SARS-CoV-2 Monoclonal Antibody Treatment Followed by Vaccination Shifts Human Memory B-Cell Epitope Recognition, Suggesting Antibody Feedback

Abstract

Therapeutic monoclonal antibodies (mAbs) have been studied in humans, but the impact on immune memory of mAb treatment during an ongoing infection remains unclear. We evaluated the effect of infusion of the anti-severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spike receptor-binding domain (RBD) mAb bamlanivimab on memory B cells (MBCs) in SARS-CoV-2-infected individuals. Bamlanivimab treatment skewed the repertoire of MBCs targeting spike toward non-RBD epitopes. Furthermore, the relative affinity of RBD MBCs was weaker in mAb-treated individuals compared to placebo-treated individuals over time. Subsequently, after mRNA coronavirus disease 2019 vaccination, MBC differences persisted and mapped to a specific reduction in recognition of the class II RBD site, the same RBD epitope recognized by bamlanivimab. These findings indicate a substantial role of antibody feedback in regulating MBC responses to infection, and single mAb administration can continue to impact MBC responses to additional antigen exposures months later.

Keywords: COVID-19; SARS-CoV-2; bamlanivimab; mAb therapy; memory B cells.

Figure 1.

Effect of bamlanivimab after 28 days of treatment in coronavirus disease 2019 (COVID-19) cases. A, Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)–infected people were treated with 700 mg intravenous dose of bamlanivimab. Created with BioRender. B and C, Spike S1 (B) and receptor-binding domain (RBD) (C) immunoglobulin G (IgG) titers in the placebo and monoclonal antibody (mAb)–treated groups. D, Representative gating strategy for spike-binding memory B cells (MBCs) in SARS-CoV-2–infected (top image) and uninfected (bottom image) humans (refer also to Supplementary Figure 8). E, Representative gating strategy for RBD-binding MBCs in SARS-CoV-2–infected (top image) and uninfected (bottom image) humans. F and G, Frequency of spike (F) and RBD (G) MBCs among total MBCs (IgD⁻, CD27⁺). H, Frequency of RBD MBCs within the spike MBC population. I, Frequency of non-RBD MBCs within the spike MBC population. A different gating strategy was applied compared to (H), and these data are not reciprocal to the findings shown in (H). J, Affinity of RBD MBCs determined by the ratio mean fluorescence intensity (MFI) from the RBD probe (PE-CY7)/MFI from immunoglobulin-associated beta (CD79b). K–M, Frequency of RBD MBCs expressing IgG (K), immunoglobulin M (IgM) (L) or immunoglobulin A (IgA) (M). Medians are shown. An unpaired, nonparametric Mann-Whitney test was used to compare the different groups. P values are shown in the graphs and considered significant if <.05. Sample size: treatment group, n = 25; placebo group, n = 21. Image in A was made using BioRender.

Figure 2.

Effect of bamlanivimab after 24 weeks of treatment in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)–infected adults. A and B, Frequency of spike (A) and receptor-binding domain (RBD) (B) memory B cells (MBCs) among total MBCs (IgD⁻, CD27⁺). C, Frequency of RBD MBCs within the spike MBC population. D, Frequency of non-RBD MBCs within the spike MBC population. E, Affinity of RBD MBCs determined by the ratio mean fluorescence intensity (MFI) from the probe PE-CY7 divided by the MFI from immunoglobulin-associated beta (CD79b). F–H, Frequency of RBD MBCs expressing immunoglobulin G (IgG) (F), immunoglobulin M (IgM) (G), or immunoglobulin A (IgA) (H). Medians are plotted in the graphs. An unpaired, nonparametric Mann-Whitney test was used to compare the different groups. P values are shown in the graphs and considered significant if <.05. Sample size: treatment group, n = 18; placebo group, n = 11.

Figure 3.

Bamlanivimab treatment followed by messenger RNA (mRNA) vaccination decreases the frequency of receptor-binding domain (RBD) memory B cells (MBCs) in the total MBC population and reduces the affinity of RBD and spike MBCs. A, Scheme of study design. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)–infected humans receiving bamlanivimab were immunized with mRNA vaccines between 20 and 126 days after monoclonal antibody (mAb) treatment. Created with biorender. B and C, Frequency of spike (B) and RBD (C) MBCs among total MBC population. D and E, Frequency of RBD (D) and non-RBD (E) MBCs among the spike MBC population. A different gating strategy was applied for (D) vs (E), and the data in (D) are not reciprocal to the findings shown in (E). F, Affinity of RBD MBCs determined by the ratio mean fluorescence intensity (MFI) from the RBD PE-CY7/MFI from immunoglobulin-associated beta (CD79b). G–I, Frequency of immunoglobulin G (IgG) (G), immunoglobulin M (IgM) (H), and immunoglobulin A (IgA) (I) RBD MBCs. Medians are plotted in the graphs. An unpaired, nonparametric Mann-Whitney test was used to compare the different groups. P values are shown in the graphs and considered significant if <.05. Sample size: treatment group, n = 9; placebo group, n = 15. Image in A was made using BioRender.

Figure 4.

Bamlanivimab treatment followed by messenger RNA vaccination promotes escape of class II receptor-binding domain (RBD) memory B cells (MBCs). A, The most common mutations associated with class I–IV monoclonal antibody escape are mapped onto the RBD structure (PDB ID 7KMG) and represent the 4 RBD constructs used in this study. B, LY-CoV555 (bamlanivimab; PDB ID 7KMG) epitope on RBD (purple) is shown in the context of the mutations described in (A), with identical color coding. C, Graphs show the frequency of B cells binding to RBD knockout (KO) constructs 24 weeks posttreatment in unvaccinated individuals. D, Graphs show the frequency of B cells binding to RBD KO constructs 24 weeks posttreatment in unvaccinated individuals. Medians are plotted in the graphs. An unpaired, nonparametric Mann-Whitney test was used to compare the different groups. P values are shown in the graphs and considered significant if <.05. Sample size: vaccinated (treatment group, n = 14; placebo group, n = 14); unvaccinated (treatment group, n = 14; placebo group, n = 14).