Convalescent plasma anti-SARS-CoV-2 spike protein ectodomain and receptor-binding domain IgG correlate with virus neutralization

Abstract

The newly emerged severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) highlights the urgent need for assays that detect protective levels of neutralizing antibodies. We studied the relationship among anti-spike ectodomain (anti-ECD), anti-receptor-binding domain (anti-RBD) IgG titers, and SARS-CoV-2 virus neutralization (VN) titers generated by 2 in vitro assays using convalescent plasma samples from 68 patients with COVID-19. We report a strong positive correlation between both plasma anti-RBD and anti-ECD IgG titers and in vitro VN titers. The probability of a VN titer of ≥160, the FDA-recommended level for convalescent plasma used for COVID-19 treatment, was ≥80% when anti-RBD or anti-ECD titers were ≥1:1350. Of all donors, 37% lacked VN titers of ≥160. Dyspnea, hospitalization, and disease severity were significantly associated with higher VN titer. Frequent donation of convalescent plasma did not significantly decrease VN or IgG titers. Analysis of 2814 asymptomatic adults found 73 individuals with anti-ECD IgG titers of ≥1:50 and strong positive correlation with anti-RBD and VN titers. Fourteen of these individuals had VN titers of ≥1:160, and all of them had anti-RBD titers of ≥1:1350. We conclude that anti-RBD or anti-ECD IgG titers can serve as a surrogate for VN titers to identify suitable plasma donors. Plasma anti-RBD or anti-ECD titers of ≥1:1350 may provide critical information about protection against COVID-19 disease.

Keywords: Adaptive immunity; Cellular immune response; Immunology; Infectious disease; Molecular pathology.

Figure 1. Patterns of VN and ELISA titers.

(A) Violin plot of distribution of VN titers at initial donation. The number of donor cases above (n = 43) and below (n = 25) the VN ≥160 cut-off value is reported. (B) Violin plots showing similar patterns of distribution of titers at initial donation for the 2 VN assays, together with the reciprocal ELISA IgG titers for plasma anti-ECD protein (ECD) and anti-RBD IgG (RBD). (C) Pair-wise Pearson correlations showing the correlation coefficient (r) and related significance value (***P < 0.001) above the diagonal as well as the bivariate scatter plots (jittered points represented as black dots) with linear regression fit (red line), CIs (gray shading), correlation value (red points), and correlation ellipse (black ellipses) below the diagonal. The density plot (black line) and histogram of each variable is reported along the diagonal. Data are presented in log₂ scale of reciprocal titers for VN, anti-ECD IgG, and anti-RBD IgG and in IC₅₀ units for VN2. The sample sizes for which the correlation coefficients were derived are as follows: VN-VN2, 86; VN-ECD, 91; VN-RBD, 91; VN2-ECD, 84; VN2-RBD, 84; ECD-RBD, 91.

Figure 2. Prevalence of donors with VN >160 for VN2, ECD, and RBD.

Probabilities of VN ≥160 were plotted for 6 range classes, with an interclass interval of 1.8 log₂ IC₅₀ values (class 1, <2; class 2, 2–12; class 3, 12–42; class 4, 42–147; class 5, 147–512; and class 6, >512) or observed classes for ECD (n = 6) and RBD (n = 8) reciprocal ELISA titers. A spline curve (dotted line, smoothness shape = 1) has been fitted to the probability values and standard errors (bars) are reported. The numbers of donor samples are shown above the bars.

Figure 3. Distributions of VN, VN2, anti-ECD, and anti-RBD titers based on convalescent plasma donor self-reported clinical characteristics.

Box plots of VN, VN2, anti-ECD and anti RBD titers by (A) dyspnea, (B) hospitalization, and (C) disease severity (1, low severity; 5, high severity) at initial plasma donation from the 68 individual donors. The median, minimum, maximum, first and third quartile, and extreme values are reported. Case counts of donors above and below the VN ≥160 threshold were stratified by whether they self-reported (A) occurrence of dyspnea during symptomatic phase of disease, (B) hospitalization, and (C) disease severity. Pairwise t test (*P < 0.05, **P < 0.01, ***P < 0.001), OR, and relative risk (RR) with CI are also reported.