Raloxifene, a Mixed Estrogen Agonist/Antagonist, Induces Apoptosis through Cleavage of BAD in TSU-PR1 Human Cancer Cells*

Selective estrogen receptor modulator is a proven agent for chemoprevention and chemotherapy of cancer. Raloxifene, a mixed estrogen agonist/antagonist, was developed to prevent osteoporosis and potentially reduce the risk of breast cancer. In this study, we examined the effect of raloxifene on the TSU-PR1 cell line. This cell line was originally reported to be a prostate cancer cell line, but recently it has been shown to be a human bladder transitional cell carcinoma cell line. The TSU-PR1 cell line contains high levels of estrogen receptor β. Following treatment with raloxifene, evidence of apoptosis, including change in nuclear morphology, DNA fragmentation, and cytochrome c release, was observed in a dose-dependent manner in the TSU-PR1 cells (10−9 to 10−6 m range). We observed no detectable change in the steady-state levels of Bax, Bcl-2, and Bcl-XL following raloxifene treatment. However, raloxifene induced caspase-dependent cleavage of BAD to generate a 15-kDa truncated protein. Overexpression of a double mutant BAD resistant to caspase 3 cleavage blocked raloxifene-induced apoptosis. These results demonstrate that raloxifene induces apoptosis through the cleavage of BAD in TSU-PR1 cells. This molecular mechanism of apoptosis suggests that raloxifene may be a therapeutic agent for human bladder cancer.

Raloxifene is the prototypical selective estrogen receptor modulator (SERM) 1 that has been shown to prevent osteoporosis and breast cancer (1,2). Raloxifene binds to both ER␣ and ER␤ with high affinity (3,4); however, the binding affinity to ER␣ is four times higher than ER␤. Among the SERMs, raloxifene is unique in that it is an estrogen antagonist in the uterus (5). In the breast and bone, however, all SERMs act as estrogen antagonists and agonists, respectively (6). The mechanism for the observed tissue-specific effect of SERMs is currently unknown.
Two estrogen receptors (ERs) are known to mediate the physiological responses to estrogen, ER␣ (7) and ER␤ (8 -10). ER␣ and ER␤ have revealed some overlap of function, but also have significant differences in their ligand-binding and transcriptional properties (11,12). ER␣ and ER␤ exhibit both differential and overlapping tissue distribution. Whereas ER␣ has been predominantly found in the anterior pituitary, uterus, vagina, testis, liver, and kidney, ER␤ is predominantly expressed in thyroid, ovary, prostate, skin, bladder, lungs, gastrointestinal tract, cartilage, and bone (13,14). In the epithelium of the urinary bladder of adult rat, high levels of ER␤, but not ER␣, were detected (15), suggesting that ER␤ plays an important role in the bladder. The finding that older ER␤ knock-out mice exhibit prostate and bladder hyperplasia (16) also suggests that the estrogen receptor is a reasonable target for therapeutic intervention in prostate and bladder cancer patients.
Apoptotic stimuli activate caspases through mitochondriadependent and mitochondria-independent pathways. The Bcl-2 family proteins serve as critical regulators of mitochondria-dependent pathways. Antiapoptotic proteins (Bcl-2 and Bcl-X L ) reside in the mitochondria, whereas proapoptotic proteins (BAX, BAD, and BID) reside in the cytosol. Upon apoptotic stimuli, proapoptotic proteins translocate to the mitochondrial membrane, and mitochondria lose their membrane potential and release cytochrome c (17,18). In IL-3-dependent lymphoid cells, BAD is a key regulator of apoptosis (19). The function of BAD is regulated by reversible phosphorylation and binding of 14-3-3 proteins (20). Deprivation of survival factors induces BAD dephosphorylation by the specific serine/threonine phosphatase PP1␣ (21), resulting in dissociation of BAD from 14-3-3 proteins and translocation to the mitochondria, where BAD interacts with Bcl-X L and Bcl-2 and antagonizes their antiapoptotic functions. BAD is cleaved by a caspase(s) at its N terminus to generate a 15-kDa truncated protein following IL-3 deprivation-induced apoptosis in murine myeloid precursor 32Dcl3 cells. The 15-kDa truncated BAD is a more potent inducer of apoptosis than the wild-type BAD, whereas a mutant BAD, resistant to caspase 3 cleavage, is a weaker inducer of apoptosis (22).
In this study, we have investigated the effect of raloxifene on TSU-PR1, a cell line that was originally reported to be a human prostate cancer cell line but has been shown recently (23) to be a bladder cancer cell line. Raloxifene treatment induces apoptosis in these cells and induces caspase-dependent cleavage of BAD. A mutant BAD resistant to caspase 3 cleavage blocked raloxifene-induced apoptosis.

MATERIALS AND METHODS
Cell Culture-Human TSU-PR1 and PC3 M cells were maintained in RPMI 1640 containing 10% fetal bovine serum (FBS), penicillin (100 * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ¶ units/ml), and streptomycin (100 g/ml). Raloxifene, provided by Lily, was diluted to 10 Ϫ3 M in 70% ethanol and added to the culture medium at selected concentrations. The 293-derived Phoenix E (kind gift of Lisa Choy, University of California) cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated FBS. For experiments, cells were plated in RPMI 1640 supplemented with 10% FBS and allowed to adhere for 24 h. Then the cultures were washed two times with phosphate-buffered saline, and cells were cultured for 3 days in phenol red-free RPMI 1640 supplemented with either 1 or 10% charcoal-stripped FBS (cFBS) containing raloxifene at 10 Ϫ7 M in the presence or absence of estradiol. Raloxifene was added such that the ratio of 70% ethanol to the culture medium was 1:1000. As a control, 70% ethanol was added to culture at 1:1000. All experiments were repeated three times, and similar results were obtained each time.
Retrovirus Infection and Generation of Stable Cell Lines-To generate retroviruses, Phoenix E packaging cells were plated at 6 ϫ 10 5 cells/60-mm tissue culture dish and transfected with the pLXSP retroviral vector (empty or containing HA-WT-BAD, HA-DM56/61 BAD, or tBAD 68 cDNA), by the calcium phosphate method. Immediately after transfection, cells were treated with 25 M chloroquine to increase transfection efficiency. Six hours after transfection, the medium was replaced with fresh Dulbecco's modified Eagle's medium containing 10% FBS, and cells were grown for an additional 24 h. The conditioned medium containing recombinant retroviruses was collected and filtered through 0.45-m polysulfonic filters. Eight ml of these supernatants were applied immediately to TSU-PR1 cells, which had been plated 24 h before infection at a density of 1 ϫ 10 6 cells/100-mm tissue culture dish. Polybrene (Sigma) was added immediately to a final concentration of 8 g/ml. Twenty four h after infection, the cells were placed in fresh growth medium, and selection with 0.5 g/ml puromycin was initiated. After 14 days, individual clones picked from plates of the recombinant retrovirus-infected cells were transferred to 6-microtiter wells and expanded to generate cell clones stably expressing WT-type, double mutant (DM56/61), or truncated BAD.
RNA Isolation and RT-PCR-RT-PCR for ER␣ and ER␤ was carried out as described previously (24). Cells were harvested, and total RNA was isolated using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Once isolated, total RNA was reverse-transcribed using Superscript (Invitrogen) and random hexamer employing the following conditions: 42°C for 50 min and 70°C for 15 min. Following reverse transcription, the samples were incubated with RNase H for 30 min at 37°C. Subsequently, PCR amplification was performed as follows: 94°C for 1 min, 55°C for 1 min, and 72°C for 1 min for 35 cycles followed by 10 min of incubation at 72°C. The following primers were used: ER␣, 5Ј primer, tactgcatcagatccaaggg, and 3Ј primer, gtgggaatgatgaaaggtgg; ER␤, 5Ј primer, tgaaaaggaaggttagtgggaacc, and 3Ј primer, tggtcagggacatcatcatgg; glyceraldehyde-3-phosphate dehydrogenase, 5Ј primer, accacagtccatgccatcac, and 3Ј primer, tccaccaccctgttgctgta. To visualize the PCR products, the samples were subjected to eletrophoresis in 1% agarose gel followed by staining with ethidium bromide. The authenticity of the products was confirmed by Southern blot analysis.
TUNEL Assay-WT-BAD, DM56/61-BAD, and t-BAD TSU-PR1 cells were plated at 5 ϫ 10 4 cells/8-well chamber slide (Nalge Nunc, Rochester, NY) and incubated for 24 h prior to treatment with raloxifene. The medium was replaced with phenol red-free RPMI 1640 containing 1% charcoal-stripped FBS. The cells were treated with 10 Ϫ7 M raloxifene for 48 h and fixed with 4% paraformaldehyde, pH 7.4, for 10 min. Apoptotic cells were assessed by measuring DNA fragmentation in a standard deoxynucleotidyltransferase-mediated dUTP nick-end labeling (TUNEL) assay according to the instructions with the kit (In Situ Cell Death Detection Kit, POD, Roche Molecular Biochemicals).
Immunoblot Analysis-Whole-cell extracts were obtained in a 1% Triton X-100 lysis buffer (50 mM Tris-Cl, pH 8.0, 150 mM sodium chloride, 1 mM EDTA, 1 mM EGTA, 2.5 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 1 mM ␤-glycerophosphate, 1 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride). Western blotting was per- Assessment of Mitochondrial Transmembrane Potential-Changes in mitochondrial membrane potential were determined by staining cells with the fluorochrome, chloromethyl-X-rosamine (CMX-Ros) (Molecular Probes, Eugene, OR). One million cells were incubated with 100 nM CMX-Ros in growth medium for 30 min at 37°C in the dark and analyzed in a FACSCalibur flow cytometer (BD PharMingen). As a control, the protonophore carbonyl cyanide m-chlorophenylhydrazone (50 M; Sigma) was used. The fluorescence was excited with an argon laser (excitation wavelength 488 nm) and analyzed in FL-1 (wavelength 520 nm and PMT 437 V). A minimum 2 ϫ 10 4 events were acquired in list mode and analyzed with CELLQuest software (BD PharMingen).
Analysis of Mitochondrial Cytochrome c Release-WT-BAD and DM56/61-BAD TSU-PR1 cells were scraped off in isotonic isolation buffer (10 mM HEPES, 1 mM EDTA, 250 mM Sucrose, pH 7.6), collected by centrifugation at 2,500 ϫ g for 5 min at 4°C, and resuspended in hypotonic isolation buffer (10 mM HEPES, 1 mM EDTA, 50 mM sucrose, pH 7.6). Cells were disrupted by passing through a 27-gauge needle 5-10 times and checked for cracked cells by trypan blue staining. Hypertonic isolation buffer (10 mM HEPES, 1 mM EDTA, 450 mM sucrose, pH 7.6) was added to balance the tonicity of the buffer. Samples were centrifuged at 1,000 ϫ g (2,100 rpm) at 4°C for 10 min. Supernatants were recovered and centrifuged again at 100,000 ϫ g. The mitochondrial pellet proteins were extracted in isotonic isolation buffer, and the supernatant contained the cytosolic protein extract. Protein concentration of lysates was determined using the Bio-Rad protein assay kit (Bio-Rad) according to the manufacturer's instructions. After an electrophoretic separation of 50 g of protein/condition in SDS-16% polyacrylamide, gels were transferred by semidry transfer (Bio-Rad) to nitrocellulose membranes. Immunoblots were blocked in TBS-T (10 mM Tris/HCl, 150 mM NaCl, pH 7.5, 0.05% Tween 20) containing 5% non-fat dried milk and incubated overnight with the primary antibody (monoclonal anti-cytochrome c diluted 1:1000 in TBS-T 5% bovine serum albumin). After washing, membranes were incubated with peroxideconjugated anti-mouse immunoglobulin (1:3000 in TBS-T 0.5% non-fat dried milk) for 1 h, and the blot was developed with the ECL kit (Pierce).

Effect of Raloxifene on Proliferation of TSU-PR1 Cells-RT-
PCR and Western blot analysis were initially carried out to determine the status of ER expression in the androgen-independent human prostate cancer cell line PC3 M and human bladder cancer cell line TSU-PR1. Authenticity of the PCR products was confirmed by Southern blot analysis. The results demonstrated that only PC3 M cells expressed ER␣, whereas both cell lines were positive for ER␤ (Fig. 1, A and B). To determine the effect of raloxifene in these cell lines, cell numbers were determined after treatment with increasing doses of raloxifene in the presence of either 1% cFBS or 10% cFBS for 4 days. Raloxifene inhibited the proliferation of both PC3 M and TSU-PR1 cell lines in a dose-dependent manner under both serum concentrations (Fig. 1, C and D). At 1% cFBS, the minimum concentration of raloxifene needed to detect a significant decrease in cell number was 10 Ϫ9 M. The cell count at 10 Ϫ6 M raloxifene was ϳ25 and 1% of control in PC3 M and TSU-PR1 cells, respectively.
Raloxifene-induced Apoptosis and Cleavage of BAD in TSU-PR1 Cells-To investigate whether the raloxifene-induced cell death in human bladder cancer cell line was an apoptosis, TUNEL assay was carried out in TSU-PR1 cells. The following experiments were all carried out at 10 Ϫ7 M raloxifene in the presence of either 1 or 10% cFBS for 48 h because this concentration for 4 days induced near-complete cell death in TSU-PR1 cells. As indicated by the number of dark brown positive cells, there was a significant increase in the rate of apoptosis in the presence of raloxifene (Fig. 2, A and B). Raloxifene induced apoptosis in the presence of both 1 and 10% cFBS. To examine the effect of estradiol on the raloxifene-induced apoptosis, TSU-PR1 cells were incubated with raloxifene in the presence and absence of estradiol. As shown in Fig. 2, C and D, estradiol partially blocked the raloxifene-induced apoptosis.
To characterize the mechanism of raloxifene-mediated apo-ptosis, the steady-state levels of several Bcl-2 family proteins were measured by Western blot analysis. We detected neither significant change in the expression nor cleaved products of Bax, Bcl-2, or Bcl-X L over the raloxifene treatment time course (Fig. 3A). Because commercially available antibody against BAD protein does not recognize the human BAD protein, we generated stable cell lines expressing WT-BAD, DM56/61-BAD, and truncated BAD by infecting cells with retroviruses encoding the HA-tagged WT-BAD, DM56/61, or tBAD 68 in order to determine the involvement of BAD in raloxifene-induced apoptosis. Puromycin-resistant colonies were selected, and expression levels of HA-tagged WT-BAD, DM56/61 BAD, or t-BAD 68 were examined by Western blotting with an antibody raised against the HA epitope (Fig. 3B). Treatment of WT-BAD-expressing TSU-PR1 cells with raloxifene induced an ϳ15-kDa cleavage product of BAD as early as 24 h after raloxifene treatment, and cleavage was increased over the raloxifene treatment time course (Fig. 3C). Effect of Caspase Inhibitor and Cycloheximde on the Cleavage of BAD-It has been shown previously that treatment of 32Dc13 myeloid precursor cells with the general caspase inhibitor Z-VAD-fmk prevents the generation of the 15-kDa cleaved form of BAD (25). Therefore, we assessed the possible involvement of caspases in the cleavage of BAD in TSU-PR1 cells. Pretreatment of TSU-PR1 cells with Z-VAD-fmk abrogated the generation of the 15-kDa cleaved form of BAD (Fig. 4A), suggesting that caspases also play a role in BAD cleavage after raloxifene treatment of TSU-PR1 cells. To test whether the raloxifene-induced cleavage of BAD is dependent on new protein synthesis, WT-BAD TSU-PR1 cells were pretreated with the protein synthesis inhibitor, cycloheximide, for 1 h followed by treatment with raloxifene for 48 h. Pretreatment with cycloheximide inhibited the raloxifene-induced cleavage of BAD in a dose-dependent fashion (Fig. 4B), suggesting that a de novo protein synthesis is required for the generation of the 15-kDa cleaved form of BAD by raloxifene.
Effects of WT-BAD, DM56/61-BAD, and t-BAD on Raloxifeneinduced Apoptosis and Cleavage of BAD-To compare the effects of DM56/61 and truncated BAD on raloxifene-induced apoptosis, stable cell lines expressing WT-BAD, DM56/61-BAD, and t-BAD were treated with raloxifene for 48 h. The extent of raloxifene-induced apoptosis was estimated by the TUNEL assay. Overexpression of DM56/61-BAD reduced raloxifene-induced apoptosis, whereas overexpression of tBAD 68 increased in raloxifene-induced apoptosis, as compared with WT-BAD (Fig. 5, A and B). Also, overexpression of DM56/ 61-BAD blocked generation of the 15-kDa cleaved form of BAD by raloxifene (Fig. 5C).
Alteration of Mitochondrial Membrane Potential and the Release of Cytochrome c in Cells Expressing WT-BAD and DM56/61 BAD-Data suggest that t-BAD regulates mitochondrial membrane potential (25). We examined changes in the mitochondrial membrane potential of cells expressing WT-BAD or DM56/61 BAD using CMX-Ros as a fluorescent probe in flow cytometry. Increased CMX-Ros accumulation reflects an increase in mitochondrial membrane potential. Cells were incubated in the absence or presence of raloxifene for 48 h followed by staining with CMX-Ros for flow cytometry. CMX-Ros fluorescence was reduced in the vector control and WT-BADexpressing TSU-PR1 cells. However, these changes in mitochondrial membrane potential were not seen in DM56/61 BADexpressing TSU-PR1 cells (Fig. 6A). Based on these results, we examined whether the loss in mitochondrial membrane potential could be coincident with the release of cytochrome c. After incubation of the cells for 48 h in the absence or presence of 10 Ϫ7 M of raloxifene, mitochondria were separated from the cytosol, and cytochrome c content was analyzed by Western blot. In WT-BAD-expressing TSU-PR1 cells, cytochrome c content decreased markedly in the mitochondria, corresponding to increased levels in the cytosol. However, raloxifene-induced cytochrome c release from mitochondria to cytosol was markedly reduced in DM56/61 BAD-expressing TSU-PR1 cells (Fig. 6B).

DISCUSSION
In this study, data suggest that the mixed estrogen agonist/ antagonist raloxifene induces apoptosis in a dose-and time-dependent manner in ER␤-positive TSU-PR1 through the cleavage of BAD. The cleaved products of BAD, in turn, caused cytochrome c release in TSU-PR1 cells. Because TSU-PR1 cells have been demonstrated recently to be a human bladder transitional cell carcinoma cell line (23), these observations provide a new potential therapeutic target for human bladder cancer.
Raloxifene, an SERM that binds to both ER␣ and ER␤ with high affinity (3,4), is a mixed estrogen agonist/antagonist. Raloxifene is a safe agent for prevention of both osteoporosis and breast cancer (1,2). ER␤, first cloned in the rat prostate, is expressed in various human tissues. ER␣ is the predominant isoform in breast and uterine tissue, whereas ER␤ is expressed in significant quantities in the urogenital tract, the central nervous system, and endothelial cells. ER␣ and ER␤ exhibit differences in binding affinity and potency and might be differentially regulated in estrogen-sensitive tissues (26). In this study, we used TSU-PR1 cells to investigate the effect of raloxifene. TSU-PR1 cells were originally isolated as a human prostate cancer cell line, but recently van Bokhoven et al. (23) demonstrated the cell line to be a human bladder cancer cell line. We observed that TSU-PR1 expresses ER␤, not ER␣. Therefore, the results of the present study suggest that estrogen/estrogen receptors may be potential targets for therapeutic intervention in bladder cancer patients.
Bcl-2 family proteins are key regulators of apoptosis via heterodimerization between prosurvival Bcl-2 proteins and proapoptotic Bcl-2 proteins (27,28). Proapoptotic Bcl-2 proteins consist of multidomain members (Bax and Bak) and BH3 domain-only members (BAD, Bid, Bim, and Noxa). A BH-3 domain-only member interacts with a multidomain member up-stream of an adaptor and of caspases. BH-3 domain-only members require multidomain Bax or Bak to initiate cytochrome c, Apaf-1-driven caspase activation, and caspase-independent mitochondrial dysfunction (18,29). BAD is a BH3 only proapoptotic member that shares substantial sequence homology only within the BH3 amphipathic ␣-helical domain (28,31). The activity of BAD is regulated by changes in phosphorylation and subcellular localization. It is known that BAD is phosphorylated on one or more of three serine residues, Ser-112, Ser- 136, or Ser-155, in response to survival factors, and phosphorylation of these serines sequesters BAD in the cytosol, bound to 14-3-3 proteins (32-33, 35, 36). Recently, it was shown that protein phosphatase 2A (PP2A) or a PP2A-like phosphatase catalyzes BAD dephosphorylation and regulates its proapoptotic activity in IL-3-dependent lymphoid cells by a mechanism requiring dissociation from 14-3-3 (37). Upon increase in calcium influx or growth factor deprivation, phosphorylated BAD is rapidly dephosphorylated by the specific serine-phosphatase calcineurin (38) or PP1␣ (39) and translocates to the mitochondrial outer membrane where, through its BH3 domain, it interacts with antiapoptotic Bcl-2 and Bcl-X L (18,28,31).
Following IL-3 deprivation of 32Dc13 myeloid precursor cells, BAD is cleaved at its N terminus to generate two smaller products, one very similar in size to the full-length protein (26 kDa) and the second ϳ15 kDa. Transforming growth factor-␤1 also induces caspase-dependent cleavage of BAD at its N terminus to generate a 15-kDa truncated protein (40). These results suggest that cleavage of BAD in response to various apoptotic stimuli may be one of the major mechanisms in the process of apoptosis. In TSU-PR1 cells, raloxifene generates only the ϳ15-kDa cleaved form of BAD. In vitro studies have shown that BAD is cleaved by capases 2, 3, 7, 8, and 10 (22). In this study, generation of the ϳ15-kDa truncated form of BAD was blocked by the caspase inhibitor, Z-VAD-fmk, suggesting that it is the product of caspase activity. Cleavage of BAD enhances sensitivity to apoptosis because raloxifene induces cell death more rapidly in cells expressing truncated BAD than in cells overexpressing wild-type BAD, whereas raloxifene-induced cell death is reduced in cells expressing caspase-resistant mutant BAD.
Proapoptotic Bcl-2 family proteins, including BAD and Bak, induce changes in mitochondrial membrane permeability, often accompanied by cytochrome c release into the cytosol and activation of downstream caspases (28,39). In contrast, Bcl-X L inhibits cytochrome c release through a physical interaction with components of the voltage-dependent anion channel via a mechanism inhibited by several pro-apoptotic Bcl-2 proteins (30,41). Recently, it was shown that the transforming growth factor-␤-induced mitochondrial cytochrome c release is inhibited by Bcl-X L (34). Truncated BAD has a similar or higher affinity for Bcl-X L than wild-type BAD, and it is a more potent inducer of cytochrome c release than full-length wild-type BAD. It has been suggested that truncated BAD may be poorly phosphorylated at Ser-155 because phosphorylation of Ser-136 seems to be required for Ser-155 phosphorylation (22). This poor phosphorylation would enhance the interaction of truncated BAD with Bcl-X L . Therefore, increased generation of truncated BAD, induced by raloxifene, may result in enhanced cytochrome c release by antagonizing antiapoptotic Bcl-X L .
In conclusion, we have demonstrated that raloxifene, a mixed estrogen agonist/antagonist, induces apoptosis in TSU-PR1 cells. Cleavage of BAD, leading to the cytochrome c release, serves as a novel mechanism of raloxifene-induced apoptosis. These results provide valuable insight concerning the role of estrogen/estrogen receptor in bladder cancer cells. Because raloxifene has minimal side effects while effectively preventing osteoporosis and breast cancer, this study suggests raloxifene as a potential treatment in patients with advanced bladder cancer.