Protective Efficacy of an Inactivated H5N8 Avian Influenza Virus Vaccine Against Highly Pathogenic Avian Influenza H5N1 Viruses Belonging to Clade 2.3.4.4b in Chickens and Ducks Open Access

Avian influenza virus (AIV) belongs to the family Orthomyxoviridae and the genus Influenzavirus A (1). The virus has a negative-sense, single-stranded, and segmented RNA genome and contains eight gene segments encoding at least 10 proteins: polymerase basic 1 (PB1), PB2, polymerase acid, hemagglutinin (HA), nucleoprotein, neuraminidase (NA), matrix 1 (M1), M2, and nonstructural 1 and 2. The HA and NA proteins are surface glycoproteins and important for virus infectivity. The HA protein is responsible for virus attachment to the host cell and the major target of the humoral immune response. The NA protein plays a role in release and spread of progeny virions by removing sialic acid from glycoproteins. The natural reservoirs of the virus are aquatic birds, with ducks, gulls, and shorebirds being the primary hosts, which has contributed to the wide geographic spread and distribution of circulating viruses (2). On the basis of antigenic specificity, 16 HA types and 9 NA types have been detected in viruses isolated from wild waterfowl.

Highly pathogenic avian influenza (HPAI) virus is an economically important pathogen of poultry worldwide (3). Aquatic birds play an important role in transmitting the virus into susceptible poultry. The outbreaks involving H5 and H7 subtypes of HPAI viruses have resulted in lethal infections in poultry, thus affecting poultry production and trade (4). In particular, continuous circulation of HPAIV has led to a reassortant H5N8 virus (clade 2.3.4.4) in which the HA gene segment is from an H5N1 HPAI virus and other gene segments are from several other AI viruses circulating in eastern China (5). Since 2014, H5N8 viruses have spread rapidly via migratory aquatic birds. Reassortment events of H5N8 virus with low pathogenic avian influenza viruses further led to the divergence of H5 viruses into distinct subtypes, including H5N1, H5N2, and reassortant H5N8 (6). The phylogenetic clade 2.3.4.4 of these reassortant H5Nx viruses has further evolved into eight subclades (2.3.4.4a–2.3.4.4h) (7). In particular, two distinct genetic groups of H5 HPAI viruses, clades 2.3.4.4a and 2.3.4.4c, caused intercontinental outbreaks in 2014 to 2015 and 2016 to 2017, respectively (8). In the winter season of 2020 and 2021, H5Nx viruses (clade 2.3.4.4b) have emerged and spread globally. The virus was detected in many countries in Africa, Asia, and Europe (9). The epizootic has led to unprecedented numbers of deaths in wild birds and caused outbreaks in domestic poultry. In late 2021 these viruses crossed to North America and subsequently South America in the autumn of 2022. Phylogenetic analyses of the viruses confirmed a close relationship to HPAI H5N1 genotypes between Europe and North America (10). Importantly, there has been an increased spillover to nonavian species including wild terrestrial and marine mammals and, more recently, the detection of an outbreak in a mink farm in Spain (11).

Many countries are experiencing the most devastating HPAI epidemic. For instance, Europe had a total of 2,520 outbreaks in poultry, 227 outbreaks in captive birds, and 3,867 HPAI virus detections in wild birds between October 2021 and September 2022 (12). The unprecedented geographical extent resulted in 50 million birds culled in affected establishments. Depopulation of infected flocks has been commonly used to control the spread of avian influenza viruses in many countries. However, the number of HPAI virus (HPAIV) outbreaks has greatly increased, causing a devastating impact on commercial poultry. Therefore, vaccination of poultry against HPAI viruses could play an important role in reducing virus shedding and raising the threshold for infection and transmission to other birds and mammalian species including humans (13). However, rapid change in the antigenicity of H5 HPAIV has been a challenge for efficient control of HPAIV infection. Antigen matching between the vaccine and the currently circulating field strains is a critical factor in AIV vaccine efficacy (14). We have previously developed an AIV vaccine by inactivating H5N8 (clade 2.3.4.). Based on HI data and antigenic mapping, the OFFLU network has shown that currently circulating H5N1 viruses are antigenically similar to H5N8 viruses of the same clade (https://www.offlu.org/wp-content/uploads/2023/11/OFFLU-AIM-REPORT-2023.pdf). In this study, we report the protective efficacy of the vaccine against currently circulating H5N1 (clade 2.3.4.4.b) by vaccinating chickens and ducks with a single vaccination or after two vaccinations with an inactivated H5N8 vaccine.

Forty white leghorn chickens (Gallus gallus) were hatched from specific-pathogen-free (SPF) embryonated chicken eggs incubated in the animal facilities of the Istituto Zooprofilattico Sperimentale delle Venezie (IZSVe), Legnaro, Italy. Forty one-day-old SPF Muscovy ducks (Cairina moschata) were purchased from a local hatchery. Birds were housed in negative high-efficiency particulate air filtered poultry isolators (Montair, The Netherlands) with feed and water ad libitum. Animal experimental procedures were conducted in strict accordance with the Decree of the Italian Ministry of Health no. 26 of 4 March 2014 on the protection of animals used for scientific purposes, implementing Directive 2010/63/EU, and approved by the Institute’s Ethics Committee.

The birds were immunized with the inactivated oil-adjuvanted vaccine Vaxigen® Flu-H5N8 2.3.4.4 (Avimex, CDMX, Mexico), a reverse genetic engineered virus expressing the HA and NA genes of the H5N8 virus A/green-winged teal/Egypt/877/2016 of the 2.3.4.4b clade (EPI_ISL_267136) in the PR8 backbone. For the challenge of SPF chickens, we used the H5N1 HPAI A/turkey/Italy/21VIR9520-3/2021 strain (ch/9520-3) (GISAID EPI ISL 17684608), and we selected the H5N1 HPAI A/mallard/Italy/22VIR9219-6/2022 (mal/9219-6) (EPI ISL 14760557) for the challenge of Muscovy ducks. Both the ch/9520-3 and mal/9219-6 isolates are part of IZSVe’s virus repository and have been previously genetically characterized as belonging to the 2.3.4.4b clade. The HA gene segments of vaccine and challenge viruses were aligned using MUSCLE (15). Amino acid changes between the challenge and vaccine viruses at putative antigenic sites of the HA1 protein were evaluated (16). The prediction of potential N-glycosylation sites was performed using NetNGlyc server 1.0 (http://www.cbs.dtu.dk/services/NetNGlyc). Titration of the virus stock was carried out by inoculation of serially diluted virus into 9-to-11-day-old SPF embryonated chicken eggs. Allantoic fluids from dead embryos were tested by the HA assay, and positivity was recorded up to 7 days postinoculation (pi). The mean embryo infectious dose (EID₅₀) was calculated according to the Reed and Muench method (17).

For experiment 1, 40 one-day-old chicks were divided into four groups of 10 birds. The chicks in the nonvaccinated control groups were injected subcutaneously (sc) with 0.5 ml of phosphate buffered saline (PBS) at 10 days of age (Prime sham) or at 10 and 31 days of age (Boost sham), respectively. In the groups with a single vaccination (Prime vaccine) or with two vaccinations (Prime-Boost vaccine), birds received 0.5 ml of the inactivated vaccine by sc injection at 10 days of age or at 10 and 31 days of age, respectively. Sham and vaccine injections were performed in the distal part of the neck. The Prime sham and Prime vaccine groups were challenged at 31 days of age, 21 days after vaccination. Groups that received two injections were challenged at 52 days of age, 21 days after the second injection. The chickens were challenged with 10^6.2 EID₅₀/100 µl by the oronasal route, while ducks were challenged with 10⁶ EID₅₀/150 µl by the same route. The challenged birds were daily monitored for the appearance of clinical signs and mortality. Oropharyngeal and cloacal swabs were collected at 2, 4, 6, and 8 days postchallenge (dpc). Serum samples were collected from the birds on the day of challenge and 14 dpc to monitor antibody responses to both vaccination and infection. At the end of the study, the surviving animals were humanely euthanatized.

For experiment 2, 40 Muscovy ducks were immunized, challenged, and evaluated as described for the chickens in experiment 1 with minor modifications. The ducks were challenged with 10⁶ EID₅₀/150 µl instead of 10^6.2 EID₅₀/100 µl, and the oral swabs were collected from the oropharynx.

The oropharyngeal and cloacal swabs were placed in 750 µl and 1000 µl of PBS for experiments 1 and 2, respectively, and vortexed for 30 sec. Viral RNA was isolated from 200 µl of sample suspension using the MagMAX™ Pathogen RNA/DNA Kit (Applied Biosystems, Waltham, MA) and the KingFisher™ instrument (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s instructions. The purified viral RNA samples were further subject to real-time quantitative reverse transcription polymerase chain reaction (RT-qPCR) targeting the matrix protein gene (18) with the QuantiTect Multiplex RT-PCR Kit (Qiagen, Hilden, Germany) using a CFX 96 Deep Well Real-Time PCR System (Biorad, Hercules, CA, USA). Tenfold serial dilutions (10⁻¹ to 10⁻⁹) of the challenge viruses were processed in triplicate along with each testing run to develop standard curves for virus quantification. Viral load was expressed in terms of EID₅₀ equivalents. The limit of detection (LoD) of the RT-qPCR was 10.0 EID₅₀/100 µl for the mal/9219-6 and 14.8 EID₅₀/100 µl for the ch/9520-3 viruses, respectively. Data were analyzed with the Bio-Rad CFX Manager software (Version 3.1). Triplicates at a concentration of 10^2.0 EID₅₀/100 µl in the standard curve had mean RT-qPCR Ct values of 33.80 and 34.42 for the mal/9219-6 and for the ch/9520-3 viruses, respectively. Based on the expert opinion of researchers and diagnosticians at IZSVe, samples recording these values or greater are highly unlikely to yield viral progeny upon isolation attempts in embryonated eggs. For this reason, in this study swabs with a genomic load ≤ 10^2.0 EID₅₀/100 µl equivalents were considered as low viral load samples. This assumption is supported by results from a large validation study conducted by Munster et al. (19), who demonstrated that samples with mean Ct values ≥ 32.1 negatively correlated with the ability to isolate virus in embryonated hens’ eggs. Viral shedding was also quantified as a cumulative measure over time using the trapezoidal method to infer the area under the curve (AUC) (20).

The hemagglutination inhibition (HI) tests were carried out according to standard procedures (21) to determine pre- and postchallenge antibody titers. Both homologous vaccine antigens and the challenge viruses were used as HI antigens. In addition, sera were analyzed with a competitive ELISA kit for the detection of influenza A nucleoprotein using the ID Screen® Influenza A Antibody Competition Multi-Species (NP-ELISA, IDvet, France). For the NP-ELISA using sample-to-negative (S/N), sera with percentages ≤ 45% were scored as positive, between 45% and 50% sera were considered doubtful, while sera with percentages ≥ 50% were recorded as negative. To monitor the kinetics of the antibody responses against the H5 hemagglutinin, we used the ID Screen® Influenza H5 Indirect ELISA (FLUH5S, IDvet, France) according to the manufacturer’s instructions operating a 1:500 dilution of the sample. For the FLUH5S kit, sera with a sample-to-positive (S/P) ratios ≥ 0.5 were scored as positive.

In chickens, the FLUH5S was also used using an experimental higher dilution of 1:5000 to explore differences in the dynamic range for high-titer sera. An indirect measure of the dynamic range of this assay was obtained by calculating the coefficient of variation of S/P at the two dilutions and was compared to the one of the HI values.

To monitor seroconversion to the H5N1 challenge virus, both pre- and postchallenge sera were tested with an in-house ELISA assay for the detection of anti-N1 antibodies (22). In brief, the assay is a liquid-phase blocking ELISA (LPBE) in which plates are coated with a specific anti-N1 monoclonal antibody (mab) developed against the A/goose/Italy/296426/03 H1N1 virus. Plates are incubated with mixtures of antigen and test sera previously diluted. Upon washing of unbound material, plates are incubated with the homologous an anti-N mab conjugated with horseradish peroxidase to allow a colorimetric reaction. For S/N percentages ≤ 25%, sera were scored as positive.

Viral loads, HI titers, and ELISA S/N and S/P ratios were tested for their distribution and according to the results were analyzed using either Student’s t-test or the Wilcoxon test. Survival curves were compared using Kaplan-Meier survival analysis with log rank significance test. P values at or below 0.05 were considered statistically significant. Graphs show individual values and standard deviations of the mean and geometric mean.

At the level of the HA gene segment, both viruses have a nucleotide identity with the vaccine seed strain of 97.2% accounting for three amino acid substitutions, namely, T140A, N236D, and E268G. Only the mutation at the 140 position falls in the antigenic site A, within the globular head of the hemagglutinin. Moreover, these mutations do not alter the number and position of the potential N-linked glycosylation sites observed for the vaccine seed strain. In particular, predictive algorithms identified for both the challenge and the vaccine viruses three potential N-glycans at positions 11–14, 23–26, and 542–545 in nonantigenic sites of the HA protein.

In experiment 1, protection by inactivated H5N8 vaccine was evaluated in SPF chickens after a single vaccination (Prime vaccine) or after two vaccinations (Prime-Boost vaccine). One day postchallenge all chickens in the Prime sham control group showed clear pathological signs, such as depression, hemorrhages at the combs and limbs, ruffled feathers, and immobility. All birds succumbed to infection within 48 hr with a mean death time (MDT) of 2 days. In contrast, all chickens in the Prime vaccine group survived (Fig. 1A) without showing any clinical signs. Mortality in the Prime-Boost sham group was 50%, 90%, and 100% on 2, 3, and 6 dpc, respectively, with a MDT of 2.8 days. All Prime-Boost–vaccinated chickens survived (Fig. 1A) without showing clinical signs.

The efficacy of protection in Muscovy ducks was tested in Experiment 2. Muscovy ducks in the Prime sham group started to show a modest decrease in the activity levels on 3 dpc. Neurological signs, such as head tilt, incoordination, abnormal gait, and severe depression, started on 4 dpc in one duck, which died the following day (Fig. 1B). On 6 dpc, two ducks exhibited similar neurological clinical signs, and one of those birds died on day 7. Between 8 and 10 dpc all remaining ducks showed symptoms but did not die and were recovered at day 14 dpc. Overall, 20% mortality and 100% morbidity were recorded for this group. All ducks receiving a single vaccination showed mild depression starting at 6 dpc, with one bird showing a moderate head tilt at dpc 9. There was 0% mortality and 10% morbidity at the termination of the experiment at 14 dpc with the absence of severe clinical disease (Fig. 1B).

In the Prime-Boost sham group, the H5N1 challenge caused severe disease and 100% mortality (Fig. 1B). At 3 dpc, three ducks started to show moderate-to-severe neurological signs such as opisthotonos, head tilt, and incoordination, and at 4 dpc all ducks manifested clear signs of infection. Two ducks died on 4 dpc, and the remaining ducks died on 5 dpc with a MDT of 4.8 days. In contrast, all ducks in the Prime-Boost vaccine group survived the challenge (Fig. 1B) without showing clinical signs.

Detection of virus shedding in chickens was assessed using oropharyngeal and cloacal swabs in RT-qPCR assays. Because of the fast progression of disease in chickens in the control group for the single vaccination, only one bird was still available on 2 dpc. Oropharyngeal and cloacal swabs of this bird had a viral load of approximately 10^4.5 EID₅₀/100 µl. In the Prime vaccine group, only 1/10 birds had positive oropharyngeal swabs with loads ranging 25 to 660 EID₅₀/100 µl (Fig. 2A). In the Prime-Boost sham group, all oropharyngeal and cloacal swabs were positive on 2 dpc with viral loads ranging from 10³ to 10⁶ EID₅₀/100 µl. On the other hand, all oropharyngeal and cloacal swabs remained negative after two vaccinations (Fig. 2B).

Virus shedding in the Prime sham group of ducks was detected in all oropharyngeal and cloacal swabs at 2, 4, 6, and 8 dpc (Fig. 3A). The highest level of viral shedding was detected at 4 dpc with oropharyngeal and cloacal viral loads of 10^5.5 and 10^4.2 EID₅₀/100 µl, respectively. On 6 and 8 dpc shedding declined throughout the group. In these unvaccinated ducks challenged at 31 days the cumulative viral load expressed as AUC reached a value of 10^6.3 in the oropharyngeal swabs and it was higher than the cumulative viral load of 10^5.0 detected in the cloacal swabs (Fig. 4A).

In the Prime vaccine group, both oropharyngeal and cloacal swabs had lower GMTs at all sampling times than in the Prime sham group. Nonetheless, the shedding kinetic was comparable to the control birds, as the peak of shedding occurred at 4 dpc in both groups with a gradual decline at 6 and 8 dpc. On days 4 and 6 pi, the differences in viral load in the oropharynx between the control and vaccinated groups were significant (P < 0.05), recording mean viral loads of 10^4.7 and 10^2.9 EID₅₀/100 µl, respectively. This difference corresponds to 6- and 18-fold reductions compared to the Prime sham group. On days 6 and 8 pi, 7/10 and 3/10 ducks in the Prime vaccine group had viral loads > 10^2.0 EID₅₀ from the oropharynx, respectively, while 3/10 and 7/10 ducks had titers < 10^2.0 EID₅₀, respectively. In contrast, 10/10 and 6/10 sham vaccinated ducks had viral loads > 10^2.0 EID₅₀ from the oropharynx (Fig. 3A), suggesting a reduction in viral shedding in the vaccinated ducks.

The number of ducks in the Prime vaccine group shedding low viral loads (<10^2.0 EID₅₀) from the cloaca on 2, 4, 6, and 8 dpc was 4/10, 2/10, 4/10, and 7/10, respectively. On 8 dpc cloacal shedding among vaccinated birds was significantly lower with a ninefold reduction compared to the sham group (P < 0.005), and only 1/10 birds in the sham group was negative (<10^2.0 EID₅₀).

Comparing the cumulative oropharyngeal shedding between the Prime sham (AUC = 10^6.3) and the Prime vaccine group (AUC = 10^5.7), the AUC was significantly reduced (P < 0.005) in the latter group by a factor of 3.7. However, there was no significant difference between the two groups for cumulative cloacal shedding, with AUC values of 10^5.0 for both groups (Fig. 4A).

All ducks in the Prime-Boost sham groups had positive oropharyngeal and cloacal swabs on 2 and 4 dpc. The peak of shedding occurred on 4 dpi with GMT of 10^5.9 and 10^3.9 EID₅₀/100 µl for the oropharyngeal and cloacal swabs, respectively (Fig. 3B). The AUC for oropharyngeal and cloacal shedding was 10^6.5 and 10^4.2, respectively, which is a significant difference (P < 0.0005) (Fig. 4A). In the Prime-Boost vaccine group, the reduction in the shedding levels was significant (P < 0.0005) for the oropharyngeal swabs at all sampling times with 10^2.7 and 10^1.7 EID₅₀/100 µl on 2 and 4 dpc (Fig. 3B), respectively, representing 10²- and 10^4.2-fold reductions compared to the sham group.

The reduction in cloacal shedding was also highly significant (P < 0.0005). Cloacal shedding was completely suppressed in all Prime-Boost vaccinated ducks on 2 dpc, and only one duck had a positive cloacal swab with <10² EID₅₀/100 µl on 4 dpc. On 6 and 8 dpc all birds had negative oropharyngeal and cloacal swabs. In the Prime-Boost vaccine group, cumulative shedding reached AUC values of 10⁴ and 10² for the oropharyngeal and cloacal swabs, respectively, representing 10^2.6- and 10^2.2-fold reductions compared to the AUC recorded for the corresponding sham vaccinated group (P < 0.00005) (Fig. 4A).

To better characterize the pathobiology of challenge at 31 and 52 days of age in influenza naïve Muscovy ducks, we compared the shedding data for the two sham vaccinated groups. Older birds shed higher loads via the oropharynx but lower loads from the cloaca than the younger birds at all sampling times with a significant difference (P < 0.05) in the mean shedding loads only for the cloacal swabs on 2 dpc. Given the survival differences after challenge, we were also interested in comparing the cumulative shedding and found that older ducks had an AUC for cloacal shedding of 10^4.2, which represents a significant (P < 0.005) 7.8-fold reduction compared to the value for ducks of 31 days of age, while oropharyngeal shedding did not differ significantly (Fig. 4B).

At the time of challenge all chickens had seroconverted after one vaccination with a geometrical mean HI titer (GMT) of 7.4 and 7.2 log₂ against the H5N8 homologous vaccine antigen and the H5N1 challenge antigen, respectively (Fig. 5A). After two vaccinations, the geometric mean HI titers were 10.6 log₂ against both the H5N8 and H5N1 antigens (Fig. 5B). Both the FLUH5S and the NP-ELISA kits recorded high positive values in all birds after one or two vaccinations groups (Fig. 5A, 5B). For the Prime vaccine group the NP-ELISA and the FLUH5S had values of 7.3 S/N and 6.5 S/P ratios, respectively. In the Prime-Boost vaccine, the NP-ELISA and the FLUH5S were 5.7% for S/N and 6.9 for the S/P ratio, respectively. At the time of challenge all chickens in the Prime vaccine and in the Prime-Boost vaccine groups were negative in the N1-ELISA assay with mean S/N values of 64% and 58%, respectively. The two control groups were antibody negative at the time of challenge. At the termination of the experiment at 14 dpc the HI titers were 9.8 and 9.6 log₂ against H5N8 and H5N1, respectively, in the Prime vaccine group, which is a significant 5.3-fold increase compared to the HI titers before challenge (P < 0.005) (Fig. 5A).

To further investigate the importance of vaccine-induced antibodies, we tested postchallenge sera using the following assays: FLUH5S, NP-ELISA, and N1-ELISA. Data were compared with those recorded before the challenge. In the Prime vaccine group, antibodies against H5 detected with the FLUH5S kit resulted significantly higher (P < 0.005) compared to the antibodies detected before the challenge. When an experimental 1:5000 dilution was used, the same ELISA kit recorded an even greater mean difference of 2.0 in the S/P ratio between pre- and postchallenge titers (P < 0.0005), and all samples were positive. Before and after challenge, mean NP-ELISA S/N ratio values did not differ significantly.

In the Prime vaccine group, postchallenge sera were all negative by the N1-ELISA except for one bird with a S/N value of 20. The difference between N1-ELISA S/N percentages before (S/N = 63.9%) and after the challenge (S/N = 48.3%) were not significant. Interestingly, the N1-positive bird was the only subject that shed virus.

The HI GMTs slightly declined in chickens receiving two vaccinations from a prechallenge value of 10.6 log₂ to 10.5 log₂ and 10.3 log₂ against H5N8 and H5N1 antigens, respectively. A negligible nonstatistically significant decline in mean anti-H5 antibody titer was observed using the FLUH5S kit, at both dilutions of 1:500 (S/P = 6.8) and 1:5000 (S/P = 5.8). Similarly, postchallenge mean anti-NP antibodies (S/N = 5.8%) were found to be nonsignificantly different from those recorded before challenge. After challenge, the mean S/N values measured by the N1-ELISA assay were equal to 54%. The difference between these values and those recorded before challenge was not significant.

The prechallenge serological responses to inactivated H5N8 vaccine in SPF Muscovy ducks were assessed after a single and after two vaccinations. After a single vaccination, all but one duck had an HI titer <1 log₂. The only positive duck had HI titers against the vaccine and challenge antigens of 3.0 and 2.0 log₂, respectively (Fig. 6A). Antibodies against the H5 hemagglutinin were not detected by the FLUH5S kit. The NP-ELISA kit gave positive results for all prechallenge sera with a mean S/N value of 11.5%. Ducks that received a booster vaccination had an HI GMT of 4.8 log₂ and 4.4 log₂ against the H5N8 and H5N1 antigens, respectively. The FLUH5S kit detected 6/10 positive birds with a mean S/P ratio of 1.5. The NP-ELISA kit showed a robust seroconversion in all ducks with a mean S/N value of 4.1%. Prior to challenge, none of the vaccinated birds in both groups were positive by the N1-ELISA assay with mean S/N values of 49% (Prime vaccine) and 62% (Prime-Boost vaccine).

After challenge, among birds that received a single vaccination, a significant (P < 0.005) increase was observed in mean antibody titers for all tests. In particular, all of the birds seroconverted scoring an HI GMT of 6.2 log₂ and 5.3 log₂ against the challenge and homologous vaccine antigens, respectively. For these birds, the FLUH5S kit recorded mean S/P values of 3.3, indicating a robust seroconversion for all birds. When the same sera were tested by the anti-NP ELISA, we observed a significant (P < 0.0005) generalized postinfection increase of the S/N% in all animals reaching mean S/N values of 4.2%. Nine of the ducks were positive by the N1-ELISA assay. Anti-N1 antibodies significantly increased in all birds with a mean S/N value of 12%, which represent a significant increase in antibody titers compared to the prechallenge sera (P < 0.005) (Fig. 6A).

For birds that received a booster vaccination, challenge induced a significant increase in the HI GMT (P < 0.05) against the H5N1 and the H5N8 antigens with titers of 6.6 log₂ and 6.4 log, respectively (Fig. 6B). However, the postchallenge titers were equal to or lower in 4/10 ducks compared to the prechallenge titers. Postchallenge sera tested with the FLUH5S kit did not show a significant increase when compared to prechallenge sera. Anti-NP antibodies were minimally decreased after the challenge (P < 0.05), recording a mean S/N value of 4.4%.

For 5/10 ducks, postchallenge sera resulted to be positive by the N1-ELISA assay. Nonetheless, in 2/5 subjects for which a negative result was observed, the S/N values were close to the threshold of positivity, ranging between 28% and 31%. The remaining 3/5 negative ducks recorded S/N values ranging between 59% and 69%. The comparison of values before and after the challenge indicated a significant difference in the Prime-Boost vaccine group (P < 0.005).

The global spread of HPAI H5N1 clade 2.3.4.4b is currently causing a devastating economic impact in the poultry industry. The recent infections in dairy cattle in the United States (23) may increase the risks of widespread infections in other agriculturally important species. Thus far, in most countries the only control methods for the poultry industry are focused on biosecurity, education, surveillance, rapid diagnosis, and depopulation of affected flocks (24,25,26). Vaccination can be more cost-effective to control AIV than the traditional depopulation approach, but its application in poultry has been limited mainly because of its negative impact on the international trade of poultry products. However, the ongoing outbreaks of diverse H5 clade 2.3.4.4 viruses in poultry has questioned the traditional approach to control HPAI outbreaks.

Experimental vaccines are currently evaluated in several countries in Europe and North America. Given the high similarity of the HA gene of the reverse genetic inactivated H5N8 vaccine and the currently circulating H5Nx viruses belonging to clade 2.3.4.4b, we decided to test its protective efficacy in chickens and ducks against highly virulent H5N1 viruses. A single injection of chickens provided complete protection against challenge with H5N1 A/turkey/Italy/21VIR9520-3/2021. Moreover, shedding of the challenge virus was greatly reduced, with only one vaccinated bird having a low virus titer in oropharyngeal swabs. Complete protection against clinical signs and shedding was obtained with the Prime-Boost vaccination strategy. Similarly, a single vaccination prevented clinical disease in 9/10 ducks against challenge with A/mallard/Italy/22VIR9219-6/2022, while a booster vaccination prevented morbidity and mortality in all birds. Efficient and effective vaccination of poultry can also reduce or prevent shedding hence secondary spread (27).

Avian influenza viruses generally do not cause mortality in ducks; however, some HPAIV H5N1 clades are pathogenic and can result in mortality (28). This pathogenicity is not uniformly observed across clades, and factors such as age and dose appear to influence infection (29). In our Experiment 2, duck mortality rates were 20% and 100% in the Prime sham and Prime-Boost sham groups, respectively, similar to findings from other studies where clade 2.3.4.4 viruses can reach 75% mortality in ducks (30), but also show rates of 50% (31), 20% (32), and 0% (33) with similar doses (10^4.0 to 10^6.0 EID₅₀) and ages ranging from 4 to 6 wk old.

In ducks, the Prime vaccine group showed lower viral shedding levels than the sham group after 4 dpc. Furthermore, a booster effect was observed on viral shedding in the Prime-Boost vaccine group at 2 and 4 dpc, with significant reductions in viral levels in OS and CS, and absence of shedding after 4 dpc. This is similar to what was previously reported by Kim et al. (32). Nevertheless, it is worthwhile to investigate further whether a significant reduction in viral shedding can decrease viral transmission among vaccinated ducks. Lewis et al. (24) examined the antigenic cartography of a large number of isolates and predicted that the conventional vaccines would provide suboptimal protection. However, the list of viruses they examined did not include the 2016 H5N8 isolate from Egypt (34). Clearly, the antigenic similarity between the currently circulating clade of AIV and the Egypt 2016 H5N8-based vaccine formulation provides a strong protection.

In general, the presence of neutralizing antibodies specific for the HA protein at systemic or mucosal sites of infection provides immediate protection against influenza viruses, whereas the clearance of the viruses depends on cell-mediated immunity (35). Our HI analysis in vaccinated chickens and ducks demonstrated that the vaccine can induce similar levels of antibodies against the vaccine homologous H5N8 strain and the H5N1 challenge virus. The lack of antibodies against the N1 protein in vaccinated chickens that did not shed virus suggests that vaccination may have induced a sterilizing immunity. The use of inactivated H5N8 vaccine proved compatible with strategies that aim at the differentiation of vaccinated chickens infected with wild-type H5N1 virus based on the absence of antibodies to the N1 protein.

A more detailed analysis of the chicken sera using the experimental 1:5000 dilution instead of the 1:500 dilution showed some interesting results. With the Prime vaccine group, when sera are used at 1:500 dilution high-titer samples are dispersed around the plateau (upper end of the dynamic range of the assay, values around 6/7 S/P) so the coefficient of variation (CV) in the group is very limited. If a greater dilution of 1:5000 is applied, sera are more dispersed, and so the CV is higher, reflecting the distribution and the CV shown by the HI assay at the same sampling time. After the challenge, antibody levels were much higher, not because of the challenge but because of the age of the birds and the natural increase in Ab titers, and so even in the case of a higher dilution the kit is working in the plateau of its dynamic range and the CV is almost the same irrespective of dilution applied. It will be important to further determine the optimal dilution to better quantitate the antibody responses in vaccinated animals.

In conclusion, the reverse genetic inactivated H5N8 avian influenza virus vaccine provides strong protection, with significantly reduced shedding and a strong antibody response in chickens. Additionally, the vaccine alone conferred protection from mortality in vaccinated ducks, although it did not fully protect against infection. However, when used in a Prime-Boost strategy, it significantly reduced viral shedding.

The authors appreciate the editorial assistance provided by Dr. K.A. Schat.

AIV =

avian influenza virus;

AUC =

area under the curve;

CS =

cloacal swab;

CV =

coefficient of variation;

dpc =

days postchallenge;

EID₅₀ =

embryo infectious dose;

GMT =

geometrical mean HI titers;

HA =

hemagglutinin;

HI =

hemagglutination inhibition;

HPAI(V) =

highly pathogenic avian influenza (virus);

LoD =

limit of detection;

LPBE =

liquid-phase blocking ELISA;

M =

matrix;

mab =

monoclonal antibody;

MTD =

mean time to death;

NA =

neuraminidase;

OS =

oropharyngeal swab;

PB =

polymerase basic;

PBS =

phosphate-buffered saline;

pi =

postinoculation;

RT-qPCR =

quantitative reverse transcription polymerase chain reaction;

sc =

subcutaneously;

S/N =

sample-to-negative;

S/P =

sample-to-positive;

SPF =

specific-pathogen-free