Reproductive Hormone-Induced, STAT3-Mediated Interleukin 6 Action in Normal and Malignant Human...

Reproductive Hormone-Induced, STAT3-MediatedInterleukin 6 Action in Normal and Malignant HumanOvarian Surface Epithelial Cells

Viqar Syed, Gregory Ulinski, Samuel C. Mok, Shuk-Mei Ho

Background: Reproductive hormones are associated withrisk for epithelial ovarian cancer. To determine the effect ofsuch hormones on the activation of interleukin 6 (IL-6)/STAT3(signal transducer and activator of transcription-3) signal-ing, which may be involved in ovarian cancer, we investi-gated the status of STAT3, IL-6, and its receptor in immor-talized human ovarian surface epithelial (HOSE) andovarian cancer (OVCA) cell lines. Methods: Two immortal-ized HOSE cell lines and two OVCA cell lines were culturedwith gonadotropins, sex steroid hormones, and/or IL-6,alone or with specific inhibitors or IL-6-neutralizing anti-bodies. Expression of IL-6, the IL-6 receptor � chain (IL-6R�),and phosphorylated and unphosphorylated STAT3 messen-ger RNAs (mRNAs) and proteins in all cells was determined.Cell proliferation and soft-agar colony formation were as-sessed. STAT3 activity was investigated in OVCA cells trans-fected with a dominant negative STAT3 (Dn-STAT3), wild-type STAT3, or an empty control vector. All statistical testswere two-sided. Results: Levels of IL-6 mRNA and proteinincreased in all cells treated with follicle-stimulating hor-mone (FSH), luteinizing hormone (LH), 17�-estradiol, or es-trone but increased only in OVCA cells treated with testos-terone and 5�-dihydrotestosterone. For all lines, IL-6antibodies partially inhibited hormone-stimulated cell pro-liferation but completely inhibited IL-6-enhanced cell pro-liferation. IL-6 induced STAT3 phosphorylation and activa-tion in HOSE cells; STAT3 was constitutively activated inOVCA cells. Higher levels of IL-6R� and STAT3 transcrip-tion factors were observed in OVCA cells than in HOSEcells. After transfection, Dn-STAT3 suppressed endogenousSTAT3 and inhibited all forms of IL-6-stimulated OVCAcell proliferation (OVCA 429 cells, P<.001; OVCA 432 cells,P<.006), whereas wild-type STAT3 enhanced HOSE cell pro-liferation (wild-type STAT3 at 0.5 µg/mL in HOSE 306 cells,P<.002; STAT3 at 1.0 µg/mL in HOSE 306 or both concen-trations of wild-type STAT3 in HOSE 642 cells, P<.001).Conclusions: The IL-6/STAT3 signaling pathway may medi-ate FSH-, LH-, and estrogen-stimulated HOSE cell prolif-eration. Increased IL-6R� expression and constitutiveSTAT3 activation may be associated with ovarian cancer.[J Natl Cancer Inst 2002;94:617–29]

Ovarian cancer is responsible for more than 14 500 deathsannually (1,2). More than 90% of ovarian cancers arise from thesingle layer of human ovarian surface epithelial (HOSE) cellscovering the ovary. Despite the magnitude of this clinical prob-lem, little is known about the mechanism of neoplastic transfor-mation.

Higher levels of gonadotropins after menopause and long-term use of estrogen/hormone-replacement therapy are associ-ated with an increased incidence of ovarian cancer in perimeno-

pausal and postmenopausal women (3–11). High levels ofandrogens, such as androstenedione and testosterone, are asso-ciated with an increased risk for ovarian cancer (1,6). Alteredlevels of receptors for follicle-stimulating hormone (FSH), lu-teinizing hormone (LH), estrogen, androgen, and progesteroneare detected in OVCA cells compared with HOSE cells (12–17).Recently, we have found that gonadotropins, androgens, andestrogens stimulate the proliferation of normal HOSE cells andOVCA cells (17). According to a current theory of hormonalcarcinogenesis (18), hormone-stimulated cell proliferation fa-vors the accumulation of genetic errors, which can result inmalignant transformation. Thus, to identify potential targets forovarian cancer therapy, it is necessary to determine whethergonadotropins and sex steroid hormones transform HOSE cellsdirectly or indirectly, by activating intermediary growth factor/cytokine signaling pathways.

Interleukin 6 (IL-6) is a pleiotropic cytokine that regulatesmany cellular functions, including cell proliferation, cell differ-entiation, immune defense mechanisms, and hematopoiesis (19–21). IL-6 also is involved in the malignant transformation andprogression of various cancers, including renal-cell, prostate,lung, and breast carcinomas and leukemia, and the progressionof Kaposi’s sarcoma (22–28). IL-6 acts through a hexametriccytokine receptor complex composed of an IL-6-specific recep-tor subunit (� chain, IL-6R�) and a signal transducer, gp130(� chain), that is also used by other cytokine receptors (29,30).The binding of IL-6 to gp130 activates the Janus kinase/STAT3(signal transducer and activator of transcription-3) signal trans-duction cascade, in which STAT3 transmits signals from the cellmembrane to the nucleus (31,32). STAT3 is essential for gp130-mediated cell survival, the G1/S-phase cell cycle transition, cellmovement, and cell differentiation (33). Constitutive activationof STAT3 results in the ligand-independent transcription ofgenes in antiapoptotic/cell cycle progression pathways, a mecha-nism that is likely involved in oncogenic transformation (34).

Epithelial ovarian cancers produce IL-6 (35,36), and the con-centration of IL-6 in the serum and peritoneal fluid is elevated inovarian cancer patients (37–39). OVCA cells expressing IL-6are resistant to cisplatin cytotoxicity (40), and a high level ofserum IL-6 is associated with poor prognosis and survival for

Affiliations of authors: V. Syed (Department of Surgery), S.-M. Ho (Depart-ment of Surgery and Cell Biology), University of Massachusetts Medical School,Worcester; G. Ulinski, Worcester Polytechnic Institute, Worcester; S. C. Mok,Laboratory of Gynecologic Oncology, Department of Obstetrics, Gynecology,and Reproductive Biology, Brigham and Women’s Hospital, Harvard MedicalSchool, Boston.

Correspondence to: Shuk-Mei Ho, Ph.D., Department of Surgery, Universityof Massachusetts Medical School, 55 Lake Ave. North, Worcester, MA 01655(e-mail: [email protected]).

See “Notes” following “References.”

© Oxford University Press

Journal of the National Cancer Institute, Vol. 94, No. 8, April 17, 2002 ARTICLES 617

by guest on August 3, 2015

http://jnci.oxfordjournals.org/D

ownloaded from

http://jnci.oxfordjournals.org/

patients with ovarian cancer (41,42). OVCA cell lines NOM1and SKOV cultured with IL-6 have increased chemotactic and/orchemokinetic activity and increased overall invasiveness (43).Thus, IL-6 appears to be an important regulator of OVCA cellgrowth and/or survival. However, to the best of our knowledge,no information is available on 1) the action of IL-6 on normalHOSE cell function, 2) the upstream regulators of IL-6 expres-sion in HOSE cells and OVCA cells, and 3) the characteristics ofthe IL-6/STAT3 signaling cascade of OVCA cells.

It is well documented that, apart from gonadotropins andovarian steroids, the ovary is regulated by many cytokines, in-cluding IL-6, that play important roles in folliculogenesis andoocyte maturation. The effects of exogenous gonadotropins onthe release of IL-6 into the human follicular fluid have beenstudied (44). Gonadotropins have been shown to regulate ovar-ian secretion of IL-6 because IL-6 levels in follicular fluid wereelevated after gonadotropin treatment (44,45). Gonadotropinshave been shown to stimulate IL-6 production in Sertoli cells(46). The effect of androgen on the expression of IL-6 varies. Instromal cells derived from murine bone marrow (47), brain glialcells and astrocytes (48), and a human osteoblastic cell line (49),androgens inhibit IL-6 production. However, in Sertoli cells, lowdoses of testosterone stimulate IL-6 expression but higher dosesinhibit IL-6 expression (46). Estrogens have been reported toenhance IL-6 expression in blood mononuclear cells in vitro (50)but to inhibit IL-6 synthesis in a human osteoblastic cell line,vascular smooth muscle cells, and bone marrow-derived stromalcells (51–53).

Regulation of IL-6 in HOSE cells has not been studied. Wethus investigated whether gonadotropins, estrogens, and andro-gens affect IL-6 expression in HOSE and OVCA cells andwhether IL-6 participates in the proliferation of HOSE andOVCA cells by activating STAT3. We also examined the levelof IL-6R� expression and the state of STAT3 activation inHOSE and OVCA cells.

MATERIALS AND METHODS

Cell Cultures and Cell Lines

The origin and culture conditions of HOSE and OVCA celllines have been previously described (12,17,54). In short, HOSEcell lines HOSE 301, HOSE 306, HOSE 642, and HOSE 12–12were derived from normal ovaries obtained from women withnoncancer gynecologic indications—specifically, a 47-year-oldwoman with endometrial adenocarcinoma, a 53-year-old womanwith breast cancer, a 47-year-old normal woman, and a 39-year-old woman with ovarian stromal hyperplasia, respectively. Pri-mary cultures were established from surface scrapings of thesenormal ovaries and immortalized with human papillomavirus E6and E7 genes (55). All primary cultures had an epithelial cellphenotype, i.e., cobblestoned morphology, epithelial cytokeratinstaining, responsiveness to transforming growth factor-�, no de-tectable CA125 production, and lack of in vivo tumorigenicity(55). OVCA cell lines OVCA 420, OVCA 429, OVCA 432, andOVCA 433 were established cell lines derived from patientswith late-stage serous ovarian adenocarcinomas, as described byBast et al. (56). These cell lines were cultured at 37 °C in ahumidified atmosphere of 5% CO2/95% air in a 1 : 1 mixture ofmedium 199/MCDB 105 (Sigma Chemical Co., St. Louis, MO)supplemented with 10% fetal bovine serum (FBS; Sigma), peni-cillin (100 U/mL; Sigma), and streptomycin (100 �g/mL;Sigma).

Hormone Treatment of HOSE and OVCA Cell Lines

Cell lines cultured in medium 199/MCDB 105 were har-vested when 80% confluent and washed twice in phosphate-buffered saline (PBS), and then 2 × 105 cells were cultured perT-25 flask (culture area � 25 cm2; Falcon; Becton DickinsonLabware, Bedford, MA) and allowed to attach for 24 hours.Forty-eight hours later, the medium was replaced with medium199/MCDB 105 containing FBS treated with activated charcoal(Sigma)/dextran-70 (hereafter, charcoal-stripped FBS; Pharma-cia, Peapack, NJ), and 1 �M FSH (Calbiochem, San Diego, CA),1 �M LH (Calbiochem), 1 �M 17�-estradiol (Sigma), 1 �Mestrone (Sigma), 1 �M 5�-dehydrotestosterone (Sigma), or 1 �Mtestosterone (Sigma). Steroid solutions were in absolute ethanol,and gonadotropin solutions were in aqueous saline. The finalconcentration of ethanol in the medium was 0.1%. Untreatedcontrol cultures were exposed to equal concentrations of ethanolor aqueous saline vehicle alone. Cells were treated daily for 5days with hormones or cultured without hormones. Because 5�-dehydrotestosterone was rapidly metabolized, 5�-dehydrotestos-terone was added every 12 hours.

Hormone Treatment of Normal Immortalized andMalignant HOSE Cells in the Absence and Presence ofHormone Receptor Antagonists

HOSE 306, HOSE 642, OVCA 429, and OVCA 432 cells(2 × 105 cells per T-25 flask) were allowed to attach for 24hours, and then 10 nM FSH, 10 nM LH, 10 nM testosterone, or10 nM 17�-estradiol was added in the presence or absence of thecorresponding receptor or signaling inhibitor. A dose of 10 nMwas chosen because it fell at the mid-point of dose–responsecurves from the cell proliferation assay. The protein kinase A-selective inhibitor H-89 (N-[2-((p-bromocinnaml)amino)ethyl]-5-isoquinolinesulfonamide, HCL; Calbiochem) with an inhibi-tion constant of 0.048 �M was added to a final concentration of0.1 mM 30 minutes before treatment with 10 nM FSH or 10 nMLH. H-89 at 0.1 mM effectively blocks the action of FSH andLH (17). Receptor-specific inhibitors were used for sex steroidhormones as follows: 0.1 mM 4-hydroxyflutamide (Schering-Plough, Kenilworth, NJ), an androgen receptor inhibitor (17),and 0.1 mM ICI 182,780 (a gift from Zeneca Pharmaceuticals,Macclesfield, U.K.), a pure estrogen receptor inhibitor (17). Cellcultures were treated daily with hormones or hormone inhibitorsfor a period of 5 days, cells were collected, and total RNA wasextracted. Each experiment was carried out twice.

Treatment of HOSE and OVCA Cell Lines With Hormoneand IL-6 Antibody

HOSE 306, HOSE 642, OVCA 429, and OVCA 432 cellswere cultured in wells of 96-well plates at 1000 cells per well inmedium containing charcoal-stripped FBS. Forty-eight hourslater, the medium was replaced with the same medium contain-ing human FSH, human LH, 5�-dehydrotestosterone, testoster-one, 17�-estradiol, or estrone (each at 10 nM), as describedabove (17). Control cells were exposed to PBS or ethanol ve-hicle alone. To neutralize hormone-induced IL-6, cells were cul-tured for 5 days with hormones and a rabbit anti-human IL-6polyclonal antibody (Santa Cruz Biotechnology, Inc., SantaCruz, CA) at a final concentration of 50 �g/mL. The specificityof this anti-IL-6 antibody was confirmed by its complete inhi-bition of the mitogenic effect of recombinant IL-6 (100 ng/mL)

618 ARTICLES Journal of the National Cancer Institute, Vol. 94, No. 8, April 17, 2002



ownloaded from


in the TF-1 cell line. To ensure stable availability, hormones andanti-IL-6 antibody were added to cells daily. Five days aftertreatment, cell proliferation was measured with the 3-(4,5-dimeth-ylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) cellproliferation kit (Roche Diagnostics, Indianapolis, IN).

IL-6 Treatment of HOSE and OVCA Lines

Cells lines were cultured, as described above, in 96-wellplates at 1000 cells per well in medium 199/MCDB 105 con-taining 10% charcoal-stripped FBS. Forty-eight hours later, hu-man recombinant IL-6 (Upstate Biotechnology, Lake Placid,NY) at 1–32 ng/mL was added to wells, and the same concen-tration of IL-6 was added daily for 5 days. On day 6, the numberof cells in each well was measured by use of an MTT cellproliferation assay (Roche Diagnostics). For one set of experi-ments, rabbit anti-human IL-6 polyclonal antibody (Santa CruzBiotechnology) at 50 �g/mL was added to the cultures to neu-tralize IL-6, and 5 days later cell proliferation was assessed withthe MTT assay. For other experiments, 2 × 105 HOSE cells or2 × 105 OVCA cells per T-25 flask (Falcon) were allowed toattach for 24 hours. Forty-eight hours later, cells were treatedwith IL-6 at 32 ng/mL for 6 hours or 5 days, and then cell lysateswere prepared.

Cell Growth Assay

The number of cells in each culture was measured with anMTT cell proliferation kit (Roche Diagnostics). The MTT la-beling reagent (10 �L, final concentration � 0.5 mg/mL) wasadded to the medium in each culture well and incubated for 4hours in a humidified atmosphere, 100 �L of a solubilizationsolution (10% sodium dodecyl sulfate in 0.01 M HCl) was addedto each well, and plates were incubated overnight at 37 °C. Thenumber of cells in each culture is linearly related to the conver-sion of the tetrazolium compound to the colored formazan prod-uct with an absorbance at 570 nm. Assays were performed intriplicate. The relative cell growth for each treatment regimen isexpressed as the fold increase over that of untreated controlcultures. The number of cells in the untreated control cultureswas arbitrarily assigned a value of one, and values from treat-ment groups were normalized to values from the control cul-tures.

RNA Isolation and Quantification of IL-6, IL-6R�, andgp130 Messenger RNA (mRNA) by a SemiquantitativeReverse Transcription–Polymerase Chain Reaction(RT–PCR)

Total RNA was isolated from treated and untreated HOSEand OVCA cell lines with Tri reagent (Sigma) according to themanufacturer’s instructions and quantified by UV absorbance at260–280 nm. The integrity of RNA was confirmed as previouslydescribed (54). Total RNA (1–3 �g) from each sample wasreverse transcribed by use of a GeneAmp RNA PCR kit (Perkin-Elmer/Cetus, Norwalk, CT). The resulting complementaryDNAs (cDNAs) were stored at –20 °C until use.

To investigate the relative expression of IL-6, IL-6R�, andgp130 mRNA in each culture, semiquantitative RT–PCRs wereperformed. Sequences of the oligonucleotide primers specific forhuman IL-6, IL-6R�, and gp130 were from published sequences(41). The forward primer sequence for IL-6 was 5�-CTTCGGT-CCAGTTGCCTTCT-3� and the reverse primer was 5�-AGGA-ACTCCTTAAAGCTGCG-3�. For human IL-6R�, the sense

primer was 5�-GAGGGAGACAGCTCTTTCTAC-3� and theantisense primer was 5�-CCGTTCAGCCCGATATCTGAG-3�.For gp130, the sense primer was 5�-ACCTATGAAGATAGAC-CATCTAAA-3� and the antisense primer was 5�-GGTTCTATA-AAATATAGTATAATT-3�. For amplification of human 18Sribosomal RNA (rRNA), the sense primer was 5�-TGAGGC-CATGATTAAGAGGG-3� and the antisense primer was 5�-CG-CTGAGCCAGTCAGTGTAG-3�. These primers are expectedto yield products of 609 base pairs (bp) for IL-6, 768 bp forIL-6R�, 720 bp for gp130, and 623 bp for 18S rRNA. Aliquotsof 1–3 �L of first-strand cDNAs were amplified by PCR. Hot-start PCR with AmpliTaq Gold DNA polymerase (PerkinElmer/Cetus) was used to minimize the amplification of nonspecificproducts. The enzyme was activated by preheating to 95 °C for6 minutes. Initially, to determine conditions for logarithmic-phase PCR amplification of IL-6, IL-6R�, gp130, and 18SrRNA, various amounts of total RNA were reverse transcribed,and aliquots were amplified for various numbers of PCR cycles.A linear relationship was observed between the amount of RNAand PCR products when 3 �g of total RNA was used in thereverse transcription reaction and 32, 24, 28, and 18 PCR am-plification cycles, respectively, were performed for IL-6, IL-6R�, gp130, and 18S rRNA. The PCR program was 1 minute at94 °C, 1 minute at 60 °C (annealing temperature), and 1 minuteat 72 °C. mRNA-specific modifications included annealing tem-peratures of 58 °C for IL-6 cDNA, 58 °C for IL-6R� cDNA, and60 °C for gp130 cDNA. PCR products were fractionated on a 2%agarose gel and quantified after ethidium bromide staining andUV transillumination as follows. Images were captured with adigital camera (Kodak EDAS 290; Eastman Kodak, Rochester,NY) and converted to digitized signals. Intensities of each bandwere quantified by the Kodak 10 Image analysis software ver-sion 3.5. (Eastman Kodak). Signal intensities of target products(IL-6, IL-6R�, and gp130) were normalized to that of the 18SrRNA product to obtain arbitrary units of relative message abun-dance. Reproducibility was evaluated by synthesis of three in-dependent cDNAs and PCR runs from each RNA preparation.Means of the replicated measurements were calculated.

Immunoblot Analysis

To obtain cellular extracts, cells were lysed in lysis buffer(50 mM Tris [pH 8.0], 250 mM NaCl, 0.5% Nonidet P-40), andproteins were quantified with a protein determination kit (Pierce,Rockford, IL). Proteins (25 �g) from HOSE 306, HOSE 642,OVCA 429, and OVCA 432 cells were separated by sodiumdodecyl sulfate–polyacrylamide gel electrophoresis in 10% gelsand transferred to nitrocellulose membranes. Blots were blockedwith 5% bovine serum albumin or 5% nonfat dry milk in Tris-buffered saline/0.1% Tween 20 for 1 hour. Blots were probedwith anti-STAT3 antibody (1.5 �g/mL; Upstate Biotechnology),anti-phospho-STAT3 antibody (1 : 1000 dilution; Cell Signaling,Beverly, MA), anti-IL-6 antibody (1 : 600 dilution, Santa CruzBiotechnology), anti-IL-6R� antibody (1 : 500 dilution; SantaCruz Biotechnology), or glyceraldehyde-3-phosphate dehydro-genase (GAPDH) antibody (1 : 500 dilution; Biogenesis, Poole,U.K.) overnight at 4 °C and then incubated with a secondaryantibody (1 : 6000 dilution) in 3% bovine serum albumin or3% nonfat dry milk in Tris-buffered saline/0.1% Tween 20 for1 hour at room temperature. After washing, bound antibodieswere detected by modified chemiluminescence with enhancedchemiluminescence (ECL) reagents (Amersham Pharmacia Bio-




ownloaded from


tech, Little Chalfont, Buckinghamshire, U.K.) and visualized byradioautography. The signal intensity was quantified by Kodak10 Image analysis software (version 3.5; Eastman Kodak) andnormalized to that of GAPDH to obtain arbitrary units of relativeabundance.

Electrophoretic Mobility Shift Assays (EMSAs)

The sequence of the double-stranded Stat3 consensus oligo-nucleotide used as the probe in gel-shift experiments correspondedto the highest affinity binding STAT3 site (5�-GATCCTTCTG-GGAATTCCTAGATC-3� and 3�-CTAGGAAGACCCTTAAG-GATCTAG-5�). This oligonucleotide is very specific for STAT3and has been extensively used to identify STAT3 transcriptionfactors (27). The Stat3 consensus oligonucleotide and the mutantSTAT3 oligonucleotide, which contained an AAT-to-CCG sub-stitution in the STAT3 binding motif, were purchased fromSanta Cruz Biotechnology. Probes were labeled with[�-32P]ATP (Amersham Pharmacia Biotech) and T4 polynucleo-tide kinase (Promega, Madison, WI).

Cells were homogenized in buffer containing 10 mM HEPES(pH 7.9), 1.5 mM MgCl2, 10 mM KCl, 0.5 mM dithiothreitol, 0.5mM phenylmethylsulfonyl fluoride, and 0.05% Nonidet P-40.Protease inhibitors (aprotinin at 1 �g/mL, leupeptin at 1 �g/mL,pepstatin at 1 �g/mL) were added to minimize protein degrada-tion. After incubation on ice for 15 minutes, the homogenate wascentrifuged at 10 000 rpm in a microcentrifuge at 4 °C for 10minutes. The pellet was resuspended in buffer containing 20 mMHEPES (pH 7.9), 0.42 M NaCl, 1.5 mM MgCl2, 0.2 mM EDTA,0.5 mM dithiothreitol, and 6.25% glycerol. After incubation onice for 40 minutes, the nuclear suspension was centrifuged at15 000 rpm in a microcentrifuge at 4 °C for 10 minutes. Thesupernatant was collected and protein concentrations determinedby the Bio-Rad protein assay. Nuclear extracts were incubatedwith 1 �g of poly(dI-dC) (Roche Diagnostics Corp., Indianapo-lis, IN) and labeled oligonucleotides for 25 minutes, and thenprotein–DNA complexes were resolved by electrophoresis in a6% nondenaturing polyacrylamide gel. Gels were dried and ra-dioautographed at –70 °C.

To investigate the sequence specificity of the protein–DNAinteraction, a 100-fold molar excess of unlabeled STAT3 oligo-nucleotides (specific competitor) or mutant STAT3 oligonucle-otides (nonspecific competitor) was added to the binding reac-tion, and displacement of the radioactive probe was determinedfor both oligonucleotides.

Transient Transfection

To monitor transient transfection, HOSE 306, HOSE 642,OVCA 429, and OVCA 432 cells (1 × 10–5 cells) were trans-fected with the pSV-�-galactosidase control vector (Promega,Madison, WI) at 1 �g/mL with LipofectAMINE PLUS (LifeTechnologies, Inc., Rockville, MD). Cells transfected with thepSV-�-galactosidase control vector express �-galactosidase andthus can be identified and used to calculate transfection effi-ciency. Briefly, after transfection, cells were fixed in glutaral-dehyde for 15 minutes, extensively washed in 1× PBS, andincubated with 0.2% X-Gal (5-bromo-4-chloro-3-indolyl �-D-galactopyranoside) in PBS containing 2 mM MgCl2, 5 mMK4Fe(CN)6•3H2O, and 5 mM K3Fe(CN)6. After fixation andincubation with X-Gal, transfected cells contained a blue reac-

tion product. Transfection efficiencies of HOSE 306, HOSE642, OVCA 429, and OVCA 432 cells transfected with thepSV-�-galactosidase vector, estimated by X-Gal staining, were71%, 68%, 75%, and 73%, respectively. After establishing thesetransfection efficiencies, HOSE 306, HOSE 642, OVCA 429,and OVCA 432 cells were cultured (each at 1 × 105 cells perwell) in six-well plates for 24 hours. Cultures 60%–80% con-fluent were transfected with the vector containing dominantnegative STAT3 (Dn-STAT3; 1 �g/mL) or the empty controlvector (1 �g/mL) by the use of LipofectAMINE PLUS (LifeTechnologies, Inc.) in serum-free medium for 3 hours at 37 °C.Cultures were incubated with complete medium for 24 hours andthen stimulated with one of two doses of IL-6 (16 or 32 ng/mL)for 3 days. The Dn-STAT3-containing vector and empty vectorwere gifts from Dr. Koichi Nakajima (Osaka University MedicalSchool, Osaka, Japan). The dominant negative construct carriesa tyrosine-to-phenylalanine substitution at position 705 thatcauses a reduction of tyrosine phosphorylation of wild-typeSTAT3 and inhibits the action of endogenous STAT3 (23,57).

In separate experiments, HOSE 306 and HOSE 642 cellswere transfected with a vector containing wild-type STAT3(0.25–1 �g/mL; a gift from Dr. Koichi Nakajima)—three dishesfor each dose. After 72 hours, cells were harvested by scraping,washed with PBS, pelleted, and counted with a hemocytometer.

Soft Agar Assay

OVCA 429 and OVCA 432 cells were transiently transfectedwith the Dn-STAT3 vector or the empty vector as describedabove, and 24 hours later, transfected cells were cultured at 500cells per 100-mm plate (four plates per sample) in 0.3% agarabove an underlayer of 0.6% Noble agar, both containing com-plete medium (58). Plates were incubated for up to 4 weeks, andthen the number of colonies was scored to establish the soft-agarcolony-formation frequency and expressed as the number ofcolonies per plate cultured. Data are expressed as the mean ±95% confidence intervals from four separate plates.

Data Analysis

Data are expressed as the mean of two to four experiments,each in triplicate samples for individual treatments or dosageregimens. Statistical analysis was carried out with a one-wayanalysis of variance, followed by Tukey’s post hoc test. Valuesare presented as the mean ± 95% confidence intervals. All sta-tistical tests were two-sided and were considered to be statisti-cally significant at P<.05.

RESULTS

Effect of Gonadotropins and Steroid Hormones on theExpression of IL-6 mRNA and Protein in HOSE andOVCA Cells

HOSE (HOSE 301, HOSE 306, HOSE 642, and HOSE 12–12) and OVCA (OVCA 420, OVCA 429, OVCA 432, andOVCA 433) cell lines were treated with medium alone or with1 �M FSH, 1 �M LH, 1 �M 17�-estradiol, 1 �M estrone, 1 �Mtestosterone, or 1 �M 5�-dehydrotestosterone for 5 days. Rela-tive levels of IL-6 mRNA were then measured with semiquan-titative RT–PCR. Data from two cell lines from each group,HOSE 306 and HOSE 642 cells and OVCA 429 and OVCA 432




ownloaded from


cells, are shown in Fig. 1, A. Levels of IL-6 mRNA in all HOSEand OVCA cell lines exposed to FSH, LH, 17�-estradiol, orestrone were 1.5-fold and 2.5-fold higher than in their respectiveuntreated control cultures. Treatment of HOSE and OVCA celllines with testosterone or 5�-dehydrotestosterone increased theexpression of IL-6 mRNA in the OVCA cell lines OVCA 420,OVCA 429, and OVCA 432 but not in HOSE cell lines. OVCA433 cells did not respond to androgens (results not shown).OVCA 433 cells do not express the androgen receptor (12) andthus cannot respond to androgen (17).

Levels of IL-6 protein were assessed in untreated and treatedHOSE and OVCA cell lines by western blotting and ECL (Fig.1, A). Gonadotropins (FSH and LH) and the two estrogens in-creased levels of IL-6 protein in HOSE 306 and HOSE 642 cellsby 1.5-fold to 2.5-fold and in OVCA 429 and OVCA 432 cellsby 3-fold to 4-fold. 5�-Dehydrotestosterone and testosteroneincreased the level of IL-6 protein in OVCA cell lines by two-fold to threefold. Overall, gonadotropins appeared to be morepotent than sex steroid hormones in stimulating the expressionof IL-6 in all cells.

Fig. 1. Effect of reproductive hormones on the ex-pression of human interleukin 6 (IL-6) messengerRNA (mRNA) and protein in normal immortalizedhuman ovarian surface epithelial (HOSE) cell linesHOSE 306 and HOSE 642 and ovarian cancer(OVCA) cell lines OVCA 429 and OVCA 432. Fol-licle-stimulating hormone (FSH), luteinizing hor-mone (LH), testosterone (T), 5�-dehydrotestoster-one (DHT), estrone (E1), or estradiol (E2) (each at 1�M) was cultured with the cells for 5 days. A) TotalRNA was prepared from all cells, reverse tran-scribed, and amplified by polymerase chain reaction(PCR). The products were separated by electropho-resis in a 2% agarose gel. Relative mRNA levels areexpressed as arbitrary units (intensities of ethidiumbromide-stained PCR products of the target IL-6complementary DNA (cDNA) normalized to thoseof 18S ribosomal RNA [rRNA] cDNA). Proteins incell lysates were separated by electrophoresis, andimmunoblots were prepared and probed with ananti-IL-6 antibody. Relative protein levels are ex-pressed as arbitrary units (intensities of IL-6 proteinbands after radioautography normalized to the inten-sity of glyceraldehyde-3-phosphate dehydrogenase[GAPDH] bands). Solid bars � IL-6 mRNA; openbars � IL-6 protein. Data are the means of threeexperiments. Error bars are the 95% confidenceintervals. Statistically significant increases in levelsof IL-6 mRNA and protein are shown by an asterisk(P<.001 to .02). B) To confirm the specificity ofFSH, LH, testosterone, and estradiol, 2 × 105 cellsper T-25 flask were cultured with 10 nM FSH, 10nM LH, 10 nM testosterone, or 10 nM estradiolalone (solid bars) or treated with 10 nM FSH, 10 nMLH, 10 nM testosterone, or 10 nM estradiol and theprotein kinase A inhibitor H-89 at 0.1 mM (for FSHand LH), 0.1 mM 4-hydroxyflutamide (for andro-gens), or 0.1 mM ICI 182,780 (for estrogens) (openbars) for 5 days. Control cells (hatched bars) weretreated with vehicle alone. After 5 days, total cellu-lar RNA was isolated and analyzed as describedabove. Data are the means of two experiments. Er-ror bars are the 95% confidence intervals. A statis-tically significant decrease in IL-6 mRNA is indi-cated by an asterisk (P<.002 to .01).




ownloaded from


Blocking Gonadotropin-Induced and SteroidHormone-Induced IL-6 Expression

To ascertain whether the observed gonadotropin-stimulatedor steroid-stimulated IL-6 expression is mediated through a re-ceptor-mediated pathway, HOSE and OVCA cell lines weretreated with FSH, LH, testosterone, or 17�-estradiol (each at10 nM) for 5 days in the presence of the protein kinase A-selective inhibitor H-89 at 0.1 mM (for FSH and LH), 4-hy-droxyflutamide at 0.1 mM (for testosterone), the anti-estrogenICI 182,780 at 0.1 mM (for 17�-estradiol), or with mediumalone. For all cell lines, addition of H-89 abolished the expres-sion of gonadotropin-stimulated expression of IL-6 mRNA, andaddition of 4-hydroxyflutamide or ICI 182,780 resulted in amarked decrease in testosterone-stimulated or 17�-estradiol-stimulated IL-6 mRNA expression, respectively (Fig. 1, B).

Enhanced Expression of IL-6R� in OVCA Cell Lines

IL-6 acts through a membrane-associated receptor complexcomposed of the IL-6-specific subunit IL-6R� and the signaltransducer gp130, which is also used by other cytokines (30).OVCA 429 and OVCA 432 cells expressed higher levels ofIL-6R� mRNA and protein than HOSE 642 and HOSE 306 cells(Fig. 2). It is interesting that the relative abundance of gp130mRNA in all cells was comparable. Expression patterns ofIL-6R� mRNA (Fig. 2, A) and protein (Fig. 2, B) were consis-tent. A twofold to threefold increase in IL-6R� protein levelswas noted in OVCA cells (P<.001).

Effect of Anti-IL-6 Antibody on IL-6-Induced Proliferationof HOSE and OVCA Cells

To determine the effect of IL-6 on cell proliferation, increas-ing doses of human recombinant IL-6 (1–32 ng/mL) were addedto HOSE 306, HOSE 642, OVCA 429, and OVCA 432 cellcultures for 5 days, and the number of cells was assessed withthe MTT assay. The number of HOSE cells treated with IL-6was fourfold to fivefold higher than that of untreated HOSE cells(Fig. 3, A). The number of OVCA cells treated with IL-6 wassixfold to sevenfold higher than that of untreated OVCA cells

(Fig. 3, A). Addition of anti-IL-6 antibodies (50 �g/mL) to allHOSE and OVCA cells treated with human recombinant IL-6(1–32 ng/mL) completely abolished IL-6-induced proliferationof HOSE and OVCA cells (Fig. 3, A).

Effect of Anti-IL-6 Antibody on Hormone-InducedProliferation of HOSE and OVCA Cells

HOSE 306, HOSE 642, OVCA 429, and OVCA 432 cellswere cultured with 10 nM FSH, 10 nM LH, 10 nM testosterone,10 nM 5�-dehydrotestosterone, 10 nM estrone, or 10 nM 17�-estradiol alone or with an IL-6 antibody at 50 �g/mL for 5 days,and cell proliferation was assessed by an MTT assay. HOSEcells treated with gonadotropins (FSH and LH) or estrogenscontained fourfold to fivefold more cells than untreated HOSEcells. Addition of IL-6 antibody markedly reduced HOSE cellproliferation stimulated by FSH, LH, estrone, or 17�-estradiol(Fig. 3, B). It is interesting that cell proliferation induced bytestosterone or 5�-dehydrotestosterone was not inhibited byanti-IL-6 antibody in HOSE cells but was reduced by anti-IL-6antibody in OVCA cells. Anti-IL-6 antibody alone or nonim-mune immunoglobulin G (IgG) had essentially no effect on cellproliferation (results not shown).

Overexpression and Constitutive Activation of STAT3 inOVCA Cells

Binding of IL-6 to its receptor induces STAT3 phosphoryla-tion (32); however, in many cancer cells, STAT3 is constitu-tively activated in the absence of IL-6 (34). Consequently, wefirst examined endogenous levels of phosphorylated (activated)and total STAT3 in untreated HOSE and OVCA cells by im-munoblot analysis (Fig. 4, A). Blots of HOSE 306, HOSE 642,OVCA 429, and OVCA 432 cell lysates were probed with aphosphorylation-specific STAT3 antibody or a pan-STAT3 an-tibody that recognizes both phosphorylated and unphosphory-lated STAT3. In the absence of IL-6, OVCA cells expressedhigher levels of total STAT3 protein and phosphorylated STAT3protein than HOSE cells. OVCA cells have the main phosphor-ylated STAT3 band (Fig. 4, A, labeled P-STAT3�) and a fastermigrating phosphorylated STAT3 isoform (labeled P-STAT3�).

Fig. 2. Levels of interleukin 6 receptor (IL-6R) mes-senger RNA (mRNA) and protein and gp130 mRNAin normal immortalized human ovarian surface epi-thelial (HOSE) cells (HOSE 306 and HOSE 642)and ovarian cancer (OVCA) cells (OVCA 429 andOVCA 432). A) Total cellular RNA (1 �g) wasisolated, reverse transcribed, and amplified by poly-merase chain reaction (PCR). Products were sepa-rated by electrophoresis in a 2% agarose gel. Threereverse transcription (RT)-PCRs were performedfrom total RNA of each cell line. Relative mRNAlevels are expressed as arbitrary units as described inFig. 1. Data are the means of three experiments.Error bars are the 95% confidence intervals. Openbars � gp130 mRNA; solid bars � IL-6R�

mRNA. Statistically significant increases are indi-cated by asterisks (*P<.001). B) Proteins in celllysates were separated by electrophoresis, and im-munoblots were prepared and probed with an anti-IL-6R� antibody. Data are the mean of three experi-ments. Error bars are the 95% confidence intervals. The relative abundance of IL-6R� protein of the indicated cell lines is as described in Fig. 1. Statisticallysignificant increases in IL-6R� protein are shown by asterisks (*P<.001).




ownloaded from


The phosphorylation-specific antibody used in this study recog-nizes an epitope common to STAT3� and STAT3� and appearsto have detected two bands with molecular masses correspond-ing to previously described STAT3 isoforms (97 kDa and 87kDa, respectively). The fast migrating isoform may represent apreviously reported dominant negative STAT3 isoform (59).

Levels of STAT3 Transcription Factor in OVCA andHOSE Cells

To determine the levels of STAT3-specific transcription fac-tors in HOSE and OVCA cells, nuclear extracts were preparedfrom HOSE 306, HOSE 642, OVCA 432, and OVCA 429 cells.

After incubation of nuclear extracts from HOSE and OVCAcells with the labeled STAT3 consensus oligonucleotide, whichhas been used to identify STAT3 transcription factors, com-plexes with approximately the same mobility were identified byradioautography (Fig. 4, B). Nuclear extracts from the twoOVCA cell lines had more STAT3 binding activity than nuclearextracts from HOSE cells. In all extracts, addition of a 100-foldexcess of a specific STAT3 competitor oligonucleotide com-pletely abolished the binding of nuclear proteins to the labeledSTAT3 oligonucleotide. In contrast, incubation with a nonspe-cific competitor did not reduce STAT3 binding. Thus, OVCAcells appear to have higher levels of STAT3 transcription factorsthan HOSE cells.

Fig. 3. A) Effect of interleukin 6 (IL-6) on theproliferation of normal immortalized humanovarian surface epithelial (HOSE) cells (HOSE306 and HOSE 642) and ovarian cancer (OVCA)cells (OVCA 429 and OVCA 432). Cells werecultured with IL-6 (1–32 ng/mL) for 5 days orwithout IL-6, as a control, and then cell prolif-eration was measured with a 3-(4,5-dimethyl-thiazol-2-yl)-2,5-diphenyltetrazolium bromide(MTT) assay. Proliferation of IL-6-treated cellswas expressed as a fold increase of untreatedcontrol cells. An anti-IL-6 antibody (Ab) wasadded at a final concentration of 50 �g/mL, withor without IL-6 (1–32 ng/mL), to inhibit IL-6-induced cell proliferation, and 5 days later, cellproliferation was assessed by an MTT assay.Data are the means of two experiments with trip-licate samples. Error bars are the 95% confi-dence intervals. Statistically significant in-creases in cell proliferation are indicated byasterisks (P<.001 to .01). B) Effect of anti-IL-6antibody on gonadotropin-induced and steroidhormone-induced growth of cells. Anti-IL-6 an-tibody was added to cultures to a final concen-tration of 50 �g/mL in the presence or absenceof follicle-stimulating hormone (FSH), luteiniz-ing hormone (LH), testosterone (T), 5�-dihydrotestosterone (DHT), estrone (E1), or es-tradiol (E2) (each at 10 nM). Five days later, cellproliferation was assessed with an MTT assay.Open bars � cells treated with hormone; solidbars � cells treated with hormones and anti-IL-6 antibody; horizontal hatched bars � con-trol cells; vertical hatched bars � cells treatedwith anti-IL-6 antibody. Data are the means oftwo experiments with triplicate samples. Errorbars are the 95% confidence intervals. Statisti-cally significant decreases in cell proliferation,compared with that of hormone-treated cells, areindicated by asterisks (P<.001 to .03). ns � notstatistically significant; † � comparison be-tween untreated cells and cells cultured withanti-IL-6 antibody only, which is not statisticallysignificant. (Data from cells cultured with non-immune immunoglobulin G are not shown.)




ownloaded from


IL-6-Induced STAT3 Expression in HOSE and OVCACells

STAT3 is activated by phosphorylation. To determine theeffect of IL-6 on the activity of STAT3, HOSE 306, HOSE 642,OVCA 429, and OVCA 432 cells were treated with IL-6 at32 ng/mL for 6 hours, and cell lysates were prepared for immu-noblot analysis with the phospho-STAT3 antibody and the pan-

STAT3 antibody. HOSE cells treated with IL-6 had a higherlevel of STAT3 expression than untreated HOSE cells (Fig. 4,C). In HOSE cells, phosphorylation of STAT3 was entirely de-pendent on the presence of IL-6, and in the absence of IL-6,STAT3 was not phosphorylated. In marked contrast, STAT3 inOVCA cells was phosphorylated in the presence or absence ofIL-6, and addition of IL-6 to OVCA cells did not increase thelevel of total or phosphorylated STAT3.

Fig. 4. A) Activity of signal transducer and activator of transcription 3 (STAT3)in ovarian cancer (OVCA) cells (OVCA 429 and OVCA 432) and normalimmortalized human ovarian surface epithelial (HOSE) cells (HOSE 306 andHOSE 642). Cell lysates were prepared. Proteins were separated by electropho-resis, and immunoblots were prepared and probed with phosphorylation-independent and phosphorylated STAT3 antibodies. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control. Arepresentative experiment of at least four others is presented, all with similarresults. B) STAT3 binding activity. Nuclear extracts were prepared, and 10 �gof protein was examined in an electrophoretic mobility shift assay (EMSA) byusing the 32P-labeled 24-base-pair double-stranded STAT3 consensus oligo-nucleotide as the probe. This oligonucleotide has been used to identify STAT3transcription factors. Complexes were resolved on native gels and visualized byautoradiography. To confirm binding specificity, an unlabeled STAT3 consensusoligonucleotide (specific competitor) or mutated STAT3 oligonucleotide (non-specific competitor) was added to the reaction at a 100-fold molar excess. DNA–protein complexes are indicated by the arrow to the right. The experimentshown is representative of at least three experiments, all with similar results.Comp. � specific competitor; NS comp. � nonspecific competitor. C) Inter-leukin 6 (IL-6) activation of STAT3. Cells were treated with IL-6 (32 ng/mL) for6 hours, and lysates were prepared. Proteins were separated by electrophoresis,and immunoblots were prepared and probed with phosphorylation-independentand phosphorylation-dependent STAT3 antibodies. GAPDH was used as a load-ing control. The experiment shown is representative of at least three experiments,all with similar results.




ownloaded from


IL-6-Stimulated Growth of HOSE and OVCA Cells andthe Ectopic Expression of Dn-STAT3

To study the role of IL-6-induced STAT3 in cell prolifera-tion, HOSE 306, HOSE 642, OVCA 429, and OVCA 432 cellswere transiently transfected with 1 �g of an expression vectorcarrying a Dn-STAT3 construct or 1 �g of empty vector. Cul-tures were then treated with IL-6 (16 or 32 ng/mL) for 3 days;control cultures were incubated without IL-6 for 3 days. Prolif-eration in the four cell lines transfected with the empty vectorwas not altered (Fig. 5, A). Treatment with IL-6 increased pro-liferation of all cell lines in a dose-dependent manner. A greaterreduction in IL-6-induced cell proliferation was noted in all cellstransfected with Dn-STAT3 than in corresponding cells trans-fected with the empty vector (Fig. 5, A). A slight statisticallynonsignificant decrease in cell proliferation was observed whenHOSE cells were transfected with Dn-STAT3 alone. OVCAcells, however, transfected with Dn-STAT3 alone showed amarked decrease (P<.04) in cell proliferation (results notshown).

Overexpression of STAT3 and HOSE Cell Proliferation

To investigate whether overexpression of wild-type STAT3affects HOSE cell proliferation, HOSE 306 and HOSE 642 cellswere transfected with a vector containing wild-type STAT3, and72 hours later, cell proliferation was assessed. A greater dose-dependent increase in cell proliferation was seen in the twowild-type STAT3-transfected HOSE cell lines than in untreatedcells (STAT3 at 0.5 �g/mL in HOSE 306 cells, P<.002; STAT3at 1.0 �g/mL in HOSE 306 or both concentrations of STAT3 inHOSE 642 cells, P<.001) or cells transfected with the emptyvector alone (Fig. 5, B).

Colony-Forming Potential of OVCA Cells and the EctopicExpression of Dn-STAT3

Transformed cells can grow in semisolid medium (e.g., softagar). To determine whether STAT3 activation is responsible forthis behavior, OVCA 429 and OVCA 432 cells were transfectedwith 1 �g of Dn-STAT3 DNA or 1 �g of the empty vector DNAand then plated on soft agar. OVCA cells formed fewer soft-agarcolonies after transfection with Dn-STAT3 than after transfec-tion with the empty vector (OVCA 429 cells, P<.001; OVCA432 cells, P<.006) (Fig. 5, C).

DISCUSSION

The goal of this study was to elucidate the role of hormonesin the regulation of IL-6 expression and the effect of IL-6 onnormal (HOSE) and malignant (OVCA) human ovarian surfaceepithelial cells. We capitalized on our access to several immor-talized but nontumorigenic HOSE cell lines, established OVCAcell lines from late-stage serous ovarian cancer, and an expres-sion vector carrying the dominant negative STAT3 construct(Dn-STAT3). Our results demonstrate that 1) gonadotropins andsex steroid hormones increased the expression of IL-6 in HOSEand OVCA cells, 2) an IL-6-neutralizing antibody markedly di-minished gonadotropin-stimulated and sex hormone-stimulatedcell proliferation of HOSE and OVCA cells, 3) IL-6 stimulatedHOSE and OVCA cell proliferation in a dose-dependent,STAT3-mediated manner, 4) IL-6 induced the expression andphosphorylation (activation) of STAT3 in HOSE cells, 5)STAT3 was constitutively phosphorylated (activated) and over-

expressed in OVCA cells in an IL-6-independent manner, 6)OVCA cells had a higher level of STAT3 transcriptional factorsand IL-6R� than HOSE cells, 7) blockade of STAT3 action byectopic expression of a dominant negative STAT3 suppressedthe IL-6-stimulated proliferation of HOSE and OVCA cells andthe anchorage-independent growth of OVCA cells, and 8) over-expression of wild-type STAT3 increased the proliferation ofHOSE cells. To our knowledge, these findings present the firstevidence in support of a growth-regulatory mechanism in HOSEand OVCA cells that is stimulated by reproductive hormonesand mediated by IL-6/STAT3 signaling. Defects in this signalingpathway may be involved in HOSE cell oncogenesis and confermalignant phenotypes to OVCA cells.

Epidemiologic data suggest that gonadotropins and sex ste-roid hormones, especially estrogens, are risk factors for epithe-lial ovarian cancer (3–11). However, the mechanism of hor-mone-induced neoplastic transformation remains unknown. Wehave recently demonstrated that gonadotropins, estrogens, andandrogens promote cell proliferation in primary cultures of nor-mal HOSE cells and immortalized HOSE cell lines (17). Be-cause incessant cell proliferation is a plausible mechanism ofcellular transformation (18), HOSE cell proliferation enhancedby reproductive hormones—during perimenopausal and post-menopausal periods and the use of estrogen/hormone replace-ment therapy—may contribute to tumor initiation. Results fromthis study strongly suggest that reproductive hormones may actthrough a convergent growth-stimulatory pathway that involvesincreased IL-6 expression and STAT3 activation in HOSE cells.We have demonstrated that gonadotropins and estrogens inducehigher levels of IL-6 expression in HOSE cells. Increased IL-6production after hormone treatment coincided with the observedhormone-induced proliferation, which was partially inhibited byan IL-6-neutralizing antibody. The latter finding indicates thatthe hormone-induced increase in cell proliferation is mediatedprincipally by IL-6, although other growth factors may make acontribution. Enhanced IL-6 expression activates STAT3, whichis associated with oncogenesis in other cancers (34). Thegrowth-promoting effect of IL-6 on HOSE cells was clearlydependent on IL-6 and mediated by STAT3, because this effectwas blocked by an IL-6-neutralizing antibody or by the ectopicexpression of Dn-STAT3 in these cells. IL-6-treated HOSE cellsalso had higher levels of STAT3 binding factors than untreatedcells. Pharmacologic concentrations of estrogens have been re-ported to enhance IL-6 expression in mononuclear cells in vitro(50) but to inhibit IL-6 synthesis in a human osteoblastic cellline, vascular smooth muscle cells, and bone marrow-derivedstromal cells (51–53). Gonadotropins, on the other hand, havebeen reported to stimulate IL-6 production in Sertoli cells and inthe ovaries of Cynomolgus fascicularis (45,46). To our knowl-edge, no information was available on the hormonal regulationof IL-6 production in HOSE cells before this study. Hence,findings from this study are the first to establish that reproduc-tive hormones stimulate HOSE cell proliferation (17) through acommon signaling cascade involving IL-6 and STAT3.

Normal IL-6 signal transduction involves formation of a com-plex between IL-6 and its cell surface receptor, which is com-posed of IL-6R� and the signal transducer gp130. IL-6R� isspecific for IL-6, whereas gp130 is also a component of cytokinereceptors (30). After IL-6 signaling, Janus kinase mediates theactivation of STAT3, and activated STAT3 transmits a messagefrom the cell surface to the nucleus that turns on specific gene




ownloaded from


Fig. 5. A) Effect of the dominant negative signal transducer and acti-vator of transcription 3 (STAT3) and interleukin 6 (IL-6) on the pro-liferation of normal immortalized human ovarian surface epithelial(HOSE) cells (HOSE 306 and HOSE 642) and ovarian cancer (OVCA)cells (OVCA 429 and OVCA 432). Cells were transfected with a vectorcontaining STAT3 (Dn-STAT3) or an empty vector, as indicated, andtreated with IL-6 (16 or 32 ng/mL) for 3 days; control cells were neithertransfected nor treated with IL-6. After the 3-day incubation, cells werecounted with a hemocytometer, and proliferation of IL-6-treated cellswas expressed as a fold increase of that of control cells. B) STAT3-induced proliferation of normal immortalized HOSE 306 and HOSE642 cells. Cells were transfected with various concentrations of vectorcontaining STAT3 (Dn-STAT3) or empty vector, and 72 hours later,cells were counted with a hemocytometer. Proliferation of STAT3-transfected cells is expressed as a fold increase of that of control cells.C) Dominant negative STAT3 and the colony formation ability ofOVCA 429 and OVCA 432 cells. Cells were transiently transfectedwith a vector containing a dominant negative STAT3 (Dn-STAT3) oran empty vector. Twenty-four hours later, cells were cultured on softagar, and 14 days later, colonies were counted. A and B) Data are themeans of two experiments with triplicate samples. C) Data are themeans of two experiments, with four plates for each point. Error barsare the 95% confidence intervals. Statistically significant increases(A and B) or decreases (C) in colony number are indicated with as-terisks (for all cells in A, *P<.004 to .02; for B, wild-type STAT3 at0.5 �g/mL in HOSE 306 cells, *P<.002, and for wild-type STAT3 at1.0 �g/mL in HOSE 306 cells and both concentrations in HOSE 642cells, **P<.001; for C, OVCA 429 cells, *P<.001, and OVCA 432cells, **P<.006).




ownloaded from


programs for cell proliferation and other functions. Inappropriateactivation of this signal transduction pathway is found in manyhuman cancers (34). It has been widely reported that IL-6 iselevated in serum and ascites fluid from patients with ovariancancer, and high levels of IL-6 in body fluids are associated withpoor prognosis and survival (35–37). In this study, we haveidentified several alterations in the IL-6/STAT3 signaling cas-cade in OVCA cells. 1) Higher levels of IL-6R� were detectedin OVCA cells than in HOSE cells. This observation was con-sistent with the slightly higher IL-6-stimulated proliferation inOVCA cells than in HOSE cells. 2) High levels of phosphory-lated STAT3 were detected in OVCA cells in the absence ofIL-6. 3) EMSA results demonstrated higher STAT3 binding ac-tivity in nuclear extracts of OVCA cells than in nuclear extractsof HOSE cells. 4) Inhibition of STAT3 action by the ectopicexpression of Dn-STAT3 demonstrated the importance ofSTAT3 in maintaining OVCA cell proliferation and anchorage-independent colony formation. Thus, the IL-6/STAT3 signalingpathway is constitutively activated and is more active in OVCAcells than in HOSE cells. These findings are in agreement withthose reported by Huang et al. (60), who demonstrated that theconstitutive activation of STAT3 was accompanied by overex-pression of Bcl-XL (i.e., the BCL2L1 gene) and cyclin D1 inseveral OVCA cell lines. Elevation of nuclear STAT3 bindingactivity has also been reported in cancers of the breast, prostate,lymphoid tissues, and the head and neck (27,34,61–63). Thus, itis now widely accepted that constitutively activated STAT3 sig-naling directly contributes to oncogenesis and maintenance ofthe malignant state (34). STAT3 regulates an array of genes,such as those for Bcl-XL, cyclin D1, p21WAF1/CIP1, and c-Myc,that are regulators of cell cycle progression and apoptosis (34).Aberrant expression of these genes, as a result of the inappro-priate activation STAT3, could promote tumor initiation andprogression (62,64,65). Although constitutive action of STAT5,like STAT3, has been shown directly to result in cellular trans-formation, IL-6 primarily activates STAT3 and STAT1, andseveral studies have shown that STAT1 does not contribute tooncogenesis. We, therefore, studied only STAT3-mediated IL-6activities in HOSE and OVCA cells.

HOSE and OVCA cells exhibited differential responses toandrogen (testosterone and 5�-dehydrotestosterone)-regulatedIL-6 expression. HOSE cells did not respond to either androgenby increasing the expression of IL-6. However, these two an-drogens do increase HOSE cell proliferation (17). Thus, andro-gen-induced HOSE cell proliferation appears to result fromthe direct action of androgen and not the indirect induction ofIL-6. The inability of an IL-6-neutralizing antibody to inhibitHOSE cell proliferation supports this conclusion. It is interestingthat androgens stimulated cell proliferation and the expressionof IL-6 in OVCA cells. Androgen-stimulated proliferationof OVCA cells, in contrast to the proliferation of HOSE cells,could be partially blocked by an IL-6-neutralizing antibody.At present, the importance of this finding is uncertain anddefinitely warrants further investigation because many ovariantumors produce androgens (66). However, studies in other celland tissue types clearly show great variability in the mechanismof androgen-regulated IL-6 expression. For example, both tes-tosterone and 5�-dehydrotestosterone inhibit IL-6 productionin stromal cells derived from murine bone marrow (47), brainglial cells and astrocytes (48), and a human osteoblastic cell line(49). Yet, a low dose of testosterone in Sertoli cells stimulates

IL-6 expression, whereas higher doses inhibit IL-6 expression(46).

In conclusion, our data demonstrate, to our knowledge for thefirst time, that gonadotropins and sex steroid hormones increasethe expression of IL-6 in HOSE and OVCA cells. By activatingSTAT3, IL-6 serves as a common growth regulator for bothHOSE and OVCA cells. More importantly, increased STAT3activity may directly contribute to HOSE cell transformation andthe progression of ovarian cancer. Lastly, enhanced expressionof IL-6R�, constitutive activation of STAT3, and acquired an-drogen regulation of IL-6 production have been identified asdefects associated with ovarian malignancy. Thus, reproductivehormone-stimulated activation of the IL-6/STAT3 pathway mayplay a role in the pathogenesis and progression of ovarian can-cer.

REFERENCES

(1) Risch HA. Hormonal etiology of epithelial ovarian cancer, with a hypoth-esis concerning the role of androgens and progesterone. J Natl Cancer Inst1998;90:1774–86.

(2) Auersperg N, Wong AS, Choi KC, Kang SK, Leung PC. Ovarian surfaceepithelium: biology, endocrinology, and pathology. Endocr Rev 2001;22:255–88.

(3) Whittemore AS, Harris R, Itnyre J. Characteristics relating to ovarian can-cer risk: collaborative analysis of 12 US case–control studies. IV. Thepathogenesis of epithelial ovarian cancer. Collaborative Ovarian CancerGroup. Am J Epidemiol 1992;136:1212–20.

(4) Rodriguez C, Calle EE, Coates RJ, Miracle-McMahill HL, Thun MJ, HeathCW Jr. Estrogen replacement therapy and fatal ovarian cancer. Am J Epi-demiol 1995;141:828–35.

(5) Shushan A, Paltiel O, Iscovich J, Elchalal U, Peretz T, Schenker JG.Human menopausal gonadotropin and the risk of epithelial ovarian cancer.Fertil Steril 1996;65:13–8.

(6) Clinton GM, Hua W. Estrogen action in human ovarian cancer. Crit RevOncol Hematol 1997;25:1–9.

(7) Garg PP, Kerlikowske K, Subak L, Grady D. Hormone replacementtherapy and the risk of epithelial ovarian carcinoma: a meta-analysis. Ob-stet Gynecol 1998;92:472–9.

(8) Helzlsouer KJ, Alberg AJ, Gordon GB, Longcope C, Bush TL, HoffmanSC, et al. Serum gonadotropins and steroid hormones and the developmentof ovarian cancer. JAMA 1995;274:1926–30.

(9) Coughlin SS, Giustozzi A, Smith SJ, Lee NC. A meta-analysis of estrogenreplacement therapy and risk of epithelial ovarian cancer. J Clin Epidemiol2000;53:367–75.

(10) Sittisomwong T, Suneja A, Kudelka AP, Verschraegen CF, Kavanagh JJ.Estrogen replacement therapy and ovarian cancer. Eur J Gynaecol Oncol2000;21:348–54.

(11) Rodriguez C, Patel AV, Calle EE, Jacob EJ, Thun MJ. Estrogen replace-ment therapy and ovarian cancer mortality in a large prospective study ofUS women. JAMA 2001;285:1460–5.

(12) Lau KM, Mok SC, Ho SM. Expression of human estrogen receptor-alphaand -beta, progesterone receptor, and androgen receptor mRNA in normaland malignant ovarian epithelial cells. Proc Natl Acad Sci U S A 1999;96:5722–7.

(13) Zheng W, Magid MS, Kramer EE, Chen YT. Follicle-stimulating hormonereceptor is expressed in human ovarian surface epithelium and fallopiantube. Am J Pathol 1996;148:47–53.

(14) Mandai M, Konishi I, Kuroda H, Fukumoto M, Komatsu T, Yamamoto S,et al. Messenger ribonucleic acid expression of LH/hCG receptor gene inhuman ovarian carcinomas. Eur J Cancer 1997;33:1501–7.

(15) Konishi I, Kuroda H, Mandai M. Review: gonadotropins and developmentof ovarian cancer. Oncology 1999;57 Suppl 2:45–8.

(16) Zheng W, Lu JJ, Luo F, Zheng Y, Feng Y, Felix JC, et al. Ovarian epi-thelial tumor growth promotion by follicle-stimulating hormone andinhibition of the effect by luteinizing hormone. Gynecol Oncol 2000;76:80–8.




ownloaded from


(17) Syed V, Ulinski G, Mok SC, Yiu GK, Ho SM. Expression of gonadotropinreceptor and growth responses to key reproductive hormones in normal andmalignant human ovarian surface epithelial cells. Cancer Res 2001;61:6768–76.

(18) Henderson BE, Feigelson HS. Hormonal carcinogenesis. Carcinogenesis2000;21:427–33.

(19) Seiler P, Plenz G, Deng MC. The interleukin-6 cytokine system in embry-onic development, embryo-maternal interactions and cardiogenesis. EurCytokine Netw 2001;12:15–21.

(20) Vanden Berghe W, Vermeulen L, De Wilde G, De Bosscher K, Boone E,Haegeman G. Signal transduction by tumor necrosis factor and gene regu-lation of the inflammatory cytokine interleukin-6. Biochem Pharmacol2000;60:1185–95.

(21) Horn F, Henze C, Heidrich K. Interleukin-6 signal transduction and lym-phocyte function. Immunobiology 2000;202:151–67.

(22) Miki S, Iwano M, Miki Y, Yamamoto M, Tang B, Yokokawa K, et al.Interleukin-6 (IL-6) functions as an in vitro autocrine growth factor in renalcell carcinomas. FEBS Lett 1989;250:607–10.

(23) Nakajima K, Yamanaka Y, Nakae K, Kojima H, Ichiba M, Kiuchi N, et al.A central role for Stat3 in IL-6-induced regulation of growth and differ-entiation in M1 leukemia cells. EMBO J 1996;15:3651–8.

(24) Okamoto M, Lee C, Oyasu R. Interleukin-6 as a paracrine and autocrinegrowth factor in human prostatic carcinoma cells in vitro. Cancer Res 1997;57:141–6.

(25) Kurebayashi J. Regulation of interleukin-6 secretion from breast cancercells and its clinical implications. Breast Cancer 2000;7:124–9.

(26) Barille S, Bataille R, Amiot M. The role of interleukin-6 and interleukin-6/interleukin-6 receptor-alpha complex in the pathogenesis of multiplemyeloma. Eur Cytokine Netw 2000;11:546–51.

(27) Lou W, Ni Z, Dyer K, Tweardy DJ, Gao AC. Interleukin-6 induces prostatecancer cell growth accompanied by activation of stat3 signaling pathway.Prostate 2000;42:239–42.

(28) Smith PC, Hobisch A, Lin DL, Culig Z, Keller ET. Interleukin-6 andprostate cancer progression. Cytokine Growth Factor Rev 2001;12:33–40.

(29) Yamasaki K, Taga T, Hirata Y, Yawata H, Kawanishi Y, Seed B, et al.Cloning and expression of the human interleukin-6 (BSF-2/IFN beta 2)receptor. Science 1988;241:825–8.

(30) Hirano T. Interleukin 6 and its receptor: ten years later. Int Rev Immunol1998;16:249–84.

(31) Schindler C, Darnell JE Jr. Transcriptional responses to polypeptide li-gands: the JAK–STAT pathway. Annu Rev Biochem 1995;64:621–51.

(32) Heinrich PC, Behrmann I, Muller-Newen G, Schaper F, Graeve L. Inter-leukin-6-type cytokine signalling through the gp130/Jak/STAT pathway.Biochem J 1998;334(Pt 2):297–314.

(33) Hirano T, Ishihara K, Habi M. Roles of STAT3 in mediating the cellgrowth, differentiation and survival signals relayed through the IL-6 familyof cytokine receptors. Oncogene 2000;19:2548–56.

(34) Bowman T, Garcia R, Turkson J, Jove R. STATs in oncogenesis. Oncogene2000;19:2474–88.

(35) Watson JM, Sensintaffar JL, Berek JS, Martinez-Maza O. Constitutiveproduction of interleukin 6 by ovarian cancer cell lines and by primaryovarian tumor cultures. Cancer Res 1990;50:6959–65.

(36) Lidor YJ, Xu FJ, Martinez-Maza O, Olt GJ, Marks JR, Berchuck A, et al.Constitutive production of macrophage colony-stimulating factor and in-terleukin-6 by human ovarian surface epithelial cells. Exp Cell Res 1993;207:332–9.

(37) Plante M, Rubin SC, Wong GY, Federici MG, Finstad CL, Gastl GA.Interleukin-6 level in serum and ascites as a prognostic factor in patientswith epithelial ovarian cancer. Cancer 1994;73:1882–8.

(38) Scambia G, Testa U, Panici PB, Martucci R, Foti E, Petrini M, et al.Interleukin-6 serum levels in patients with gynecological tumors. IntJ Cancer 1994;57:318–23.

(39) Tempfer C, Zeisler H, Sliutz G, Haeusler G, Hanzal E, Kainz C. Serumevaluation of interleukin 6 in ovarian cancer patients. Gynecol Oncol 1997;66:27–30.

(40) Eskandari N, Gage J, Johnson MT, Martínez-Maza O. Cytokine-mediatedmodulation of cisplatin sensitivity in ovarian cancer cells. Obstet Gynecol2001;97(4 Suppl 1):S2.

(41) Penson RT, Kronish K, Duan Z, Feller AJ, Stark P, Cook SE, et al. Cy-tokines IL-1beta, IL-2, IL-6, IL-8, MCP-1, GM-CSF and TNFalpha in

patients with epithelial ovarian cancer and their relationship to treatmentwith paclitaxel. Int J Gynecol Cancer 2000;10:33–41.

(42) Moradi MM, Carson LF, Weinberg B, Haney AF, Twiggs LB, Ra-makrishnan S. Serum and ascitic fluid levels of interleukin-1, interleukin-6,and tumor necrosis factor-alpha in patients with ovarian epithelial cancer.Cancer 1993;72:2433–40.

(43) Obata NH, Tamakoshi K, Shibata K, Kikkawa F, Tomoda Y. Effects ofinterleukin-6 on in vitro cell attachment, migration and invasion of humanovarian carcinoma. Anticancer Res 1997;17:337–42.

(44) Loret de Mola JR, Goldfarb JM, Hecht BR, Baumgardner GP, Babbo CJ,Friedlander MA. Gonadotropins induce the release of interleukin-1 beta,interleukin-6 and tumor necrosis factor-alpha from the human preovulatoryfollicle. Am J Reprod Immunol 1998;39:387–90.

(45) Machelon V, Gougeon A, Duquenne C, Testart J, Wallon C, Delattre RM,et al. Ovarian production of IL6 and its potential inhibitory effect on pro-gesterone secretion in Cynomolgus fascicularis. C R Acad Sci III 1995;318:1111–8.

(46) Stephan JP, Syed V, Jegou B. Regulation of Sertoli cell IL-1 and IL-6production in vitro. Mol Cell Endocrinol 1997;134:109–18.

(47) Bellido T, Jilka RL, Boyce BF, Girasole G, Broxmeyer H, Dalrymple SA,et al. Regulation of interleukin-6, osteoclastogenesis, and bone mass byandrogens. The role of the androgen receptor. J Clin Invest 1995;95:2886–95.

(48) McCarty MF. Interleukin-6 as a central mediator of cardiovascular riskassociated with chronic inflammation, smoking, diabetes, and visceral obe-sity: down-regulation with essential fatty acids, ethanol and pentoxifylline.Med Hypotheses 1999;52:465–77.

(49) Hofbauer LC, Ten RM, Khosla S. The anti-androgen hydroxyflutamide andandrogens inhibit interleukin-6 production by an androgen-responsive hu-man osteoblastic cell line. J Bone Miner Res 1999;14:1330–7.

(50) Li ZG, Danis VA, Brooks PM. Effect of gonadal steroids on the productionof IL-1 and IL-6 by blood mononuclear cells in vitro. Clin Exp Rheumatol1993;11:157–62.

(51) Kassem M, Harris SA, Spelsberg TC, Riggs BL. Estrogen inhibits inter-leukin-6 production and gene expression in a human osteoblastic cell linewith high levels of estrogen receptors. J Bone Miner Res 1996;11:193–9.

(52) Okubo T, Urabe M, Tsuchiya H, Iwasa K, Yokota K, Kikuchi N, et al.Effect of estrogen and progesterone on gene expression of growth regula-tory molecules and proto-oncogene in vascular smooth muscle cells.Endocr J 2000;47:205–14.

(53) Girasole G, Jilka RL, Passeri G, Boswell S, Boder G, Williams DC, et al.17 beta-estradiol inhibits interleukin-6 production by bone marrow-derivedstromal cells and osteoblasts in vitro: a potential mechanism for the anti-osteoporotic effect of estrogens. J Clin Invest 1992;89:883–91.

(54) Rauh-Adelmann C, Lau KM, Sabeti N, Long JP, Mok SC, Ho SM. Alteredexpression of BRCA1, BRCA2, and a newly identified BRCA2 exon 12deletion variant in malignant human ovarian, prostate, and breast cancercell lines. Mol Carcinog 2000;28:236–46.

(55) Tsao SW, Mok SC, Fey EG, Fletcher JA, Wan TS, Chew EC, et al.Characterization of human ovarian surface epithelial cells immortalized byhuman papilloma viral oncogenes (HPV-E6E7 ORFs). Exp Cell Res 1995;218:499–507.

(56) Bast RC Jr, Feeney M, Lazarus H, Nadler LM, Colvin RB, Knapp RC.Reactivity of a monoclonal antibody with human ovarian carcinoma. J ClinInvest 1981;68:1331–7.

(57) Kaptein A, Paillard V, Saunders M. Dominant negative stat3 mutant in-hibits interleukin-6-induced Jak–STAT signal transduction. J Biol Chem1996;271:5961–4.

(58) Salmon SE, Hamburger AW, Soehnlen B, Durie BG, Alberts DS, MoonTE. Quantitation of differential sensitivity of human-tumor stem cells toanticancer drugs. N Engl J Med 1978;298:1321–7.

(59) Caldenhoven E, van Dijk TB, Solari R, Armstrong J, Raaijmakers JA,Lammers JW, et al. STAT3beta, a splice variant of transcription factorSTAT3, is a dominant negative regulator of transcription. J Biol Chem1996;271:13221–7.

(60) Huang M, Page C, Reynolds RK, Lin J. Constitutive activation of stat 3oncogene product in human ovarian carcinoma cells. Gynecol Oncol 2000;79:67–73.

(61) Watson CJ, Miller WR. Elevated levels of members of the STAT family oftranscription factors in breast carcinoma nuclear extracts. Br J Cancer 1995;71:840–4.




ownloaded from


(62) Catlett-Falcone R, Landowski TH, Oshiro MM, Turkson J, Levitzki A,Savino R, et al. Constitutive activation of Stat3 signaling confers resistanceto apoptosis in human U266 myeloma cells. Immunity 1999;10:105–15.

(63) Spiotto MT, Chung TD. STAT3 mediates IL-6-induced growth inhibition inthe human prostate cancer cell line LNCaP. Prostate 2000;42:88–98.

(64) Karni R, Jove R, Levitzki A. Inhibition of pp60c-Src reduces Bcl-XLexpression and reverses the transformed phenotype of cells overexpressingEGF and HER-2 receptors. Oncogene 1999;18:4654–62.

(65) Bromberg JF, Wrzeszczynska MH, Devgan G, Zhao Y, Pestell RG, Alba-nese C, et al. Stat3 as an oncogene. Cell 1999;98:295–303.

(66) Abrahamsson G, Janson PO, Kullander S. Steroid release from two humanepithelial ovarian tumors: evidence for an intrinsic production in vitro.Gynecol Oncol 1997;64:99–104.

NOTES

Supported by Army Ovarian Cancer Research Program grant DAMD17-99-1-9563 (to S. C. Mok and S.-M. Ho) and Public Health Service grant CA78523from the National Cancer Institute, National Institutes of Health, Department ofHealth and Human Services (to S. C. Mok).

We thank Dr. Koichi Nakajima of Osaka University Medical School (Osaka,Japan) for permitting us to use his STAT3 and dominant negative STAT3 con-structs and Dr. Allen C. Gao, University of Pittsburgh Medical Center (Pitts-burgh, PA) for providing these constructs to us.

Manuscript received October 9, 2001; revised February 5, 2002; acceptedFebruary 21, 2002.




ownloaded from


Reproductive Hormone-Induced, STAT3-Mediated Interleukin 6 Action in Normal and Malignant Human...

Documents

Transcript of Reproductive Hormone-Induced, STAT3-Mediated Interleukin 6 Action in Normal and Malignant Human...