Abstract
Neoantigens (neoAg) offer a unique opportunity for directing a patient’s adaptive immune system against tumors while avoiding damage to normal tissues. Current methods for their accurate identification and therapeutic targeting suffer from three main drawbacks: 1) they predict rather than confirm neoAg, 2) their “hit rate” is often too low to be useful in tumors of low to moderate mutational burden such as HNSCC, and, 3) they fail to inform on whether a given mutation is a natural target that can be recognized by CD8+ T cells on tumor cells or by CD4+ T cells on local antigen-presenting cells (APC). This last aspect is perhaps most critical, as a T cell that recognizes a peptide but not a target cell is of little therapeutic value to a cancer patient. We have developed a new unbiased functional approach to neoAg identification that combines bioinformatic analysis of genomic sequence data with functional T-cell assays from a patient’s own PBMC and tumor-infiltrating lymphocytes (TIL). It is based on the concept that, although in silico algorithms may seek to model one or another features of a given mutation’s predicted immunogenicity, the intact immune system can provide proof as to which mutations have been targeted by T cells, and does so according to defined rules of antigen processing and presentation that operate in both cross-presenting antigen-presenting cells (APC) that prime the response and in the tumor cells that express the source antigen. Our approach does not rely on in silico prediction of MHC binding nor any other speculative aspect of peptide immunogenicity and has minimal tissue and peripheral blood requirements. Our preliminary data show it can identify neoAg at a 10x higher rate than current methods and reveals both CD4+ and CD8+ responses. We have performed this analysis for the 4 neoAg identified for the syngeneic murine SCC VII model, and find that the prime/boost vaccination protocol results in significant but incomplete protection from challenge with live SCC VII tumor cells, thus demonstrating that, as a pool, these peptides induce T cells capable of tumor recognition. Furthermore, we have performed preliminary experiments demonstrating that, although SCC VII tumor responds to both PD-1 and CTLA4 blockade immunotherapy, the growing tumors are eradicated much faster in mice that had a pre-existing neoAg-specific T cell response induced in them prior to challenge and treatment. Mice that were vaccinated twice with neoAg peptides + polyI:C but did not receive immune checkpoint blockade showed some initial tumor control that was subsequently lost, leading to progressive growth. A phase 1b clinical trial is now enrolling subjects with advanced cancer that will produce a personalized vaccine for each patient based on our neoantigen identification methodology. The trial combines the vaccine with the anti-PD1 antibody, pembrolizumab, and enrolls patients into 2 consecutive cohorts to elucidate the effects on neoepitope specific T-cell responses of the vaccine versus anti-PD1 targeting. This presentation will summarize our efforts to identify neoantigens, efficacy of a neoantigen-specific vaccine in the murine model, and the current phase 1b study rationale and design.
Citation Format: Ezra Cohen. Augmenting anti-PD1 activity via an HLA-agnostic, mutation-burden independent, personalized neoantigen vaccine strategy [abstract]. In: Proceedings of the AACR-AHNS Head and Neck Cancer Conference: Optimizing Survival and Quality of Life through Basic, Clinical, and Translational Research; 2019 Apr 29-30; Austin, TX. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(12_Suppl_2):Abstract nr IA16.