Discussion While p21 was initially characterized as an important cell cycle inhibitor, recent studies suggest that cytoplasmic p21 has anti apoptotic LDP-341 and actin cytoskeleton regulatory functions. The accumulation of cytoplasmic p21 is associated with Ras and HER2/neu activated tumorigenic transformation. Moreover, overex pression of p21 is associated with poor prognosis of many types of cancer. However, the function of p21 in breast cancer has not been established. In our study, we assessed p21 levels with clinical outcomes in breast can cer patients. High p21 expression correlates with poor overall survival and distant metastasis free survival. Furthermore, using an in vivo model of mammary fat pad transplantation of metastatic human breast cancer cells in mice, we showed that while silencing p21 gene expres sion did not affect the primary tumor formation, it potently prevented primary tumor cells to invade into surrounding tissues.

Together, our results provide evi dence of a tumor promoting role for p21 in primary tumor local invasion. Previous studies have indicated that during breast can cer progression, TGFb cytostatic responses are lost while pro migratory and pro invasive effects are maintained. Here, we found that all invasive breast cancer cell lines tested were resistant to growth inhibition by TGFb and that while TGFb did not induce any change in p15 or c myc expression levels, it strongly up regulated p21 expression arguing that in advanced breast cancer p21 functions independently of cell cycle regulation.

This is in contrast to the effect observed in human immortalized the keratinocyte cell line HaCaT, where TGFb mediated p21 gene expression leads to cell cycle arrest. Indeed, we found that the induction of p21 in invasive breast cancer cells is required for the pro migratory and pro TGFB invasive effects of TGFb. In accordance with these results, depletion of p21 did not modulate primary tumor growth in vivo but strongly blocked tumor invasion capa city. These findings together support the notion of Smad4 a direct oncogenic role for p21 in breast cancer progression. We further report that the TGFb mediated increase in p21 expression is Smad dependent and Smad3 spe cific. This is interesting in light of previous reports indicating that overexpression of a dominant negative form of Smad3 reduced the ability of cancer cells to metastasize and that Smad3, but not Smad2, pro motes breast cancer metastasis in mice.

Further more, while Smad2 mutations in cancer have been described, no mutations in Smad3 or p21 have yet been reported. Together these data suggest that in breast cancer Smad3 pro invasive functions are mediated by p21 and that targeting p21 may prove useful to improve the clinical course of metastatic patients. Tumor cell migration and invasion are critical initiation Cilengitide steps in the process of breast cancer metastasis.