Abstract
Mammary MCF-10A cells seeded on reconstituted basement membrane form spherical structures with a hollow central lumen, termed acini, which are a physiologically relevant model of mammary morphogenesis. Bcl-2-associated athanogene 1 (Bag-1) is a multifunctional protein overexpressed in breast cancer and ductal carcinoma in situ. When present in the nucleus Bag-1 is predictive of clinical outcome in breast cancer. Bag-1 exists as three main isoforms, which are produced by alternative translation initiation from a single mRNA. The long isoform of Bag-1, Bag-1L, contains a nuclear localisation sequence not present in the other isoforms. When present in the nucleus Bag-1L, but not the other Bag-1 isoforms, can interact with and modulate the activities of estrogen-, androgen- and vitamin D-receptors. Overexpression of Bag-1 mRNA in MCF-10A is known to produce acini with luminal filling reminiscent of ductal carcinoma in situ. As this mRNA predominantly overexpresses the short isoform of Bag-1, Bag-1S, we set out to examine whether the nuclear Bag-1L isoform is sufficient to drive premalignant change by developing a Bag-1L-overexpressing MCF-10A model. Two clones differentially overexpressing Bag-1L were grown in two-dimensional (2D) and three-dimensional (3D) cultures and compared with an established model of HER2-driven transformation. In 2D cultures, Bag-1L overexpression reduced proliferation but did not affect growth factor responsiveness or clonogenicity. Acini formed by Bag-1L-overexpressing cells exhibited reduced luminal clearing when compared with controls. An abnormal branching morphology was also observed which correlated with the level of Bag-1L overexpression, suggesting further malignant change. Treatment with Thio-2, a small-molecule inhibitor of Bag-1, reduced the level of branching. In summary, 3D cultures of MCF-10A mammary epithelial cells overexpressing Bag-1L demonstrate a premalignant phenotype with features of ductal carcinoma in situ. Using this model to test the small-molecule Bag-1 inhibitor, Thio-2, reveals its potential to reverse the atypical branched morphology of acini that characterizes this premalignant change.