Abstract S03-01: Pan-HLA prediction of SARS-CoV-2 epitopes

Abstract

Abstract Introduction: Delineating antiviral T-cell responses to SARS-CoV-2 may shed light on the heterogeneity of clinical outcomes and inform vaccine or therapeutic approaches. Viral antigens can be predicted using computational tools that calculate the binding affinity between viral peptides and antigen presentation machinery. However, in order to account for the role of host genetics in the diversity of responses, this analysis must be performed with consideration of the global diversity of the human leukocyte antigen (HLA) proteins responsible for antigen presentation. Methods: We deployed binding predictions across the SARS-CoV-2 peptidome for 9,360 Class I HLA alleles (2,987 HLA-A; 3,707 HLA-B; 2,666 HLA-C; 9-mers) using a consensus approach of 7 algorithms and 3,486 Class II HLA alleles (15-mers) using a consensus approach of 4 algorithms. All pMHC predictions were filtered to include only those with consensus binding less than 500 nM. Results: There were 368,145 unique combinations of peptides and HLA alleles (pMHCs) with a predicted binding affinity of less than 500nM, including 1,103 unique 9-mer and 2,547 15-mer peptides and 1,022 MHC Class I and 8,075 MHC Class II HLA proteins. Of these pMHCs, 82% of 9-mers overlapped with 15-mers, suggesting cross-presentation to both CD4 and CD8 T cells in a subset of individuals. We evaluated this filtered dataset with respect to the population frequency of HLA haplotypes. While the predicted susceptibility of SARS-CoV-2 antigen presentation differed greatly across countries, there was a subset of 21 Class I antigens shared by common HLA types across 30 or more countries (out of 79 countries with reported population frequency data). Our database has been made publicly available, and we have developed a user interface to explore the results based upon viral proteins, HLA alleles, or country populations of interest. Conclusions: With the ongoing SARS-CoV-2 pandemic, there are worldwide efforts to generate a successful vaccine and to evaluate clinical samples to understand the viral pathogenesis and diverse outcomes in patients. This application can serve as a guide to identify responses of putative SARS-CoV-2-specific T cells across patients with a broad range of HLA haplotypes internationally. Citation Format: Katie M. Campbell, Gabriela Steiner, Daniel K. Wells, Antoni Ribas, Anusha Kalbasi. Pan-HLA prediction of SARS-CoV-2 epitopes [abstract]. In: Proceedings of the AACR Virtual Meeting: COVID-19 and Cancer; 2020 Jul 20-22. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(18_Suppl):Abstract nr S03-01.