Staged stromal extracellular 3D matrices differentially regulate breast cancer cell responses through PI3K and beta1-integrins

Early and late three-dimensional matrices were produced by NIH-3T3 and tumor-associated murine fibroblasts, respectively. After removing fibroblasts, extracted matrices were re-cultured with breast epithelial cells of assorted characteristics: MCF-10A (non-tumorigenic), MCF-7 (tumorigenic, non-invasive), and MDA-MB-231 (tumorigenic, invasive). Effects prompted by staged matrices on epithelial cell's growth, morphology and invasion were determined. Also, matrix-induced velocity, directionality and relative track orientation of invasive cells were assessed in the presence or absence of inhibitors of phosphoinositide-3 kinase (PI3K) and/or beta-1 integrin.We observed that assorted breast epithelial cells reacted differently to two-dimensional vs. staged, control (early) and tumor-associated (late), three-dimensional matrices. MCF-10A had a proliferative advantage on two-dimensional substrates while MCF-7 and MDA-MB-231 showed no difference. MCF-10A and MCF-7 formed morphologically distinguishable aggregates within three-dimensional matrices, while MDA-MB-231 exhibited increased spindle-shape morphologies and directional movements within three-dimensional matrices. Furthermore, MDA-MB-231 acquired a pattern of parallel oriented organization within tumor-associated, but not control matrices. Moreover, tumor-associated matrices induced PI3K and beta1-integrin dependent Akt/PKB activity in MDA-MB-231 cells. Interestingly, beta1-integrin (but not PI3K) regulated tumor-associated matrix-induced mesenchymal invasion which, when inhibited, resulted in a change of invasive strategy rather than impeding invasion altogether.We propose that both cells and matrices are important to promote effective breast cancer cell invasion through three-dimensional matrices and that beta1-integrin inhibition is not necessarily sufficient to block tumor-matrix induced breast cancer cell invasion. Additionally, we believe that characterizing stroma staging (e.g., early vs. late or tumor-associate