Glycan repositioning of influenza hemagglutinin stem facilitates the elicitation of protective cross-group antibody responses

Abstract

The conserved hemagglutinin (HA) stem has been a focus of universal influenza vaccine efforts. Influenza A group 1 HA stem-nanoparticles have been demonstrated to confer heterosubtypic protection in animals; however, the protection does not extend to group 2 viruses, due in part to differences in glycosylation between group 1 and 2 stems. Here, we show that introducing the group 2 glycan at Asn38_HA1 to a group 1 stem-nanoparticle (gN38 variant) based on A/New Caledonia/20/99 (H1N1) broadens antibody responses to cross-react with group 2 HAs. Immunoglobulins elicited by the gN38 variant provide complete protection against group 2 H7N9 virus infection, while the variant loses protection against a group 1 H5N1 virus. The N38_HA1 glycan thus is pivotal in directing antibody responses by controlling access to group-determining stem epitopes. Precise targeting of stem-directed antibody responses to the site of vulnerability by glycan repositioning may be a step towards achieving cross-group influenza protection.

Fig. 1. Design and characterization of group 1 H1 stem nanoparticle variants.

a Comparison of N-linked glycosylation pattern of group 1 and group 2 HA stem. Surface renderings of HA trimer for a representative group 1 HA (A/Solomon Islands/06 (H1N1), PDB: 3SM5, left) and representative group 2 HA (A/Finland/486/04 (H3N2), PDB: 2YP5, right). One monomer is colored in blue for visibility. The terminal N-acetylglucosamine (GlcNAc) moiety for each glycan is modeled in green and residue positions are labeled according to the H3 numbering. Yellow areas designate the approximate location of the CR6261 epitope on each HA. b Amino acid sequence alignment of N-terminal portion of HA1 of H1 (A/New Caledonia/20/99), H3 (A/Finland/486/04) and designed variants gN38 and R38. Dots and dashes indicate residues identical to H1 and gaps, respectively. c Cryo-electron microscopy structures of H1 stem nanoparticle (H1ssF WT) and its variants H1ssF gN38 and H1ssF R38. Nanoparticles are depicted along the 2-fold symmetry axis. Side views of HA stem trimeric spikes are shown below each nanoparticle. The maps were low-pass filtered to a resolution of 7 Å for comparison. Scale bar indicates 5 nm. d Antigenicity of H1 stem nanoparticle variants. ELISA binding curves are shown for mAbs specific to group 1 stem (CR6261, 02-1H01 and C179), group 2 stem (CR8020), both group 1 and 2 stems (FI6v3, MEDI8852), or irrelevant RSV F protein (D25). Source data are provided as Source Data file.

Fig. 2. Immunogenicity of H1 stem nanoparticle variants in mice.

BALB/c mice (N = 10) were immunized with 2 µg of H1ssF WT, H1ssF gN38 or H1ssF R38 on weeks 0, 4, and 8. Serum antibody titers to HAs were measured at 2 weeks after the third immunization (week 10). a ELISA endpoint titers to H1N1 NC99 HA (left) and its gN38 variant (right). b Serum neutralizing antibody titers to NC99, SG86, and CA09 (H1N1) pseudoviruses. Shown are half-maximal inhibitory dilution of serum (IC₅₀). c ELISA endpoint titers to H5N1 VN04 (left) and H7N9 AN13 HAs (right). d Serum neutralizing antibody titers to H5N1 VN04 and H7N9 AN13 pseudoviruses. Dotted lines indicate the lower detection limit of the assay. Data are presented as scattered dot plots with horizontal lines indicating geometric mean for each group. Statistical analysis was carried out by using nonparametric Kruskal–Wallis test with Dunn’s multiple comparisons. Displayed results are representative of two independent experiments (H1ssF WT and H1ssF gN38) or based on one experiment (H1ssF R38). Source data are provided as Source Data file.

Fig. 3. Specificity of serum neutralizing activity elicited by H1 stem nanoparticle variants.

a Serum pseudovirus-neutralization activity in the presence of competitor proteins. Immune sera were pre-incubated with either NC99 WT HA, HA gN38, or irrelevant RSV F proteins prior to evaluation of SG86 (H1N1) pseudovirus neutralization. Serum neutralization IC₅₀ titers were calculated in the absence or presence of competitor proteins. b Percent inhibition of virus neutralization was calculated for each competitor protein. Dotted lines indicate the lower detection limit of the assay. Data are presented as scattered dot plots with horizontal lines indicating geometric mean for each group. Statistical analysis was carried out by using nonparametric Kruskal–Wallis test with Dunn’s multiple comparisons. Displayed results are representative of two independent experiments. Source data are provided as Source Data file.

Fig. 4. Homotypic, heterosubtypic, and cross-group protection of mice immunized with H1 stem nanoparticle variants.

BALB/c mice (N = 10) were immunized with 2 µg of H1ssF WT, H1ssF gN38 or H1ssF R38 on weeks 0, 4, and 8, and challenged between weeks 16 and 18 with 20 × LD₅₀ of A/California/07/09 (H1N1) virus (a), 25 × LD₅₀ of A/Vietnam/1203/04 (H5N1) virus (b), or 10 × LD₅₀ of A/Anhui/1/13 (H7N9) virus (c). All viruses were given intranasal inoculation. Mice were monitored twice daily for their weight and activity for 14 days post infection. Survival curve (left) and associated weight loss curve (right) were plotted for each virus challenge (a–c). For weight loss curves, lines indicate group mean ± s.d. (shaded). Statistical test to compare multiple Kaplan–Meier curves was carried out by using Mantel-Cox log-rank test with Bonferroni correction. Displayed results are representative of two independent experiments (H1N1 and H5N1) or cumulative of two experiments (H7N9). Source data are provided as Source Data file.

Fig. 5. Cross-group neutralization and protective capacity of immunoglobulins isolated from mice immunized with H1 stem nanoparticle gN38 variant.

Immune Ig were obtained from H1ssF gN38-immunized mice by affinity-purifying with H7 (AN13) HA. a ELISA binding profile of purified Ig to HA from NC99, H5N1 IN05, H3N2 HK68, H7N9 AN13, and H10N8 JD13. b Neutralization profile of purified Ig to H5N1 VN04 (left), H3N2 HK68 (center), and H7N9 AN13 (right) pseudoviruses. Data are shown as mean ± s.d. c Passive immunization of purified Ig in naïve BALB/c mice (N = 10). Mice were given 0.2 mg of purified Ig from either naïve or H1ssF gN38-immunized mice, or bNAb FI6v3 (0.1 mg) intraperitoneally 24 h prior to H7N9 virus challenge (10 × LD₅₀). Mice were monitored twice daily for their weight and activity for 14 days post infection. Statistical test to compare multiple Kaplan–Meier curves was carried out by using Mantel-Cox log-rank test with Bonferroni correction. Displayed results are from one experiment. Source data are provided as Source Data file.