Research ArticleINFECTIOUS DISEASE

Preferential induction of cross-group influenza A hemagglutinin stem–specific memory B cells after H7N9 immunization in humans

+ See all authors and affiliations

Science Immunology  14 Jul 2017:
Vol. 2, Issue 13, eaan2676
DOI: 10.1126/sciimmunol.aan2676
  • Fig. 1 Cross-reactivity of the memory IgG+ B cell response to H5N1 or H7N9 vaccination.

    (A) Phylogenetic tree showing relationship of HA protein from different influenza A subtypes. HA sequences were aligned, and a maximum likelihood tree based on the amino acid sequence was produced using MEGA 7.0 (53). The influenza strain and respective accession numbers of each HA used are listed in table S1. (B to D) Representative flow cytometry plot of the percentage of IgG+ B cells that bound full-length HA or stabilized stem of the influenza strain HA as indicated. Doublets and dead cells were excluded followed by gating on CD3CD8CD14CD56CD19+ CD38lo IgG+ cells. Cells were further gated on H5 HA+ (top) or H7 HA+ (bottom) for (C) and (D). (E) Percentage of IgG+ B cells in each donor that bound HA of the vaccinating influenza strain (H5 or H7 HA). The line indicates the mean percentage. (F) IgG+ H5 HA+ in H5N1 vaccinated volunteers and H7 HA+ cells in H7N9 vaccinated volunteers were tested for the ability to bind H3 (A/Texas/50/2012) and/or H1 (A/California/04/2009) HA as shown in (C). Percentage of all H5 HA+ or H7 HA+ B cells, depending on the vaccine trial, that were (left) subtype-specific (H5+H1H3 or H7+H1H3), (middle) cross-reactive with HA from a subtype within the same group (intragroup reactive; H5+H1+H3 or H7+H3+H1), or cross-reactive with HA from subtypes across groups (cross-group reactive; H5+H1+H3+ or H7+H3+H1+). Each dot represents the data from one donor with the line indicating the mean. (G) Same data as in (F), summarized as the proportion of the total vaccinating HA response with the indicated level of cross-reactivity to influenza subtypes. The mean in each vaccine group with SEM is shown. (H and I) Percentage (H) or number per 106 IgG+ cells (I) that bound H5 or H7 HA plus the group 1 H1 stabilized stem probe and full-length H3 and H1, as shown in (D). Data were log10-transformed for (E) and (I). Statistical significance was determined for all plots using unpaired nonparametric Mann-Whitney t test.

  • Fig. 2 VH usage of cross-reactive B cells induced by H5N1 or H7N9 vaccination.

    (A) Flow cytometry plots showing the gate used to single-cell sort cross-reactive H5H1+ or H7H3+ B cells from H5N1 (top) or H7N9 (bottom) vaccinated donors. (B) Graphs showing the percentage of IgG+ cells that are H5 or H7 HA+ (left) and proportion of HA+ cells that also cross-react with HA from another subtype in the same group (right) in each donor as detected by flow cytometry. The red dots indicate the donors from which single-cell sorting and repertoire analysis were performed. (C and D) Comparison of percentage of Ig heavy chain sequences from all single cell–sorted H5H1+ or H7H3+ cells encoded by VH1-69, VH1-18, or VH6-1 genes, as indicated. In (C), the percentage of all sequences from each vaccine trial is shown. Differences in frequency for each of these three VH genes between vaccine studies were significant (P < 0.0001, Fisher’s exact test). In (D), each dot represents the mean percentage for each donor, with the line indicating the mean for all 20 donors in each vaccine trial. P values were determined using unpaired nonparametric Mann-Whitney t test. n.s., not significant.

  • Fig. 3 Evaluation of multidonor class Ig sequences obtained from H5N1 and H7N9 vaccinated donors.

    (A) Criteria used to classify paired heavy and light chains from single cells into the three multidonor classes. (B and C) Percentage of H5H1+, H7H3+, H5+H1, or H7+H3 single cell–sorted cells that expressed a VH1-18 or VH6-1 heavy chain and had an Ig molecular signature that fit one of the three classes indicated in (A). No H5+H1 or H7+H3 B cells expressing VH6-1 were found. The total number of sequences analyzed in each category is indicated above the bar. (D) Percentage of all H5H1+ or H7H3+ cells whose Ig heavy chain sequence matched the molecular signature of one of the three multidonor classes. Statistical significance between the H5H1+ and H7H3+ cells by Fisher’s exact test was P < 0.0001. (E) Each pie chart shows the proportion of H5H1+ (top) or H7H3+ (bottom) cross-reactive B cells in each donor that expressed an Ig heavy/light chain that matched one of the indicated multidonor classes or VH1-69. The volunteer number and total number of sequences obtained from that donor are indicated above each pie chart.

  • Fig. 4 Functional characterization of multidonor class antibodies.

    (A and B) mAbs belonging to the multidonor VH6-1, VH1-18+ DH3-9, and VH1-18 QXXV classes or control antibodies were tested for binding to full-length recombinant HA or the stabilized stem of the indicated influenza strains (A) or neutralization in a microneutralization assay with the indicated strains (B). Full influenza strain names are detailed in Materials and Methods. Antibodies in bold were published previously (24) and shown for comparison. Control antibodies were previously characterized as group 1 stem binding (CR6261) (17), group 2 stem binding (CR8020) (21), or group 1 and group 2 stem binding (CR9114) (2). All other antibodies were cloned from B cells isolated after H7N9 vaccination. In (A), binding of each antibody to each strain is color-coded according to the AUC of the titration curve as indicated in the color legend. In (B), the mAbs are color-coded according to the microneutralization IC50 for each strain, as indicated in the legend. Actual values are in table S4. Representative data are shown from at least two independent experiments. (C) Mice were passively administered with the indicated antibody (5 mg/kg) (10 mice per group) 24 hours before infection with A/Anhui/01/2013 (H7N9) or A/California/07/2009 (H1N1) as indicated. As a control, one group received VRC01 IgG, an HIV-specific antibody. Mice were euthanized when they reached 80% of their initial weight. The left graph shows the Kaplan-Meier survival curve. There was greater survival in all groups that received an HA stem–binding antibody (P ≤ 0.001, Fisher’s exact test) compared to the control group. The middle graph shows the percent weight loss over time with each line representing one mouse. On the right, the percent weight loss is shown for animals in each group on the last day when all animals infected with that strain had a recorded weight. The weight loss is statistically lower (P < 0.0001, Mann-Whitney t test) in all groups that received an HA stem–binding antibody compared to the control group.

  • Fig. 5 Characterization of a novel HA stem–binding multidonor class.

    (A) Pie charts showing proportion of sequences obtained from stem-binding B cells in four donors that expressed Igs with a canonical 18–amino acid CDRH3. Clonally related Igs are separated by white lines in each pie chart. Clones with the canonical CDRH3 motif are in green, and all other sequences are in gray. (B) Table showing genetic characteristics of a representative antibody from each lineage. (C) Negative-stain EM images of A/Shanghai/02/13 H7 HA, an antibody Fab alone, and H7 HA in complex with Fabs of indicated antibodies. (D and E) Binding AUC (D) and microneutralization IC50 (E) of representative mAbs as described in Fig. 4. (F) Kaplan-Meier survival and weight loss over time of H7N9- or H1N1-infected mice 24 hours after passive transfer of 02-1F07 (5 mg/kg) as in Fig. 4. (G) Crystal structures of three representative antibodies are overlaid with the CDRH3 loop shown in the inset panels. The overlaid structures are rotated to allow visualization of the highly similar structures as well as the protruding nature of the hydrophobic CDRH3 tip.

  • Fig. 6 Characterization of unique cross-group HA stem–binding lineages.

    (A) Pie charts showing proportion of volunteer-unique cross-group HA stem–binding lineages in two donors. White lines separate clonally related sequences. The lineages described here are in orange. (B) Table showing genetic characteristics of a representative antibody from each lineage. (C) Negative-stain EM images of H7 HA in complex with Fabs of indicated antibodies. (D and E) Binding AUC (D) and microneutralization IC50 (E) as in Fig. 4 for one representative antibody of each lineage. (F) Kaplan-Meier survival curves and weight loss over time of mice infected with an H7N9 or H1N1 influenza strain 24 hours after passive transfer of the 53-1A09 antibody (5 mg/kg). (G) Crystal structures of 13-1B02 and 53-1A09 antibodies are shown in ribbon representation with select CDRH3 residues shown in stick representation (top). CDR regions 1 to 3 for both the heavy and light chain are also indicated. Both structures reveal a protruding hydrophobic CDRH3. The two antibodies are overlaid (bottom) with the most protruding CDRH3 residue for each shown in stick representation. The disulfide bond of 13-1B02 is also shown in stick.

  • Fig. 7 Longitudinal analysis of cross-group HA stem–binding lineages.

    (A) Cells were single cell–sorted from the same donor from blood collected on the day of the H7 prime (PreVacc) and 2 weeks after the H7N9 boost (PostVacc). Each bar represents the total number of sequences from each donor amplified from cells sorted after vaccination that were H7H3+ (left) or bound H7 alone (right) by flow cytometry. The sequences found after vaccination that belonged to lineages detected before vaccination as well are in black. (B) For each of the 36 lineages found before and after vaccination, we calculated the approximate percentage of this lineage of all IgG+ cells at each time point by flow cytometry and single-cell Ig sequencing. Each line connects the proportion of a single lineage in the IgG+ B cell population for each time point. Significance was determined using Wilcoxon paired nonparametric t test. (C) Mean percent mutation level of the rearranged VH gene amplified from single cell–sorted cells that were cross-group HA stem–specific (H7H3H1+, Stem+), group 2 only–specific (H7H3+ H1), or bound H7 HA alone as indicated. Each dot connected by lines represents the mean mutation percent in each donor for each B cell specificity group. (D) For each donor, the percentage of Ig sequences from B cells with the indicated specificity as in (C) that were clonally related to another sequence is plotted. Statistical significance for (B) to (D) was determined using a paired nonparametric Wilcoxon t test. (E, G, and I) Pie charts showing sequences from all cross-reactive H5H1+ or H7H3+ B cells generated after H5N1 or H7N9 vaccination, respectively, in the same donor that received both vaccines. Individual clones are separated by white lines. Colored slices of the pie charts represent all cross-group HA stem–binding B cells as detected by flow cytometry. Representative mAbs from some of the lineages were also expressed, and cross-group binding was confirmed (Figs. 4 and 6). All other cross-reactive B cells are in gray. The total number of cross-reactive B cells sorted and sequenced is indicated in the center of the chart. (F and H) Approximate percentage of each of the indicated lineages in the total IgG+ population 2 weeks after the H5N1 boost, the day of the H7 vaccine prime (PreVacc), and 2 weeks after the H7N9 boost.

Supplementary Materials

  • immunology.sciencemag.org/cgi/content/full/2/13/eaan2676/DC1

    Fig. S1. Immunization protocols and response.

    Fig. S2. Ig repertoire analysis of heterosubtypic B cells.

    Fig. S3. Analysis of Ig heavy chain junction of VH1-2/VH3-53 class antibodies.

    Fig. S4. B cell response of volunteers that received both the H5N1 and H7N9 vaccine.

    Fig. S5. Gating strategy for analyzing and sorting HA-specific B cells by flow cytometry.

    Table S1. List of influenza strain HA proteins.

    Table S2. Demographic information for donors used for single-cell sorting and Ig sequencing.

    Table S3. Genetic characteristics of Ig heavy and light chains of expressed and characterized mAbs.

    Table S4. IC50 microneutralization titers.

    Table S5. IC80 neutralization titers by pseudovirus entry inhibition assay.

    Table S6. Crystallographic data collection and refinement statistics.

    Table S7. Ig sequences of cross-group HA-specific antibodies.

    Raw data for Figs. 1, 2, and 7 and figs. S1 and S2

  • Supplementary Materials

    Supplementary Material for:

    Preferential induction of cross-group influenza A hemagglutinin stem–specific memory B cells after H7N9 immunization in humans

    Sarah F. Andrews,* M. Gordon Joyce, Michael J. Chambers, Rebecca A. Gillespie, Masaru Kanekiyo, Kwanyee Leung, Eun Sung Yang, Yaroslav Tsybovsky, Adam K. Wheatley, Michelle C. Crank, Jeffrey C. Boyington, Madhu S. Prabhakaran, Sandeep R. Narpala, Xuejun Chen, Robert T. Bailer, Grace Chen, Emily Coates, Peter D. Kwong, Richard A. Koup, John R. Mascola, Barney S. Graham, Julie E. Ledgerwood, Adrian B. McDermott*

    *Corresponding authors. Email: adrian.mcdermott{at}nih.gov (A.B.M.); sarah.andrews2{at}nih.gov (S.F.A.)

    Published 14 July 2017, Sci. Immunol. 2, eaan2676 (2017)
    DOI: 10.1126/sciimmunol.aan2676

    This PDF file includes:

    • Fig. S1. Immunization protocols and response.
    • Fig. S2. Ig repertoire analysis of heterosubtypic B cells.
    • Fig. S3. Analysis of Ig heavy chain junction of VH1-2/VH3-53 class antibodies.
    • Fig. S4. B cell response of volunteers that received both the H5N1 and H7N9 vaccine.
    • Fig. S5. Gating strategy for analyzing and sorting HA-specific B cells by flow cytometry.
    • Table S1. List of influenza strain HA proteins.
    • Table S2. Demographic information for donors used for single-cell sorting and Ig sequencing.
    • Table S3. Genetic characteristics of Ig heavy and light chains of expressed and characterized mAbs.
    • Table S4. IC50 microneutralization titers.
    • Table S5. IC80 neutralization titers by pseudovirus entry inhibition assay.
    • Table S6. Crystallographic data collection and refinement statistics.

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    Files in this Data Supplement: