The human antibody CR9114, encoded by the IGHV1-69 germline gene, remains the only known hemagglutinin (HA) stem broadly neutralizing antibody (bnAb) that cross-reacts with both influenza A and influenza B viruses, more than a decade after its discovery. Dailey et al. (bioRxiv 2025) used four deep mutational scanning experiments to compare the binding-affinity landscapes of germline and somatic CR9114 against H1 HA (group 1), H3 HA (group 2), and influenza B HA, and found that breadth evolution is highly vulnerable: many mutations neutral or beneficial for H1 binding abolish H3 and influenza B binding, mutational tolerance correlates with binding affinity, and epistasis is prevalent. In particular, the paratope mutation VH I73W, modeled to create a steric clash with VH F74, destroys H3 and influenza B HA binding, and removing the clash with the second mutation F74S does not rescue it. The resulting questions remain open in the literature: what biophysical mechanism couples mutational tolerance to affinity; whether the W73-class paratope constraint can be overcome by further mutations, so that CR9114-like breadth evolution is not derailed; and which vaccination strategy, if any, can reliably elicit cross-group HA stem bnAbs — the lack of a reliable method for doing so being a major obstacle toward a pan-influenza A vaccine. Clinical HA-stem candidates still elicit only group-specific responses.

Background

Influenza virus entry is mediated by hemagglutinin (HA), a trimeric surface glycoprotein with a variable head domain and a far more conserved stem (stalk) domain. Antibodies to the HA stem can neutralize across HA subtypes, making them prime targets for broadly protective ('universal') influenza vaccines. Influenza A HAs divide into two phylogenetic groups: group 1 (H1, H2, H5, and others) and group 2 (H3, H7, and others); 'cross-group' breadth means binding HA of both groups, and full CR9114-like breadth additionally covers influenza B HA (often abbreviated BHA). CR9114, isolated by Dreyfus et al. (Science, 2012), is a human HA stem bnAb that neutralizes influenza A viruses of both groups and influenza B viruses; despite extensive searching it remains the only HA stem bnAb known to cross-react with both influenza A and B. It is encoded by the common IGHV1-69 heavy-chain germline gene and lacks the extreme somatic mutation of HIV-class bnAbs, so its rarity was long considered puzzling. During an immune response, B cells diversify their antibody genes by somatic hypermutation and compete for antigen binding in germinal centers; breadth therefore emerges through an evolutionary process on a fitness landscape defined by antibody–antigen binding affinity. Deep mutational scanning (DMS) measures such landscapes at scale: a library of antibody variants is displayed (here on yeast, quantified by the Tite-Seq method, which links yeast-display fluorescence to a binding constant), and the binding effect of each single mutation is inferred from library enrichment. Epistasis means a mutation's effect depends on other mutations present (e.g., the same change behaving differently in the germline and somatic backgrounds); pleiotropy means one mutation affecting binding to several antigens differently. Phillips et al. (eLife, 2021) had previously shown, through a similar landscape approach, that CR9114's breadth must be acquired sequentially — affinity for H1, then H3, then influenza B HA — under nested evolutionary constraints. Dailey et al. (bioRxiv, 2025) extended this with four DMS experiments comparing germline CR9114 (the inferred unmutated precursor antibody) with somatic CR9114 (the mature, somatically mutated antibody) against an H1 stem, an H3 HA, and an influenza B HA, validated by biolayer interferometry, a label-free optical method that measures binding kinetics from interference shifts as soluble antigen binds antibody immobilized on sensor tips. Their central findings are: (i) many mutations that minimally affect or improve H1 binding are detrimental for H3 and BHA binding, so a large fraction of somatic hypermutations that a primed H1-reactive lineage can acquire destroy the possibility of later breadth; (ii) mutational tolerance of an antibody–antigen pair decreases in correlation with its binding affinity, a phenomenon whose biophysical cause is unexplained; (iii) epistasis is prevalent — e.g., the paratope (antigen-contacting surface) mutation VH I73W, which replaces isoleucine 73 of the heavy-chain variable domain (VH) with tryptophan, improves germline binding to H1 (where residue 74 is a small Ser) but abolishes somatic binding to H3 and BHA, with structural modeling attributing this to a steric clash with VH F74; the double mutant I73W/F74S, which removes the clash, still fails to bind H3 or BHA, indicating a constraint not easily overcome by a secondary mutation. The source further postulates separate immunogens (antigen preparations used for immunization) devoted to pan-influenza A and pan-influenza B bnAbs. In parallel, clinical HA-stem vaccine candidates (e.g., the group-1 stem nanoparticle tested by Widge et al., Science Translational Medicine, 2023) robustly elicit group-specific bnAbs but have shown only limited ability to elicit cross-group bnAbs, and recent germinal-center work (DeWitt et al., Cell, 2026) shows antibody evolutionary outcomes are in principle predictable from binding-affinity landscapes and somatic hypermutation biases — a tool that makes the elicitation problem more tractable in principle but has not solved it for cross-group HA stem bnAbs. A comprehensive August 2026 review of universal-influenza-vaccine candidates (Mudaly et al., Antibodies, 2026) confirms that no current candidate reliably achieves the broad, durable protection these findings aim to enable.

Problem Statement

Can the constraints that make CR9114-like breadth nearly unachievable be understood and overcome? The unifying objective is to render cross-group HA stem antibody breadth — binding HA of both influenza A group 1 and influenza A group 2 — and, in the fuller CR9114-like sense, breadth that additionally covers influenza B HA, attainable by rational design and by vaccination, where it is currently blocked by the epistatic and pleiotropic constraints mapped by Dailey et al. (2025). Concretely, the open question from the source comprises four interlocking parts, each of which states the breadth scope it concerns: (1) Biophysical mechanism — what causes the observed correlation between mutational tolerance and binding affinity (tolerance shrinking as affinity weakens from H1 stem to H3 HA to influenza B HA), and does this principle generalize beyond IGHV1-69 HA stem antibodies to other antibody classes? (2) Mutational routes — the double mutation VH I73W/F74S failed to restore somatic CR9114 binding to H3 HA and influenza B HA, and the source tested only this single secondary mutation; can the W73-class paratope constraint be overcome at all, that is, do further or alternative mutations in a lineage that has acquired VH W73 restore binding to influenza A group 2 HA and influenza B HA without losing H1 binding (CR9114-like, influenza A+B scope)? More broadly, which combinations of mutations allow an IGHV1-69 HA stem antibody to reach cross-group and influenza B breadth despite the wider set of pleiotropic constraints that silently destroy H3 and BHA binding? (3) Vaccination strategy — sequential (e.g., H1-priming then heterologous) schemes are unlikely to reliably elicit CR9114-like bnAbs (influenza A+B scope) because many mutations are deleterious for H3/BHA binding yet invisible to H1-driven selection; can any vaccination strategy — prime–boost or immunogen design — reliably elicit cross-group HA stem bnAbs (influenza A group 1 + group 2 scope), whose absence the source identifies as a major obstacle toward a pan-influenza A vaccine, and, in the more demanding influenza A+B scope, CR9114-like bnAbs? (4) Landscape-guided design: continued characterization of binding-affinity landscapes of diverse HA stem bnAbs is needed to ground rational vaccine design, since no reliable method currently exists for eliciting cross-group (influenza A group 1 + group 2) HA stem bnAbs, the obstacle the source identifies for a pan-influenza A vaccine. In these parts, 'reliably' is used in the source's sense — elicitation that is not merely occasional or limited, in contrast to the group-specific bnAbs that current HA-stem vaccine candidates robustly elicit.

References

  1. Dailey KE, Wang Y, Teo QW, Wang C, Lei R, Tong M, Rodriguez LA, Wang L, Wu NC (2025). Vulnerability in the breadth evolution of an influenza broadly neutralizing antibody. bioRxiv preprint. doi:10.1101/2025.11.03.686356. https://doi.org/10.1101/2025.11.03.686356
  2. Dreyfus C, Laursen NS, Kwaks T, Zuijdgeest D, Khayat R, Ekiert DC, Lee JH, Metlagel Z, Bujny MV, Jongeneelen M, van der Vlugt R, Lamrani M, Korse HJ, Geelen E, Sahin Ö, Sieuwerts M, Brakenhoff JP, Vogels R, Li OT, Poon LL, Peiris M, Koudstaal W, Ward AB, Wilson IA, Goudsmit J, Friesen RH (2012). Highly conserved protective epitopes on influenza B viruses. Science 337(6100):1343–1348. doi:10.1126/science.1222908. https://doi.org/10.1126/science.1222908
  3. Phillips AM, Lawrence KR, Moulana A, Dupic T, Chang J, Johnson MS, Cvijovic I, Mora T, Walczak AM, Desai MM (2021). Binding affinity landscapes constrain the evolution of broadly neutralizing anti-influenza antibodies. eLife 10:e71393. doi:10.7554/eLife.71393. https://doi.org/10.7554/eLife.71393
  4. DeWitt WS, Vora AA, Araki T, Galloway JG, Alkutkar T, Bortolatto J, Castro TBR, Dumm W, Jennings-Shaffer C, Jia T, Mesin L, Ozorowski G, Pae J, Ralph DK, Bloom JD, Nourmohammad A, Song YS, Ward AB, Starr TN, Matsen FA 4th, Victora GD (2026). Replaying germinal center evolution on a quantified affinity landscape. Cell 189(16):4980–4996.e8. doi:10.1016/j.cell.2026.05.013. https://doi.org/10.1016/j.cell.2026.05.013
  5. Widge AT, Hofstetter AR, Houser KV, Awan SF, Chen GL, Burgos Florez MC, Berkowitz NM, Mendoza F, Hendel CS, Holman LA, Gordon IJ, Apte P, Liang CJ, Gaudinski MR, Coates EE, Strom L, Wycuff D, Vazquez S, Stein JA, Gall JG, Adams WC, Carlton K, Gillespie RA, Creanga A, Crank MC, Andrews SF, Castro M, Serebryannyy LA, Narpala SR, Hatcher C, Lin BC, O'Connell S, Freyn AW, Rosado VC, Nachbagauer R, Palese P, Kanekiyo M, McDermott AB, Koup RA, Dropulic LK, Graham BS, Mascola JR, Ledgerwood JE, VRC 321 study team (2023). An influenza hemagglutinin stem nanoparticle vaccine induces cross-group 1 neutralizing antibodies in healthy adults. Science Translational Medicine 15(692):eade4790. doi:10.1126/scitranslmed.ade4790. https://doi.org/10.1126/scitranslmed.ade4790
  6. Beukenhorst AL, Rogiers R, Rice KL, Booth G, Haasnoot J, Nkolola J, Ayala J, Wang L, Julg B, Pastini AK, Vietsch H, Magyarics Z, Chatzifotiou E, Temperton N, da Costa KAS, Koldijk MH, Crawford J, Rivas SA, Barouch DH, Koch CM, Goudsmit J (2026). Phase 1 and preclinical studies reveal safety, pharmacokinetics, and efficacy of intranasal delivery of the influenza antibody CR9114. Science Translational Medicine 18(835):eadz1580. doi:10.1126/scitranslmed.adz1580. https://doi.org/10.1126/scitranslmed.adz1580
  7. Cai X, Fan M, Zhai Y, Luo W, Liu S, Lv K, Zhao M, Liu L, Zhou S, Xu M, He B, Sun Y, Zhao R, Zhu X, Kuang Y, Lu H, Yue X, Li K, Zhang Y, Yuan L, Zhou N, Ye F, Liang H, Chen Y, Zheng M, Hong Z, Wang D, Zhang Z, Shu YL, Ge J, Chen YQ (2026). Human monoclonal antibodies isolated after seasonal vaccination broadly neutralize antigenically drifted influenza B viruses. Cell Host & Microbe 34(5):860–872.e5. doi:10.1016/j.chom.2026.02.018. https://doi.org/10.1016/j.chom.2026.02.018
  8. Mudaly P, Lee M, Bhatta A, Thompson C (2026). Advances in the Development of a Universal Influenza Vaccine. Antibodies 15(4):75. doi:10.3390/antib15040075. https://doi.org/10.3390/antib15040075