Skip to main content
Open Access Publications from the University of California

Integrating form and function in the mammary gland: crosstalk of extracellular matrix, extracellular matrix degrading enzymes and epithelial cells in 3D cultures and engineered mice

  • Author(s): Acosta, Jamie
  • Advisor(s): Bissell, Mina J
  • Liu, Fenyong
  • et al.

Tissues exist in a highly ordered three dimensional structure that is critical to inform function. When cells are explanted into culture, the tissue-specific functions are often lost or drastically altered suggesting that the lowest functional unit in higher organisms is not a cell but a cell associated with a particular extracellular matrix (ECM) (Bissell, 1981; Bissell and Hall, 1987). Indeed the unit of function in a larger context is the organ itself (Bissell and Hall, 1987). In our studies of the mammary gland, its ECM and ECM-degrading enzymes we have employed culture models that begin to recreate the tissue structure and aspects of the normal mammary gland microenvironment. These three dimensional (3D) culture models provided a platform to ask questions- and elucidate functions- of mammary gland development and mechanisms of tumorigenesis within a 3D structure.

We examined the ultrastructural features of a progression series where cells became tumorigenic outside the human body (Briand et al., 1987; Petersen et al., 1992; Rizki et al., 2008). In 3D culture the non-malignant cells form growth arrested acinus-like structures and the malignant cells form disorganized masses that do not growth arrest. Electron microscopy imaging of the non-malignant S1 acini revealed that the cell line thought to represent highly polarized luminal epithelial cells of the human breast displayed features consistent with a luminal-basal hybrid cell line. The acini had well-organized basal polarity and basement membrane (BM). The BM is a specialized ECM that separates epithelial, endothelial and fat cells from the surrounding connective tissue and functions as a selective barrier and organizer of organs. The acini had partially apically oriented tight junctions (ZO-1 protein) suggestive of less apical polarization. They also displayed irregular microvilli projections, primary cilia and nuclear invaginations characteristic of basal cells. We found that the three dimensional acini were growth arrested and demonstrated polarized distribution of organelles and proteins (Petersen et al., 1992). Overall the features suggested that S1 cells are a hybrid of a luminal and basal cell perhaps representing a progenitor cell of the human breast.

The BM protein laminin-111 has been shown to be necessary for polarization of mammary epithelium (Gudjonsson et al., 2002) and functional differentiation (Streuli et al., 1991). Here, we knocked down the unique chain of laminin-111, laminin α1, in the HMT-3522-S1 and found that polarity and growth arrest were disrupted despite the presence of laminin-111 in the 3D assay. We show endogenous laminin α1 regulates levels of the matrix degrading enzyme MMP9, activation of signaling pathways and secretion of other ECM proteins. We have shown previously that modulating malignant pathways reverts the malignant phenotype (Weaver et al., 1997; Wang et al., 2002; Beliveau et al., 2010). Expression of the laminin α1 protein in the syngeneic tumor cell line, T4, revealed that cells transfected with the LAMA1 construct phenotypically revert in culture resembling non-malignant HMT-3522-S1 cells. The results demonstrate that deposition of endogenous BM and proper remodeling of the BM is critical to generating a microenvironment that fosters growth arrest, polarization and tissue-specific gene expression.

We examined the role of laminin α1 in mouse mammary gland development and found extracellular laminin α1 is not necessary for mammary gland development and function. The mammary gland of mice where laminin α1 is knocked out of mammary epithelium develops normally, branches and extends into the fat pad at the same rate as control littermates. Female laminin α1 knockout mice undergo complete lobuloalveolar development and are able to feed their entire litters. Our findings pertain only to extracellular laminin α1 as we present evidence of intracellular laminin α1 transcribed from the LAMA1 gene downstream of the knockout in the mammary epithelial cells of laminin α1 knockout mice.

We employed a novel 3D model of mammary gland branching morphogenesis (Nelson et al., 2006; Nelson et al., 2008) as well as a transgenic mouse model (Yana et al., 2007) to ask how the matrix degrading enzyme MMP14 functions in branching morphogenesis of the mammary gland. We found that MMP14 is highly expressed at the terminal end buds and branching ducts of the developing mammary gland in vivo (Mori et al., 2013). In our micromolded collagen assay, MMP14 mediates sorting of cells within the model tissue. Sorting occurs through differential cellular motility such that the subpopulation highest for MMP14 expression segregates to the ends of the tubules where branching originates. MMP14-driven sorting is independent of its catalytic activity and required the hemopexin domain. We describe a signaling cascade through Rho kinase that allows cells to sort. The results indicate that differential directional persistence can give rise to patterns within model developing tissues.

Main Content
Current View