Cellular senescence is a state of permanent replicative arrest that is linked to telomere erosion and dysfunction that engages at least two mechanisms, the p53 and the p16INK4a -pRB pathways (1, 2). In cancer cells, the presence of oncogenic mutations, chemotherapeutic drugs, and oxidative stress can cause an acutely inducible, telomere-independent, stress-responsive form of cellular senescence, termed premature senescence (PS) 2 (3, 4). PS is considered a physiologic mechanism of DNA damage response occurring in chemotherapy (5-7), and the senescence induction could be an effective in vivo mechanism to limit tumor progression by preventing cancer cell proliferation or by blocking the cells at risk of neoplastic transformation (8). However, the physiological consequences of prematurely induced senescent (PIS) cancer cells remain elusive. PIS cancer cells have been shown to promote the growth of neighboring cells, and they are intrinsically resistant to chemotherapeutic agents (5, 9, 10). Importantly, cells in prematurely senescent tumors are capable of escaping growth arrest and re-entering the cell cycle, leading to tumor relapse (5, 9, 11). As for the mechanism underlying escape from DNA damage-induced senescence, overexpression of the cyclin-dependent kinase Cdc2 has been found in clones that bypassed replicative arrest in human non-small cell H1299 carcinoma (5) and in MCF-7 breast cancer cells (9). It was recently shown that survivin is the immediate downstream effector of Cdc2/Cdk1 and that phosphorylated survivin is necessary for the escape of senescent cells (12). Moreover, Twist1, which is involved in the metastatic dissemination of cancer cells, was shown to override oncogeneinduced senescence by abrogating cell cycle inhibition by p21 and p16 (11, 13), leading to complete epithelial-mesenchymal transition and implicating a direct link between escape from senescence and the acquisition of invasive features by cancer (11, 13). These data collectively suggest that there exist mechanisms that foster survival of PIS cancer cells and promote escape of these cells from the senescent state, which are likely detrimental to the overall therapeutic efficacy of cancer treatment. Because the p53 gene is frequently inactivated in 50% of * This work was supported, in whole or in part, by National Institutes of Health Grants R01 ES015323 (to M. A.) and K01-AR048582 and R03 CA125855 (to A. L. K.). This work was also supported by an Irving Scholar Award and Skin Cancer Foundation grant (to A. L. K. 2 The abbreviations used are: PS, premature senescence; PIS, prematurely induced senescent; ATM, ataxia telangiectasia, mutated; ATR, ATM and Rad3-related; Bfl-1/A1, Bcl2-related protein A1; Chk1/2, checkpoint kinase 1/2; DOX, doxorubicin; DYRK, the dual specificity, tyrosine phosphorylation-regulated kinase; MBP, maltose-binding protein; mTOR, mammalian target of rapamycin; mTORC1, the rapamycin-sensitive mTOR complex