New methods to study human mammary development and breast cancer

Abstract

Breast cancer is fundamentally a disease of aberrant differentiation. Breast cancers arise from within the normal architecture of the mammary gland and resemble normal mammary epithelial cell types on a molecular and gene expression level. Many tumors become dependent on the signaling pathways that guide mammary differentiation and proliferation, and may be driven by transcription factors and signaling pathways that enforce cell state. It's no wonder then that many fundamental insights into breast cancer biology derive from study of the normal mammary gland. Mouse models of mammary gland development have helped identify many of the key genetic and hormonal drivers of mammary differentiation. However, these systems have some limitations. First, study of stem and progenitor cell differentiation decisions has been hampered by a lack of definitive markers of cell state. Second, validation of these pathways in human mammary tissue has been challenging due to a paucity of human model systems. This thesis describes work to overcome these limitations. Here I describe a computational method to identify regulators of cell state transitions without the need for definitive markers of cell state. Using this method, we identified RUNXI as a regulator of mammary stem cell differentiation, and demonstrated that RUNX1 is required for exit from the stem/progenitor state. RUNX1 inhibition expanded the pool of stem cells and blocked mammary morphogenesis. This thesis also describes the development of a 3D hydrogel culture system that supports the growth of primary human mammary epithelial tissues. The tissues exhibit all major cell types found in the mammary gland and are hormone responsive. We further adapted the culture system to study the early stages of breast cancer progression by injecting tumor cells into the tissues. Tumor cells interacted and intercalated with normal mammary epithelial cells before invading out of the tissues. We utilized this system to validate SMARCE1 as a regulator of human breast cancer progression. SMARCE1 expression is predictive of progression in early-stage epithelial tumors, and SMARCE1 is functionally required for basement membrane degradation. In our tissue model of tumor progression SMARCE1 was dispensable for cell growth and in situ spreading but was required for invasion.