Background: The difference in the immune response to severe acute respiratory syndrome coro-navirus 2 (SARS-CoV-2) in patients with mild versus severe disease remains poorly understood. Recent scientific advances have recognised the vital role of both B cells and T cells; however, many questions remain unanswered, particularly for T cell responses. T cells are essential for helping the generation of SARS-CoV-2 antibody responses but have also been recognised in their own right as a major factor influencing COVID-19 disease outcomes. The examination of T cell receptor (TCR) family differences over a 12-month period in patients with varying COVID-19 disease severity is crucial for understanding T cell responses to SARS-CoV-2.
Methods: We applied a machine learning approach to analyse TCR vb family responses in COVID-19 patients (n = 151) across multiple timepoints and disease severities alongside SARS-CoV-2 infection-naïve (healthy control) individ-uals (n = 62).
Results: Blood samples from hospital in-patients with moderate, severe, or critical disease could be classified with an accuracy of 94%. Furthermore, we identified significant variances in TCR vb family specificities between disease and control subgroups.
Conclusions: Our findings suggest advantageous and disadvantageous TCR repertoire patterns in relation to disease severity. Following validation in larger cohorts, our methodology may be useful in detecting protective immunity and the assessment of long-term outcomes, particularly as we begin to unravel the immunological mechanisms leading to post-COVID complications.
Keywords: COVID-19; SARS-CoV-2; T cell receptor; TCR repertoire; flow cytometry; immune response; machine learning; severity model.