Elimination of receptor binding by influenza hemagglutinin improves vaccine-induced immunity

Abstract

The binding of influenza hemagglutinin (HA) to sialic acid (SA) receptors plays a well-defined role in shaping infection but the impact of such binding on vaccine responses has not yet been explored. We generated a virus-like particle (VLP) vaccine bearing the HA of H1N1 A/California/07/09 that is unable to bind to its α(2,6)-linked SA receptor (H1_Y98F-VLP) and compared its immunogenicity and efficacy to a wild-type H1-VLP (H1_WT-VLP) in mice. The H1_Y98F-VLP elicited significantly stronger and more durable antibody responses (hemagglutination inhibition and microneutralization titers) and greater avidity maturation, likely attributable to improved germinal center formation. H1_Y98F-VLP also resulted in a robust population of IL-2⁺TNFα⁺IFNγ^- CD4⁺ T cells that correlated with antibody responses. Compared to H1_WT-VLP vaccination, mice immunized with H1_Y98F-VLP had 2.3-log lower lung viral loads and significantly lower pulmonary inflammatory cytokine levels 5 days post-challenge. These findings suggest that abrogation of HA-SA interactions may be a promising strategy to improve the quality and durability of influenza vaccine-induced humoral responses.

Fig. 1. Y98F mutation abrogates SA binding without affecting HA expression or VLP structure.

Wild-type (WT) H1 (H1_WT) and Y98F H1 (H1_Y98F) were expressed in Nicotiana benthamiana. a Representation of the H1_WT (top) and H1_Y98F (bottom) expression cassettes. 2X35S/CPMV160 promoter, double 35S promoter fused to the 5′ UTR of a cowpea mosaic virus (CPMV) expression enhancer; SpPDI signal peptide from alfalfa protein disulfide isomerase, NOS nopaline synthase terminator signal. b Expression of HA was confirmed by SDS-PAGE of crude leaf extracts followed by immunoblot analysis. Commercially available recombinant H1 expressed in HEK-293 cells (recH1, Immune Technologies) was included as a positive control (1 µg). 1° ab: rabbit polyclonal anti-H1 1:500 (Cat. No. IT-003-SW, Immune Technology); 2° ab: horseradish peroxidase-conjugated goat anti-rabbit IgG 1:20000 (Cat. No. IT-200-01, Immune Technology). c VLP composition and purity were evaluated by SDS-PAGE of purified leaf digests followed by Coomassie G-250 staining. d Representative TEM images show the similar size and morphology of H1_WT- and H1_Y98F-VLP. Images were acquired using a Tecnai G2 Spirit transmission electron microscope. Scale bar = 100 nm. e Sialic acid (SA) binding was evaluated based on hemagglutination of turkey red blood cells following incubation (30 min) with serial two-fold dilutions of H1_WT- and H1_Y98F-VLP (starting at 1:150 and 1:10, respectively). SA binding was further quantified by SPR. f SPR sensorgram showing the binding response of H1_WT- (red) and H1_Y98F-VLP (blue) to α-2,6 SA captured on a streptavidin-coated chip surface. g Relative binding of H1_Y98F-VLP when adjusted for HA content.

Fig. 2. H1_Y98F-VLP elicits a more robust and durable antibody response.

Mice were vaccinated (IM) with H1_WT- or H1_Y98F-VLP (3 µg/dose). Sera were collected at day 21 (left panel) or on a monthly basis (right panel) to measure (a) total H1-specific IgG by ELISA, (b) hemagglutination inhibition titers and (c) microneutralization titers. d Avidity indices of sera obtained 2–7 mpv following incubation with 8 M urea. e H1-specific IgG-producing plasma cells (PC) in the bone marrow measured by ELISpot. Representative wells from each group are shown on the right. f Spearman’s rank correlation technique was applied to evaluate the relationship between the frequency of PC and IgG titers (left), HI titers (middle) and MN titers (right). Error bars represent the mean ± SEM. At day 21, N = 40–70/group and data are pooled from six independent experiments. For long-term studies, N = 7–8/group. Statistical significance between vaccine groups was determined by Mann–Whitney test (*p < 0.033, **p < 0.01, ***p < 0.001). Statistical significance between time points within the same group was determined by two-way ANOVA with the Geisser-Greenhouse correction and Sidak’s multiple comparisons (^♦p < 0.033).

Fig. 3. H1_Y98F-VLP promotes enhanced germinal center selection.

Mice were immunized with 0.5 µg H1_WT- or H1_Y98F-VLP in the right hind limb footpad and popliteal lymph nodes (pLN) were harvested at indicated time points. The mean frequency (±SEM) of (a) CD19⁺Fas⁺GL7⁺ GC B cells and (b) CD3⁺CD4⁺CXCR5⁺PD-1⁺ T_FH cells was determined by flow cytometry. Representative plots showing GC B cell and T_FH gating are shown on the right. To identify H1-specific GC B cells, freshly isolated cells were incubated with 1 µg/ml H1_Y98F-VLP (30 min, 4 °C) and cognate GC B cells were detected following staining with anti-H1 FITC. c Representative plots and d mean frequency (±SEM) of HA⁺ cells among GC B cells. Data are pooled from three independent experiments, n = 7–13/group at each time point. Statistical significance between groups at each time point was determined by Mann–Whitney test (*p < 0.033).

Fig. 4. H1_Y98F-VLP elicits robust CD4⁺ T cell responses with enhanced recruitment of antigen-specific CD4⁺ T cells to the bone marrow.

Splenocytes and BM immune cells were stimulated for 18 h with 2.5 µg/ml H1_WT-VLP. Flow cytometry was used to quantify H1-specific CD4⁺ T cells in a splenocytes isolated at 28 days post-vaccination (3 µg) and in b splenocytes and c BM immune cells at 28 days post-boost (0.5 µg/dose). Background values obtained from non-stimulated samples were subtracted from values obtained following stimulation with H1_WT-VLP. a–c The left panel shows the mean frequency (±SEM) of CD4⁺ T cells expressing CD44 and at least one of IL-2, TNFα or IFNγ. The right panel shows the individual cytokine signatures of responding CD4⁺ T cells obtained by Boolean analysis. Color-matched pie charts depict relative distributions of cytokine-producing CD4⁺ T cell populations and IFNγ⁺ populations are highlighted by a green arc. Statistical significance was determined by Kruskal–Wallis test with Dunn’s multiple comparisons (total response) or two-way ANOVA with Tukey’s multiple comparisons (cytokine signatures) (*p < 0.033, **p < 0.01, ***p < 0.001). Spearman’s rank correlation technique was applied to evaluate the relationship between (d) the frequency of H1-specific CD4⁺ T cells in the spleen and BM, e the frequency of IL-2⁺TNFα⁺IFNγ^- CD4⁺ T cells in the spleen and IgG avidity index (4 M urea) and f the frequency of IL-2⁺TNFα⁺IFNγ⁻ CD4⁺ T cells in the BM and HI titer (*p < 0.033, **p < 0.01).

Fig. 5. H1_Y98F-VLP improves viral clearance and reduces pulmonary inflammation upon lethal challenge.

Female Balb/c mice were challenged with 1.6 × 10³ TCID₅₀ of H1N1 (A/California/07/09) 28 days post-vaccination with 3 µg H1_WT- or H1_Y98F-VLP or an equivalent volume of PBS. a Mean weight loss (left) and survival (right) were monitored for 12 dpi (n = 12–14/group, error bars represent SEM). Mice weighing <80% of their pre-challenge weight were euthanized. A subset of challenged mice (n = 9/group) were euthanized at 3 dpi and 5 dpi for evaluation of the viral load and pulmonary inflammation. b Viral titers in the supernatant of lung homogenates calculated using the Karber method and reported as TCID₅₀/100 µl supernatant (GMT ± 95% CI). c Concentrations of cytokines and chemokines in the supernatant of lung homogenates measured by multiplex ELISA (mean ± SEM). The horizontal line represents the mean of mock-infected mice as a baseline. d Radar plots showing the cytokine profiles of mock-infected and infected lungs at 3 dpi and 5 dpi. e H&E stains of lungs collected 4 dpi (×10 magnification, scale bar = 100 µm). Statistical significance for b and c were determined by Kruskal–Wallis test with Dunn’s multiple comparisons. Comparisons between groups at the same time point are represented by * (*p < 0.033, **p < 0.01, ***p < 0.001) and comparisons within the same group over time are represented by ♦(^♦♦p < 0.01, ^♦♦♦p < 0.001).