ABSTRACT
Background. T cells are central to the early identification and clearance of viral infections and support antibody generation by B cells, making them desirable for assessing the immune response to SARS-CoV-2 infection and vaccines. We combined 2 high-throughput immune profiling methods to create a quantitative picture of the SARS-CoV-2 T-cell response that is highly sensitive, durable, diagnostic, and discriminatory between natural infection and vaccination. Methods. We deeply characterized 116 convalescent COVID-19 subjects by experimentally mapping CD8 and CD4 T-cell responses via antigen stimulation to 545 Human Leukocyte Antigen (HLA) class I and 284 class II viral peptides. We also performed T-cell receptor (TCR) repertoire sequencing on 1815 samples from 1521 PCR-confirmed SARS-CoV-2 cases and 3500 controls to identify shared public TCRs from SARS-CoV-2-associated CD8 and CD4 T cells. Combining these approaches with additional samples from vaccinated individuals, we characterized the response to natural infection as well as vaccination by separating responses to spike protein from other viral targets. Results. We find that T-cell responses are often driven by a few immunodominant, HLA-restricted epitopes. As expected, the SARS-CoV-2 T-cell response peaks about 1-2 weeks after infection and is detectable at least several months after recovery. Applying these data, we trained a classifier to diagnose past SARS-CoV-2 infection based solely on TCR sequencing from blood samples and observed, at 99.8% specificity, high sensitivity soon after diagnosis (Day 3-7 = 85.1%;Day 8-14 = 94.8%) that persists after recovery (Day 29+/convalescent = 95.4%). Finally, by evaluating TCRs binding epitopes targeting all non-spike SARS-CoV-2 proteins, we were able to separate natural infection from vaccination with > 99% specificity. Conclusion. TCR repertoire sequencing from whole blood reliably measures the adaptive immune response to SARS-CoV-2 soon after viral antigenic exposure (before antibodies are typically detectable) as well as at later time points, and distinguishes post-infection vs. vaccine immune responses with high specificity. This approach to characterizing the cellular immune response has applications in clinical diagnostics as well as vaccine development and monitoring.