APS March Meeting 2024
Monday–Friday, March 4–8, 2024;
Minneapolis & Virtual
Session N00: Poster Session II (11:30am-2:30pm CST)
11:30 AM,
Wednesday, March 6, 2024
Room: Hall BC
Sponsoring
Unit:
APS/SPS
Abstract: N00.00320 : Predicting Peptide Immunogenicity in Cancer*
Abstract
Presenter:
Marcus Thomas
(Mount Sinai Sch of Med)
Author:
Marcus Thomas
(Mount Sinai Sch of Med)
Within the last decade there has been remarkable progress in cancer therapies based on utilizing a patient’s own immune system to recognize and target cancer cells. Unfortunately, cancer mutated peptides - neoantigens - often exist in an ambiguous space in terms of their immunological similarity to self (as characterized by the subset of MHC-presented epitopes from among the human proteome) and to non-self. Tumor vaccines in particular rely on the fact that immune cells can be taught to recognize neoantigens as non-self. This recognition process involves a number of separate molecular processes which, in aggregate, contribute to the evolutionary fitness of the clonal populations to which a neoantigen belongs. Primary among these processes are antigen presentation — MHC molecules associated with human leukocyte alleles (HLAs) present potentially foreign antigens to the immune system — and subsequent T-cell receptor (TCR) binding. We develop a neoantigen quality model by considering how immune tolerance, i.e., the natural suppression of TCRs that bind to immunologically-self peptides thereby preventing auto-immune reactions, affects their binding to presented neoantigens in the context of patient specific HLAs. Our new approach incorporates terms relating neoantigen-HLA pairs to the presented human proteome. These terms are combined in a thermodynamics inspired equation governing the binding dynamics of neoantigens and T-cell receptors. The overall fitness of a clonal population within a tumor depends on the contributions from every neoantigen, allowing a direct connection between accumulated mutations, the immune system and predictions of evolutionary trajectories. We present initial results illustrating the framework's ability to distinguish patient survival duration based on differential clone fitness.
*NCI Training Grant