BNT162b2 induced neutralizing and non-neutralizing antibody functions against SARSCoV-2 diminish with age

Abstract

Each novel SARS-CoV-2 variant renews concerns about decreased vaccine efficacy caused by evasion of vaccine induced neutralizing antibodies. However, accumulating epidemiological data show that while vaccine prevention of infection varies, protection from severe disease and death remains high. Thus, immune responses beyond neutralization could contribute to vaccine efficacy. Polyclonal antibodies function through their Fab domains that neutralize virus directly, and Fc domains that induce non-neutralizing host responses via engagement of Fc receptors on immune cells. To understand how vaccine induced neutralizing and non-neutralizing activities synergize to promote protection, we leverage sera from 51 SARS-CoV-2 uninfected health-care workers after two doses of the BNT162b2 mRNA vaccine. We show that BNT162b2 elicits antibodies that neutralize clinical isolates of wildtype and five variants of SARS-CoV-2, including Omicron BA.2, and, critically, induce Fc effector functions. FcγRIIIa/CD16 activity is linked to neutralizing activity and associated with post-translational afucosylation and sialylation of vaccine specific antibodies. Further, neutralizing and non-neutralizing functions diminish with age, with limited polyfunctional breadth, magnitude and coordination observed in those ≥65 years old compared to <65. Thus, studying Fc functions in addition to Fab mediated neutralization provides greater insight into vaccine efficacy for vulnerable populations such as the elderly against SARS-CoV-2 and novel variants.

Figure 1:

BNT162b2 induced IgG mediates age-dependent neutralization of wildtype and SARS-CoV-2 clinical variants. (A) Live SARS-CoV-2 neutralization (FRNT50) wildtype WA1 and receptor binding domain (RBD) specific IgG/M/A, IgG and IgA EC50 values (Supplemental Figure 1A) are plotted with relationship assessed by linear regression. (B) Neutralization of live SARS-CoV-2 wildtype WA1 and variants (Supplemental Figure 1B) are depicted with each dotted line representing a single individual and statistical significance calculated by Wilcoxon matched pair signed rank test. Live SARS-CoV-2 neutralization (FRNT50) for (C) wildtype (WT) and variants (D) B.1.1.7 (Alpha), (E) B.1.351 (Beta), (F) P.1 (Gamma), (G) B.1.617.2 (Delta) and (H) BA.2 (Omicron) and age in years are plotted with relationship assessed by linear regression and p values adjusted for sex.

Figure 2:

Vaccine specific IgG induction of FcγRIIIa/CD16 effector functions correlate with neutralization of wildtype and SARS-CoV-2 clinical variants. Relationships between live SARS-CoV-2 WA1 neutralization (FRNT50) and receptor binding domain (RBD) specific relative binding to (A) FcγRIIIa/CD16a, (B) FcγRIIa/CD32a and (C) FcγRIIb/CD32b, RBD specific antibody dependent natural killer cell activation (ADNKA) determined by (D) CD107a expression, (E) IFNγ production and (F) TNFα secretion, and (G) RBD specific antibody dependent neutrophil phagocytosis (ADNP) are shown. (H) Heatmap summarizes Spearman correlations (Supplemental Figure 2) between neutralization of SARS-CoV-2 wildtype WA1 and variants with relative binding of RBD specific IgG to activating (FcγRIIIa/CD16a and FcγRIIa/CD32a), inhibitory (FcγRIIb/CD32b) and ratios of activating:inhibitory FcγR (FcγRIIIa/CD16a:FcγRIIb/CD32b and FcγRIIa/CD32a:FcγRIIb/CD32b) binding and Fc effector functions ADNKA, antibody dependent cellular phagocytosis (ADCP), ADNP and antibody dependent complement deposition (ADCD). * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001.

Figure 3:

Differential fucose and sialic acid on vaccine specific IgG link to FcγRIIIa/CD16a effector functions. Stacked column graphs depict the relative abundance of individual glycoforms (Supplemental Figure 3A–B) with respect to (A) total bulk non-antigen specific and (B) receptor binding domain (RBD) specific IgG. Each column represents one individual study participant. Dot plots summarize differences between bulk non-antigen specific and RBD specific IgG in the collective relative abundance of all individual glycoforms (Supplemental Figure 3C and D) containing (C) fucose, (D) total sialic acid, (E) di-sialic acid, (F) mono-sialic acid and (G) di-galactose with statistical significance calculated by Wilcoxon matched-pairs signed rank test. Data for (H) asialylated fucosylated, (I) total sialic and (J) total di-sialic acid, the three RBD specific glycoforms that have a statistically significant relationship across all markers of ADNKA activation, are plotted with CD107a expression per RBD specific IgG, as well as IFNγ and TNFα (Supplemental Figure 4). For comparison, data for (K) total mono-sialic acid is plotted.

Figure 4:

Age influences some but not all vaccine specific antibody FcγR functions. The relationships between relative binding of receptor binding domain (RBD) specific IgG to (A) FcγRIIIa/CD16a, (B) FcγRIIa/CD32a and (C) inhibitory FcγRIIb/CD32b and age are shown. (D) Heatmap of the coefficient of determination (r²) summarizes the goodness of fit across RBD and control respiratory syncytial virus (RSV), influenza (Flu) and anthrax antigens in FcγR binding and age. * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; N/A not available given absence of significant detectable levels. The relationship between age and RBD antibody dependent natural killer cell activation (ADNKA) as measured by (E) CD107a and (F) TNFα, and (G) RBD antibody dependent cellular phagocytosis (ADCP) and (H) RBD antibody dependent neutrophil phagocytosis (ADNP) are shown. Linear regression with p values adjusted for sex are reported. (I) Radar plots depict vaccine specific IgG glycoforms calculated from the Z scored data for each individual RBD specific IgG glycoforms relative to bulk non-antigen specific IgG glycoforms with lines representing the median for each age group.

Figure 5:

Enhanced BNT162b2 induced polyfunctional antibody breadth and magnitude against SARS-CoV-2 in younger compared to older adults. For each individual, the neutralization breadth across variants (Supplemental Figure 6A) and cumulative vaccine specific Fc functional magnitude from the sum of the Z scores for each of the individual effector functions (Supplemental Figure 6B) were calculated. (A) Grouped by neutralization breadth (top), each column shows the cumulative Fc functional score for one individual. Median, minimum and maximum ages characterizing each neutralization breadth group are shown (bottom). Polyfunctional antibody breadth was calculated for each individual (Supplemental Figure 6C) and used to categorize individuals into high (90–100%), medium (80–90%) or low (<80%) responders. The proportions of high, median and low responders are grouped by (B) age and (C) sex. (D) Radar plots depict vaccine specific polyfunctional antibody magnitude calculated from the Z scored data for each antibody function (Supplemental Figure 6C) with lines representing the median for each age group. (E) Heatmap summarizes Spearman correlations (Supplemental Figure 2) between neutralization of SARS-CoV-2 wildtype WA1 and variants with RBD specific IgG/M/A, IgG and IgA levels, relative binding of RBD specific IgG to activating (FcγRIIIa/CD16a and FcγRIIa/CD32a), inhibitory (FcγRIIb/CD32b) and ratios of activating:inhibitory FcγR (FcγRIIIa/CD16a:FcγRIIb/CD32b and FcγRIIa/CD32a:FcγRIIb/CD32b) binding and Fc effector functions antibody dependent natural killer cell activation (ADNKA), antibody dependent cellular phagocytosis (ADCP), antibody dependent neutrophil phagocytosis (ADNP) and antibody dependent complement deposition (ADCD) for each age group. * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001.