Fig. 1 Serological survey of SARS-Cov-2 antibodies in human serum samples. Four serological assays were used to detect anti–SARS-Cov-2 antibodies in serum samples: (top row) individuals sampled between 2017 and 2019 (before pandemic), (second row) hospitalized cases with confirmed COVID-19, (third row) mildly symptomatic individuals from the Crépy-en-Vallois pandemic cluster with suspected COVID-19, and (bottom row) healthy blood donors. ELISA N and ELISA tri-S assays are conventional ELISAs using either N protein or the trimeric ectodomain of S protein as antigens. S-Flow is an assay detecting antibodies bound to cells expressing S protein by flow cytometry. The LIPS S1 and N assays detect either S1 or N protein fused to luciferase by immunoprecipitation. Pre-pandemic serum samples were used to determine the cutoff for each assay, which is indicated by a dotted line and a green area. The two ELISA assays were set to 95% specificity; the specificity of the S-Flow assay and LIPS assay was 99%. The number of positive samples is indicated. Each dot represents one sample. OD, optical density; StoN, signal-to-noise ratio.
Fig. 2 Antibody detection in serum samples from five hospitalized patients with COVID-19. The kinetics of seroconversion in serum samples from five hospitalized patients with COVID-19 (B10 to B14) were measured by four different serological assays. At least five longitudinal serum samples were collected for each patient up to 20 days after symptom onset. All patients were admitted to the intensive care unit. Each line represents one patient. Dotted lines and green areas indicate cutoff for positivity in the seroprevalence assays.
Fig. 3 Comparison of positive serum samples. The number of positive serum samples detected by each serological assay is shown for the three cohorts: hospitalized patients with COVID-19, mildly symptomatic individuals, and healthy blood donors. Correspondence of the positive results is shown among the four assays. For a given assay, each row indicates the number of positive samples that were also positive with the other three assays. Bold numbers indicate the number of positive samples for a given assay. The number of positive samples is color-coded: White corresponds to lower numbers, and green corresponds to higher numbers.
Fig. 4 Correlations among the four serological assays. To compare the four serological assays, results from serum samples from mildly symptomatic individuals and hospitalized patients with COVID-19 (n = 329) were pooled. (A) Results obtained with one assay were correlated with those of the other three assays. Dotted lines indicate assay cutoff values for positivity. Values in pale green areas are positive in one assay, and values in darker green areas are positive in two assays. Each dot represents one study participant. (B) Pearson correlation coefficient (R2) of each comparison is shown. R2 values are color-coded, with white corresponding to the lowest value and dark blue corresponding to the highest value. All correlations are significant (P < 0.0001).
Fig. 5 Virus neutralizing activity in human serum samples. (A) Virus-neutralizing activity (dilution 1:100) of 12 serum samples from the mildly symptomatic cohort of individuals with suspected COVID-19 (C1 to C12) and 9 serum samples from hospitalized patients with COVID-19 (B1 to B9). Virus-neutralizing activity was determined by the pseudovirus neutralization assay and compared to serology data obtained with the four serological assays. Numbers in the top left quadrant indicate the Spearman correlation coefficient, r. All correlations are significant (P < 0.0001). (B) Neutralization activity of serum samples (B1 to B9) from hospitalized patients with COVID-19 was plotted against days after symptom onset. The black line corresponds to a nonlinear fit of the data.
- Table 1 The four serological assays used in this study.
Assay Antigen Serum dilution Readout ELISA N N 1:200 Optical density ELISA tri-S Trimeric S 1:400 Optical density S-Flow S at the cell surface 1:300 Flow cytometry LIPS S1 and N 1:10 Bioluminescence (luciferase) - Table 2 Characteristics of the four cohorts.
Cohorts n Samples Date Area COVID-19 Pre-pandemic
individuals491 491 2017–2019 France Naïve Hospitalized patients
with COVID-1951 161 January to March 2020 Paris, France Confirmed Mildly symptomatic
individuals209 209 3–4 March 2020 Crépy-en-Vallois, France Suspected Healthy blood donors 200 200 20–24 March 2020 Lille, France Unknown - Table 3 SARS-CoV-2 seroprevalence in hospitalized patients with COVID-19, mildly symptomatic individuals, and healthy blood donors.
Cohort ELISA N ELISA tri-S S-Flow LIPS S1 + N Antibody prevalence Pre-pandemic
individuals (specificity)23/491 (95%) 5/100 (95%) 0/134 (100%) 3/280 (99%) Hospitalized patients
(seroprevalence)33/51 (65%) 35/51 (69%) 21/29 (72%) 35/51 (69%) 69% (65–72%) Mildly symptomatic
individuals56/209 (27%) 75/209 (36%) 73/209 (35%) 68/209 (32%) 32% (27–36%) Healthy blood donors 0/200 2/16 6/200 0/200 3%
Supplementary Materials
stm.sciencemag.org/cgi/content/full/scitranslmed.abc3103/DC1
Fig. S1. Seroreactivity of the SARS-CoV-2 tri-S ELISA assay.
Fig. S2. Seroreactivity of the S-Flow assay.
Fig. S3. Seroreactivity of the SARS-CoV-2 LIPS assays.
Fig. S4. Characteristics of the pseudovirus neutralization assay.
Table S1. Clinical status of hospitalized patients with COVID-19.
Table S2. Characteristics of mildly symptomatic individuals with suspected SARS-CoV-2 infection.
Data file S1. Individual-level data for figures.
Additional Files
The PDF file includes:
- Fig. S1. Seroreactivity of the SARS-CoV-2 tri-S ELISA assay
- Fig. S2. Seroreactivity of the S-Flow assay
- Fig. S3. Seroreactivity of the SARS-CoV-2 LIPS assays
- Fig. S4. Characteristics of the pseudovirus neutralization assay
- Table S1. Clinical status of hospitalized patients with COVID-19
- Table S2. Characteristics of mildly-symptomatic individuals with suspected SARS-CoV-2 infection
Other Supplementary Material for this manuscript includes the following:
- Data File S1. Individual level data for figures
