BRCA1 Functions As a Differential Modulator of Chemotherapy-Induced Apoptosis1

[CANCER RESEARCH 63, 6221–6228, October 1, 2003] BRCA1 Functions as a Differential Modulator of Chemotherapy-induced Apoptosis1

Jennifer E. Quinn,2 Richard D. Kennedy,2 Paul B. Mullan, Paula M. Gilmore, Michael Carty, Patrick G. Johnston, and D. Paul Harkin3 Department of Oncology, Cancer Research Centre, Queen’s University Belfast, Belfast, Northern Ireland BT9 7AB, United Kingdom

ABSTRACT It has also been suggested that BRCA1 functions as a sensor of DNA damage relaying signals to either the cell cycle checkpoint or We have evaluated the role played by BRCA1 in mediating the pheno- cell death machinery. A number of studies have correlated BRCA1 typic response to a range of chemotherapeutic agents commonly used in deficiency with defects in cell cycle checkpoint control. Human tumor cancer treatment. Here we provide evidence that BRCA1 functions as a differential mediator of chemotherapy-induced apoptosis. Specifically, we cells lacking functional BRCA1 demonstrate a high frequency of demonstrate that BRCA1 mediates sensitivity to apoptosis induced by chromosome aneuploidy, characteristic of a defective G2/M check- antimicrotubule agents but conversely induces resistance to DNA-damag- point (9). In addition it has been demonstrated that BRCA1 is required ing agents. These data are supported by a variety of experimental models for both S phase and G2 arrest after ionizing irradiation, an effect that including cells with inducible expression of BRCA1, siRNA-mediated is dependent on differential phosphorylation by ATM (10, 11). Fur- inactivation of endogenous BRCA1, and reconstitution of BRCA1-defi- thermore, genetic instability has been observed in BRCA1 exon 11 cient cells with wild-type BRCA1. Most notably we demonstrate that isoform-deficient MEFs resulting from a defective G2/M checkpoint BRCA1 induces a 10–1000-fold increase in resistance to a range of DNA- and centrosome amplification (12). We have reported previously that damaging agents, in particular those that give rise to double-strand breaks inducible expression of BRCA1 can activate both the G and mitotic such as etoposide or bleomycin. In contrast, BRCA1 induces a >1000-fold 2 increase in sensitivity to the spindle poisons, paclitaxel and vinorelbine. checkpoints after treatment with paclitaxel, an antimicrotubule agent Fluorescence-activated cell sorter analysis demonstrated that BRCA1 me- that functions by inhibiting the depolymerization of tubulin, thereby disrupting the mitotic spindle (13). Therefore, it appears that BRCA1 diates G2/M arrest in response to both antimicrotubule and DNA-damaging agents. However, poly(ADP-ribose) polymerase and caspase-3 cleav- can act in a more general capacity to activate cell cycle checkpoints in age assays indicate that the differential effect mediated by BRCA1 in response to different types of cellular stress. response to these agents occurs through the inhibition or induction of In addition to cell cycle arrest, BRCA1 has also been implicated in apoptosis. Therefore, our data suggest that BRCA1 acts as a differential the regulation of apoptosis. It was initially demonstrated that exoge- modulator of apoptosis depending on the nature of the cellular insult. nous expression of BRCA1 induced apoptosis, an effect that was dependent on JNK4/SAPK activation (14). It was reported subse- INTRODUCTION quently that BRCA1 modulates stress-induced apoptotic signaling Substantial evidence exists to support a role for BRCA1 in medi- through a pathway that sequentially involves the H-ras oncogene, ating the cellular response to DNA damage and, in particular, double- mitogen-activated protein kinase kinase kinase 4, JNK, Fas, and Fas strand DNA breaks. BRCA1 becomes hyperphosphorylated in re- ligand, and the activation of caspase-9 (15, 16). We have reported sponse to various DNA-damaging agents including ␥-irradiation, an recently that BRCA1 dramatically sensitizes breast cancer cell lines to effect that is mediated in part by the ATM (1) and chk2 kinases (2). IFN-␥-mediated apoptosis, indicating that BRCA1 may regulate ap- BRCA1 has been shown to colocalize at sites of DNA damage with optosis in response to diverse stress signals (17). In contrast, BRCA1- RAD51, the human homologue of bacterial RecA, which is involved deficient cells exhibit a radiosensitive phenotype after exposure to a in homologous recombination repair after ionizing radiation (3). Fur- range of DNA-damaging agents including ionizing radiation (2) and thermore, BRCA1 is a component of the RAD50, Mre11, NBS1 the DNA cross-linking agents, cisplatin and mitomycin C (18, 19). complex implicated in homologous recombination and nonhomolo- This radiosensitive phenotype, which can be reversed after exogenous gous end joining (4). Genetic studies support a role for BRCA1 in the expression of wild-type BRCA1, has been associated with a failure in repair of double-strand breaks. Significantly, ES cells from BRCA1 BRCA1-dependent DNA damage repair pathways. Consistent with knockout mice exhibit a defect in the repair of double-strand breaks this, antisense inhibition of BRCA1 has been reported to decrease by homologous recombination (5). More recently, it has been reported repair proficiency and enhance apoptosis in response to cisplatin (20). that BRCA1 resides within a large DNA repair protein complex called Therefore, it appears that BRCA1 in general confers an antiapoptotic- BRCA1-associated genome surveillance complex that includes vari- resistant phenotype after DNA damage. ous mismatch repair proteins including MLH1, MSH2, and MSH6, In the current study we provide new evidence to clarify the role suggesting a role for BRCA1 in mismatch repair (6). Interestingly, played by BRCA1 as a mediator of apoptosis and show that BRCA1 BRCA1 has also been implicated in the transcription-coupled repair of functions as a molecular determinant of response to a range of oxidative-induced DNA damage (7) and global genomic repair of cytotoxic chemotherapeutic agents. Specifically, we demonstrate that UV-induced cyclobutane pyrimidine dimers (8). Taken together, these BRCA1 abrogates the apoptotic phenotype induced by a range of reports suggest that BRCA1 is a component of multiple repair path- DNA-damaging agents including cisplatin, etoposide, and bleomycin, ways, which remain to be fully investigated. whereas inducing dramatic sensitivity to a range of antimicrotubule agents including paclitaxel and vinorelbine. Furthermore, we also Received 3/19/03; revised 5/26/03; accepted 6/13/03. provide mechanistic evidence to indicate that BRCA1 regulates the The costs of publication of this article were defrayed in part by the payment of page G /M checkpoint in response to both spindle poisons and DNA- charges. This article must therefore be hereby marked advertisement in accordance with 2 18 U.S.C. Section 1734 solely to indicate this fact. damaging agents. Therefore, these data indicate that BRCA1 func- 1 Supported by grants from The Breast Cancer Campaign, The Research and Devel- tions as a differential regulator of chemotherapy-induced apoptosis opment Office, Northern Ireland, and Cancer Research United Kingdom. 2 These authors contributed equally to this work. 3 4 To whom requests for reprints should be addressed, at Department of Oncology, The abbreviations used are: JNK, c-Jun NH2-terminal kinase; SAPK, stress activated Cancer Research Centre, The Queen’s University of Belfast, University Floor, Belfast protein kinase; CMV, cytomegalovirus; RT-PCR, reverse transcription-PCR; FACS, City Hospital, Lisburn Road, Belfast BT9 7AB, Northern Ireland, United Kingdom. fluorescence-activated cell sorter; PARP, poly(ADP-ribose) polymerase; siRNA, short Phone: 28-9032-9241; Fax: 28-90263744; E-mail: [email protected]. interfering RNA; ATM, ataxia telangiectasia mutated. 6221

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 2003 American Association for Cancer Research. BRCA1 IS A DIFFERENTIAL MEDIATOR OF APOPTOSIS and suggest that BRCA1 is worthy of additional investigation as a RESULTS potential predictive marker of response to these agents. BRCA1 Enhances Paclitaxel-induced Apoptosis. To study the functional properties of BRCA1 we generated inducible, tetracycline- MATERIALS AND METHODS regulated expression in a MDA435 breast cancer-derived cell line, termed MBR62-bcl2 (13, 14). BRCA1 expression levels were shown Generation of Cell Lines. MBR62-bcl2 cells were generated as described to increase ϳ4-fold above endogenous levels in this cell line after previously (13). HCC-empty vector and HCC-BRCA1 cells were generated by tetracycline withdrawal (Fig. 1A). This is well within the physiolog- the stable transfection of the HCC1937 breast cancer cell line with the ical levels observed for this protein in mouse models where BRCA1 Rc/CMV (Invitrogen) or Rc/CMV-BRCA1 constructs under selection with expression has been shown to increase 10-fold during pregnancy, a geneticin (G418; Sigma). The Rc/CMV construct that expresses the neomycin level that is maintained through postlactational involution (21). We resistance gene was obtained from Invitrogen, and the Rc/CMV-BRCA1 demonstrated previously that induction of BRCA1 in the presence of construct was generated as described previously (17). These constructs were transfected into the HCC1937 cell line using Genejuice methodology (Invitro- paclitaxel resulted in acute arrest at the G2/M phase of the cell cycle, gen) according to the manufacturer’s instructions. Transfected cells were an effect that correlated with BRCA1-mediated induction of GADD45 selected in 0.2 mg/ml G418, assessed by RT-PCR for expression of BRCA1, (13). To extend these studies additionally we evaluated the effect of and the resultant HCC-EV1, HCC-EV2, and HCCEV3, and HCC-BR116, inducible BRCA1 expression on the sensitivity of MBR62-bcl2 cells

HCC-BR18, and HCC-mixed clones (derived from a mixed population of to paclitaxel. We generated IC50 curves after treatment of the MBR62- transfected cells) were selected. Exogenous BRCA1 was detected by vector bcl2 cells with a range of different paclitaxel concentrations in the insert RT-PCR using a forward primer to BRCA1 and a reverse primer to the presence and absence of inducible BRCA1 expression. Inducible Rc/CMV 3-prime untranslated region (Rc/CMV-R). expression of BRCA1 increased paclitaxel sensitivity 100-fold in the BRCA1-F, 5Ј-AGGAGCTTTCATCATTCACCC -3Ј MBR62-bcl2 cells with an observed decrease in IC from 7.7 nM to RcЈCMV-R 5Ј-AACTAGAAGGCACAGTCGAGG -3Ј 50 96.4 pM (Fig. 1B). To determine whether BRCA1-induced sensitivity Cell Culture. The MBR62-bcl2 cell line was maintained as described previously (13). The HCC1937 and T47D breast cancer cell lines were main- to paclitaxel might reflect activation of an apoptotic pathway in these tained in RPMI 1640 supplemented with 20% FCS, 1 mM sodium pyruvate, cells we carried out PARP cleavage assays. The presence of the and 100 ␮g/ml penicillin-streptomycin (all from Life Technologies, Inc.). The HCC-EV1, EV2, EV3, BR116, BR18, and BR-mixed cell lines were grown in HCC1937 medium supplemented with 0.2 mg/ml G418. Flow Cytometry Analysis. DNA content was evaluated after propidium iodide staining of cells as described previously (13). In all of the cases FACS analysis was carried out using an EPICS ELITE flow cytometer (Coulter). Drug Sensitivity Assays. For assays in MBR62-bcl2 cells were seeded in 24-well plates, and induced to express BRCA1 by withdrawal of tetracycline from medium and at the same time were treated or untreated with 10 nM paclitaxel (Bristol Myers Squibb). For drug sensitivity assays in HCC-EV and HCCBR cells were seeded onto 24-well tissue culture plates at a density of 100,000 cells/well. After 24 h, cells were incubated in HCC1937 medium supplemented with the described concentration ranges of paclitaxel (Bristol Myers Squibb), vinorelbine (Novartis), cisplatin (Faulding Pharmaceuticals), etoposide “etopo-phos” (Bristol Myers Squibb), and 5- fluorouracil (David Bull Laboratories). After 72 h of continuous drug exposure, cells were detached with trypsin, and counted using a Z2 particle and size analyzer (Coulter,

Miami, FL). IC30–50 values were calculated for each drug from the respective sigmoidal dose-response curves using Prism software. Antibodies and Western Blot Analysis. MBR62-bcl2, HCC-EV, and HCC-BR116 cells were treated in the presence of 10 nM paclitaxel, 1 nM vinorelbine, 1 ␮M cisplatin, 1 ␮M etoposide, and 10 ␮M 5-fluorouracil. Protein lysates were extracted in radioimmunoprecipitation assay buffer [50 ␮M HEPES (pH 7.0), 150 mM NaCl, 0.1% Triton X-100, 0.1% sodium deoxy- cholate, and 0.1% SDS] separated on a 12% SDS polyacrylamide gel and transferred to polyvinylidene difluoride membrane followed by immunoblot- ting. BRCA1 Westerns were carried out using the rabbit polyclonal, BRCA1

NH2-terminal antibody, D-20 (Santa Cruz Biotechnology). BRCA1 IP West- erns were carried out using the rabbit polyclonal C-20 (Santa Cruz Biotech- nology) and the mouse monoclonal AB1 (Calbiochem). For PARP assays, the monoclonal antibody #556494 (BD Bioscience) was used, which specifically recognizes the full-length Mr 116,000 PARP protein and its Mr 85,000 and 25,000 cleaved products. Cleaved caspase-3 was detected using the rabbit polyclonal cleaved caspase-3 (Asp 175) antibody #9661 (Cell Signaling) that ␤ specifically recognizes the cleaved Mr 17,000 caspase-3 products. -Tubulin Fig. 1. Inducible expression of BRCA1 sensitizes MBR62-bcl2 cells to paclitaxel- was detected using the monoclonal antibody T-4026 (Sigma). induced apoptosis. A, IP-Western demonstrating inducible expression of BRCA1 in the ϩ Ϫ siRNA. T47D cells were transfected with a BRCA1-specific siRNA oligo- MBR62-bcl2 cells at 12 and 24 h after withdrawal of tetracycline ( tet, BRCA1 off; tet, BRCA1 on; B). Dose response curves detailing the effect of inducible expression of nucleotide or a control oligonucleotide and treated with a range of paclitaxel BRCA1 on the cytotoxicity of paclitaxel in the MBR62-bcl2 cells. The IC50 values were concentrations. After 72 h of continuous drug exposure, cells were detached calculated from the sigmoidal dose response curves. (Ⅺ, ϩtet, BRCA1 off; Œ, Ϫtet, using trypsin, and counted using a Z2 particle and size analyzer (Coulter). BRCA1 on). C, PARP cleavage assay demonstrating that BRCA1 sensitizes MBR62-bcl2 Ϫ IC values were calculated from the respective sigmoidal dose response cells to paclitaxel induced apoptosis. MBR62-bcl2 cells were induced ( tet) or uninduced 30-50 (ϩtet) to express BRCA1 for 24 or 48 h in the presence or absence of 10 nM paclitaxel. curves: BRCA1 siRNA oligonucleotide 5Ј-aac ctg tct cca caa agt gtg-3Ј and The presence of the cleaved Mr 85,000 PARP protein is indicative of apoptosis. The control siRNA oligonucleotide 5Ј-aaa acc cgu cua ggc ugu uac-3Ј. identical blot was reprobed with ␤-tubulin as a loading control. 6222

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 2003 American Association for Cancer Research. BRCA1 IS A DIFFERENTIAL MEDIATOR OF APOPTOSIS cleaved PARP Mr 85,000 product, which is an early hallmark of increased expression of wild-type BRCA1 in the HCC-BR116 cells apoptosis, was initially observed as early as 24 h after inducible relative to the mutant BRCA1 in vector only transfected controls expression of BRCA1 and was dramatically increased 48 h after (Fig. 2B). BRCA1 induction in the presence of paclitaxel, suggesting that To investigate the ability of BRCA1 to sensitize these cells to BRCA1 can sensitize these cells to paclitaxel-induced apoptosis (Fig. paclitaxel-mediated apoptosis we carried out PARP and caspase-3 1C). Paclitaxel treatment alone resulted in moderate PARP cleavage cleavage assays. PARP assays demonstrated that the empty vector- 48 h after treatment as expected (Figs. 1C). Similarly, induction of transfected HCC-EV1 and EV2 cells were resistant to paclitaxel- BRCA1 in the absence of paclitaxel treatment also resulted in mod- mediated apoptosis after treatment with 10 nM paclitaxel (Fig. 2C). In erate PARP cleavage consistent with our previous report that exoge- contrast, the BRCA1 reconstituted HCC-BR18 and HCC-BR116 cells nous expression of BRCA1 alone can induce apoptosis in these cells displayed the presence of the Mr 85,000 cleaved PARP product, (14). indicative of apoptosis after treatment with identical concentrations of BRCA1 Is Required for Paclitaxel-mediated Apoptosis. We paclitaxel (Fig. 2C). Western blots were reprobed with an anti-␤- used the HCC1937 breast cancer cells that express a single copy of a tubulin antibody to confirm equal loading of protein lysates (Fig. 2C). COOH-terminal truncated BRCA1 protein, which is transcriptionally Identical data were obtained after caspase-3 cleavage assays. Recon- inactive, as a model to investigate the requirement for BRCA1 in stitution of BRCA1 in the HCC-BR18 and HCC-BR116 cells resulted paclitaxel-mediated apoptosis (9). We generated the HCC1937- in a dramatic increase in the Mr 17,000 cleaved caspase-3 products derived cell lines, termed HCC-BR18, HCC-BR116, and HCCBR- after treatment with 10 nM paclitaxel (Fig. 2D). In contrast, the mixed (derived from a mixed transfected population) with constitutive HCC-EV1 and HCC-EV2 cells failed to display the presence of expression of exogenous wild-type BRCA1 and the control cell lines cleaved caspase-3 suggesting that these cells are resistant to pacli- HCC-EV1, HCC-EV2, and HCC-EV3 reconstituted with empty Rc/ taxel-mediated apoptosis (Fig. 2D). We extended these studies to look CMV vector. To confirm expression of exogenous BRCA1 we carried at paclitaxel sensitivity in the T47D breast cancer cell line after out vector-insert RT-PCR from cDNA derived from the various cell inhibition of endogenous BRCA1 using a siRNA approach. T47D lines. RT-PCR analysis confirmed expression of exogenous BRCA1 cells transfected with a BRCA1-specific siRNA oligonucleotide dis- expression in both the HCC-BR18 and HCC-BR116 clones (Fig. 2A). played a Ͼ50-fold increase in resistance to paclitaxel compared with

To confirm exogenous BRCA1 expression at the protein level we cells transfected with a control siRNA oligonucleotide with the IC50 carried out Western blot analysis to detect both endogenous mutant values calculated as Ͼ0.1 mM and 2.2 ␮M, respectively (Fig. 2E). and exogenous wild-type BRCA1. Western blot analysis confirmed Inhibition of endogenous BRCA1 after transfection with the BRCA1-

Fig. 2. BRCA1 is required for paclitaxel-induced apoptosis. A, vector insert RT-PCR demonstrating the presence of exogenous BRCA1 in the HCC-BR116 (Lane 3) and HCC-BR18 (Lane 4) compared with EV-1 (Lane 1) and EV-2 (Lane 2). A glyceraldehyde-3-phosphate dehydrogenase RT-PCR was included as a control to confirm equal amounts of template. M ϭ 123 bp marker. B, Western blot analysis confirming increased expression of BRCA1 in the HCC-BR116 cells (Lane 2) compared with the control vector transfected HCC-EV1cells (Lane 1). A ␤-tubulin Western was carried out on the identical protein lysate as a control for loading. C, PARP cleavage assay demonstrating the requirement for functional BRCA1 in paclitaxel-mediated apoptosis. HCC-EV1 (Lane 3), HCC-EV2 (Lane 5), HCC-BR116 (Lanes 4), and HCC-BR18 (Lane 6) were treated with paclitaxel for 24 h. The presence of the cleaved Mr 85,000 PARP protein is indicative of apoptosis and is only observed in the BRCA1 reconstituted HCC-BR116 and HCC-BR18 cells after paclitaxel treatment (Lanes 4 and 6). Untreated HCC-EV1 (Lane 1) and HCC-BR116 cells (Lane 2) are included as a control. The identical blot was reprobed with ␤-tubulin as a loading control. D, caspase-3 cleavage assay demonstrating the requirement for functional BRCA1 in paclitaxel-mediated apoptosis. HCC-EV1 (Lanes 3), HCC-EV2 (Lane 5), HCC-BR116 (Lanes 4), and HCC-BR18 (Lane 6) were treated with paclitaxel for 24 h. The presence of the cleaved Mr 17,000 caspase-3 fragments is indicative of apoptosis and is only observed in the BRCA1 reconstituted HCC-BR116 and HCC-BR18 cells after paclitaxel treatment. Untreated HCC-EV1 (Lane 1) and HCC-BR116 cells (Lane 2) are included as a control. The identical blot was reprobed with ␤-tubulin as a loading control. E, dose-response curve demonstrating a Ͼ50-fold increase in resistance of T47D cells to paclitaxel after transfection with a BRCA1-specific siRNA oligonucleotide compared with a control siRNA oligonucleotide. IC50 values for paclitaxel were calculated from the respective sigmoidal dose response curves. (Œ ϭ BRCA1 siRNA; Ⅺ ϭ control siRNA). F, IP-Western demonstrating a marked reduction in endogenous BRCA1 protein levels in T47D cells transfected with a BRCA1 specific siRNA oligonucleotide (Lane 2) compared with a control siRNA oligonucleotide (Lane 1). A ␤-tubulin Western was carried out using the identical protein lysate as a control for the specificity of the BRCA1 siRNA. 6223

specific siRNA oligonucleotide was confirmed by immunoprecipita- Specifically, the HCC-BR116 cells displayed an IC30 value of 1.9 nM ␮ tion-Western blot analysis relative to T47D cells transfected with the compared with an IC30 value of 17.0 M in the HCC-EV1 cells, control siRNA oligonucleotide (Fig. 2F). representing a 10,000-fold increase in sensitivity after reconstitution BRCA1 Functions as a Differential Modulator of Sensitivity to with BRCA1 (Fig. 3B). Chemotherapeutic Agents. The ability of BRCA1 to mediate sen- We observed the opposite effects when these cell lines are exposed

sitivity to paclitaxel-induced apoptosis prompted us to examine the to DNA damaging agents. The IC50 value for the HCC-EV1 cells after role played by BRCA1 in mediating sensitivity or resistance to a range bleomycin treatment was calculated at 0.5 mM. However, we failed to

of chemotherapeutic agents commonly used in cancer treatment. To obtain an IC50 value for the HCC-BR116 cells obtaining a value of do this we generated dose response curves for the HCC-BR116 and Ͼ1.0 mM, suggesting that BRCA1 induces extreme resistance to HCC-EV1 cells after treatment with the antimicrotubule agents pacli- bleomycin in these cells (Fig. 3C). Similarly, treatment with the ␮ taxel and vinorelbine, which disrupt the mitotic spindle, the radiom- topoisomerase II poison etoposide resulted in an IC50 value of 0.9 M Ͼ imetic bleomycin, which gives rise to double-strand breaks, the to- in the HCC-EV1 cells compared with an IC50 value of 0.1 mM in the poisomerase II poison etoposide, the inter and intra-strand DNA HCC-BR116 cells representing at least a 1000-fold increase in resist- cross-linking agent cisplatin, and the antimetabolite, 5-fluorouracil. ance (Fig. 3D). BRCA1 also induced 20-fold resistance to cisplatin ␮ The HCC-EV1 and HCC-BR116 cells were plated at equal densities with an IC50 value of 0.2 M observed in the HCC-EV1 cells com- ␮ and exposed to a range of concentrations of each drug for 72 h after pared with an IC50 value of 4.1 M in the HCC-BR116 cells (Fig. 3E). which time cell counts were carried out. Reconstitution of BRCA1 in Interestingly, BRCA1 failed to modulate resistance or sensitivity to

the HCC1937 cells resulted in a 1000-fold increase in sensitivity of the antimetabolite 5-fluorouracil with equivalent IC50 values obtained the HCC-BR116 cells to paclitaxel, with the IC50 value decreasing for both cell lines (Fig. 3F), probably reflecting the distinct mode of from 6.2 ␮M in the HCC-EV1 cells compared with 7.7 nM in the action of this agent. HCC-BR116 cells (Fig. 3A). Similar to that observed for paclitaxel, BRCA1 Is a Differential Modulator of Chemotherapy-induced the HCC-BR116 cells displayed a dramatic increase in sensitivity to Apoptosis. To investigate whether the observed chemosensitivity or vinorelbine over many concentrations compared with HCC-EV1 cells. resistance induced by BRCA1 was because of activation or inhibition

Fig. 3. BRCA1 functions as a differential modulator of sensitivity to chemotherapeutic agents. Dose-response curves comparing the IC30–50 concentrations of a range of chemotherapeutic agents in the BRCA1 mutant HCC1937 cells reconstituted with exogenous wild-type BRCA1 (HCC-BR116) compared with HCC1937 cells reconstituted with empty vector (HCC-EV1). The HCC-BR116 and HCC-EV1 cells were treated with a range of concentrations of (A) paclitaxel, (B) vinorelbine, (C) bleomycin, (D) etoposide, (E) cisplatin, and (F) 5-fluorouracil. In each case the HCC-BR116 cells display a marked increase in sensitivity to the antimicrotubule agents paclitaxel and vinorelbine compared with the HCC-EV1 cells and conversely display a marked resistance to DNA damaging agents compared with the HCC-EV1 cells. IC30–50 values for the individual chemotherapeutic agents were calculated from the respective sigmoidal dose response curves. (Ⅺ, HCC-BR116; Œ, HCC-EV1).

6224

of apoptosis we carried out PARP assays 24 h after treatment with of cell cycle checkpoints we carried out cell cycle analysis. The ␮ paclitaxel (10 nM), vinorelbine (1 nM), cisplatin (1 M), etoposide (1 HCC-BR116 and HCC-EV1 cells were treated with IC50 concentra- ␮M), bleomycin (0.1 mM), and 5-fluorouracil (10 ␮M). PARP assays tions of paclitaxel (10 nM) or etoposide (1 ␮M) for 12 h, followed by

demonstrated cleavage of the Mr 116,000 PARP product into its Mr FACS analysis to determine DNA content. Reconstitution of BRCA1 85,000 cleaved fragment when HCC-BR116 cells were treated with in the absence of drug treatment had no effect on cell cycle profile, either paclitaxel or vinorelbine (Fig. 4, A and B). In agreement with compared with vector-transfected controls (Fig. 6A). However, a

the drug sensitivity assays the HCC-EV1 cells were completely re- significant increase in the G2/M population was observed in the sistant to paclitaxel and vinorelbine-induced apoptosis at these con- HCC-BR116 (50%) cells compared with HCC-EV1 cells (21%) after centrations (Fig. 4, A and B). We observed the exact opposite after treatment with paclitaxel (Fig. 6B). Similar results were obtained with treatment of the HCC-BR116 cells with the DNA damaging agents vinorelbine (data not shown), suggesting that BRCA1 can regulate the

bleomycin, etoposide, and cisplatin. HCC-BR116 cells were resistant G2/M checkpoint after treatment with agents that disrupt the mitotic to apoptosis induced by these agents, whereas the HCC-EV1 cells spindle. In agreement with previous studies using ionizing radiation

displayed the characteristic Mr 85,000 PARP cleavage product indic- (10, 22), treatment of the HCC-BR116 cells with etoposide also ative of apoptosis (Fig. 4A). We repeated the drug sensitivity exper- resulted in a significant increase in the G2/M content in HCC-BR116 iments with two additional clones, termed HCC-BR118 and HCCBR- cells (50%) compared with HCC-EV1 cells (31%; Fig. 6C). Similar mixed, using paclitaxel and etoposide as representative agents. In results were obtained for bleomycin (data not shown), suggesting that

keeping with our previous observations, reconstitution of BRCA1 in BRCA1 regulates the G2/M checkpoint in response to agents that the HCC-BR18 and HCCBR-mixed cells resulted in an ϳ5,000-fold induce double-strand breaks in DNA. and 10,000 fold increase in sensitivity to paclitaxel, respectively, compared with the HCC-EV2 and HCC-EV3 cells (Fig. 5, A and C). DISCUSSION Similarly, the HCC-BR18 cells were observed to be 100-fold more resistant and the HCCBR-mixed cells ϳ1,000 fold more resistant to In the current study we provide evidence to suggest that BRCA1 etoposide compared with the HCC-EV2 and HCC-EV3 cells, respec- acts as a molecular determinant of response to a range of chemother- tively (Fig. 5, A and C). Finally, to illustrate that this difference apeutic agents commonly used in cancer treatment. We demonstrate

reflected a corresponding increase or decrease in apoptosis we carried that BRCA1 mediates G2/M arrest in response to agents that disrupt out PARP cleavage assays in these cells. The HCC-BR18 and HC- the mitotic spindle including paclitaxel and to DNA-damaging agents

CBR-mixed cells displayed the characteristic Mr 85,000 PARP cleav- such as etoposide. In contrast, BRCA1 dramatically sensitizes breast age product after treatment with paclitaxel but not with etoposide, cancer cell lines to apoptosis induced by the antimicrotubule agents whereas the reverse was true for the HCC-EV2 and HCC-EV3 cells paclitaxel and vinorelbine, yet induced resistance to a range of DNA- (Fig. 5, B and D). damaging agents, most notably those that induce double-strand breaks

BRCA1 Induces G2/M Arrest in Response to Antimicrotubule in DNA including bleomycin and etoposide. The observation that

and DNA Damaging Agents. To determine whether the differential BRCA1 mediates G2/M arrest in response to both antimicrotubule effects on apoptosis mediated by BRCA1 in response to DNA- agents and DNA-damaging agents is intriguing considering the dif- damaging agents and spindle poisons reflected differential modulation ferential effects on apoptosis. Standard FACS analysis does not dis-

tinguish between the G2 and mitotic populations of cells. Therefore, it is possible that the differential effect on apoptosis may reflect differ-

ent subpopulations of cells in the G2/M and mitotic phases in response to the different agents. In support of this argument we have reported previously that the inducible expression of BRCA1 can activate the mitotic checkpoints in response to paclitaxel (13). In addition a dominant-negative BRCA1 construct encoding the COOH terminus

was shown to abrogate G2/M arrest in response to the spindle poison colchicine (23). Moreover, BRCA1 has been shown to localize to centrosomes via a physical interaction with ␥-tubulin, suggesting that BRCA1 plays a direct role in the accurate segregation of duplicated chromosomes during mitosis (24). It has also been reported that BRCA1 is required for the decatenation and, therefore, accurate separation of sister chromatids at anaphase lending additional support for a role in mitotic checkpoint regulation (25). Genetic studies support these observations revealing aneuploidy characteristic of a Fig. 4. BRCA1 is a differential modulator of chemotherapy induced apoptosis. A, defect in the mitotic checkpoint in BRCA1 exon 11 isoform-deficient PARP cleavage assay demonstrating the absence of the cleaved PARP Mr 85,000 fragment in HCC-BR116 cells (Lanes 8, 10, and 12) compared with HCC-EV1 cells (Lanes 7, 9, cells (12). Our observation that BRCA1 can specifically sensitize and 11), 24 h after treatment with etoposide, bleomycin, or cisplatin, respectively. In direct breast cancer cell lines to apoptosis induced by antimicrotubule agents contrast HCC-BR116 cells display a marked increase in cleaved PARP indicative of apoptosis, 24 h after treatment with paclitaxel (Lane 4) compared with HCC-EV1 cells is intriguing in the context of these other studies. Antimicrotubule (Lane 3). 5-FU treatment induced equivalent levels of PARP cleavage in both the agents function by promoting the stabilization or destabilization of HCC-BR116 and HCC-EV1 cells (Lanes 5 and 6) consistent with the distinct mode of tubulin, which, in turn, prevents sister chromatid separation and arrest action of this agent. Untreated HCC-BR116 (Lane 1) and HCC-EV1 (Lane 2) lysates are included as a control. The identical blot was reprobed with ␤-tubulin as a control for at the metaphase to anaphase transition (26). The mechanisms through loading. B, PARP cleavage assay demonstrating the presence of the cleaved PARP Mr which spindle poisons such as paclitaxel induce apoptosis are still a 85,000 fragment indicative of apoptosis in HCC-BR116 cells (Lanes 6 and 8) compared matter of speculation; however, it has been suggested that different with the HCC-EV1 cells (Lanes 5 and 7) 24 h after treatment with paclitaxel or vinorelbine. In direct contrast the HCC-BR116 cells fail to undergo PARP cleavage 24 h pathways are activated depending on the concentrations used. Higher after treatment with etoposide (Lane 4) compared with the HCC-EV1 cells, which exhibit concentrations of paclitaxel in the region of 0.2–3.0 mM give rise to marked PARP cleavage after treatment with the identical concentration of etoposide (Lane 3). Untreated HCC-BR116 (Lane 1) and HCC-EV1 (Lane 2) lysates are included as a microtubule damage and activation of JNK-dependent apoptosis (26). control. The identical blot was reprobed with ␤-tubulin as a control for loading. In contrast, paclitaxel concentrations of 10 ␮M to 0.1 mM result in 6225

Fig. 5. BRCA1 is a differential modulator of chemotherapy induced apoptosis. Dose-response curves comparing the IC50 concentrations of paclitaxel and etoposide in the HCC-BR18 (A) and HCC-BR mixed (C) cells reconstituted with exogenous wild-type BRCA1 compared with HCC-EV2 (A) and HCC-EV3 (C) cells reconstituted with empty vector. The HCC-BR18/HCC-BR mixed and HCC-EV2/EV3 cells were treated with a range of concentrations of (i) paclitaxel and (ii) etoposide. The HCC-BR18 and HCC-BR mixed cells display a marked increase in sensitivity (5000-fold and 7000-fold, respectively) to paclitaxel compared with the HCC-EV2 and HCC-EV3 cells ,and conversely display a marked resistance Ͼ to etoposide (100-fold and 1000-fold, respectively) compared with the HCC-EV2 and HCC-EV3 cells. IC50 values for the individual chemotherapeutic agents were calculated from the respective sigmoidal dose response curves. (Ⅺ, HCC-BR18/HCC-BR mixed; Œ, HCC-EV2/HCC-EV3). PARP cleavage assays demonstrating the differential effects of paclitaxel and etoposide on the induction of apoptosis in the HCC-BR18 and HCC-BR mixed cells compared with the HCC-EV2 and HCCEV3 cells. The HCC-BR18 and HCC-BR mixed cells exhibit marked PARP cleavage after treatment with paclitaxel (B, Lane 4; D, Lane 6) but are resistant to apoptosis induced by etoposide (B, Lane 6; D, Lane 4). In contrast the HCC-EV2 and HCC-EV3 cells are resistant to apoptosis induced by paclitaxel (B, Lane 3; D, Lane 5) but are sensitive to apoptosis induced by etoposide (B, Lane 5; D, Lane 3). Untreated HCC-BR18/HCC-BRmixed (B, Lane 1; D, Lane1) and HCC-EV3/HCC-EV3 (B, Lane 2; D, Lane 2) lysates are included as a control. The identical blots were reprobed with ␤-tubulin as a control for loading in each case. suppression of microtubule dynamics and mitotic arrest in the absence may play a role in the regulation of the spindle checkpoint, which, in of microtubule damage (26). turn, results in the induction of apoptosis through a number of poten- In this study we demonstrate that BRCA1 sensitizes HCC1937 cells tial pathways. The mechanistic basis of BRCA1-induced apoptosis in to apoptosis at concentrations ranging from 1 nM to 10 ␮M suggesting response to paclitaxel and vinorelbine remains to be defined; however, that BRCA1 may specifically play a role in mediating apoptosis at it is reasonable to speculate that BRCA1 may regulate key proapo- lower concentrations. Indeed, cellular proliferation assays clearly dis- ptotic target genes after disruption of the mitotic spindle. Additional play a biphasic decrease in cell viability with BRCA1 mediating elucidation of this pathway may therefore provide critical new infor- sensitivity at paclitaxel concentrations as high as 0.1 ␮M followed by mation on the mechanism of paclitaxel sensitivity and resistance in a BRCA1-independent paclitaxel sensitivity at concentrations higher breast tumors. than 10 ␮M (Fig. 3A; Fig. 5, A and C). Furthermore, we demonstrate The observation that BRCA1 sensitizes HCC1937 cells to antimi- that inhibition of endogenous BRCA1 results in a dramatic increase in crotubule-induced apoptosis is in direct contrast to that observed after resistance of breast cancer cells to paclitaxel-induced apoptosis (Fig. exposure to DNA-damaging agents where BRCA1 clearly mediates 2, E and F) suggesting that BRCA1 expression levels are critical for resistance. We observed a dramatic increase in resistance to apoptosis mediating apoptosis in response to this class of chemotherapeutic in HCC1937 cells reconstituted with wild-type BRCA1 after treat- agent. Irrespective of the mechanistic basis of antimicrotubule- ment with the radiomimetic bleomycin and the topoisomerase II induced apoptosis it appears that the initial step involves the activation inhibitor etoposide, both of which introduce double-strand breaks in of the spindle checkpoint (27). A number of potential scenarios can DNA, although through different mechanisms. We also demonstrate then occur that ultimately result in the induction of apoptosis, includ- that BRCA1 can induce resistance to cisplatin-mediated apoptosis ing the generation of multinucleated cells and the onset of mitotic suggesting that BRCA1 may also be involved in the repair of inter and catastrophe or entrance of cells into a tetraploid G1 state that also intrastrand cross-links. These data are in agreement with previous results in apoptosis. The observation that BRCA1 enhances antimi- reports, which demonstrate that reconstitution of BRCA1 in HCC1937 crotubule-induced G2/M arrest and apoptosis suggests that BRCA1 cells can rescue the radioresistant phenotype, although this had not 6226

Fig. 6. BRCA1 mediates G2/M arrest in response to antimicrotubule and DNA damaging agents. Cell cycle distribution analysis of propidium iodide-stained HCC-BR116 and HCC-EV1 cells untreated (A), or after treatment with 10 nM paclitaxel (B)or1␮M etoposide (C) for 12 h.

been linked to inhibition of apoptosis (2, 18, 19). The mechanism a range of different DNA-damaging agents. Many studies have also through which BRCA1 rescues the resistant phenotype in response to suggested that BRCA1 is required for the efficient repair of DNA these agents has yet to be defined; however, it is likely to be linked to damage and, in particular, double-strand breaks in DNA. BRCA1 has the activation of BRCA1 dependent G2/M arrest and or DNA repair been shown to colocalize at sites of DNA damage with RAD51 and pathways (5). In support of this hypothesis, BRCA1 has been shown RAD50, which are involved in homologous recombination and non- to be essential for activation of the Chk1 kinase and regulation of the homologous end joining, respectively (3, 4). In addition, ES cells from

G2/M checkpoint after DNA damage (22). Similarly, ATM-dependent BRCA1 knockout mice exhibit a defect in the repair of double-strand phosphorylation of BRCA1 on serine 1423 has been shown to be breaks by homologous recombination (5). Therefore, it is possible that required for BRCA1-induced activation of the G2/M checkpoint after BRCA1 functions to couple G2/M arrest with repair pathways, thereby ionizing radiation (10). In light of the above studies, our observation mediating survival after DNA damage. The exact mechanism through that BRCA1 can mediate G2/M arrest in response to etoposide and which BRCA1 regulates DNA repair pathways is unclear, but may bleomycin suggests that BRCA1 plays an important role in the reg- involve its ability to function as a transcriptional coactivator or ulation of the G2/M checkpoint after exposure of breast cancer cells to repressor. BRCA1 has been shown to mediate the induction of genes 6227

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 2003 American Association for Cancer Research. BRCA1 IS A DIFFERENTIAL MEDIATOR OF APOPTOSIS involved in nucleotide excision repair in a p53-independent manner 10. Xu, B., Kim S-T., and Kastan, M. B. Involvement of Brca1 in S-phase and G2-phase after exposure to UV-irradiation (8). In addition, it has been demon- checkpoints after ionizing irradiation. Mol. Cell. Biol., 21: 3445–3450, 2001. 11. Xu, B., O’Donnell A. H., Kim S-H., and Kastan, M. B. Phosphorylation of serine strated that exogenous expression of BRCA1 mediates the induction 1387 in Brca1 is specifically required for the ATM mediated S-phase checkpoint after of growth arrest and DNA repair genes in a p53-dependent manner in ionizing irradiation. Cancer Res., 62: 4588–4591, 2002. the absence of DNA damage (28). Because the HCC1937 cell line also 12. Xu, X., Weaver, Z., Linke, S. P., Li, C., Gotay, J., Wang X-W, Harris, C. C., Reid, T., and Deng, C. X. Centrosome amplification and a defective G2-M cell cycle harbors mutant p53, our data would suggest that BRCA1-mediated checkpoint induce genetic instability in BRCA1 exon 11 isoform-deficient cells. Mol. inhibition of apoptosis in response to DNA damage and conversely Cell, 3: 389–395, 1999. 13. Mullan, P. B., Quinn, J. E., Gilmore, P. M., Mc Williams, S., Andrews, H. N., Gervin, induction of apoptosis in response to spindle poisons occurs inde- C., McCabe, N., McKenna, S., White, P., Song, Y. H., Maheswaran, S., Liu, E., pendently of a fully functional p53 protein. Haber, D. A., Johnston, P. G., and Harkin, D. P. BRCA1 and GADD45 mediated Therefore, in summary, the data presented suggest that BRCA1 G2/M cell cycle arrest in response to antimicrotubule agents. Oncogene, 20: 61: 23–31, 2001. functions as a molecular determinant of response to a range of 14. Harkin, D. P., Bean, J. M., Miklos, D., Song Y-H, Truong, V. B., Englert, C., different chemotherapeutic agents and, in particular, functional Christians, F. C., Ellisen, L. W., Maheswaran, S., Oliner, J. D., and Haber, D. A. BRCA1 may be required for an effective apoptotic response to pacli- Induction of GADD45 and JNK/SAPK-dependent apoptosis following inducible expression of BRCA1. Cell, 97: 575–586, 1999. taxel chemotherapy. With the recent adoption of paclitaxel containing 15. Takekawa, M., and Saito, H. A family of stress-inducible GADD45-like proteins regimens for potentially curable breast cancer (29) it is likely that mediate activation of the stress-responsive MTK1/MEKK4 MAPKKK. Cell, 95: determinants of response to this agent will become increasingly im- 521–539, 1998. 16. Thangaraju, M., Kaufmann, S. H., and Couch, F. J. BRCA1 facilitates stress-induced portant. Therefore, we suggest that BRCA1 expression level could be apoptosis in breast and ovarian cancer cell lines. J. Biol. Chem., 275: 33487–33496, considered as a potential predictive marker for response to chemo- 2000. 17. Andrews, H. N., Mullan, P. B., McWilliams, S., Seblova, S., Quinn, J. E., Gilmore, therapy in both sporadic and inherited breast cancer. P. M., McCabe, N., Pace, A., Koller, B., Johnston, P. G., Haber, D. A., and Harkin, D. P. BRCA1 regulates the interferon ␥ -mediated apoptotic response. J. Biol. Chem., 277: 26225–32, 2002. ACKNOWLEDGMENTS 18. Moynahan, M. E., Cui, T. Y., and Jasin, M. Homology-directed DNA repair, Mito- mycin-C resistance, and chromosome stability is restored with correction of a BRCA1 We thank Prof. Daniel Haber, Prof. Peter Hall, and Dr. Shyamala Maheswa- mutation. Cancer Res., 61: 4842–4850, 2001. ran for critical review of this manuscript. We also thank Stephen McNeice, 19. Bhattacharyya, A., Ear, U. S., Koller, B. H., Weichselbaum, R. R., and Bishop, D. K. Stephanie Tohill, Michael Jackson, and Claire Hillen of the Belfast City The breast cancer susceptibility gene BRCA1 is required for subnuclear assembly of RAD51 and survival following treatment with the DNA cross-linking agent Cisplatin. Hospital Pharmacy Department for their advice and support. J. Biol. Chem., 275: 23899–903, 2000. 20. Husain, A., He, G., Venkitaraman, E. S., and Spriggs, D. R. BRCA1 up-regulation is associated with repair-mediated resistance to cis-diamminedichloroplatinum (II). REFERENCES Cancer Res., 58: 1120–1123, 1998. 1. Cortez, D., Wang, Y., Qin, J., and Elledge, S. J. Requirement of ATM dependent 21. Marquis, S. T., Rajan, J. V., Wynshaw-Boris, A., Xu, J., YinG-J, Abel, K. J., Weber, phosphorylation of BRCA1 in the DNA damage response to double strand breaks. B. L., and Chodosh, L. A. The developmental pattern of brca1 expression implies a Science (Wash. DC), 286: 1162–1166, 1999. role in differentiation of the breast and other tissues. Nat. Genet., 11: 17–26, 1995. 2. Lee, J. S., Collins, K. M., Brown, A. L., Lee, C. H., and Chung, J. H. hCds1-mediated 22. Yarden, R. I., Pardo-Reoyo, S., Sgagias, M., Cowan, K. H., and Brody, L. C. BRCA1 phosphorylation of BRCA1 regulates the DNA damage response. Nature (Lond.), regulates the G2/M checkpoint by activating Chk1 kinase upon DNA damage. Nat. 404: 201–204, 2000. Genet., 3: 285–289, 2002. 3. Scully, R., Chen, J., Plug, A., Xiao, Y., Weaver, D., Feunteun, J., Ashley, T., and 23. Larson, J. S., Tonkinson, J. L., and Lai, M. T. A BRCA1 mutant alters G2-M cell Livingston, D. M. Association of BRCA1 with RAD51 in mitotic and meiotic cells. cycle control in human mammary epithelial cells. Cancer Res., 57: 335 1–55, 1997. Cell, 88: 265–275, 1997. 24. Hsu, L. C., and White, R. L. BRCA1 is associated with the centrosome during mitosis. 4. Zhong, Q., Chen, C. F., Li, S., Chen, Y., Wang C-C, Xiao, J., Chen, P. L., Sharp, Proc. Natl. Acad. Sci. USA, 95: 12983–88, 1998. Z. D., and Lee, W. H. Association of BRCA1 with the hRad50-hMre11-p95 complex 25. Deming, P. B., Cistuilli, C. A., Zhao, H., Graves, P. R., Piwnica-Worms, H., Paules, and the DNA damage response. Science (Wash. DC), 285: 2747–2750, 1999. R. S., Downes, C. S., and Kaufmann, W. K. The human decatenation checkpoint. 5. Moynahan, M. E., Chiu, J. W., Koller, B. H., and Jasin, M. Brca1 controls homology- Proc. Natl. Acad. Sci. USA, 21: 12044–49, 2001. directed DNA repair. Mol. Cell, 4: 511–518, 1999. 26. Wang, T. H., Wang, H. S., and Soong, Y. K. Paclitaxel-induced cell death: where the 6. Wang, Y., Cortez, D., Yazdi, P., Neff, N., Elledge, S. J., and Qin, J. BASC, a super cell cycle and apoptosis come together. Cancer (Phila.), 88: 2619–2628, 2000. complex of BRCA1-associated proteins involved in the recognition and repair of 27. Blajeski, A. L., Kottke, T. J., and Kaufmann, S. H. A multistep model for paclitaxel aberrant DNA structures. Genes Dev., 14: 927–939, 2000. induced apoptosis in human breast cancer cells. Exp. Cell Res., 270: 277–280, 2002. 7. Gowen, L. C., Avrutskaya, A. V., Latour, A. M., Koller, B. H., and Leadon, S. A. 28. MacLachlan, T. K., Takimoto, R., and El-Deiry, W. S. BRCA1 directs a selective BRCA1 is required for the transcription-coupled repair of oxidative DNA damage. p53-dependent transcriptional response towards growth arrest and DNA repair tar- Science (Wash. DC), 281: 1009–1012, 1998. gets. Mol. Cell. Biol., 22: 4280–4292, 2002. 8. Hartman A-R., and Ford, J. BRCA1 induces DNA damage recognition factors and 29. Henderson, C., Berry, D. A., Demetri, G. D., Cirrincione, C. T., Goldstein, L. J., enhances nucleotide excision repair. Nat. Genet., 32: 180–184, 2002. Martino, S., Ingle, J. N., Cooper, M. R., Hayes, D. F., Tkaczuk, K. H., Fleming, G., 9. Tomlinson, G. E., Chen TT-L, Stastny, V. A., Virmini, A. K., Spillman, M. A., Tonk, Holland, J. F., Duggan, D. B., Carpenter, J. T., Frei, E., Schilsky, R. L., Wood, W. C., V., Blum, J. L., Schneider, N. R., Wistuba, I. I., Shay, J. W., Minna, J. D., and Gazdar, Muss, H. B., and Norton, L. Improved outcomes from adding sequential paclitaxel but A. F. Characterization of a breast cancer cell line derived from a germ-line BRCA1 not from escalating doxorubicin dose in an adjuvant chemotherapy regimen for mutation carrier. Cancer Res., 58: 3237–3242, 1998. patients with node-positive primary breast cancer. J. Clin. Oncol., 21: 976–983, 2003.

6228

Jennifer E. Quinn, Richard D. Kennedy, Paul B. Mullan, et al.

Cancer Res 2003;63:6221-6228.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/63/19/6221

Cited articles This article cites 27 articles, 15 of which you can access for free at: http://cancerres.aacrjournals.org/content/63/19/6221.full#ref-list-1

Citing articles This article has been cited by 45 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/63/19/6221.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 Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/63/19/6221. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.