Isolation of a panel of ultra-potent human antibodies neutralizing SARS-CoV-2 and viral variants of concern

Abstract

In the absence of virus-targeting small-molecule drugs approved for the treatment and prevention of COVID-19, broadening the repertoire of potent SARS-CoV-2-neutralizing antibodies represents an important area of research in response to the ongoing pandemic. Systematic analysis of such antibodies and their combinations can be particularly instrumental for identification of candidates that may prove resistant to the emerging viral escape variants. Here, we isolated a panel of 23 RBD-specific human monoclonal antibodies from the B cells of convalescent patients. A surprisingly large proportion of such antibodies displayed potent virus-neutralizing activity both in vitro and in vivo. Four of the isolated nAbs can be categorized as ultrapotent with an apparent IC₁₀₀ below 16 ng/mL. We show that individual nAbs as well as dual combinations thereof retain activity against currently circulating SARS-CoV-2 variants of concern (such as B.1.1.7, B.1.351, B.1.617, and C.37), as well as against other viral variants. When used as a prophylactics or therapeutics, these nAbs could potently suppress viral replication and prevent lung pathology in SARS-CoV-2-infected hamsters. Our data contribute to the rational development of oligoclonal therapeutic nAb cocktails mitigating the risk of SARS-CoV-2 escape.

Fig. 1. Donor selection and sorting of RBD-specific B lymphocytes from the blood of convalescent COVID-19 patients.

a, b Sera of four select donors were characterized by ELISA for anti-RBD binding potency (a) and pseudovirus neutralization activity (b). K- stands for a healthy donor. c Short medical information for these donors. d Gating strategy used for selection of individual RBD+IgG+ B lymphocytes, shown are the representative FACS plots for donor C34.

Fig. 2. Comparative characteristics of the SARS-CoV-2 RBD-binding human monoclonal antibodies.

Heat maps are shown for monoclonal antibody blocking of ACE2/RBD interaction (%), staining of cell-surface expressed SARS-CoV-2 Spike (%), neutralization of Spike-pseudotyped lentiviral particles (IC₅₀) or authentic SARS-CoV-2 virus (IC₁₀₀), KD for the interaction with recombinant RBD (BLI measurements), autoreactivity (fold MFI increase above background), and SHM rate (number of nucleotide substitutions). Antibodies are grouped and ordered by neutralization potency and were either obtained in this study (iB0-26) or reported earlier (10933, REGN10933; 10987, REGN10987; 2504, COV2-2504; 2015, COVA2-15)^,,. All numerical values can be found in Supplementary Table S2.

Fig. 3. In vitro virus neutralization by the SARS-CoV-2 nAbs.

Pseudovirus (a) or live (b, c) SARS-CoV-2 virus were used. nAbs were either published before (dark blue columns) or obtained in this study (cyan). iB19* neutralization was incomplete (gray). glVRC01 is an irrelevant HIV-specific antibody. 10933, REGN10933; 10987, REGN10987; 2-15, COVA2-15; 2504, COV2-2504. Antibody pair (iB6 + iB20) displaying synergistic activity is marked by a red frame (c).

Fig. 4. Epitope binning of SARS-CoV-2 RBD-specific nAbs.

The analysis includes either previously published (REGN10987, REGN10933, COVA2-15) or iB-series of nAbs. a BLI analysis of competition between nAbs iB20, REGN10987, and iB6 for RBD binding. b Venn diagram showing mutual overlap between the tentative epitopes of nAbs forming eight distinct bins. The diagram was plotted based on the BLI data (Supplementary Figs. S6 and S7). c Positions of nine mutations escaping neutralization by at least one of the tested antibodies. Residues mutated are highlighted red or pink on the 3D model of SARS-CoV-2 RBD. d Tentative positions of nAb footprints are circled.

Fig. 5. Neutralization profiles for the iB-series of nAbs as well as for some of the previously reported reference nAbs tested against WT and selected amino acid residue mutant pseudoviruses (positions 406–501 aa).

a Heatmap displaying nAb IC₅₀ fold change, when tested against a single amino acid mutant virus relative to WT. Asterisks denote mutations that are absent from the currently circulating viral isolates and which have been shown to confer resistance to some of the clinical-stage nAbs. b Heatmap displaying nAb sensitivity to double and triple mutations in the RBD, which correspond to the substitutions found in several currently circulating VOCs.

Fig. 6. Efficacy of iB12, iB14, and iB6 + iB20 cocktail in treatment and prophylaxis of SARS-CoV-2 infection in a hamster model.

In a prophylactic regimen, antibodies were administered i/p 1 day before infection at a dose of 1 or 0.1 mg/animal (i.e., ~10 or 1 mg/kg weight). Control animals received 1 mg of total human IgG. In the therapeutic regimen, iB12 antibody (1 mg/animal) was given 6 h following the viral challenge (Experiment 1), whereas iB14 antibody was injected either 6 h (0.1 or 1 mg/animal) or 24 h (1 mg/animal) following the viral challenge (Experiment 2). Intact animals formed a separate control group. Animals were sacrificed on day 5 following infection. Lungs were weighted and used for RT-qPCR quantification of the viral load. a Relative weight dynamics in the prophylaxis and treatment groups following administration of iB nAbs or control total human IgG. Error bars represent means ± SD. b Impact of iB nAbs on the levels of viral transcripts in hamster lungs (RT-qPCR with RdRp primers). RT-qPCR with E gene-specific primers produced very similar results (not shown). Animals with strongly reduced human nAb titers in the blood were considered as outliers due to the technical error of nAb injection and were excluded from the analysis. Error bars represent means ± SEM. c Lung pathology scores for hamster lungs at 5 dpi (Supplementary Fig. S9).