SARS-CoV-2 Omicron boosting induces de novo B cell response in humans

Abstract

The primary two-dose SARS-CoV-2 mRNA vaccine series are strongly immunogenic in humans, but the emergence of highly infectious variants necessitated additional doses of these vaccines and the development of new variant-derived ones ^1-4 . SARS-CoV-2 booster immunizations in humans primarily recruit pre-existing memory B cells (MBCs) ^5-9 . It remains unclear, however, whether the additional doses induce germinal centre (GC) reactions where reengaged B cells can further mature and whether variant-derived vaccines can elicit responses to novel epitopes specific to such variants. Here, we show that boosting with the original SARS- CoV-2 spike vaccine (mRNA-1273) or a B.1.351/B.1.617.2 (Beta/Delta) bivalent vaccine (mRNA-1273.213) induces robust spike-specific GC B cell responses in humans. The GC response persisted for at least eight weeks, leading to significantly more mutated antigen-specific MBC and bone marrow plasma cell compartments. Interrogation of MBC-derived spike-binding monoclonal antibodies (mAbs) isolated from individuals boosted with either mRNA-1273, mRNA-1273.213, or a monovalent Omicron BA.1-based vaccine (mRNA-1273.529) revealed a striking imprinting effect by the primary vaccination series, with all mAbs (n=769) recognizing the original SARS-CoV-2 spike protein. Nonetheless, using a more targeted approach, we isolated mAbs that recognized the spike protein of the SARS-CoV-2 Omicron (BA.1) but not the original SARS-CoV-2 spike from the mRNA-1273.529 boosted individuals. The latter mAbs were less mutated and recognized novel epitopes within the spike protein, suggesting a naïve B cell origin. Thus, SARS-CoV-2 boosting in humans induce robust GC B cell responses, and immunization with an antigenically distant spike can overcome the antigenic imprinting by the primary vaccination series.

Extended Data Figure 1.. Robust GC and Tfh responses to mRNA-1273 and mRNA-1273.213 boosters.

(a) Frequencies of S-binding IgA-producing PB in PBMC 1 week post-boost measured by ELISpot in participants who received mRNA-1273 (left) and mRNA-1273.213 (right). (b) Plasma IgG binding to indicated strains of SARS-CoV-2 S measured by multiplex bead binding array in participants who received mRNA-1273 (upper) and mRNA-1273.213 (lower). (c) Gating strategy for analyzing S⁺ GC B cells and Tfh in FNA. (d) Frequencies of T follicular helper cells (Tfh) from FNA of draining lymph nodes. (e) Correlation between frequencies of S⁺ GC B cells and Tfh. (f) Representative ELISpot wells coated with BSA, and developed in blue (IgG) and red (IgA) after plating the indicated numbers of BMPCs. (g) Frequencies of IgA-secreting BMPCs specific for the indicated antigens 26 weeks post-boost. Black lines indicate medians. Symbols at each time point represent one sample. For mRNA-1273 and mRNA-1273.213 respectively, n = 7 and 38 (a), n = 6 and 28 (b), n = 5 and 20 (d), n = 3 and 10 (g).

Extended Data Figure 2.. Breadth of MBC-derived mAbs after mRNA-1273 and mRNA-1273.213 boosters.

(a) Gating strategy for sorting S⁺ MBC from PBMC. (b) Binding of mAbs to indicated antigens by ELISA performed in duplicate, presented as OD₄₉₀ minus two times the background signal to BSA.

Extended Data Figure 3.. Maturation of S+ MBCs in response to mRNA-1273 or mRNA-1273.213 booster.

(a) Gating strategy for sorting PB from PBMC. (b, c, e, g) UMAPs showing scRNA-seq transcriptional clusters of total cells (b, c) or B cells (e, g) from all participants (b, e) or from each participant separately (c, g). (d, f) Dot plots for the marker genes used for identifying the annotated clusters in (b, c) (d) and in (e, g) (f). (h) SARS-CoV-2 S+ clones visualized in red on UMAPs of B cells from each participant separately and faceted by time point. (i) Clonal overlap between S-binding PBs and GC B cells at indicated time points. Arc length corresponds to the number of BCR sequences and chord width corresponds to clone size. Purple chords correspond to clones spanning both compartments. Percentages are of GC B cell clones related to PBs at each time point. (j) Percentages of S-specific GC clones related to week 1 PBs. Symbols at each time point represent one sample, n = 6. (k) Representative flow cytometry plots of WA1/2020 and B.1.351 (left) or B.1.617.2 (right) staining of SA⁻ BCL6⁺CD38^int IgD^lo CD19⁺ CD3⁻ live singlet lymphocytes (top) or SA⁻ CD20⁺CD38^lo IgD^lo CD19⁺ CD3⁻ live singlet lymphocytes (bottom) in FNA samples from participants who received mRNA-1273.213. (l) Frequencies of B.1.351⁺ WA1/2020⁻ and B.1.617.2⁺ WA1/2020⁻ GC B cells (left) and MBC (right) from FNA of draining lymph nodes from participants who received mRNA-1273.213. Black lines indicate medians. Symbols represent one sample; n = 9.

Extended Data Figure 4.. Characterization of BA.1-specific mAbs.

(a) Frequencies of S-binding IgA-producing PB in PBMC 1 week post-boost measured by ELISpot in participants who received mRNA-1273.529. Black lines indicate medians. Symbols represent one sample; n = 7. (b) Neutralizing activity of mAbs from week 17 S⁺ MBCs against chimeric vesicular stomatitis virus in which the native envelope glycoprotein was replaced with S from WA1/2020 (with D614G mutation). (c) Binding of mAbs to BA.1 S and its constituent domains by ELISA performed in duplicate, presented as OD₄₉₀ values. S2-specific mAb 368–22 1B08 was described previously. (d) Titration of mAbs to determine neutralizing concentrations against VSV-SARS-CoV-2. Data are representative of two independent experiments.

Figure 1.. B cell response to mRNA-1273 and mRNA-1273.213 booster immunizations.

(a) Study design. Seven and thirty-nine healthy adult volunteers were enrolled and received an mRNA-1273 or mRNA-1273.213 booster, respectively. Blood was collected at baseline and at 1, 2, 4, 8, 17, and 26 weeks post-boost. Fine needle aspirates (FNAs) of ipsilateral axillary lymph nodes were collected 2 and 8 weeks post-boost from 5 and 20 participants, and bone marrow aspirates were collected from 3 and 11 participants 26 weeks post-boost in the mRNA-1273 and mRNA-1273.213 cohorts, respectively. (b) Frequencies of S-binding IgG-producing plasmablasts (PB) in PBMC 1 week post-boost measured by enzyme-linked immunosorbent spot (ELISpot) in participants who received mRNA-1273 (left) and mRNA-1273.213 (right). (c) Representative flow cytometry plots of BCL6 and CD38 staining of streptavidin (SA)⁻ IgD^lo CD19⁺ CD3⁻ live singlet lymphocytes in FNA samples (left), pooled (WA1/2020, 1.351, 1.617.2, and 1.1.529 all on BV421 and BV650) S probe staining on BCL6⁺CD38^int GC B cells (left center), and frequencies of S⁺ GC B cells from FNA of draining lymph nodes from participants who received mRNA-1273 (right center) and mRNA-1273.213 (right). (d) Representative flow cytometry plots of CD20 and CD38 staining of SA⁻ IgD^lo CD19⁺ CD3⁻ live singlet lymphocytes in PBMC (left), pooled S probe staining on CD20⁺CD38^lo/int B cells (left center), and frequencies of S⁺ memory B cells (MBCs) from PBMC 17 weeks post-boost in participants who received mRNA-1273 (right center) and mRNA-1273.213 (right). (e) Representative ELISpot wells coated with the indicated antigens, and developed in blue (IgG) and red (IgA) after plating the indicated numbers of bone marrow plasma cells (BMPCs). (f) Frequencies of IgG-secreting BMPCs specific for the indicated antigens 26 weeks post-boost in participants who received mRNA-1273 (left) and mRNA-1273.213 (right). Black lines indicate medians. Symbols at each time point represent one sample. For mRNA-1273 and mRNA-1273.213 respectively, n = 7 and 38 (b), n = 5 and 20 (c), n = 6 and 28 (d), n = 3 and 10 (f). (g, i) Binding of mAbs from S⁺ MBCs 17 weeks post-boost from participants who received mRNA-1273 (g) and mRNA-1273.213 (i) to indicated strains of SARS-CoV-2 S measured by multiplex bead binding array. (h, j) Summary of mAb cross-reactivity from participants who received mRNA-1273 (h) and mRNA-1273.213 (j).

Figure 2.. Maturation of S+ MBCs and BMPCs in response to mRNA-1273 or 1273.213 booster immunizations.

(a, b, e) Uniform manifold approximation and projection (UMAP) of scRNA-seq transcriptional clusters of B cells either from sorted circulating PBs 1 week post boost with log-normalised XBP1 gene expression (a) and S-specific clones (b) overlaid, or from FNAs of draining lymph nodes with S-specific clones overlaid (e). Each dot represents a cell. (c) Clonal overlap and percentages of S-specific PB clones related to clones generated during the primary vaccine response among participants receiving mRNA-1273 (upper) and mRNA-1273.213 (lower). Arc length corresponds to the number of B cell receptor sequences and chord width corresponds to clone size. Purple chords correspond to overlapping clones. Percentages are of PB clones related to pre-boost S-specific clones. (d, f, g) Paired median immunoglobulin heavy chain variable region gene (IGHV) mutation frequencies of S-specific clones found in PB both 1 week after the 2^nd dose of the primary vaccine series and boost (d), MBCs identified both 6 months after primary vaccination and 17 weeks after boost (f), and BMPCs identified both 6 and/or 9 months after primary vaccination and 6 months after boost (g). Each symbol represents the median mutation frequency of a clone; horizontal lines indicate medians. For mRNA-1273 and mRNA-1273.213 respectively, n = 52 and 104 (d), n = 44 and 41 (f), n = 7 and 16 (g). P values from two-sided Wilcoxon test.

Figure 3.. Neutralization capacity of MBC-derived mAbs.

(a) Frequencies of S-binding IgG-producing plasmablasts (PB) in PBMC 1 week post-boost measured by ELISpot in participants who received mRNA-1273.529. Horizontal lines indicate medians. Each symbol represents 1 sample, n = 7. (b) Binding of mAbs from S⁺ MBCs 17 weeks post-boost to indicated strains of SARS-CoV-2 S measured by multiplex bead binding array. (c) Summary of mAb cross-reactivity. (d) Neutralizing activity of mAbs from week 17 S⁺ MBCs against indicated strains of authentic SARS-CoV-2 virus from participants who received indicated booster vaccines. Each symbol represents an individual mAb, n = 39 for mRNA-1273, n = 49 for mRNA-1273.213, and n = 52 for mRNA-1273.529. Percentages are of mAbs below the 90% infection reduction threshold. P values from chi-squared tests between vaccine cohorts.

Figure 4.. Characterization of BA.1-specific mAbs.

(a) Gating strategy for sorting BA.1⁺ WA1/2020⁻ MBC from 17 weeks post-boost PBMC. (b) Binding of mAbs from BA.1⁺ WA1/2020⁻ sorted MBCs to indicated strains of SARS-CoV-2 S measured by multiplex bead binding array. (c) Summary of mAb binding. (d) Neutralizing activity of BA.1⁺ WA1/2020⁻ binding mAbs against indicated strains of authentic SARS-CoV-2 virus. Numbers above each virus are of mAbs below the 90% infection reduction threshold. (e) IGHV mutation frequencies of clones related to mAbs from participants 382–54 and 382–55 that neutralized D164G (left) and BA.1 but not D614G (right). Black lines indicate medians. Each symbol represents a sequence; n = 39 for D614G⁺ and n = 7 for BA.1⁺ D614G⁻. (f) Plaque assays on Vero E6 cells with indicated mAb in the overlay to isolate escape mutants (red arrows). Images are representative of three experiments per mAb. (g) Structure of RBD with hACE2 footprint highlighted in brown, BA.1 mutations highlighted in blue, and amino acids whose substitution confers resistance to indicated mAbs in plaque assays highlighted in red.