T-cell receptor sequencing identifies prior SARS-CoV-2 infection and correlates with neutralizing antibody titers and disease severity

Abstract

Measuring the adaptive immune response to SARS-CoV-2 can enable the assessment of past infection as well as protective immunity and the risk of reinfection. While neutralizing antibody (nAb) titers are one measure of protection, such assays are challenging to perform at a large scale and the longevity of the SARS-CoV-2 nAb response is not fully understood. Here, we apply a T-cell receptor (TCR) sequencing assay that can be performed on a small volume standard blood sample to assess the adaptive T-cell response to SARS-CoV-2 infection. Samples were collected from a cohort of 302 individuals recovered from COVID-19 up to 6 months after infection. Previously published findings in this cohort showed that two commercially available SARS-CoV-2 serologic assays correlate well with nAb testing. We demonstrate that the magnitude of the SARS-CoV-2-specific T-cell response strongly correlates with nAb titer, as well as clinical indicators of disease severity including hospitalization, fever, or difficulty breathing. While the depth and breadth of the T-cell response declines during convalescence, the T-cell signal remains well above background with high sensitivity up to at least 6 months following initial infection. Compared to serology tests detecting binding antibodies to SARS-CoV-2 spike and nucleoprotein, the overall sensitivity of the TCR-based assay across the entire cohort and all timepoints was approximately 5% greater for identifying prior SARS-CoV-2 infection. Notably, the improved performance of T-cell testing compared to serology was most apparent in recovered individuals who were not hospitalized and were sampled beyond 150 days of their initial illness, suggesting that antibody testing may have reduced sensitivity in individuals who experienced less severe COVID-19 illness and at later timepoints. Finally, T-cell testing was able to identify SARS-CoV-2 infection in 68% (55/81) of convalescent samples having nAb titers below the lower limit of detection, as well as 37% (13/35) of samples testing negative by all three antibody assays. These results demonstrate the utility of a TCR-based assay as a scalable, reliable measure of past SARS-CoV-2 infection across a spectrum of disease severity. Additionally, the TCR repertoire may be useful as a surrogate for protective immunity with additive clinical value beyond serologic or nAb testing methods.

Figure 1:. SARS-CoV-2–specific T cell responses correlate with nAb titer.

Correlation of T-cell clonal breadth (A) and depth (B) with nAb titer evaluated by Spearman’s rank-order correlation.

Figure 2:. Association of T-cell clonal breadth with clinical variables.

Association of clonal breadth with (A) hospitalization, (B) fever, (C) difficulty breathing, (D) sex, and (E) age. P values indicated by p and $\widehat{\text{p}}$ are for univariate Mann-Whitney U test and multivariable linear regression (with variables age, sex, hospitalization, fever, difficulty breathing, and TCR rearrangements) respectively.

Figure 3:. Comparison of TCR-based assay with serological assays.

(A) Number of enhanced sequences associated with SARS-CoV-2 infection for RT-PCR–confirmed samples collected at visit 1 (orange) or visit 2 (green). The axes are the two parameters comprising the T-cell Test: enhanced sequence count out of 4,287 enhanced sequences (y-axis) and number of unique TCR rearrangements (x-axis). Blue dots represent 1,657 control samples collected before January 2020. All samples were held out from training the classifier. (B) T-cell Test score, (C) EUROIMMUN IgG OD ratio, and (D) Abbott index as a function of days from symptom onset to sample, indicated as hospitalized (blue) or non-hospitalized (red) individuals, with trend lines connecting visit 1 and visit 2 sample points from the same subject. Blue and red bold trend lines indicate smoothed mean (locally estimated scatterplot smoothing = LOESS [Cleveland 1981]) for hospitalized and non-hospitalized individuals, respectively. (E) T-cell Test scores in RT-PCR-confirmed samples compared to EUROIMMUN IgG OD ratio and (F) Abbott index. Samples classified negative by all three antibody tests (EUROIMMUN, Abbott, and nAb titer) are highlighted in orange. Black dashed lines indicate cutoffs for positivity/negativity. The cutoff used for positivity for nAb is 1:40 recognizing that some samples could have neutralizing titers of < 1:40.