Broadly neutralizing antibodies against sarbecoviruses generated by immunization of macaques with an AS03-adjuvanted COVID-19 vaccine

Abstract

The rapid emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants that evade immunity elicited by vaccination has placed an imperative on the development of countermeasures that provide broad protection against SARS-CoV-2 and related sarbecoviruses. Here, we identified extremely potent monoclonal antibodies (mAbs) that neutralized multiple sarbecoviruses from macaques vaccinated with AS03-adjuvanted monovalent subunit vaccines. Longitudinal analysis revealed progressive accumulation of somatic mutation in the immunoglobulin genes of antigen-specific memory B cells (MBCs) for at least 1 year after primary vaccination. Antibodies generated from these antigen-specific MBCs at 5 to 12 months after vaccination displayed greater potency and breadth relative to those identified at 1.4 months. Fifteen of the 338 (about 4.4%) antibodies isolated at 1.4 to 6 months after the primary vaccination showed potency against SARS-CoV-2 BA.1, despite the absence of serum BA.1 neutralization. 25F9 and 20A7 neutralized authentic clade 1 sarbecoviruses (SARS-CoV, WIV-1, SHC014, SARS-CoV-2 D614G, BA.1, and Pangolin-GD) and vesicular stomatitis virus-pseudotyped clade 3 sarbecoviruses (BtKY72 and PRD-0038). 20A7 and 27A12 showed potent neutralization against all SARS-CoV-2 variants and multiple Omicron sublineages, including BA.1, BA.2, BA.3, BA.4/5, BQ.1, BQ.1.1, and XBB. Crystallography studies revealed the molecular basis of broad and potent neutralization through targeting conserved sites within the RBD. Prophylactic protection of 25F9, 20A7, and 27A12 was confirmed in mice, and administration of 25F9 particularly provided complete protection against SARS-CoV-2, BA.1, SARS-CoV, and SHC014 challenge. These data underscore the extremely potent and broad activity of these mAbs against sarbecoviruses.

Fig. 1.. AS03-adjuvanted RBD-NP/Hexapro-NP vaccination elicits progressive memory B cell maturation.

(A) Study overview. Rhesus macaques received AS03-adjuvanted RBD-NP or Hexapro-NP on day 0 and day 21 and received a booster with RBD (beta)-NP with AS03 at 12 months. Analysis of banked blood samples was performed as illustrated in the diagram. (B) Pseudovirus neutralizing antibody responses against viruses indicated on X-axis are shown. Each symbol represents an animal [RBD-NP (blue; n = 5) and Hexapro-NP (red; n = 6)], and paired samples are connected with a dashed line. The numbers within the graphs show geometric mean titers (GMTs). The statistical differences were calculated using two-way ANOVA and the statistical differences between indicated viruses and SARS-CoV-2 Wuhan strain at the same time points were labeled as (**P < 0.01 and ****P < 0.0001). (C) Representative flow cytometry plots show dual RBD and RBD (beta) binding B cells for RBD-NP vaccinated animals (blue), and dual spike protein and RBD (beta) binding B cells for Hexapro-NP vaccinated animals (red). Cells were pre-gated on live, CD3⁻ CD14⁻ CD16⁻ CD20⁺ IgD⁻ IgM⁻ IgG⁺ B cells. (D) The frequency of antigen-specific IgG⁺ MBCs relative to CD20⁺ B cells is shown for samples from the RBD-NP (blue) and Hexapro-NP (red) groups. The binding region is indicated on the X-axis. The statistical differences were calculated using two-way ANOVA (**P < 0.01). (E and F) Shown are the somatic hypermutation (SHM) rates of the productive IGHV genes (E) and IGLV genes (F) of B cells isolated from RBD-NP (blue) or Hexapro-NP (red) vaccinated animals at indicated time points. In (E) and (F), the boxes inside the violin plot show median, upper, and lower quartiles. The whiskers represent minimum and maximum values. Each dot represents an individual gene. The statistical differences between timepoints were calculated using one-way ANOVA (ns > 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001).

Fig. 2.. Maturation of the B cell response generates antibodies with greater potency and breadth.

(A) Diagram depicting the strategy for antigen (Ag)-specific MBCs sorting, mAb isolation, and characterization. (B) Graphs showing the neutralizing activity of monoclonal antibodies measured by pseudovirus neutralization assays. Bar graphs show IC₅₀s of all neutralizing antibodies against SARS-CoV-2 Wuhan (blue), BA.1 (purple), and BA.4/5 (orange) pseudoviruses. Each dot represents one antibody. Pie charts illustrate the fraction of non-neutralizing (no valid IC₅₀ or IC₅₀ > 100 μg/ml) antibodies (white slices); the inner circles show the number of antibodies tested. The frequency of neutralizing antibodies against SARS-CoV-2 Wuhan (blue slices), BA.1 (purple slices), or BA.4/5 (orange slices) pseudoviruses are shown on top of each pie chart, respectively. Bars and whiskers indicate geometric mean and geometric standard deviation (SD). Statistical significance was determined by the two-tailed Kruskal–Wallis test with subsequent Dunn’s multiple-comparisons test (*P < 0.05 and ****P < 0.0001). (C) Heatmap shows the neutralization activity of mAbs isolated at indicated time points against pseudotyped SARS-CoV-2 Wuhan, BA.1, and BA.4/5 respectively. Each unit within the heatmap represents one antibody. The color gradient indicates IC₅₀ values ranging from 0.001 (red) to 100 (white). Pie charts illustrate the fraction of non-neutralizers (white slices), SARS-CoV-2 Wuhan only (blue slices), Wuhan and BA.1 double (purple slices), Wuhan and BA.4/5 double (orange slices), and Wuhan, BA.1, BA.4/5 triple (red slices) neutralizing antibodies; the inner circle shows the number of antibodies tested at indicated time points. Statistical significance between the frequencies of the five categories of antibodies isolated from five different time points was determined using a two-tailed chi-square test (*P < 0.05). (D) The graphs show kinetic change of the potency, reported as IC₅₀ (μg/ml), of neutralizing antibodies against pseudotyped SARS-CoV-2 Wuhan (blue), BA.1 (orange) and BA.4/5 (purple), respectively. Each dot represents one antibody. Bars and whiskers indicate geometric mean and geometric SD. The statistical differences between timepoints were calculated using one-way ANOVA (**P < 0.01, ***P < 0.001 and ****P < 0.0001). (E) The heat map shows the IC₅₀ values of 15 selected mAbs against the indicated pseudoviruses. The heatmap range from 0.01 to 30 μg/ml is represented by white to dark red. (F) Graphs show the neutralization of authentic SARS-CoV-2 D614, SARS-CoV-2 BA.1, Pangolin, SARS-CoV, SHC014, and WIV-1 by indicated antibodies. Symbols are means ± SD. Dashed lines indicate IC₅₀ values. N=4. (G) 25F9 (red) and 20A7 (purple) mediated neutralization of VSV pseudoviruses containing spike proteins of Clade 3 African sarbecoviruses, BtKY72 (top) and PRD-0038 (bottom). Symbols are means ± SD. Dashed lines indicate IC₅₀ values. N=3. (H) The fold changes in neutralization IC₅₀ values of BQ.1, BQ.1.1, XBB relative to BA.1 are shown, with resistance colored from white to dark red.

Fig. 3.. 25F9 recognizes a conserved region on SARS-CoV-2 RBD.

The SARS-CoV-2 RBD is shown in white and human ACE2 is in pale green throughout all the figures; the heavy and light chains of 25F9 are in blue and lavender, respectively. For clarity, only variable domains of the antibodies are shown in all figures. (A) Shown are the relative positions of epitopes on SARS-CoV-2 RBD (white). The RBS is shown in pale green, the CR3022 site is shown in yellow, and the S309 site is shown in pink. Epitopes of 25F9, 20A7, and 21B6 are highlighted in blue, orange, and red outlines, respectively. RBS and epitope residues are defined as buried surface area (BSA > 0 Ų) as calculated by Proteins, Interfaces, Structures and Assemblies (PISA, http://www.ebi.ac.uk/pdbe/prot_int/pistart.html). (B) Shown is the crystal structure of 25F9 in complex with SARS-CoV-2 RBD. (C) The surface area of SARS-CoV-2 buried by heavy and light chains of 25F9 is shown. (D) 25F9 interacts with RBD using CDRs H2, H3, L1, and L3, as well as LFR3. (E) SARS-CoV-2 RBD with 25F9 superimposed onto an RBD-ACE2 complex structure (PDB 6M0J) shows that 25F9 would clash (indicated with a red circle) with ACE2. (F) Shown is sequence alignment of epitope residues in a subset of SARS-like viruses. Residues that differ from wild-type SARS-CoV-2 are indicated in red. SARS2, wild-type SARS-CoV-2; SARS1, SARS-CoV-1. (G to I) Molecular interactions between RBD and CDR H2 (G), CDR H3 (H), and light chain (I) are shown. Hydrogen bonds and salt bridges are indicated by dashed lines.

Fig. 4.. 20A7 accommodates mutations of SARS-CoV-2 variants.

The SARS-CoV-2 wild-type and BA.2 RBDs are shown in white and black; the heavy and light chains of 20A7 are in orange and yellow. Hydrogen bonds and salt bridges are indicated by dashed lines. (A) Shown is the crystal structure of 20A7 in complex with SARS-CoV-2 wild-type RBD. (B) SARS-CoV-2 RBD in complex with 20A7 superimposed onto an RBD-ACE2 complex structure (PDB 6M0J) shows that 20A7 would clash (red circle) with ACE2. (C) Shown is the surface area of SARS-CoV-2 wild-type RBD that would be buried by heavy and light chains of 20A7. (D) Molecular interactions between wild-type SARS-CoV-2 RBD and 21B6 CDR H2. (E) Molecular interactions between wild-type SARS-CoV-2 RBD and 21B6 CDR H3. (F) The crystal structure of 20A7 with RBD (BA.2) shows that 20A7 targets BA.2 in the same binding mode as wild-type SARS-CoV-2. Mutated residues in the Omicron BA.2 subvariant are indicated by red spheres. (G) Shown is a structural comparison of the interaction of 20A7 with wild-type (white) and BA.2 (black) RBDs. (H) 20A7 V_H E33 forms a hydrogen bond with RBD. (I) All alleles of VH3-73*01 encode Glu at position 33 whereas those of VH3-73*02 encode Val. (J) Biolayer interferometry (BLI) binding assay results showed that E33V reduced the binding of 20A7 to SARS-CoV-2 RBD by about 300-fold.

Fig. 5.. 25F9, 20A7, and 27A12 protect aged mice from SARS-CoV-2 and other sarbecovirus-induced pathology.

(A) Shown is a diagram depicting the challenge study in mice. 25F9, 20A7, 27A12, or a DENV2(2D22) control antibody were administered intraperitoneally at 200 μg per animal into 14 groups of aged mice (10 animals per group). Animals were challenged intranasally 12 hours after antibody infusion with one of the indicated sarbecoviruses (mouse-adapted SARS-CoV-2, 1 × 10³ PFU; mouse-adapted SARS-CoV-2-BA.1, 1 × 10⁵ PFU; mouse-adapted SARS-CoV, 1 × 10⁴ PFU; or SHC014 MA15, 1 × 10⁵ PFU). Lungs were collected on day 2 or 4 after infection. As a control, groups of mice were exposed only to phosphate-buffered saline (PBS) in the absence of virus. (B) Shown is the body weight change of mice after challenge with moused-adapted SARS-CoV-2, SARS-CoV-2 BA.1, SARS-CoV, and SHC014, respectively. Data are presented as mean ± SEM from 10 animals per group from days 0 to 2, or 5 animals per group from days 3 to 4. Data were analyzed using a mixed-effects model with post hoc Dunnett’s multiple tests in comparison with the control group; significance is indicated as **p < 0.01, ***p < 0.001, and ****p < 0.0001 or ns when not significant. The dotted horizontal line at 80% designates a weight loss amount at which softened mouse food is added to the cages. (C) Lung gross pathology was scored at the collection on day 2 and 4 post-infection in mice prophylactically treated with indicated bnAbs or the isotype control mAb (n = 5 per group). Individual mice are represented by the dot plots. Data are presented as mean values ± SEM. (D) Lung virus titers (PFU per lung) were determined by plaque assay of lung tissues collected at days 2 or 4 after infection (n = 5 individuals per time point for each group). Data are shown as scatter dot plots with bar heights representing the mean and whiskers representing SEM. Data were analyzed with a mixed-effects model with post hoc Dunnett’s multiple tests in comparison with the control group; significance is indicated as **p < 0.01, ***p < 0.001, and ****p < 0.0001 or ns when not significant. The dotted horizontal line indicates the limit of detection (50 PFU) for the plaque assay. For samples with values below this, data is plotted at half the limit of detection.