Given these limitations the current

Given these limitations, the current knowledge regarding the identity of normal human mammary stem dextromethorphan hydrobromide is based on markers that associate with the highest enrichment in stem-like functional properties, such as the ability to differentiate along both luminal and myoepithelial lineages, branching morphogenesis in 3D culture, and generation of outgrowths in xenotransplantation experiments. Combinations of cell surface markers that have been used to detect cell populations enriched in these properties include CD49fhighEpCAMlow (Eirew et al., 2008; Lim et al., 2009), CD73+CD90– (Roy et al., 2013), CD10+ (Keller et al., 2012), and CD49f+DLL1+DNER+ (Pece et al., 2010). Functional properties used to identify stem cells are high aldehyde dehydrogenase (ALDH) activity (Ginestier et al., 2007) and the ability to survive and proliferate in anchorage-independent conditions (Dontu et al., 2003; Pece et al., 2010). Some of these markers (i.e., ALDH+ and CD49f+) correlate with poor clinical outcome when highly expressed in breast tumors (Ali et al., 2011; Ginestier et al., 2007), possibly because they also identify a cancer stem cell population. Other stem cell markers validated in in vitro assays include SSEA4+ and CK14+CK19+ (Villadsen et al., 2007). All these phenotypes identify heterogeneous cell populations that contain more differentiated cells in addition to stem cells. The combination of assays and markers listed above have not led to a consensus regarding the identity and localization of human mammary stem cells (Visvader and Stingl, 2014). To address this issue, we adopted an alternative, theoretical approach based on modeling mammary morphogenesis. We utilized 1D cell-replacement rules as well as computer-generated 3D fractals for modeling the human mammary lobule. This approach allowed us to formulate hypotheses for the localization of stem and progenitor cells within the branching structure of the gland. We compared predictions of these theoretical models with the pattern of marker expression in situ, as determined by immunostaining of sections of normal breast. Several proposed stem cell markers were co-expressed and their localization in situ coincided with the predictions of one of the models put forward in this study, in which stem cells are primarily present in clusters at the growing ends of intralobular branching ductules.
Results
Discussion The stem cell model of cancer development posits that stem and progenitor cells present in adult tissues constitute the main target of malignant transformation. Additionally, it proposes that intra- and inter-tumoral heterogeneity can be attributed in part to the molecular features contributed by the cell of cancer origin, superimposed on aberrant differentiation and genomic instability. Experimental evidence supporting this concept in the context of breast cancer has been provided by studies from several groups (Keller et al., 2012; Molyneux et al., 2010). Recent findings by Santagata and colleagues showed that classification of breast tumors based on cell types found in the normal breast reflect differences in patient survival (Santagata et al., 2014). The clinical implications of these models are profound. However, controversies regarding identity of the normal human mammary stem cells and their relationship with the cells of cancer origin constitute a major roadblock in validating these models and translating their concepts into clinical applications. We argue that definitive experimental proof for the existence and identity of human mammary stem cells cannot be produced with the current experimental tools. In this study we developed a new, complementary approach to address some of these challenges. We utilized theoretical modeling together with plausible combinations of cell fate decisions to examine the distribution of stem cells in different settings, and we compared the predictions of the generated models with our experimental observations. We used a combination of markers with well-defined disposition in the ductal and lobular regions of the mammary tree to compare predictions of the theoretical models with the observations in the human normal breast. Additionally, we utilized proliferation markers, under the assumption that proliferation is more frequent in the compartment of undifferentiated progenitor cells, as is the case in the intestinal and epidermal epithelia (Pinto and Clevers, 2005). Only a subset of growth rules was found to be consistent with the known 1D and 3D patterns of marker expression in normal human mammary lobules. The model, consistent with observed patterns of marker expression, combined the cell fate choices of asymmetric self-renewal, high rate of stem cell division, and distal orientation of the more undifferentiated cell progeny. The outcome was a repetitive gradient of differentiation in the growing lobule, with proliferating cells all along the developing structure and pools of stem cells at branching points and the tip of the developing structure. This repetitive pattern is consistent with a fractal-like, self-reiterative structure of the lobule, indicated by in situ measurements (Russo et al., 1992).