Cross-Neutralization of Emerging SARS-CoV-2 Variants of Concern by Antibodies Targeting Distinct Epitopes on Spike

Abstract

Several severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants have arisen that exhibit increased viral transmissibility and partial evasion of immunity induced by natural infection and vaccination. To address the specific antibody targets that were affected by recent viral variants, we generated 43 monoclonal antibodies (mAbs) from 10 convalescent donors that bound three distinct domains of the SARS-CoV-2 spike. Viral variants harboring mutations at K417, E484, and N501 could escape most of the highly potent antibodies against the receptor binding domain (RBD). Despite this, we identified 12 neutralizing mAbs against three distinct regions of the spike protein that neutralize SARS-CoV-2 and variants of concern (VOCs), including B.1.1.7 (alpha), P.1 (gamma), and B.1.617.2 (delta). Notably, antibodies targeting distinct epitopes could neutralize discrete variants, suggesting that different variants may have evolved to disrupt the binding of particular neutralizing antibody classes. These results underscore that humans exposed to the first pandemic wave of prototype SARS-CoV-2 possess neutralizing antibodies against current variants and that it is critical to induce antibodies targeting multiple distinct epitopes of the spike that can neutralize emerging variants of concern. IMPORTANCE We describe the binding and neutralization properties of a new set of human monoclonal antibodies derived from memory B cells of 10 coronavirus disease 2019 (COVID-19) convalescent donors in the first pandemic wave of prototype SARS-CoV-2. There were 12 antibodies targeting distinct epitopes on spike, including two sites on the RBD and one on the N-terminal domain (NTD), that displayed cross-neutralization of VOCs, for which distinct antibody targets could neutralize discrete variants. This work underlines that natural infection by SARS-CoV-2 induces effective cross-neutralization against only some VOCs and supports the need for COVID-19 vaccination for robust induction of neutralizing antibodies targeting multiple epitopes of the spike protein to combat the current SARS-CoV-2 VOCs and any others that might emerge in the future.

Keywords: SARS-CoV-2; humoral immunity; immune memory; infectious disease; monoclonal antibodies; single cell; variants of concern.

FIG 1

Analyses of serum antibody responses in COVID-19 convalescent individuals. (a and b) Total IgG endpoint antibody titers from 10 convalescent subjects against SARS-CoV-2 full-length spike variants (a) and RBD recombinant antigens (b). The dashed line is the mean IgG titer. (c) Neutralization titers from 10 convalescent donors against WT SARS-CoV-2, B.1.1.7, P.1, B.1.617.2, and B.1.617.1. The dashed line represents the mean neutralization titer. Data in panels a to c were analyzed using nonparametric Friedman’s test with Dunnett’s multiple-comparison test. Fold changes in relative mAb binding to variants or mutants compared to the WT in panels a and b are indicated above the statistical asterisks. ns, not significant.

FIG 2

Characterization of spike-reactive mAbs. (a and b) Uniform manifold approximation and projection (UMAP) of SARS-CoV-2 spike non-RBD binding (a) and spike RBD binding (b) B cells isolated from PBMCs of 10 convalescent subjects. (c) Proportion of spike non-RBD- and spike RBD-specific binding B cells. The number in the center of the pie chart indicates the number of antigen-specific binding B cells. (d) mAbs generated from selected B cells (n = 43) were tested for binding to full-length spike, S1, S2, and the RBD and neutralization potential against WT SARS-CoV-2. Binding data are represented as areas under the curve (AUC). Neutralizing activities of <10,000 ng/ml are considered neutralizing. (e and f) Pie charts of mAb domain specificity (e) and neutralizing capability (f). The numbers in the center of the pie graphs indicate the number of antibodies tested. (g) Comparison of neutralizing capabilities of mAbs targeting the spike RBD and spike non-RBD. (h and i) IC₅₀s of the neutralization potencies of spike-reactive antibodies against WT virus based on domain specificity (h) and by subject (i). The mean in panel h is indicated as a solid line. Data in panels h and i are colored based on domain specificity, and the dashed lines shown in panels h and i indicate the limit of detection (10,000 ng/ml). Data in panels d to i are representative of results from two independent experiments performed in duplicate. Genetic characterization of each mAb is further detailed in Table S2 in the supplemental material.

FIG 3

Binding breadth and neutralization of spike non-RBD mAbs. (a) Full-length spike protein binding to ACE2 (PDB accession number 7KJ2). (b to g) Locations of mutations found on B.1.1.7 (b), B.1.351 (c), P.1 (d), B.1.617.2 (e), B.1.526 (f), and B.1.617.1 (g) (modified from the structure under PDB accession number 6XM4). (h) Binding reactivity and neutralization capabilities 5 of NTA-A (pink)-, NTD-B (green)-, and S2 (purple)-reactive mAbs. The color gradients indicate the percentages of relative binding compared to WT spike. The neutralization potencies (IC₅₀) of spike non-RBD mAbs against the WT and the B.1.1.7, P.1, B.1.617.2, and B.1.617.1 variants are indicated in nanograms per milliliter. The panel of SARS-CoV-2s is detailed in Table S4 in the supplemental material. Data in panel h are representative of results from two independent experiments performed in duplicate. Genetic information for each mAb can be found in Table S2.

FIG 4

Binding and neutralization profiles of RBD binding mAbs against a panel of RBD escape mutants and variants. (a) Structural model of RBD “up” binding with ACE2 (PDB accession number 7KJ2) and RBD antibody classes and associated escape mutants. (b) RBD colored by antibody classes and associated mutations. Pink, class 1; purple, overlap of classes 1 and 2; blue, class 2; teal, class 3. (c) Heat map detailing the binding reactivity of RBD mAbs (n = 29) against single key escape sites for class 1, class 2, and class 3 antibodies; combinations of RBD mutants; and the RBDs from SARS-CoV-1 and MERS-CoV. a refers to class 3-like antibodies, which are defined by mAbs that compete with a class 3 mAb (see Fig. S2c in the supplemental material). b to f refer to mutations in the RBD of each full-length spike variant, B.1.1.7 with N501Y (b), B.1.351 with K417N:E484K:N501Y (c), P.1 with K417T:E484K:N501Y (d), B.1.617.2 with T478K:L452R (e), B.1.526 with E484K (f), and B.1.617.1 with L452R:E484Q (g). The panel of recombinant antigens in panel c is detailed in Table S3, including mutations found in circulating SARS-CoV-2 variants (boldface type), mutations that escape/reduce binding by polyclonal serum/potent neutralizing mAbs (italic type), mutations found in both circulating SARS-CoV-2 variants and the in vitro escape map (boldface and italic type), and artificial mutants at key contact residues of the RBD-ACE2 interaction (normal typeface). The neutralization potencies (IC₅₀) of spike RBD mAbs against the WT and the B.1.1.7, P.1, B.1.617.2, and B.1.617.1 variants are indicated in nanograms per milliliter. The panel of SARS-CoV-2s is detailed in Table S4. Data in panel c are representative of results from two independent experiments performed in duplicate. Genetic information for each antibody can be found in Table S2.