Bridging Predictive Models and Precision Immunotherapies

ASPIRE Award, PHASE II (2022-Present)

Benjamin Greenbaum, PhD, Memorial Sloan Kettering Cancer Center

It has long been appreciated that the immune system can recognize tumors, but only in the past decade have we seen remarkable progress in our ability to exploit such anti-tumor immunity therapeutically. A breakthrough approach has been the use of immune-checkpoint blockade (ICB) therapies to activate the immune system, allowing T cells to recognize neoantigens presented on the surface of cancer cells. While ICBs and other immunotherapies have permanently changed cancer care across multiple malignancies, most patients do not respond to immunotherapies. Therefore, to advance the next generation of immunotherapies, it is critical to improve our understanding of the underlying mechanisms of immune recognition in tumors so we may exploit these vulnerabilities. In previous work supported by The Mark Foundation for Cancer Research as a The Pershing Square Sohn Prize—Mark Foundation Fellow, Benjamin Greenbaum and his team modelled specific features underlying the recognition of passenger neoantigens and aberrantly transcribed self-RNA from repetitive elements by T cells. With this ASPIRE II award, they will build on that work by showing how specific neoantigens generated by driver mutations are constrained to maintain their immunogenicity despite evolutionary pressure to eliminate it, and how the growth of DNA repeats during tumor evolution creates new therapeutic vulnerabilities.

They will first characterize the immunological vulnerabilities of driver mutations using a combination of mechanism-driven machine learning models and patient data from clinical collaborations. After identifying immunogenic mutations in driver genes, they will validate the immunogenicity of these mutations in patient samples and identify T-cell receptors targeting those neoantigens. Initially they will focus their work on mutations in TP53, which will serve as a paradigm for extending this model to other driver genes in a precise, clinically translatable manner.

In parallel, they will study how the immunogenic properties of DNA repeat expansions in unstable genomes can be targeted as a therapeutic vulnerability in cancer. For this effort, the team will use computational tools they previously developed to analyze whole genome and targeted long-read sequencing from curated patient data. Greenbaum and his team will use these data to probe how immunogenic repeat expansions affects immunotherapy. Importantly, they will examine whether such expansions are mechanistic drivers of response via innate immune engagement as nucleic acids or adaptive immune engagement as tumor antigens.

The work will be done in the context of the newly established Program in Computational Immuno-oncology at Memorial Sloan Kettering Cancer Center which ensures that the tools described herein will be integrated into public resources and used in translational research. These studies combine analytical approaches with novel sequencing modalities, immunological assays, and curated data from rationally assembled patient cohorts and animal models. In doing so, the team will create mechanism-driven models that leverage immunological vulnerabilities in the cancer genome for precision immunotherapies.