Protective effect and molecular mechanisms of human non-neutralizing cross-reactive spike antibodies elicited by SARS-CoV-2 mRNA vaccination

Abstract

Neutralizing antibodies correlate with protection against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Recent studies, however, show that binding antibody titers, in the absence of robust neutralizing activity, also correlate with protection against disease progression. Non-neutralizing antibodies cannot directly protect against infection but may recruit effector cells and thus contribute to the clearance of infected cells. Additionally, they often bind conserved epitopes across multiple variants. Here, we characterize 42 human monoclonal antibodies (mAbs) from coronavirus disease 2019 (COVID-19)-vaccinated individuals. Most of these antibodies exhibit no neutralizing activity in vitro, but several non-neutralizing antibodies provide protection against lethal challenge with SARS-CoV-2 in different animal models. A subset of those mAbs shows a clear dependence on Fc-mediated effector functions. We have determined the structures of three non-neutralizing antibodies, with two targeting the receptor-binding domain and one that binds the subdomain 1 region. Our data confirm the real-world observation in humans that non-neutralizing antibodies to SARS-CoV-2 can be protective.

Keywords: ADCC; ADCP; CP: Immunology; effector functions; immunology; mAbs; non-neutralizing mAbs; spike.

Figure 1.. Characterization of mAb binding and neutralization against SARS-CoV-2 variants

(A–D) Binding of RBD-specific (A), NTD-specific (B), S2-specific (C), and unknown-epitope-specific (D) SARS-CoV-2 mAbs to the spike proteins of SARS-CoV-2 variants. Neutralizing mAbs are denoted with red Ns. (E) Neutralizing activity of the neutralizing mAbs against authentic SARS-CoV-2 variants. Crosses denote a complete loss of binding/neutralization.

Figure 2.. Investigating the protection elicited by the mAbs using lethal animal challenge models

(A–D) Maximum weight loss and survival of BALB/cAnNHsd mice treated with 10 mg/kg RBD-specific (A), NTD-specific (B), S2-specific (C), or unknown-epitope-specific (D) SARS-CoV-2 mAbs prior to lethal challenge with a 5×LD₅₀ dose of mouse-adapted SARS-CoV-2. Mice were infected and their body weight monitored for 8 days. Each point represents the maximum weight lost for each mouse (mean ± SD). For all groups n = 5, except PVI.V3–17, PVI.V3–18, PVI.V6–6, PVI.V6–7, PVI.V6–8, PVI.V6–12, and PVI.V6–15 where n = 4, and the negative control group where n = 9. Survival data are shown as Kaplan Meier curves to the right. The negative control group is shared and is displayed in all panels. (E) Golden Syrian hamster weight loss following treatment with the six most protective mAbs and challenge with 10⁵ PFU SARS-CoV-2 WA1/2020 (n = 4, mean ± SD). Neutralizing mAbs are shown as crosses, while non-neutralizing mAbs are shown as circles. p values are shown only for statistically significant comparisons as determined by a Kruskal-Wallis with Dunn’s multiple comparisons test.

Figure 3.. Mouse lung titers following mAb treatment and SARS-CoV-2 challenge

(A–D) Murine lung titers (n = 4, mean ± SD) at 3 dpi after treatment with 10 mg/kg RBD-specific (A), NTD-specific (B), S2-specific (C), or unknown-epitope-specific (D) SARS-CoV-2 mAbs and challenge with a 1×LD₅₀ dose of mouse-adapted SARS-CoV-2. (E–H) lung titers at 5 dpi (n = 4, mean ± SD) after treatment with RBD-specific (E), NTD-specific (F), S2-specific (G), and unknown-epitope-specific (H) SARS-CoV-2 mAbs and challenge with mouse-adapted SARS-CoV-2. Neutralizing mAbs are shown as crosses, while non-neutralizing mAbs are shown as circles. p values are shown only for groups with titers statistically significant compared to the negative control, as determined by an ordinary one-way ANOVA with Dunnett’s multiple comparisons test with a single pooled variance. The control groups are shared and are displayed in all panels.

Figure 4.. Cryo-EM structures of PVI.V6–12, PVI.V3–21, and PVI.V5–4 in complex with SARS-CoV-2 spike

(A and B) Global, modest-resolution cryo-EM maps of PVI.V6–12 and PVI.V3–21 Fabs in complex with spike. The structures identify the binding epitopes on the spike RBD. Fabs are colored green, RBD sienna, and the rest of the spike blue. (C) Cryo-EM map of PVI.V5–4 in complex with spike at 3.67 Å nominal resolution. The structure defines the PVI.V5–4 epitope on the subdomain 1 (SD1). Spike S1 including RBD, SD1, and SD2 domains are colored sienna, S2 dark gray, and PVI.5–4 variable heavy and light chains blue and green, respectively. (D) Cartoon representation of the atomic model of PVI.V5–4 in complex with spike SD1. Major interactions are defined by the antibody HCDR3, HCDR2, and LCDR3 loops. The N318 glycan is shown for context. (E) Surface representation of PVI.V5–4 epitope on spike SD1. The Fab interacting loops are shown as cartoon, and individual interacting residues are labeled and shown as sticks. The spike interacting residues are underlined. The N318 glycan is shown for context.

Figure 5.. Characterization of the Fc-mediated activity of the mAbs

(A) FcγRIIIa reporter activity of the RBD-, NTD-, S2-, and unknown-epitope-binding mAbs. (B) FcγRIIa reporter activity of the RBD-, NTD-, S2-, and unknown-epitope-binding mAbs. (C) Binding of complement protein C1q to the Fc portion of the mAbs in the presence of wild-type SARS-CoV-2 spike protein. Dotted lines denote the limit of detection. Each assay was performed at least twice, and the geometric mean and geometric standard deviation of these replicates is shown. The control groups for each assay are shared and are displayed in all panels belonging to the respective assay.

Figure 6.. Investigating the protection elicited by Fc-silent mutant mAbs

(A) FcγRIIIa reporter activity of the seven most protective mAbs and their LALA and LALAPG mutant counterparts. (B) FcγRIIa reporter activity of the wild-type, LALA, and LALAPG mAbs. (C) Binding of complement protein C1q by the wild-type, LALA, and LALAPG mAbs in the presence of wild-type SARS-CoV-2 spike protein. In (A)–(C), dotted lines denote the limit of detection. Each assay was performed twice, and the geometric mean and geometric standard deviation of these replicates is shown. (D) Weight loss and survival of BALB/cAnNHsd mice (n = 4, mean ± SD) treated intraperitoneally with a 10 mg/kg dose of the seven most protective mAbs and their LALA and LALAPG counterparts following lethal challenge with a 5×LD₅₀ dose of mouse-adapted SARS-CoV-2. The negative control was shared across experimental groups and is shown in each respective panel.