Discovery and characterization of SARS-CoV-2 reactive and neutralizing antibodies from humanized CAMouseHG mice through rapid hybridoma screening and high-throughput single-cell V(D)J sequencing

Abstract

The coronavirus disease 2019 pandemic has caused more than 532 million infections and 6.3 million deaths to date. The reactive and neutralizing fully human antibodies of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) are effective detection tools and therapeutic measures. During SARS-CoV-2 infection, a large number of SARS-CoV-2 reactive and neutralizing antibodies will be produced. Most SARS-CoV-2 reactive and neutralizing fully human antibodies are isolated from human and frequently encoded by convergent heavy-chain variable genes. However, SARS-CoV-2 viruses can mutate rapidly during replication and the resistant variants of neutralizing antibodies easily survive and evade the immune response, especially in the face of such focused antibody responses in humans. Therefore, additional tools are needed to develop different kinds of fully human antibodies to compensate for current deficiency. In this study, we utilized antibody humanized CAMouse^HG mice to develop a rapid antibody discovery method and examine the antibody repertoire of SARS-CoV-2 RBD-reactive hybridoma cells derived from CAMouse^HG mice by using high-throughput single-cell V(D)J sequencing analysis. CAMouse^HG mice were immunized by 28-day rapid immunization method. After electrofusion and semi-solid medium screening on day 12 post-electrofusion, 171 hybridoma clones were generated based on the results of SARS-CoV-2 RBD binding activity assay. A rather obvious preferential usage of IGHV6-1 family was found in these hybridoma clones derived from CAMouse^HG mice, which was significantly different from the antibodies found in patients with COVID-19. After further virus neutralization screening and antibody competition assays, we generated a noncompeting two-antibody cocktail, which showed a potent prophylactic protective efficacy against SARS-CoV-2 in cynomolgus macaques. These results indicate that humanized CAMouse^HG mice not only provide a valuable platform to obtain fully human reactive and neutralizing antibodies but also have a different antibody repertoire from humans. Thus, humanized CAMouse^HG mice can be used as a good complementary tool in discovery of fully human therapeutic and diagnostic antibodies.

Keywords: SARS-CoV-2; antibody humanized mice; fully human antibody; reactive and neutralizing antibodies; single-cell sequencing.

Figure 1

Integrated workflow for rapid isolation of anti-SARS-CoV-2 RBD antibodies from CAMouse^HG mice.

Figure 2

Antibody repertoire features of RBD-reactive hybridoma cells derived from CAMouse^HG mice. (A) Isotype usage in single-cell repertoire sequencing data of hybridoma clones specifically recognized SARS-COV-2 RBD. (B) CDR3 AA lengths of heavy and light chains. (C) Human V and J genes of heavy- and light-chain utilization. Circos plots comparing VJ gene association of heavy (D) and light chains (E). VJ combinations are linked based on their relative frequency. (F) Paired antibody repertoire. Circle size and color correspond to the number of heavy and light chains present. (G) Amino acid sequence alignment of the five SARS-COV-2 neutralizing mAbs isolated from CAMouse^HG mice.

Figure 3

Characterization of mAbs against SAR-CoV-2 RBD. (A) Antibody binding competition assay. Immobilized first mAb (30 μg/mL) was mixed with SAR-CoV-2 RBD (51.5 μg/mL), and secondary mAb (30 μg/mL) was used to compete for binding. (B) Grouping of mAbs. Group 1, MAb 18-4A; group 2, the other mAbs. Structural analyses of SARS-CoV-2 spike and Fab complex. (C) Structure of SARS-CoV-2 spike (PDB code: 7TPH) in complex with 11-2G-Fab. (D) Structure of SARS-CoV-2 spike (PDB code: 7TPH) in complex with 18-4A-Fab. The different subunits of the spike protein are colored green, orange, purple; two Fab domains of antibodies are colored blue and red, respectively. Binding details between and SARS-CoV-2 spike and 11-2G-Fab (E) or 18-4A-Fab (F) are presented with amino acids from SARS-CoV-2 spike, which forms hydrophobic or hydrogen bond interactions with amino acids from 11-2G-Fab. (G, H) Analysis of affinity of 11-2G and 18-4A for the SARS-CoV RBD protein. (I) Neutralizing activity of 11-2G and 18-4A against SARS-CoV-2 pseudoviruses. (J) Neutralizing activity of 11-2G/18-4A cocktail (1:1) against SARS-CoV-2 pseudoviruses. (K) Neutralizing activity of 11-2G and 18-4A against authentic SARS-CoV-2. (L) Neutralizing activity of 11-2G/18-4A cocktail (1:1) against authentic SARS-CoV-2.

Figure 4

The 11-2G/18-4A cocktail showed potent prophylactic efficacy in SARS-CoV-2-infected cynomolgus macaques. (A) Schematic diagram of antibody treatment and SARS-CoV-2 challenge in cynomolgus macaques. (B) Nasal, throat and anal swabs were collected at 2, 4, 6 days post-infection. (C) Trachea and (D) Lungs (both left lobe and right lobe) were obtained at 7 days post-infection. Viral load was assessed by detection of SARS-CoV-2 genomic RNA and subgenomic RNA using qRT-PCR. Data are shown as mean ± SD. An unpaired Student’s t-test (two-tailed) was used for statistical analysis, and the relevant p-values are indicated (not indicated in graph; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). (E) H&E staining of lung tissues. ns, not significant.