Pre-existing T cell-mediated cross-reactivity to SARS-CoV-2 cannot solely be explained by prior exposure to endemic human coronaviruses

Several studies have reported the presence of pre-existing humoral or cell-mediated cross-reactivity to SARS-CoV-2 peptides in healthy individuals unexposed to SARS-CoV-2. In particular, the current literature suggests that this pre-existing cross-reactivity could, in part, derive from prior exposure to 9common cold9 endemic human coronaviruses (HCoVs). In this study, we characterised the sequence homology of SARS-CoV-2-derived T-cell epitopes reported in the literature across the entire diversity of the Coronaviridae family. Slightly over half (54.8%) of the tested epitopes did not have noticeable homology to any of the human endemic coronaviruses (HKU1, OC43, NL63 and 229E), suggesting prior exposure to these viruses cannot explain the full cross-reactive profiles observed in healthy unexposed individuals. Further, we find that the proportion of cross-reactive SARS-CoV-2 epitopes with noticeable sequence homology is extremely well predicted by the phylogenetic distance to SARS-CoV-2 (R2 = 96.6%). None of the coronaviruses sequenced to date showed a statistically significant excess of T-cell epitope homology relative to the proportion of expected random matches given the sequence similarity of their core genome to SARS-CoV-2. Taken together, our results suggest that the repertoire of cross-reactive epitopes reported in healthy adults cannot be primarily explained by prior exposure to any coronavirus known to date, or any related yet-uncharacterised coronavirus.