hsp90inhibitorspromotep53 …...cancer therapeutics insights...

Cancer Therapeutics Insights

Hsp90 Inhibitors Promote p53-Dependent Apoptosis throughPUMA and Bax

Kan He1,3, Xingnan Zheng1,3, Lin Zhang2,3, and Jian Yu1,3

AbstractHsp90 is widely overexpressed in cancer cells and believed to be essential for the maintenance of

malignant phenotypes. Targeting Hsp90 by small molecules has shown promise in solid and hema-

tologic malignancies, which likely involves degradation of client oncoproteins in a cell-type–specific

manner. In this study, we found that structurally unrelated Hsp90 inhibitors induce DNA damage and

apoptosis via p53-dependent induction of PUMA, which indirectly triggers Bax activation and mito-

chondrial dysfunction in colon cancer cells. Deficiency in PUMA, BAX, or p53, at lesser extent, abrogated

17-allylamino-17-demethoxygeldanamycin (17-AAG)-induced apoptosis and mitochondrial dysfunction,

and enhanced clonogenic cell survival. Furthermore, suppression of p53-dependent p21 induction or

enhanced p53 activation synergized with 17-AAG to induce PUMA-dependent apoptosis. Finally,

PUMA was found to mediate apoptotic and therapeutic responses to the 17-AAG analog 17-DMAG in

xenografts. These results show an important role of the p53/PUMA/Bax axis in Hsp90 inhibitor–induced

killing of p53 wild-type cells, and have important implications for their clinical applications. Mol Cancer

Ther; 12(11); 2559–68. �2013 AACR.

IntroductionHsp90 is one of the most abundant molecular cha-

perones in eukaryotes and regulates many cellularprocesses including signal transduction, protein deg-radation, protein folding, and maturation of clientproteins (1). More than 200 proteins have been reportedto be Hsp90 clients, including many oncogenic pro-teins, such as mutant p53, Raf-1, Akt (1, 2), and othersassociated with hallmarks of cancer (3). Interestingly,Hsp90 is constitutively expressed at 2- to 10-fold higherlevels in tumor cells than their normal counterparts (4),and seems to be required for malignant transformation(5), and a selective therapeutic target in cancer cells(1, 2).Hsp90 targeting has attracted considerable attention

since the late 1990s. Currently, more than 17 Hsp90inhibitors have entered clinical trials as potential anti-cancer agents (1, 5). These inhibitors, such as the naturalproduct geldanamycin and its less toxic analog 17-ally-

lamino-17-demethoxygeldanamycin (17-AAG), inhibitthe molecular chaperone function of Hsp90 by bindingto its ATP/ADP pocket and causing destabilization ofits complexes with client proteins (2). Treatment withgeldanamycin and 17-AAG can induce apoptosis, inhib-it metastasis or angiogenesis in both cell and animalmodels (1, 5, 6). 17-AAG is one of the very first Hsp90inhibitors to complete phase II clinical trials (1, 5). OtherHsp90 inhibitors with improved pharmaceutical prop-erties, such as 17-DMAG and NVP-AUY922, can pro-mote apoptosis in preclinical models and are currentlyin phase I/II clinical trials (1, 7). However, the mechan-isms underlying Hsp90 inhibition-induced cell killingare currently not well understood, and may vary bycell type. Several mechanisms have been suggested,including Akt downregulation (8), inhibition of NF-kBactivation due to inhibitor of IkB kinase destabilization(9), endoplasmic reticulum (ER) stress (10), and morerecently p53 activation (11, 12).

The Bcl-2 family proteins are the central regulators ofmitochondria-mediated apoptosis (13–15). The BH3-onlysubfamilymembers areproximal signalingmolecules thatrespond to distinct as well as overlapping signals, andconsist of at least 10 members. Several of BH3-only pro-teins, such as PUMAandBim, activate Bax/Bak followingthe neutralization of all known antiapoptotic Bcl-2 familymembers (13–15) to promote mitochondrial dysfunctionand caspase activation (16–18). PUMA is a critical medi-ator of p53-dependent and -independent apoptosis inmultiple tissues and cell types (18). Transcription ofPUMA is mainly dependent on p53 activity in responseto DNA-damaging agents such as g-irradiation, common

Authors' Affiliations: Departments of 1Pathology, 2Pharmacology andChemical Biology, University of Pittsburgh School of Medicine, 3Universityof Pittsburgh Cancer Institute, Pittsburgh, Pennsylvania

Note: Supplementary data for this article are available at Molecular CancerTherapeutics Online (http://mct.aacrjournals.org/).

K. He and X. Zheng contributed equally to this work.

Corresponding Author: Jian Yu, Hillman Cancer Center Research Pavil-ion, Suite 2.26h, 5117 Centre Ave, Pittsburgh, PA 15213. Phone: 412-623-7786; Fax: 412-623-7778; E-mail: [email protected]

doi: 10.1158/1535-7163.MCT-13-0284

�2013 American Association for Cancer Research.

MolecularCancer

Therapeutics

www.aacrjournals.org 2559

on September 21, 2020. © 2013 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst August 21, 2013; DOI: 10.1158/1535-7163.MCT-13-0284

http://mct.aacrjournals.org/

chemotherapeutic drugs (18, 19), and even some kinaseinhibitors (20). Upon exposure to nongenotoxic stimuli,PUMA is induced by other transcriptional factors suchas p73 (21, 22), NF-kB (23), and FoxO3a (24, 25). Bimregulates apoptosis in immune cells, normal and malig-nant hematopoietic cells, and some epithelial cells gener-ally following nongenotoxic stresses (13).

In this study, we investigated the mechanisms of apo-ptosis induced by Hsp90 inhibitors in colon cancer cells.We found that p53-mediated induction of PUMA and Baxis required for the apoptotic responses toHsp90 inhibitorsvia the mitochondrial pathway in vitro and in vivo. Incontrast, p53-mediated p21 induction suppresses 17-AAG–induced apoptosis. Enhanced p53 activation orp21 inhibition sensitized p53wild-type (WT) colon cancercells to PUMA-dependent apoptosis following treatmentwith Hsp90 inhibitors. These results show an importantrole of the p53/PUMA/Bax axis in the therapeuticresponses to Hsp90 inhibitors in p53 WT cancers, andhave important implications for their future developmentand applications.

Materials and MethodsCell culture and treatment

The human colorectal cancer cell lines, includingHCT116, RKO, LoVo (WT p53), HT29, and DLD1 (mutantp53), were obtained from the American Type CultureCollection. Isogenic HCT 116 knockout cell lines includ-ing p53-KO (26), PUMA-KO (27), BAX-KO (28), p21-KO(29), p21/PUMA double KO (DKO; ref. 27), and PUMA-KO DLD1 cells (30) have been described. We examineloss of expression of targeted proteins by Westernblotting and conduct mycoplasma testing by PCR dur-ing culture routinely. No additional authentication wasdone by the authors. Details on cell culture and drugtreatments are found in the Supplementary Materialsand Methods.

Western blottingWestern blotting was conducted as previously

described (16). More details on antibodies are found inthe Supplementary Materials and Methods.

Real-time reverse transcriptase PCRTotal RNAwas isolated fromuntreated or drug-treated

cells using theMini RNA Isolation II Kit (Zymo Research)according to the manufacturer’s protocol. Of note, 1 mg oftotal RNAwas used to generate cDNA using SuperScriptII reverse transcriptase (Invitrogen). Real-time PCR wascarried out as before (24). Details on primers are found inthe Supplementary Materials and Methods.

siRNA knockdownTransfection was conducted 24 hours before 17-AAG

treatment using 400pmol of siRNA inonewell of a 12-wellplate (20). All siRNA, including FoxO3a (ON-TARGET-plus J-003007-10), PUMA (ACGTGTGACCACTGGCAT-TdTdT,andACCTCAACGCACAGTACGAdTdT;ref.31),

and Bim (GACCGAGAAGGUAGACAAUdTdT; ref. 32)were synthesized by Dharmacon (Lafayette), and siRNAswere transfected with lipofectamine 2000 (Invitrogen)according to the manufacturer’s instructions.

Luciferase assaysPUMA luciferase reporter constructs (Fragments A-E)

have been described previously (22). For reporter assays,cells were transfectedwith PUMA reporter alongwith thetransfection control b-galactosidase reporter pCMVb(Promega). Cell lysates were collected and luciferaseactivities were measured as previously described (33).All reporter experiments were carried out in triplicateand repeated three times.

Analysis of apoptosis and cell deathApoptosis was analyzed by nuclear staining with

Hoechst 33258 (Invitrogen), and Annexin V/propidiumiodide (Invitrogen) staining followed by flow cytometryas described (30). For colony formation assays, the samenumber of cells were treated and plated in 12-well platesat appropriate dilutions, and allowed to grow for 10 to 14days before stainingwith crystal violet (Sigma). For detec-tion of mitochondrial membrane potential change, thetreated cells were stained by JC-1 (30001, Biotium) for15 minutes according to the manufacturer’s instruction,and then analyzed by flow cytometry.

Analysis of cytochrome c release and Baxmultimerization, conformational change, andinteracting proteins

Cytoplasmic and mitochondrial fractions were sepa-rated by Mitochondrial Fractionation Kit (Active Motif)according to the manufacturer’s instructions. Cyto-chrome c in both cytoplasmic and mitochondrial frac-tions was detected by Western blotting. To detect Baxmultimerization, purified mitochondrial fractions werecross-linked with 1 mmol/L of Dithiobis (succinimidyl)propionate (DSP; Pierce) as described (34, 35), followedby Western blotting under nondenaturing conditions.Methods on detection of Bax conformational change (27)and interacting protein (16) have been described anddetails are found in the Supplementary Materials andMethods.

Xenograft studiesAll animal experiments were approved by the Univer-

sity of Pittsburgh Institutional Animal Care and UseCommittee (Pittsburgh, PA). Female 5- to 6-week-oldNu/Nu mice (Charles River) were housed in a sterileenvironment withmicroisolator cages and allowed accessto water and chow ad libitum. Mice were injected subcu-taneously in both flanks with 4 � 106 WT or PUMA-KOHCT116 cells. After implantation, tumorswere allowed togrow 7 days before treatment was initiated. Mice wererandomized into two groups (n ¼ 6 per group) receivingeither vehicle (saline) or 17-DMAG (15mg/kg/d) on days1 to 4 and 7 to 11, respectively (36). Detailed methods on

He et al.

Mol Cancer Ther; 12(11) November 2013 Molecular Cancer Therapeutics2560




tumormeasurements, harvests and histologic analysis arefound in the Supplementary Materials and Methods andas described (24, 31).

Statistical analysisStatistical analyses were carried out using GraphPad

Prism IV software. P values were calculated by the Stu-dent t test and were considered significant if P < 0.05. Themean � one SD are displayed in the figures.

Resultsp53-dependent induction of PUMA and Bax by17-AAG in colon cancer cells

To determine a potential role of the Bcl-2 family inHsp90 inhibitor–induced apoptosis, we first examinedprotein levels of its major members following 17-AAGtreatment in HCT116 colon cancer cells (Fig. 1A). Amongthem, PUMA, Bim, and Bax were significantly induced,and induction was detected as early as 24 hours after

Figure1. p53-dependentPUMAandBax inductionby17-AAG in coloncancer cells. Indicated cellswere treatedwith 1mmol/L 17-AAGandanalyzed for proteinor mRNA expression at indicated times. A, the expression of Bcl-2 family members was analyzed byWestern blotting. b-actin was used as a loading control.B, mRNA levels of PUMA, BIM, BAX were analyzed by real-time reverse transcriptase (RT) PCR. b-actin was used as a control. C, WT and p53-KOHCT116 and RKO cells were treated with 1 mmol/L 17-AAG for 48 hours. PUMA, Bim, Bax, and p53 expression was analyzed by Western blotting.D, top, schematic representation of the genomic structure of PUMA, highlighting the PUMA promoter fragments A–E used in the luciferase experiment. Twop53-binding sites (BS1/2) are indicated by vertical lines. Bottom,WT and p53-KOHCT116 cells were transfected overnight with a luciferase reporter plasmidand then treated with 1 mmol/L 17-AAG for 24 hours. Reporter activities were measured by luciferase assay. The ratios of normalized relative luciferaseactivities (to the empty vector pBV-Luc as 1) were plotted. E, the indicated proteins were analyzed by Western blotting.

p53 and PUMA in Hsp90 Inhibitor–Induced Apoptosis

www.aacrjournals.org Mol Cancer Ther; 12(11) November 2013 2561




treatment (Fig. 1A). The levels of PUMA, BIM, and BAXtranscripts were also elevated (Fig. 1B). The expression ofother Bcl-2 family members such as Bid, Bcl-2, and Bcl-xLremained unchanged, whereas that of Noxa and Mcl-1decreased (Fig. 1A).We then treatedparental, and p53-KOHCT116 or RKO cells with 17-AAG to determine whetherp53 is involved in the induction of PUMA, Bim, and Bax.Induction of PUMA and Bax, but not Bim, was largelyblocked in p53-KO cells, compared with parental cells(Fig. 1C). Furthermore, we used a series of PUMA pro-moter luciferase reporters (containing �2 kb upstreamfrom the transcriptional initiation site; ref. 22) to deter-mine whether p53 directly activates PUMA transcription.We found that the reporters containing the two p53-binding sites (Frag A and E) had much higher activitiesafter 17-AAG treatment in p53 WT cells than in p53-KOHCT116 cells (Fig. 1D).Moreover, the induction of PUMAand other p53 targets Bax and p21 paralleled that of p53,p53S15 phosphorylation, and g-H2AX, a marker of DNAdouble-strand breaks (Fig. 1E). Consistent with reports in

other systems (37, 38), 17-AAG treatment decreased AKTand p-ERK (Supplementary Fig. S1). These findings showthat 17-AAG treatment leads to p53-dependent inductionof PUMA and Bax, besides modulation of survivalpathways.

PUMA mediates Hsp90 inhibitor–induced apoptosisTo examine a potential role of PUMA in 17-AAG–

induced apoptosis, we compared the responses of paren-tal HCT116 cells with isogenic PUMA knockout (PUMA-KO) cells (27). 17-AAG treatment induced apoptosis inHCT116 cells as early as 24hours,which increased tomorethan 30% at 72 hours, whereas apoptosis and activation ofcaspase-3 and -9 were abolished in PUMA-KO cells (Fig.2A). Annexin V/propidium iodide staining confirmedapoptotic resistance of PUMA-KO cells (SupplementaryFig. S2A and S2B). PUMA-KO cells were also found to besignificantly more resistant to 17-AAG than WT cells in along-term clonogenic assay (Fig. 2B). PUMA knockdownsuppressed 17-AAG–induced apoptosis in p53 WT cell

Figure 2. PUMA is required forHsp90 inhibitor–inducedapoptosis. A, WT and PUMA-KOHCT116 cells were treated with 1mmol/L 17-AAG. Left, apoptosis atthe indicated times was analyzedfor by counting condensed andfragmented nuclei. �, P < 0.05;��, P < 0.001, WT versus PUMA-KO. Right, active caspase-3 andcaspase-9 at 72 hours wereanalyzed by Western blotting. B,colony formation assay was doneby seeding an equal number of WTand PUMA-KO HCT116 cellstreated with 1 mmol/L 17-AAG for48 hours in 12-well plates, and theattached cells were stained withcrystal violet after 14 days.Representative pictures ofcolonies (left) and quantification ofcolony numbers (right) are shown.��, P < 0.001, WT versus PUMA-KO. C, LoVo and RKO cells weretransfectedwith either a scrambledsiRNA orPUMA siRNA for 24 hoursand then treated with 1 mmol/L17-AAG for 48 hours. Top,apoptosis was analyzed bycounting condensed andfragmented nuclei. ��, P < 0.001,si-PUMA versus scrambled.Bottom, Western blottingconfirmed PUMA depletion bysiRNA. D, WT and PUMA-KOHCT116 cells were treated with0.25 mmol/L 17-DMAG and 0.5mmol/L NVP-AUY922 for 48 hours.Apoptosis was analyzed bycounting condensed andfragmented nuclei. ��, P < 0.001,WT versus PUMA-KO.

He et al.





lines, LoVo and RKO (Fig. 2C). In contrast, BIM knock-down did not inhibit 17-AAG–induced apoptosis (Sup-plementary Figs. S2C and S2D). Furthermore, 17-DMAGandNVP-AUY922 induced significant apoptosis in HCT116 cells, which was largely blocked in PUMA-KO cells(Fig. 2D and Supplementary Fig. S3). Collectively, theseresults indicate that PUMA is required for Hsp90 inhib-itor-induced apoptosis in colon cancer cells with WTp53.

PUMA mediates 17-AAG–induced apoptosis via themitochondrial pathway and Bax activation17-AAG–induced apoptosis in HCT 116 was associ-

ated with mitochondrial membrane depolarization andrelease of cytochrome c and SMAC, which was sup-

pressed in PUMA-KO cells (Fig. 3A). 17-AAG treatmentalso induced PUMA-dependent Bax conformationalchanges and oligomerization (Fig. 3B and C). BAX-KOHCT116 cells were found to be resistant to 17-AAG–induced apoptosis, and activation of caspase-3 and -9(Fig. 3D), consistent with PUMA acting upstream of Bax(13–15). We further probed interactions of Bax withseveral Bcl-2 family members. 17-AAG treatmentinduced dissociation of Bax from Bcl-xL and Bcl-2,which was suppressed in PUMA KO cells (Fig. 3E). Incontrast, Bim binding to Bax was largely unaffected byPUMA status or 17-AAG treatment. No interaction wasdetected between Bax and Mcl-1, PUMA or Noxa. Inter-estingly, levels of Mcl-1 and Noxa decreased signifi-cantly upon 17-AAG treatment in HCT 116 cells, which

Figure 3. PUMAmediates 17-AAG–

induced apoptosis via themitochondrial pathway and Baxactivation in HCT116 cells. Theindicated cells were treated with1 mmol/L 17-AAG or vehicle(untreated) for 48 hours. A, top,mitochondrial membrane potentialwas analyzed by flow cytometryfollowing staining with JC-1. Thecircled lower populations indicatedecreased red/green ratio andmembrane depolarization. Bottom,the distribution of cytochrome cand SMAC in mitochondrial orcytosolic fraction was analyzed byWestern blotting. a-Tubulin andcytochrome oxidase subunit IV(COX IV) were used as the controlfor fractionation. B, Baxconformational change wasdetected by immunopercipiation(IP) with anti-Bax 6A7 (activated)antibody followed by Westernblotting. C, Bax multimerization inisolated mitochondria wasanalyzedbyWesternblottingundernondenaturing conditionsfollowing DSP cross-link. D,apoptosis was analyzed bycounting condensed andfragmented nuclei (left) andWestern blotting of activecaspase-3, caspase-9 (bottom).b-actin was used as the control forloading. ��, P < 0.001, WT versusBAX-KO. E, interactions ofendogenous Bax with indicatedBcl-2 family members wereanalyzed by immunoprecipitation(IP) followed by Western blotting.The rabbit (N20, top) and mouse(mAb, top) anti-Bax antibodieswere used.






was reduced in PUMA KO cells (Figs. 1A and Fig. 3E).These results show that PUMA functions upstream ofBax to induce Bax activation and mitochondrial dys-function in 17-AAG–induced apoptosis. PUMA likelyactivates Bax indirectly by replacing it from Bcl-xL orBcl-2 (16, 39), and altered interactions among Bcl-2family might affect their stabilities.

p53-dependent p21 induction inhibits 17-AAG–induced apoptosis

p21 also showed p53-dependent induction following17-AAG treatment, whereas p53-KO cells were less resis-tant than PUMA-KO to 17-AAG–induced apoptosis (Figs.1E, 4A and 4B and Supplementary Fig. S4). We hypoth-esized that p21-inductionmight inhibit apoptosis inducedby Hsp90 inhibitors, much like it inhibits apoptosisinduced by DNA damage (40, 41). Indeed, p21-KO HCT116 cells were found to be more sensitive to 17-AAG–induced apoptosis comparedwith parental cells. Apopto-sis was significantly reduced in p21/PUMADKO cells thatphenocopied p53-KO cells (Fig. 4B and SupplementaryFig. S4). UCN-01 was shown to selectively block p53phosphorylation on S20 but not S15 by targeting Chk1after DNA damage (42, 43). We found that UCN-01eliminated p21 expression and significantly enhanced17-AAG–induced apoptosis (Fig. 4C). The apoptosisinduced by 17-AAG and UCN-01 combination was par-tially dependent on PUMA (Fig. 4D). Together, these datasuggest that p53-dependent induction of p21 suppressesthe apoptotic response to 17-AAG, andselective inhibitionof p21 while preserving PUMA induction enhances cellkilling.

Enhanced p53 activity synergizes with 17-AAG toinduce PUMA-dependent apoptosis

Chemosensitization effects of 17-AAGhave been exten-sively documented (1, 7). We found that HCT116 cellswere relatively resistant to cisplatin-induced apoptosis orPUMA expression, whereas p53 stabilization and serine15 (S15) phosphorylation were readily detected (Fig. 5Aand B). The combination of 17-AAG with cisplatin syn-ergistically induced apoptosis, caspase activation, andPUMA expression in HCT 116 cells (Fig. 5A and B). Thecombination of 17-AAGwith nutlin-3, a p53 activator thatdisrupts MDM2/MDMX and p53 interactions, also syn-ergistically induced apoptosis and PUMA induction(Fig. 5C and D). Remarkably, the combination treat-ment-induced apoptosis was almost completely abol-ished inPUMA-KOcells (Fig. 5AandC). PUMAinductionwas reduced in p53-KO cells compared withWTHCT 116cells (Fig. 5B and D). Loss of p21 induction likely explainswhy p53-KO cells were more sensitive than PUMA-KO toapoptosis induced by these combinations (Fig. 5A and Band Supplementary Fig. S5). In addition, p53-KO cellswere also partially resistant to apoptosis induced by 17-DMAG and NVP-AUY922, and showed significantlyreduced induction of PUMA and Bax, but not Bim (Sup-plementary Fig. S5). These data show a general role of thep53/PUMA/Bax axis in Hsp90 inhibitor–induced apo-ptosis and chemosensitization in colon cancer cells.

Antitumor activities of Hsp90 inhibitors aremediated by PUMA-dependent apoptosis in vivo

To determine whether PUMA-mediated apoptosisplays a critical role in the antitumor activities of Hsp90

Figure 4. p53-dependent p21induction suppresses 17-AAG–

induced apoptotic. A, HCT116 celllines with the indicated genotypeswere treatedwith 1 mmol/L 17-AAGfor 48 hours. Western blotting ofPUMA, p21, p53, and activecaspase-3. B, apoptosis wasanalyzed by counting condensedand fragmented nuclei in theindicated cell lines at 48 hours.��, P < 0.001, p21 KO versus WT,p53-KO, PUMA-KO, or p21/PUMA-DKO. C, HCT116 cells weretreated with 1 mmol/L UCN-01or 1 mmol/L 17-AAG or theircombination for 48 hours. Left,apoptosis was analyzed bycounting condensed andfragmented nuclei. Right, theexpression levels of total p53,p-p53 (S20), p21, and PUMA wereanalyzed by Western blotting. D,WT and PUMA-KO HCT116 cellswere treatedwith 1 mmol/L 17-AAGand 1 mmol/L UCN-01 for 48 hours.Apoptosis was analyzed bycounting condensed andfragmented nuclei. ��, P < 0.001,WT versus PUMA-KO.

He et al.





inhibitors in vivo, we established WT and PUMA-KOHCT116 xenograft tumors in nude mice. Tumor-bearingmice were then treated with 15 mg/kg 17-DMAG, whichis water soluble, or the vehicle intraperitoneally, andtumor volumes were monitored every 2 days for 3 weeks.The WT tumors, but not PUMA-KO HCT116 tumors,responded well to 17-DMAG treatment and reachedroughly 1/4 the size of the tumors in vehicle-treated miceon day 21 (Fig. 6A). PUMA, Bax, Bim, as well as p53phosphorylation (S15) were significantly induced by 17-DMAG treatment in WT tumors (Fig. 6B), consistent withour in vitro data. TUNEL and active caspase-3 stainingrevealed marked apoptosis in these tumors, whichdecreased by more than 55% in the PUMA-KO tumorstreated identically (Fig. 6C and D). These data showed animportant role of the p53/PUMA/Bax axis in the antitu-mor activities of 17-DMAG in vivo.

DiscussionHsp90 helps mammalian cells and lower eukaryotes

to cope with environmental stresses (44). Cancer cellsare constantly under a variety of stresses (3) andHsp90 is believed to help buffer such stresses for theirsurvival, while different client proteins might beinvolved depending on cellular context (1, 7). Hsp90inhibitors induce apoptosis in a variety of preclinicalmodels through degradation of oncogenic client pro-teins including mutant p53, Raf-1, Akt, Her-2, and ALKfusion proteins (1, 7, 45). Using multiple cell lines andstructurally unrelated compounds, we show a role of

the p53/PUMA/Bax axis in Hsp90 inhibitor–inducedapoptosis and chemosensitization in colon cancer cellsusing cell culture and xenograft models. p53-dependentPUMA expression directly activates Bax and the mito-chondrial apoptotic pathway. Modulation of other Bcl-2family members and survival pathways (37, 38), and ERstress (Supplementary Fig. S6) can also contribute toapoptosis or PUMA induction. However, the cell-killingmechanisms by Hsp90 inhibitors are likely different inp53 WT and mutant cancer cells where mutant p53 istargeted for degradation (46–48).

Several recent studies showed that functional p53 isrequired for the induction of apoptosis by 17-AAG (11,12), while the mechanism has not been well defined. Wediscovered distinct roles of p53-mediated PUMA and p21induction in this process. In contrast with p53-mediatedinduction of PUMA/Bax, induction of p21 protectedagainst 17-AAG–induced apoptosis, consistent with theknown antiapoptotic function of p21 in p53-dependentapoptosis (40, 41). The induction of g-H2AX further sug-gests a classical p53-dependent DNA damage responseassociated with double-strand breaks (Fig. 1E; refs. 40,41, 49), whereas the precise cause of DNA damage uponHsp90 inhibition remains to be defined. 17-AAG is com-monly used as a chemosensitizer and enhances cell killingwhen combinedwith a variety of cytotoxic agents, such ascisplatin, UCN-01, taxol, gemcitabine, cytarabine, or pro-teasome inhibitors and HDAC inhibitors (7). Our resultsprovidemechanistic insights for such a synergy inp53WTcells. For examples, cisplatin or nutlin-3 enhances 17-

Figure 5. Enhanced p53 activitysynergizes with 17-AAG to inducePUMA and apoptosis. A, WT,PUMA-KO, and p53-KO HCT116cells were treated with 1 mmol/L17-AAG, 10 mmol/L cisplatin, ortheir combination for 48 hours.Apoptosis was analyzed bycounting condensed andfragmented nuclei. B, Westernblotting of active caspase-3,PUMA, p53, and p-p53 (S15) forsamples in A. C, WT, PUMA-KO,and p53-KO HCT116 cells weretreated with 1 mmol/L 17-AAG or10 mmol/L nutlin-3 or theircombination for 48 hours.Apoptosis was analyzed bycounting condensed andfragmented nuclei. �, P < 0.01,WT versus p53-KO; ��, P < 0.001,WT versus PUMA-KO. D, Westernblotting of active caspase-3,PUMA, phosphorylation of p53on serine 15 for samples in C.






AAG–induced cell killing via enhanced p53 activity andPUMA/Bax induction. On the other hand, UCN-01enhances 17-AAG–induced cell killing by blocking p21expression. Therefore, selective modulation of p53-medi-ated apoptosis or cell-cycle arrest can be exploited toenhance cancer cell killing.

The results of Hsp90 inhibitors in clinical trials aremixed (1, 5), likely reflecting lack of patient stratificationor biomarkers. Improved response rates were observed inpatientswithHER2-positivemetastatic breast cancer trea-tedwith 17-AAGand trastuzumab (50), aswell as inALK-rearranged non–small cell lung cancer patients treatedwithHsp90 inhibitor IPI-504 (51). This has been attributedto the degradation of the Hsp90 clients as driver onco-proteins. Our study together with others (11, 12) wouldstrongly suggests that this class of agents works differ-ently in cancers with WT p53, and examining the modu-

lation of PUMA, Bax, Bim, and p21 might help betterdistinguish responders from nonresponders in thesetumors.Nomutation in thePUMAgenehas been reportedin human cancer, whereas reduced PUMAexpression hasbeen reported in cutaneous melanoma and Burkitt lym-phomas (18). Therefore reduced PUMA expression canpotentially contribute to resistance to Hsp90 inhibitors.Toxicity to normal tissues is also a concern of Hsp90inhibitors. It is tempting to speculate that blocking p53-dependent apoptosis might selectively reduce normaltissue toxicity without compromising efficacy in tumorsharboring mutant p53.

In conclusion,wehave shown that thep53/PUMA/Baxaxis plays an important role in the therapeutic andapoptotic response to Hsp90 inhibitors in p53WTcancer cells. Loss of PUMA, BAX, or p53 leads to resis-tance at varying degrees, whereas loss of p21 leads to

Figure 6. PUMA contributes to theantitumor activities of 17-DMAG ina xenograft model. A, nude miceafter 1 week of implantation of4 � 106 WT or PUMA-KO HCT116cells were treated with 15 mg/kg of17-DMAG or the control buffer byintraperitoneal injection on days 1to 4 and 7 to 11. Tumor volume atindicated time points aftertreatment was calculated andplotted (n ¼ 6 in each group). Topinset, representative tumors at theend of the experiment. B, HCT116WT xenograft tumors from micetreated with 15mg/kg of 17-DMAGor the control buffer for 5consecutive days. The indicatedproteins were analyzed byWesternblotting in representative tumors.C, paraffin-embedded WT andPUMA-KO tumor sections frommice treated as in B were analyzedby TUNEL staining. Left,representative TUNEL stainingpictures; right, TUNEL-positivecells were counted and plotted. D,paraffin-embedded WT andPUMA-KO tumor sections frommice treated as in B were analyzedby active caspase-3 staining. Left,representative staining pictures.Right, active caspase-3–positivecells were counted and plotted.��,P<0.001,WTversusPUMA-KO.

He et al.





sensitization. These findings provide novel mechanisticinsights for Hsp90 inhibitor–induced apoptosis, and canhave important implications for their future develop-ment and application.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: K. He, J. YuDevelopment of methodology: K. He, J. YuAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): K. He, J. YuAnalysis and interpretation of data (e.g., statistical analysis, biostatis-tics, computational analysis): K. He, X. Zheng, L. Zhang, J. YuWriting, review, and/or revision of themanuscript:K.He, X. Zheng, J. YuAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): K. He, X. Zheng, L. Zhang, J. YuStudy supervision: J. Yu

AcknowledgmentsThe authors thank Bert Vogelstein (Howard Hughes Medical Institute,

Johns Hopkins University) for p53-KO, p21-KOHCT116 cells, and p53-KORKO cells, and other members of Yu and Zhang laboratories for helpfuldiscussions.

Grant SupportThis work was supported by NIH grant CA129829, American Cancer

Society grant RGS-10-124-01-CCE, FAMRI (J. Yu), and by NIH grantsCA106348, CA121105, and American Cancer Society grant RSG-07-156-01-CNE (L. Zhang). This project used shared facilities thatwere supported,in part, by UPCI CCSG award P30CA047904.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received April 15, 2013; revised July 21, 2013; accepted August 14, 2013;published OnlineFirst August 21, 2013.

References1. Neckers L,Workman P. Hsp90molecular chaperone inhibitors: are we

there yet? Clin Cancer Res 2012;18:64–76.2. Kamal A, Thao L, Sensintaffar J, Zhang L, BoehmMF, Fritz LC, et al. A

high-affinity conformation of Hsp90 confers tumour selectivity onHsp90 inhibitors. Nature 2003;425:407–10.

3. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation.Cell 2011;144:646–74.

4. Ferrarini M, Heltai S, Zocchi MR, Rugarli C. Unusual expression andlocalization of heat-shock proteins in human tumor cells. Int J Cancer1992;51:613–9.

5. Workman P, Burrows F, Neckers L, Rosen N. Drugging the cancerchaperone HSP90: combinatorial therapeutic exploitation of onco-geneaddiction and tumor stress. AnnNYAcadSci 2007;1113:202–16.

6. Hostein I, Robertson D, DiStefano F, Workman P, Clarke PA. Inhibitionof signal transduction by the Hsp90 inhibitor 17-allylamino-17-demethoxygeldanamycin results in cytostasis and apoptosis. CancerRes 2001;61:4003–9.

7. Lu X, Xiao L,Wang L, RudenDM. Hsp90 inhibitors and drug resistancein cancer: the potential benefits of combination therapies of Hsp90inhibitors and other anti-cancer drugs. Biochem Pharmacol 2012;83:995–1004.

8. Munster PN, Marchion DC, Basso AD, Rosen N. Degradation of HER2by ansamycins induces growth arrest and apoptosis in cells with HER2overexpression via a HER3, phosphatidylinositol 30-kinase-AKT-dependent pathway. Cancer Res 2002;62:3132–7.

9. Rakitina TV, Vasilevskaya IA, O'Dwyer PJ. Additive interaction ofoxaliplatin and 17-allylamino-17-demethoxygeldanamycin in coloncancer cell lines results from inhibition of nuclear factor kappaBsignaling. Cancer Res 2003;63:8600–5.

10. Davenport EL, Moore HE, Dunlop AS, Sharp SY, Workman P, MorganGJ, et al. Heat shock protein inhibition is associated with activation ofthe unfolded protein response pathway in myeloma plasma cells.Blood 2007;110:2641–9.

11. Ayrault O, Godeny MD, Dillon C, Zindy F, Fitzgerald P, Roussel MF,et al. Inhibition of Hsp90 via 17-DMAG induces apoptosis in a p53-dependent manner to prevent medulloblastoma. Proc Natl Acad Sci US A 2009;106:17037–42.

12. Vaseva AV, Yallowitz AR, Marchenko ND, Xu S, Moll UM. Blockade ofHsp90 by 17AAG antagonizes MDMX and synergizes with Nutlin toinduce p53-mediated apoptosis in solid tumors. Cell DeathDis 2011;2:e156.

13. YouleRJ,StrasserA. TheBCL-2protein family: opposingactivities thatmediate cell death. Nat Rev Mol Cell Bio 2008;9:47–59.

14. Yu J, Zhang L. Apoptosis in human cancer cells. Curr Opin Oncol2004;16:19–24.

15. Leibowitz B, Yu J. Mitochondrial signaling in cell death via the Bcl-2family. Cancer Biol Ther 2010;9:417–22.

16. Ming L, Wang P, Bank A, Yu J, Zhang L. PUMA dissociates Bax andBCL-XL to induce apoptosis in colon cancer cells. J Biol Chem 2006;281:16034–42.

17. Yu J,Wang P, Ming L,WoodMA, Zhang L. SMAC/Diablomediates theproapoptotic function of PUMA by regulating PUMA-induced mito-chondrial events. Oncogene 2007;26:4189–98.

18. Yu J, Zhang L. PUMA, a potent killer with or without p53. Oncogene2008;27 Suppl 1:S71–83.

19. Yu J, Zhang L. No PUMA, no death: implications for p53-dependentapoptosis. Cancer Cell 2003;4:248–9.

20. Zheng X, He K, Zhang L, Yu J. Crizotinib induces PUMA-dependentapoptosis in colon cancer cells. Mol Cancer Ther 2013;12:777–86.

21. Melino G, Bernassola F, Ranalli M, Yee K, ZongWX, Corazzari M, et al.p73 Induces apoptosis via PUMA transactivation and Bax mitochon-drial translocation. J Biol Chem 2004;279:8076–83.

22. Ming L, Sakaida T, Yue W, Jha A, Zhang L, Yu J. Sp1 and p73activate PUMA following serum starvation. Carcinogenesis 2008;29:1878–84.

23. WangP, QiuW, DudgeonC, Liu H, HuangC, Zambetti GP, et al. PUMAis directly activated by NF-kappa B and contributes to TNF-alpha-induced apoptosis. Cell Death Differ 2009;16:1192–202.

24. DudgeonC,WangP,SunX,PengR,SunQ,YuJ, et al. PUMA inductionby FoxO3a mediates the anticancer activities of the broad-rangekinase inhibitor UCN-01. Mol Cancer Ther 2010;9:2893–902.

25. Sun J, SunQ, BrownMF,DudgeonC,Chandler J, Xu X, et al. Themulti-targeted kinase inhibitor sunitinib induces apoptosis in colon cancercells via PUMA. PLoS One 2012;7:e43158.

26. Bunz F, Hwang PM, Torrance C,Waldman T, Zhang Y, Dillehay L, et al.Disruption of p53 in human cancer cells alters the responses totherapeutic agents. J Clin Invest 1999;104:263–9.

27. Yu J, Wang Z, Kinzler KW, Vogelstein B, Zhang L. PUMAmediates theapoptotic response to p53 in colorectal cancer cells. Proc Natl AcadSci U S A 2003;100:1931–6.

28. Zhang L, Yu J, Park BH, Kinzler KW, Vogelstein B. Role of BAX in theapoptotic response to anticancer agents. Science 2000;290:989–92.

29. Waldman T, Kinzler KW, Vogelstein B. p21 is necessary for the p53-mediated G(1) arrest in human cancer cells. Cancer Res 1995;55:5187–90.

30. Wang P, Yu J, Zhang L. The nuclear function of p53 is required forPUMA-mediated apoptosis induced by DNA damage. Proc Natl AcadSci U S A 2007;104:4054–9.

31. Sun Q, Ming L, Thomas SM, Wang Y, Chen ZG, Ferris RL, et al. PUMAmediates EGFR tyrosine kinase inhibitor-induced apoptosis in headand neck cancer cells. Oncogene 2009;28:2348–57.

32. Han J, Goldstein LA, Gastman BR, RabinowichH. Interrelated roles forMcl-1 and BIM in regulation of TRAIL-mediated mitochondrial apo-ptosis. J Biol Chem 2006;281:10153–63.






33. Yu J, Zhang L, Hwang PM, Kinzler KW, Vogelstein B. PUMA inducesthe rapid apoptosis of colorectal cancer cells. Mol Cell 2001;7:673–82.

34. Wu B, Qiu W, Wang P, Yu H, Cheng T, Zambetti GP, et al. p53independent induction of PUMA mediates intestinal apoptosis inresponse to ischaemia-reperfusion. Gut 2007;56:645–54.

35. QiuW, Carson-Walter EB, Liu H, Epperly M, Greenberger JS, ZambettiGP, et al. PUMA regulates intestinal progenitor cell radiosensitivity andgastrointestinal syndrome. Cell Stem Cell 2008;2:576–83.

36. Moser C, Lang SA, Kainz S, Gaumann A, Fichtner-Feigl S, Koehl GE,et al. Blocking heat shock protein-90 inhibits the invasive propertiesand hepatic growth of human colon cancer cells and improves theefficacy of oxaliplatin in p53-deficient colon cancer tumors in vivo. MolCancer Ther 2007;6:2868–78.

37. Sato S, Fujita N, Tsuruo T. Modulation of Akt kinase activity by bindingto Hsp90. Proc Natl Acad Sci U S A 2000;97:10832–7.

38. Jia WT, Yu CR, Rahmani M, Krystal G, Sausville EA, Dent P, et al.Synergistic antileukemic interactions between 17-AAG and UCN-01involve interruption of RAF/MEK- and AKT-related pathways. Blood2003;102:1824–32.

39. Peng R, Tong JS, Li H, Yue B, Zou F, Yu J, et al. Targeting Baxinteraction sites reveals that only homo-oligomerization sites areessential for its activation. Cell Death Differ 2013;20:744–54.

40. Yu J, Zhang L. The transcriptional targets of p53 in apoptosis control.Biochem Biophys Res Commun 2005;331:851–8.

41. Vousden KH, Prives C. Blinded by the light: the growing complexity ofp53. Cell 2009;137:413–31.

42. YuQ, LaRose J, ZhangH, TakemuraH,KohnKW,PommierY.UCN-01inhibits p53 up-regulation and abrogates gamma-radiation-inducedG(2)-M checkpoint independently of p53 by targeting both of the check-point kinases, Chk2 and Chk1. Cancer Res 2002;62:5743–8.

43. Arlander SJH, Eapen AK, Vroman BT, McDonald RJ, Toft DO, KarnitzLM. Hsp90 inhibition depletes Chk1 and sensitizes tumor cells toreplication stress. J Biol Chem 2003;278:52572–7.

44. Jarosz DF, Lindquist S. Hsp90 and environmental stress transform theadaptive value of natural genetic variation. Science 2010;330:1820–4.

45. Porter JR, Fritz CC, Depew KM. Discovery and development of Hsp90inhibitors: a promising pathway for cancer therapy. Curr Opin ChemBiol 2010;14:412–20.

46. Peng Y, Chen L, Li C, Lu W, Chen J. Inhibition of MDM2 by hsp90contributes to mutant p53 stabilization. J Biol Chem 2001;276:40583–90.

47. Muller P, Hrstka R, Coomber D, Lane DP, Vojtesek B. Chaperone-dependent stabilization and degradation of p53 mutants. Oncogene2008;27:3371–83.

48. Li D, Marchenko ND, Schulz R, Fischer V, Velasco-Hernandez T, TalosF, et al. Functional inactivation of endogenous MDM2 and CHIP byHSP90 causes aberrant stabilization of mutant p53 in human cancercells. Mol Cancer Res 2011;9:577–88.

49. Leibowitz BJ, Qiu W, Liu H, Cheng T, Zhang L, Yu J. Uncoupling p53functions in radiation-induced intestinal damage via PUMA and p21.Mol Cancer Res 2011;9:616–25.

50. Modi S, Stopeck A, Linden H, Solit D, Chandarlapaty S, Rosen N, et al.HSP90 inhibition is effective in breast cancer: a phase II trial oftanespimycin (17-AAG) plus trastuzumab in patients with HER2-pos-itive metastatic breast cancer progressing on trastuzumab. Clin Can-cer Res 2011;17:5132–9.

51. Sequist LV, Gettinger S, Senzer NN,Martins RG, Janne PA, LilenbaumR, et al. Activity of IPI-504, a novel heat-shock protein 90 inhibitor, inpatients with molecularly defined non-small-cell lung cancer. J ClinOncol 2010;28:4953–60.

He et al.





2013;12:2559-2568. Published OnlineFirst August 21, 2013.Mol Cancer Ther Kan He, Xingnan Zheng, Lin Zhang, et al. and BaxHsp90 Inhibitors Promote p53-Dependent Apoptosis through PUMA

Updated version

10.1158/1535-7163.MCT-13-0284doi:

Access the most recent version of this article at:

Material

Supplementary

http://mct.aacrjournals.org/content/suppl/2013/08/21/1535-7163.MCT-13-0284.DC1

Access the most recent supplemental material at:

Cited articles

http://mct.aacrjournals.org/content/12/11/2559.full#ref-list-1

This article cites 51 articles, 27 of which you can access for free at:

Citing articles

http://mct.aacrjournals.org/content/12/11/2559.full#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mct.aacrjournals.org/content/12/11/2559To request permission to re-use all or part of this article, use this link



http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-13-0284

http://mct.aacrjournals.org/content/suppl/2013/08/21/1535-7163.MCT-13-0284.DC1

http://mct.aacrjournals.org/content/12/11/2559.full#ref-list-1

http://mct.aacrjournals.org/content/12/11/2559.full#related-urls

http://mct.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mct.aacrjournals.org/content/12/11/2559


hsp90inhibitorspromotep53 …...cancer therapeutics insights...

Documents