Down-regulation of Wild-type p53-induced Phosphatase 1 (Wip1) Plays a Critical Role in Regulating Several p53-dependent Functions in Premature Senescent Tumor Cells*

Background: Wip1 is a phosphatase involved in DNA-damage response. Results: Wip1 expression is down-regulated in premature senescent cancer cells. Failure to down-regulate Wip1 expression results in cell death and polyploidy. Conclusion: Wip1 down-regulation is important for maintenance of permanent cell cycle arrest in premature senescent tumor cells. Significance: These findings improve our understanding of the mechanism by which Wip1 promotes tumor progression. Premature or drug-induced senescence is a major cellular response to chemotherapy in solid tumors. The senescent phenotype develops slowly and is associated with chronic DNA damage response. We found that expression of wild-type p53-induced phosphatase 1 (Wip1) is markedly down-regulated during persistent DNA damage and after drug release during the acquisition of the senescent phenotype in carcinoma cells. We demonstrate that down-regulation of Wip1 is required for maintenance of permanent G2 arrest. In fact, we show that forced expression of Wip1 in premature senescent tumor cells induces inappropriate re-initiation of mitosis, uncontrolled polyploid progression, and cell death by mitotic failure. Most of the effects of Wip1 may be attributed to its ability to dephosphorylate p53 at Ser15 and to inhibit DNA damage response. However, we also uncover a regulatory pathway whereby suppression of p53 Ser15 phosphorylation is associated with enhanced phosphorylation at Ser46, increased p53 protein levels, and induction of Noxa expression. On the whole, our data indicate that down-regulation of Wip1 expression during premature senescence plays a pivotal role in regulating several p53-dependent aspects of the senescent phenotype.

Premature or drug-induced senescence is a major cellular response to chemotherapy in solid tumors. The senescent phenotype develops slowly and is associated with chronic DNA damage response. We found that expression of wild-type p53induced phosphatase 1 (Wip1) is markedly down-regulated during persistent DNA damage and after drug release during the acquisition of the senescent phenotype in carcinoma cells. We demonstrate that down-regulation of Wip1 is required for maintenance of permanent G 2 arrest. In fact, we show that forced expression of Wip1 in premature senescent tumor cells induces inappropriate re-initiation of mitosis, uncontrolled polyploid progression, and cell death by mitotic failure. Most of the effects of Wip1 may be attributed to its ability to dephosphorylate p53 at Ser 15 and to inhibit DNA damage response. However, we also uncover a regulatory pathway whereby suppression of p53 Ser 15 phosphorylation is associated with enhanced phosphorylation at Ser 46 , increased p53 protein levels, and induction of Noxa expression. On the whole, our data indicate that downregulation of Wip1 expression during premature senescence plays a pivotal role in regulating several p53-dependent aspects of the senescent phenotype.
Cellular senescence was initially described as a growth-arrest program that limits the lifespan of normal mammalian cells (1). Additional work has demonstrated that, apart from aging, various physiologic stresses induce a rapid onset of cell senescence, often referred to as "stress-induced" or "premature senescence." Such stresses include oncogene activation, DNA-damaging agents, and several different stimuli (2). Although cancer cells bypass both replicative and oncogene-induced senescence, some senescence pathways remain intact and can be reactivated by expression of critical regulators (3,4). In particular, treatment with sublethal concentrations of conventional DNA-damaging anticancer agents readily induces premature senescence in cancer cells (5). Premature senescence is a major cellular response to chemotherapy and radiotherapy in solid tumors (6,7), and an intact senescence pathway critically contributes to the success of chemotherapy (8). Hence, the induction of senescence has been proposed as a potential strategy for therapeutic intervention in cancer (9,10). Senescent cells remain viable, metabolically active, and acquire a complex phenotype, including distinctive morphological alterations (flat and enlarged morphology), the expression of acidic ␤-galactosidase, and a specific increase in the secretion of cytokines, chemokines, and other factors, which has been termed the "senescence-associated secretory phenotype" or SASP 2 (11,12). p53 has a well defined role in establishment and maintenance of growth arrest during senescence (13,14). More recently, it has also been reported that p53 negatively modulates the SASP (11). As a result, cells lacking p53 secrete higher levels of several SASP components (11). PPM1D or Wip1 (wildtype p53-induced phosphatase) belongs to the Ser/Thr PP2C family of phosphatases (15). Members of this evolutionarily conserved family are frequently involved in the regulation of cellular stress responses (15,16). Wip1 is both a direct transcriptional target of p53 and an important negative regulator of p53, thus creating a negative regulatory feedback loop (17,18). Wip1 also dephosphorylates several other DNA damage-responsive proteins, such as ATM, ATR, Chk1, Chk2, and p38MAPK (19).
In this work we investigated the role of Wip1 phosphatase in premature senescence in cancer cells. Here we show that Wip1 levels are decreased in chemotherapy-induced senescence. We demonstrate that down-regulation of Wip1 is required for maintenance of permanent G 2 arrest. Accordingly, forced expression of Wip1 in premature senescent tumor cells suppresses phosphorylation of p53 at serine 15 and induces inappropriate re-initiation of mitosis, uncontrolled polyploid progression, and cell death by mitotic failure. Interestingly, dephosphorylation of p53 at Ser 15 is associated with enhanced phosphorylation at Ser 46 and induction of Noxa gene expression. Finally, premature senescent cells forced to express Wip1 develop an amplified SASP.

EXPERIMENTAL PROCEDURES
Cell Culture and Biological Reagents-A549 cells were obtained from American Type Culture Collection and cultured according to its instructions. MCF-7 cells were cultured in DMEM. All media were supplemented with 10% fetal bovine serum. The cell culture media and reagents were purchased from Invitrogen. Doxorubicin (Merck) was dissolved in sterile water. Bleomycin sulfate (Merck) was dissolved in sterile water. Pifithrin-␣ (Sigma) was dissolved in DMSO. z-VAD-fmk (Merck) was dissolved in DMSO.
Induction of Premature Senescence, Senescence-associated ␤-Galactosidase Activity, and Micronuclei Detection-Unless otherwise stated, senescence was induced by treating cells with the DNA-damaging agents doxorubicin (200 nM for MCF-7 cells and 600 nM for A549 cells) for 72 h as previously described (20). Because forced Wip1 expression leads to death of premature senescent cells, all the experiments were conducted from 1 to 7 days after drug release. Staining for acidic ␤-galactosidase was performed as previously described (21). Micronuclei were detected in senescent cells by phase-contrast microscopy. For each sample at least 300 nuclei were analyzed.
Lentiviral Constructs and Infection of Cells-The Wip1 and the Wip1-FLAG plasmids were generously provided by Prof. Galit Lahav and Prof. Xiongbin Lu, respectively. Wip1-FLAG cDNA was subcloned into pWPT lentiviral vector at BamHI site. The construct was sequenced to confirm correct DNA sequence and orientation. Subconfluent 293T lentivirus packaging cells were cotransfected with either pWPT-GFP or Wip1-FLAG-pWPT and pMD2G and pCMV-R8.91 by calcium phosphate precipitation. After 24 h the medium was changed, and supernatant was harvested after 48 and 72 h. Lentiviral supernatant, cleared of cell debris, was concentrated by centrifugation for 90 min at 23,000 rpm.
For transduction, MCF7 and A549 cells were plated on 12-well plates and infected with lentiviruses in the presence of 10% fetal bovine serum and 8 g/ml Polybrene (hexadimethrine bromide; Sigma). After 48 h of incubation, infection efficiency was determined by analyzing GFP expression by flow cytometry.
Treatment with siRNAs-MCF7 cells were treated with 200 nM doxorubicin for 72 h, extensively washed, and released in drug-free media. Cells were transfected with either control siRNA (Dharmacon, ON-TARGETplus Non-targeting pool) or with Wip1 siRNA (Sigma MISSION esiRNA #EHU009271) using Oligofectamine (Invitrogen) according to the manufacturer's protocol.
Cell Cycle Analysis-Cells were fixed with 70% ethanol in PBS and routinely kept at Ϫ20°C overnight. Cells were washed twice with PBS, resuspended in PBS, 40 g/ml propidium iodide (Sigma), and 50 g/ml RNase DNase-free (Roche Applied Science), and incubated at room temperature for 20 min. Cells were analyzed using a CyAn ADP Flow Cytometer (Beckman Coulter, Inc., Milano, Italy) and Summit Software.
Measurement of Mitochondrial Membrane Potential-Measurements of mitochondrial membrane potential were performed by staining cells with tetramethylrhodamine-ethyl ester (TMRE) (Invitrogen). A TMRE stock was prepared at a concentration of 10 mM in DMSO and stored at Ϫ20°C. Cells were loaded with TMRE by incubating cells in media containing 50 nM TMRE for 30 min at 37°C. After two washes, cells were kept at 4°C and analyzed immediately using a CyAn ADP Flow Cytometer (Beckman Coulter) and Summit Software.
Live/Dead Assay-LIVE/DEAD Fixable Green Dead Cell Stains (Invitrogen) was used according to the manufacturer's instructions to measure the viability of the cells at different times after induction of senescence. Cells were analyzed using a CyAn ADP Flow Cytometer (Beckman Coulter) and Summit Software.
Measurement of Apoptosis and Necrosis-Apoptosis was detected by flow cytometry using PE Annexin V Apoptosis Detection Kit I (BD Biosciences). Briefly, cells were doublestained with annexin V-PE and 7-AAD following the manufacturer's instructions. Early apoptosis is defined by annexin V-positive/7-AAD negative staining, and late apoptosis/necrosis is defined by annexin V-positive/7-AAD positive staining. Cells were analyzed using a CyAn ADP Flow Cytometer (Beckman Coulter) and Summit Software.

Decrease in Wip1 Protein Level during Premature Senescence-
To investigate the role of Wip1 phosphatase in modulating premature senescence in tumor cells, we treated the lung adenocarcinoma cell line A549 with doxorubicin for 72 h as previously described (20). We analyzed Wip1 protein level during drug treatment and thereafter during the development of the senescent phenotype over several days after the initial DNA damage. An increase in Wip1 protein was readily detectable 24 h after the start of the treatment, then followed by a gradual decrease starting from 48h treatment (Fig. 1A). Wip1 is amplified in various human cancers (22)(23)(24)(25). To investigate the regulation of Wip1 in the context of gene amplification, we induced premature senescence in the breast cancer cell line MCF7. MCF7 cells carry an amplified PPM1D/Wip1 gene and overexpress the phosphatase (22). In line with the data obtained in the A549 cell lines, treatment with doxorubicin resulted in increased Wip1 protein at 24 h, then followed by a progressive decrease under the base line (Fig. 1B). The phosphatase was almost undetectable in fully senescent MCF7 cells (Fig. 1B, 15 days). These data were confirmed by immunoblot analysis with a different anti-Wip1 antibody (supplemental Fig. 1). Hence, Wip1 is down-regulated during persistent DNA damage and after drug release during the acquisition of the senescent phenotype. This down-regulation is detected also in cells with deregulated Wip1 gene expression resulting from gene amplification. Next, we induced senescence in all cells by treatment with doxorubicin. A characterization of the senescent cells is illustrated in supplemental Fig. 1. Both GFP-and FLAG-Wip1-expressing cells acquired a fully senescent phenotype, with morphological alterations and acidic ␤-galactosidase staining typical of premature senescent cells (supplemental Fig. 1B). In addition, both GFP-and FLAG-Wip1 cells developed a senescence-associated secretory phenotype (data not shown). It is important to note that, different from endogenous Wip1, elevated levels of transduced FLAG-Wip1 were maintained in both cell lines after induction of senescence ( Fig. 1, C and E). Hence, Wip1 does not prevent drug-induced senescence. However, a gradual loss of cell viability during senescence, assessed by MTT assay, was observed in MWIP1 cells as compared with control cells (MGFP and MCF7 cells) (Fig. 2A). The progressive loss of viability was further confirmed in both senescent MWIP1 and AWIP1 cells by means of a live/dead assay (Fig. 2, B and C).

Effects of Wip1 Overexpression in Premature Senescent
Because senescent cells accumulate persistent DNA damage foci (26) and becauseWip1 directly dephosphorylates several DNA damage response proteins (27), we next analyzed phospho-ATM (Ser1981) and ␥-H2AX foci formation in FLAG-Wip1-expressing lines. As shown in supplemental Fig. 2, whereas persistent P-ATM and ␥-H2AX foci were readily detected in senescent MCF-7 and A549 cells, no foci were observed in FLAG-Wip1-expressing cells.
Wip1 Overexpression Overrides G 2 Phase Arrest and Promotes Mitotic Cell Death-Recent studies revealed a critical role for Wip1 in conferring G 2 checkpoint recovery competence by counteracting p53-dependent transcriptional repres- sion of mitotic regulators (28). Because senescent tumor cells mainly arrest in the G 2 phase of the cell cycle (29) (Fig. 3A), we analyzed the expression of cyclin B1 in proliferating and senescent MGFP, MWIP1, AGFP, and AWIP1 cells. As shown in Fig.  4A, acquisition of the senescent phenotype is accompanied by suppression of cyclin B1 protein in GFP control cells. In contrast, elevated levels of cyclin B1 were detected in FLAG-Wip1expressing cells. In addition, a clear correlation between cyclin B1 levels and pRb phosphorylation status was detected (Fig.  4A). Interestingly, Wip1 protein levels appeared to progressively decline in the senescent cells ( Fig. 4A and data not shown), an effect likely attributable to a selection against Wip1expressing senescent cells. Notably, under the conditions used for routine propagation of the cells, i.e. in the absence of senescence induction, cells maintain a relatively stable level of FLAG-Wip1 expression.
Data in Fig. 4A raise the possibility that down-regulation of Wip1 in premature senescence may be required to inhibit inap-propriate cell cycle re-entry, with unrepaired DNA damage. Indeed, flow cytometric analyses of histone H3 phosphorylation at serine 10 revealed that a significant subset of FLAG-Wip1 senescent cells progress from G 2 into mitosis (Fig. 4B). Furthermore, a significant polyploid cell fraction, characterized by a Ͼ4N DNA content, appeared in senescent cells forced to express Wip1 (Fig. 3B). We employed 5-bromo-2-deoxyuridine (BrdU) incorporation as a measure of the numbers of cells engaging in DNA synthesis. As shown in Fig. 4C, a significant fraction of G 2 -arrested cells, forced to express Wip1, enter S phase, likely resulting in polyploidy.
To determine the role of endogenously overexpressed Wip1 in the regulation of p53 functions in premature senescent tumor cells, we silenced Wip1 expression in senescent MCF7 cells using RNA interference. As shown in Fig. 5A, treatment with Wip1-specific siRNA reduced the expression of Wip1 in senescent cells at all concentrations tested. Inhibition of Wip1 increased the phosphorylation of p53 at serine 15 in a dose-dependent manner. Next, we analyzed the ability of endogenous Wip1 to counteract the p53-dependent prolonged G 2 arrest. Senescent MCF7 cells were transfected with control siRNA or Wip1 siRNA and analyzed for the expression of cyclin B1 and for polyploid progression. In line with the increased phosphorylation and activation of p53, treatment with Wip1-specific siRNA resulted in down-regulation of cyclin B1 in the senescent cells (Fig. 5B). More importantly, silencing of endogenous Wip1 resulted in a significant decrease in the frequency of polyploid cells (8N) (Fig. 5C).
On the whole, these data suggest that, by suppressing both ATM (supplemental Fig. 2) and p53 phosphorylation, Wip1 induces inappropriate re-initiation of mitosis from G 2 phase, uncontrolled polyploid progression, and cell death by mitotic failure.
Mitotic catastrophe is characterized by the occurrence of aberrant mitosis, resulting in the accumulation of large cells with several micronuclei (30). Accordingly, we observed a significant increase in the number of micronucleated senescent cells when Wip1 was constitutively expressed (Fig. 6). Cells undergoing mitotic catastrophe can die by either apoptosis or necrosis (31). Hence, we induced senescence in all cells and analyzed cell death by annexin V/7-AAD staining. As shown in Fig. 7, in both cell lines, forced expression of Wip1 induced a significant increase in both early apoptotic (annexin V-positive, 7-AAD-negative) and late apoptotic/necrotic cells (annexin V-positive, 7-AAD-positive). Treatment with the pan-caspase inhibitor z-VAD-fmk significantly reduced apoptosis in both MWIP1 and AWIP1 cells. Interestingly, z-VAD-FMK also partially inhibited late apoptosis in MWIP1 senescent cells (Fig.  7A). Finally, Western blot analyses showed both caspase-9 activation and PARP-1 cleavage in senescent MWIP1 cells (supplemental Fig. 3). In contrast we did not observe activation of either caspase-9 or caspase-8 or caspase-3 or PARP cleavage in senescent AWIP1 cells undergoing mitotic catastrophe (data not shown). These data suggest that both apoptosis and necrosis are induced in senescent cells forced to express Wip1.
Wip1 Overexpression Affects p53 Phosphorylation Status-To get more insight into the ability of Wip1 to cause cell death in premature senescent tumor cells, we examined p53 and p21 CIP1 protein levels in proliferating and senescent cells. Both AWIP1 and MWIP1 senescent cells showed increased levels of p53 and reduced amounts of p21 CIP1 proteins as compared with control cells (Fig. 8A and supplemental Fig. 4B). In addition, p53 was found to accumulate not only in the nucleus but also in the cytoplasm of FLAG-Wip1-expressing cells (supplemental Fig. 4, C and D  increased levels of p53 and the accumulated p53 protein was not phosphorylated at Ser 15 , we decided to further investigate p53 post-translational modifications in FLAG-Wip1-expressing cells. First, we used phage -phosphatase to analyze the phosphorylation status of p53 in senescent A549, AGFP, and AWIP1 cells. Both in controls (A549 and AGFP) and in AWIP1 cells, a pronounced phosphatase-dependent shift in p53 electrophoretic mobility was observed, indicating that in premature senescent tumor cells p53 is phosphorylated, even in the presence of constitutively active FLAG-Wip1 (supplemental Fig. 4F). Phosphorylation of p53 at Ser 46 regulates the ability of p53 to induce apoptosis (35,36). Hence, we analyzed p53 phosphorylation at Ser 46 during the development of the senescent phenotype in AGFP and AWIP1 cells. It was found that phosphorylation at Ser 46 is selectively induced in senescent cells forced to express Wip1 (Fig. 8B).
To investigate if the increased p53 protein in these cells was functional and to investigate the role of Ser 46 phosphorylation in regulating the transcription of pro-apoptotic genes in premature senescent tumor cells, we analyzed the expression of a number of proapoptotic p53 target genes (such as Puma, Noxa, Perp, Pig3, Apaf1). Of the genes tested, Noxa (PMAIP1) was significantly induced in both senescent MWIP1 and AWIP1 cells, as compared with control cells (Fig. 8, C and D, and data not shown). During apoptosis, Noxa induces mitochondrial dysfunction (37). Hence, we measured the changes in mitochondrial membrane potential (⌬⌿m) by staining with TMRE. As shown in Fig. 9, A and B, forced Wip1 expression induces premature senescent MCF7 cells to undergo mitochondrial depolarization. The same results were obtained in A549 cells (data not shown). In addition, we examined the effects of Pifithrin-␣ (PFT-␣), an inhibitor of p53 transactivation (38), on Wip1-dependent up-regulation of Noxa. We first confirmed the ability of PFT-␣ to prevent the up-regulation of p21 CIP1 in proliferating cells treated with doxorubicin (supplemental Fig.  5A). Next, to specifically inhibit pro-apoptotic p53 (i.e. phospho-Ser 46 p53), we treated deep senescent AWIP1 and MWIP1 cells (percentage of cells showing reduced mitochondrial membrane potential Ͼ25%) with PFT-␣ and analyzed Noxa expression by real time PCR. As shown in Fig. 9C, PFT-␣ significantly suppressed Noxa expression and attenuated mitochondrial depolarization in Wip1-overexpressing cells (Fig. 9, D and E).
These data indicate that constitutive Wip1 expression during premature senescence counteracts p53-dependent repression of mitotic genes and compromises permanent G 2 arrest, likely through dephosphorylation of p53 at Ser 15 . However, phosphorylation of different residues, such as Ser 46 , still allows p53 proapoptotic signaling.

DISCUSSION
In this manuscript we present evidence pointing to Wip1 as a critical regulator of cell fate after induction of premature senescence. We demonstrate that Wip1 expression is down-regulated during the acquisition of the senescent phenotype and that this down-regulation is required for a permanent arrest in the cell cycle. Indeed, ectopic expression of Wip1 in premature senescent cells results in an inappropriate re-entry in the cell cycle (via cyclin B1), polyploid progression, or cell death. These effects seem to be mediated by different p53 phosphorylation (decreased Ser 15 and increased Ser 46 phosphorylation) and induction of the proapoptotic gene Noxa.
Wip1 belongs to the conserved PP2C phosphatase family (15), whose members are frequently involved in the regulation of cellular stress responses (15,16). Accordingly, expression of Wip1 is readily induced in response to DNA damage in a p53dependent manner (15). Wip1 dephosphorylates several DNA damage-responsive proteins, such as ATM, ATR, Chk1, Chk2, and p38MAPK (19). Furthermore, Wip1 also dephosphorylates p53 at Ser 15 , thus attenuating the DNA damage response (DDR) (41). Cytotoxic drugs as well as ionizing radiation are able to induce senescence in tumor cells expressing wild-type p53 both in vitro and in vivo (5-7  develop after transient DNA damage but develops slowly, over several days, and is associated with chronic DDR (11). We show that when the DNA damage signal lasts for a long time, i.e. during a persistent DNA damage that induces premature senescence in tumor cells, Wip1 protein is reduced. Interestingly, persistent DNA damage results in Wip1 down-regulation also in MCF-7 cells, which overexpress the phosphatase as a consequence of gene amplification. Repression of Wip1 protein during chronic DDR and in pathological aging has been recently demonstrated in a mouse model of progeria (42). In this model, suppression of Wip1 has been related to miR-29 up-regulation (42). We are currently investigating if a similar mechanism is also responsible for Wip1 down-regulation in our experimental system. To investigate the biological significance of Wip1 down-regulation in premature senescence, we studied the effects of forced expression of Wip1. Wip1 protein levels do not prevent drug-induced senescence; in fact both AWIP1 and MWIP1 cells develop a full senescent phenotype after treatment with doxorubicin. However, our results demonstrate that down-regulation of Wip1 is required for maintenance of permanent G 2 arrest in premature senescent tumor cells. Forced expression of Wip1 suppresses DDR and induces inappropriate re-initiation of mitosis, as demonstrated by analyses of histone H3 phosphorylation. Because Wip1 dephosphorylates most DDR proteins, i.e. ATM, Chk2 (data not shown), and ␥-H2AX, cells enter mitosis with unrepaired DNA and undergo cell death likely by mitotic failure. This effect appears to be tumor cell-specific. In fact, although normal senescent cells arrest with a G 1 DNA content, premature senescent tumor cells characteristically arrest in the G 2 phase of the cell cycle (29), likely due to defective G 1 checkpoint of cancer cells (43). Accordingly, it has been shown that introduction of Wip1 in normal human mesenchymal stem cells (G 1 arrested) allows bypass of senescence and extends cellular life span (44).
Mitotic catastrophe has recently been defined as a mechanism that senses mitotic failure and drives the cells to death (45). In tumor cells, mitotic catastrophe has been associated with compromised G 2 /M checkpoint signaling (46). In our experimental system, forced expression of Wip1 induces dephosphorylation of p53 at Ser 15 and the inappropriate expression of mitotic regulators, which prevents G 2 arrest and triggers mitotic catastrophe. In particular, we show that the induction of premature senescence is associated with suppression of cyclin B1 protein in control cells, whereas elevated levels of cyclin B1 are detected in cells expressing Wip1. the essential role played by Wip1 in regulating p53-dependent transcriptional repression of mitotic regulators (28). Notably, silencing of endogenously overexpressed Wip1 in premature senescent MCF-7 cells results in increased p53 phosphorylation at Ser 15 and decreased cyclin B1 expression. Hence, although expression of Wip1 during acute DNA damage is critical to successfully recover from the arrest (28), down-regulation of Wip1 is required during persistent DNA damage to establish a permanent G 2 /M cell cycle arrest in tumor cells.
Cells that undergo mitotic catastrophe can die by two separate mechanisms, i.e. apoptosis or necrosis, and the cell type appears to determine the final pattern of cell death (30). In our study, forced expression of Wip1 in premature senescent tumor cells induces both apoptosis and necrosis, as estimated by annexin V/7-AAD staining. In MWIP1 senescent cells undergoing mitotic catastrophe caspase-9 was clearly activated (and PARP-1 cleaved). Although we did not observe activation of either caspase-9 or caspase-8 or caspase-3 in senescent AWIP1 cells, treatment with the pan-caspase inhibitor z-VAD-fmk significantly reduced apoptosis in both cell lines, thus suggesting the involvement of other caspases (e.g. caspase-2; Ref. 30) in AWIP1 cells. The involvement of caspases in mitotic catastrophe is controversial and appears to depend on the experimental system (47). Finally, we found that expression of Wip1 induces premature senescent cells to undergo mitochondrial depolarization, and mitochondrial membrane permeabilization may cause "caspase-independent death" (48). It has been demonstrated that tetraploid cells may be generated through catastrophic mitosis followed by mitotic slippage (30). In addition, persistent DNA damage response also induces tetraploidization (49). According with a central role of p53 in preventing polyploidization (50), forced expression of Wip1 in premature senescent tumor cells also induces the accumulation of a significant polyploid cell fraction, characterized by a Ͼ4N DNA content. Hence, expression of Wip1 in premature senescent tumor cells abrogates the G 2 /M checkpoint resulting in cell death by mitotic catastrophe or generating polyploid cells.
Although most of the effects of Wip1 described above may be attributed to its ability to dephosphorylate and to inhibit p53, we uncover a regulatory pathway whereby suppression of p53 Ser 15 phosphorylation was associated with increased p53 protein levels both in the nucleus and in the cytoplasm, enhanced phosphorylation at Ser 46 , and induction of Noxa expression. Phosphorylation of Ser 46 is critical for the induction of proapoptotic genes and is not required for the induction of cell cycle arrest (32). To investigate the role of phosphorylated p53 (Ser 46 ), we inhibited p53-dependent transcription with Pifithrin-␣ (38) in deep senescent tumor cells. Interestingly, treatment of premature senescent AWIP1 and MWIP1 cells with Pifithrin-␣ reduced Noxa expression and abrogated mitochondrial depolarization. Hence, although Wip1 inhibits p53 activation through dephosphorylation at serine 15, the p53-dependent apoptotic signaling is not affected. In addition, cytosolic p53 can induce apoptosis through transactivation-independent mechanisms (51).
Senescent cells secrete growth factors, cytokines, and chemokines that affect both senescent cells and neighboring cells, functioning in an autocrine and/or paracrine mode (39,40). Notably, SASP occurs in vivo in response to DNA-damaging chemotherapy (11). Evidence points to p53 as one of the critical regulators of SASP: p53 restrains the SASP, and cells lacking p53 secrete higher levels of several SASP components (11). Interestingly, in line with the ability of Wip1 to inhibit some critical p53 functions in our cellular system, preliminary data from our laboratory suggest that forced expression of Wip1 results in enhanced expression of a subset of cytokines and chemokines in premature senescent cells (data not shown). The precise role of Wip1 and the mechanisms of this modulation need further investigations.
Wip1 has been shown to be amplified and overexpressed in various human cancers (22)(23)(24)(25). Based on our results, overexpression of Wip1 in cancer cells may result in a selective advantage as cells bypass the G 2 checkpoint, re-enter the cell cycle, and become polyploid, potentially acquiring additional mutation.
Our study shows that deregulated expression of Wip1 during a persistent DNA damage, which induces premature senescence in tumor cells, results in the accumulation of polyploid cells. This activity provides potential mechanisms by which Wip1 may promote tumor progression.