CDDO-Imidazolide Induces DNA Damage, G2/M Arrest and Apoptosis in BRCA1-Mutated Breast Cancer Cells

The breast cancer-associated gene 1 (BRCA1) is the most frequently mutated tumor suppressor gene found in familial breast cancers, and 27 missense variants, affecting 15 codons and over 670 truncating mutations have been identified for BRCA1 (1). BRCA1 mutations also predispose to other types of cancers, pointing to a fundamental role of this pathway in tumor suppression. BRCA1 is a large protein with multiple functional domains and associates with a number of proteins that are involved in many important biological processes/pathways (2). BRCA1 plays important roles in DNA damage and repair response, homologous recombination, cell-cycle regulation, protein ubiquitination, and apoptosis (3). Loss of BRCA1 causes a defective DNA repair response and a G2/M cell-cycle checkpoint. Because of their high risk for developing invasive cancer, BRCA1-mutation carriers have special needs for effective preventive and therapeutic agents.

We used primary cell lines from mammary tumors of Brca1^Co/Co; MMTV-Cre; p53^+/− mice containing a targeted deletion of full-length Brca1 in this study, as the tumors and cell lines from these mice mimic human BRCA1-mutated breast cancer as shown by extensive microarray experiments comparing mouse and human breast tumors. Gene set enrichment analysis (GSEA), used to search for shared relationships between mouse and human BRCA1 tumors, showed that functional gene sets exhibit similar levels of correlation between gene expression and the class labels (i.e., BRCA1-proficient vs. BRCA1-deficient) in both species. Hence, gene sets that were predictive for the human tumor samples were also likely to be predictive for the mouse samples (4). Therefore, it is likely that studies using primary cell lines from mammary tumors of Brca1^Co/Co; MMTV-Cre;p53^+/− mice should have significant value for preclinical evaluation of new drugs.

Synthetic triterpenoid derivatives of oleanolic acid, including 2-cyano-3,12-dioxooleana-1,9-dien-28-oic acid (CDDO), CDDO-imidazolide (CDDO-Im) and CDDO-methyl ester (CDDO-Me), have strong antiinflammatory and apoptotic activity in various cancer cells and are effective for the prevention and treatment of cancer in a variety of preclinical animal models (5). Among them, CDDO-Im is known as one of the most potent cytoprotective agents to induce phase 2 enzymes via Nrf2-ARE signaling in several tissues (5). In this study, CDDO-Im inhibits proliferation and induces apoptosis in BRCA1-deficient mammary cancer cells. However, precise molecular mechanisms that initiate the selective induction of apoptosis by triterpenoids in cancer cells remain incompletely clarified (6). Some reports propose that triterpenoids induce apoptosis via the intrinsic cell death pathway, whereas others have found that other triterpenoids act through the extrinsic apoptotic pathway. Here, we have investigated the relationship among CDDO-Im, reactive oxygen species (ROS), DNA damage and apoptosis in BRCA1-mutated breast cancer cells. Our results show that CDDO-Im rapidly and selectively induces intracellular ROS, which causes DNA damage and the activation of checkpoint signaling. In particular, we show that a target of CDDO-Im is checkpoint kinase1 (Chk1), which functions as a feedback link between ROS and DNA damage. This is the first report to show that the ability of CDDO-Im to inhibit BRCA1 mutant tumor cell growth is associated with oxidative DNA damage followed by the activation of the G2/M checkpoint pathway, triggering selective apoptosis of the tumor cells. Thus, CDDO-Im would be expected to have beneficial effects in the prevention or treatment of BRCA1-mutated breast cancer.

CDDO-Im and other derivatives (CDDO-Me, CDDO-EA, and CDDO-TFEA) were synthesized as described (7–9). For cell culture studies, compounds were dissolved in DMSO, and controls containing equal concentrations of DMSO (<0.1%) were included in all experiments. Sources of reagents and antibodies were as follows: H₂DCFDA from Molecular Probes (Eugene); Chk1 siRNA, scrambled siRNA and antibodies against p21^Waf1/Cip1 from Santa Cruz Biotechnologies; γH2AX from R&D Systems; pChk1, pChk2, and pCdc2 from Cell Signaling Technology; PARP from Upstate. UCN-01 and uric acid were obtained from Sigma. Uric acid was solubilized in 1 M NaOH with heat.

The Brca1 mutant cell lines W0069 and W780 were derived from mammary tumors of Brca1^Co/Co; MMTV-Cre;p53^+/− mice containing a targeted deletion of full-length Brca1 (10). To generate a new cell line from mammary tumors of Brca1^Co/Co; MMTV-Cre;p53^+/− mice, a tumor was minced with a scalpel and digested in phosphate-buffered saline (PBS) supplemented with collagenase (300 units/mL, Sigma) for 30 minutes at 37°C with gentle agitation from a stir bar. The cell suspension was filtered through a 40-μm Cell Strainer (BD Bioscience), centrifuged at 220 g for 10 minutes, and plated in DMEM + 20% fetal bovine serum. After 10 passages in this medium, a stable cell line (B8701) was obtained and then used for mechanistic studies. All mouse Brca1 mutant cell lines (10) were cultured in DMEM with 5% FBS and supplemented with antibiotics (Invitrogen); HCT1937 cells (ATCC) were grown in RPMI with 10% FBS; MCF-10A cells (ATCC) were cultured in DMEM/F12 + 5% HS; and E18–14C-27 cells (provided by Powel Brown, MD Anderson Cancer Center) and 3T3 cells (ATCC) were grown in DMEM with 10% FBS. Cells were treated with CDDO-Im dissolved in DMSO in a 10 mmol/L stock solution and used at the final concentrations indicated in the text and in figure legends.

For the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay, cells were seeded into 48-well plates at 1–2 × 10⁴ cells/well overnight before synthetic triterpenoids were added. Then the cells were incubated with different concentrations of triterpenoids for 24 hours. Cells were incubated with MTT for 2 hour (Sigma) and read at OD570. For proliferation assays, cells were treated with compounds for 24 hours, pulsed with ³H-thymidine for 2 hours, and its incorporation into cells measured by scintillation counting. Apoptosis was analyzed by fluorescence-activated cell sorting using the TACS Annexin V-FITC Apoptosis Detection kit (R&D Systems) and Cell Quest software (Becton Dickinson) or by Western blotting with PARP antibodies.

Cells were treated with CDDO-Im or DMSO as a control. After trypsinization, cells were fixed in 70% ethanol for 30 minutes at 4°C. The cells were washed twice with PBS, and then incubated for 30 minutes under dark conditions at 37°C in 1 mL of PBS containing 100-μg propidium iodide and 100-μg RNase A. Flow cytometric analysis was performed with histograms generated using the computer program, Cell Quest and Mod-Fit.

Small interfering RNA (siRNA) was purchased from Santa Cruz Biotechnologies and transfected into W780 cells using Lipofectamine 2000 according to the manufacturer's procedure (Invitrogen).

Cellular ROS contents were measured by incubating the control or drug-treated W0069, W780, B8701, or NIH3T3 cells with 10-μmol/L H₂DCFDA for 30 minutes, followed by flow cytometry using a FACS Calibur equipped with Cell Quest software.

DNA damage was measured using single cell gel electrophoresis (Comet Assay) (Trevigen). After electrophoresis, the slides were stained with silver and then covered with 100 μL of 5% acetic acid to stop the reaction prior to drying in the dark overnight. Measurement of mean tail moment was from 100 cells/slide from 15–20 randomly selected fields representing the whole area of each slide. Comet tail moments were determined using Comet Assay IV software (Perceptive Instruments).

When necessary, data were expressed as means ± SD of at least 3 independent experiments, and statistical analysis for single comparison was conducted using the Student's t-test. The criterion for statistical significance was P < 0.05.

To determine the cytotoxic activity of synthetic triterpenoids against cultured BRCA1-mutated breast cancer cells, W780 and W0069 cells were treated with CDDO-Im, CDDO-Me and CDDO- trifluoroethyl amide (TFEA) for 24 hours. CDDO-Im was the most potent inhibitor of cellular proliferation, as measured by ³H-thymidine incorporation (Fig. 1A). CDDO-Im inhibited proliferation in a concentration-dependent way in all 3 BRCA1-mutated human breast cancer lines tested (Fig. 1B). To clarify whether the inhibition of cell growth by CDDO-Im is attributed to its ability to induce apoptosis, we examined the effect of CDDO-Im on the staining of FITC-Annexin V, a biochemical hallmark of apoptosis. CDDO-Im induced apoptosis in 3 different BRCA1-deficient cancer cells but had no effect in normal mouse 3T3 fibroblasts (Fig. 1C). W780 and B8701 were more sensitive to the apoptotic stimuli of CDDO-Im than W0069 cells. Treatment of W780 cells with CDDO-Im caused cleavage of poly(ADP-ribose) polymerase (PARP) (Fig. 1D) and significantly inhibited PARP activity (data not shown). We also conducted the proteome profiler antibody array analysis (R&D Systems) that can measure the relative levels of expression of 35 apoptosis-related proteins. We found that CDDO-Im significantly induced cleavage of caspase-3 and the expression of HSP70, and downregulated the expression of claspin, survivin, Bclx, and Bad (supp. 1).

To investigate further the inhibitory effect of CDDO-Im on cell growth, we examined the cell-cycle distribution in W780 cells by flow cytometry. As shown in Fig. 2A and B, quantitation of cells in G1, S, and G2/M phase of the cell cycle showed that W780 cells accumulated in G2/M phase after exposure to CDDO-Im, whereas the cell population in G1 was decreased in a concentration-dependent manner. Collectively, these observations suggest that CDDO-Im causes G2/M arrest in W780 cells. Chk1 has emerged as the part of the cellular machinery that recognizes and responds to DNA damage, the signaling cascade that specifically regulates cell-cycle arrest after DNA damage (11). Because Chk1 is known as one of the most important members in the G2/M DNA damage checkpoint (12), the effects of UCN-01 on cell cycle delay were investigated in W780 cells exposed to CDDO-Im (Fig. 3A). The staurosporin analog UCN-01 is a potent inhibitor of Chk1 but not Chk2 (13). Quantification of the percentage of cells in each phase of the cell cycle showed that UCN-01 decreased the proportion of CDDO-Im-treated W780 cells in G2/M. To confirm that Chk1 is required for G2/M arrest, W780 cells were transfected with either control or Chk1 small interfering RNA (siRNA) before CDDO-Im-stimulated cell cycle analysis (Fig. 3B). Protein immunoblotting confirmed siRNA specificity for Chk1 in W780 cells (data not shown). Chk1 siRNA abolished the effect of CDDO-Im to induce G2/M arrest in these cells (Fig. 3C), indicating that activation of Chk1 might be crucial for G2/M arrest induced by CDDO-Im.

To assess further the involvement of the G2/M checkpoint pathway, we conducted immunoblotting using antibodies recognizing phospho-specific Chk1, Chk2, Cdc2, or p21 (Fig. 3D). We chose these particular proteins because they are known to regulate cell-cycle arrest. Treatment of BRCA1-mutated breast cancer cells with either CDDO-Im or CDDO-Me markedly increased phosphorylation of Chk1 at Ser345 and Chk2 at Thr68. Several reports suggest that p21^waf1/Cip1 regulates the entry of cells at the DNA damage-induced G2/M checkpoint as an inhibitor of cyclin-dependent protein kinases (CDK) and induces apoptosis (14, 15). Western analysis revealed that treatment with synthetic triterpenoids for 24 hours resulted in a marked induction of the expression of p21^waf1/Cip1, indicating its involvement in CDDO-Im-mediated G2/M arrest. The activation of Cdc2 (also named as CDK1) is the rate-limiting factor for cells to enter into mitosis, whereas its inactivation leads to G2/M arrest (15). Cdc2 is activated by phophorylation of Thr160 and the dephosphorylation of Thr14 and Tyr15. As shown in Fig. 3D, CDDO-Im and CDDO-Me–induced dephosphorylation of Cdc2 at Tyr15 in W0069 and W780 cells. Synthetic triterpenoids also induced the activation of G2/M checkpoint signaling in our new primary cell line B8701 (data not shown). Next, we determined the kinetics for this activation of the G2/M checkpoint by treatment with CDDO-Im for varying times in W780 cells. As short as a 2-hour treatment with CDDO-Im resulted in an increase in the phosphorylation of Chk1 protein (data not shown). Upregulation of p21 and phosphorylation of Chk2 were observed 4 hours after treatment with CDDO-Im during the time course (data not shown). Collectively, these results indicate that synthetic triterpenoids activate Chk1 and Chk2 kinases and block the G2/M phase transition.

To investigate the mechanisms of G2/M checkpoint signaling activation by CDDO-Im, the genotoxicity of this compound was investigated using the alkaline comet assay to measure DNA damage and Western blotting to monitor phosphorylation of the nuclear histone H2AX, a sensitive marker for breaks of double stranded DNA. As shown in Fig. 4A, BRCA1-deficient W780 cells showed greatly increased levels of DNA damage after 1 hour of exposure to CDDO-Im. However, CDDO-Im failed to induce DNA damage in normal mouse 3T3 fibroblasts. Interestingly, CDDO-Me did not induce DNA damage in W780 cells under identical conditions (data not shown). Tail Moment, a damage measurement method combining the amount of DNA in the tail with the distance of migration was used as the measure of DNA damage. We also focused on the phosphorylation on serine 139 of histone H2AX (γH2AX), which is an early molecular event in response to DNA damage and a marker of double strand breaks (16). As shown in Fig. 4B, induction of γH2AX was barely detectable in CDDO-Im-treated NIH3T3 cells whereas it was clearly induced by CDDO-Im in W780 cells (Fig. 4B). These observations suggest that CDDO-Im causes DNA damage that is recognized by the cell prior to triggering G2/M DNA damage response pathways.

To define the mechanisms contributing to the DNA damage caused by CDDO-Im, we first examined the effect of synthetic triterpenoids on intracellular ROS levels in W780 cells. As shown in Fig. 5A, exposure of cells to CDDO-Im (0.1–3 μmol/L) for 1 hour caused a significant concentration-dependent increase in ROS accumulation. In contrast, CDDO-Me did not cause any increase in cellular ROS levels. The ability of CDDO-Im to elevate intracellular ROS was also observed in 2 other BRCA1-mutated breast cancer cell lines, W0069 and B8701 (Fig. 5B) and in HCT1937 human BRCA1-mutated breast cancer cells (Supp. 2). Accumulating evidence suggests that, compared with their normal counterparts, many types of cancer cells have increased levels of ROS and a decreased capacity to eliminate ROS (17). Compared with NIH3T3 cells, W780 cells exhibited higher basal ROS content, as quantified by flow cytometry (Fig. 5C). Importantly, CDDO-Im caused ROS production in W780 cells, whereas there was no significant change in H₂DCFDA fluorescence after treatment with CDDO-Im in NIH3T3 cells. Although CDDO-Im also induced apoptosis of E18–14C-27 breast cancer cells, which were derived from transgenic mice overexpressing the erbB2 protein in the mammary gland but have wildtype BRCA1 expression, it did not induce ROS formation or G2/M arrest in these cells (Supp. 3). We also found similar effects on ROS generation with this triterpenoid in MCF-10A and MCF-10AT1 (18), human breast epithelial and Ha-ras transfected cells, respectively. MCF-10AT1 cells transformed with Ha-ras exhibited a significant increase of ROS after treatment with CDDO-Im compared with their parental MCF-10A cells (data not shown). However, CDDO-Im does not induce ROS in the MCF-10A cells (Supp. 4), which also express wildtype BRCA1. These results support previous reports showing that oncogenically transformed cells are more sensitive to ROS-mediated damage due to their high basal ROS generation (19). Transformation of ovarian epithelial cells with H-Ras and Bcr-Abl led to increased basal ROS content and elevated ROS generation in response to a ROS-inducing agent compared to their parental cells (19).

The above observations suggest that ROS might mediate the anticancer activity of CDDO-Im in BRCA1-mutated breast cancer cells. To further test whether CDDO-Im could induce ROS-mediated DNA damage in BRCA1-mutated cells, we treated W780 cells with uric acid, an antioxidant acting as an electron donor followed by CDDO-Im. Indeed, uric acid decreased ROS levels induced by CDDO-Im (Fig. 6A) and abrogated CDDO-Im-induced DNA damage (Fig. 6B) in W780 cells, suggesting DNA damage induced by this triterpenoid was a response to ROS stress. To determine whether Chk1 is also responsible for the accumulation of ROS, as well as G2/M arrest in CDDO-Im-treated W780 cells, we employed Chk1 siRNA to knock down the chk1 gene expression (Fig. 6C). Loss of Chk1 prevented the accumulation of ROS in response to CDDO-Im in W780 cells, showing that Chk1 is required for the CDDO-Im-induced increase in ROS levels.

Although CDDO-Imidazolide was first shown in 2003 (9) to inhibit DNA synthesis and to be useful for treatment of melanoma and leukemia in mice, this is the first report that any synthetic oleanane triterpenoid induces DNA damage and activates cell-cycle checkpoint signaling in BRCA1-mutated breast cancer cells. In this study, we have shown that CDDO-Imidazolide induced ROS accumulation, causing DNA damage, G2/M arrest, and finally apoptosis in these cancer cells. Cancer cells usually exhibit increased levels of intracellular ROS, which in turn can initiate a vicious cycle in which ROS induce gene mutations leading to further metabolic malfunction and ROS generation (17, 20). ROS cause oxidative damage to DNA, proteins, lipids, and other cellular components and therefore are a significant cellular stress (20). In line with this notion, a proposed therapeutic strategy against cancer is to treat cancer cells with pharmacological agents that have pro-oxidant properties, which increase the intracellular ROS generation to a toxic threshold that triggers cell death in the cancer cells without harming normal cells (17). This hypothesis is strengthened by the recent findings that ROS-generating agents showed selective toxicity in tumor cells with increased endogenous ROS, raising oxidative stress over the threshold of toxicity as antioxidant systems become overwhelmed both in vivo and in vitro (19, 21, 22).

In this study, we used a relatively high dose of CDDO-Im (1 μmol/L for 1 hour) compared with much lower doses used in previous reports that have focused on the antiinflammatory properties of the triterpenoids (23, 24). However, this higher dose is commonly used to induce apoptosis of cancer cells (25, 26). The purpose of this study is to show the ability of CDDO-Im to induce apoptosis through generation of ROS and DNA damage. As noted previously, low concentrations of triterpenoids are cytoprotective and antiinflammatory, whereas high concentrations are apoptotic (5, 24). These differences in activity seem to be correlated with increasing induction of oxidative stress as concentrations of triterpenoids are increased (5). These higher concentrations of triterpenoids are still relevant as we have reported concentrations of CDDO-Im in tissue of nearly 1 μmol/L (27), and tissue levels well above 1 μmol/L can be safely achieved in vivo with other triterpenoids with better pharmacokinetic properties.

Furthermore, CDDO-Im selectively increased intracellular ROS levels and subsequently induced apoptosis in mouse and human BRCA1-mutated breast cancer cells, but was unable to induce ROS production in nonmalignant MCF-10A breast epithelial cells, in NIH3T3 fibroblasts, or in E18–14C-27 cells; these 3 cell lines all express a wildtype BRCA1 gene. Moreover, CDDO-Im induced a significant increase in ROS in Ha-Ras transfected human breast epithelial cells but did not increase ROS in the parental MCF-10A cells. Therefore, CDDO-Im could be considered as a possible therapeutic compound to selectively cause ROS stress overload in cancer cells.

The G2/M checkpoint plays an important role in cellular response to genotoxic stimuli. The G2/M checkpoint prevents cells from entering mitosis when DNA is damaged, providing an opportunity for repair and stopping the proliferation of damaged cells, which helps to maintain genomic stability (28). As the signal transducer to mediate DNA damage signal to the downstream effectors, Chk1 is required for cell-cycle arrest in response to damaged DNA. Transfection with Chk1 siRNA revealed that Chk1 is essential to cell-cycle arrest in G2/M phase induced by CDDO-Im in W780 cells. This is consistent with the observations with UCN-01, a Chk1 inhibitor. Interestingly, knock down of chk1 gene expression using Chk1 siRNA prevented ROS generation in CDDO-Im–treated BRCA1-mutated breast cancer cells. This result supports the hypothesized vicious cycle of ROS production that, once initiated, leads to accumulation of gene mutations, which in turn exacerbates ROS generation. In this study, the high level of ROS induced by CDDO-Im in BRCA1-deficient cells caused DNA damage, which subsequently activated Chk1 and the G2 checkpoint. However, the signal to repair damaged DNA may not be transmitted to downstream effectors such as p53, cdc2, or cdc25 in Chk1 siRNA-transfected cells. Accordingly, cells without Chk1 would not recognize or restore DNA damage, and would not trigger a vicious cycle. In addition, it is reported that Chk1 is associated with chromatin, and that a phosphorylation event is likely to take place on chromatin when Chk1-containing complexes and the sensor complexes come together on DNA (29). Based on this observation, it is likely that Chk1 has a role in detection of DNA damage and monitoring genomic integrity.

Over the past few years, evidence has accumulated to support the role of synthetic triterpenoids in the induction of apoptosis in many types of cancer cells. In addition to their pivotal role in inhibiting cell proliferation and inducing apoptosis, CDDO and its derivatives, at high concentrations, show pro-oxidative properties that are strongly reflected by rapid oxidation of the intracellular GSH pool (25). The molecular links between the generation of ROS and apoptosis by the triterpenoids, however, remain incompletely defined. The generation of ROS induced by CDDO and CDDO-Im has been reported to result in apoptosis of leukemic (26) and multiple myeloma cells (30). Furthermore, several reports have shown that CDDO (31) and CDDO-Me (32), again at high concentrations, promoted the release of cytochrome c from mitochondria and inhibited mitochondrial electron transport via perturbations in inner mitochondrial membrane integrity, indicating that this organelle is a direct target of these agents. In addition, CDDO and CDDO-Im have been reported to form unregulated, constitutively open, permeability transition pores by the formation of high molecular weight protein aggregates through direct interaction with numerous mitochondrial protein thiols (33). Interestingly, different mechanisms have been proposed for the cytotoxic activity of CDDO-Im and CDDO-Me in various types of cells. CDDO-Me has been reported to induce apoptosis by down-regulation of the Akt pathway (34, 35), inhibition of IKK (36) and the mitochondrial pathway (26, 32, 37), whereas CDDO-Im has been reported to induce apoptosis in a caspase- (38) or ROS-dependent manner (26, 30).

Cancer chemoprevention has emerged as an effective and rational approach for modulating multi-stage carcinogenesis (39). An effective chemopreventive agent should intervene early in the process of carcinogenesis to suppress or eliminate premalignant cells before they become malignant. New evidence indicates that some chemopreventive agents can trigger apoptosis in transformed cells in vivo and in vitro, which seems to be associated with their effectiveness in controlling the process of carcinogenesis (40, 41). Our observations suggest that CDDO-Im, at high concentrations, induces ROS and subsequent DNA damage that is strengthened through a feedback loop, thereby facilitating the activation of the DNA damage checkpoint, G2/M arrest, and finally apoptosis in BRCA1-mutated cancer cells. BRCA1-mutated cells are defective in repairing their DNA damage, which makes the mutant cells more sensitive to apoptosis induced by CDDO-Im, whereas normal cells with normal BRCA1 alleles have a normal cell-cycle checkpoint, and they repair their DNA effectively. In this scenario, cells with a BRCA1 mutation selectively undergo apoptosis induced by CDDO-Im, whereas normal cells are spared from this action of the drug. In addition, because malignant cells have high levels of intrinsic ROS (17), CDDO-Im may be considered even further as a selective chemopreventive agent for BRCA1 mutation carriers. Our findings thus now indicate that there is a new rationale to use CDDO-Im as a chemopreventive agent. We are now exploring this possibility with new in vivo studies in mice bearing BRCA1 mutations.

M.B. Sporn: commercial research grant, Reata Pharmaceuticals; M.B. Sporn and K. Liby: patent interest in synthetic triterpenoids. The other authors disclosed no potential conflicts of interest.

These studies were supported by the Breast Cancer Research Foundation, the NIH (RO1 CA78814), and Reata Pharmaceuticals.