Next generation platforms for interrogating TCR specificity

Project Abstract: Peptide-MHC tumor antigens provide actionable value for cancer detection, and therapeutic targeting. But the process of identifying tumor antigens and then detecting them on tumor tissue at the protein level, and then targeting them therapeutically remains a major bottleneck. Technologies that can interrogate, and mimic, how T cell receptors recognize the composite peptide-MHC surface are important for clinical translation to antigen- specific immunotherapies. We wish optimize development of two validated complementary methods for the identification, and subsequent detection and targeting of pMHC tumor antigens at the protein level. In the first approach, the PI has developed a technology, yeast peptide-MHC display, to identify peptide ligands of ‘orphan’ TCRs in the natural immune system or from pathogenic systems (e.g. cancer, autoimmunity, infectious disease). We have deployed this technology to discover new tumor antigens, but the platform has several technical liabilities that limit its impact. We wish to engineer improved ‘next generation’ versions of the pMHC display technology that will increase the speed and yield of screening and help solve the problem of MHC restriction ambiguity by TCRs. In the second approach, to make the best use of new ligands from Aim #1, we wish to optimize experimental and computational platforms for the rapid isolation of high-affinity “TCR mimic” antibodies and machine-learning designed pMHC mini-binders, respectively, that specifically recognize pMHC ligands and enables us to track tissue expression and induce selective killing of cells expressing these antigens. While exome sequencing reveals neoantigens, it does not tell us if the antigen is expressed on the tumor cell surface. Using our validated technology we are able to quantify the number of tumor antigens expressed on the tumor cell. The pMHC display tumor antigen discovery platform goes with the TCR-mimic binder discovery as a “one-two punch” for protein-level detection and targeting of cancer antigens. We request support to further develop the platforms to increase hit rates and make the workflow more user friendly for adoption by the wider community.