Spike protein antibodies mediate the apparent correlation between SARS-CoV-2 nucleocapsid antibodies and neutralization test results

ObjectivesSARS-CoV-2 infection induces the formation of different antibodies. However, not all of which might prevent the virus from entering the cell, although their concentrations correlate with the titers of viral neutralization tests (NTs). Antibodies against the viral nucleocapsid (NC), e.g., can be classified as such. We aimed to prove the hypothesis that the apparent correlation between NC-antibody levels and NT-titers is mediated by simultaneously occurring antibodies against viral spike-protein components. MethodsWe included 64 individuals with previous SARS-CoV-2 infection (>14d after symptom onset). SARS-CoV-2 antibodies against the NC (Roche total antibody ECLIA, Abbott IgG CMIA) and spike-protein (Technozym RBD ELISA, DiaSorin S1/S2 CLIA) were measured, and neutralization tests were performed. The effect of spike-protein antibodies on the correlation between NC-antibodies and NT-titers was evaluated by partial correlation and mediation analyses. ResultsBoth tested assays assessing antibodies against the NC correlated significantly with NT titers: Abbott {rho}=0.742, P<0.0001; Roche {rho}=0.365, P<0.01. However, when controlling the rank correlations for the presence of RBD or S1/S2 antibodies, correlation coefficients dropped to {rho}=0.318/{rho}=0.329 (P<0.05/P<0.01), respectively for Abbott and vanished for Roche. As a result, only a maximum of 11% of NT titer variability could be explained by NC-antibody levels. ConclusionsOur data suggest that the apparent correlation between NC antibodies and NT titers is strongly mediated by co-occurring RBD antibody concentrations. To avoid falsely implied causal relationships, all correlation analyses of non-spike-associated antibody assays and neutralization assays should include a partial correlation analysis to exclude a possible mediator effect of spike-associated antibodies.