PROJECT SUMMARY/ABSTRACT Understanding the molecular mechanisms that drive prostate cancer progression has profound significance for defining the biology of lethal prostate cancer. The cancer stem cell model proposes that cells within a tumor are organized in a hierarchical lineage relationship and display different tumorigenic potential. This model has important translational implications since it suggests that effective therapeutics should target cancer stem cells that sustain tumor malignancy. However, despite intensive investigation, long-standing questions about the existence and properties of prostate cancer stem cells remain unresolved. To identify, characterize, and therapeutically target this population, we will combine the expertise of cancer stem cell researchers with computational/mathematical approaches. Leveraging techniques from theoretical physics (Random Matrix Theory) and algebraic topology (Topological Data Analysis), we will provide a statistical framework to dissect a molecular signature (termed CasPro) implicated in progenitor activity in prostate tumors. We will experimentally validate our findings by investigating CasPro-high cells in genetically-engineered mouse models of prostate cancer and analyzing their functional properties in assays of stem cell activities. These findings will then be further validated in studies of human prostate tumor samples, particularly from patients with treatment-resistant castration-resistant prostate cancer (CRPC). In the longer-term, we anticipate that our studies of key regulators of the stem cell signature as well as the epigenomic landscape of CasPro-high cells will identify novel targets for therapy, and lead to novel candidate drugs for these targets. Further development of candidate drugs might result in inhibition of cancer stem cell activity, thereby providing potential treatments for CRPC.

PROJECT NARRATIVE Although prostate cancer stem cells have been investigated for the past twenty years, it has been difficult to provide definitive evidence for their existence or properties. In part, previous studies have been hampered by the lack of high-resolution methods to examine the behavior of individual cells within tumors, such as single-cell RNA-sequencing, as well as sophisticated computational approaches for the precise identification of candidate markers. Leveraging techniques from theoretical physics and algebraic topology, we will provide a statistical framework to identify and characterize tumor stem cells using single-cell data, and will validate our findings using mouse models and human prostate tumor samples.