Subdominance and poor intrinsic immunogenicity limit humoral immunity targeting influenza HA stem

Abstract

Both natural influenza infection and current seasonal influenza vaccines primarily induce neutralizing antibody responses against highly diverse epitopes within the "head" of the viral hemagglutinin (HA) protein. There is increasing interest in redirecting immunity toward the more conserved HA stem or stalk as a means of broadening protective antibody responses. Here we examined HA stem-specific B cell and T follicular helper (Tfh) cell responses in the context of influenza infection and immunization in mouse and monkey models. We found that during infection, the stem domain was immunologically subdominant to the head in terms of serum antibody production and antigen-specific B and Tfh cell responses. Similarly, we found that HA stem immunogens were poorly immunogenic compared with the full-length HA with abolished sialic acid binding activity, with limiting Tfh cell elicitation a potential constraint to the induction or boosting of anti-stem immunity by vaccination. Finally, we confirm that currently licensed seasonal influenza vaccines can boost preexisting memory responses against the HA stem in humans. An increased understanding of the immune dynamics surrounding the HA stem is essential to inform the design of next-generation influenza vaccines for broad and durable protection.

Keywords: Adaptive immunity; B cells; Immunology; Influenza; Vaccines.

Figure 1. Serological and B cell responses in experimentally infected mice.

(A) Serum endpoint total IgG titers were measured by ELISA using HA-FL (blue) or stabilized HA stem (red) in mice infected intranasally with PR8 (n = 6 per time point). Dotted lines denote the detection cutoff (1:100 dilution). Data represent the mean ± SEM. (B) Frequency of GC B cells (B220⁺IgD^–CD38^loGL7⁺) and memory B cells (B220⁺IgD^–CD38^hiGL7^–) binding HA-FL (blue) or HA stem (red) (n = 6). Data represent the mean ± SEM. (C) Frequency of plasma cells (CD138⁺B220^–IgD^–) binding HA-FL (blue) or HA stem (red) (n = 6). Data represent the mean ± SEM. (D) HA bioavailability visualized by monoclonal anti–HA head or anti–HA stem antibody staining (white) and B220⁺ B cell staining (green). Scale bars: 100 μm.

Figure 2. Serological and B cell responses in experimentally infected macaques.

(A) Serum endpoint total IgG titers were measured by ELISA using CA09 HA-FL (blue) or stabilized CA09 HA stem (red) in macaques (n = 8) infected intranasally with A/Auckland/1/2009. Note that 2 animals were sacrificed on day 23. Dotted lines denote the detection cutoff (dilution 1:100). (B) Frequency of IgG⁺ memory B cells (CD19⁺IgD^–IgG⁺) binding CA09 HA-FL (blue) or stabilized CA09 HA stem (red) was measured by flow cytometry within cryopreserved PBMC samples from infected macaques (n = 6). Note that the 2 animals sacrificed on day 23 were excluded.

Figure 3. Serological and B cell responses in primary vaccinated mice.

Serum endpoint total IgG titers were measured by ELISA using HA-FL, HA stem, or PE proteins in mice that were immunized (A) 5 times with unadjuvanted immunogens or (B) 2 times with adjuvanted (Addavax) immunogens (n = 10 from 2 independent experiments with groups of 5 animals). Mice were immunized at 3-week intervals, and serum was collected every 2 weeks after immunization. Dotted lines denote the detection cutoff (dilution 1:400). Box boundaries represent 25^th and 75^th percentiles, the inner line represents the median, and whiskers represent minimum and maximum values. (C) Representative flow cytometric plots and (D) frequency of GC B cells (B220⁺IgD^–CD38^loGL7⁺) from mice vaccinated once with adjuvanted (Addavax) immunogens double stained with HA stem probes (PR8) (n = 10 from 2 independent experiments with groups of 5 animals). Data indicate the mean ± SEM. Adj., adjuvant.

Figure 4. Antigen specificity of Tfh cells following immunization or infection in mice.

(A) Tfh cells were quantified in the 2 draining ILNs on day 14 after vaccination with PR8 HA-FL or stem KLH antigens (n = 5). (B–D) Antigen-specific Tfh cells were identified either by OX-40 upregulation in combination with ICOS⁺⁺ or CD25 coexpression or CD154 expression following 18 hours of stimulation with HA head or HA stem peptide pools. Antigen-specific responses are presented after background subtraction using a DMSO control (dotted line indicates no change above background). Samples were collected on day 14 after immunization or infection with PR8 HA-FL (n = 12), stem KLH protein (n = 10), or 50 TCID₅₀ PR8 virus (n = 5). Error bars indicate the median and IQR. **P < 0.01, by Wilcoxon matched-pairs test.

Figure 5. Stem immunogens fail to selectively recall stem antibodies in preimmune mice and macaques.

Mice infected intranasally with PR8 and immunized on day 56 were analyzed for (A) serum endpoint total IgG titers measured by ELISA on day 56 (black box shows before immunization) and day 70 (white box shows 2 weeks after immunization) using HA-FL or HA stem proteins (n = 10 from 2 independent experiments with groups of 5 animals) and (B) frequency of GC B cells (B220⁺IgD^–CD38^loGL7⁺) stained with HA-FL and HA stem probes (PR8) measured by flow cytometry (n = 5). Box boundaries represent the 25^th and 75^th percentiles, the inner line represents the median, and whiskers represent minimum and maximum values. Data indicate the mean ± SEM. *P < 0.05, by Mann-Whitney U test. Macaques (n = 6) infected intranasally with A/Auckland/1/2009 and immunized on day 56 with seasonal IIV4 or HA stem immunogen were analyzed for (C) serum endpoint total IgG titers measured by ELISA using CA09 HA-FL (blue) or stabilized CA09 HA stem (red) proteins and (D) frequency of IgG⁺ memory B cells (CD19⁺IgD^–IgG⁺) binding CA09 HA-FL (blue) or stabilized CA09 HA stem (red) probes measured by flow cytometry within cryopreserved PBMC samples. Dotted lines denote the detection cutoff (dilution 1:100).

Figure 6. Stem antibody and memory B cells are expanded by seasonal influenza vaccination in humans.

(A) Serum endpoint total IgG titers of antibody binding the HA-FL (blue) or the stabilized HA stem (red) proteins at baseline and following immunization (day 28 [d28]) with seasonal IIV3 (n = 29) or IIV4 (2016: n = 18; 2017: n = 21). (B) Representative flow cytometric plots of IgG⁺ memory B cells from IIV4 (2016 season) recipients double stained with recombinant HA-FL or HA stem probes (CA09). Memory B cells were defined as CD19⁺IgD^–IgG⁺ after prior exclusion of doublets, dead cells, and CD3⁺, CD14⁺, CD16⁺, CD8⁺, and CD10⁺ cells. Frequency (C) and percentage change (D) of IgG⁺ memory B cells binding HA-FL or HA stem between baseline and after IIV4 (2016 season) immunization (n = 18). (E) Representative flow cytometric plots of IgG⁺ memory B cells from IIV4 (2017 season) recipients costained with recombinant HA-FL (A/Michigan/45/2015) or HA stem probes (CA09). Frequency (F) and percentage change (G) of IgG⁺ memory B cells binding either HA-FL or HA stem between baseline and after IIV4 (2017 season) immunization (n = 21). Data represent the mean ± SEM. *P < 0.05, by Mann-Whitney U test (A, C, and E).