RBD-Based ELISA and Luminex Predict Anti-SARS-CoV-2 Surrogate-Neutralizing Activity in Two Longitudinal Cohorts of German and Spanish Health Care Workers

Abstract

The ability of antibodies to neutralize severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is an important correlate of protection. For routine evaluation of protection, however, a simple and cost-efficient anti-SARS-CoV-2 serological assay predictive of serum neutralizing activity is needed. We analyzed clinical epidemiological data and blood samples from two cohorts of health care workers in Barcelona and Munich to compare several immunological readouts for evaluating antibody levels that could be surrogates of neutralizing activity. We measured IgG levels against SARS-CoV-2 spike protein (S), its S2 subunit, the S1 receptor binding domain (RBD), and the full length and C terminus of nucleocapsid (N) protein by Luminex, and against RBD by enzyme-linked immunosorbent assay (ELISA), and assessed those as predictors of plasma surrogate-neutralizing activity measured by a flow cytometry assay. In addition, we determined the clinical and demographic factors affecting plasma surrogate-neutralizing capacity. Both cohorts showed a high positive correlation between IgG levels to S antigen, especially to RBD, and the levels of plasma surrogate-neutralizing activity, suggesting RBD IgG as a good correlate of plasma neutralizing activity. Symptomatic infection, with symptoms such as loss of taste, dyspnea, rigors, fever and fatigue, was positively associated with anti-RBD IgG positivity by ELISA and Luminex, and with plasma surrogate-neutralizing activity. Our serological assays allow for the prediction of serum neutralization activity without the cost, hazards, time, and expertise needed for surrogate or conventional neutralization assays. Once a cutoff is established, these relatively simple high-throughput antibody assays will provide a fast and cost-effective method of assessing levels of protection from SARS-CoV-2 infection. IMPORTANCE Neutralizing antibody titers are the best correlate of protection against SARS-CoV-2. However, current tests to measure plasma or serum neutralizing activity do not allow high-throughput screening at the population level. Serological tests could be an alternative if they are proved to be good predictors of plasma neutralizing activity. In this study, we analyzed the SARS-CoV-2 serological profiles of two cohorts of health care workers by applying Luminex and ELISA in-house serological assays. Correlations of both serological tests were assessed between them and with a flow cytometry assay to determine plasma surrogate-neutralizing activity. Both assays showed a high positive correlation between IgG levels to S antigens, especially RBD, and the levels of plasma surrogate-neutralizing activity. This result suggests IgG to RBD as a good correlate of plasma surrogate-neutralizing activity and indicates that serology of IgG to RBD could be used to assess levels of protection from SARS-CoV-2 infection.

Keywords: ELISA; Luminex; SARS-CoV-2; antibodies; immunoglobulin G; neutralization; receptor binding domain; spike protein; symptoms.

FIG 1

Radar charts comparing antigen-specific antibody levels (IgG, IgA, and IgM) between Munich and Barcelona cohorts in symptomatic (A) and asymptomatic (B) participants at the different sampling time points. The dots represent the median values of log₁₀ MFI antibody levels. Wilcoxon rank sum test was used to assess statistically significant differences in antibody levels. *, P < 0.05; **, P < 0.01; ***, P < 0.001. Samples sizes in the symptomatic group, 0 months (n = 336), 1 month (n = 123), 3 months (n = 79), and 6 months (n = 131). Samples sizes in the asymptomatic group, 0 months (n = 451), 1 month (n = 499), 3 months (n = 84), 6 months (n = 424).

FIG 2

Correlations between antibody levels measured by Luminex and RBD-ACE2 neutralization capacity. Spearman’s rank correlation test between levels (log₁₀ of median fluorescence intensity [MFI]) of IgA, IgG, and IgM against each antigen (nucleocapsid full-length protein [N-FL], its C-terminal region [N-Cterm], the receptor binding domain [RBD], the full S protein, and its subregion S2) and plasma neutralizing capacity (as a percentage of RBD-ACE2 binding inhibition). Data are presented for both sites together (n = 315 samples) and stratified by site (Barcelona, n = 265; Munich, n = 50). Each dot represents an individual measurement. Black lines represent the fitted curve calculated using the linear model method. Shaded areas represent 95% confidence intervals. Black, gray, and white dots correspond to positive, indeterminate, and negative MFIs, respectively. Indeterminate results correspond to MFIs between the positivity threshold (10 to the mean plus 3 standard deviations [SD] of log₁₀-transformed MFI of the prepandemic controls) and an upper limit at 10 to the mean plus 4.5 SD of the log₁₀-transformed MFIs of prepandemic samples.

FIG 2

Correlations between antibody levels measured by Luminex and RBD-ACE2 neutralization capacity. Spearman’s rank correlation test between levels (log₁₀ of median fluorescence intensity [MFI]) of IgA, IgG, and IgM against each antigen (nucleocapsid full-length protein [N-FL], its C-terminal region [N-Cterm], the receptor binding domain [RBD], the full S protein, and its subregion S2) and plasma neutralizing capacity (as a percentage of RBD-ACE2 binding inhibition). Data are presented for both sites together (n = 315 samples) and stratified by site (Barcelona, n = 265; Munich, n = 50). Each dot represents an individual measurement. Black lines represent the fitted curve calculated using the linear model method. Shaded areas represent 95% confidence intervals. Black, gray, and white dots correspond to positive, indeterminate, and negative MFIs, respectively. Indeterminate results correspond to MFIs between the positivity threshold (10 to the mean plus 3 standard deviations [SD] of log₁₀-transformed MFI of the prepandemic controls) and an upper limit at 10 to the mean plus 4.5 SD of the log₁₀-transformed MFIs of prepandemic samples.

FIG 3

Correlations of IgG-RBD levels measured by Luminex and ELISA between them and by ELISA with plasma neutralizing activity. (A) Correlation between IgG-RBD levels measured by Luminex and ELISA (n = 710 samples, 266 from Barcelona and 444 from Munich). (B) Correlation between IgG-RBD levels measured by ELISA and plasma neutralization activity (n = 315 samples, 265 from Barcelona and 50 from Munich). Data is presented for both sites together and stratified. Each dot represents an individual measurement. Black lines represent the fitted curve calculated using the linear model method. Shaded areas around the regression lines represent the 95% confidence intervals of the mean. Black, gray and white dots correspond to positive, indeterminate, and negative ELISA measurements, respectively. Indeterminate results in ELISA data correspond to a 0.5 to 1.0 ratio.

FIG 4

Relative importance of each predictor variable for the R² in the multiple linear regression model assessing the association of IgG against RBD, N-FL, S, and S2 with plasma neutralization activity.

FIG 5

Forest plots showing the factors associated to IgG-RBD levels measured by ELISA and Luminex (A) and plasma neutralization activity (B). Sample size in panel A is 302 for the anti-RBD IgG ELISA and 778 for the anti-RBD IgG Luminex. Sample size in panel B is 111 for the neutralization data. Multiple linear regression models were used, and all estimates were adjusted for age and sex. The dots represent the point estimates, while the whiskers depict the corresponding 95% confidence intervals.

FIG 6

Comparison of IgG-RBD levels between symptomatic and asymptomatic individuals at different time points after a positive diagnostic by rRT-PCR. (A) IgG-RBD measured by ELISA (n = 335 samples). (B) IgG-RBD measured by Luminex (n = 316 samples). Plots show IgG-RBD levels measured as ratio for ELISA and log₁₀ MFI for Luminex data. The center line of boxes depicts the median values, the lower and upper hinges correspond to the first and third quartiles, the distance between the first and third quartiles corresponds to the interquartile range (IQR), and the whiskers extend from the hinge to the highest or lowest value within 1.5× IQR of the respective hinge. Wilcoxon rank sum test was used to assess statistically significant differences in antibody levels. ns, nonsignificant; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 7

Comparison of plasma neutralization activity between symptomatic and asymptomatic individuals at different time points after a positive diagnostic by rRT-PCR. (A) Comparison between symptomatic and asymptomatic (n = 114). (B) Comparison between presence or absence of the specific symptoms positively associated to neutralization in the multiple linear regression model (loss of taste [n = 113], dyspnea [n = 113], rigors [n = 98], fever [n = 113], fatigue [n = 114]). Plots show percentage of plasma neutralization activity. The center line of boxes depicts the median values, the lower and upper hinges correspond to the first and third quartiles, the distance between the first and third quartiles corresponds to the interquartile range (IQR), and whiskers extend from the hinge to the highest or lowest value within 1.5× IQR of the respective hinge. Wilcoxon rank sum test was used to assess statistically significant differences in percent neutralization. ns, nonsignificant; *, P < 0.05; **, P < 0.01; ***, P < 0.001.