Comparison of Four SARS-CoV-2 Neutralization Assays

Abstract

Neutralizing antibodies are a major correlate of protection for many viruses including the novel coronavirus SARS-CoV-2. Thus, vaccine candidates should potently induce neutralizing antibodies to render effective protection from infection. A variety of in vitro assays for the detection of SARS-CoV-2 neutralizing antibodies has been described. However, validation of the different assays against each other is important to allow comparison of different studies. Here, we compared four different SARS-CoV-2 neutralization assays using the same set of patient samples. Two assays used replication competent SARS-CoV-2, a focus forming assay and a TCID₅₀-based assay, while the other two assays used replication defective lentiviral or vesicular stomatitis virus (VSV)-based particles pseudotyped with SARS-CoV-2 spike. All assays were robust and produced highly reproducible neutralization titers. Titers of neutralizing antibodies correlated well between the different assays and with the titers of SARS-CoV-2 S-protein binding antibodies detected in an ELISA. Our study showed that commonly used SARS-CoV-2 neutralization assays are robust and that results obtained with different assays are comparable.

Keywords: SARS-CoV-2; neutralization assay; neutralizing antibodies; pseudotype virus.

Figure 1

Focus forming neutralization assay. (a) Vero-TMPRSS2 cells were infected with replication competent SARS-CoV-2 virus, fixed after indicated incubation times and stained with the serum of a SARS-CoV-2 convalescent patient. Representative microscopic pictures are shown. (b) Neutralization titers for a set of 11 samples were determined using replication competent SARS-CoV-2 in a focus forming assay. Cells were fixed and stained after 10 h incubation and neutralization titers were determined using ImmunoSpot S6 Ultra-V reader and CTL analyzer BioSpot^® 5.0 software (CTL Europe GmbH, Bonn, Germany). The data presented are the result of three independent experiments. Each experiment is represented by a different symbol. Dotted line indicates detection limit for the assay; titers ≤ 1:4 were regarded as negative. (c) Low interassay variability and high reproducibility of the assay was demonstrated by analyzing neutralization titers of a positive control (P3, 1:64 dilution) in 18 independent experiment. Percent neutralization was calculated relative to virus only wells. Shown are individual replicates and mean.

Figure 2

TCID₅₀ neutralization assay. (a) Parental Vero cells, Vero cells expressing TMPRSS2, or Vero cells expressing TMPRSS2 and ACE2 were infected with serial dilutions of SARS-CoV-2. A total of 48 h after infection CPE was analyzed under the microscope and representative microscopic pictures are shown. Additionally, cells were fixed and stained with the serum of a SARS-CoV-2 convalescent patient. Upper panel: noninfected cells; middle panel: immunostaining (10⁻¹ dilution for Vero and Vero-TMPRSS2 cells, 10⁻⁵ dilution for Vero-TMPRSS2/ACE2 cells); lower panel: higher magnification of middle panel to evaluate CPE. (b) Neutralization titers for a set of 11 samples were determined using replication competent SARS-CoV-2 in a TCID₅₀ assay. Neutralization titers of three independent experiments are displayed with different symbols. Dotted line indicates detection limit for the assay; titers ≤ 1:4 were regarded as negative.

Figure 3

Vesicular stomatitis virus (VSV)-based neutralization assay. (a) VSVΔG-GFP-S particles were titrated on parental 293T, BHK21, and Vero cells or on cell lines with stable ACE2 overexpression. For 293T and BHK-21 cells two different lentiviral vectors were used for ACE2 overexpression, either with a blasticinin resistance (B) or a hygromycin resistance cassette (H). A total of 16 h after infection, the number of GFP positive cells was determined using an immunospot reader. Shown are mean ± SEM of duplicate samples from one representative of two independent experiments. (b) Neutralization titers for a set of 11 samples were determined using a GFP encoding VSVΔG-S variant. Each sample was analyzed in four independent experiments. Each experiment is represented by a different symbol. Neutralization titers ≤ 1:4 are regarded negative. (c) A positive control (P3 in a 1:64 dilution) was analyzed using the VSVΔG-GFP-S virus on 66 independent plates. Percent neutralization was calculated relative to virus only wells. Shown are individual replicates and mean. (d) Neutralization titers for a set of 11 samples were determined using a luciferase encoding VSVΔG-S variant. Each sample was analyzed in three independent experiments. Each experiment is represented by a different symbol. Dotted line in b + d indicate threshold for neutralizing antibody titers, i.e., titers ≤ 1:4 are regarded as negative. Dotted line in c indicates 50% neutralization.

Figure 4

Lentiviral particle-based neutralization assay. Lentiviral particles encoding GFP were pseudotyped with SARS-CoV-2 S. Neutralization titers for a set of 11 samples were determined. Each sample was analyzed in three independent experiments. Each experiment is represented by a different symbol. Dotted line indicates detection limit for the assay; titers ≤ 1:4 were regarded as negative.

Figure 5

Neutralization titers for the different assays correlate well with each other and with binding antibody titers. For neutralization assays mean neutralizing titers of three (focus forming, TCID₅₀, and lentiviral particle assay) or four independent experiments (VSV assay) were calculated. (a) Correlation of neutralizing antibody titers and binding antibodies (anti-S IgA, anti-S IgG, and anti-N IgG) were determined. (b) Neutralizing antibody titers determined in each assay were correlated to titers in all three other neutralizing antibody assays. Dotted lines indicate detection limit for the different assays; neutralization assay ≤ 1:4 negative, anti-S IgA and anti-S IgG OD < 0.8 negative, OD 0.8–1.1 borderline, OD > 1.1 positive, anti-N IgG RLU > 1.4 positive. Some of the dots in the blots represent multiple samples.