Cancer Immunoediting is the process by which the immune system controls and shapes cancer. We originally envisaged and subsequently showed that, in its most complex form, cancer immunoediting occurs in three phases: Elimination (also known as cancer immunosurveillance, the host protective phase of the process), Equilibrium (the phase in which tumor cells that survive immune elimination remain under immunologic growth control resulting in a state of functional tumor dormancy) and Escape (the phase where clinically apparent tumors emerge because immune sculpting of the tumor cells has produced variants that display either reduced immunogenicity or enhanced immunosuppressive activity) (1-3). Strong experimental data have been obtained using mouse cancer models to demonstrate the existence of each phase of the cancer immunoediting process and compelling clinical data suggests that a similar process also occurs during the evolution of certain types of human cancer. Our efforts now focus on elucidating the molecular and cellular mechanisms that underlie each phase of cancer immunoediting and identifying the critical checkpoints that regulate progression from one phase of the process to the next. We recently used a combination of exome sequencing and epitope prediction algorithms to show that mutant proteins in highly immunogenic tumor cells derived from methylcholanthrene treated immunodeficient mice represent immunodominant, tumor specific antigens for CD8+ T cells and that immunoselection is a major mechanism of immunoediting (4). More recently, we asked whether our approach could identify antigens in progressively growing tumors that render them susceptible to checkpoint blockade immunotherapy. T cell lines generated from anti-PD-1 treated mice that had rejected d42m1-T3 progressor sarcoma cells displayed restriction to H-2Kb but not to H-2Db, suggesting that anti-PD-1 promotes T cell responses to only a limited number of antigens. We then identified expressed nonsynonymous mutations in d42m1-T3 cells using exome sequencing and generated a prioritized list of potential H-2Kb binding epitopes. This analysis predicted two unequivocal “best candidates”—a mutant form of Laminin a subunit 4 (mLama4) and a mutant glucosyltransferase (mAlg8). When tested in vitro, these two epitopes were the only ones among the 61 top predicted H-2Kb binding sequences that stimulated either d42m1-T3 specific T cell lines or CD8+ tumor infiltrating T cells freshly isolated from d42m1-T3 tumors growing in mice treated with anti-PD-1. These findings were further validated by showing that: (i) mLama4 and mAlg8 epitopes were detected in association with H-2Kb on d42m1-T3 tumor cells; (ii) CTLs expressing TCRs for mLama4 and mAlg8 accumulated over time in d42m1-T3 tumors in anti-PD-1-treated, tumor-bearing mice; (iii) vaccination of naïve WT mice with mutant but not WT forms of Lama4 or Alg8 induced strong CD8+ T cell responses; and (iv) naïve mice vaccinated against mLama4 plus mAlg8 controlled outgrowth of d42m1-T3 tumors when vaccination was performed not only prophylactically but also therapeutically. Interestingly, we found that mLama4 and mAlg8 CD8+ T cells were also present in d42m1-T3 tumors growing progressively in WT mice treated with control mAb. However, as determined by RNA Seq and immunostaining/flow cytometry, the functional profiles of antigen specific T cells from tumor bearing mice treated with control versus anti-PD-1 mAb were very different. Antigen specific CD8+ TILs from tumor bearing mice treated with control mAb displayed a gene expression pattern and surface markers suggesting an “exhausted” T cell phenotype. In contrast, antigen specific CD8+ TILs from mice treated with anti-PD-1 displayed patterns suggesting a reversal of the “exhausted” phenotype and acquisition of anti-tumor effector functions. These patterns were only observed in antigen specific CD8+ TILs and not in CD8+ TILs that showed no specificity for tumor antigens. These results thus reveal that our genomics approach is useful in (a) developing personalized cancer immunotherapies, (b) identifying individuals who would best benefit from checkpoint blockade cancer immunotherapy and (c) identifying tumor antigen specific T cells that function as a predictive biomarker of successful immunotherapy.
Citation Format: Robert D. Schreiber. Using genomics to define the antigenic targets of checkpoint blockade cancer immunotherapy. [abstract]. In: Proceedings of the AACR Special Conference: Tumor Immunology and Immunotherapy: A New Chapter; December 1-4, 2014; Orlando, FL. Philadelphia (PA): AACR; Cancer Immunol Res 2015;3(10 Suppl):Abstract nr IA08.
- ©2015 American Association for Cancer Research.