Jin-Ha Song

Efficient replication of influenza A viruses (IAVs) requires balanced activities of hemagglutinin (HA), neuraminidase (NA), and the RNA polymerase complex, whose functions are strongly influenced by PB2 mutations. We previously revealed three distinct evolutionary pathways for PB2 mutations, with two pathways leading to the emergence of viral strains responsible for human seasonal infections and the 2009 pandemic, and a third pathway giving rise to H5Nx highly pathogenic avian influenza viruses (HPAIVs) defined by a triad of mutations (I147T, K339T, and A588T) that occasionally spill over to humans. Here, we investigated the zoonotic risk posed by this triad and elucidated its evolutionary relationship with HA, NA, and vaccination. Recombinant PR8 and clade 2.3.2.1c H5N1 viruses carrying the triad replicated efficiently in embryonated chicken eggs and had moderate replication efficiency in mammalian cells; moreover, mice infected with these viral strains exhibited milder weight loss and lower lung titers than those infected with the E627K-carrying strain. Sequence analysis of H5Nx viruses revealed early emergence and long-term persistence of the triad across diverse genotypes, which was closely linked to HA glycosylation and NA-stalk truncation. However, the prevalence of these viral strains declined significantly after successive H5 poultry-vaccination campaigns. These data indicate that the triad provides a replication advantage compatible with both poultry and mammalian hosts but confers only moderate mammalian pathogenicity and that sustained vaccination can restrain the spread of viral strains with these mutations. Continuous molecular surveillance of PB2 alongside HA and NA remains essential for preventing H5Nx zoonotic threats.Please check and confirm that the authors and their respective affiliations have been correctly identified and amend if necessary.I have checked and modifed the email addresses of the corresponding authors. I confirm the information of other authors are correct.

Avian influenza A viruses (IAVs) pose a persistent threat to the poultry industry, causing substantial economic losses. Although traditional vaccines have helped reduce the disease burden, they typically rely on multivalent antigens, emphasize humoral immunity, and require intensive production. This study aimed to establish a genetically matched host–cell system to evaluate antigen-specific immune responses and identify conserved CD8+ T cell epitopes in avian influenza viruses. To this end, we developed an MHC class I genotype (B21)-matched host (Lohmann VALO SPF chicken) and cell vector (DF-1 cell line) model. DF-1 cells were engineered to express the hemagglutinin (HA) gene of clade 2.3.4.4b H5N1 either transiently or stably, and to stably express the matrix 1 (M1) and nucleoprotein (NP) genes of A/chicken/South Korea/SL20/2020 (H9N2, Y280-lineage). Following prime-boost immunization with HA-expressing DF-1 cells, only live cells induced strong hemagglutination inhibition (HI) and virus-neutralizing (VN) antibody titers in haplotype-matched chickens. Importantly, immunization with DF-1 cells transiently expressing NP induced stronger IFN-γ production than those expressing M1, demonstrating the platform’s potential for differentiating antigen-specific cellular responses. CD8+ T cell epitope mapping by mass spectrometry identified one distinct MHC class I-bound peptide from each of the HA-, M1-, and NP-expressing DF-1 cell lines. Notably, the identified HA epitope was conserved in 97.6% of H5-subtype IAVs, and the NP epitope in 98.5% of pan-subtype IAVs. These findings highlight the platform’s utility for antigen dissection and rational vaccine design. While limited by MHC compatibility, this approach enables identification of naturally presented epitopes and provides insight into conserved, functionally constrained viral targets.

During the 2009 H1N1 pandemic (pdm09), the poor replication of PR8-derived vaccine strains in embryonated chicken eggs (ECEs) delayed vaccine production, necessitating costly adjuvants. To improve egg-based yield, we generated PB2-substituted H1N1 strains via reverse genetics, replacing PR8 PB2 with a PB2 lacking mammalian-adaptive mutations (dtxPB2), cognate pdm09 PB2 (19PB2), or avian PB2. All PB2-substituted strains achieved over tenfold higher titers than the conventional PR8 PB2-containing strain (rGD19), with rGD19/dtxPB2 and rGD19/19PB2 exhibiting significantly higher titers and reduced murine virulence. Among these, rGD19/19PB2 produced the highest hemagglutinin (HA) yield and, when administered intranasally as a binary ethyleneimine (BEI)-inactivated whole-virion vaccine, elicited a significantly stronger broncho-alveolar IgA response than rGD19. Both rGD19 and rGD19/19PB2 provided comparable protection against a homologous H1N1 challenge, yet only rGD19/19PB2 conferred full survival protection after a lethal heterologous H3N2 challenge. These findings show that incorporation of cognate PB2 enhances H1N1 replication in ECEs and antigen yield, reduces murine virulence, and confers robust homo- and heterosubtypic protection via intranasal immunization, underscoring the promise of PB2-modified H1N1 strains as inactivated mucosal whole-virion vaccines for future vaccine development.

Background/Objective: Highly pathogenic (HP) H5Nx and low-pathogenicity (LP) H9N2 avian influenza viruses (AIVs) pose global threats to the poultry industry and public health, highlighting the critical need for a dual-protective vaccine. Methods: In this study, we generated a model PR8-derived recombinant H5N2 vaccine strain with hemagglutinin (HA) and neuraminidase (NA) genes from clade 2.3.2.1c H5N1 and Y439-like H9N2 viruses, respectively. To enhance the immunogenicity of the recombinant H5N2 vaccine strain, N-glycans of the HA2 subunit, NA, and M2e were modified. Additionally, we replaced M2e with avian M2e to enhance the antigenic homogeneity of AIVs for better protection. We also replaced PR8 PB2 with 01310 PB2, which is the PB2 gene derived from an LP H9N2 avian influenza virus, to eliminate pathogenicity in mammals. The productivity of the model vaccine strain (rvH5N2-aM2e-vPB2) in embryonated chicken eggs (ECEs), its potential risk of mammalian infection, and the immunogenicity associated with different inactivation methods (formaldehyde (F/A) vs. binary ethyleneimine (BEI)) were evaluated. Results: The rvH5N2-aM2e-vPB2 strain demonstrated high productivity in ECEs and exhibited complete inhibition of replication in mammalian cells. Furthermore, compared with using F/A inactivation, inactivation using BEI significantly enhanced the immune response, particularly against NA. This enhancement resulted in increased virus neutralization titers, supporting its efficacy for dual protection against H5Nx and H9N2 avian influenza viruses. Furthermore, we demonstrated that M2e-specific immune responses, difficult to induce with inactivated vaccines, can be effectively elicited with live vaccines, suggesting a strategy to enhance M2e immunogenicity in whole influenza virus vaccines. Conclusions: Finally, the successful development of the model rH5N2 vaccine strain is described; this strain provides dual protection, has potential applicability in regions where avian influenza is endemic, and can be used to promote the development of versatile H5N2 recombinant vaccines for effective avian influenza control.

Highly pathogenic (HP) H5Nx and low pathogenicity (LP) H9N2 avian influenza viruses (AIVs) pose threats to the poultry industry and public health. HP clade 2.3.2.1c AIVs and clade 2.3.4.4b H5N1 AIVs, along with various lineages of H9N2 AIVs, are problematic worldwide, and vaccines that provide dual protection against both H5N1 AIVs and H9N2 AIVs need to be developed. In this study, we generated a model PR8-derived recombinant H5N2 vaccine strain with hemagglutinin (HA) and neuraminidase (NA) genes from clade 2.3.2.1c H5N1 and Y439-like H9N2 viruses, respectively. To increase the immunogenicity of the common epitopes located in the HA2 subunit, the 154N-glycan, which eventually masks the common epitope, was removed. Additionally, we selected an NA that has no N-glycans in the stalk and removed N-glycan in the M2e of the PR8 strain to increase the accessibility of epitopes that are masked by the N-glycans. We also replaced PR8 PB2 with 01310 PB2 to eliminate pathogenicity in mammals and replaced the M2e with avian M2e to increase antigenic homogeneity to AIVs for better protection. The efficacy of the model vaccine strain (rvH5N2-aM2e-vPB2) was compared with that of rH5N2-vPB2, which contains unmodified HA2 and M2e region, and a recombinant clade 2.3.2.1c H5N1 strain in which only the PB2 segment was replace with 01310 PB2. The rvH5N2-aM2e-vPB2 strain was highly productive in embryonated chicken eggs (ECEs), and inactivation with binary ethyleneimine (BEI) significantly enhanced the immune response against NA compared to formaldehyde (F/A) inactivation, resulting in an increased virus neutralization titer for dual protection. We also demonstrated that while the immune response against M2e is challenging to induce with an inactivated vaccine, it can be effectively elicited via a live vaccine. Thus, the successful generation of a model vaccine strain for dual protection described herein may promote the development of various combinations of recombinant H5N2 vaccine strains for use in countries where H5Nx and H9N2 AIVs are endemic.

H9N2 avian influenza A viruses (AIVs) cause economic losses in the poultry industry and provide internal genomic segments for the evolution of H5N1 and H7N9 AIVs into more detrimental strains for poultry and humans. In addition to the endemic Y439/Korea-lineage H9N2 viruses, the Y280-lineage spread to Korea since 2020. Conventional recombinant H9N2 vaccine strains, which bear mammalian pathogenic internal genomes of the PR8 strain, are pathogenic in BALB/c mice. To reduce the mammalian pathogenicity of the vaccine strains, the PR8 PB2 was replaced with the non-pathogenic and highly productive PB2 of the H9N2 vaccine strain 01310CE20. However, the 01310CE20 PB2 did not coordinate well with the hemagglutinin (HA) and neuraminidase (NA) of the Korean Y280-lineage strain, resulting in a 10-fold lower virus titer compared to the PR8 PB2. To increase the virus titer, the 01310CE20 PB2 was mutated (I66M-I109V-I133V) to enhance the polymerase trimer integrity with PB1 and PA, which restored the decreased virus titer without causing mouse pathogenicity. The reverse mutation (L226Q) of HA, which was believed to decrease mammalian pathogenicity by reducing mammalian receptor affinity, was verified to increase mouse pathogenicity and change antigenicity. The monovalent Y280-lineage oil emulsion vaccine produced high antibody titers for homologous antigens but undetectable titers for heterologous (Y439/Korea-lineage) antigens. However, this defect was corrected by the bivalent vaccine. Therefore, the balance of polymerase and HA/NA activities can be achieved by fine-tuning PB2 activity, and a bivalent vaccine may be more effective in controlling concurrent H9N2 viruses with different antigenicities.

For the development of an optimized Egyptian H9N2 vaccine candidate virus for poultry, various recombinant Egyptian H9N2 viruses generated by a PR8-based reverse genetics system were compared in terms of their productivity and biosafety since Egyptian H9N2 avian influenza viruses already possess mammalian pathogenicity-related mutations in the hemagglutinin (HA), neuraminidase (NA), and PB2 genes. The Egyptian HA and NA genes were more compatible with PR8 than with H9N2 AIV (01310) internal genes, and the 01310-derived recombinant H9N2 strains acquired the L226Q reverse mutation in HA after passages in eggs. Additionally, the introduction of a strong promoter at the 3′-ends of PB2 and PB1 genes induced an additional mutation of P221S. When recombinant Egyptian H9N2 viruses with intact or reverse mutated HA (L226Q and P221S) and NA (prototypic 2SBS) were compared, the virus with HA and NA mutations had high productivity in ECES but was lower in antigenicity when used as an inactivated vaccine due to its high binding affinity into non-specific inhibitors in eggs. Finally, we substituted the PB2 gene of PR8 with 01310 to remove the replication ability in mammalian hosts and successfully generated the best recombinant vaccine candidate in terms of immunogenicity, antigenicity, and biosafety.

Clade 2.3.4.4c H5N6 avian influenza A viruses (AIVs) may have originally adapted to infect chickens and have caused highly pathogenic avian influenza (HPAI) in poultry and human fatalities. Although A/Puerto Rico/8/1934 (H1N1) (PR8)-derived recombinant clade 2.3.4.4c H5N6 vaccine strains have been effective in embryonated chicken eggs-based vaccine production system, they need to be improved in terms of immunogenicity and potential mammalian pathogenicity. We replaced the PB2 gene alone or the PB2 (polymerase basic protein 2), NP (nucleoprotein), M (matrix protein) and NS (non-structural protein) genes together in the PR8 strain with corresponding genes from AIVs with low pathogenicity to remove mammalian pathogenicity and to match CD8+ T cell epitopes with contemporary HPAI viruses, respectively, without loss of viral fitness. Additionally, we tested the effect of the H103Y mutation of hemagglutinin (HA) on antigen productivity, mammalian pathogenicity and heat/acid stability. The replacement of PB2 genes and the H103Y mutation reduced the mammalian pathogenicity but increased the antigen productivity of the recombinant vaccine strains. The H103Y mutation increased heat stability but unexpectedly decreased acid stability, probably resulting in increased activation pH for HA. Interestingly, vaccination with inactivated recombinant virus with replaced NP, M and NS genes halted challenge virus shedding earlier than the recombinant vaccine without internal genes replacement. In conclusion, we successfully generated recombinant clade 2.3.4.4c H5N6 vaccine strains that were less pathogenic to mammals and more productive and heat stable than conventional PR8-derived recombinant strains by optimization of internal genes and the H103Y mutation of HA.