henhaobio
Sign in

Abstract

The rapid evolution of avian influenza A/H5N1, including the recent U.S. clade 2.3.4.4b outbreak, highlights its pandemic potential and the urgent need for durable, broadly protective vaccines. Given the capacity of CD8+ T cells to mediate cross-strain immunity, we investigated whether geographically distinct HLA-A33 allotypes, HLA-A*33:01 in East/Southeast Asia and HLA-A*33:03 in South Asia, differentially shape the influenza immunopeptidome and influence antiviral immunity. Antigen-presenting cells overexpressing HLA-A*33:01 or HLA-A*33:03 were transfected with single A/H5N1 antigens or infected with A/X-31 (H3N2) as a control comparison representing current seasonal influenza virus. We identified novel ligands restricted to HLA-A*33:01 (57 from A/H5N1; 55 from A/X-31) and HLA-A*33:03 (29 from A/H5N1; 45 from A/X-31). Although fewer peptides were recovered for HLA-A*33:03, a larger proportion of A/X-31-derived peptides were predicted as high-affinity binders (74%) compared with HLA-A*33:01 (61%), indicating qualitative differences in antigen presentation. To determine immunogenicity, peripheral blood lymphocytes from HLA-A*33:03-positive, A/H5N1-naive donors were stimulated with four conserved peptides: PB2GTF, PB2KTY, NPSVQ and PB1MTK. All elicited robust CD8 T cell activation despite the absence of prior A/H5N1 exposure, demonstrating cross-recognition by memory T cells primed against seasonal influenza. These findings define HLA-A33-restricted influenza epitopes and reveal allotype-specific presentation features that shape CD8+ T cell immunity. Conserved, immunogenic peptides identified here represent promising candidates for rational design of broadly cross-reactive vaccines to protect HLA-A33-expressing populations against severe A/H5N1 disease. Data are available via ProteomeXchange with identifier PXD078870. Author SummaryAvian influenza A/H5N1 continues to pose a significant pandemic threat because of its ability to infect humans and its potential to acquire sustained human-to-human transmissibility. While current influenza vaccines primarily target rapidly evolving viral surface proteins, CD8+ T cells can recognize more conserved internal viral proteins and may provide broader protection against diverse influenza strains. In this study, we investigated how two common HLA-A33 variants, which are prevalent in South, East, and Southeast Asian populations, present influenza-derived peptides to CD8+ T cells. We identified novel influenza peptides presented by HLA-A*33:01 and HLA-A*33:03. Importantly, several conserved A/H5N1-derived peptides were recognized by memory CD8+ T cells from healthy individuals with no prior exposure to A/H5N1, suggesting that previous infection with seasonal influenza viruses can generate cross-reactive immune responses. Our findings expand the current repository of influenza T cell targets and provide new insights into antiviral immunity in HLA-A33-expressing populations. The conserved and immunogenic peptides identified in this study may help guide the development of broadly protective influenza vaccines and contribute to future pandemic preparedness efforts.

In-depth summary

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (13 Jul 2026) · CC BY 4.0

Immunopeptidomics shows HLA-A*33:01 and HLA-A*33:03 present largely distinct influenza peptide repertoires, and four HLA-A*33:03-restricted epitopes triggered CD8+ T cell responses in A/H5N1-naive donors.

Current vaccines mostly target fast-evolving surface glycoproteins, while conserved internal-protein CD8+ T cell epitopes depend on HLA type, and HLA-A33 epitopes from A/H5N1 were poorly defined. This work maps naturally presented peptides for two HLA-A33 allotypes common in Asia, most of them absent from the Immune Epitope Database (IEDB).

Key results

  1. ShownMost naturally presented viral peptides were new to the field: A/X-31 infection and single-protein A/H5N1 expression yielded large peptide sets for both HLA-A33 allotypes, and the great majority of A/H5N1 peptides were absent from IEDB.

    A/X-31: 73 (HLA-A*33:01) and 58 (HLA-A*33:03) peptides; 18/73 (24.6%) and 13/58 (22.4%) already in IEDB. A/H5N1 (HA, NP, PB2, M1 only): 61 and 32 peptides; 57/61 (93%) and 29/32 (91%) unreported in IEDB. · Differential peptide presentation by HLA-A33…; HLA-A33 immunopeptidomes reveal novel…

  2. ShownDespite highly similar binding motifs, the two allotypes present largely non-overlapping peptide sets, and HLA-A*33:03 shows more predicted binders and a preference for longer peptides.

    Shared by both allotypes: 22.9% of human peptides (8-12mer) and 31% of A/X-31 peptides (8-14mer); 0.68% of viral peptides shared with HLA-A*11:01 too. Predicted binders, A*33:03 vs A*33:01: 70% vs 48% (human), 74% vs 61% (A/X-31). · Differential peptide presentation by HLA-A33…

  3. SuggestedIn two A/H5N1-naive HLA-A*33:03 donors, in vitro expanded CD8+ T cells recognised several tested peptides, indicating cross-reactive memory probably from seasonal influenza.

    29 candidate peptides tested in six pools; PB2GTF, PB2KTY and NPSVQ (A/H5N1) gave significant IFNγ responses in both donors; PB1MTK (A/X-31) in Donor 1 only (untested in Donor 2). Single peptides from pool 6 were not evaluated; one T-cell culture per condition, technical triplicates. · Identification of immunogenic…; Figure legends

  4. ShownAbout a third of canonical A/X-31 peptides were highly conserved across influenza A strains. Among the four validated epitopes, PB2GTF was the most conserved, while PB1MTK showed lower conservation overall.

    23 (34%) HLA-A*33:01 and 20 (37%) HLA-A*33:03 canonical A/X-31 peptides >90% conserved; M1FTL and PB2DLN >90% conserved in human and avian A/H5N1; PB2GTF >98.6% in all strains; PB2KTY >96.5% in A/H5N1 and A/H1N1 but 89% of A/H3N2 carried E5D. · High epitope conservation supports broad vaccine…

  5. SuggestedNPSVQ-specific CD8+ T cells in the two donors were dominated by different (private) paired TCR clonotypes that nonetheless shared CDR3 lengths.

    Donor 1: TRAV20.TRAJ13_TRBV27.TRBJ2-3 in n=20/36 pairs; Donor 2: TRAV14/DV4.TRAJ43_TRBV10-2.TRBJ2-1 in n=7/20. Both donors: CDR3α 12mer and CDR3β 15mer. · Private dominant TCR clonotypes target the…; Discussion

Methods, data and reproducibility

HLA-A33 transfectants of the C1R cell line were infected with A/X-31 or expressed single A/H5N1 proteins; HLA-I peptides were sequenced by LC-MS/MS, scored with NetMHCpan-4.2, tested in donor T cell IFNγ assays, and checked for sequence conservation.

  • HLA-I immunopeptidomes of C1R.A*33:01 and C1R.A*33:03 cells, mock vs A/X-31 (H3N2) infected, biological duplicates · human (C1R cells); influenza A/X-31 · n = 12 datasets · PXD078870
  • Immunopeptidomes of C1R.A*33:01 and C1R.A*33:03 expressing single A/H5N1 proteins (HA, NP, PB2, M1) from A/Chicken/Laos/Xaythiani-26/2006 · human (C1R cells); influenza A/H5N1 · n = 61 and 32 A/H5N1 peptides (HLA-A*33:01, HLA-A*33:03) · PXD078870
  • PBMCs from A/H5N1-naive, HLA-A*33:03-positive healthy donors for peptide-specific T cell expansion and IFNγ assays · human · n = 2 donors
  • Paired αβTCR sequences from NPSVQ-specific CD8+IFNγ+ T cells · human · n = 36 pairs (Donor 1); 20 pairs (Donor 2)
  • Influenza A protein isolates from NCBI Virus (captured 15/02/2026) for conservation analysis · influenza A virus (human and avian isolates)
Methods & tools
PEAKS Xpro, NetMHCpan-4.2, Seq2logo2.0, MAFFT G-INS-I, TCR_Explore, IMGT/V-QUEST, GraphPad Prism 9.0.0, FlowJo version 10, IEDB, NCBI Virus, UniProt/Swiss-Prot v2024
Reagents & environment
PEAKS Xpro; NetMHCpan-4.2; Seq2logo2.0; MAFFT G-INS-I; GraphPad Prism 9.0.0; FlowJo version 10; UniProt/Swiss-Prot v2024; NCBI Virus data captured 15/02/2026.
Assay conditions & compute
Sequence alignments ran on the M3 HPC cluster (Monash eResearch); core counts and runtime not stated.
Reproducibility gaps
No code repository stated; mass spectrometry data deposited in PRIDE as PXD078870 (no URL printed); peptide and TCR tables are in Supplementary Data 1, 2 and 3.
Limitations (4)
  • Direct functional comparison of the two allotypes was not possible because T cell response data exist only for HLA-A*33:03, not HLA-A*33:01. · stated by the authors
  • Productive A/H5N1 infection of the cell lines was not achievable, so only four proteins (NP, PB2, M1, HA) were expressed individually and the A/H5N1 immunopeptidome likely does not cover the whole viral proteome. · noted in this summary
  • Single peptides from pool 6 were not tested in either donor because of limited PBMC availability, and PB1MTK was not tested in Donor 2. · stated by the authors
  • Immunogenicity rests on two HLA-A*33:03 donors with one T-cell culture per condition; viral peptides were accepted at a 5% FDR versus 1% for human peptides, so some viral assignments may be false positives. · noted in this summary

Read this if you work on HLA-restricted T cell epitope discovery, immunopeptidomics, or conserved-epitope influenza vaccine design for Asian HLA alleles.

Citations
0

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.