial No. 19/079,402 and PCT/US25/19870 and are currently pending. These filings cover aspects of the immunogenic composition described in this study, including its formulation, methods of production, methods for eliciting an immune response, therapeutic applications, and assay-based detection of pathogens using the composition. All other authors declare no conflict of interest.

Supplementary information

The online version contains supplementary material available at 10.1038/s41541-025-01369-6.