RAD51C-Deficient Cancer Cells Are Highly Sensitive to the PARP Inhibitor Olaparib

Published OnlineFirst March 19, 2013; DOI: 10.1158/1535-7163.MCT-12-0950

Molecular Cancer Chemical Therapeutics Therapeutics

RAD51C-Deﬁcient Cancer Cells Are Highly Sensitive to the PARP Inhibitor Olaparib

Ahrum Min1, Seock-Ah Im1,2, Young-Kwang Yoon1, Sang-Hyun Song1, Hyun-Jin Nam1, Hyung-Seok Hur1, Hwang-Phill Kim1, Kyung-Hun Lee1,2, Sae-Won Han1,2, Do-Youn Oh1,2, Tae-You Kim1,2,4, Mark J. O'Connor5, Woo-Ho Kim1,3, and Yung-Jue Bang1,2

Abstract A PARP inhibitor is a rationally designed targeted therapy for cancers with impaired DNA repair abilities. RAD51C is a paralog of RAD51 that has an important role in the DNA damage response. We found that cell lines sensitive to a novel oral PARP inhibitor, olaparib, had low levels of RAD51C expression using microarray analysis, and we therefore hypothesized that low expression of RAD51C may hamper the DNA repair process, resulting in increased sensitivity to olaparib. Compared with the cells with normal RAD51C expression levels, RAD51C-deﬁcient cancer cells were more sensitive to olaparib, and a higher proportion

underwent cell death by inducing G2–M cell-cycle arrest and apoptosis. The restoration of RAD51C in a sensitive cell line caused attenuation of olaparib sensitivity. In contrast, silencingofRAD51Cinaresistant cell line enhanced the sensitivity to olaparib, and the number of RAD51 foci decreased with ablated RAD51C expression. We also found the expression of RAD51C was downregulated in cancer cells due to epigenetic changes and RAD51C expression was low in some gastric cancer tissues. Furthermore, olaparib significantly suppressed RAD51C-deficient tumor growth in a xenograft model. In summary, RAD51C- deficient cancer cells are highly sensitive to olaparib and offer preclinical proof-of-principle that RAD51C deficiency may be considered a biomarker for predicting the antitumor effects of olaparib. Mol Cancer Ther; 12(6); 865–77. 2013 AACR.

Introduction mutations (3–5). PARP inhibitors block the repair of DNA The DNA repair system is critical for maintaining single-strand breaks (SSB); unrepaired SSBs lead to the genomic integrity. Synthetic lethality is defined as the formation of DNA double-strand breaks (DSB). If DSBs loss of cell viability when multiple genes lose their func- cannot be repaired because of homologous recombination tions altogether, especially when compensatory genes are dysfunction, genomic instability or cell death can result also defective. The concept of synthetic lethality has been (6). Therefore, PARP inhibitors may be effective against shown using the novel PARP inhibitor in patients with various human cancer cells with defective DNA repair breast and ovarian cancer harboring mutations in the genes. An example of that is CDK1 depletion increased BRCA1 or BRCA2 genes (1, 2). There is clinical evidence PARP inhibitor sensitivity in cancer cells because the showing that olaparib (AZD2281; KU-0059436), a small- DSBs induced by PARP inhibition could not be repaired molecule inhibitor of PARP, has potential as a therapeutic because of inactivation of homologous recombination– agent alone and in combination with radiotherapy or associated repair in CDK1-depleted cells (7). In addition, chemotherapy to treat cancers with BRCA1 and BRCA2 recent study suggests that nonhomologous end joining (NHEJ) plays an important role in the genomic instability and hypersensitivity to PARP inhibitors in homologous Authors' Affiliations: 1Cancer Research Institute, Seoul National Univer- recombination–deficient cells (8). 2 3 sity; Departments of Internal Medicine and Pathology, Seoul National RAD51C is a RAD51-like gene that has a key role in University College of Medicine; 4Department of Molecular Medicine and Biopharmaceutical Sciences, Graduate School of Convergence Science maintaining genomic stability (9–12). The functional role and Technology, Seoul, Republic of Korea; and 5AstraZeneca UK Ltd., of RAD51C in DNA damage repair has also been exam- fi Maccles eld, Cheshire, United Kingdom ined (13). The results of these studies suggest that Note: Supplementary data for this article are available at Molecular Cancer RAD51C acts sequentially with RAD51 at the DNA Therapeutics Online (http://mct.aacrjournals.org/). damage site to repair DNA damage. Therefore, RAD51C Corresponding Authors: Seock-Ah Im and Yung-Jue Bang, Department depletion leads to impaired RAD51 foci formation, of Internal Medicine, Seoul National University College of Medicine, 101 Daehak-ro, Jongno-gu, Seoul, 110-744, Republic of Korea. Phone: 82-2- resulting in impaired DNA repair (13, 14). In addition, 2072-0850; Fax: 82-2-762-9662; E-mail: Seock-Ah Im, [email protected]; some studies have shown that RAD51C is required for Yung-Jue Bang, [email protected] the checkpoint response to DNA damage (13, 15). Fur- doi: 10.1158/1535-7163.MCT-12-0950 thermore, recent studies have found that germline muta- 2013 American Association for Cancer Research. tions of RAD51C are associated with cancers. In these

www.aacrjournals.org 865

Min et al.

RAD51C cancers, germline mutations of impede RAD51 (Bid-Rad); the cell survival rate and IC50 of olaparib were foci formation results in blocking the homologous determined using SigmaPlot software. recombination–mediated DNA repair. RAD51C-defective cancers can therefore potentially be treated with cDNA microarray analysis olaparib because the DNA damage induced by olaparib The SNU-601 and SNU-668 cells were treated for dif- cannot be effectively repaired by homologous recombi- ferent periods of time (24, 48, and 72 hours) with 1 mmol/L nation, as RAD51C deficiency interferes with RAD51- olaparib or dimethyl sulfoxide (DMSO). The total RNA mediated homologous recombination. was extracted and hybridized to an Affymetrix GeneChip In the present study, we evaluated the antitumor activ- human Gene 1.0 ST array (Affymetrix, Inc.). The results ity of olaparib in cancer cell lines in vitro and observed that were normalized to the robust multi-array average (RMA) olaparib-sensitive cell lines had low levels of RAD51C and analyzed by DNALINK, Inc.. The microarray data expression using microarray analysis. Subsequently, we were deposited in the ArrayExpress database (accession evaluated whether the silencing of RAD51C-sensitize number: E-MTAB-1012). olaparib sensitivity and restoration of RAD51C cause decreased sensitivity to olaparib. We also characterized BRCA1 and BRCA2 mutation analysis the mechanisms of RAD51C silencing in human cancers. Genomic DNA (gDNA) was extracted from gastric RAD51C expression using immunohistochemistry was cancer cell lines using an Accuprep Genomic DNA Extrac- evaluated in gastric cancer tissue with paired normal tion Kit (Bioneer) according to the manufacturer’s proto- gastric mucosa. This is the first report to show that col. The BRCA1 and BRCA2 mutations were analyzed RAD51C-deficient cancer cells are selectively sensitive to using fluorescent-conformation sensitive gel electropho- PARP inhibitor olaparib and olaparib promotes cell death resis (F-CSGE), as described previously (17). by inducing G2–M cell-cycle arrest and apoptosis. Reverse transcription PCR and real-time PCR Materials and Methods Total RNA was isolated using TRI reagent (Molecular Reagents Research Center, Cincinnati, OH). cDNA was synthesized Olaparib was kindly provided by AstraZeneca Ltd.. The by conducting reverse transcription PCR (RT-PCR) with chemical structure of olaparib is shown in Fig. 1A. 5-Aza- ImProm-II reverse transcriptase (Promega) and amplified 20-deoxycytidine (5-aza-dc) was purchased from Sigma- using AmpliTaq Golf DNA polymerase (Applied Biosys- Aldrich. tems) with gene-specific primers. Quantitative real-time PCR was conducted using an iCycler iQ detection system Cell lines and cell culture (Bio-Rad Laboratories, Inc.) with SYBR Green. All data Human gastric cancer cells (SNU-1, -5, -16, -216, -484, from all samples were normalized to the actin cDNA -601, -620, -638, -668, -719, and KATO-III) were purchased levels. The sequences of the primers used for the RT-PCR from the Korean Cell Line Bank; the identities of the cell and qRT-PCR are listed in Supplementary Table S1. cDNA lines were authenticated by DNA fingerprinting analysis was synthesized at least 3 times from 3 independent sets of (16). Human breast cancer cells (BT-549 and MCF-7) samples, and all PCR analyses were conducted in authenticated using short tandem repeat analysis were triplicate. purchased from the American Type Culture Collection. All cell lines were stored in liquid nitrogen, passaged for Western blot analysis less than 6 months before use in this study, and cultured in Proteins were extracted and equal amount of proteins RPMI-1640 supplemented with 10% FBS and 10 mg/mL were separated on 8% to 15% SDS-PAGE as previously gentamicin at 37 C in a 5% CO2 atmosphere. described (18). The resolved proteins were transferred onto nitrocellulose membranes, the blots were probed Cell growth inhibition assay overnight at 4C with appropriate primary antibodies Cells were seeded in 96-well plates and exposed to [RAD51C and XRCC3 (Novus Biologicals), RAD51 and increasing concentrations of olaparib (doses ranged from RAD51D (Santa Cruz Biotechnology), caspase-3, cyclin 0–10 mmol/L) for 5 days. After drug treatment, the absor- B1, p21, phosphorylated p53, cdc2, and PTEN (Cell bance of MTT dye was measured at 540 nm with a Signaling Technology), phosphorylated histone H2AX VersaMax microplate reader (Molecular Devices). The (Millipore), RAD51B (Abcam), PARP (BD Biosciences), absorbance and IC50 of olaparib were analyzed using and a-tubulin (Sigma-Aldrich)]. Antibody binding was SigmaPlot software [Statistical Package for the Social detected using an enhanced chemiluminescence system Sciences, Inc. (SPSS)]. according to the manufacturer’s protocol (Amersham For the colony formation assay (CFA), the cells were Biosciences). seeded in 6-well plates and treated with various concentrations (5, 1, 0.1, 0.01, and 0.001 mmol/L) of olaparib for 5 Cell-cycle analysis days and were cultured until colonies formed (14 days). The cells treated with olaparib and/or radiation were The cell colonies were stained with 0.1% Coomassie blue harvested, fixed with cold 70% ethanol, and then stored solution (Sigma-Aldrich) and counted using Gel-Doc at 20C for at least 24 hours. The cells were washed in

866 Mol Cancer Ther; 12(6) June 2013 Molecular Cancer Therapeutics

Olaparib Sensitivity in RAD51C-Deﬁcient Cancer Cells

A Olaparib O

N N Figure 1. RAD51C expression O affects olaparib sensitivity in human cancer cell lines. A, the chemical N structure of olaparib. B, RAD51C N F O silencing restored olaparib sensitivity in the insensitive cell line SNU-668. The SNU-668 cells were B transfected with nonspecific control or RAD51C-specific siRNA. The cells 100 were exposed to increasing doses of olaparib and a CFA was conducted. 80 After 14 days, the colonies were SNU-668 stained and counted. Cell survival siControl siRAD51C was calculated as a percentage and 60 presented as a graph with SE bars RAD51C (left; n ¼ 3). The IC50 was also 40 determined (SNU-601, 0.037 α-Tubulin 0.001 mmol/L; SNU-668, 3.80 fi SNU-601 0.670 mmol/L; nonspeci c control- 20 SNU-668 SNU-668, 3.70 0.620 mmol/L; and survival (%) Percentage siControl / SNU-668 siRAD51C-SNU-668, 0.061 0.007 siRAD51C / SNU-668 mmol/L). C, SNU-601 cells were 0 0 0.001 0.01 0.1 15 transfected with pcDNA3 or μ pcDNA3-RAD51C plasmids to Olaparib concentration ( mol/L) establish the stable clones and the CFA was conducted. After 14 days of C olaparib treatment, cell survival was calculated and is presented in the 100 graph (left). The IC50 was also calculated (pcDNA3 #1, 0.098 SNU-601 0.002 mmol/L; pcDNA3 #2, 0.065 80 pcDNA3 pcDNA3-RAD51C 0.002 mmol/L; RAD51C #1, 3.250 #1 #2 #1 #2 0.80 mmol/L; and RAD51C #2, 3.860 60 0.370 mmol/L). All data presented RAD51C represent the results of 3 independent experiments. The α-Tubulin 40 transfection efficacy was verified by Western blot analysis with anti- SNU-668 Percentage survival (%) Percentage RAD51C and anti-a-tubulin 20 pcDNA3 #1 / SNU-601 antibodies (right column of B and C). pcDNA3 #2 / SNU-601 pcDNA3-RAD51C #1 / SNU-601 pcDNA3-RAD51C #2 / SNU-601 0 0 0.001 0.01 0.1 15 Olaparib concentration (μmol/L)

PBS and incubated with 10 mg/mL RNase A (Sigma- PBS-T (0.5% Triton X-100 in PBS) for 5 minutes, and Aldrich) at 37C for 20 minutes. Next, the cells were incubated with primary antibody for 24 hours at 4C. stained with 20 mg/mL propidium iodide (Sigma- The primary antibodies used in this study were rabbit Aldrich), and the DNA content of the cells (10,000 cells polyclonal anti-RAD51 (H-92; Santa Cruz Biotechnolo- per experimental group) was quantified using a FACS gy) and mouse monoclonal anti-phosphorylated histone Calibur flow cytometer (BD Biosciences). H2AX (clone JBW301; Millipore) at a dilution of 1:100. Thecoverslipswererinsed3timesfor10minutesin Immunofluorescence assay (RAD51 foci formation) PBS, followed by incubation with the appropriate fluor- SNU-668 cells were plated on 0.01% poly-L-lysine ophore-conjugated secondary antibody (Invitrogen). (Sigma-Aldrich)–coated coverslips, transfected with The cells were counterstained with 40,6-diamidino-2- RAD51C-specific or nonspecific control siRNA, and trea- phenylindole (DAPI; 300 nmol/L; Invitrogen) and the ted with 1 mmol/L olaparib. After 2 days, the cells were coverslips were mounted on slides using Faramount exposed to 10 Gy of radiation for 2 hours. Afterward, the aqueous mounting medium (DAKO). Immunofluores- coverslips were rinsed once in PBS (37C), fixed in 3.7% cence was visualized using a Zeiss LSM 510 laser scan- paraformaldehyde for 10 minutes, permeabilized with ning microscope.

www.aacrjournals.org Mol Cancer Ther; 12(6) June 2013 867

Min et al.

Comet assays stored in liquid nitrogen and then used for RNA and SNU-601, SNU-668, and RAD51C-specific or nonspe- protein extraction, and the other portion was embedded cific control siRNA-transfected SNU-668 cells were trea- in paraffin. The RAD51C expression in the normal and ted with 1 mmol/L olaparib or 10 Gy of radiation. After cancer tissues was analyzed via immunohistochemical treatment, cells were trypsinized and subjected to an (IHC) staining using the anti-mouse monoclonal anti- alkaline comet assay using the Trevigen Comet Assay Kit body against RAD51C (Novus Biologicals) at a dilution (Trevigen) following the manufacturer’s protocol. Tail of 1:50. The quantification of the IHC slides was con- lengths were measured with Comet assay IV program. ducted in a blinded fashion, and 3 staining patterns were distinguished relative to the number of positive Plasmid and siRNA transfection cells and the staining intensity: the scores are negative The pcDNA3-RAD51C expression plasmid (14) was (negative; no staining to weak staining in 10% tumor kindly provided by Dr. Masson (Laval University Cancer cells), 1þ (weak staining in >10% of tumor cells), 2þ Research Center, Que´bec, Canada). siRNAs specific for (moderate staining in >10% of tumor cells), and 3þ RAD51C and PTEN were obtained from Qiagen. SNU-668 (strong staining in >10% of tumor cells). cells were transfected with siRNA at a final concentration of 80 nmol/L or 4 mg of RAD51C plasmid using Lipofec- In vivo study tamine 2000 (Invitrogen) according to the manufacturer’s The animal experiments were carried out at the ani- instructions. After 48 hours, the cells were harvested and mal facility of Seoul National University (Seoul, Repub- subjected to Western blot analysis. The sequence of the lic of Korea) according to the institutional guidelines RAD51C-specific siRNA was 50-CACCTTCTGTTCAG- with prior approval from the Institutional Animal Care CACTAGA-30, the sequence of the PTEN-specific siRNA and Use Committee (IRB number: SNU-100816-2). A was 50-AAGGCGTATACAGGAACAATA-30, the seque- total of 16 female Balb/c athymic nude mice of ages nce of the XRCC3-specific siRNA was 50-CACAGAAT- 4- to 6-week were purchased from Central Lab Animal TATTGCTGCAATTAA-30, and the sequence of the Ezh2- Inc.. The mice were injected subcutaneously in the right specific siRNA was 50-AAGACTCTGAATGCAGTTGCT- flank with 1 108 of SNU-601 cells in 100 mLofPBS.After 30. The sequence of the control (nonspecific) siRNA was 50- implantation of the tumor cells, the sizes of the resulting AATTCTCCGAACGTGTCACG-30. tumors were measured every other day using calipers; the body weight of each mouse was also determined Stable overexpression of RAD51C twice per week. The tumor volume was calculated using The pcDNA3 and pcDNA3-RAD51C plasmids (14) the following formula: (width2 height)/2. When the were used to transfect SNU-601 cells with Lipofectamine tumor volume reached 150 to 200 mm3, the mice were 2000 (Invitrogen) according to the manufacturer’s instruc- randomly divided into 2 groups (8 mice per group). tions. The cells were treated with 300 mg/mL G418 (Cell- One group of mice was treated daily with 50 mg/kg gro) to select cells that had stably integrated the plasmid. olaparib for 28 consecutive days via oral gavages. The After selection in G418, RAD51C expression was mea- control group was treated with a 10% 2-hydroxyl-pro- sured in the clones by Western blot analysis. pyl-b-cyclodextrine/PBS solution alone. The mice were euthanized with CO2 when the tumor volume reached Bisulfite modification and genomic sequencing 1,500 mm3. The tumors were excised and stored in liquid gDNA was isolated using the QIAamp DNA Mini Kit nitrogen until further analysis. (Qiagen) according to the manufacturer’s instructions and dissolved in H2O to a final volume of 20 mL. Bisulfite Immunohistochemistry gDNA modification was conducted using an EpiTect The histologic sections from individual paraffin- Bisulfite kit (Qiagen) according to the manufacturer’s embedded xenograft tumor tissues were deparaffinized protocol. The modified gDNA was amplified with and dehydrated. IHC detection of proliferating cells was RAD51C-specific primers. These primers were designed conducted using the anti-rabbit polyclonal antibody to amplify the CpG island of RAD51C. The PCR products against Ki-67 (GeneTex) at a dilution of 1:100. Terminal were loaded onto a 1.5% agarose gel (Lonza) and sepa- deoxynucleotidyl transferase–mediated dUTP nick end rated at 100 V for 10 minutes in 1 tris-borate EDTA (TBE). labeling (TUNEL) assay was conducted for IHC detection DNA was extracted from the gel and sequenced. The DNA of apoptosis using ApopTag In situ Apoptosis Detection sequencing was conducted by Bionics, Inc.. Kit (Chemicon International), in accordance with manufacturer’s protocol. RAD51C expression in cancer tissue Thirteen gastric tumor tissue and matched adjacent Statistical analysis normal tissue samples were obtained from Seoul Statistical analyses were conducted using SigmaPlot National University Hospital [Seoul, Republic of Korea; version 9.0. A two-sided Student t test was used when Institutional Review Board (IRB) number: H-1004-078- appropriate. The results are expressed as the mean SD 316]. Each sample was collected immediately after or SE. A P value less than 0.05 was considered to be surgery and divided into 2 portions. One portion was statistically significant.

868 Mol Cancer Ther; 12(6) June 2013 Molecular Cancer Therapeutics

Olaparib Sensitivity in RAD51C-Deﬁcient Cancer Cells

Results olaparib (IC50, 0.44 0.05 mmol/L by MTT). Exogenous Olaparib selectively targets RAD51C-deficient RAD51C overexpression was confirmed by Western blot human cancer cells analysis (Fig. 1C and Supplementary Fig. S5B). BT-549 We examined olaparib sensitivity in gastric and breast cells overexpressing RAD51C were less sensitive to ola- cancer cells using CFA. Several cell lines (SNU-601, parib compared with those transfected with the empty KATO-III, and BT-549) were highly sensitive to olaparib vector (Supplementary Fig. S5A). These findings suggest compared with other cell lines (Supplementary Fig. S1A). that RAD51C deficiency plays a role in the growth inhib- We chose SNU-601 and BT-549 as the sensitive cell lines itory activity of olaparib. and SNU-668 and MCF-7 as the resistant cell lines for further study. The SNU-601 and BT-549 cells were highly Olaparib sensitivity in RAD51C-deficient cells is sensitive to olaparib compared with the SNU-668 and associated with G2–M cell-cycle arrest and apoptosis MCF-7 cells (Supplementary Fig. S1B). To identify the RAD51C deficiency leads to G2–M cell-cycle arrest and predictive markers of olaparib sensitivity, we conducted a subsequent genomic instability (9, 12). We therefore inves- microarray analysis of the sensitive SNU-601 and resistant tigated whether olaparib promotes G2–M cell-cycle arrest SNU-668 cells. We found that RAD51C expression was and/or apoptosis in RAD51C-deficient cells by conduct- much lower in the SNU-601 cells. RAD51C protein expres- ing a fluorescence-activated cell sorting (FACS) analysis. sion was also low in olaparib-sensitive SNU-601 and BT- SNU-601 and SNU-668 cells were exposed to increasing 549 cells (Supplementary Fig. S2A). Next, the mRNA doses of olaparib for 4 days, the DNA content was mea- BRCA1/2 RAD51 MRE11 expression of , , and was measured with propidium iodide staining. The G2–M and sub- sured. No significant differences were observed between G1 populations were significantly increased in a dose- the sensitive and resistant cell lines (Supplementary Fig. dependent manner in the RAD51C-deficient SNU-601 S2B). In addition, mutational analysis of BRCA1 and cells (Fig. 2A). In the SNU-601 cells, PARP cleavage was BRCA2 indicated that olaparib sensitivity was not asso- detected in a dose-dependent manner (Fig. 2A). The G2–M ciated with BRCA mutations, at least in these cell lines population also increased in the BT-549 cells (Supplemen- (Supplementary Table S2). Because the BT-549 cells were tary Fig. S6). In addition, the SNU-668 cells were subjected characterized by a PTEN deficiency, we determined to siRNA transfection to reduce RAD51C expression and whether PTEN overexpression affected olaparib sensitiv- treated with increasing concentrations of olaparib for 96 ity in the BT-549 cells. Our data showed that PTEN hours. We then analyzed cell-cycle progression using expression did not affect olaparib sensitivity in the BT- FACS analysis. The siRNA-mediated knockdown of 549 cells (Supplementary Fig. S3). We therefore hypoth- RAD51C increased the olaparib-induced apoptosis and esized that RAD51C deficiency may be linked with the G2–M cell-cycle arrest compared with the control siRNA synthetic lethality associated with olaparib; thus, a lack transfection (Fig. 2B). The olaparib-induced apoptosis in of RAD51C expression could be a marker of olaparib RAD51C-depleted SNU-668 cells was also confirmed by sensitivity. detecting the PARP cleavage (Fig. 2B). Conversely, SNU- 601 cells that stably overexpressed RAD51C failed to Olaparib sensitivity is mediated by the absence of undergo cell-cycle arrest and apoptosis when treated with RAD51C expression olaparib. Populations in the G2–M and sub-G1 phases did To determine whether olaparib sensitivity was a direct not increase with olaparib treatment in the SNU-601 cells result of RAD51C deficiency, we measured the viability of overexpressing RAD51C (Fig. 2C). In addition, PARP SNU-668 cells treated with olaparib. These cells were cleavage associated with apoptosis did not change (Fig. transfected with control or RAD51C-specific siRNA. Cell 2C). Our results indicate that olaparib-induced G2–M viability was measured in the presence of olaparib with a arrest and apoptosis are a result of RAD51C deficiency. CFA. RAD51C knockdown enhanced olaparib sensitivity to a level comparable with that of olaparib-sensitive SNU- Olaparib sensitizes cancer cells to ionizing radiation 601 cells (Fig. 1B and Supplementary Fig. S4A). The Olaparib has also been reported to enhance the effects of successful knockdown of RAD51C expression was vali- conventional cytotoxic agents in homologous recombina- dated by Western blot analysis (Fig. 1B). RAD51C-defi- tion–defective cells (19–21). The suppression of homolo- cient SNU-601 cells were transfected with an empty gous recombinational repair due to RAD51C deficiency is pcDNA3 vector or pcDNA3-RAD51C plasmid to over- associated with disturbed RAD51 foci formation at irra- express RAD51C, after which stable clones were selected diation-induced DSB sites (14). Thus, olaparib may using G418. CFA was conducted with these stable clones increase the cytotoxic effects of irradiation in RAD51C- to measure the growth inhibitory activity of olaparib after deficient cells, especially during the G2–M phase, due to RAD51C overexpression. Two SNU-601 clones that stably the presence of unrepaired DSBs. To address whether overexpressed RAD51C had attenuated olaparib sensitiv- olaparib administered in combination with radiation can ity compared with SNU-601 clones transfected with emp- increase the percentage of RAD51C-depleted cells in the ty pcDNA3 vector (Fig. 1C and Supplementary Fig. S4B). G2–M phase, we conducted FACS analysis on SNU-601, To further validate our findings, we transiently overex- SNU-668, and RAD51C-depleted SNU-668 cells treated pressed RAD51C in BT-549 cells, which are sensitive to with olaparib, radiation, or both. The numbers of SNU-601

www.aacrjournals.org Mol Cancer Ther; 12(6) June 2013 869

Min et al.

A SNU-601 SNU-668 30 30 DMSO DMSO c μ 25 0.1 μmol/L of olaparib 25 0.1 mol/L of olaparib μ 1 μmol/L of olaparib 1 mol/L of olaparib μ 20 5 μmol/L of olaparib 20 5 mol/L of olaparib b 15 15

10 a 10

% Cell-cycle phases 5 % Cell-cycle phases 5

0 0 Sub-G1 G2–M Sub-G1 G2–M

Sub-G1 phase (%) 4.52 5.81 6.24 9.18 Sub-G1 phase (%) 2.37 2.35 2.47 2.94

Figure 2. Olaparib induces G2–M cell-cycle G2–M phase (%) 10.39 13.57 16.50 24.02 G2–M phase (%) 10.54 11.57 11.11 13.21 arrest and apoptosis in RAD51C-deficient SNU-601 SNU-668 cell lines. The cells were treated with the indicated concentrations of olaparib for 96 C 0.1 1 5 C 0.1 1 5 hours, and the percentage of cells in the RAD51C G2–M phase and those undergoing PARP apoptosis are determined using FACS analysis. The columns represent the mean of α-Tubulin 3 independent experiments and are shown with error bars (SE). A, the proportion of the B siControl / SNU-668 siRAD51C / SNU-668 SNU-601 and SNU-668 cells undergoing 30 30 G –M phase and apoptosis are presented in DMSO DMSO e 2 25 0.1 μmol/L of olaparib 25 0.1 μmol/L of olaparib the bar graphs (top). a, P ¼ 0.038; 1 μmol/L of olaparib 1 μmol/L of olaparib d b, P ¼ 0.032; and c, P ¼ 0.005. The total 5 μmol/L of olaparib 20 5 μmol/L of olaparib 20 c cellular proteins were extracted and Western 15 15 blotting was conducted with the indicated antibodies. PARP cleavage was observed in 10 10 the olaparib-sensitive cells in a dose- b 5 5 a dependent manner (bottom). B, SNU-668 % Cell-cycle phases % Cell-cycle phases fi 0 0 cells transfected with nonspeci c control or Sub-G1 G2–M Sub-G1 G2–M RAD51C-specific siRNA were harvested and

the percentages of cells in the G2–M phase Sub-G phase (%) 1.90 2.35 2.50 3.48 Sub-G phase (%) 1.96 3.13 3.40 4.23 1 1 and undergoing apoptosis were calculated G –M phase (%) 8.26 8.43 9.39 10.42 and are shown in the bar graph (top). 2 G2–M phase (%) 11.46 15.67 20.25 24.85 a, P ¼ 0.013; b, P ¼ 0.0001; c, P < 0.0001; siControl siRAD51C d, P < 0.0001; and e, P < 0.0001. Western blot analysis showed that PARP cleavage was C 0.1 1 5 C 0.1 1 5 observed after olaparib treatment in the RAD51C RAD51C-depleted SNU-668 cells in a dose- dependent manner (bottom). C, the pcDNA3 PARP or pcDNA3-RAD51C plasmids were stably α-Tubulin transfected to the SNU-601 cells and the cell-cycle distribution was analyzed. The – C pcDNA3 / SNU-601 pcDNA3-RAD51C / SNU-601 percentages of cells in the G2 M phase and 30 30 those undergoing apoptosis are presented in DMSO DMSO μ P ¼ 25 0.1 mol/L of olaparib 25 0.1 μmol/L of olaparib the bar graphs (top). a, 0.0028; μ 1 mol/L of olaparib 1 μmol/L of olaparib b, P < 0.0001; c, P < 0.0001; and 5 μmol/L of olaparib μ 20 d 20 5 mol/L of olaparib d, P < 0.0001. Western blot analysis of 15 c 15 SNU-601 cells overexpressing RAD51C revealed that RAD51C expression can lead 10 10 b a to evasion of apoptosis (bottom). 5 5 % Cell-cycle phases % Cell-cycle phases 0 0 Sub-G1 G2–M Sub-G1 G2–M

3.25 4.93 5.48 7.03 Sub-G1 phase (%) Sub-G1 phase (%) 3.93 4.31 4.1 3.71 11.27 12.745 13.42 17.7 G2–M phase (%) G2–M phase (%) 10.89 11.89 10.72 11.45

pcDNA3 pcDNA3-RAD51C

C 0.1 1 5 C 0.1 1 5 RAD51C

PARP

α-Tubulin

870 Mol Cancer Ther; 12(6) June 2013 Molecular Cancer Therapeutics

Olaparib Sensitivity in RAD51C-Deﬁcient Cancer Cells

and SNU-668 cells transfected with RAD51C-specific ry Fig. S8B). Similar to the Western blotting results, the siRNA in the G2–M and sub-G1 phases were increased IHC analysis showed that 5 of 13 tumors (38%) were IHC compared with untreated SNU-668 cells and SNU-668 1þ and that 2 tumors (15.3%) were IHC 2þ, in contrast to cells transfected with nonspecific control siRNA (Fig. the normal tissues, which were 3þ according to the IHC 3A). Olaparib significantly enhanced the cell-cycle arrest analysis. We next explored DNA methylation as a poten- and apoptosis induced by irradiation, indicating that tial mechanism underlying the loss of RAD51C expression there is synergy between olaparib and radiation. Increas- in cancer cells. To determine whether the absence of ed PARP cleavage was clearly detected in RAD51C-defi- RAD51C is due to DNA methylation in the SNU-601 cells, cient cells following radiation exposure combined with we analyzed the induction of RAD51C expression in SNU- 1 mmol/L olaparib (Fig. 3B). Overall, our data indicate that 601 and SNU-668 cells treated with 5-aza-dc, an inhibitor olaparib sensitized the RAD51C-deficient cancer cells to of DNA methylation. The 5-aza-dc increased RAD51C radiation. expression in the SNU-601 cells more than 160-fold (Sup- plementary Fig. S9A). We also observed that the RAD51C RAD51C silencing impairs the efficiency of gene was densely methylated in the SNU-601 and KATO- homologous recombinational DSB repair III cells using bisulfite-modified sequencing (Fig. 5A). To We hypothesize that olaparib sensitivity results from further investigate whether the low expression levels of reduced homologous recombinational repair efficiency of RAD51C in gastric tumor tissues are caused by DNA olaparib-induced DSBs in cells lacking RAD51C. We con- methylation, we conducted DNA sequencing with bisul- ducted an immunofluorescence study to examine RAD51 fite modification. Interestingly, almost every tumor tissue foci formation, indicative of DNA repair activity, in sample with reduced RAD51C expression showed a level RAD51C-deficient SNU-668 cells exposed to ionizing of RAD51C methylation ranging from 19% to 56% (Sup- radiation. RAD51C protein expression was depleted by plementary Fig. S9B). These data suggest that the loss of siRNA (Fig. 4A). Control and RAD51C siRNA-transfected RAD51C mediated by DNA methylation may frequently SNU-668 cells were treated with 10 Gy of radiation and occur in cancer. stained with antibodies against RAD51 (a marker of Unlike the gastric cancer cell lines, SNU-601 and KATO- homologous recombinational repair) and phosphorylated III, RAD51C expression was absent in the BT-549 cells H2AX (a marker of DNA damage). The number of RAD51 regardless of the DNA methylation status (Fig. 5B). Thus, foci in the RAD51C knockdown cells was significantly we speculated that histone modification might also be reduced at the sites of DNA damage even when the degree responsible for controlling RAD51C downregulation. of damage was comparable with that of the controls (Fig. Ezh2 is a component of the polycomb group protein that 4B and C). In addition, accumulation of DNA damage in is essential for the transcriptional regulation of a number individual cells was detected using a comet assay. SNU- of genes involved in the DNA repair process. Recently, 601 and RAD51C-depleted SNU-668 cells showed accu- increased Ezh2 expression was found to be correlated with mulation of DNA DSBs when treated with olaparib and 10 the epigenetic repression of DNA repair genes, including Gy of radiation (Supplementary Fig. S7A and S7B). The RAD51 (22–24). We therefore evaluated the possible effect knockdown of RAD51C expression was confirmed by of Ezh2 on RAD51C expression. First, we conducted real- Western blot analysis (Supplementary Fig. S7C). These time PCR to quantify the basal expression levels of data indicate that RAD51C downregulation leads to the RAD51C and Ezh2 to determine whether RAD51C expres- accumulation of DNA damage due to homologous recom- sion was inversely correlated with Ezh2 expression in BT- bination inactivation. Our results support a role for 549 and MCF-7 breast cancer cells. Cells overexpressing RAD51C in the repair of DSBs in human cancer cells. Ezh2 had downregulated RAD51C expression (Fig. 5C). Ezh2 depletion in BT-549 cells resulted in markedly RAD51C is epigenetically repressed in human cancer increased RAD51C expression (Fig. 5D). Thus, the deple- Our previous experiments showed that there was an tion of RAD51C mediated by epigenetic silencing resulted inverse relationship between RAD51C protein expression in the attenuated formation of RAD51 foci and increased and olaparib sensitivity in cancer cell lines. Next, we asked sensitivity to DNA damage in cancer cells. whether we can find the tumors with RAD51C deficiency using human cancer tissues for future use in clinic as a Olaparib impedes the growth of RAD51C-defective predictive marker and what mechanism might underlie cells in an in vivo mouse model RAD51C silencing in human cancers. RAD51C protein Olaparib showed significant antitumor activity in a expression was significantly downregulated in 4 of 11 SNU-601 gastric cancer xenograft model. Olaparib signif- primary gastric tumor tissues (36%) compared with the icantly delayed tumor growth not only during treatment corresponding normal tissues (Supplementary Fig. S8A). but also after treatment was ceased (Fig. 6A). Tumor For future clinical application, an IHC analysis was con- tissues from mice treated with olaparib showed lower ducted to examine the RAD51C protein expression in Ki-67 expression, which suggests lower proliferation abil- paraffin-embedded tissues. We assigned scores between ity compared with the tumor tissue from mice treated with negative and 3þ according to the intensity and extent of vehicle control and it was associated with increased apo- RAD51C protein expression in the tissues (Supplementa- ptosis by TUNEL assay (Fig. 6B). There were no signs of

www.aacrjournals.org Mol Cancer Ther; 12(6) June 2013 871

Min et al.

A SNU-601 SNU-668 30 30 DMSO b DMSO 1 μmol/L of olaparib 1 μmol/L of olaparib 25 25 10 Gy of irratiation (IR) 10 Gy of irratiation (IR) 1 μmol/L of olaparib + 10 Gy of IR 1 μmol/L of olaparib + 10 Gy of IR 20 20 d a 15 15

10 10 % Cell-cycle phases % Cell-cycle phases c 5 5

0 0 Sub-G1 G2–M Sub-G1 G2–M

Sub-G1 phase (%) 2.89 11.79 5.17 16.20 Sub-G1 phase (%) 1.14 1.25 1.55 2.72

G2–M phase (%) 7.34 19.37 15.06 28.07 G2–M phase (%) 6.63 8.03 12.06 16.11

siControl / SNU-668 siRAD51C / SNU-668 30 40 h DMSO DMSO μ 25 1 μmol/L of olaparib 1 mol/L of olaparib 10 Gy of irratiation (IR) 10 Gy of irratiation (IR) 30 μ 1 μmol/L of olaparib + 10 Gy of IR 1 mol/L of olaparib + 10 Gy of IR 20 f

15 20

10 g

% Cell-cycle phases 10 e 5 % Cell-cycle phases

0 0 Sub-G1 G2–M Sub-G1 G2–M

Sub-G1 phase (%) 1.48 1.71 2.03 2.65 Sub-G1 phase (%) 2.16 3.55 3.87 6.93

G2–M phase (%) 6.71 9.62 9.87 15.24 G2–M phase (%) 12.64 21.88 25.51 39.88

B SNU-601 SNU-668 siControl (668) siRAD51C (668) Olaparib (1 μmol/L) IR (10 Gy)

RAD51C

PARP

Caspase-3

Cyclin B1

cdc2

P–p53

p21

α-Tubulin

Figure 3. Olaparib sensitizes RAD51C-deficient cancer cells to radiation-induced G2–M arrest and apoptosis. A, the wild-type SNU-668 and SNU-601 cells along with SNU-668 cells transfected with nonspecific control or RAD51C-specific siRNA were treated with or without 1 mmol/L olaparib. The cells were either irradiated (10 Gy) or not 48 hours after the treatment. The cells were allowed to recover and grow for another 48 hours regardless of treatment with olaparib.

Next, the percentages of cells in the G2–M phase and those undergoing apoptosis were measured and are presented in the bar graph. Column, the mean of 3 independent experiments; bars, SE; a, P ¼ 0.0003; b, P < 0.0001; c, P ¼ 0.014; d, P ¼ 0.0066; e, P ¼ 0.009; f, P ¼ 0.0002; g, P ¼ 0.0006; and h, P < 0.0001. B, SNU-601, SNU-668, and nonspeciﬁc or RAD51C-speciﬁc siRNA-transfected SNU-668 cells were exposed to 1 mmol/L olaparib, irradiation, or left untreated. The cells were treated with olaparib for 48 hours, followed by irradiation (10 Gy) for 48 hours. The total cellular proteins were subjected to Western blot analysis with the indicated antibodies.

872 Mol Cancer Ther; 12(6) June 2013 Molecular Cancer Therapeutics

Olaparib Sensitivity in RAD51C-Deﬁcient Cancer Cells

A SNU-668 Control siControl siRAD51C Radiation – + + RAD51C

RAD51

Figure 4. RAD51C silencing impairs γ-H2AX RAD51 foci formation during the DNA damage response. A, the α-Tubulin siRNA-mediated reduction of RAD51C expression was confirmed by Western blot analysis. B, B SNU-668 SNU-668 cells transfected with nonspecific control or RAD51C- Control siControl + 10 Gy of Radiation siRAD51C + 10 Gy of Radiation specific siRNA were treated with DAPI RAD51 DAPI RAD51 DAPI RAD51 1 mmol/L olaparib. The cells were irradiated (10 Gy) 48 hours after transfection and allowed to recover for 2 hours before the immunofluorescence analysis with the indicated antibodies. Confocal microscopy was used to observe the γ-H2AX Merge γ-H2AX Merge γ-H2AX Merge signals corresponding to RAD51 (green) and g-H2AX (red). The DNA was counterstained with DAPI (blue). The merged pictures represent overlays of the 3 channels. The scale bars indicate 5 mm (top). C, the percentage of cells containing more C a than 5 foci of RAD51 and g-H2AX 80 over 3 experiments is presented in a RAD51 γ-H2AX bar graph. At least 100 nuclei were analyzed for each experiment 60 (bottom). Columns, the mean of 3 independent experiments; bars, P SE; a, < 0.0001. 40

20 % of cells with ≥ 5 foci

0 Control siControl + IR siRAD51C + IR SNU-668 toxicity in mice undergoing extended treatment (Fig. 6C). of several proteins, including the deficiency of RAD51, This study showed the antitumor effect of olaparib in a involved in homologous recombination on the sensitivity RAD51C-deficient gastric cancer xenograft model. to PARP inhibition was also reported (1), but the role of RAD51C deficiency on the sensitivity to the PARP inhib- Discussion itor in cancer cells has not yet been reported. Importantly, The use of a PARP inhibitor is a promising new strategy models of sensitivity to PARP inhibition suggest that the for treating cancers using the concept of synthetic lethal- key ingredient is a deficiency in homologous recombina- ity. There is preclinical and clinical evidence showing that tion, indicating that this approach may be more widely olaparib (AZD2281; KU-0059436), a small-molecule inhib- applicable in the treatment of sporadic cancers sharing itor of PARP, has potential as a therapeutic agent to treat homologous recombination impairments. In the present cancers with BRCA1 and BRCA2 mutations that have an study, we first reported that RAD51C expression was homologous recombination deficiency (1–5, 25, 26). In low in olaparib-sensitive cancer cells. We also found addition to the BRCA mutations, a recently published that olaparib was able to enhance the cytotoxic effects phase II trial showed that 24% of high-grade ovarian of irradiation-induced DNA damage in RAD51C-defi- carcinoma or triple-negative breast cancers without a cient cells. This effect subsequently led to the accumula- BRCA mutation also respond to olaparib (4, 5). The effects tion of unrepaired DSBs due to dysfunctional RAD51 foci

www.aacrjournals.org Mol Cancer Ther; 12(6) June 2013 873

Min et al.

A SNU-601 KATO-III Exon 1 Exon 1

Clone 1 Clone 1 Clone 2 Clone 2 Clone 3 Clone 3

SNU-668 MKN-45 Exon 1 Exon 1

Clone 1 Clone 1 Clone 2 Clone 2 Clone 3 Clone 3

Methylated Methylated Unmethylated Unmethylated

BT-549 Exon 1 0.07 RAD51C Ezh2 0.06

Clone 1 0.05

Clone 2 0.04

0.03 Methylated

mRNA expression 0.02 Unmethylated 0.01

0.00 BT-549 MCF-7 D 2.5 siControl siEzh2

2.0

1.5

1.0 mRNA expression 0.5

0.0 Ezh2 RAD51C BT-549

Figure 5. RAD51C deficiency is caused by epigenetic silencing or the overexpression of a transcriptional repressor. A, methylation levels of the RAD51C gene in the SNU-601 and SNU-668 cells was analyzed by genomic sequencing of bisulfite-modified DNA (left). The methylation levels in the KATO-III and MKN-45 cell lines were determined using bisulfite-modified sequencing (right). B, hypomethylation of the RAD51C gene in the BT-549 cells was determined using bisulfite-modified sequencing. C, the basal expressions of Ezh2 and RAD51C in the BT-549 and MCF-7 cells were analyzed by quantitative real-time PCR. The results were normalized to actin expression and are presented in the bar graph. Columns, the mean of 3 independent experiments; bars, SE. D, Ezh2 negatively regulates RAD51C expression. The siRNA-mediated depletion of Ezh2 induced RAD51C mRNA expression. The expression levels of RAD51C and Ezh2 were determined by quantitative real-time PCR in BT-549 cells transfected with nonspecific control or Ezh2-specific siRNA.

874 Mol Cancer Ther; 12(6) June 2013 Molecular Cancer Therapeutics

Olaparib Sensitivity in RAD51C-Deﬁcient Cancer Cells

A 1,800 )

3 Vehicle 1,600 Olaparib 50 mg/kg 1,400

1,200 Finished 1,000 treatment Started 800 treatment 600 Figure 6. Olaparib signiﬁcantly inhibits tumor growth in a xenograft 400 200 model of SNU-601 human gastric (mm Mean tumor volume cancer. A, Balb/c nude mice were 0 8 injected with 1 10 SNU-601 cells 10 13 15 17 20 22 24 27 30 32 34 36 39 41 43 46 49 51 53 56 59 61 63 66 69 76 83 and the drug treatment was initiated Day after transplantation 13 days after inoculation of the cancer cells. The mice were treated B with 50 mg/kg olaparib (n ¼ 8) or vehicle alone (n ¼ 8) daily for 28 days. H&E Ki-67 TUNEL The tumor volumes of each mouse were measured every other day and Vehicle are presented as a graph with SD bars. Olaparib inhibited tumor growth in a SNU-601 mouse xenograft model (, P < 0.001). B, the tumors were removed from the mice H&E Ki-67 TUNEL 42 days after the drug treatment ended, and pathologic examination was done using hematoxylin and Olaparib 50 mg/kg eosin (H&E) slides (400). The scale bars represent 25 mm. IHC stain for Ki-67 and TUNEL assay showed decreased Ki-67 with increased apoptosis in olaparib-treated tumors. C, changes of mouse body C 25 weight. The body weight of each Vehicle mouse was measured twice weekly. 24 Bars, SD. No signiﬁcant difference Olaparib 50 mg/kg 23 of body weight was detected. 22 21 20 Mouse weight (g) Mouse weight 19 18 10 15 17 22 24 30 32 39 41 46 48 53 55 61 63 69 76 83 Day after transplantation

formation along with increased cell death and G2–M in RAD51C downregulated cancer cells, resulting in arrest. Moreover, olaparib had significant in vivo antitu- hypersensitivity to PARP inhibition. Our findings suggest mor effects in a RAD51 normal and BRCA wild-type that RAD51C plays an essential role in DNA damage RAD51C-deficient SNU-601 xenograft model. Overall, response as a homologous recombination component, and our findings indicate that sensitivity to olaparib relies on that can promote synthetic lethality between PARP inhi- RAD51C inactivation. bition and RAD51C loss in cancer. RAD51C maintains genome integrity by participating Some previous reports that have described the inter- in branch migration and Holliday junction resolution action between RAD51C and XRCC3, and the role of (12, 13, 27). Germline mutations in RAD51C confer RAD51C in the resolution of homologous recombina- an increased risk of breast and ovarian cancer (28–30). tion intermediates have also been evaluated (15, 31). In According to the previous studies, the role of RAD51C in addition, some studies have suggested that RAD51C homologous recombination activation via the regulation expression can affect XRCC3 expression. We also of RAD51 recruitment was understood. Here, we first observed that the levels of XRCC3 expression were determined that inactivation of homologous recombina- proportional to the expression levels of RAD51C in the tional repair–mediated by RAD51 increases DNA damage olaparib-sensitive cancer cell lines. In the present study,

www.aacrjournals.org Mol Cancer Ther; 12(6) June 2013 875

Min et al.

we examined the potential physical interaction between with adjacent normal tissues via DNA methylation. RAD51C and XRCC3 with an IP assay and found that Using a novel synthetic lethal approach with a PARP depletion of RAD51C expression reduced XRCC3 inhibitor, we determined that RAD51C participated in expression (data not shown). Although we found that the DNA repair pathway. Our findings have potential RAD51C expression can affect XRCC3 protein levels, clinical implications for treating cancers with RAD51C the mechanism through which XRCC3 expression may deficiencies. Furthermore, RAD51C may serve as a be regulated by RAD51C is not clearly understood. novel biomarker for identifying tumors that are sensi- Elucidating the mechanism responsible for the regula- tive to olaparib, thereby allowing physicians to select tion of XRCC3 expression by RAD51C and the effect of patient populations who would receive the maximal this regulation on the homologous recombination pro- benefit from olaparib treatment. cess would be very interesting and worthy of further study. Disclosure of Potential Conflicts of Interest The downregulation of genes associated with the epi- S.-A. Im has commercial research grant from AstraZeneca through fellowship program and is a consultant/advisory board member of Astra- genetic mechanism involved in the DNA repair pathway Zeneca. W.-H. Kim has commercial research grant from AstraZeneca and promotes tumorigenesis (24, 32–35). We observed receives travel expense from AstraZeneca. Y.-J. Bang has commercial RAD51C downregulation in gastric and breast cancer cell research grant, honoraria from speakers bureau, and is a consultant/ advisory board member of AstraZeneca. No potential conflicts of interest lines and gastric tumor tissue samples. Interestingly, were disclosed by the other authors. RAD51C expression was significantly decreased in cancer compared with the adjacent normal tissue, and the lack of Authors' Contributions RAD51C was attributed to DNA methylation and histone Conception and design: A. Min, S.-A. Im, Y.-K. Yoon, D.-Y. Oh, M.J. modification. Gene silencing mediated by DNA methyl- O’Connor, Y.-J. Bang Development of methodology: A. Min, S.-A. Im, Y.-K. Yoon, H.-S. Hur, ation has been well characterized (36, 37). Similar to H.-P. Kim, D.-Y. Oh BRCA1/2 mutations, BRCA1/2 hypermethylation is asso- Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Min, S.-A. Im, D.-Y. Oh, T.-Y. Kim, W.-H. Kim ciated with the sensitivity to treatment with PARP inhi- Analysis and interpretation of data (e.g., statistical analysis, biostatis- bitors (32–35). Thus, the epigenetic inactivation of DNA tics, computational analysis): A. Min, S.-A. Im, H.-P. Kim, K.-H. Lee, S.-W. repair genes by CpG island hypermethylation could be a Han, D.-Y. Oh, T.-Y. Kim, Y.-J. Bang Writing, review, and/or revision of the manuscript: A. Min, S.-A. Im, K.- marker that may be used to improve the response to H. Lee, S.-W. Han, D.-Y. Oh, T.-Y. Kim, M.J. O’Connor, W.-H. Kim, Y.-J. cancer treatment and provide more personalized thera- Bang pies with PARP inhibitors. We found that RAD51C was Administrative, technical, or material support (i.e., reporting or orga- nizing data, constructing databases): A. Min, S.-A. Im, S.-H. Song, H.-J. densely methylated in cancer cell lines and tumor tissue Nam, D.-Y. Oh samples, resulting in RAD51C silencing. Moreover, our Study supervision: S.-A. Im, S.-W. Han, D.-Y. Oh, Y.-J. Bang findings showed that the loss of RAD51C expression due to epigenetic silencing leads to olaparib sensitivity in Acknowledgments The authors thank Drs. Masson for providing the pcDNA3-RAD51C cancer cells. plasmid, Alan Barge, Jee-Woong Son, Vasanti Natarajan, Charlotte Conventional therapies induce drug resistance due to Knights (former Astrazeneca Inc. employees), and Alan Lau (AstraZeneca decreased drug accumulation, increased drug degrada- Inc.) for supporting this study. The authors also thank Debora Kim at Seoul International School for English editing assistance. tion, and consolidation of DNA repair pathways. PARP inhibition could overcome these problems by impairing the DNA repair pathway. In our study, we found that the Grant Support This study was supported by the Basic Science Research Program combined use of radiation and olaparib enhanced irradi- through the National Research Foundation of Korea (NRF) funded by the ation-induced cytotoxicity by inducing cell-cycle arrest at Ministry of Education, Science and Technology (2010-0022299; S.-A. Im); the G –M phase. Thus, olaparib can be potentially used as and supported by research grants from the Oncology Research collabo- 2 ration of AstraZeneca Korean Cancer Study Group (06-2007-301-0; Y.-J. a therapeutic agent alone and in combination with radi- Bang). ation or with other DNA damaging agents. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked In summary, this report is the first to show that advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate epigenetically silenced RAD51C-deficient cancer cells this fact. are highly sensitive to the PARP inhibitor olaparib in vitro in vivo and . Moreover, RAD51C expression was Received September 27, 2012; revised March 12, 2013; accepted March downregulated in some gastric cancer tissues compared 12, 2013; published OnlineFirst March 19, 2013.

References 1. McCabe N. Deficiency in the repair of DNA damage by homologous 3. Fong PC, Boss DS, Yap TA, Tutt A, Wu P, Mergui-Roelvink M, et al. recombination and sensitivity to poly(ADP-ribose) polymerase inhibi- Inhibition of poly(ADP-ribose) polymerase in tumors from BRCA mutation. Cancer Res 2006;66:8109–15. tion carriers. N Engl J Med 2009;361:123–34. 2. Bryant HE, Schultz N, Thomas HD, Parker KM, Flower D, Lopez E, et al. 4. Tutt A, Robson M, Garber JE, Domchek SM, Audeh MW, Weitzel Specific killing of BRCA2-deficient tumours with inhibitors of poly JN, et al. Oral poly(ADP-ribose) polymerase inhibitor olaparib (ADP-ribose) polymerase. Nature 2005;434:913–7. in patients with BRCA1 or BRCA2 mutations and advanced

876 Mol Cancer Ther; 12(6) June 2013 Molecular Cancer Therapeutics

Olaparib Sensitivity in RAD51C-Deﬁcient Cancer Cells

breast cancer: a proof-of-concept trial. Lancet 2010;376: 21. Hay T, Matthews JR, Pietzka L, Lau A, Cranston A, Nygren AOH, et al. 235–44. Poly(ADP-ribose) polymerase-1 inhibitor treatment regresses autoch- 5. Audeh MW, Carmichael J, Penson RT, Friedlander M, Powell B, Bell- thonous Brca2/p53-mutant mammary tumors in vivo and delays tumor McGuinn KM, et al. Oral poly(ADP-ribose) polymerase inhibitor ola- relapse in combination with carboplatin. Cancer Res 2009;69:3850–5. parib in patients with BRCA1 or BRCA2 mutations and recurrent 22. Zeidler M, Varambally S, Cao Q, Chinnaiyan AM, Ferguson DO, ovarian cancer: a proof-of-concept trial. Lancet 2010;376:245–51. Merajver SD, et al. The polycomb group protein EZH2 impairs DNA 6. Plummer ER, Calvert H. Targeting poly(ADP-ribose) polymerase: a repair in breast epithelial cells. Neoplasia 2005;7:1011–9. two-armed strategy for cancer therapy. Clin Cancer Res 2007;13: 23. Zeidler M, Kleer CG. The polycomb group protein enhancer of Zeste 2: 6252–6. its links to DNA repair and breast cancer. J Mol Histol 2006;37:219–23. 7. Johnson N, Li Y-C, Walton ZE, Cheng KA, Li D, Rodig SJ, et al. 24. Stefansson OA, Esteller M. EZH2-mediated epigenetic repression of Compromised CDK1 activity sensitizes BRCA-proficient cancers to DNA repair in promoting breast tumor initiating cells. Breast Cancer PARP inhibition. Nat Med 2011;17:875–82. Res 2011;13:309. 8. Patel AG, Sarkaria JN, Kaufmann SH. Nonhomologous end joining 25. Farmer H, McCabe N, Lord CJ, Tutt AN, Johnson DA, Richardson TB, drives poly(ADP-ribose) polymerase (PARP) inhibitor lethality in et al. Targeting the DNA repair defect in BRCA mutant cells as a homologous recombination-deficient cells. Proc Natl Acad Sci U S A therapeutic strategy. Nature 2005;434:917–21. 2011;108:3406–11. 26. Gelmon KA, Tischkowitz M, Mackay H, Swenerton K, Robidoux A, 9. Smeenk G, de Groot AJ, Romeijn RJ, van Buul PP, Zdzienicka MZ, Tonkin K, et al. Olaparib in patients with recurrent high-grade serous or Mullenders LH, et al. Rad51C is essential for embryonic development poorly differentiated ovarian carcinoma or triple-negative breast can- and haploinsufficiency causes increased DNA damage sensitivity and cer: a phase 2, multicentre, open-label, non-randomised study. Lancet genomic instability. Mutat Res 2010;689:50–8. Oncol 2011;12:852–61. 10. Kuznetsov SG, Haines DC, Martin BK, Sharan SK. Loss of Rad51c 27. Abe K, Osakabe K, Nakayama S, Endo M, Tagiri A, Todoriki S, et al. leads to embryonic lethality and modulation of Trp53-dependent Arabidopsis RAD51C gene is important for homologous recombina- tumorigenesis in mice. Cancer Res 2009;69:863–72. tion in meiosis and mitosis. Plant Physiol 2005;139:896–908. 11. Kuznetsov S, Pellegrini M, Shuda K, Fernandez-Capetillo O, Liu Y, 28. Vaz F, Hanenberg H, Schuster B, Barker K, Wiek C, Erven V, et al. Martin BK, et al. RAD51C deficiency in mice results in early prophase I Mutation of the RAD51C gene in a Fanconi anemia–like disorder. Nat arrest in males and sister chromatid separation at metaphase II in Genet 2010;42:406–9. females. J Cell Biol 2007;176:581–92. 29. Meindl A, Hellebrand H, Wiek C, Erven V, Wappenschmidt B, Nieder- 12. French CA. Role of mammalian RAD51L2 (RAD51C) in recombination acher D, et al. Germline mutations in breast and ovarian cancer and genetic stability. J Biol Chem 2002;277:19322–30. pedigrees establish RAD51C as a human cancer susceptibility gene. 13. Badie S, Liao C, Thanasoula M, Barber P, Hill MA, Tarsounas M. Nat Genet 2010;42:410–4. RAD51C facilitates checkpoint signaling by promoting CHK2 phos- 30. Loveday C, Turnbull C, Ramsay E, Hughes D, Ruark E, Frankum JR, phorylation. J Cell Biol 2009;185:587–600. et al. Germline mutations in RAD51D confer susceptibility to ovarian 14. Rodrigue A, Lafrance M, Gauthier MC, McDonald D, Hendzel M, West cancer. Nat Genet 2011;43:879–82. SC, et al. Interplay between human DNA repair proteins at a unique 31. Liu Y, Tarsounas M, O'Regan P, West SC. Role of RAD51C and XRCC3 double-strand break in vivo. Embo J 2006;25:222–31. in genetic recombination and DNA repair. J Biol Chem 2007;282: 15. Lio YC, Schild D, Brenneman MA, Redpath JL, Chen DJ. Human 1973–9. Rad51C deficiency destabilizes XRCC3, impairs recombination, and 32. Veeck J, Ropero S, Setien F, Gonzalez-Suarez E, Osorio A, Benitez J,

radiosensitizes S/G2-phase cells. J Biol Chem 2004;279:42313–20. et al. BRCA1 CpG island hypermethylation predicts sensitivity to poly 16. Ku JL, Park JG. Biology of SNU cell lines. Cancer Res Treat 2005; (adenosine diphosphate)-ribose polymerase inhibitors. J Clin Oncol 37:1–19. 2010;28:e563–4. 17. Ahn SH, Son BH, Yoon K-S, Noh D-Y, Han W, Kim S-W, et al. BRCA1 33. Drew Y, Mulligan EA, Vong WT, Thomas HD, Kahn S, Kyle S, et al. and BRCA2 germline mutations in Korean breast cancer patients at Therapeutic potential of poly(ADP-ribose) polymerase inhibitor high risk of carrying mutations. Cancer Lett 2007;245:90–5. AG014699 in human cancers with mutated or methylated BRCA1 or 18. Nam H-J, Kim H-P, Yoon Y-K, Song S-H, Min A-R, Han S-W, et al. The BRCA2. J Natl Cancer Inst 2011;103:334–46. irreversible pan-HER inhibitor PF00299804 alone or combined with 34. Horton JK, Stefanick DF, Zeng JY, Carrozza MJ, Wilson SH. Require- gemcitabine has an antitumor effect in biliary tract cancer cell lines. ment for NBS1 in the S phase checkpoint response to DNA methylation Invest New Drugs 2012;30:2148–60. combined with PARP inhibition. DNA Repair (Amst) 2011;10:225–34. 19. Rottenberg S, Jaspers JE, Kersbergen A, van der Burg E, Nygren 35. Ibragimova I, Cairns P. Assays for hypermethylation of the BRCA1 AOH, Zander SAL, et al. High sensitivity of BRCA1-deﬁcient mam- gene promoter in tumor cells to predict sensitivity to PARP-inhibitor mary tumors to the PARP inhibitor AZD2281 alone and in combi- therapy. Methods Mol Biol 2011;780:277–91. nation with platinum drugs. Proc Natl Acad Sci U S A 2008;105: 36. Fullgrabe€ J, Kavanagh E, Joseph B. Histone onco-modiﬁcations. 17079–84. Oncogene 2011;30:3391–403. 20. Ratnam K, Low JA. Current development of clinical inhibitors of 37. Bae YK, Shim YR, Choi JH, Kim MJ, Gabrielson E, Lee SJ, et al. Gene poly(ADP-ribose) polymerase in oncology. Clin Cancer Res 2007;13: promoter hypermethylation in tumors and plasma of breast cancer 1383–8. patients. Cancer Res Treat 2005;37:233–40.

www.aacrjournals.org Mol Cancer Ther; 12(6) June 2013 877

RAD51C-Deficient Cancer Cells Are Highly Sensitive to the PARP Inhibitor Olaparib

Ahrum Min, Seock-Ah Im, Young-Kwang Yoon, et al.

Mol Cancer Ther 2013;12:865-877. Published OnlineFirst March 19, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-12-0950

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2013/03/21/1535-7163.MCT-12-0950.DC1

Cited articles This article cites 37 articles, 14 of which you can access for free at: http://mct.aacrjournals.org/content/12/6/865.full#ref-list-1

Citing articles This article has been cited by 13 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/12/6/865.full#related-urls

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://mct.aacrjournals.org/content/12/6/865. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.