Potent SARS-CoV-2 neutralizing antibodies with therapeutic effects in two animal models

Abstract

The use of therapeutic neutralizing antibodies against SARS-CoV-2 infection has been highly effective. However, there remain few practical antibodies against viruses that are acquiring mutations. In this study, we created 494 monoclonal antibodies from patients with COVID-19-convalescent, and identified antibodies that exhibited the comparable neutralizing ability to clinically used antibodies in the neutralization assay using pseudovirus and authentic virus including variants of concerns. These antibodies have different profiles against various mutations, which were confirmed by cell-based assay and cryo-electron microscopy. To prevent antibody-dependent enhancement, N297A modification was introduced. Our antibodies showed a reduction of lung viral RNAs by therapeutic administration in a hamster model. In addition, an antibody cocktail consisting of three antibodies was also administered therapeutically to a macaque model, which resulted in reduced viral titers of swabs and lungs and reduced lung tissue damage scores. These results showed that our antibodies have sufficient antiviral activity as therapeutic candidates.

Keywords: Unology; immune response; virology.

Graphical abstract

Figure 1

Patient selection and cell sorting Serum neutralization titers of 47 patients with COVID-19 convalescent were measured by (A) cell-based Spike-ACE2 inhibition assay. The binding quantity of soluble ACE2 to Spike-expressing cells without serum/antibody is defined as 100%, and the binding quantities of soluble ACE2 to Spike-expressing cells after incubation with serum/antibody are calculated as the ACE2-binding rate. (B) The neutralization ability of patient serum for each severity is shown. Samples used for antibody production are labeled with their ID. (C) The sorting strategy is shown. CD19⁺ cells were size gated, and CD19⁺CD27⁺IgD⁻ cells were selected. Antigen-specific memory B cells (red) and antigen-nonspecific plasma cells (blue) were single-cell sorted. Asy, asymptomatic; Mil, mild; Mod, moderate; Sev, severe; Cri, critical.

Figure 2

Screening of neutralizing antibodies (A) The binding ability of the antibodies against Spike-expressing cells, as indicated by mean fluorescence intensity (MFI), and the inhibiting ability of ACE2 to bind to Spike-expressing cells, as indicated by ACE2-binding rate, are shown. Antibodies that have binding ability without neutralization ability are encircled by a dotted line, and antibodies that have a correlation between binding and neutralization ability are encircled by a solid line. Color indicates the individual and shape indicates the cell type from which the antibodies were derived. Spearman’s rank correlation coefficient was calculated from all samples. (B) The percentage of antibodies having high binding ability to Spike and high ability to inhibit ACE2 binding to Spike is shown by the source cell. Chi-squared test. (C) The neutralization ability of recombinant monoclonal antibody was measured by fusion inhibition assay. The quantity of cell fusion without the antibody is defined as 100%, and the quantity of cell fusion after incubation with antibodies is shown as the fusion rate. The correlation between the ACE2-binding rate and the fusion rate for antibodies with the ACE2-binding rate of 50% or less is shown. Spearman’s rank correlation coefficient. (D) The correlation between the end-point micro-neutralization titer and the ACE2-binding rate is shown. Spearman’s rank correlation coefficient.

Figure 3

Effect of point mutation of Spike protein on antibody neutralizing ability The ACE2-binding rate (%) of recombinant monoclonal antibodies against Spike protein with various point mutations (A) within RBD, (B) outside RBD, and (C) against Spike proteins of VOCs and SARS-CoV-1 were measured by the cell-based Spike-ACE2 inhibition assay. The variant column indicates that the mutation is contained in the VOC and Variants of interest. The numbers indicate the binding quantities of soluble ACE2 to Spike-expressing cells after incubation with antibodies. The color indicates the grade of neutralization ability.

Figure 4

Pseudovirus and authentic virus neutralization assay The IC₅₀ (ng/mL) of selected antibodies was measured using pseudovirus that expresses Spike of the Wuhan strain or Alpha, Beta, Gamma, or Delta variants. (A) The neutralization curve and (B) IC₅₀ values are shown. The color indicates the neutralizing ability. The data are represented as mean ± SD. The IC₅₀ (ng/mL) of the selected antibodies was examined using authentic virus of the Wuhan strain or Alpha, Beta, Gamma, Delta, Kappa, Omicron (BA.1 and BA.2) variants. (C) Neutralization curve and (D) IC₅₀ values are shown. The data are represented as mean ± SD. ∗Measured separately.

Figure 5

Cryo-EM structure of neutralizing antibodies (A) The structures of RBD and Ab159, Ab188, Ab326, Ab354, Ab445, and Ab496 are shown. Only the variable domains of antibodies are modeled and drawn as a cartoon tube (individual color) on the RBD surface (gray), and the epitope of each antibody is colored the same as each antibody. The red area in the central RBD is the binding residue of ACE2 (PDB: 7A94), showing the relationship between the binding sites of the antibodies, which are roughly divided into three groups. The positions of key amino acids are indicated by black arrows. (B) The residues 400-506 of Spike are shown. The epitopes revealed by cryo-EM are colored in red, and the residues affected by the mutation described in Figure 3A are shown in squares.

Figure 6

Fc-engineering for the prevention of antibody-dependent enhancement Pseudovirus (Wuhan strain) was incubated with serially diluted antibody with or without N297A modification, and the mixture was applied to Raji cells. After incubation for 3 days at 37°C, the cells were lysed and subjected to luciferase assay. The lines represent the mean value.

Figure 7

Therapeutic efficacy of neutralizing antibodies in two animal models (A) Overview of the experiment with the Syrian hamster model. Hamsters were inoculated intranasally with 10³ PFU of NCGM02. On day 1 post-infection, hamsters were injected intraperitoneally with 50 mg/kg BW of neutralizing antibodies or human IgG1 as a control. (B) On day 4 post-infection, serum and lung samples were collected. Serum-neutralizing titer and viral RNA levels in the lung tissues were measured. Color indicates serum neutralization titer. ∗p < 0.05 by Dunnett’s test using all samples. (C) Overview of the experiment using the cynomolgus macaque model. Six cynomolgus macaques were inoculated with 2 × 10⁷ TCID₅₀ SARS-CoV-2 JP/TY/WK-521/2020 into the conjunctiva, nasal cavity, oral cavity, and trachea on day 0. On day 1, 20 mg (5-7 mg/kg) of antibody cocktail 3Mix (⅓ each of an N297A-modified Ab326, Ab354, and Ab496) or human IgG1 as a control were injected intravenously. Nasal swabs and peripheral blood were collected on days 1, 3, 5, and 7, and lung tissues were collected on day 7. (D) Viral titers of nasal swabs were measured. (E) Plasma neutralization titers were measured. The numbers of plaques (%) after antibody administration (day 3) compared to before antibody administration (day 1) are shown. (F) Viral titers of each lung lobe were measured. “<” indicates under the detection limit. The shaded cells indicate that the lung lobe was not present. ^aCPE-positive wells were 2/4 at no dilution, and 2/4 at 10¹ dilution. ^bCPE-positive wells were 1/4 at no dilution. ^cCPE-positive wells were 1/4 at 10¹ dilution. RU; right upper lobe, RM; right middle lobe, RL; right lower lobe, LU; left upper lobe, LM; left middle lobe, LL; left lower lobe. (G and H) Representative H&E stained sections and (I and J) immunohistochemical staining of SARS-CoV-2 N protein. (G and I) Lung sections of 1925F were treated with the control antibody. (H and J) Lung sections of 1938F were treated with the mixed SARS-CoV-2 specific neutralizing antibodies. The bars indicate 100 μm. (K) Histological damage score was evaluated according to the previously published criteria. The bars indicate mean. ∗p < 0.05 by Student’s t test.