Oncogene-induced senescence (OIS), the proliferative arrest engaged in response to persistent oncogene activation, serves as an important tumor-suppressive barrier. We show here that finite lifespan human mammary epithelial cells (HMEC) undergo a p16/ RB-and p53-independent OIS in response to oncogenic RAS that requires TGF-β signaling. Suppression of TGF-β signaling by expression of a dominant-negative TGF-β type II receptor, use of a TGF-β type I receptor inhibitor, or ectopic expression of MYC permitted continued proliferation upon RAS expression. Surprisingly, unlike fibroblasts, shRNA-mediated knockdown of ATM or CHK2 was unable to prevent RAS-mediated OIS, arguing that the DNA damage response is not required for OIS in HMEC. Abrogation of TGF-β signaling not only allowed HMEC lacking p53 to tolerate oncogenic RAS but also conferred the capacity for anchorage-independent growth. Thus, the OIS engaged after dysregulated RAS expression provides an early barrier to malignant progression and is mediated by TGF-β receptor activation in HMEC. Understanding the mechanisms that initiate and maintain OIS in epithelial cells may provide a foundation for future therapies aimed at reengaging this proliferative barrier as a cancer therapy. (ii) constitutive growth signaling, (iii) unlimited replication potential, and (iv) invasive potential (1). Early studies using normal mouse cells indicated that a limited set of genetic manipulations could confer neoplastic potential (2). However, normal human cells have been more difficult to transform to malignancy, indicative of their more stringent tumor-suppressive pathways. Extensive study of cultured human mammary epithelial cells (HMEC) has identified two senescence barriers. One involves the stress-associated induction of the cyclin-dependent kinase inhibitor p16 before attaining critically short telomeres. This stasis barrier can be overcome by inhibiting p16, allowing continued proliferation, which results in agonescence, a proliferative barrier mediated by telomere depletion (3). Additionally, the ability of dysregulated oncogenic signaling to induce senescence in human cells has implicated oncogene-induced senescence (OIS) as an important tumor-suppressive barrier. A number of recent studies have demonstrated the physiological relevance of OIS in human tumorigenesis and in vivo tumor mouse models (4). Additionally, the presence of senescent cells in benign but not advanced tumors argues that OIS serves as an early tumorsuppressive barrier that needs to be dismantled for full oncogenic progression (4). In human fibroblasts, OIS could be bypassed by disabling p16 or molecular components of the DNA damage response (DDR), including ATM, CHK2, or p53, before RAS, MOS, or STAT5 overexpression (5-9). However, OIS in HMEC has been shown to be independent of p53 and the p16-RB pathway after oncogenic RAF-1 expression (10). The contrasting responses between epithelial and fibroblast cells argue that the signaling networks responsible for OIS have tissue specificity. Indeed, fibrob...