Human Ovarian Cancer and Cisplatin Resistance: Possible Role of Inhibitor of Apoptosis Proteins 1

Human Ovarian Cancer and Cisplatin Resistance:Possible Role of Inhibitor of Apoptosis Proteins*

JULANG LI†, QIANG FENG, JONG-MIN KIM‡, DANIELLE SCHNEIDERMAN§,PETER LISTON, MING LI, BARBARA VANDERHYDEN, WYLAM FAUGHT,MICHAEL FUNG KEE FUNG, MARY SENTERMAN, ROBERT G. KORNELUK, AND

BENJAMIN K. TSANG

Reproductive Biology Unit (J.L., Q.F., J.-M.K., D.S., M.L., B.V., B.K.T.) and Division of GynecologicOncology (B.V., W.F., M.F.K.F., M.S., B.K.T.), Departments of Obstetrics and Gynecology, Cellular andMolecular Medicine (J.L., D.S., B.V., B.K.T.) and Pathology (M.S.), University of Ottawa; Loeb HealthResearch Institute (J.L., Q.F., D.S., J.-M.K., M.L., B.K.T.), The Ottawa Hospital (W.F., M.F.K.F., M.S.,B.K.T.), ApoptoGen, Inc. (P.L., R.G.K.), and Children’s Hospital of Eastern Ontario, Ottawa RegionalCancer Center (B.V.), Ottawa, Ontario, Canada K1Y 4E9

ABSTRACTThe inhibitor of apoptosis proteins (IAPs) constitutes a family of

highly conserved apoptosis suppressor proteins that were originallyidentified in baculoviruses. Although IAP homologs have recentlybeen demonstrated to suppress apoptosis in mammalian cells, theirexpression and role in human ovarian epithelial cancer and chemo-therapy resistance are unknown. In the present study we used cis-platin-sensitive and -resistant human ovarian surface epithelial(hOSE) cancer cell lines and adenoviral antisense and sense comple-mentary DNA expression to examine the role of IAP in the regulationof apoptosis in human ovarian cancer cells and chemoresistance.Antisense down-regulation of X-linked inhibitor of apoptosis protein(Xiap), but not human inhibitor of apoptosis protein-2 (Hiap-2), in-duced apoptosis in cisplatin-sensitive and, to a lesser extent, in -re-sistant cells. Cisplatin consistently decreased Xiap content and in-

duced apoptosis in the cisplatin-sensitive, but not cisplatin-resistant,cells. Hiap-2 expression was either unaffected or inhibited to a lesserextent. The inhibition of IAP protein expression and induction ofapoptosis by cisplatin was time and concentration dependent. Infec-tion of cisplatin-sensitive cells with adenoviral sense Xiap comple-mentary DNA resulted in overexpression of Xiap and markedly atten-uated the ability of cisplatin to induce apoptosis. Immunohistochemicallocalization of the IAPs in hOSE tumors demonstrated the presence ofXiap and Hiap-2, with their levels being highest in proliferative, but notapoptotic, epithelial cells. These studies indicate that Xiap is an impor-tant element in the control of ovarian tumor growth and may be a pointof regulation for cisplatin in the induction of apoptosis. These resultssuggest that the ability of cisplatin to down-regulate Xiap content maybe an important determinant of chemosensitivity in hOSE cancer. (En-docrinology 142: 370–380, 2001)

HUMAN OVARIAN surface epithelial cancer (hOSE)ranks fifth among the most common female cancers

and is the leading cause of death from gynecologic malig-nancy in the western world. Although the clinical and his-tological prognostic factors (e.g. tumor grade and surgicalstage) are well understood, less is known about the biologicalprocess that leads to uncontrolled cellular growth. The con-trol of cell number during tissue growth is thought to be theresult of a balance of cell proliferation and cell death.Whereas cisplatin is currently a front line chemotherapeuticagent for ovarian epithelial cancer, chemoresistance remainsa major barrier to successful therapy. Ovarian epithelial can-cer cell apoptosis has been demonstrated to be involved in

cisplatin-induced cellular responses (1, 2). The action of cis-platin is thought to involve the formation of inter- and in-trastrand DNA cross-links (3), although the events leading tocell death after cisplatin treatment are unclear. Understand-ing the molecular mechanism by which this drug induces celldeath should provide a fundamental approach for increasingthe sensitivity of cells to this anti-cancer agent.

Apoptosis plays an important role in the maintenance ofphysiological homeostasis in response to stimuli that indi-cate that a cell is potentially harmful or abnormal (4–6).When the apoptosis machinery fails, abnormal cells can sur-vive, and unopposed tissue growth, as in the case of cancer,can result. Thus, it is conceivable that carcinomas may becaused or promoted in part by factors inhibiting cell death.In this regard, considerable work has been focused on therole of bcl-2 (7–9) as a negative regulator of apoptotic celldeath (10). In ovarian cancer, expression of the bcl-2 gene isan important modulator of drug-induced apoptosis and apotential determinant of chemoresistance (8) and survival(9). However, evidence also indicates that bcl-2 overexpres-sion or mutation cannot adequately account for the etiologyof existing ovarian cancer (7, 11), suggesting that other cellsurvival factors may also be involved in this complexprocess.

Inhibitor of apoptosis proteins (IAPs) were first identified

Received August 9, 2000Address all correspondence and requests for reprints to: Benjamin K.

Tsang, Ph.D., Loeb Health Research Institute, The Ottawa Hospital (Civ-ic Campus), 725 Parkdale Avenue, Ottawa, Ontario, Canada K1Y 4E9.E-mail: [email protected].

* This work was supported by grants from the Canadian Institutes ofHealth Research (MOP-15691), the University of Ottawa-IndustryGrants Program, and the Ottawa Civic Hospital Foundation.

† Recipient of a Cancer Research Society Postdoctoral Fellowship.‡ Recipient of a Corinne Boyer Fund for Ovarian Cancer Research

Postdoctoral Fellowship.§ Recipient of a Natural Science and Engineering Research Council of

Canada Graduate Scholarship.

0013-7227/01/$03.00/0 Vol. 142, No. 1Endocrinology Printed in U.S.A.Copyright © 2001 by The Endocrine Society

370

in baculoviruses, where they function to keep the host cellalive while the viruses replicate (12, 13). Five IAPs haverecently been identified in mammalian cells: neuronal apo-ptosis inhibitory protein (Naip) (14), X-linked inhibitor ofapoptosis protein (Xiap) (15–17), human inhibitor of apo-ptosis protein-1 (Hiap-1) (15, 16, 18), Hiap-2 (15, 16, 18), andsurvivin (19). The apoptosis-suppressive actions of mamma-lian IAPs have been previously shown. For example, theoverexpression of IAPs, including Naip, Xiap, Hiap-1, andHiap-2, protects Chinese hamster ovary (CHO) and Rat-1cells from apoptosis triggered by menadione, a potent in-ducer of free radicals, or by growth factor withdrawal (15).Overexpression of Xiap or Hiap-2 protected HeLa cells fromapoptosis induced by transient transfection of prointerleu-kin-1b-converting enzyme (16), and Xiap was shown to sup-press apoptosis induced by Sindbis virus (17). In most ofthese studies, the protection mediated by the IAPs was com-parable to that mediated by known antiapoptotic proteins,such as Bcl-2.

In the present study we examined the possible role of IAPsin the regulation of apoptosis in hOSE cancer cells and thepossible involvement of these proteins in the action of thechemotherapeutic agent cisplatin and in chemoresistance.This communication represents the first report supporting animportant role for Xiap in chemoresistance in human cancer.

Materials and MethodsCell lines and culture

hOSE cancer cells were cultured in DMEM/Ham’s F-12 (for A2780sand A2780cp) or RPMI 1640 (for OV2008 and C13 cells) supplementedwith 10% (vol/vol) FBS, nonessential amino acids (0.1 mm), penicillin(100 U/ml), and streptomycin (100 mg/ml) at 37 C under 5% CO2 and95% air. After a 24-h cell plating period, the culture medium waschanged, and cisplatin was added, and the cells were cultured for up toan additional 24 h. These cisplatin-sensitive (OV2008 and A2780s) celllines were derived from ovarian carcinomas from two separate patientswithout prior chemotherapy, and their respective resistant variants (C13and A2780cp) were established after in vitro cisplatin challenges (20–22).These cell lines have been previously used as a model for ovarian cancerresearch (21, 23–29). At the end of the culture period, cells weretrypsinized and aliquoted for the assessment of nuclear morphology andprotein extraction. Cell number in each treatment group was determinedby hemocytometry. Cell viability was determined by the trypan blue dyeexclusion test as previously described (30).

Adenoviral infection

After 24 h of plating, OV2008 and C13 were infected with an adeno-viral Xiap or Hiap-2 sense or antisense at a multiplicity of infection(MOI) of 5 or 10. The infection efficiency at a MOI of 5, evaluated byX-galactosidase test, was approximately 90% for both cell lines. Cellswere either infected with adenoviral sense Xiap complementary DNA(cDNA; MOI 5 5) for 48 h before the addition of cisplatin or for 60–72h with the adenoviral Xiap antisense (MOI 5 5 and 10). To prepare theadenoviral expression vectors for the full-length sense and antisensecDNAs of Xiap and Hiap-2, the open reading frames (ORFs) of Xiap andHiap-2 were PCR amplified with primers corresponding to the amino-and carboxyl-termini. These PCR products were cloned in the pCR2.1vector (Invitrogen, San Diego, CA), and sequenced. The ORFs were cutout with EcoRI, blunt ended with Klenow fragment of DNA polymeraseI, and ligated into Swa-1-cut pAdex1CAwt cosmid DNA. Packaging wasperformed with the Promega Corp. cosmid packaging extracts (Madi-son, WI) and used to infect Escherichia coli. Colonies were picked andscreened for the presence of the insert in both the sense and antisenseorientations relative to the chicken b-actin promoter. CsCl-purified cos-mid DNA was cotransfected with wild-type adenovirus DNA that con-

tains the terminal protein complexed to the ends of the DNA. Wild-typeadenovirus DNA was cut with Nsi-1, such that only homologous re-combination with the cosmid DNA will generate infectious adenovirusDNA. The final recombinant adenovirus contains a linear, doublestranded genome of 44,820 bp plus the insert size (;1500 bp for Xiap,;1800 bp for Hiap-2). This method has been shown to be an efficientmeans of generating recombinant adenoviruses (31).

Western blot

Total cell protein extracts were prepared as follows. Cells were son-icated (8 sec/cycle, three cycles) on ice in 10 mm HEPES buffer (pH 7.4)containing 1 mm EGTA and 2 mm phenylmethylsulfonylfluoride. Thesonicates were stored at 220 C until electrophoretic analyses were per-formed. Protein concentration was determined by protein assay (Bio-Rad Laboratories, Inc., Richmond, CA). Equal amounts of proteins(60–80 mg, depending on individual experiments) in cell extracts wereresolved by SDS-PAGE and transferred to nitrocellulose membrane (32).Membranes were blocked with 5% nonfat milk and subsequently incu-bated with rabbit polyclonal antihuman Xiap or Hiap-2 antibody (1:2000and 1:1000 dilutions, respectively). An ECL kit (Amersham PharmaciaBiotech) was used to visualize immunopositive proteins. To ascertainwhether protein loading to each lane was even within each acrylamidegel, the overall protein bands in the gels were routinely stained withCoomassie blue after electrotransfer and analyzed densitometrically forstaining intensity. Results from experiments with uneven protein load-ing were excluded in the final analysis.

To generate polyclonal antibodies for the present studies, full-lengthXiap- or Hiap-2-coding regions were ligated into the pGEX (Pharmacia-LKB, Piscataway, NJ) vector system and expressed as glutathione-S-transferase (GST) fusion proteins in Escherichia coli. The recombinantproteins were affinity purified using glutathione-Sepharose and used toimmunize New Zealand White rabbits. Fifty micrograms of protein weresuspended in 0.6 ml RIBI adjuvant (Sigma, St. Louis, MO) and injectedin 0.1-ml aliquots at each of six sites. Rabbits were injected weekly for4 weeks, at which point harvesting of serum was initiated. Antibody wasaffinity purified using a GST-Xiap or GST-Hiap-2 glutathione-Sepharosecolumn containing a minimum of 400 mg recombinant protein. Depletedantibody was prepared by passing the affinity-purified antibody overthe same column and collecting the flow-through. Studies on the spec-ificity of these antibodies indicated that absorption of the respectiveantibody on affinity columns effectively ablated the signals of the IAPproteins on Western blots and that no cross-reactivity with other IAPswas evident (Fig. 1). The use of rabbits in the generation of the antibodieswas approved by the University of Ottawa animal care committee andwas in compliance with the Guide to the Care and Use of ExperimentalAnimals of the Canadian Council on Animal Care.

Nuclear staining

Cells were fixed in 4% neutral buffered formalin and then resus-pended in Hoechst 33248 staining solution (0.1 mg/ml, 10 min). Cellswith typical apoptotic nuclear morphology were identified and counted.

Preparation of hOSE carcinomas, immunohistochemistry,and in situ localization of apoptotic cells

Samples of hOSE carcinomas were obtained fresh as pathological spec-imens from three patients undergoing their initial laparotomy for surgicalstaging and tumor debulking. The patients (56–76 yr old) had stage III(disseminated intraabdominal) disease, but no previous chemotherapy (Ta-ble 1). The tumor tissues were fixed and paraffin embedded by routineprocedures. Xiap, Hiap-2, and proliferating cell nuclear antigen (PCNA; anauxiliary protein of DNA polymerase a highly expressed at the G1/Sinterphase) was immunolocalized as previously reported (32). Briefly, thetumor sections were incubated in series with H2O2, normal goat serum, andrabbit polyclonal antihuman Xiap (0.5 mg/ml), Hiap-2 (0.5 mg/ml), ormouse monoclonal antihuman PCNA (0.5 mg/ml; Santa Cruz Biotechnol-ogy, Inc., Santa Cruz, CA) antibodies. Control tumor sections carried outwith the omission of the primary antibody exhibited negative immuno-staining. Immunoreactivities were detected by a Vector ABC Elite Kit (Vec-tor Laboratories, Inc., Burlingame, CA). Terminal deoxynucleotidyl trans-ferase-mediated dUTP-biotin end labeling (TUNEL) was performed as

IAPs IN OVARIAN CANCER AND CHEMORESISTANCE 371

described previously (33). Briefly, tumor sections were incubated in serieswith proteinase K, H2O2, and terminal deoxynucleotidyltransferase plusbiotinylated 16-deoxy-UTP. The incorporated deoxy-UTP molecules werevisualized with horseradish peroxidase-conjugated streptavidin anddiaminobenzidene.

Statistical analysis

Results were expressed as the mean 6 sem of three or four experi-ments. Statistical analysis were carried out by one- or two-way ANOVAand Tukey’s test. When necessary, data were logarithmically trans-formed to remove heterogeneity of variance before analysis. Statisticalsignificance was inferred at P , 0.05.

Results

Adenoviral Xiap antisense expression induces apoptosis inhOSE cancer cells

To test whether IAPs are indeed key elements in the reg-ulation of apoptosis in ovarian cancer cells, cisplatin-sensi-tive hOSE cancer cells (OV2008) were infected with eitheradenoviral Xiap antisense or vector expressing LacZ (as con-trol) for up to 60 h, and changes in cell morphology, apoptoticcell number, cell viability, as well as total cell number weredetermined. Xiap antisense infection significantly increased

FIG. 1. Immunospecificity of polyclonal antihuman Xiap and Hiap-2 antibodies. Expression plasmids for Xiap, Hiap1, Hiap2, mouse IAP-2(Miap2), and rat IAP-2 (Riap2) were generated by blunt end ligation of the respective coding regions into the sma1 site of pCI (PromegaCorp.). HeLa cells were transfected with Lipofectamine Plus (Life Technologies, Inc., Grand Island, NY) and harvested at 48 h post-transfection. Protein samples were quantitated and electrophoresed on 10% SDS-polyacrylamide gels. A, Western blot analysis ofendogenous and transfected XIAP protein in HeLa cells. Primary antibody was rabbit polyclonal anti-GST-XIAP used at 1:2000. Secondaryantibody was goat antirabbit HRP (Amersham Pharmacia Biotech, Arlington Heights, IL). Proteins were visualized with the ECL system.The migration of the approximately 55-kDa XIAP protein is indicated. B, A sample of the anti-XIAP antibody was affinity depleted ona column containing 100 mg immobilized GST-XIAP protein. The flow-through was reapplied three times before being used on the sameplot presented in A. The blot was developed for the same length of time with ECL reagent as in A. C, Western blot analysis of transfectedHiap1 and Hiap2 proteins in HeLa cells. Primary antibody was rabbit polyclonal anti-GT-Hiap2 used at a 1:2000 dilution. Secondaryantibody was goat antirabbit horseradish peroxidase(Amersham Pharmacia Biotech). Proteins were visualized with the ECL system. Themigration of the approximately 68-kDa Hiap2 protein is indicated. Note that the antibody does not detect the Hiap1 protein, but doescross-react with both mouse and rat homologues of Hiap2. D, A sample of the anti-Hiap2 antibody was affinity depleted on a columncontaining 100 mg immobilized GST-Hiap2 protein. The flow-through was reapplied three times before being used on the same plotpresented in C. Note that virtually all signal was lost using the same exposure as in C.

372 IAPs IN OVARIAN CANCER AND CHEMORESISTANCE Endo • 2001Vol. 142 • No. 1

the percentage of dead cells compared with the control value(vector; P , 0.001), as determined by trypan blue exclusiontests (Fig. 2, top left panel). Although there also appeared tobe a slight increase in percentage of dead cells with Hiap-2antisense infection, it was not statistically significant (Fig. 2,top left panel; P . 0.05). Infection of the cisplatin-resistantvariant of OV2008 cells (C13) with antisense of Xiap, but notof Hiap-2, also significantly, although to a lesser extent, de-creased cell viability (Fig. 2, top right panel). The cell deathinduced in both OV2008 and C13 by Xiap antisense was alsoaccompanied by decreases in total cell number (Fig. 2, bottompanels). The decrease in cell number could not be completelyaccounted for by increased cell death, as Xiap antisense in-

fection (especially at MOI 5 10) produced more cell death inOV2008 than C13, whereas the total cell numbers for both celllines were similar. In addition, although Xiap antisense in-fection (MOI 5 5) failed to significantly induced cell deathin C13 cells (;2.5%), it suppressed cell growth by 25% (P ,0.001). These findings suggest that ovarian epithelial cancercell proliferation may possibly be also under the regulationof Xiap and that this response may be attenuated by Xiapdown-regulation.

Xiap antisense infection of OV2008 decreased Xiap proteincontent, induced extensive cell detachment (Fig. 3A, b, blackarrows) and nuclear fragmentation (white arrows), and in-creased the number of apoptotic cells as well as the abun-

TABLE 1. Clinical and pathological features of three human ovarian surface epithelial carcinomas

Specimen Age ofpatients (yr) Distribution of disease Histologic diagnosis Surgical procedure

OC-4 56 RO, RT, omentum Grade III endometrioid carcinoma TAH, BSO omentectomyOC-7 56 RO, LO, RT, LT serosa, uterine

serosa, omentumGrade III serous carcinoma TAH, BSO omentectomy

OC-12 76 RO, LO, uterine & transverse colonserosa, omentum, umbilicus

Grade III serous carcinomaa TAH, BSO, omentectomy, resection oftransverse colon 1 umbilicus

RO, Right ovary; LO, left ovary; RT, right fallopian tube; LT, left fallopian tube; TAH, total abdominal hysterectomy; BSO, bilateralsalpingoophorectomy.

a Tumor showed heterogeneity in the degree of differentiation.

FIG. 2. Effect of down-regulation ofXiap and Hiap-2 on hOSE cancer cellviability and number. Cisplatin-sensi-tive OV2008 cells and its resistant vari-ant (C13) were infected with adenovirus(MOI 5 5 or 10) containing antisenseXiap or Hiap-2 cDNA or LacZ (as vectorcontrol) for 60 h. Top panels, Dead cellsas a percentage of the total cell popu-lation; lower panels, total cell number atthe end of the infection period. Datarepresent the mean 6 SEM of four ex-periments. **, P , 0.01; ***, P , 0.001(compared with LacZ group).


dance of apoptotic bodies (Fig. 3B, b and d compared witha and c). ANOVA indicates that there were highly significanteffects of the antisense on Xiap protein content (P , 0.01; Fig.3, D and E) and apoptosis (P , 0.001; Fig. 3C). Indeed,infection of these cells with a higher titer of the adenoviral

antisense (MOI 5 10) further increased the number of cellsundergoing apoptosis (Fig. 3C).

Cisplatin down-regulates IAP expression and inducesapoptosis in chemo-sensitive but not -resistant hOSE cancercells in vitro

To study whether IAP expression is a target for the che-motherapeutic action of cisplatin, OV2008 and C13 cells werecultured in the absence and presence of cisplatin (10–30 mm)for 24 h, and apoptosis and Naip, Xiap, and Hiap-2 expres-sion levels were assessed morphologically and by Westernanalysis, respectively. As the 150-kDa Naip protein wasbarely detectable regardless of the treatment (data notshown), subsequent studies were focused on Xiap andHiap-2. Like adenoviral Xiap antisense infection, the pres-ence of cisplatin induced morphological features of apopto-sis in OV2008 cells, including decreased cell volume, chro-matin condensation, and nuclear fragmentation (Fig. 4A),and was accompanied by decreased IAP contents (Fig. 5). The

FIG. 3. Influence of Xiap down-regulation on general (A; phase con-trast; black arrows indicate cell detachment) and nuclear (B; Hoechststaining; white arrows show nuclear fragmentation) morphology andon the extent of apoptosis (C) in OV2008 cells after 60 h of adenoviralinfection [MOI 5 5 (a and b); MOI 5 10 (c and d)] with adenoviralantisense Xiap (A and B, b and d) or LacZ (vector control; A and B,a and c). Magnification, 3400. D, A representative Western blot filtershowing effective antisense infection. E, Changes in Xiap proteincontent as analyzed densitometrically. The data in C and E representthe mean 6 SEM of three or four experiments. *, P , 0.05; **, P , 0.01(compared with respective LacZ group).

FIG. 4. Cisplatin-induced apoptosis in cisplatin-sensitive (OV2008),but not cisplatin-resistant (C13), hOSE cancer cells during a 24-hculture period. A, Apoptotic nuclear morphology (Hoechst staining;magnification, 3400; arrows show fragmented nuclei); B, concentra-tion-response data (extent of apoptosis; mean 6 SEM) from threeexperiments. **, P , 0.01 (compared with control).


increase in apoptotic cell number in response to cisplatin wasconcentration dependent and significant (50% vs. 2%; P ,0.05) even at a concentration of 10 mm cisplatin (Fig. 4B).However, cisplatin failed to induce apoptosis in C13 cells inthis concentration range (Fig. 4B). In addition, time-coursestudies on the induction of apoptosis by cisplatin in both celllines indicated that although OV2008 cells were responsive[apoptotic cells, ,20% (6 h), ;45% (12 h), and ;80% (24 h)]to the anticancer agent (30 mm), it had no significant effect onits resistant variant (C13; ,2% at all times) regardless of theculture period.

As shown in Fig. 5, although both Xiap and Hiap-2 arepresent in this cisplatin-sensitive hOSE cancer cell line (55 and68 kDa, respectively), their cellular contents were decreased bycisplatin in a concentration-dependent manner, with Xiap ap-pearing more responsive to the anticancer agent. Although theXiap level was decreased by almost 80% (P , 0.01) in thepresence of 20 mm cisplatin, the decrease in Hiap-2 proteincontent was not statistically significant at concentrations lowerthan 30 mm. However, the expression of Xiap and Hiap-2 in C13(the cisplatin-resistant variant of OV2008) was not suppressedby cisplatin (Fig. 5), and no morphological characteristics ofapoptosis could be detected (Fig. 4A). Although Xiap andHiap-2 contents in C13 appeared to be higher in the presenceof the anticancer agent, the differences were statistically non-significant (P . 0.05). Time-course experiments on IAP expres-sion demonstrated that the suppression of Xiap and Hiap-2protein levels in OV2008 by cisplatin was time dependent; a

significant decrease was observed between 12–24 h of culture(Fig. 6). Expression of the IAPs in C13 cells was not influencedby cisplatin regardless of the duration of treatment.

To determine whether the observed Xiap responses inOV2008 and C13 cells were unique to this pair of cell lines, theinfluence of cisplatin in vitro on Xiap and Hiap-2 protein con-tents in another cisplatin-sensitive hOSE cancer cell line(A2780s) and its resistant variant (A2780cp) was studied (Fig.7). Interestingly, whereas the Hiap-2 content in both the sen-sitive and resistant cells was not significantly altered by thepresence of the cisplatin, the level of Xiap was markedly de-creased in A2780s, but not in A2780cp (as in OV2008 and C13cells), in the presence of the chemotherapeutic agent at 30 mm(Fig. 7). Although cisplatin significantly increased the incidenceof apoptosis in the resistant cells, and the response was higherthan that observed in C13 cells under identical conditions, it wasconsiderably lower than that of its cisplatin-sensitive counter-part (OV2008; 7% vs. 50%). Taken together, these data suggestthat the apoptotic responsiveness of ovarian cancer cells tocisplatin may be related to the ability of the chemotherapeuticagent to down-regulate Xiap expression and that Hiap-2 mayplay a minor role or no role in cisplatin-induced apoptosis.

Xiap overexpression prevents cisplatin-induced apoptosis inchemo-sensitive hOSE cancer cells in vitro

To determine whether Xiap expression is indeed an im-portant determinant in chemo-resistance in hOSE cancer, the

FIG. 5. Concentration-dependent decreases in Xiap and Hiap-2 protein contents in the presence of cisplatin. Cisplatin-sensitive hOSE cancercells (OV2008) and its resistant variant (C13) were cultured with different concentration of cisplatin (0–30 mM) for 24 h, and equal amountsof solubilized cell proteins were analyzed by Western blot. Representative filters (A) and densitometric scans (B; mean 6 SEM) of Xiap and Hiap-2protein contents from three experiments are shown. *, P , 0.05; **, P , 0.01 (compared with control).


influence of cisplatin on Xiap protein content and apoptosisin OV2008 cells after adenoviral Xiap sense infection wasinvestigated (Fig. 8). Although cisplatin reduced Xiap pro-tein content in OV2008 cells infected with adenoviral LacZ(control vector; LacZ 1 cisplatin), overexpression of the pro-tein with adenoviral sense Xiap cDNA 48 h before treatmentwith the chemotherapeutic agent in vitro attenuated the cis-platin effects not only on Xiap protein expression (Fig. 8, Cand D), but also on apoptotic nuclear fragmentation (Fig. 8A,d compared with c, and Fig. 8B), suggesting that Xiap maybe an important element in hOSE cancer chemoresistance.

In situ immunolocalization of Xiap and Hiap-2 in humanovarian carcinomas

The in vitro studies with hOSE cancer cell lines stronglysuggest an important role of IAPs, particularly of Xiap, in thecontrol of apoptosis in human ovarian cancer. To determinewhether IAPs are indeed expressed in ovarian carcinomasand thus of clinical relevance, Xiap and Hiap-2 were immu-nolocalized in hOSE tumors obtained as pathological sam-ples from patients during surgical debulking, using poly-clonal antibodies against human Xiap and Hiap-2,respectively (Table 1). In addition, in situ TUNEL and im-munohistochemistry for PCNA (an auxiliary protein of DNApolymerase a highly expressed at the G1/S interphase) were

performed to examine whether and how the expression ofthese IAPs relates to epithelial cell apoptosis and/or prolif-eration. As shown in Fig. 9, which illustrates these param-eters from one patient (OC-4), ovarian epithelial tumorsexhibit considerable cellular heterogeneity, and PCNA-pos-itive cells were evident in cell nuclei throughout the tumorsection. In general, most of the cells were TUNEL negative,and the expression of IAPs was highly correlated to theproliferative state of the cells and inversely related to epi-thelial cell death. Xiap and Hiap-2 immunoreactivity (Fig. 9),specifically localized in cytoplasm or perinuclear region, washighest in proliferatively active (PCNA-positive) cells andwas low or absent in the apoptotic (TUNEL-positive) cellsoccasionally found in the tumor specimens.

Discussion

IAPs are a family of intracellular apoptosis suppressorproteins. Although the inhibitory effects of IAPs on apoptosisinduced by a variety of stimuli have been demonstrated(15–17), their role in tumorigenesis is unknown. Tumor pro-gression may be a consequence of excessive cell proliferationand/or suppressed physiological cell death, resulting in animbalance between the two processes (11, 34, 35). We reporthere that Hiap-2 and Xiap are expressed in human ovarianepithelial carcinomas and that their cellular contents are high

FIG. 6. Time-course studies of the influence of cisplatin on Xiap and Hiap-2 protein expression in hOSE cancer cells. Cisplatin-sensitive OV2008and its resistant variant (C13) were cultured in the absence and presence of cisplatin (30 mM) for up to 24 h. Equal amounts of solubilized cellproteins were analyzed by Western blot. Representative filters (A) and densitometric scans (B; mean 6 SEM) of Xiap and Hiap-2 protein contentsfrom three experiments are shown. **, P , 0.01 (compared with control at same time point).


in proliferatively active, but not apoptotic, cells. In addition,these cell survival proteins are present in human ovarianepithelial cancer cell lines, and down-regulation of Xiap byantisense cDNA expression results in apoptosis. These find-ings provide direct evidence for an apoptosis-suppressiverole of IAP in ovarian epithelial cancer. Although the mech-anism(s) by which Xiap exert its antiapoptotic action is stillnot fully understood, it has recently been demonstrated that

IAPs can inhibit caspase-3 and caspase-7, two cell deathproteases downstream of many proapoptotic signaling path-ways (36, 37). Interestingly, our preliminary studies haveshown that caspase-3 is present in hOSE cancer cells and isactivated in the sensitive, but not the resistant, cells by thepresence of cisplatin (Schneiderman, D., et al., unpublisheddata).

A decrease in ovarian epithelial cancer cell number wasobserved after Xiap down-regulation and could not be fullyaccounted for by increased apoptosis, suggesting that Xiapmay also important in the control cancer cell growth. The

FIG. 7. Influence of cisplatin on Xiap and Hiap-2 protein expressionand apoptosis in hOSE cancer cells. Cisplatin-sensitive A2780s cellsand its resistant variant (A2780cp) were cultured in the absence andpresence of cisplatin (30 mM) for 24 h, and equal amounts of solubilizedcell proteins were analyzed by Western blot. The extent of apoptosiswas determined based on nuclear morphology (Hoechst nuclear stainshowing fragmented nuclei). Representative Western blot filters (toppanels) and densitometric scans (middle panels; mean 6 SEM) of Xiapand Hiap-2 protein contents from three experiments are shown. Thelower panel shows the extent of apoptosis in the cell population. *, P ,0.05; **, P , 0.01; ***, P , 0.001 (compared with control).

FIG. 8. The effects of Xiap overexpression on the proapoptotic actionof cisplatin (30 mM) in cisplatin-sensitive OV2008 cells after 48-hinfection with adenoviral sense Xiap cDNA or LacZ (vector control).A, Nuclear morphology (magnification, 3400): a, LacZ; b, sense Xiap;c, LacZ and cisplatin; d, sense Xiap and cisplatin. B, Percentage oftotal cell population undergoing apoptosis (mean 6 SEM). C, A rep-resentative Western blot filter. D, Densitometric scans (mean 6 SEM)of Xiap protein contents. Data in B and D are from three experiments.*, P , 0.05; ***, P , 0.001 (compared with LacZ). 21, P , 0.01; 211,P , 0.001 (compared with LacZ and cisplatin group).


dual roles of regulators in controlling cell proliferation andapoptosis have been reported. For example, transforminggrowth factor-a and PGs have been shown to be both mi-togenic and antiapoptotic in ovarian granulosa cells (38, 39).Interestingly, down-regulation of survivin, another memberof the IAP family, can also induce apoptosis and suppress cellproliferation in HeLa cells (40). Whether and how Xiap isinvolved in the mitogenic signaling in human ovarian epi-thelial cancer cells remain to be elucidated.

The cytotoxicity of cisplatin is believed to be due to its rolein the formation of DNA adducts, including DNA-proteincross-links, DNA monoadducts, and interstrand DNA cross-links (41, 42), and its ability to trigger apoptosis. Apoptosisinduced by cisplatin can be inhibited by cycloheximide, sug-gesting that synthesis of new proteins may be required (43,44). However, the events that link DNA damage to the celldeath pathway are not known. In the current study the con-centration-dependent induction of apoptosis in OV2008 bycisplatin is accompanied by decreases in Xiap and Hiap-2protein content, suggesting that the modulation of IAP ex-pression and/or stability may be one of the mechanisms bywhich cisplatin induces apoptosis in the chemo-sensitive

cells. This decrease in IAP levels did not appear to be due toa general cytotoxic effect of the anticancer agent, becausestudies carried out in our laboratory under identical exper-imental conditions have shown that the cisplatin treatmentincreased Fas and Fas ligand expression (28). Moreover,DNA damage-inducible genes, such as gadd153, gadd45,p21, and c-jun, have also been shown to be up-regulated bycisplatin in this cell system (25), suggesting that the observeddecrease in IAP content was not a nonspecific response tocisplatin. In the present study we observed that whereasapproximately 80% of OV2008 cells underwent apoptosis inthe presence of cisplatin (30 mm), only 35–55% of the cellpopulation were apoptotic when Xiap expression was down-regulated by antisense expression. Although the reasons forthis apparent discrepancy remain to be determined, it ispossible that cisplatin may have multiple points of action inthe induction of cell death. For example, in addition to sup-pressing IAP expression, cisplatin has been shown to up-regulate the expression of a number of apoptosis inducers,including p53 (45), Fas, and Fas ligand (28). In the lattercontext, cisplatin (30 mm) significantly increased cell-associ-ated Fas (300% over the control value) and Fas ligand (50%

FIG. 9. Representative photomicrographs illustrating in situ detection of apoptosis (TUNEL) and immunolocalization of PCNA, Xiap, andHiap-2 in a hOSE tumor tissue from a patient (OC-4). a, Localization of apoptotic cells. b–d, Immunoreactivities for PCNA, Xiap, and Hiap-2,respectively. The regions of tumor shown in the circle and rectangle in each panel were TUNEL positive and TUNEL negative, respectively.Magnification, 3400.


over the control value) as well as soluble Fas ligand (300%over the control value) in cisplatin-sensitive cells (OV2008)during a 24-h culture period (28). Alternately, the relativelylow incidence of apoptosis could be due to insufficient Xiapdown-regulation, as the antisense was only able to suppressXiap by about 50%. In addition, in light of the fact that onlyone apoptosis suppressor (i.e. Xiap) was down-regulated inthe antisense infection experiments (which only partiallysuppressed Xiap protein levels), it is not surprising that cis-platin was more effective in inducing cell death comparedwith the adenoviral Xiap antisense expression.

The relative importance of various IAP members in con-trolling hOSE cancer cell apoptosis is unclear. In addition tothe slightly higher concentration of cisplatin required todown-regulate Hiap-2 than Xiap (30 vs. 20 mm) in OV2008cells, our present studies demonstrated that adenoviralHiap-2 antisense infection of OV2008 cells failed to signifi-cantly induce cell death. Although it is possible that only aminimal level of Hiap-2 is required to prevent the cells fromundergoing apoptosis, a more likely possibility could be thatHiap-2 is less important or is unrelated to the regulation ofovarian epithelial cancer cell apoptosis. The latter is consis-tent with the present observation that Hiap-2 protein inA2780cp cells content not only was not decreased by cisplatintreatment, but appeared slightly, although not significantly,elevated. Further studies including double IAP antisenseexpression are required to clearly define the relative roles ofthe different IAP family members as well as their mecha-nisms of action in ovarian carcinoma.

Chemoresistance is a major therapeutic problem, and un-derstanding the molecular mechanisms involved is a majorfocus for cancer research. The biology of chemoresistance hastraditionally been thought of in terms of altered drug deliv-ery, modified drug target, increased rate of DNA repair, ordecreased rate of drug-induced DNA or macromoleculedamage (45) as well as enhanced replicative bypass of plat-inum-DNA adducts in the resistant cells (46). Although re-cent evidence has suggested that drug resistance may in partbe attributed to decreased cellular susceptibility to apopto-genic insults (47, 48), the molecular basis of the latter phe-nomenon remains to be fully defined. We have tested thehypothesis that the inability of cisplatin to down-regulateIAP may be an important contributing factor for chemore-sistance in human ovarian cancer. We hereby report that 1)cisplatin failed to alter Xiap and Hiap-2 contents or induceapoptosis in the cisplatin-resistant cells; 2) overexpression ofXiap by adenoviral sense cDNA infection is effective in at-tenuating cisplatin-induced apoptosis in chemosensitivecells; and 3) Xiap antisense expression decreased C13 cellviability. These findings suggest that down-regulation ofXiap may increase the sensitivity of the chemoresistant cellsto cisplatin and may be a potential strategy for gene therapyfor cisplatin-resistant ovarian cancer. In this context, we haverecently observed that down-regulation of Xiap by adeno-viral antisense expression not only induced apoptosis inchemoresistant hOSE cancer cells (C13) in vitro, but alsosensitized the cells to the cytotoxic action of cisplatin (29).

In summary, our findings suggest that Xiap may be atarget for chemotherapeutic agents in the induction of apo-ptosis, and that the ability of cisplatin to down-regulate IAP

may be an important factor in chemosensitivity. These stud-ies provide important insight on the involvement of IAPs inthe regulation of apoptosis in ovarian cancer cells, the mech-anisms of action of chemotherapeutic agents, and the mo-lecular basis of chemoresistance.

Acknowledgments

The generous gifts of the hOSE cancer cell lines from Dr. R. Goel(OV2008 and C13) and Dr. C. E. Ng (A2780s and A2780cp), OttawaRegional Cancer Center (Ottawa, Canada), are most appreciated. Wealso thank Ms. Angela Tonary for her assistance with the collection ofovarian tumors and Ms. Keiko Hicks for the production of and forconducting the specificity analysis of the IAP antibodies used in thepresent studies.

References

1. Havrilesky LJ, Elbebdary A, Hurteau JA, Whitaker RS, Rodriguez GC, Ber-chuck A 1995 Chemotherapy-induced apoptosis in epithelial ovarian cancer.Obstet Gynecol 85:1007–1010

2. Anthoney DA, McIlwrath AJ, Gallagher WM, Edlin AR, Brown R 1996Microsatellite instability, apoptosis, and loss of p53 function in drug-resistanttumor cells. Cancer Res 56:1374–1387

3. Sherman SE, Lippard SJ 1987 Structural aspects of platinum anticancer druginteraction with DNA. Chem Rev 87:1153–1157

4. Ellis MC, Yuan JY, Horvitz HR 1991 Mechanisms and function of cell death.Annu Rev Cell Biol 7:663–698

5. Steller H 1995 Mechanisms and genes of cellular suicide. Science 267:1445–1449

6. Thompson CB 1995 Apoptosis in the pathogenesis and treatment of disease.Science 267:1456–1462

7. Diebold J, Baretton G, Felchner M, Meier W, Karin D, Schmidt M, Lohrs U1996 bcl-2 expression, p53 accumulation, and apoptosis in ovarian carcinomas.Am J Clin Pathol 105:341–349

8. Eliopoulos AG, Kerr DJ, Herrod J, Hodgkins L, Krajewski S, Reed JC,Lawrence SY 1995 The control of apoptosis and drug resistance in ovariancancer: influence of p53 and bcl-2. Oncogene 11:1217–1228

9. Herod JO, Eliopoulos AG, Warwick J, Niedobitek G, Young LS, Kerr DJ 1996The prognostic significance of Bcl-2 and p53 expression in ovarian carcinoma.Cancer Res 56:2178–2184

10. Vaux DL, Cory S, Adams JM 1988 Bcl-2 gene promotes haemopoietic cellsurvival and cooperates with c-myc to immortalize pre-B cells. Nature 335:440–442

11. Stewart BW 1994 Mechanisms of apoptosis: integration of genetic, biochem-ical, and cellular indicators J Natl Cancer Inst 86:1286–1296

12. Birnbaum MJ, Clem RJ, Miller LK 1994 An apoptosis inhibiting gene from anuclear polyhedrosis virus encoding a peptide with Cys/His sequence motifs.J Virol 68:2521–2528

13. Crook NE, Clem RJ, Miller LK 1993 An apoptosis-inhibiting baculovirus genewith a zing finger-like motif. J Virol 67:2168–2174

14. Roy N, Mahadevan MS, Mclean M, Shutler G, Yaraghi Z, Farahni R, BairdS, Besner-Johnston A, Lefebvre C, Kang X, Salih M, Aubry H, Tamai K, GuanX, Ioannou P, Crawford TO, Jong PJD, Korneluk RG, MacKenzie A 1995 Thegene for neuronal apoptosis inhibitory protein is partially deleted in individ-uals with spinal muscular atrophy. Cell 80:167–178

15. Liston P, Roy N, Tamai K, Lefebvre C, Baird S, Cherton-Horvat G, FarahaniR, McLean M, Ikeda J-E, MacKenzie A, Korneluk RG 1996 Suppression ofapoptosis in mammalian cells by Naip and a related family. Nature 379:349–353

16. Uren AG, Pakusch M, Hawkins CJ, Puls KL, Vaux DL 1996 Cloning andexpression of apoptosis inhibitory protein homologs that function to inhibitapoptosis and/or bind tumor necrosis factor receptor-associated factors. ProcNatl Acd Sci USA 93:4974–4978

17. Duckett CS, Nava VE, Gedrich RW, Clem RJ, Van Dongen JL, Gilfilan MC,Sheils H, Hardwick JM, Thompson CB 1996 A conserved family of cellulargenes related to the baculovirus iap gene and encoding apoptosis inhibitors.EMBO J 15:2685–2694

18. Rothe M, Pan M-G, Henzel WJ, Ayres TM, Goeddel DV 1995 The TNFR2-TRAF signalling complex contains two novel proteins related to baculovirusinhibitor of apoptosis proteins. Cell 83:1243–1252

19. Ambrosini G, Adida C, Altieri DC 1997 A novel anti-apoptosis gene, survivin,expressed in cancer and lymphoma. Nat Med 3:917–921

20. Hamilton TC, Winker MA, Louie KG, Batist G, Behrens BC, Tsuruo T,Grotzinger KR, McKoy WM, Young RC, Ozols RF 1985 Augmentation ofadriamycin, melphalan, and cisplatin cytotoxicity in drug-resistant and sen-sitive human ovarian carcinoma cell lines by buthomine sulfoximine mediatedglutathione depletion. Biochem Pharmacol 34:2583–2586


21. Hamilton TC, Lai G-M, Rothenberg ML, Fojo AT, Young RC, Ozols RF 1989Mechanisms of resistance to cisplatin and alkylating agents. Cancer Treat Res48:151–169

22. Louie KG, Behrens BC, Kinsella TJ, Hamilton TC, Grotzinger KR, McKoyWM, Winker MA, Ozols RF 1985 Radiation survival parameters of antineo-plastic drug-sensitive and -resistant human ovarian cancer cell lines and theirmodification by buthionine sulfoximine. Cancer Res 45:2110–115

23. Vaisman A Varchenko M, Said I, Chaney SG 1997 Cell cycle changes asso-ciated with formation of P-DNA adducts in human ovarian carcinoma cellswith different cisplatin sensitivity. Cytometry 27:54–64

24. Liu Y, Bhalla K, Hill C, Priest DG 1994 Evidence for involvement of tyrosinephosphorylation in taxol-induced apoptosis in a human ovarian tumor cellline. Biochem Pharmacol 48:1265–1272

25. Delmastro DA, Li Jibin, Vaisman A, Solle M, Chaney SG 1997 DNA damageinducible-gene expression following platinum treatment in human ovariancarcinoma cell lines. Cancer Chemother Pharmacol 39:245–253

26. Behrens RC, Hamilton TC, Masuda H, Grotzinger KR, Whang-Peng J, LouieKG, Knutsen T, McKoy WM, Young RC, Ozols RF 1987 Characterization ofa cis-diamminedichloroplatinum(II)-resistant human ovarian cancer cell lineand its use in evaluation of platinum analogues. Cancer Res 47:414–418

27. Andrews PA, Albright KD 1992 Mitrochondrial defects in cis-diamminedi-chloroplatinum(II)-resistant human ovarian carcinoma cells. Cancer Res 52:1895–1901

28. Schneiderman D, Kim J-M, Senterman M, Tsang BK 1999 Sustained sup-pression of Fas ligand expression in cisplatin-resistant human ovarian surfaceepithelial cancer cells. Apoptosis 4:271–282

29. Sasaki H, Sheng Y-L, Kotsuji H, Tsang BK 2000 Down-regulation of XIAPinduces apoptosis in chemoresistant human ovarian cancer cells. Cancer Res60:5659–5666

30. Simmons KR, Caffrey JO, Phillips JL, Abel JH, Niswender GD 1976 A simplemethod for preparing suspension of luteal cells. Proc Soc Exp Biol Med148:123–126

31. Miyake S, Makimura M, Kanegae Y, Harada S, Sato Y, Takamori K, TokudaC, Saito I 1996 Efficient generation of recombinant adenoviruses using ade-novirus DNA-terminal protein complex and a cosmid bearing the full-lengthvirus genome. Proc Natl Acad Sci USA 93:1320–1324

32. Li J, Kim J-M, Liston P, Li M, Mackenzie AE, Korneluk RG, Tsang BK 1998Expression of apoptosis inhibitory proteins (IAPs) in rat granulosa cells duringfollicular development and atresia. Endocrinology 139:1321–1328

33. Gavrieli Y, Sherman Y, Bensasson SA 1992 Identification of programmed celldeath in situ via specific labeling of nuclear DNA fragmentation. I. Cell Biol119:493–501

34. Kerr JF, Winterford CM, Harmon BV 1994 Apoptosis: its significance in cancerand cancer therapy. Cancer 73:2013–2026

35. Reed JC 1994 Bcl-2 and the regulation of programmed cell death. J Cell Biol124:1–6

36. Daveraux QL, Takahashi R, Saivesen QS, Reed JC 1997 X-linked IAP is adirect inhibitor of cell-death proteases. Nature 388:300–303

37. Roy N, Deveraux QL, Takahashi R, Salvesen GS, Reed JC 1997 The c-IAP-1and c-IAP-2 proteins are direct inhibitors of specific caspases. EMBO J 16:6914–6925

38. Li J, Tsang BK 1995 Prostaglandins mediate the stimulation of deoxyribonu-cleic acid synthesis by transforming growth factor a in hen granulosa cellsduring ovarian follicular development. Biol Reprod 52:1050–1058

39. Manchanda R, Kim J-M, Tsang BK, Transforming growth factor a inhibitsapoptosis in hen granulosa cells. Proceedings of the 11th Ovarian Workshop:Ovarian Cell Growth Apoptosis and Cancer, London, Canada, 1996, p 17

40. Ambrosini G, Adida C, Sirugo G, Altieri DC 1998 Induction of apoptosis andinhibition of proliferation by survivin. J Biol Chem 273:11177–11182

41. Fichtinger-Schepmen AMJ, van der Veer JL, den Hartog JHJ, Lohman PHM,Reedijk J 1985 Adducts of the antitumor drug cis-diamminedichloroplatinum(II) with DNA: formation, identification and quantitation. Biochemistry 24:707–713

42. Eastman A 1986 Reevaluation of interaction of cis-dechloro(ethylenediamine)platinum (II) with DNA. Biochemistry 25:3912–3915

43. Sorenson CM, Barry MA, Eastman AJ 1990 Analysis of events assosiated withcell cycle arrest at G2 phase and cell and cell death induced by cisplatin. J NatlCancer Inst 82:749–755

44. Barry MA, Behnke CA, Eastman A 1990 Activation of programmed cell deathby cisplatin, other anti-cancer drugs, toxins and hyperthermia. Biochem Phar-macol 40:2353–2362

45. Reed JC, Toshiyuki M, Takayama S, Wang H-G, Sato T, Krajewski S, Aime-Sempe C, Bodrug S, Kitada S, Hanada M 1996 Bcl-2 family proteins: regulatorsof cell death involved in the pathogenesis of cancer and resistance to therapy.J Cell Biochem 60:23–32

46. Mamenta EL, Poma EE, Kaufmann WK, Delmastro DA, Grady HL, ChaneySG 1994 Enhanced replicative bypass of platinum-DNA adducts in cisplatin-resistant human ovarian carcinoma cell lines. Cancer Res 54:3500–3505

47. Perego P, Giarola M, Righeti SC, Supino R, Caserini C, Delia D, Pierotti MA,Miyashita T, Reed J, Zunino F 1996 Association between cisplatin resistanceand mutation of p53 gene and reduced bax expression in ovarian carcinomacell system. Cancer Res 56:556–562

48. Dive C, Hickman JA 1991 Drug-target interactions: only the first step in thecommitment to a programmed cell death? Br J Cancer 64:192–196


Human Ovarian Cancer and Cisplatin Resistance: Possible Role of Inhibitor of Apoptosis Proteins 1

Documents

Transcript of Human Ovarian Cancer and Cisplatin Resistance: Possible Role of Inhibitor of Apoptosis Proteins 1