UC Berkeley

It is now known, the microenvironment including the stroma play an important role in both organ specificity and mammary cancer. In characterizing the interactions between stroma and epithelium, it is useful to develop an ex vivo model to more freely dissect out the intricate network of signals that are necessary to allow functional differentiation in vivo. With this in mind, my first aim was to develop such models to study the mammary gland outside the animal in designer microenvironments. It has been known for some time that once cells are removed from their native tissue environment and placed into traditional two-dimensional (2D) cultures, cells lose functional performance and relevant morphology (M.J. Bissell 1981). In 1977, Emerman and Pitelka using a technique developed by Michalopoulas and Pitot placed mammary cells on top of collagen-1 gels and allowed it to float (Michalopoulos & Pitot 1975; Emerman & Pitelka 1977). In the presence of lactogenic hormones, mammary cells were able to produce milk proteins. These studies were reproduced and colleagues showed that in the floating collagen gel the important component produced by the cells was laminin-111 (Danielson et al. 1984; Parry et al. 1987). Following these studies, the Bissell laboratory discovered in 1989 that a gel mimicking the properties of the basement membrane, a specialized form of extracellular matrix in glandular tissues, allows mammary cells to produce milk and secrete it vectorially (Barcellos-Hoff et al. 1989; Streuli et al. 1991; Streuli et al. 1995). In this dissertation, I utilized an organoid technique developed in the Bissell laboratory to recapitulate both form and function of mammary gland from small pieces of mammary tissue. By using these culture models, we were able to systematically define the biochemical and environmental signaling cues that are important in mammary gland form and function. As such, we have composed and detailed a number of matrices to reproduce the developmental processes ex vivo similar to what is observed in vivo in the developing mammary gland (Lo et al. 2012). These methodologies illustrate a way to investigate elaborately the epithelium outside the complex microenvironment of the tissue, and provide a system for investigating not only normal developmental processes but also diseases such as cancer. We then applied the use of these three-dimensional (3D) culture models to investigate the developmental processes of mammary gland branching.

Invasion is a key step of branching morphogenesis, the process by which simple epithelial structures form elaborate branched networks (Williams & Daniel 1983; Montesano et al. 1991; Hirai et al. 1998; Simian et al. 2001; Fata et al. 2007). This process requires invasion through a type-I collagen rich stroma in vivo. Matrix metalloproteinases were shown to be expressed both in the epithelium and stroma of the invading terminal end buds, suggesting that these enzymes enable epithelial invasion into the mammary fat pad (Talhouk et al. 1991; Simian et al. 2001; Wiseman et al. 2003; Mori et al. 2009; Mori et al. 2013). To dissect whether matrix metalloproteinase 14 (MMP14) is a key signaling molecule in branching morphogenesis, we utilized 3D culture models comprised of primary mammary organoids and mammary epithelial cell (MEC) line for our study. Motivated by data from a genetic MMP14 mutant mouse, we were able to use our 3D models to uncover reciprocal pathways required for mammary branching morphogenesis (Yana et al. 2007). We found that MMP14 is required for invasion of MECs through stroma and these interactions drive MEC invasion through a collagen-1 microenvironment. Additionally, we identified signals downstream of MMP14 and uncovered the interaction between MMP14 and integrin-β1 (ITGB1) that is essential for MEC invasion to occur. Given the high expression levels of MMP14 in breast cancer, we proposed that the mechanisms we uncovered for branching of normal mammary epithelium are also relevant to the invasion of breast cancer cells through the stroma that surrounds the mammary carcinoma (Mori et al. 2013).

From using 3D models to study development and the interactions of MMP14 in branching morphogenesis, it became apparent that we could further utilize this assay as a system to elucidate the means by which cells become cancerous. Utilizing an elaborate genetic backcross study, we sought to analyze the genetic contributions involved in mammary cancer susceptibility in response to a stimulus such as low dose radiation. Using our 3D culture model and a genome-wide single nucleotide polymorphisms (SNPs) analysis, we revealed how treatment with ionizing radiation led to interactions with the genetic loci and identified TGF-1 as a factor regulating cancer susceptibility. Our ex vivo models allowed us to assess the particular signaling components that provide resistance to cancer risk thus opening possible new avenues to identify individual risk for environmental exposure and cancer (Zhang P*, Lo A* et al. 2015).