Project 3 - Drug Mechanism of Action-based targeting of tumor subpopulations

Investigators: Andrea Califano (Project Leader), Cory Abate-Shen, Kenneth Olive, Andrew Kung

In Project 3 we will elucidate the proteome-wide Mechanism of Action (MoA) of clinically relevant compounds via network-based analyses, to target non-oncogene dependencies of molecularly distinct cell states comprising either malignant or immunosuppressive tumor microenvironment (TME) cells coexisting in the same tumor mass. Targeting subpopulations corresponding to distinct cell states is becoming increasingly critical because the heterogeneity and plasticity of both malignant and other TME-related subpopulations have emerged as, perhaps, the most fundamental obstacles to achieving durable responses in cancer patients. To study inter-tumor heterogeneity of drug sensitivity and to develop methodologies to identify high-fidelity, patient-matched models, we will focus on metastatic castration resistant prostate cancer (mCRPC) and on its aggressive neuroendocrine subtype (NEPC) (4, 5). Specifically, we will leverage a novel molecular triangulation methodology (OncoLoop) (6) to first generate drug perturbation profiles and then to validate predicted therapies in preclinical models that recapitulate the drug sensitivity of human tumors. Conversely, to elucidate the molecular dependencies of immunosuppressive TME subpopulations, we will focus on pancreatic ductal adenocarcinoma (PDA) (7, 8, 9), a heterogeneous cancer with a highly immunosuppressive TME. We will first elucidate and target non-oncogene dependencies in six isogenic, yet molecularly distinct malignant subpopulations present in virtually all tumors (7). We will then focus on its TME, which comprises multiple immunosuppressive subpopulations, including three cancer associated fibroblast (CAF) subtypes (9), TREM2+/C1Q+ macrophages (10), and immunosuppressive regulatory T cells (11). We propose that achieving durable response will require targeting distinct malignant compartments, to abrogate viability or reprogram cells to a drug-sensitive state, and/or rescuing immune response by targeting immunosuppressive TME subpopulations. Our goals include:

• Developing novel, fully generalizable methodologies to assess tumor model fidelity in terms of recapitulating drug mechanism of action and drug sensitivity, including for cell lines, organoids, GEMMs, and PDX models. As proof of concept, this will be applied to the elucidation and experimental validation of drugs targeting Master Regulator (MR) proteins representing non-oncogene dependencies of mCRPC and of its aggressive neuroendocrine subtype (NEPC). These methodologies will be developed to be fully generalizable to virtually any other tumor context, including at the single cell level. 
• Extending the OncoTreat (12) and OncoTarget (13) algorithms to elucidate and pharmacologically target Master Regulator dependencies in molecularly distinct, malignant cell subpopulations, at the single cell level. As proof-of-concept, we will focus on six subpopulations contributing to PDA intra-tumor heterogeneity. Drugs will be validated by assessing treatment-induced depletion/reprogramming of target subpopulation(s), first in syngeneic transplants and then in KPC GEMMs. 
• Pharmacological targeting of Master Regulator dependencies in distinct immunosuppressive subpopulations of the TME. As proof-of-concept, we will perform OncoTreat and OncoTarget analysis of subpopulations isolated from the TME of PDA patient/model-derived resections. Predictions aimed at rescuing immune sensitivity and response to immune checkpoint inhibitors will be validated in syngeneic models and KPC GEMM, as monotherapy and in combination with immune checkpoint inhibitors.