Wip1 promotes RUNX2-dependent apoptosis inp53-negative tumors and protects normal tissuesduring treatment with anticancer agentsAnastasia R. Goloudinaa, Kan Tanoueb, Arlette Hammanna, Eric Fourmauxa, Xavier Le Guezennecc, Dmitry V. Bulavinc,Sharlyn J. Mazurb, Ettore Appellab, Carmen Garridoa,d,e, and Oleg N. Demidova,d,1

aInstitut National de la Santé et de la Recherche Médicale, Unité Mixte de Recherche 866, University of Burgundy, 21078 Dijon, France; bLaboratory of CellBiology, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892; cInstitute of Molecular and Cell Biology, Cell Cycle Control andTumorigenesis Group, Singapore 138673; dFaculty of Medicine and Pharmacy, University of Burgundy, 21078 Dijon, France; and eCentre HospitalierUniversitaire Dijon, 21000 Dijon, France

Edited by Carol Prives, Columbia University, New York, NY, and approved October 17, 2011 (received for review May 11, 2011)

The inactivation of the p53 tumor suppressor pathway in manycancers often increases their resistance to anticancer therapy. Herewe show that a previously proposed strategy directed to Wip1inhibition could be ineffective in tumors lacking p53. On thecontrary, Wip1 overexpression sensitized these tumors to chemo-therapeutic agents. This effect was mediated through interactionbetween Wip1 and RUNX2 that resulted, in response to anticancertreatment, in RUNX2-dependent transcriptional induction of theproapoptotic Bax protein. The potentiating effects of Wip1 over-expression on chemotherapeutic agents were directed only to tu-mor cells lacking p53. The overexpression of Wip1 in normal tissuesprovided protection from cisplatin-induced apoptosis through de-creased strength of upstream signaling to p53. Thus, Wip1 phos-phatase promotes apoptosis in p53-negative tumors and protectsnormal tissues during treatment with anticancer agents.

caspases | Bcl-2 family | dephosphorylation | intestine

The tumor suppressor p53 is a transcriptional factor that is ac-tivated by various stresses and initiates cell cycle arrest, DNA

repair, and apoptosis (1). The importance of p53 in suppressingtumor initiation and growth is indicated by the fact that more thanhalf of all human cancers lose p53 function through mutation ordepletion of the p53 gene (2). The consequences of p53 loss aremultiple; not only does it influence growth, genomic stability, andother tumor characteristics, but it also affects the efficacy of an-ticancer treatment.Alternatively, tumor cells could use negative regulators of p53 to

suppress its activity. For example, a significant number of humancancers exhibitMdm2 gene amplification and/or overexpression (3).Mdm2 gene encodes an E3 ubiquitin ligase that interacts with p53and mediates p53 proteasomal degradation. P53 activates Mdm2expression, forming a negative feedback regulatory loop (4, 5).Another feedback mechanism in the p53 pathway is connected

to its posttranslational modification. After stress, p53 undergoesvarious posttranslational modifications that increase p53 stabilityand/or potentiate its transcriptional activity (6). Among the geneswhose expression is regulated by p53 is the PP2C serine-threoninephosphatase Wip1 (gene name PPM1D). Similarly to Mdm2, theWip1 gene (PPM1D) is amplified in many tumor types (7, 8). Wip1overexpression leads to dephosphorylation of crucial activatingphosphoserines or phosphothreonines found in the DNA damageresponse protein kinases ATM, Chk2, and Chk1, among others.DNA damage response kinases are important transducers of sig-nals from damaged DNA to p53. The inhibition of the ATM/Chk2kinase cascade by Wip1 prevents p53 activation (9–11). Addition-ally, it was shown that phospho-Ser15 in p53, which is important forp53 activity, could be directly dephosphorylated by Wip1 (12).Wip1-KO mice exhibit a tumor-suppressive phenotype in severalcancer models such as lymphomagenesis and mammary gland orintestinal tumorigenesis (13–16). Therefore, Wip1 inhibition may

have clinical implications given the potential therapeutic uses ofcompounds that interfere with Wip1 suppression of p53 activity.Indeed, expression of a Wip1 antisense transcript in the breastcarcinoma cell line MCF-7 resulted in increased p53-dependentapoptosis (17). Several studies in mice have shown that the tumor-suppressive phenotype of Wip1 deletion_is p53-dependent andthat p53 loss completely reverses the effect in these systems. Thus,Wip1 inhibition is probably only effective in tumors with preservedWT p53.Currently, the majority of anticancer therapies use the p53

pathway to induce tumor cells death (18). Following activation byanticancer drugs, p53 induces the expression of numerous genes,including proapoptotic genes such as Bax (19). Elevated expres-sion of proapoptotic genes leads to initiation of the apoptoticprogram and eventually to cell death. In contrast, tumors bearingmutant p53 often exhibit resistance to anticancer drugs (20, 21).Effective therapy of tumors with inactive p53 continues to presenta challenge for modern oncology. To overcome this issue, severalstrategies have been proposed. For example, inactivation of Chk1kinase in p53-negative tumors compromises G2 arrest in responseto anticancer therapy and induces mitotic catastrophe, eliminatingtumor cells (22). Unfortunately, Chk1 inhibition could be highlytoxic to normal tissues and may induce severe side effects (23).Here we report an alternative approach, based on activation of

Wip1 phosphatase, toward sensitizing tumors containing inactivep53 to anticancer drugs, while at the same time protecting normaltissues. In response to anticancer drugs in tumors with inactive p53,Wip1 overexpression led to the induction of Bax through de-phosphorylation-dependent activation of the transcription factorRUNX2. In normal tissues, Wip1 suppressed p53 hyperactivationin response to anticancer therapy, thereby decreasing normal tissuedamage. Thus, in tumors lacking functional p53, Wip1 acts asa sensitization factor to anticancer drugs while protecting normaltissues bearing WT p53 from side effects of anticancer therapies.

ResultsWip1 Overexpression Increases Sensitivity of Tumor Cells withInactive p53 to Anticancer Drugs. Wip1 inhibition was proposed asa novel anticancer strategy directed to nongenotoxic activation of

Author contributions: A.R.G. and O.N.D. designed research; A.R.G., K.T., A.H., E.F., andO.N.D. performed research; X.L.G., D.V.B., and E.A. contributed new reagents/analytictools; A.R.G., K.T., D.V.B., S.J.M., E.A., C.G., and O.N.D. analyzed data; and A.R.G., S.J.M.,E.A., C.G., and O.N.D. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. E-mail: oleg.demidov@u-bourgogne.fr.

See Author Summary on page 361.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1107017108/-/DCSupplemental.

E68–E75 | PNAS | January 10, 2012 | vol. 109 | no. 2 www.pnas.org/cgi/doi/10.1073/pnas.1107017108

Dow

nloa

ded

by g

uest

on

June

1, 2

020

p53 (24). At the same time, in several mouse models of tumori-genesis, the tumor-resistant phenotype of Wip1-null mice wasabrogated by p53 deletion, supporting the interpretation thateffects of Wip1 deletion are mediated through the p53 pathway(15). The p53 tumor suppressor mutated or lost in approximately50% of human tumors. Therefore, the status of p53 may affect theresponse of tumor cells to Wip1 inhibition. First, we analyzed theeffect ofWip1 depletion on the sensitivity to an anticancer agent ofseveral tumor cell lines with or without p53. As expected, Wip1depletion (Fig. S1A) increased sensitivity to cisplatin (i.e., CDDP)in tumor cell lines expressing WT p53, the osteosarcoma cell lineU2OS and the colon cancer cell line HCT116, but it had no effecton the sensitivity in HCT116 p53−/− or the p53-negative osteo-sarcoma cell line Saos-2 (Fig. 1A).

To examine the effect of Wip1 overexpression on the sensitivityof tumor cells to anticancer drugs, we generated two human celllines, U2OS-Wip1-on (WT p53) and Saos-2-Wip1-on (p53 de-letion), in which Wip1 expression was regulated by a tetracyclineinducible promoter (Fig. S1B). Induction of Wip1 in the cells withWT p53 mildly protected the cells from anticancer treatment withcisplatin. In sharp contrast, in the highly cisplatin-resistant, p53-negative Saos-2 cells, previous Wip1 induction strongly increasedSaos-2-Wip1-on cell death in response to cisplatin (Fig. 1B andFig. S1C). Interestingly, Wip1 overexpression in Saos-2-Wip1-oncells increased cell death only after cisplatin treatment. Wip1 in-duction had no effect on survival of nontreated Saos-2-Wip1-oncells. Thus, Wip1 overexpression sensitized Saos-2 cells to cis-platin-induced cell death, but did not induce cell death by itself.

Fig. 1. Wip1 sensitizes tumor cells to chemotherapeutic agents. (A) Down-regulation of Wip1 sensitized cancer cell lines to cisplatin (CDDP) only in thepresence of p53. WT p53-expressing cells (human osteosarcoma U2OS cells and colorectal cancer HCT116 cells) and p53-negative cells (human osteosarcomaSaos-2 and colorectal cancer HCT116 p53−/− cells) were transfected with control scrambled siRNA or Wip1 siRNA. Twenty-four hours after transfection, cellswere nontreated (NT) or treated with 25 μM CDDP for 48 h, harvested, and subjected to Guava ViaCount cell death assay. (B) Sensitivity of Saos-2 and U2OScells to cisplatin (CDDP) with or without doxycycline-induced Wip1 expression. Wip1 was induced by doxycycline for 24 h in several clones of established U2OS-Wip1-on and Saos-2-Wip1-on cell lines (Fig. S1B). Cells with or without Wip1 induction were treated with 25 μM cisplatin for 48 h, harvested, and subjected toGuava ViaCount cell death assay (*P < 0.05). (C) Overexpression of phosphatase inactive Wip1 does not sensitize Saos-2 cells to cisplatin. Saos-2 cells weretransfected with mutant Wip1 D314A expressing vector DNA; Saos-2 Wip-on cells were cultivated with or without doxycycline. Twenty-four hours later, cellswere treated with 25 μM CDDP for 48 h, harvested, and subjected to Guava ViaCount cell death assay. (D) Overexpression of Wip1 sensitizes Saos-2 cells tovarious anticancer drugs. Saos-2 cells with inducible Wip1 were cultivated with or without doxycycline for 24 h and then treated with 25 μM CDDP, 10 μMetoposide (VP-16), 20 μg/mL of 5-fluorouracil (5-FU), or 1 μM camptothecin (CPT) for 48 h. Cells were harvested and subjected to Guava ViaCount cell deathassay. (E) Overexpression of Wip1 sensitizes non–small-cell lung carcinoma H1299 cells to CDDP. H1299 cells with inducible Wip1 were maintained with orwithout doxycycline for 24 h and then treated with 25 μM CDDP for 48 h. Cells were harvested and subjected to Guava ViaCount cell death assay. (F) Retroviraloverexpression of Wip1 sensitizes colon cancer cells HCT116 (p53−/−) to cisplatin. Cells were infected with retrovirus expressing Wip1 and treated with 25 μMCDDP for 48 h, harvested, and subjected to Guava ViaCount cell death assay (*P < 0.05, **P < 0.01, and ***P < 0.001; ns, not statistically significant at the 95%confidence level). Data represent the mean ± SEM of two to four independent experiments.

Goloudina et al. PNAS | January 10, 2012 | vol. 109 | no. 2 | E69

MED

ICALSC

IENCE

SPN

ASPL

US

Dow

nloa

ded

by g

uest

on

June

1, 2

020

To address the involvement of Wip1 phosphatase activity inthe observed sensitization phenotype, we established the stablytransfected Saos-2-Wip1 D314A-on cell line, which features in-ducible Wip1 bearing the phosphatase activity-impaired D314Amutation (Fig. S1D) (25). As shown in Fig. 1C, the D314Amutation in the phosphatase domain of Wip1 abrogated theWip1-dependent sensitization. Thus, we concluded that the ob-served sensitization in Saos-2-Wip1-on cells depended on thephosphatase activity of Wip1.To demonstrate that the sensitization was not specific only to

cisplatin, we extended the panel of anticancer agents by treatingSaos-2 cells with etoposide, 5-fluorouracil, and camptothecin. Weobserved that Wip1 overexpression significantly (P < 0.05) in-creased cell death induced by etoposide, 5-fluorouracil, or camp-tothecin in Saos-2-Wip1-on cells (Fig. 1D), similar to the inductionof cell death by these drugs in WT p53 cells, U2OS (Fig. S1G).To address whether this effect was specific for the osteosar-

coma-derived Saos-2 cells or was applicable more broadly to otherp53-negative cells, we generated anH1299 Tet-inducibleWip1 cellline (Fig. S1E); the H1299 cell line was derived from humannonsmall cell lung carcinoma and is null for p53 (26). Overex-pression of Wip1 in H1299 cells potentiated cell death induced bycisplatin in a similar way as was observed in the Saos-2 cell line(Fig. 1E). Furthermore, a sensitization to anticancer drugs wasalso observed after retroviral overexpression of Wip1 (Fig. S1F) ineight additional cancer cell lines with p53 deletion or mutation—HCT116 p53−/−, HT-29, PC3, MDA-MB-231, T47D, Panc1, Calu-6, and SK-OV-3—although the differences were not as pro-nounced as in Saos-2 and H1299 cells (Fig. 1F and Fig. S1H).These results indicate that activation of Wip1 may potentiate

the efficiency of anticancer drugs in tumors lacking p53. Wedecided next to explore the potential mechanism of this effect.

G2 Checkpoint in p53-Negative Cells Is Unaffected by Wip1 Overexpression. Previously, Wip1 has been characterized as a negativeregulator of Chk1 kinase (12). Furthermore, our earlier studies ofp53-negative tumors, which have defective checkpoints, showedthat inhibition of Chk1 sensitized the tumors to anticancer treat-ments through abrogation of the G2 checkpoint, resulting in pre-mature mitosis entrance with unrepaired DNA and subsequentmitotic catastrophe (22). Therefore, we examined whether Wip1overexpression affected Chk1 activation in response to cisplatintreatment. We observed only a slight delay in the phosphorylationof the Chk1 activating sites Ser317 and Ser345 in response tocisplatin treatment in Saos-2 cells after Wip1 induction (Fig. 2A).We analyzed whether Saos-2-Wip1-on cells could bypass G2checkpoint during cisplatin treatment.We determined the fractionof cells in mitotic phase, using phosphorylated histone H3 as amitotic marker. Cisplatin treatment resulted in fewer mitoticphase cells, indicating an effective G2 checkpoint in Saos-2-Wip1-on cells, either with or without Wip1 overexpression (Fig. 2B).Thus, it is unlikely that inhibition of Chk1, with ensuing

compromised G2 arrest and mitotic catastrophe, is responsiblefor sensitization resulting from Wip1 overexpression throughcompromised G2 arrest and mitotic catastrophe.

Wip1 Overexpression Permits Apoptosis in Response to AnticancerDrugs in Tumor Cells Lacking p53. Next we investigated the mech-anism of enhanced cell death resulting fromWip1 overexpression-induced sensitization to anticancer treatment.We noted that Saos-2-Wip1-on cells displayed the typical apoptotic morphology aftertreatment with anticancer drugs only when also treated withdoxycycline, which induced Wip1 overexpression. To investigatethe mechanism of increased apoptosis in greater detail, we de-termined the levels of several proapoptotic proteins. As expected,in U2OS cells, which contain functional p53, cisplatin treatmentinduced activation of caspases (caspases 3 and 9) irrespective ofthe presence or absence of Wip1 overexpression (Fig. 3A). In

contrast, cisplatin treatment of Saos-2 cells, which lack functionalp53, did not induce caspase activation. We observed elevatedlevels of active caspase-3 and caspase-9 after cisplatin treatmentonly in Saos-2 cells with induced Wip1 (Fig. 3B). To investigatefurther the mechanism of increased apoptosis, we examined thelevels of selected pro- and antiapoptotic proteins. As shown in Fig.3C, cisplatin treatment strongly inducedBax protein levels in Saos-2 cells overexpressing Wip1 whereas Bax protein levels remainedunchanged in the absence of Wip1 induction, thus correlating withcaspase activation. The antiapoptotic Bcl-2 protein levels in thesecells were not strongly affected by cisplatin treatment in the ab-sence or presence of Wip1 induction (Fig. 3C). The proapoptoticproteins Bad and Puma did not increase in response to cisplatintreatment either in the absence or presence of Wip1 induction,although the level of Puma decreased at the later time in thepresence of Wip1 overexpression (Fig. S2A). Interestingly, thelevel of the antiapoptotic Bcl-xl protein increased following cis-platin treatment in the absence of Wip1 induction, but remainedlow in cells overexpressing Wip1 following cisplatin treatment(Fig. S2A). Thus, when Wip1 is overexpressed, cisplatin treatmentresults in increased levels of Bax and reduced levels of Bcl-xl.The ratio of Bax to Bcl-xl is a critical determinant in the in-

duction of apoptosis. In cells with WT p53, Bax induction in re-sponse to anticancer therapy is mediated through the binding ofp53 to a specific response element in the Bax promoter, resultingin increased Bax transcription. Saos-2 cells bear a deletion in thep53 gene that prevents p53 expression (27). As expected, p53protein was not detected in Saos-2 or Saos-2-Wip1-on cells underany condition, whereas cisplatin treatment of U2OS-Wip1-on cellsresulted in strong or attenuated increases in the level of p53 in theabsence or presence of Wip1 induction, respectively (Fig. S2B).Therefore, p53 could not be responsible for the increased tran-

scription of Bax in Wip1-overexpressing Saos-2 cells after cisplatintreatment. We concentrated our further efforts on understandingthemechanism of Bax up-regulation as a major trigger of apoptosisin response to cisplatin in the absence of functional p53.

Apoptosis Induced by Cisplatin and Wip1 Overexpression Is RUNX2-Dependent. Among other possible transcriptional activators of Baxis RUNX2, a transcription factor involved in osteoblast differen-tiation that also has been implicated in the development of breastand prostate cancers (28, 29). RUNX2 has been identified as an

Fig. 2. Dephosphorylation of activating serines in Chk1 by Wip1 and G2checkpoint. (A) Wip1 suppresses phosphorylation of Chk1. Saos-2-Wip1-oncells were cultivated with or without doxycycline. Twenty-four hours later,cells were treated with 25 μM cisplatin (CDDP), harvested at indicated timepoints and analyzed by immunoblotting. Data are representative of threeindependent experiments. (B) Wip1 overexpression does not abrogate G2checkpoint. Saos2-Wip1-on cells were cultivated with or without doxycy-cline. Twenty-four hours later, cells were treated with 25 μM CDDP for 12 h.Nocodazole was added to trap cells in mitosis. Cell were harvested, fixed,and stained for phosphohistone H3 as a mitotic marker. Data represent themean ± SEM of three independent experiments.

E70 | www.pnas.org/cgi/doi/10.1073/pnas.1107017108 Goloudina et al.

Dow

nloa

ded

by g

uest

on

June

1, 2

020

inducer of Bax expression in response to TGF signaling, and twoRUNX2 binding elements have been identified in the Bax pro-moter (30). By treating cells with RUNX2 siRNA, we abrogatedthe Wip1-dependent sensitization of Saos-2 cells to cisplatin, asdemonstrated by the inhibition of apoptosis (Fig. 4A and Fig.S3A). Furthermore, we observed that Wip1 induction resulted inan increase in Bax mRNA levels in cells treated with controlsiRNA, but observed no increase in Bax mRNA levels in theRUNX2-depleted cells (Fig. 4B).

Wip1 Interacts with RUNX2 and Activates Its Transcriptional ActivityThrough Dephosphorylation of Ser432. We performed coimmuno-precipitation experiments and found that Wip1 interacted withRUNX2 in Saos-2-Wip1-on cells (Fig. 4C) and also in HCT116p53−/− and HT29 cells (Fig. S3B). We analyzed the sequence ofRUNX2 and found several potential sites of serine phosphoryla-tion residing in amino acid contexts similar to those of previouslycharacterized Wip1 phosphatase substrates. We separately mu-

tated Ser283, Ser430, Ser432, and Ser516 of RUNX2 to Ala. Wecotransfected Saos-2 cells with a reporter plasmid expressing lu-ciferase under the regulation of Bax-promoter including the Runxbinding sites and either WT or mutated RUNX2 expressionplasmids. Expression of WT RUNX2 induced luciferase activitymore than threefold compared with the empty vector (Fig. S3C).Mutation of RUNX2 Ser432 to alanine resulted in significantlyincreased transcriptional activity of the Bax promoter, comparedwith WT RUNX2 (Fig. 4D). The S430A mutation resulted in in-creased luciferase activity, but the difference was not significant atthe 95% confidence level. The mutations S283A and S516A didnot significantly affect RUNX2-induced luciferase expression.To confirm that phospho-Ser432 of RUNX2 was a target for

Wip1, we synthesized human RUNX2 peptide containing phos-phorylated Ser432 and measured the peptide concentration de-pendence of Wip1 phosphatase activity. As shown in Fig. 4E, theresults displayed Michaelis–Menten kinetics with a Km of 136 μMand a Vmax of 0.039 pmol/ng/s. These kinetic constants were similarto those obtained with the Chk1 345pS phosphopeptide, a wellknown substrate forWip1 (12). It was reported that Runx2 activitywas negatively regulated by phosphorylation of Ser465, an SP site(31, 32). As expected, Wip1 did not show measurable activityagainst the Runx2 phospho Ser465 peptide. The ATM-1981pSpeptide, an established substrate for Wip1, was used as a positivecontrol; the kinetic constants determined (Km = 30 μM, Vmax =0.010 pmol/ng/s) were very similar to those previously reported(11). Combined with the results of luciferase assay of S432A mu-tant, these data suggested that Ser432 of RUNX2 is a substrate forWip1 phosphatase activity.Thus, following cisplatin treatment, the interaction of Wip1

with RUNX2 and subsequent dephosphorylation of Ser432 canpotentiate the transcriptional activity of RUNX2, leading toincreased levels of Bax and, consequently, increased level ofapoptosis.

Wip1 Overexpression Sensitizes p53-Negative Tumors to Cisplatin inVivo. To confirm our findings in vivo, we established Saos-2 tumorswithout or with doxycycline-inducible Wip1 expression explantedinto athymic nude mice. Tumors were allowed to grow to a visiblesize (50 mm3), at which time Wip1 was induced by addition ofdoxycycline (2 mg/mL) into the drinking water. A single i.p. in-jection of cisplatin (10 mg/kg) abrogated tumor growth only intumors with Wip1 overexpression (Fig. 5A).

Wip1 Overexpression Protects Normal Tissues from Cell Death In-duced by Cisplatin. In contrast to increased apoptosis resulting fromWip1 overexpression in tumors lacking functional p53, the over-expression of Wip1 reduced cell death in tumor cells with WT p53during treatment with anticancer drugs (Fig. 1B). Therefore, in ourproposed model, Wip1 overexpression sensitizes only tumor cellslacking functional p53. We hypothesized that normal cells, withpreserved WT p53, should be protected by Wip1 down-regulationof the p53 response to anticancer drugs. We used pUbC-Wip1transgenic mice with ubiquitous Wip1 overexpression (33) to ana-lyze the effects of Wip1 activation in normal tissues during anti-cancer treatment. Among the most sensitive tissues to anticanceragents are the intestinal epithelium and the reproductive tissues.We compared cisplatin-induced apoptosis in intestinal crypts andtestes of WT mice and mice with Wip1 overexpression. We ob-served a significant decrease in activated caspase 3-positive cells inboth organs (Fig. 5 B–E). Thus, Wip1 overexpression, while sensi-tizing p53 negative tumors to chemotherapy, can protect normaltissue from deleterious effects of cisplatin anticancer treatment.

DiscussionThe development of more efficient strategies for cancer treatmentis one of the most important tasks for modern oncology. Manytherapeutic approaches under development are targeted to the

Fig. 3. Characterization of cisplatin-induced apoptosis in U2OS-Wip1-onand Saos-2-Wip1-on cells. (A) Induction of Wip1 in U2OS-Wip1-on cells delayscaspase 9 and caspase 3 activation after cisplatin (CDDP). U2OS-Wip1-on cellswere cultivated with or without doxycycline for 24 h, then treated with 25μM CDDP for indicated time points, harvested, and analyzed by immuno-blotting. (B) Induction of Wip1 in Saos-2-Wip1-on cells leads to strong cas-pase 9 and caspase 3 activation after CDDP, Saos-2-Wip1-on cells werecultivated with or without doxycycline for 24 h, then treated with 25 μMcisplatin for indicated time points, harvested, and analyzed by immuno-blotting. (C) Overexpression of Wip1 results in Bax induction after cisplatintreatment. Saos-2-Wip1-on cells were cultivated with or without doxycyclinefor 24 h, and then treated with 25 μM cisplatin for indicated time points,harvested, and immunoblotted. Data are representative of three inde-pendent experiments.

Goloudina et al. PNAS | January 10, 2012 | vol. 109 | no. 2 | E71

MED

ICALSC

IENCE

SPN

ASPL

US

Dow

nloa

ded

by g

uest

on

June

1, 2

020

status of the tumor suppressor p53 in specific tumors, wherebydifferent strategies are engaged for WT and mutant p53 (34). Therecently proposed therapy directed toward inhibition of Wip1phosphatase, one of the main negative regulators of the tumorsuppressor protein p53, has the potential to increase the efficacy ofexisting anticancer therapies and prevent cancer recurrence, mostparticularly for tumors with WT p53 (24, 35–37) Additionally, asWip1 is often amplified in many tumor types and its amplificationor overexpression interferes with several tumor suppressor path-ways (38), the use of Wip1 inhibitors in such patients would targetthe tumor-specific pathogenetic mechanism.The data presented here indicate that the p53 status in in-

dividual tumors should be determined before administration of aWip1 inhibitor. Our data provide support to the benefit of inhib-iting Wip1 in tumors with functional p53, in agreement with pre-vious observations (17, 39), but furthermore suggest that Wip1inhibitors may be ineffective in tumors exhibiting loss of functionalp53. In these cases, on the contrary, activation of Wip1 may po-tentiate the efficacy of anticancer agents, resulting in more effec-tive treatment. Our study demonstrates that overexpression ofWip1 sensitized tumor cells to anticancer agents and that this ef-

fect required the phosphatase activity of Wip1, as overexpressionof a phosphatase-impaired mutant Wip1 abrogated the sensitiza-tion of tumor cells to cisplatin. These observations motivate us topropose a search for specific Wip1 activators. Recently, chemicalactivators were identified for another PP2C familymember, PP2Cβ(40). Thus, the search for chemical activators of PP2Cδ -Wip1 maybe rationalized in terms of potential application in cancer therapy.Many of the recently characterized direct substrates of Wip1,

including the kinases ATM, Chk1, and Chk2, as well as H2AXand p53, conform to the p-Ser/p-Thr Q motif (11, 12). Here, wereport the characterization of p-Ser432 of the transcription fac-tor RUNX2 as a substrate of Wip1 phosphatase activity. More-over, we observed a physical association between RUNX2 andWip1 and demonstrated that Wip1 activates the transcriptionalactivity of RUNX2 through dephosphorylation of Ser432.RUNX2 belongs to the family of RUNT-domain transcriptional

factors. The Runx genes family comprises three closely relatedtranscription factors: Runx1, Runx2, and Runx3, each of whichbinds a common partner, CBFβ, to form a core binding factor(CBF) complex that can activate or repress gene transcription.Runx genes are expressed in a tissue specific manner, but they have

Fig. 4. Role of RUNX2 transcriptional factor in Wip1-dependent cisplatin-induced apoptosis. (A) RUNX2 silencing abrogates Wip1-mediated cell death aftercisplatin (CDDP) treatment in Saos-2-Wip1-on cells. Saos-2-Wip1-on cells were transfected with control scrambled siRNA or RUNX2 siRNA. Forty-eight hoursafter transfection, cells were treated with CDDP for 48 h, harvested, and subjected to Guava ViaCount cell death assay. (B) RUNX2 silencing abrogates BaxmRNA induction in Saos2-Wip1-on cells after CDDP treatment. Saos2-Wip1-on cells were transfected with control scrambled siRNA or RUNX2 siRNA. Thirtyhours after transfection, Wip1 was induced by doxycycline. Forty-eight hours after transfection, cells were treated with cisplatin for 6 h and harvested, andmRNA was purified and subjected to quantitative PCR analysis. (C) Wip1 interacts with endogenous RUNX2. Saos-2-FLAG-Wip1-on cells were cultivated withor without doxycycline. Twenty-four hours later, cells were lysed and immunoprecipitated with FLAG M2 or RUNX2 antibody, and immunoblotted with FLAG,Wip1, or RUNX2 antibody as indicated. Whole-cell lysates were also immunoblotted. (D) Transactivation of the human Bax gene promoter by Runx2 mutants.Firefly luciferase reporter vectors contained a 1.2-kb fragment of the human bax gene promoter. Saos-2 cells were cotransfected with the reporter plasmid,a WT Runx2 or mutant Runx2 expression plasmid, and a renilla luciferase plasmid as a reference. The values shown are the means ± SE from threeexperiments. (E) Phosphatase activity of Wip1 on Runx2 phosphopeptides. Phosphate release by Wip1 with 0 to 400 μM 432pS (○), and 465pS (◆) phos-phopeptides was measured at 30 °C. Wip1 activity was measured in the same conditions with 0 to 100 μM ATM-1981pS (▵) phosphopeptide as a positivecontrol. The values shown represent the means ± SEM for A, B, and D, or means ± SD for E of three to four independent experiments (*P < 0.05; ns, notstatistically significant at the 95% confidence level). Data are representative of two or three independent experiments.

E72 | www.pnas.org/cgi/doi/10.1073/pnas.1107017108 Goloudina et al.

Dow

nloa

ded

by g

uest

on

June

1, 2

020

overlapping functions and the capacity for cross-regulation. Thestability and activity of RUNX2 are regulated by posttranslationalmodifications (41, 42). In particular, phosphorylation of RUNX2can activate or repress its transcriptional activity (43–45). It wasproposed that some proapoptotic genes may be important carci-nogenic targets of RUNX transcriptional factors and TGF-β–induced apoptosis is greatly reduced after depletion of RUNX insome tissues (46). Recently, it was shown that RUNX2 binds to thepromoter of the proapoptotic gene Bax and induces Bax expressionand apoptosis in response to BMP2 stimulation and etoposidetreatment in osteosarcomas Saos-2 cells (30). In our study, Wip1overexpression led to Bax induction and promoted caspase-de-pendent apoptosis in response to cisplatin treatment. This effectwas abrogated by siRNA-mediated RUNX2 depletion. Moreover,substitution of the Wip1-targeted serines to alanine, which mimicsdephosphorylated state, increased the transcriptional activity ofRUNX2. We hypothesized that, in cells in which p53 is active, it isp53 that plays the role of the main regulator of Bax expression inresponse to cisplatin. The absence of active p53 caused by muta-tions in the p53 gene (TP53) RUNX2 can be amajor transcriptionalactivator of Bax expression, resulting in increased Bax protein levelsand increased induction of apoptosis. The ratio of Bax to Bcl-xl iscritical in determining the sensitivity of cells to anticancer drugs(47–49). Thus, we have identified RUNX2 as a target of Wip1phosphatase activity and have shown that, during anticancer ther-apy, in the absence of p53, transcriptionally activated RUNX2 isresponsible for mediating the increased apoptotic response.

One of the most interesting findings of our work is that thepotentiating effect of increased Wip1 activity on anticancer agenttreatment is restricted to tumors. Moreover, the transient activa-tion of Wip1 may protect normal tissues from deleterious sideeffects of the anticancer therapy. In patients bearing tumors with aloss of functional p53 gene, the normal tissues still express WTp53, which is responsible for most of the cytotoxic effects of anti-cancer agents (50). Among the organs most sensitive to this un-desirable effect of anticancer therapy are the intestine and testes,leading to severe effects in the digestive and reproductive systems(51). Wip1 is highly expressed in testis and intestinal stem cells (15,52). Its transient activation would increase the threshold of p53activation in response to anticancer treatment and thus preventapoptosis. This hypothesis was supported by our observation fromin vivo experiments demonstrating that, compared with WT mice,transgenic mice overexpressing Wip1 exhibited significantly lesscisplatin-induced damage in the intestine and testis. In our study,we analyzed only the early, caspase 3-dependent apoptosis, whichhas been shown to be p53-dependent in response to DNA damagein the intestine. We propose that Wip1, as a negative regulator ofp53 activity, has a maximal effect on this particular type of celldeath. Other types of cell death have been implicated in the p53-dependent response to DNA-damaging treatments, and we cannotexclude the possibility that additional mechanisms are also af-fected by Wip1 overexpression (53). Previously it has been shownthat, although the immediate p53-dependent DNA damage re-sponse is irrelevant to the subsequent suppression of tumorigen-esis, it is responsible for the severe side effects (54). It was

Fig. 5. Increased antitumor efficiency of cisplatin and protection of normal tissues from cisplatin-induced cytotoxicity by Wip1 overexpression in vivo. (A)Tumor growth of Saos-2-Wip1-on cells explanted into the nude mice. Saos-2-on and Saos-2-Wip1-on were injected s.c. on day 0 into athymic nude mice (3 ×106

cells per mouse). At the indicated time (arrow), when tumor became visible, Wip1 was induced by addition of doxycycline (2 mg/mL) to the drinking water;after 24 h, cisplatin was administrated as a single dose (10 mg/kg, i.p.). Tumor sizes are plotted as mean ± SEM for five mice per group. Experiments wererepeated two times with similar results. (B–E) Decrease in cisplatin-induced apoptosis in intestinal crypt (B and D) and testes (C and E) of transgenic mice withWip1 overexpression. FVBn (control mice) and pUbC-Wip1 mice were injected with cisplatin (10 mg/kg). After 12 h, organs were harvested and fixed informalin. Tissue sections were stained for active caspase 3. Arrows indicate apoptotic caspase 3-positive cells. For intestinal sections, apoptotic cells werecounted per intestinal circumference containing approximately 100 crypts on average. For testes sections, apoptotic cells were counted per testes section.Experiments were repeated two times with similar results with three to four mice per group (*P < 0.05 and **P < 0.01).

Goloudina et al. PNAS | January 10, 2012 | vol. 109 | no. 2 | E73

MED

ICALSC

IENCE

SPN

ASPL

US

Dow

nloa

ded

by g

uest

on

June

1, 2

020

proposed that p53 activation is important for tumor suppressionduring the period following recovery from DNA damage (55).Thus, the transient activation of Wip1 before anticancer treat-ments that induce the DNA damage response could suppress p53-dependent toxicity of such treatments toward normal tissueswithout affecting the tumor-suppressive activity of p53 pathway.In conclusion, our study uncovers a proapoptotic function of

Wip1 that acts through dephosphorylation and activation ofRUNX2. We propose an anticancer strategy for the treat-ment of tumors that lack functional p53, which is based on thetransient activation of Wip1 and simultaneously protectsnormal tissues.

Materials and MethodsCell Lines and Cell Culture. The U2OS Tet-on and Saos-2 Tet-on cell lines,human osteosarcoma lines with a tetracycline-inducible gene expression sys-tem, were purchased from Clontech. To obtain cells stably expressing Wip1,the Wip1, FLAG-tagged Wip1 or FLAG-tagged Wip1 D314A mutant codingsequence was inserted in TRE2pur (Clontech). Twenty puromycin-resistantclones were isolated, amplified, and checked for inducible expression ofFLAG-Wip1 protein following addition of 5 μg/mL doxycycline (D-9891;Sigma). Clonal lines with high levels of inducible Wip1 expression were usedfor experiments. For human non–small-cell lung cancer cell line NCI-H1299,stable clones expressing pTet-ON (Clontech) were established before trans-fection with the pTRE2pur–Wip1 vector.

Mice. Female athymic Swiss nude mice were purchased from Charles River.Exponentially growing Saos-2-Wip1-on cells were harvested, resuspended inPBS solution, and injected s.c. (3×106 cells) into the rightflank.When the tumorsize reached 50mm3, mice received doxycycline in drinking water (2 mg/mL) toinduce Wip1 in tumor cells and were injected i.p. with cisplatin (10 mg/kg).Tumor volumes were evaluated every 4 d. The animals were treated accordingto theguidelines of theMinistère de la Recherche et de la Technologie (France).All experiments were approved by the Comité d’Ethique de l’Université deBourgogne. pUbC-Wip1 transgenic mice were described before (33).

Immunohistochemistry. Tissues were fixed with 10% formalin, embedded inparaffin, and sectioned at 5 μm. Anti-caspase 3 antibody (AF835; R&D Sys-tems) was used.

Plasmid Constructs and siRNAs. The phosphatase-inactive Wip1 point mutantD314A was described before (25). The pCMV-RUNX2 and pGL4.10-luc-Bax-promoter vectors were a gift from Roman Eliseev (University of Rochester,Rochester, NY). To introduce point mutations into Wip1 and RUNX2, weused a QuikChange II site-directed mutagenesis kit (Stratagene).

ON-Target Anti-RUNX2 siRNA, anti-Wip1 siRNA, and control siRNA werepurchased from Dharmacon (catalog nos. L-012665–00, L-004554–0005, andD-001810–01-05, respectively).

Cells were transfected with plasmid DNA by jetPEI (catalog no. 101–10;Polyplus Transfection). To transfect cells with siRNA, we used INTERFERin(catalog no. 409–10; Polyplus Transfection). Transfections experiments wereperforming according to manufacturer’s instructions.

Wip1 cDNA was cloned into the PINCO vector (56). Retroviral infection ofHCT116 and HCT116 p53−/− colon cancer cell lines (gifts from Bert Vogel-stein, Johns Hopkins University, Baltimore, MD) were performed accordingto the Retroviral Gene Transfer and Expression User Manual (Clontech).

Western Blot Analysis and Immunoprecipitation. Cell lysates containing 50 μg ofprotein were analyzed by Western blot by using the following primary anti-bodies: anti-Wip1 (H-300; Santa Cruz), anti-RUNX2 (M-70; Santa Cruz), anti-p53(DO-1; Santa Cruz ), anti-Chk1 (G-4; Santa Cruz ), anti–β-actin antibody (A 2103;Sigma), anti–phospho-Chk1 Ser345, anti–phospho-Chk1 Ser317, anti-Bax, anti-Bad, anti–Bcl-2, anti-Puma, anti–Bcl-xl, anti-caspase 3, and anti-caspase 9 (allfrom Cell Signaling Technologies). Secondary antibodies were from JacksonImmuno Research. Immunoreactivity was detected by using Western BlottingLuminol Reagent (Santa Cruz).

For immunoprecipitation, cells were seeded in 75-cm2flasks, treated with

5 μg/mL of doxycycline for 24 h or infectedwithWip1 virus as described earlier,and lysed in Tris-buffered saline solution containing 1% Nonidet P-40, 5%glycerol, and protease inhibitors. Cell lysates were centrifuged at 15,000 × g,and supernatant was used for immunoprecipitation. Wip1-FLAG was pre-cipitated overnight with 20 μL of Anti-FLAG M2 Affinity Gel (A2220; Sigma).RUNX2 was precipitated with 10 μg of anti-RUNX2 antibodies (M-70 X; SantaCruz) and then 40 μL of Protein A/G beads was added to cell lysate for 4 h.Immunoprecipitateswere spun down andwashed three timeswith lysis buffer,boiled in sample buffer, and loaded on 10% SDS-polyacrylamide gels. Proteinswere transferred to PVDFmembrane andwere detectedwith RUNX2 antibody(catalog no. D130-3; MBL International), anti-FLAG M2 antibody (catalog no.F1804; Sigma), and antibodies described earlier.

In Vitro Phosphatase Assay. The N-terminal histidine-tagged, catalytic domainof humanWip1 (residues 1–420), rWip1, was expressed in Escherichia coli BL21(DE3) and purified as previously reported (57). The sequences of the humanRUNX2-432pS, RUNX2-465pS and ATM-1981pS peptides, which were syn-thesized by solid phase chemistry, were PYPGSSQ(pS)QSGPF, PGGDR(pS)PSRMLGY, and AFEEG(pS)QSTTIGY, respectively. Wip1 phosphatase activitywas determined by measuring the released phosphate using a malachitegreen/molybdate based assay. For each reaction, 60 ng of rWip1was incubatedwith the phosphopeptide for 7 min at 30 °C in a reaction medium containing50 mM Tris-HCl (pH 7.5), 0.1 mM EGTA, 0.02% 2-mercaptoethanol, 40 mMNaCl, and 30 mMMgCl2. The reaction was stopped by adding Malachite greensolution (Millipore) or Biomol Green (Enzo Life Sciences), and the absorbancewas measured at 650 nm or 620 nm, respectively. The data were fitted to theMichaelis–Menten equation by using GraphPad Prism 5 software.

Luciferase Assays. The luciferase reporter plasmids, pGL4.10-luc-Bax-pro-moter (firefly), and pRL-TK (renilla; Promega) were cotransfected along withpCMV-RUNX2 or its mutant variants into SAOS-2 cells. Firefly and renillaluciferase activities were measured by using a Lumat LB 9507 Single TubeLuminometer (Berthold Technologies) and the Dual Luciferase Reporter AssaySystem (Promega) according to the manufacturer’s protocol. The firefly lu-ciferase signal was normalized to the renilla luciferase signal.

Cell Death Assay. The number of dead cells was estimated by the cell deathassay using the Guava ViaCount Reagent (catalog no. 4000–0041; Millipore).Cells were trypsinized, washed with PBS solution, and stained with GuavaViaCount Reagent for 5 min. The percentage of dead cells was measured byusing Guava EasyCyte Plus Flow Cytometry System (Millipore) and analyzedby using Guava ViaCount Software.

Real-Time RT-PCR. Total cellular RNA was prepared using RNeasy Kit (Qiagen)and reverse-transcribed into cDNAby using SuperScript II First-Strand SynthesisKit (Invitrogen) andoligo-dTprimers according to themanufacturer’s protocol.Real-time PCR was performed by using a 7500 Fast Real-Time PCR System(Applied Biosystems), KAPA SYBR FAST qPCR Kit (KAPA Biosystems), and thefollowing sets of primers: Bax (5′-TGCTTCAGGGTTTCATCCAG-3′ and 5′-GGC-GGCAATCATCCTCTG-3′), RUNX2 (5′-CCGGAATGCCTCTGCTGTTATGA-3′ and5′-ACTGAGGCGGTCAGAGAACAAACT-3′), and GAPDH (5′-GAAGGTGAAGG-TCGGAGTC-3′ and 5′-GAAGATGGTGATGGGATTTC-3′). The expression ofmRNA of interest was normalized to the expression of GAPDH.

Statistical Analysis. Statistical analysis was performed by using GraphPadPrism 5 software.

ACKNOWLEDGMENTS. We are thankful to Dr. Alexei Arnaoutov, Dr.Vladimir Kuznetsov, and Dr. Laurent Lagrost for productive discussions andmethodological help; to Dr. Roman Eliseev and Dr. Hicham Drissi for Runx2and Bax-promoter vectors. C. Garrido group has the label “La Ligue Contre leCancer.” This work was supported by the Ligue Contre le Cancer La Confér-ence de Coordination Interrégionale du Grand Est and Conseil Regional deBourgogne, and in part by the Intramural Research Program of the NationalCancer Institute, National Institutes of Health, and federal funds from theNational Cancer Institute, National Institutes of Health, under HHSN 2612008 00001E.

1. Harris SL, Levine AJ (2005) The p53 pathway: Positive and negative feedback loops.

Oncogene 24:2899–2908.2. Martin AC, et al. (2002) Integrating mutation data and structural analysis of the TP53

tumor-suppressor protein. Hum Mutat 19:149–164.3. Momand J, Jung D, Wilczynski S, Niland J (1998) The MDM2 gene amplification

database. Nucleic Acids Res 26:3453–3459.

4. Bond GL, HuW, Levine AJ (2005) MDM2 is a central node in the p53 pathway: 12 years

and counting. Curr Cancer Drug Targets 5:3–8.5. Lu X, et al. (2007) The Wip1 Phosphatase acts as a gatekeeper in the p53-Mdm2

autoregulatory loop. Cancer Cell 12:342–354.6. Meek DW, Anderson CW (2009) Posttranslational modification pf p53: cooperative

integrators of function. Cold Spring Harbor Perspect Biol 1:a000950.

E74 | www.pnas.org/cgi/doi/10.1073/pnas.1107017108 Goloudina et al.

Dow

nloa

ded

by g

uest

on

June

1, 2

020

7. Bulavin DV, et al. (2002) Amplification of PPM1D in human tumors abrogates p53tumor-suppressor activity. Nat Genet 31:210–215.

8. Hirasawa A, et al. (2003) Association of 17q21-q24 gain in ovarian clear celladenocarcinomas with poor prognosis and identification of PPM1D and APPBP2 aslikely amplification targets. Clin Cancer Res 9:1995–2004.

9. Fujimoto H, et al. (2006) Regulation of the antioncogenic Chk2 kinase by theoncogenic Wip1 phosphatase. Cell Death Differ 13:1170–1180.

10. Oliva-Trastoy M, et al. (2007) The Wip1 phosphatase (PPM1D) antagonizes activationof the Chk2 tumour suppressor kinase. Oncogene 26:1449–1458.

11. Shreeram S, et al. (2006) Wip1 phosphatase modulates ATM-dependent signalingpathways. Mol Cell 23:757–764.

12. Lu X, Nannenga B, Donehower LA (2005) PPM1D dephosphorylates Chk1 and p53 andabrogates cell cycle checkpoints. Genes Dev 19:1162–1174.

13. Shreeram S, et al. (2006) Regulation of ATM/p53-dependent suppression of myc-induced lymphomas by Wip1 phosphatase. J Exp Med 203:2793–2799.

14. Demidov ON, et al. (2007) The role of the MKK6/p38 MAPK pathway in Wip1-dependent regulation of ErbB2-driven mammary gland tumorigenesis. Oncogene 26:2502–2506.

15. Demidov ON, et al. (2007) Wip1 phosphatase regulates p53-dependent apoptosis ofstem cells and tumorigenesis in the mouse intestine. Cell Stem Cell 1:180–190.

16. Bulavin DV, et al. (2004) Inactivation of the Wip1 phosphatase inhibits mammarytumorigenesis through p38 MAPK-mediated activation of the p16(Ink4a)-p19(Arf)pathway. Nat Genet 36:343–350.

17. Kong W, Jiang X, Mercer WE (2009) Downregulation of Wip-1 phosphataseexpression in MCF-7 breast cancer cells enhances doxorubicin-induced apoptosisthrough p53-mediated transcriptional activation of Bax. Cancer Biol Ther 8(6).

18. Bouchet BP, de Fromentel CC, Puisieux A, Galmarini CM (2006) p53 as a target foranti-cancer drug development. Crit Rev Oncol Hematol 58:190–207.

19. el-Deiry WS (1998) Regulation of p53 downstream genes. Semin Cancer Biol 8:345–357.

20. Aas T, et al. (1996) Specific P53 mutations are associated with de novo resistance todoxorubicin in breast cancer patients. Nat Med 2:811–814.

21. Wattel E, et al. (1994) p53 mutations are associated with resistance to chemotherapyand short survival in hematologic malignancies. Blood 84:3148–3157.

22. Demidova AR, Aau MY, Zhuang L, Yu Q (2009) Dual regulation of Cdc25A by Chk1and p53-ATF3 in DNA replication checkpoint control. J Biol Chem 284:4132–4139.

23. Greenow KR, Clarke AR, Jones RH (2009) Chk1 deficiency in the mouse small intestineresults in p53-independent crypt death and subsequent intestinal compensation.Oncogene 28:1443–1453.

24. Bang J, Yamaguchi H, Durell SR, Appella E, Appella DH (2008) A small molecularscaffold for selective inhibition of Wip1 phosphatase. ChemMedChem 3:230–232.

25. Takekawa M, et al. (2000) p53-inducible wip1 phosphatase mediates a negativefeedback regulation of p38 MAPK-p53 signaling in response to UV radiation. EMBO J19:6517–6526.

26. Lin DL, Chang C (1996) p53 is a mediator for radiation-repressed human TR2 orphanreceptor expression in MCF-7 cells, a new pathway from tumor suppressor to memberof the steroid receptor superfamily. J Biol Chem 271:14649–14652.

27. Masuda H, Miller C, Koeffler HP, Battifora H, Cline MJ (1987) Rearrangement of thep53 gene in human osteogenic sarcomas. Proc Natl Acad Sci USA 84:7716–7719.

28. Zaidi SK, et al. (2007) Runx2 deficiency and defective subnuclear targeting bypasssenescence to promote immortalization and tumorigenic potential. Proc Natl Acad SciUSA 104:19861–19866.

29. Onodera Y, et al. (2010) Runx2 in human breast carcinoma: Its potential roles incancer progression. Cancer Sci 101:2670–2675.

30. Eliseev RA, et al. (2008) Runx2-mediated activation of the Bax gene increasesosteosarcoma cell sensitivity to apoptosis. Oncogene 27:3605–3614.

31. Wee HJ, Huang G, Shigesada K, Ito Y (2002) Serine phosphorylation of RUNX2 withnovel potential functions as negative regulatory mechanisms. EMBO Rep 3:967–974.

32. Shen R, et al. (2006) Cyclin D1-cdk4 induce runx2 ubiquitination and degradation.J Biol Chem 281:16347–16353.

33. Wong ES, et al. (2009) p38MAPK controls expression of multiple cell cycle inhibitorsand islet proliferation with advancing age. Dev Cell 17:142–149.

34. Wang Z, Sun Y (2010) Targeting p53 for novel anticancer therapy. Transl Oncol 3:1–12.

35. Tan DS, et al. (2009) PPM1D is a potential therapeutic target in ovarian clear cellcarcinomas. Clin Cancer Res 15:2269–2280.

36. Rayter S, et al. (2008) A chemical inhibitor of PPM1D that selectively kills cellsoverexpressing PPM1D. Oncogene 27:1036–1044.

37. Belova GI, Demidov ON, Fornace AJ, Jr., Bulavin DV (2005) Chemical inhibition ofWip1 phosphatase contributes to suppression of tumorigenesis. Cancer Biol Ther 4:1154–1158.

38. Lu X, et al. (2008) The type 2C phosphatase Wip1: An oncogenic regulator of tumorsuppressor and DNA damage response pathways. Cancer Metastasis Rev 27:123–135.

39. Natrajan R, et al. (2009) Tiling path genomic profiling of grade 3 invasive ductalbreast cancers. Clin Cancer Res 15:2711–2722.

40. Brautigan DL, et al. (2005) Allosteric activation of protein phosphatase 2C by D-chiro-inositol-galactosamine, a putative mediator mimetic of insulin action. Biochemistry44:11067–11073.

41. Bae SC, Lee YH (2006) Phosphorylation, acetylation and ubiquitination: The molecularbasis of RUNX regulation. Gene 366:58–66.

42. Jonason JH, Xiao G, Zhang M, Xing L, Chen D (2009) Post-translational regulation ofRunx2 in bone and cartilage. J Dent Res 88:693–703.

43. Ge C, et al. (2009) Identification and functional characterization of ERK/MAPKphosphorylation sites in the Runx2 transcription factor. J Biol Chem 284:32533–32543.

44. Kugimiya F, et al. (2007) GSK-3beta controls osteogenesis through regulating Runx2activity. PLoS ONE 2:e837.

45. Park OJ, Kim HJ, Woo KM, Baek JH, Ryoo HM (2010) FGF2-activated ERK mitogen-activated protein kinase enhances Runx2 acetylation and stabilization. J Biol Chem285:3568–3574.

46. Ito Y (2004) Oncogenic potential of the RUNX gene family: ‘Overview’. Oncogene 23:4198–4208.

47. Edlich F, et al. (2011) Bcl-x(L) retrotranslocates Bax from the mitochondria into thecytosol. Cell 145:104–116.

48. Zhang XW, Xu B (2000) Differential regulation of P53, c-Myc, Bcl-2, Bax and AFPprotein expression, and caspase activity during 10-hydroxycamptothecin-inducedapoptosis in Hep G2 cells. Anticancer Drugs 11:747–756.

49. Semaan SJ, Nickells RW (2010) The apoptotic response in HCT116BAX-/- cancer cellsbecomes rapidly saturated with increasing expression of a GFP-BAX fusion protein.BMC Cancer 10:554.

50. Komarov PG, et al. (1999) A chemical inhibitor of p53 that protects mice from the sideeffects of cancer therapy. Science 285:1733–1737.

51. Christophorou MA, et al. (2005) Temporal dissection of p53 function in vitro and invivo. Nat Genet 37:718–726.

52. Choi J, et al. (2002) Mice deficient for the wild-type p53-induced phosphatase gene(Wip1) exhibit defects in reproductive organs, immune function, and cell cyclecontrol. Mol Cell Biol 22:1094–1105.

53. Komarova EA, et al. (2004) Dual effect of p53 on radiation sensitivity in vivo: p53promotes hematopoietic injury, but protects from gastro-intestinal syndrome in mice.Oncogene 23:3265–3271.

54. Christophorou MA, Ringshausen I, Finch AJ, Swigart LB, Evan GI (2006) Thepathological response to DNA damage does not contribute to p53-mediated tumoursuppression. Nature 443:214–217.

55. Junttila MR, Evan GI (2009) p53—a Jack of all trades but master of none. Nat RevCancer 9:821–829.

56. Grignani F, et al. (1998) High-efficiency gene transfer and selection of humanhematopoietic progenitor cells with a hybrid EBV/retroviral vector expressing thegreen fluorescence protein. Cancer Res 58:14–19.

57. Yamaguchi H, et al. (2005) Substrate specificity of the human protein phosphatase2Cdelta, Wip1. Biochemistry 44:5285–5294.

Goloudina et al. PNAS | January 10, 2012 | vol. 109 | no. 2 | E75

MED

ICALSC

IENCE

SPN

ASPL

US

Dow

nloa

ded

by g

uest

on

June

1, 2

020