MHC validator boosts unbiased discovery of conserved T-Cell epitope vaccine candidates through improved immunopeptidomics sensitivity

Next-generation T-cell-directed vaccines against COVID-19 aim to induce long-lasting T cell immunity to protect against circulating and future hypermutated SARS-CoV-2 variants. Despite the potential of mass spectrometry (MS)-based immunopeptidomics in guiding T-cell vaccine design against rapidly mutating viruses, persistent computational challenges hinder the sensitive, accurate and unbiased identification of conserved, vaccine-relevant T-cell epitopes. To address these challenges, we present a comprehensive framework that integrates a novel machine learning algorithm, immunopeptidomics, intra-host data, epitope immunogenicity, and global CD8+ T-cell epitope conservation analyses. At the forefront of our approach is MHCvalidator, an innovative artificial neural network-based algorithm designed to increase the sensitivity of MS-based immunopeptidomics experiments by modeling antigen presentation and sequence-feature properties, enhancing the unbiased discovery of both self and viral peptide antigens. Notably, the application of MHCvalidator led to the identification of a new form of nonconventional SARS-CoV-2 T-cell epitope presented by B7 supertype molecules. This discovery stems from a +1-frameshift in a truncated version of the Spike (S) antigen, supported by ribo-seq data. Remarkably, analysis of SARS-CoV-2 proteomes from ~80,000 COVID-19 patients unveiled a prevalent S antigen truncation in ~65% of cases, revealing a rich source of frameshifted and untapped viral antigens. Furthermore, integral to our framework, we developed a new computational pipeline to track the global mutational dynamics of MHCvalidator-identified SARS-CoV-2 CD8+ epitopes from the T-cell-directed vaccine BNT162b4, currently undergoing clinical trials. While most CD8+ epitopes encoded by the vaccine exhibit global conservation from January 2020 to October 2023, a highly immunodominant A*01-associated epitope, especially in hospitalized patients, experiences substantial mutations in the Delta and Omicron variants. Collectively, our approach unveils unprecedented SARS-CoV-2 T-cell epitopes, elucidates novel antigenic features, and highlights the non-trivial mutational dynamics of immunodominant, vaccine-relevant epitopes. Our approach can be applied to any viruses and underscores the need for continuous vigilance in T-cell vaccine development against the evolving landscape of circulating and future hypermutated SARS-CoV-2 variants.