Microenvironmental Regulation of BRCA1 Gene Expression by C-Jun and Fra2 in Premalignant Human Ovarian Surface Epithelial Cells

Published OnlineFirst January 21, 2013; DOI: 10.1158/1541-7786.MCR-12-0395

Molecular Cancer DNA Damage and Repair Research

Microenvironmental Regulation of BRCA1 Gene Expression by c-Jun and Fra2 in Premalignant Human Ovarian Surface Epithelial Cells

Lixin Zhou1, Marcia Graves1, Gwen MacDonald4, Jane Cipollone1, Christopher R. Mueller2,3, and Calvin D. Roskelley1

Abstract Reduced BRCA1 gene expression is common in the sporadic form of ovarian carcinoma. The spread of this highly lethal cancer often begins when tumor cell clusters are shed into the fluid of the abdominopelvic cavity such that they can float freely before seeding distant sites on the peritoneal walls and organs. Thus, the microenvironment that tumor cells find themselves in changes dramatically during these early shedding and floating stages of transperitoneal metastasis. To mimic this microenvironmental change in vitro, we released premalignant human ovarian surface epithelial cells from the substratum and forced them to cluster in suspension. Under these conditions, steady state levels of BRCA1 mRNA and protein fell significantly and the transcriptional activation state of the BRCA1 promoter was suppressed. Analysis of the promoter indicated that the previously identified "CRE" element located within the "positive regulatory region" (PRR) contributed to this suppression. More specifically, we show that the suppression was mediated, at least in part, by a suspension culture–driven decrease in the levels of two members of the AP1 transcription factor complex, c-Jun and Fra2, that bind to the CRE element. Therefore, a microenvironmental change that is manifested during the initial stages of ovarian carcinoma dissemination may, potentially, help suppress BRCA1 expression in sporadic tumors and thus promote their progression. Mol Cancer Res; 11(3); 272–81. 2013 AACR.

Introduction how its expression is silenced during the development of Epithelial ovarian carcinoma is the leading cause of death high-grade ovarian tumors is an important goal. due to gynecologic malignancy (1). The great majority of BRCA1 expression is tightly regulated through its pro- these deaths occur in women who present with high-grade moter, which is located within a 218 bp intragenic region serous subtype tumors that metastasize by spreading trans- between the bidirectionally controlled BRCA1 and NBR2 peritoneally through the abdominopelvic cavity. High-grade genes (3). A strong positive regulatory region (PRR), which serous subtype tumors are also notable because they also have consists of 2 subregions, lies within a 56 bp sequence from nucleotide 204 to 148 relative to the BRCA1 start site (4, the highest rate of BRCA1 abnormalities, a considerable fi proportion of which occur due to epigenetic silencing rather 5). The rst upstream subregion in the PRR is the "RIBS" 0 element (204 nt to 182 nt), which binds to, and is than mutation (2). Given BRCA1 s myriad roles in acting to a b maintain genomic stability and differentiation, determining activated by, a complex of GABP and GABP transcription factors of the ETS family (6). The second subregion is the CRE element located just downstream (174 nt to 167 nt), which binds to, and is activated by, a complex of c- Authors' Affiliations: 1Department of Cellular and Physiological Sciences, Jun and Fra2 transcription factors of the AP1 complex (3, 7). University of British Columbia, Vancouver, British Columbia; Departments To date, the transcriptional regulation of the BRCA1 pro- of 2Biochemistry and 3Pathology and Molecular Medicine, Queens Uni- versity, Kingston, Ontario, Canada; and 4Friedrich Miescher Institute for moter has been most thoroughly characterized in breast Biomedical Research, Basel, Switzerland carcinoma cells (8). However, we have shown that the CRE Note: Supplementary data for this article are available at Molecular Cancer element is active in ovarian carcinoma cells (3). Research Online (http://mcr.aacrjournals.org/). It has long been assumed that ovarian carcinomas arise within the ovarian surface epithelium (OSE; ref. 9). How- C.R. Mueller and C.D. Roskelley are co-corresponding authors for this fi article. ever, recent ndings suggest that some ovarian carcinomas may arise in extraovarian tissues (10). In the case of high- Corresponding Author: Calvin Roskelley, Department of Cellular and fi Physiological Sciences, Life Sciences Institute, University of British Colum- grade serous subtype tumors speci cally, there is now con- bia, 2350 Health Sciences Mall, Vancouver, BC, Canada. Phone: 604-822- siderable evidence that these lesions often arise in lining 0779; Fax: 604-822-2316; E-mail: [email protected] epithelium of the distal portion of the fallopian tubes, which doi: 10.1158/1541-7786.MCR-12-0395 is very closely related, developmentally, to the OSE (11). 2013 American Association for Cancer Research. The fallopian tube origin of high-grade serous tumors has

272 Mol Cancer Res; 11(3) March 2013

BRCA1 Suppression in Ovarian Cancer Progression

many implications for clinical management, including Real-time quantitative PCR changes to surgical intervention protocols for women who Total RNA was extracted from IOSE-80pc cells main- are BRCA1 mutation carriers who choose prophylactic early tained in either monolayer or suspension culture. cDNA was lesion removal before frank tumor diagnosis (12). In addi- then synthesized using a high-capacity cDNA archive kit tion, it indicates that the cells of frank high-grade serous (Applied Biosystems) and real-time PCR (RT-PCR) analysis tumors have likely already been shed into the fluid of the was conducted using 20 ng of template in 15 mL reactions peritoneal cavity and then later attach to either to the ovarian containing TaqMan universal PCR master mix as well as surface or the abdominopelvic wall where they are most often BRCA1 primers and probe (HS00173237; Applied Biosys- first discovered (13). Given that altered interactions with tems). 18s rRNA was simultaneously quantified as the extracellular matrix (ECM) substrata have been shown to internal control. Amplification was conducted in an ABI alter BRCA1 expression in mammary epithelial cells (14, prism 7500 HT sequence detector using standard settings of 15), we reasoned that detaching premalignant ovarian cells 40 cycles with an annealing temperature of 60C. All from the substratum, in an effort to mimic their initial samples were assayed in 3 parallel reactions, and 3 indepen- shedding and floating in the peritoneal fluid, might also dent experiments were carried out with similar results. affect BRCA1 expression. Premalignant, immortal human OSE cells were cultured Antibodies either as attached, flat monolayers, or as detached cell clusters The mouse monoclonal anti-BRCA1 antibody (Ab-1) in suspension. Under the latter conditions, steady-state was obtained from Oncogene. Anti-c-Jun (H-79), anti- BRCA1 mRNA and protein levels were significantly Fra2 (Q-20), anti-GABPa (H-180), anti-GABPb (H- reduced. Furthermore, the transcriptional activity of the 265), anti-SP1, and anti-extracellular signal–regulated BRCA1 promoter was suppressed in suspension and this kinase (ERK)1/2 antibodies were all from Santa Cruz suppression was mediated, at least in part, by the decreased Biotechnology. The anti-His-tag and anti-Flag-tag anti- expression of c-Jun and Fra2, both of which bind to the CRE bodies were obtained from Applied Biological Materials element within the promoter. This suggests that the micro- Inc. and Sigma, respectively. Alexa-Fluor–conjugated sec- environmental loss of cell–substratum adhesion that occurs ondary antibodies for immunofluorescence were from during the earliest stages of transperitoneal spread may help Molecular Probes (Life Technologies). Peroxidase-conju- facilitate the decrease in endogenous BRCA1 expression that gated secondary antibodies for Western blotting were from has been noted in a significant proportion of high-grade Jackson Immuno-Research. serous ovarian tumors. Plasmids Materials and Methods The L6 proximal promoter region of BRCA1 was excised Cell culture from the L6-pRL construct (7) with BglII and HindIII and Extended lifespan human OSE IOSE-80 cells, which have ligated directly into the pGL4.11 (Luc2P; Promega) basic, escaped normal growth controls were generated by stable minimal promoter vector containing the firefly luciferase transfection with SV40 virus large T antigen in the labora- reporter. To generate the BRCA1 promoter mutants used tory of Nelly Auersperg (Women's Hospital, Vancouver, here, the respective DNA strands were synthesized, Canada) as previously described (16). We subsequently annealed, and ligated into pGL4.11 as previously described expanded a population of IOSE-80 cells that had passed (7, 19). For the PRR construct, 2 synthetic primers (50- through crisis to generate the truly immortal, premalignant GATCTTTCTTCCTCTTCCGTCTCTTTCCTTTTA- IOSE-80pc line (17). In all experiments carried out here, CGTCATCCGGGGGCAGACA-30 and 50-AGCTTGT- IOSE-80pc cells were maintained in 199:105 medium (1:1; CTGCCCCCGGATGACGTAAAAGGAAAGAGACGG- Sigma) supplemented with 5% FBS (Invitrogen) and 50 mg/ AAGAGGAAGAAA-30) were designed to include the mL gentamycin (Sigma). For monolayer culture, cells were BRCA1 promoter region from 201 to 155 and to plated on regular tissue culture plates at subconfluent den- contain adapters of the BglII and HindIII enzymes for sities (5 104 cells/cm2). For suspension culture, tissue annealing and ligation. The "RIBSm" (previously described culture plates were precoated with 5 mg/mL of nonadhesive as "CREs") and "CREm" mutants shown (see Fig. 2) were poly[2-hydroxyethyl methacrylate] (polyHEMA, Sigma; then generated by site-directed mutagenesis as previously ref. 18) and cells were plated at the density described earlier. described (6, 7). All promoter/reporter constructs were Under the latter conditions, the cells did not attach to the confirmed by sequencing. dish. Instead, they floated in suspension and, given the His-tagged c-Jun- and Fra2-containing expression vectors relatively high density of plating, aggregated and formed were gifts from Dr. Irina A. Udalova (Imperial College, small rounded clusters (see Results, later). Conditioned London; ref. 20). The dominant negative AP-1 mutant "A- medium, generated when cells were maintained in either fos" expression vector was a gift from Dr. Charles Vinson monolayer or suspension culture, was cleared of cells by (National Cancer Institute, Bethesda, MD; ref. 21). centrifugation (200 g, 5 minutes) and added to cultures immediately. Live, phase images of monolayer, and suspen- Transient transfection and luciferase reporter assays sion cultures were collected using a Nikon TMS phase Cells in monolayer culture were transfected using lipo- microscope. fectamine (Life Technologies) with 0.5 mg of BRCA1

www.aacrjournals.org Mol Cancer Res; 11(3) March 2013 273

Zhou et al.

promoter/firefly luciferase reporter constructs together with EdU proliferation assay 0.01 mg of pGL4.73 SV40 promoter/Renilla luciferase Cells monolayers transfected with His-tagged c-Jun and reporter construct. The latter reporter served as an internal Fra2 expression vectors were maintained for 24 hours control for transfection efficiency given that it the SV-40 followed by dissociation and maintenance in either mono- promoter was not affected by the changes in culture con- layer or suspension culture for a further 48 hours. The dition. Alternatively, both reporter constructs were cotrans- cells were then incubated for 2 hours with 20 mmol/L 5- fected into cells in monolayer culture with either 0.25 mg ethynyl-20-deoxyuridine (EdU; Invitrogen), which fluores- each of the expression vectors for wild-type c-Jun and Fra2 cently labels proliferating cells in S-phase of the cell cycle. together, or 0.50 mg of the A-Fos dominant negative AP1 Fluorescent, proliferating cells were then quantified using a mutant alone. Twenty-four hours after transfection, the cells FACS LSRII flow cytometer (BD Biosciences). were dissociated and cultured for a further 48 hours as either flat monolayers or as cell clusters in suspension. Firefly and Chromatin immunoprecipitation Renilla luciferase activities were then determined using a Cells were plated at a density of 2 105 cells/mL on commercial dual luciferase assay kit (Promega). Graphical normal 100-mm tissue culture plates and maintained as flat and multiple t test statistical analyses were conducted using monolayers for 48 hours. The chromatin immunoprecipi- GraphPad Prism6 (GraphPad). A P value of less than 0.05 tation (ChIP) analysis was conducted according to manu- was considered statistically significant. facturer's instructions for the ChIP-IT Express Kit (Active c-Jun- and Fra2-specific, as well as random control, Motif). This included enzymatic digestion of the chromatin siRNAs were purchased from Santa Cruz Biotechnology. and the magnetic immunoprecipitation of the antibody Cell monolayers were transfected with siRNAs using lipo- complexes. Antibodies used were directed against acetylated fectamine RNAiMAX (Life Technologies) with a total of 25 histone H3 lysine K9 (Abcam) as well as hemagglutanin, nmol/L siRNA (12.5 nmol/L c-Jun siRNA plus 12.5 nmol/L GABPa, Fra-2, and c-Jun, all of which were from Santa Fra2 siRNA; or 25 nmol/L total of random siRNAs). After Cruz. 24 hours, cells were dissociated and were then maintained in The PCR reaction was carried out as previously described either monolayer or suspension culture for a further 48 hours (22) after adaptation from Bindra and colleagues (23) using before analysis by Western blotting. the following primers: PF2 50–GATTGGGACCTCTTCT- TACG-30; pR2 50–TACCCAGAGCAGAGGGTGAA-30. Western blotting The PCR conditions were 1 Theromopol buffer, 1 Cells were lysed on ice in radioimmunoprecipitation assay mol/L betaine, 200 mmol/L dNTPs, 250 mg each oligonu- (RIPA) buffer [159 mmol/L NaCl, 50 mmol/L Tris–HCl, cleotide in a 25 mL reaction with 0.5 units of Taq poly- pH 7.5, 1% NP-40, 1% deoxycholate, 0.1% SDS, 2 mmol/ merase. The reactions conditions were 95C for 3 minutes, L EDTA, plus 1 mmol/L pepstatin, 10 mg/mL aprotinin, 10 38 cycles of 95C, 30 seconds, 60C, 30 seconds; 72C, 30 mmol/L phenylmethylsulfonylfluoride (PMSF), and 2 mg/ seconds, followed by 72C for 4 minutes. mL leupepin], sonicated to generate whole-cell lysates, cleared by centrifugation, and protein was quantified by Electrophoretic mobility shift assay BCA protein assay kit (Pierce). Twenty microgram of whole- Nuclear extracts from cells maintained either as flat mono- cell lysate protein was size-fractioned on 7% SDS-PAGE gels layers or as clusters in suspension for 48 hours were prepared to detect BRCA1 or 12% SDS-PAGE gels to detect c-Jun, as previously described (6), with the exception that nuclear Fra2, A-Fos, and ERK1/2. Gels were blotted overnight onto proteins were not concentrated by (NH3)2SO4 precipitation, polyvinylidene difluoride (PVDF) membranes (Bio-RAD), but instead were dialyzed against 10 mmol/L HEPES pH which were then blocked with 5% skim milk and incubated 7.6, 40 mmol/L KCl, 0.1 mmol/L EDTA, and 10% glycerol. with the primary antibodies indicated followed by incuba- Nuclear extracts were then combined with 32P-labeled tion with the appropriate species-specific peroxidase-labeled BRCREO oligonucleotide plus strand 50-CTTTCCTTTT- second antibody. Antibody binding was then visualized by ACGTCATCCGGGGGCAGACT-30 minus strand 50-A- enhanced chemiluminescence (Amersham Life Science). GTCTGCCCCCGGATGACGTAAAAGG-30 or with After first establishing that their steady state levels were not CREOmut 50-CTTTCCTTTTAatTCATCCGGGGGC- altered by changing the culture conditions (see Fig. 1C), AGACT-30 minus strand 50- AGTCTGCCCCCGGAT- ERK1/2 served as loading controls in subsequent GAatTAAAAGG-30 (1 ng) in binding buffer [25 mmol/L experiments. HEPES pH 7.6, 5 mmol/L magnesium chloride (MgCl2), 34 mmol/L KCl, 50 mg/mL poly dI:dC (Sigma), 0.5 mg/mL Immunofluorescence bovine serum albumin (BSA)]. Binding reactions (20 mL Cell monolayers were transfected with His-tagged c-Jun final volume) were incubated on ice for 15 minutes and Fra2 expression vectors, maintained for 48 hours, fixed before separation on a 6% acrylamide 0.25 Tris-Borate- in cold methanol, and immunostained with antibodies EDTA buffer (TBE) nondenaturing gel. For the supershift against the His-tag of the transgenes. Nuclei were counter- assay, 2 mg anti-Fra2 (sc604), c-Jun (sc1694), or Sp1(sc-59) stained with 40,6-diamidino-2-phenylindole (DAPI, Sigma) [Santa Cruz] or PBS was incubated with 4 mg of nuclear and images were collected using an Olympus FV100 con- extracts for 30 minutes on ice before the addition of 32P- focal microscope. labeled oligonucleotide in binding buffer.

274 Mol Cancer Res; 11(3) March 2013 Molecular Cancer Research

BRCA1 Suppression in Ovarian Cancer Progression

Figure 1. BRCA1 levels are reduced in suspension culture. A, in monolayer culture, premalignant IOSE-80pc cells were ﬁbroblastic, as expected given their mesodermal derivation. In suspension culture, the cells ﬂoated as small, rounded cohesive clusters. B, quantitative RT-PCR indicated that steady-state BRCA1 mRNA levels were reduced in cells maintained suspension culture compared with monolayer culture. C, Western blotting indicated that steady-state BRCA1 protein levels were reduced in cells maintained suspension culture; ERK1/2 levels remained unchanged, and therefore were used as loading controls for all subsequent Western blotting in this study. D, cells maintained in either monolayer or suspension culture conditions ("Culture") were incubated in media previously conditioned by cells maintained either in suspension or in monolayers ("Media") and BRCA1 protein levels were assessed by Western blotting. Note that conditioned media did not appreciably alter BRCA1 levels of cells in either monolayer or suspension culture conditions. E, cells were maintained in suspension culture for 48 hours and one half of the cells were then lysed; the remaining cells were replated as monolayers for a further 48 hours and lysed. Both lysates were then subjected to Western blotting. Note that the suppression of BRCA1 was reversed when the cells were replated in monolayer culture.

Results considerably when the cells were maintained in suspension BRCA1 expression is suppressed in premalignant OSE culture (Fig. 1B and C). cells maintained in suspension culture To determine if soluble factors released by IOSE-80pc Human OSE cells, which are mesodermally derived, do cells contribute to the differences in BRCA1 expression not express E-cadherin. Thus, OSE cells tend to become described earlier, we cultured them in media precondi- fibroblastic when they are explanted into monolayer tioned by cells in either monolayer or suspension culture. culture (17). Cells of the immortal, premalignant Clearly, this preconditioned media had little or no effect IOSE-80pc line took on this fibroblastic appearance when on BRCA1 levels (Fig. 1D). Therefore, it is unlikely that they were maintained as monolayers on tissue culture soluble factors contributed to the suppression of BRCA1 plastic (Fig. 1A, Mono). In contrast, when IOSE-80pc expression that occurred in suspension culture. This cells were maintained in suspension they rounded-up and tentative conclusion was supported by the observation formed loosely adherent clusters of cells (Fig. 1A, Susp) that the reduction in BRCA1 expression due to suspension that are not dissimilar in appearance to the clusters that culture was reversed when suspension cells were replated ovarian carcinoma cells form after they have been shed as monolayers using the same suspension medium (Fig. into the adominopelvic cavity during the initial stages of 1E). The latter finding also indicates that the suppression transperitoneal spread (24). Interestingly, steady-state of BRCA1 expression initiated in suspension culture was BRCA1 mRNA and protein levels were both reduced reversible.

www.aacrjournals.org Mol Cancer Res; 11(3) March 2013 275

Zhou et al.

Figure 2. Transcriptional activation of the BRCA1 promoter is reduced in suspension culture. A, schematic representation and sequence of the BRCA1 promoter constructs. The L6 construct is indicated with a single line, which represents the full-length BRCA1 promoter located between start sites of the NBR2 and BRCA1 genes. The PRR region, located between 201 nt and 155 nt relative to the BRCA1 start site, contains 2 positive-acting elements, RIBS and CRE, which are indicated with open boxes. The "RIBSm" mutant fragment has the same sequence as PRR, but there are 2 previously characterized loss-of-function point mutations, as shown in the sequence of the upstream RIBS element. The "CREm" mutant fragment also has the same sequence as PRR, but there are loss-of-function point mutations in the sequence of the downstream CRE element (3). B, the promoter constructs shown were ligated into the pGL4.11(Luc2P) luciferase reporter vector, and transiently cotransfected into IOSE-80pc cells with the pGL4.73 (hRluc/SV40) vector as an internal transfection efﬁciency control. Cells were then maintained in either monolayer or suspension culture for 48 hours, lysed, and luciferase activity was assessed. In the left graph, relative L6 promoter activities were compared where the monolayer activity normalized as 100%. Alternatively, in the right graph, relative wild-type PRR, RIBSm, and CREm and promoterless "vector" activities were compared where the wtPRR monolayer activity was normalized as 100% (see Materials and Methods for transfection and assay details; n ¼ 3, SD; , P < 0.05; data shown are from one representative experiment of 3).

Transactivation of a PRR within the BRCA1 promoter is promoter, 204nt to þ27nt) contains a very strong PRR suppressed in suspension culture (i.e., the PRR fragment, 204nt to 155nt) that is max- As the steady state levels of BRCA1 mRNA as well as imally active in both ovarian carcinoma and immortalized protein were suppressed in IOSE-80pc cells maintained in OSE (IOSE) cells (3). Importantly, the ability of both the L6 suspension, we next assessed the effect of this culture promoter and the PRR fragment to transactivate and drive condition on the transactivation state of the BRCA1 reporter gene expression was reduced when IOSE-80 cells promoter. were maintained in suspension culture compared with when The BRCA1 promoter, which acts bidirectionally, is they were maintained in monolayer culture (Fig. 2B). While located on the long arm of chromosome 17 between the the suspension culture-mediated suppression of the entire NBR2 and BRCA1 genes (Fig. 2A). Previously, we deter- L6 promoter was a trend, the decrease in the PRR was larger mined that a 231 bp form of the promoter (i.e., the "L6" (i.e., approximately 3-fold) and statistically signiﬁcant.

276 Mol Cancer Res; 11(3) March 2013 Molecular Cancer Research

BRCA1 Suppression in Ovarian Cancer Progression

The PRR fragment contains 2 positively acting binding elements; the upstream RIBS element and the downstream A CRE element (Fig. 2A). When the upstream "RIBS" ele- Input -Ab H3 HA Alpha USF-2Fra-2 c-Jun ment was disrupted by mutation (i.e., "RIBSm" in Fig. 2A), the PRR activity fell signiﬁcantly in monolayer culture. B C However, the activity of this mutant, which had a wild-type Mono Susp Mono Susp CRE-element, was further reduced in the suspension culture GABP-α c-Jun

(Fig. 2B). In contrast, when the downstream CRE-binding GABP-β Fra2 element was mutated (i.e., "CREm" in Fig. 2A), PRR activity was also greatly decreased in monolayer culture, but ERK1/2 ERK1/2 there was no further decrease in suspension culture (Fig. 2B). These data suggest that factors binding the positive-acting D Monolayer Suspension wild-type CRE element may be reduced in suspension culture. – Fra2 c-Jun Sp1 CREOmut – Fra2 c-Jun Sp1 CREOmut The binding of c-Jun and Fra2 to the CRE element of the BRCA1 promoter is reduced in suspension culture Previously, we determined that the RIBS element helps to activate the PRR fragment of the BRCA1 promoter by binding GABPa and GABPb transcription factor subunits in breast and ovarian carcinoma cells in standard monolayer Figure 3. The expression of c-Jun and Fra2, which bind to the CRE culture conditions. In contrast, the CRE element contri- element of the BRCA1 promoter, are reduced in suspension culture. A, butes to PRR activation by binding c-Jun and Fra2 members ChIP indicated that GABPa (Alpha), which interacts with the RIBS element of the PRR, as well as c-Jun and Fra-2, which interact with the of the AP1 transcription factor complex in standard mono- CRE element of the PRR, all bind to the BRCA1 promoter under native layer conditions (3, 6, 7). As such, we carried out ChIP assays conditions in IOSE-80pc cell monolayers. Histone H3 (H3) and USF-2 and confirmed that GABPa ("Alpha"), c-Jun, and Fra2 all served as positive and negative binding controls, respectively. The bound PRR-containing chromatin in IOSE-80pc cell absence of antibody (Ab) and an irrelevant antibody against monolayers (Fig. 3A). We were unable to assess GABPb hemagglutanin (HA) served as technical controls. B, Western blotting of whole-cell lysates indicated that the steady state protein levels of GABPa binding in this assay given that, in our hands, the commer- and GABPb, both of which bind to the RIBS element, did not decrease in cially available antibody we used for Western blotting (see cells maintained suspension culture (Susp) compared with cells later) did not efficiently immunoprecipitate chromatin. maintained in monolayer culture (Mono). C, Western blotting of whole-cell We next assessed the steady state levels of GABPa, lysates indicated that the levels of c-Jun and Fra2, which bind to the CRE b element, both decreased when the cells were maintained in suspension. GABP , c-Jun, and Fra2 proteins in IOSE-80pc cells ERK1/2 served as a protein loading control. D, nuclear extracts from cells maintained in the 2 different culture conditions. There was maintained in monolayer or suspension culture were used in EMSAs no decrease in the level of either GABPa or GABPb in using either a radiolabeled "BRCREO" probe, which contains the entire suspension culture (Fig. 3B). This observation, and the CRE element with short flanking sequences, or a negative control CRE observation described earlier which indicated that mutating mutant (CREOmut), which contains the nonfunctional CREm point a b mutations shown in Fig. 2A. Assays of the wild-type probe were carried the GABP / -binding RIBS element (6) did not abrogate out in the absence () or the presence of antibodies against Fra2, and c- suspension-mediated suppression of the PRR, led us to Jun as well as the Sp1 transcription factors that does not bind the CRE conclude that changes in the levels of these factors are not element (Ref. 7). Note that, in monolayer culture, the antibody free-control likely critical regulators of this phenomenon. In contrast, the and the antibody against Sp1 both produced a weak, nonspecific doublet steady state levels of c-Jun and Fra2 were both reduced in (marked by the asterisks), whereas the Fra2 and c-Jun antibodies both fi fi generated a supershift of a single species (arrowheads). In suspension suspension culture (Fig. 3C). This nding, and the nding culture only a slight Fra2 supershift remained which suggests that c-Jun described earlier, which indicated that mutating the c-Jun/ and Fra2 binding were both decreased. Fra2–binding CRE element (7) abrogated suspension-mediated suppression of the PRR led us to tentatively conclude Forced expression of c-Jun and Fra2 activates the BRCA1 that the decrease in these factors may contribute to the promoter in suspension culture suppression of BRCA1. This notion was supported by the c-Jun and Fra2 are members of the AP1 family of tran- data generated by electrophoretic mobility shift assays scription factors, and AP1 activity can be broadly suppressed (EMSA) using the CRE as the probe. Specifically, super- by a dominant negative form of Fos that acts to block all AP1 shifting with antibodies that recognized c-Jun and Fra2 activity (A-Fos mutant; ref. 20). Forced expression of the A- indicated that complexes formed from nuclear extracts of Fos mutant in IOSE-80pc cells decreased the transcriptional cells maintained in monolayer culture and that these com- activation state of the entire L6 promoter, the PRR and the plexes were reduced when the cells were maintained in upstream RIBS mutant, but not the downstream CRE suspension (Fig. 3D). Given these findings, and the changes mutant (Supplementary Fig. S1). This finding led us to in promoter activity, we decided to next determine if more specifically investigate the effects of manipulating c- manipulating c-Jun/Fra2 levels affected the suspension cul- Jun and Fra2 expression on the PRR when the cells were ture-mediated suppression of BRCA1 promoter activity. maintained in either monolayer or suspension culture.

www.aacrjournals.org Mol Cancer Res; 11(3) March 2013 277

Zhou et al.

To test the effects of c-Jun and Fra2 specifically, we We next assessed the effect of forced c-Jun and Fra2 on force expressed both transcription factors together, by IOSE-80pc cell proliferation. In general, the cell spreading dual transfection. In standard monolayer culture condi- that occurs in monolayer culture facilitates proliferative tions, this greatly increased the activation of the L6 growth factor signaling, a phenomenon that is curtailed as promoter and the wild-type form of the PRR fragment cells round and are released into suspension (25). IOSE- (Fig. 4A). Forced expression of c-Jun and Fra2 also 80pc cells fit this paradigm. Specifically, IOSE-80pc increased the activation of the RIBS mutant, but not proliferation fell significantly when the cells were switched the CRE mutant, of the PRR (Fig. 4A). The latter result from being flat and spread in monolayer culture to was expected given the fact that the former, but not the rounded and clustered in suspension culture (Fig. 5). latter, still has an intact AP1-binding site for c-Jun and Importantly, while forced c-Jun/Fra2 expression increased Fra2 (3, 7). Importantly, however, the same pattern of proliferation modestly in both the monolayer and suspen- increased activation of the L6 promoter, the wtPRR sion culture conditions, the total percentage of proliferating region and the PRR/RIBS mutant, but not the PRR/ cells remained much lower in the suspension condition. CRE mutant, also occurred when the cells were main- Therefore, the ability of forced c-Jun/Fra2 expression to tained in suspension (Fig. 4B). This further supports the robustly stimulate BRCA1 promoter transactivation in sus- notion that the decrease in endogenous c-Jun and Fra2 pension culture was not likely caused by a general ability to levels that occurs in suspension culture may contribute to stimulate proliferation back to the levels observed in mono- BRCA1 suppression. layer culture.

Figure 4. Transient transfection of c-Jun and Fra2 increases BRCA1 promoter activity in monolayer and suspension culture. Cells were cotransfected with an empty pcDNA3 expression vector or pcDNA3 vectors containing His- tagged forms of c-Jun and Fra2, together with the promoter/ luciferase reporters described in Fig. 2. Cotransfected cells were maintained in monolayer (A) or suspension (B) culture for 48 hours, lysed, and luciferase activity was assessed. The data were then normalized using either L6 promoter activity (left) or wild-type PRR activity (right) in the pcDNA3 expression vector control condition as 100% (see Materials and Methods for details; n ¼ 3, SD; , P < 0.005; data shown are from one representative experiment of 3).

278 Mol Cancer Res; 11(3) March 2013 Molecular Cancer Research

BRCA1 Suppression in Ovarian Cancer Progression

forced c-Jun/Fra2 expression increased endogenous BRCA1 protein levels modestly in monolayer cultures and strongly in suspension cultures (Fig. 6A). To conﬁrm that changes in c-Jun and Fra2 levels alter BRCA1 expression in both monolayer and suspension culture, we also knocked both transcription factors down coordinately using siRNA-mediated cotransfection (Fig. 6B). In response to this coordinated knockdown, BRCA1 protein levels were greatly reduced when the cells were maintained in monolayer culture and there was a further reduction in the already suppressed BRCA1 level when the cells were maintained in suspension culture (Fig. 6B). Therefore, we concluded that the decrease in endogenous c-Jun and Fra2 levels that occurs when IOSE-80pc cells are

Figure 5. Transient cotransfection of c-Jun and Fra2 does not appreciably increase proliferation in monolayer and suspension culture. A, cells transfected with either an empty expression vector or cotransfected with His epitope-tagged versions of c-Jun and Fra2 were immunostained with antibodies against the His tag. Note that approximately 50% of the cells were positive for the epitope-tagged transgenes. B, cells cotransfected with either empty expression vector constructs or with expression constructs containing c-Jun and Fra2 were maintained in monolayer or suspension culture. Proliferative activity was then assessed by the ability of cells to take up fluorescent EdU (see Materials and Methods for details). Note that while maintenance in suspension greatly decreased proliferative rates, there was only a slight increase in the percentage of Figure 6. c-Jun and Fra2 regulate endogenous BRCA1 levels in both EdU-positive cells in either monolayer or suspension culture after monolayer and suspension culture. A, cells transfected with an either cotransfection with c-Jun/Fra2 compared with vector control empty expression vector or cotransfected with His epitope-tagged transfectants. versions of c-Jun and Fra2 were maintained for 48 hours in either monolayer or suspension culture, lysed, and subjected to Western c-Jun and Fra2 regulate endogenous BRCA1 levels in blotting with the indicated antibodies. Note that cotransfection with c- monolayer and suspension culture Jun/Fra2 raised levels of both transcription factors (as is clear from both the anti-His and anti-c-Jun and Fra2 blots), and it specifically increased When we force-expressed wild-type c-Jun and Fra2 in BRCA1 in both monolayer and suspension culture (i.e., ERK1/2 levels IOSE-80pc cells by cotransfection, the exogenous forms of were not affected). B, cells cotransfected with either scrambled (scr) or the transcription factors were identified by Western blotting sequence-specific siRNAs (si) against c-Jun and Fra2 and were for the epitope tag (Fig. 6A, anti-His-tag blots), and the maintained for 48 hours in either monolayer or suspension culture. The overall increase for each could be observed by blotting cells were then lysed, and subjected to Western blotting with the fi indicated antibodies. Note that cotransfection with siRNAs to c-Jun/Fra2 speci cally for either c-Jun or Fra2 in both monolayer and decreased levels of both transcription factors and specifically decreased suspension culture (Fig. 6A, anti-c-Jun and Fra2 blots). This BRCA1 in both the monolayer and suspension culture condition.

www.aacrjournals.org Mol Cancer Res; 11(3) March 2013 279

Zhou et al.

maintained in suspension culture contributes to the resultant including components of the AP1 transcription factor com- suppression of BRCA1. plex (32, 33). At this point, it is not clear if a loss of integrin signaling or a decrease in cytoskeletal tension, or both, Discussion contributed to the decrease in c-Jun and Fra2 when Ovarian carcinoma, which consists of a number of his- IOSE-80pc cells were maintained as rounded cell clusters tologically distinct subtypes, is unusual in that it often in suspension. However, this decrease also occurred when spreads transperitoneally (26). During this process, tumor the cells were clustered and maintained on compliant recon- cells are first shed from the primary lesion and they float stituted basement membrane gels where integrins were freely, often as small clusters, within the peritoneal fluid of engaged but the cells remained rounded (data not shown). the abdominopelvic cavity (27). It is these floating tumor cell One way to address this issue definitively would be to alter clusters that, if they subsequently adhere to peritoneal the rigidity of the substratum after the cells are already surfaces at sites distant from the primary lesion, can expand attached such that cytoskeleton tension is modulated, where- and become locally invasive. This process is particularly as the integrins are continuously engaged with the ECM. prominent in the high-grade serous tumors, which are shed These are experiments that we are currently attempting by early in their progression and go on to form the most coating silicone rubber ECM substrata that can then be prominent and lethal ovarian carcinoma subtype by far stretched to increase its rigidity after cells are preattached. It (28). High-grade serous ovarian carcinoma tumors are also will also be interesting to determine if the suspension- notable in that they often exhibit BRCA1 abnormalities, mediated decrease in c-Jun and Fra2 affect other aspects of only a small percentage of which are caused by genetic the ovarian carcinoma phenotype, although at this point, we mutation. Thus, epigenetic suppression is a major factor in have not noticed that there is any change in adhesion or the histogenesis of these tumors (2). invasiveness (e.g., on basement membrane gels). The tumor microenvironment is known to regulate The BRCA1 promoter constitutes a large CpG island that BRCA1 expression. This has been best studied in the breast is hypermethylated in up to 30% of spontaneous ovarian where the ECM, hormones, and hypoxia have all been epithelial carcinomas (34, 35). Importantly, this hyper- shown to alter BRCA1 levels by impinging on the gene's methylation is associated with decreased BRCA1 expression transcriptional and posttranslational regulation (13, and poor clinical outcome (36). One of the CpG targets for 14, 22, 29, 30). Given that release from the ECM is a methylation in the BRCA1 promoter is the CRE element major microenvironmental change that occurs as ovarian (37). In fact, the "CREm" mutation used in this study that carcinoma begins to disseminate by floating as effusive eliminated the differential monolayer- and suspension cul- tumor cell clusters in the fluid of the peritoneal cavity ture–mediated activation and prevented c-Jun/Fra2 bind- (27) and the fact that BRCA1 levels are reduced in dissem- ing, removes the major hypermethylation site within this inated high-grade serous tumor cells compared with their element. Thus, the decreased binding of c-Jun/Fra2 at this tissue of origin (31), we released premalignant ovarian cells site in suspension culture and, presumably, during shedding from the ECM, maintained them as clusters in suspension into the abdominopelvic fluid in vivo, could, potentially, be a culture, and determined the effect on BRCA1 levels. Under form of initial "flexible" transcriptional suppression that suspension conditions, BRCA1 levels were decreased con- subsequently helps to facilitate a "stable" suppression by siderably and this decrease was mediated, at least in part, methylation (38, 39). If this turns out to be the case, it would transcriptionally. More specifically, we showed that a sus- mean that the process of tumor cell shedding that is an early pension-mediated decrease in the AP1 transcription factor hallmark of the deadly transperitoneal spread of high-grade components c-Jun and Fra2 contributed to inhibition of serous tumors may help initiate the stable epigenetic sup- BRCA1 promoter transactivation and endogenous BRCA1 pression of BRCA1 that is prominent in the sporadic form of gene expression. At this point, we cannot state definitively these tumors. that GABPa and GABPb, 2 ETS-like transcription factors that also positively regulate the BRCA1 promoter are not Disclosure of Potential Conflicts of Interest involved in this suppression. However, it was clear that No potential conflicts of interest were disclosed. releasing the cells into suspension did not affect the steady state levels of either GABP factor, which is why we focused Authors' Contributions Conception and design: L. Zhou, C.R. Mueller, C.D. Roskelley on c-Jun and Fra2 for the remainder of this initial study. It is Development of methodology: L. Zhou, C.R. Mueller, C.D. Roskelley also important to note that other, as yet unidentified factors, Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L. Zhou, G. MacDonald, J. Cipollone, C.R. Mueller may be involved in the BRCA1 suppression given the large Analysis and interpretation of data (e.g., statistical analysis, biostatistics, compu- decrease in wild-type PRR activation we observed in sus- tational analysis): L. Zhou, M. Graves, C.R. Mueller, C.D. Roskelley pension culture that we could not fully recapitulate in either Writing, review, and/or revision of the manuscript: L. Zhou, C.R. Mueller of the RIBS or CRE element mutants. Study supervision: M. Graves, C.R. Mueller, C.D. Roskelley Cells in monolayer culture have engaged integrins and, The costs of publication of this article were defrayed in part by the payment of page because they are spread on a rigid tissue culture substratum, charges. This article must therefore be hereby marked advertisement in accordance with their cytoskeletal tension is high, both of which contribute to 18 U.S.C. Section 1734 solely to indicate this fact. the mechanical augmentation of growth factor-mediated Received June 29, 2012; revised January 3, 2013; accepted January 3, 2013; signaling that leads to the expression of various factors, published OnlineFirst January 21, 2013.

280 Mol Cancer Res; 11(3) March 2013 Molecular Cancer Research

BRCA1 Suppression in Ovarian Cancer Progression

References 1. Vergote I, Trope CG, Amant F, Kristensen GB, Ehlen T, Johnson N, 20. Udalova IA, Kwiatkowski D. Interation of AP-1 with a cluster of et al. Neoadjuvant chemotherapy or primary surgery in stage IIIC or IV NF-kappa B binding elements in the human TNF promoter region. ovarian cancer. J Engl J Med 2010;363:943–53. Biochem Biophys Res Commun 2001;289:25–33. 2. The Cancer Genome Atlas Research Network. Integrated genomic 21. Olive M, Krylov D, Echlin DR, Gardner K, Taparowsky E, Vinson C. A analyses of ovarian carcinoma. Nature 2011;474:609–614. dominant negative to activation protein-1 (AP1) that abolishes DNA 3. Graves ML, Zhou L, MacDonald G, Mueller CR, Roskelley CD. Reg- binding and inhibits oncogenesis. J Biol Chem 1997;272:18586–94. ulation of the BRCA1 promoter in ovarian surface epithelial cells and 22. MacDonald G, Stramwasser M, Mueller CR. Characterization of a ovarian carcinoma cells. FEBS Lett 2007;581:1825–33. negative transcriptional element in the BRCA1 promoter. Breast Can- 4. Suen TC, Goss PE. Transcription of BRCA1 is dependent on the cer Res 2007;9:R49. formation of a specific protein-DNA complex on the minimal BRCA1 23. Bindra RS, Gibson SL, Meng A, Westermark U, Jasin M, Pierce AJ, bi-directional promoter. J Biol Chem 1999;274:31297–304. et al. Hypoxia-induced down-regulation of BRCA1 expression by 5. Thakur S, Croce CM. Positive regulation of the BRCA1 promoter. J Biol E2Fs. Cancer Res 2005;65:11597–604. Chem 1999;274:8837–43. 24. Burleson KM, Casey RC, Skubitz KM, Pambuccian SE, Oegema TR, 6. Atlas E, Stramwasser M, Whiskin K, Mueller CR. GA-binding protein Skubitz AP. Ovarian carcinoma ascites spheroids adhere to extracel- alpha/beta is a critical regulator of the BRCA1 promoter. Oncogene lular matrix components and mesothelial cell monolayers. Gynecol 2000;19:1933–40. Oncol 2004;93:170–181. 7. Atlas E, Stramwasser M, Mueller CR. A CREB site in the BRCA1 25. Mammoto A, Ingber DE. Cytoskeletal control of growth and cell fate proximal promoter acts as a constitutive transcriptional element. switching. Curr Opin Cell Biol 2009;21:864–70. Oncogene 2001;20:7110–14. 26. Flam F, Einhorn N, Sjovall K. Symptomatology of ovarian cancer. Eur 8. Suen TC, Tang MS, Goss PE. Model of transcriptional regulation of the J Obstet Gynecol Reprod Biol 1988;27:53–7. BRCA1-NBR2 bi-directional transcriptional unit. Biochim Biohys Acta 27. Davidson B, Risberg B, Reich R, Berner A. Effusion cytology in ovarian 2005;1728:126–34. cancer: new molecular methods as aids to diagnosis and prognosis. 9. Auersperg N, Wong AS, Choi KC, Kang SK, Leung PC. Ovarian surface Clin Lab Med 2003;23:729–54. epithelium: biology, endocrinology, and pathology. Endocr Rev 2001; 28. Levanon K, Crum C, Drapkin R. New insights into the pathogenesis of 22:255–88. serous ovarian cancer and its clinical impact. J Clin Oncol 2008;26: 10. Kurman RJ, Shih IeM. The origin and pathogenesis of epithelial ovarian 5284–93. cancer: a proposed unifying theory. Am J Surg Pathol 2010;34:433–43. 29. Miralem T, Avraham HK. Extracellular matrix enhances heregulin- 11. Jarboe E, Folkins A, Nucci MR, Kindelberger D, Drapkin R, Miron A, dependent BRCA1 phosphorylation and suppresses BRCA1 expres- et al. Serous carcinogenesis in the fallopian tube: a descriptive clas- sion through its C terminus. Mol Cell Biol 2003;23:579–93. sification. Int J Gynecol Pathol 2008;27:1–9. 30. Antonova L, Mueller CR. Hydrocortisone down-regulates the tumor 12. Mingels MJ, Roelofsen T, van der Laak JA, de Hullu JA, van Ham MA, suppressor gene BRCA1 in mammary cells: a possible molecular link Massuger LF, et al. Tubal epithelial lesions in salpingo-oophorectomy between stress and breast cancer. Genes Chromosome Cancer specimens of BRCA-mutation carriers and controls. Gynecol Oncol 2008;47:341–52. 2012;127:88–93. 31. McMillen BD, Aponte MM, Liu Z, Helenowski IB, Scholtens DM, Buttin 13. Callahan MJ, Crum CP, Medeiros F, Kindelberger DW, Elvin JA, Garber BM, et al. Expression analysis of MIR182 and its associated target JE, et al. Primary fallopian tube malignancies in BRCA-positive women genes in advanced ovarian carcinoma. Mod Pathol 2012;25:1644–53. undergoing surgery for ovarian cancer risk reduction. J Clin Oncol 32. Orr AW, Helmke BP, Blackman BR, Schwartz MA. Mechanisms of 2007;25:3985–90. mechanotransduction Dev. Cell 2006;10:11–20. 14. O'Connell FC, Martin F. Laminin-rich extracellular matrix association 33. Provenzno PP, Keely PJ. Mechanical signaling through the cytoskel- with mammary epithelial cells suppresses Brca1 expression. Cell eton regulates cell proliferation by coordinated focal adhesion and Rho Death Differ 2000;7:360–67. GTPase signaling. J Cell Sci 2011;124:1195–1205. 15. Murtagh J, McArdle E, Gilligan E, Thornton L, Furlon F, Martin F. 34. Baldwin RL, Nemeth E, Tran H, Shvartsman H, Cass I, Narod S, et al. Organization of mammary epithelial cells into 3F acinar structures BRCA1 promoter region hypermethylation in ovarian carcinoma: a requires glucocorticoid and JNK signaling. J Cell Biol 2004;166: population-based study. Caner Res 2000;60:5329–33. 133–43. 35. Wilcox CB, Baysal BE, Gallion HH, Strange MA, Deloia JA. High- 16. Maines-Bandiera SL, Kruk PA, Auersperg N. Simian virus 40-trans- resolution methylation analysis of the BRCA1 promoter in ovarian formed human ovarian surface epithelial cells escape normal growth tumors. Cancer Genet Cytogenet 2005;159:114–22. controls but retain morphogenetic responses to extracellular matrix. 36. Chiang JW, Karlan BY, Cass L, Baldwin RL. BRCA1 promoter meth- Am J Obstet Gynecol 1992;167:729–35. ylation predicts adverse ovarian cancer prognosis. Gynecol Oncol 17. Wu C, Cipollone J, Maines-Bandiera S, Tan C, Karsan A, Auersperg N, 2006;101:403–10. et al. The morphogenic function of E-cadherin-mediated adherens 37. DiNardo DN, Butcher DT, Robinson DP, Archer TK, Rodenhiser DI. junctions in epithelial ovarian formation and progression. Differentia- Functional analysis of CpG methylation in the BRCA1 promoter region. tion 2008;76:193–205. Oncogene 2001;20:5331–40. 18. Roskelley CD, Desprez PY, Bissell MJ. Extracellular matrix-dependent 38. Baylin S, Jones PA. A decade of exploring the cancer epigenome— tissue-specific gene expression in mammary epithelial cells requires biological and translational implications. Nat Rev Cancer 2011;11: both physical and biochemical signal transduction. Proc Natl Acad Sci 726–34. U S A 1994;91:12378–82. 39. Zhang Y, Shu J, Si J, Shen L, Estecio MRH, Issa J-PJ. Repetitive 19. Thakur S, Nakamura T, Calin G, Russo A, Tamburrino JF, Shimizu M, elements and enforced transcriptional repression co-operate to et al. Regulation of BRCA1 transcription by specific single-stranded enhance DNA methylation spreading into a promoter CpG-island. DNA binding factors. Mol Cell Biol 2003;23:3774–87. Nuclei Acid Res 2012;40:7257–68.

www.aacrjournals.org Mol Cancer Res; 11(3) March 2013 281

Microenvironmental Regulation of BRCA1 Gene Expression by c-Jun and Fra2 in Premalignant Human Ovarian Surface Epithelial Cells

Lixin Zhou, Marcia Graves, Gwen MacDonald, et al.

Mol Cancer Res 2013;11:272-281. Published OnlineFirst January 21, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-12-0395

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2013/01/21/1541-7786.MCR-12-0395.DC1

Cited articles This article cites 39 articles, 12 of which you can access for free at: http://mcr.aacrjournals.org/content/11/3/272.full#ref-list-1

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/11/3/272. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.