Drug Mechanism of Action-based targeting of tumor subpopulations

Patients with aggressive cancers often present with no pharmacologically actionable mutations and fail to respond to immune checkpoint blockade, thus deriving only modest improvement in disease-free survival from targeted therapy and immunotherapy. Tumor heterogeneity further complicates these challenges by fostering paracrine signal-mediated reprogramming, adaptation, selection, and expansion of drug-resistant cell states, as well as emergence of an immunosuppressive tumor microenvironment (TME), which are ultimately responsible for patient relapse and poor outcome . We propose that addressing these challenges—i.e., identifying more universal, mechanistic targets for pharmacological intervention and assessing their potential value in highly heterogeneous tumors—is critically dependent on the availability of accurate and comprehensive cellular networks, which underlie both the cell-autonomous behavior of cancer cells and their interaction with other TME subpopulations. In Project 3 we propose to match the proteome-wide Mechanism of Action (MoA) of clinically relevant compounds—as dissected from in our PanACEA database of genome-wide molecular perturbations in high-fidelity models of human malignancies—to the non-oncogene dependencies of molecularly distinct, yet coexisting subpopulations, representing either transformed, malignant cells or healthy cells recruited to the TME to create a pro-malignant, immunosuppressive milieu, as dissected by single cell analyses. Targeting individual subpopulations is becoming increasingly critical because the heterogeneity and plasticity of both transformed and non-transformed TME subpopulations have emerged as, perhaps, the most fundamental obstacles to achieving durable responses in cancer patients and distinct subpopulations appear to either have potentially orthogonal drug sensitivities or to represent healthy, immunosuppressive cells that will require an entirely different approach to targeting their recruitment to the TME rather than causing their demise. To accomplish these goals, Project 3 will leverage data, models and reagents generated during the prior CSBC funding cycle for the study of metastatic castration resistant prostate cancer (mCRPC) and pancreatic ductal adenocarcinoma (PDAC), two aggressive, highly heterogeneous malignancies, with ≤ 20% 5-year survival. We will focus on mCRPC and on its aggressive neuroendocrine subtype (NEPC) to explore a novel molecular triangulation methodology (OncoLoop) designed to identify high-fidelity models—i.e., cell lines, organoids, genetically engineered mouse models (GEMMs) and patient derived xenografts (PDXs)—to generate patient-relevant drug perturbation profiles in vitro and to validate drugs predicted from patient-derived sample analysis in preclinical models in vivo. We will then focus on 6 molecularly distinct malignant PDAC subpopulations—comprising Lineage, Morphogenic, and Acinar to Ductal Metaplasia-like cells, each detected in either a MAPK pathway active or inactive state—exploring multiple pro-malignant TME subpopulations (including tumor infiltrating T regulatory cells, macrophages and fibroblasts), to identify small molecule inhibitors that effectively deplete them.