Research Project
10381324
This proposal combines a team with expertise in aging, tumor immunology, tumor immunotherapy, specific genetically modified animal models and early phase clinical trials with a computational team having great expertise in analyzing and modeling aging of the immune system. We will study age effects on PD-L1/PD-1 signaling in the host and the tumor focusing on melanoma with some bladder cancer work, two tumors that are highly responsive to ?PD-1 and/or ?PD-L1 as proofs-of-concept, and residing in distinct anatomic compartments. In Aim 1 we study tumor PD-L1 intrinsic effects on ?PD-L1 and ?PD-1 treatment in melanoma and bladder cancer using transplantable B16 and inducible Nras/Cdk2n melanoma models, and transplantable MB49 and BBN-induced tumors for bladder cancer studies. We also use novel melanoma and BC models with tumor cell- specific PD-L1KO. We study 3 cohorts of elderly versus younger humans getting ?PD-L1 or ?PD-1 for melanoma or bladder cancer for human validation. We measure high-dimensional cell phenotypes and signaling responses, proteins and genes to maximize the information collected from human samples and mice using 23-color FACS, CyTOF, Luminex, Nanostring and other approaches. In Aim 2 we use all the above models and analytic strategies in young and aged PD-L1KO mice and WT or bone marrow chimeras to test hematopoietic and non- hematopoietic (host) PD-L1 signals in treatment outcomes in melanoma and bladder cancer. In Aim 3 the Systems Immunology team will use their innovative and successful computational modeling to identify age- related co-predictors of immunotherapy response and to identify candidate mechanisms for responders and non- responders. We will define a trajectory of immune system aging in mice at ultra-high resolution by performing a systems level integrative analysis of aging in Collaborative Cross and BL6 mice tracked in a combined longitudinal and cross-sectional study. This trajectory will be used to understand how tumor response and treatment outcomes vary as a function of age, and to build a simple, low parameter (i.e., easily testable and clinically translated), predictive models of treatment response. We will test insights by analyzing immune data from aged versus young patients undergoing ?PD-L1 and ?PD-1 cancer immunotherapy in novel machine learning approaches that we pioneered to identify insights from mouse data that are relevant to humans. Coupling this disease information with the healthy human aging trajectory that we recently defined will allow us to adapt our mouse data to predict optimal treatments in humans based on chronological and immune aging. This combined trans-disciplinary approach will identify common age-related disabilities that reduce PD-L1/PD-1 based immunotherapy responses and suggest tailored treatments for optimal efficacy that could later be tested in validation sets. These data can also be applied to other types of immunotherapy as we will also test.
