The adult mammary gland is comprised of a bi-layered epithelium of luminal and myoepithelial cells surrounded by an adipocyte-rich fat pad, a highly collagenous extra-cellular matrix (ECM) and a number of other stromal and endothelial cell types. Mammary stem cells (MaSCs) reside within the epithelium and these are capable of repopulating a mammary fat pad that is devoid of epithelium, upon transplantation. It was sought to recapitulate this process of MaSCs repopulating a fat pad using a synthetic fat pad, engineered from a collagen scaffold invested with adipocytes, to provide an in vitro 3D model. Fluorescently tagged murine Axin2-expressing cells were obtained from transgenic mice and seeded into these scaffolds and cultured, mimicking the process of fat pad repopulation. Immunohistochemical analysis demonstrated that Axin2+ myoepithelial cells were rarely capable of forming bi-layered structures that expressed correct myoepithelial localisation and resemblance to a luminal morphology.

Breast tumours surrounded by anisotropic (directional) collagen fibres running perpendicular to the tumour boundary are more aggressive and associated with poor patient prognosis. To recapitulate this anisotropic collagen phenotype in vitro, an ice-templating technique was used to modify the structure of the collagen scaffolds producing both non-directional (isotropic) and anisotropic internal architectures. Tumour cells from various breast cancer cell lines were seeded into both isotropic and anisotropic scaffolds to investigate whether this approach could distinguish cell type-specific migratory ability and whether anisotropy affected migration efficiency. Following analysis by confocal microscopy and ImageJ, anisotropic scaffolds were observed to enhance the migratory potential of MDA-MB-231 breast cancer cells. These results highlight the importance of collagen alignment and provide a reproducible method to quantitatively measure cell migration in 3D for cells derived from different breast cancer subtypes.

Building on these data, the protocol was adapted to permit the direct investigation of tumour biopsy material. Given the heterogeneity of breast tumours, it was considered important to maintain tumour architecture and stromal components. Thus, murine mammary tumour fragments from two different established mammary cancer models were utilised and cultured in anisotropic collagen scaffolds in the presence or absence of adipocytes to allow an investigation of their influence on tumour cell migration. Further experiments included addition of various therapeutic drugs followed by immunofluorescence microscopy coupled with an optical clearing technique. These data demonstrated the utility of the model in determining both the rate and capacity of tumour cells to migrate through the engineered stroma while shedding light also on the mode of migration. Moreover, the response of different mammary tumour types to chemotherapeutic drugs could be could be readily quantified.

To humanize the fat pad for subsequent human tissue analysis, human mesenchymal stem cells (MSC) were obtained from reduction mammoplasties and immortalised, before differentiating them into adipocytes within anisotropic collagen scaffolds. Human breast cancer cells were fluorescently tagged for tracking using lentiviral methods and were seeded into scaffolds invested with differentiated MSCs. Both cell types were successfully co-cultured for 7 days and imaged using multiphoton methods.