A cell-based ELISA as surrogate of virus neutralization assay for RBD SARS-CoV-2 specific antibodies

Abstract

SARS-CoV-2, the cause of the COVID-19 pandemic, has provoked a global crisis and death of millions of people. Several serological assays to determine the quality of the immune response against SARS-CoV-2 and the efficacy of vaccines have been developed, among them the gold standard conventional virus neutralization assays. However, these tests are time consuming, require biosafety level 3 (BSL3), and are low throughput and expensive. This has motivated the development of alternative methods, including molecular inhibition assays. Herein, we present a safe cell-based ELISA-virus neutralization test (cbE-VNT) as a surrogate for the conventional viral neutralization assays that detects the inhibition of SARS-CoV-2 RBD binding to ACE2-bearing cells independently of species. Our test shows a very good correlation with the conventional and molecular neutralization assays and achieves 100% specificity and 95% sensitivity. cbE-VNT is cost-effective, fast and enables a large-scale serological evaluation that can be performed in a BSL2 laboratory, allowing its use in pre-clinical and clinical investigations.

Keywords: SARS-CoV-2; Vaccine; cVNT; cell-based ELISA; mVNT.

Graphical abstract

Fig. 1

Cell line selection for RBD binding to ACE2 in a cell-based ELISA. Binding of RBD-hFc (A) and RBD-mFc (B) to ACE2 of Vero and Vero-E6 cells. PDL1-hFc and PDL1-mFc were used as irrelevant fusion proteins. Graphics represent the mean OD450nm for technical duplicates of two independent experiments ± SD. (C) Signal-to-noise (S/N) ratio of RBD-hFc and RBD-mFc in Vero and Vero-E6 cells is represented.

Fig. 2

Determination of optimal cell density and growth time of Vero cells for RBD reactivity cell-based ELISA. Binding of RBD-hFc (upper panel: A-C) or RBD-mFc (lower panel: D-F) to ACE2 of Vero cells. (A-F) Bars represent means of OD_450nm duplicate samples of three independent experiments ± SD. In (C) and (F) is represented the result corresponding to plates coated with 40,000 cells/well. Different letters indicate significant differences and “ns” means not significant, *p < 0.01, **p < 0.001, ****p < 0.0001 by two-way ANOVA with Tukey’s post-hoc.

Fig. 3

Cell-based ELISA to measure direct binding of RBD-Fc to ACE2 on fixed Vero cells. The illustrations were created using BioRender. HRP: horseradish peroxidase

Fig. 4

Detection of SARS-CoV-2 NAbs by a cell based ELISA-virus neutralization test (cbE-VNT). The illustrations were created using BioRender. HRP: horseradish peroxidase

Fig. 5

Specificity and sensitivity of cbE-VNT. Sera (final dilution 1/100) from COVID-19 convalescents (n = 110), or healthy controls (n = 114) were evaluated. Inhibition percentages are shown, as mean ± SD of duplicate values. The horizontal lines indicate the median values. The dotted lines represent the cutoff at 15% inhibition. The P value was calculated from unpaired two-tailed Mann Whitney test.

Fig. 6

Inhibition of RBD binding to ACE2 receptor by sera from immunized individuals, as measured by cbE-VNT. Sera from COVID-19 convalescents (n = 29) (A), recovered people immunized with FINLAY-FR1A dimeric RBD recombinant vaccine (n = 30) (B) or mice (n = 14) (C) and hamsters (n = 9) (D) immunized with RBD-TT/AlOH₃ were evaluated. Placebo and healthy donors sera were used as negative controls. Inhibition percentages are shown as mean ± SD of duplicates values (A, C-D) or of triplicates values (B).

Fig. 7

Correlation between cbE-VNT and mVNT. The neutralization titers of hyperimmune sera of different species: COVID-19 convalescents (A), immunized humans (B) mice (C) and hamsters (D) were used for the analysis. Linear regression and correlations were performed in GraphPad Prism using Pearson’s correlation coefficients. Statistical significance was calculated using the two-tailed test. The data presented are the log of the ID50 value for cbE-VNT and for mVNT. The dashed lines indicate the standard deviations of the linear regression plots.

Fig. 8

Correlation between cbE-VNT and cVNT (A-B) or mVNT and cVNT (C-D). The neutralization titers of RBD-immunized mice and humans sera were used for the analysis. Linear regression and correlations were performed in GraphPad Prism using Pearson’s correlation coefficients. Statistical significance was calculated using the two-tailed test. The data presented are the log of the ID50 value for cbE-VNT or mVNT, and the log of NT50 of cVNT. The dashed lines indicate the standard deviations of the linear regression plots.