Bcl-2 Oncoprotein Protects the Human Prostatic Carcinoma Cell Line PC3 from TRAIL-Mediated Apoptosis

Bcl-2 oncoprotein protects the human prostatic carcinoma cell line PC3 from TRAIL-mediated apoptosis

Oskar W Rokhlin1, Natalia Guseva1, Agshin Tagiyev1, C Michael Knudson1 and Michael B Cohen*,1,2,3,4

1Department of Pathology, The University of Iowa, Iowa, IA 52242, USA; 2Department of Urology, The University of Iowa, Iowa, IA 52242, USA; 3Department of Epidemiology, The University of Iowa, Iowa, IA 52242, USA; 4Veterans Aairs Medical Center, Iowa City, IA 52242, USA

The role of Bcl-2 in TRAIL-induced apoptosis has been of FADD (Fas-associated death domain protein) and investigated in lymphoid cells. Here we show that the caspase-8. Moreover, it has been shown that FADD human prostatic carcinoma cell line PC3 was sensitive to and caspase-8 were recruited to TRAIL-R1 and -R2 TRAIL treatment whereas PC3 overexpressing of Bcl-2 independently of each other (Sprick et al., 2000). It has was resistant. TRAIL receptors ligation in PC3 also been shown that TRAIL recruits endogeneous activated caspases -2, -3, -7, -8, and -9, induced Bid FADD and caspase-8 to TRAIL-R1 and -R2 and can processing, dissipation of mitochondrial transmembrane form both homomeric and heteromeric complexes with potential (DCm), and cytochrome c release. We have TRAIL-R1 and -R2 (Kischkel et al., 2000). From these detected caspases -8 and -3 only in the cytosolic fraction studies it was concluded that TRAIL utilizes a of cells, but caspases -2, -7, and -9 were found both in pathway similar to the one used by the Fas receptor. cytosolic and mitochondrial fractions. Bcl-2 overexpres- The biochemical pathways of TRAIL-induced apop- sion did not aect caspase-8 activation although it did tosis have mainly been investigated in lymphoid cells. change the processing pattern of caspase-3. At the same TRAIL-mediated apoptosis has been shown to involve time, Bcl-2 overexpression inhibited the activation of activation of caspases -8 and -3, the late dissipation of mitochondrial localized caspases -2, -7, and -9. Bcl-2 the mitochondrial transmembrane potential (DCm)and also abrogated TRAIL-induced cytochrome c release and cytochrome c release. Overexpression of Bcl-2 in the dissipation of DCm. These ®ndings suggest that TRAIL- human lymphoid cell line Jurkat did not protect induced apoptosis in the epithelial cell line PC3 depends against TRAIL-induced apoptosis. However, apoptosis both on mitochondrial integrity and caspase activation. induced by etoposide was decreased in these cells. Oncogene (2001) 20, 2836 ± 2843. These authors concluded that TRAIL induces apoptosis via a caspase cascade independent of the proapop- Keywords: Bcl-2; prostatic neoplasms; TRAIL; apoptotic machinery of mitochondria (Walczak et al., 2000). tosis Similar results were found for lymphoid cell line CEM overexpressing Bcl-2 (Keogh et al., 2000). Based on these and other data (Bonavida et al., 1999) it was Introduction suggested that TRAIL could be used as an antitumor agent in Bcl-2 overexpressing tumors since TRAIL can TNF-related apoptosis-inducing ligand (TRAIL or bypass the antiapoptotic eect of Bcl-2 overexpression Apo2 ligand) was discovered as an apoptosis-inducing (Walczak and Krammer, 2000). member of the TNF gene superfamily (Willey et al., Here we have characterized the biochemical pathway 1995; Pitti et al., 1996). TRAIL can interact with four involved in TRAIL-induced apoptosis of prostatic cellular receptors but only two of them, DR4 (also carcinoma cells. The human prostatic carcinoma cell called TRAIL-R1) and DR5 (TRAIL-R2, TRICK2, or line PC3 is sensitive to Fas-mediated apoptosis, which KILLER) contain cytoplasmic death domains (DDs) occurs via activation of caspases -8 and -7, and and signal apoptosis (Pan et al., 1997a,b; Screaton et involves cytochrome c release (Rokhlin et al., 1997, al., 1997; Sheridan et al., 1997; Walczak et al., 1997). 1998; Gewies et al., 2000). Here we show that PC3 is Analysis of the native TRAIL death-inducing signaling sensitive to treatment with TRAIL that activates complex (DISC) revealed ligand-dependent recruitment initiator caspases -2, -8, and -9, as well as executioner caspases -3 and -7, induces dissipation of DCm and caspase-dependent cytochrome c release. Bcl-2 overexpression did not aect caspase-8 activation, although *Correspondence: MB Cohen, Department of Pathology, The it did change the processing pattern of caspase-3. At University of Iowa, 200 Hawkins Drive, 1117 ML Iowa City, IA the same time, Bcl-2 overexpression inhibited the 52242-1087, USA Received 14 December 2000; revised 9 February 2001; accepted 20 activation of mitochondrial localized caspases -2, -7, February 2001 and -9. Bcl-2 also abrogated TRAIL-induced dissipa- Bcl-2 dependent apoptosis in prostate cancer OW Rokhlin et al 2837 tion of DCm and cytochrome c release. These data suggest that, in contrast to lymphoid cells, TRAIL- induced apoptosis in PC3 is dependent on mitochondrial integrity rather than on caspase activation.

Results

TRAIL-induced apoptosis is Bcl-2-dependent We have previously shown that PC3 is sensitive to treatment with agonistic anti-Fas monoclonal antibodies and that the mitochondrial pathway is involved in Fas-mediated apoptosis (Rokhlin et al., 1997, 1998; Gewies et al., 2000). To determine the role of the mitochondrial pathway in TRAIL-mediated apoptosis, we overexpressed Bcl-2 in PC3 and then compared the pattern of apoptosis using PC3-Hygro (mock transfec- tant) and PC3-Bcl-2. As shown in Figure 1a, expression of Bcl-2 in PC3-Bcl-2 was found to be much higher compared to PC3-Hygro, especially in mitochondria. We then compared cell viability of these cell lines (Figure 1b) and found that PC3-Bcl-2 was signi®cantly more resistant to TRAIL-mediated cell death compared to PC3-Hygro. As can be seen from Figure 1b, 100 ng/ml of TRAIL killed more than 90% of PC3-Hygro cells after 24 h of treatment whereas only 25% of PC3-Bcl-2 cells were killed even by 1000 ng/ml of TRAIL. As expected, cell death of PC3 was found to be caspase-dependent since the pan- caspase inhibitor Z-VAD-fmk completely prevented cell death under TRAIL treatment (Figure 1b). Cell death following TRAIL treatment is apoptotic based on the formation of a DNA ladder within 4 h of TRAIL treatment (data not shown) and the proteolysis of known caspase substrates (Utz and Anderson, 2000). Both PARP and DFF45 are proteolyzed following TRAIL treatment and this can be blocked by Bcl-2 (Figure 2).

Bcl-2 overexpression does not affect caspase-8 activation Figure 1 (a) Expression of Bcl-2 in PC3-Hygro and PC3-Bcl-2 but changes caspase-3 processing transfectants. Expression of Bcl-2 was investigated by Western blot analysis. To analyse total cytosolic proteins, cells were It has been reported that the initiator caspase-8 plays a harvested and lysed in 1% Triton X-100 buer. Mitochondria crucial role in Fas- and TRAIL-mediated apoptosis were isolated in mito-buer as described in `Materials and (Walczak and Krammer, 2000) and that the execu- methods'. Twenty mg of cytosolic proteins and 5 mgof tioner caspase-3 is responsible for proteolysis of mitochondrial proteins were separated on 4 ± 20% gradient multiple targets during apoptosis induced by dierent SDS ± PAGE, blotted to nitrocellulose membrane, blocked with 5% nonfat dry milk in PBS containing 0.1% Tween-20 and then agents (Utz and Anderson, 2000). Surprisingly, we did incubated with mouse anti-Bcl-2 antibody. The blots were not ®nd signi®cant dierences between PC3-Hygro and counterstained with antimouse IgG conjugated with HRP. The PC3-Bcl-2 in the activation pattern of caspase-8 after immunoreactive bands were visualized by incubation of the treatment with TRAIL. As can be seen from Figure 3, membrane with an enhanced chemiluminescence reagent. (b) PC3-Bcl-2 is resistant to treatment with TRAIL. Cells were caspase-8 was already activated after 20 min of plated at a density of 7000 cells/well in 96-well tissue culture ¯at- treatment with TRAIL and to the same extent in both bottom plates; TRAIL (at the indicated concentrations) was cell lines. However, caspase-3 was found to be added at the time of plating and cell death was estimated 24 h underprocessed in PC3-Bcl-2; the weak activation band later by calcein assay as described in `Materials and methods'. was observed after 1 h of treatment with TRAIL and Each point represents mean values of seven separate experiments. In two experiments, cell death was estimated in PC3-Hygro in the this band disappeared after 4 h whereas in PC3-Hygro presence of 10 mM of the pan-caspase inhibitor Z-VAD-fmk after the caspase-3 activation band was detected at the same treatment with 100 ng/ml of TRAIL (shown by column). Cell level as after 1 h of treatment. We have also examined death was also estimated by crystal violet staining (data not the presence of caspases -8 and -3 in mitochondria and shown), which gave similar results

Oncogene Bcl-2 dependent apoptosis in prostate cancer OW Rokhlin et al 2838

Figure 2 PARP and DFF45 proteolysis is prevented in PC3-Bcl- 2 during TRAIL-mediated apoptosis. Cells were treated with TRAIL (500 ng/ml) for 2 and 4 h. To isolate nuclear and cytosolic proteins, cells were gently lysed in a Dounce homogenizer with Te¯on pestle in ice-cold Mito-buer containing 250 mM of sucrose. Homogenates were spun at 16 000 g for Figure 4 Caspase activity is lower in PC3-Bcl-2 compared to 20 min at 48C and supernatants were used as mitochondria-free PC3-Hygro after treatment with TRAIL. Caspase activity was cytosol for assessment of DFF45. Pellets containing nuclei were estimated with Ac-DEVD-AMC, a substrate that is cleaved by layered over 30% sucrose in Nuclei buer and pelleted by several caspases. Cells were treated with TRAIL (500 ng/ml) for 1 centrifugation at 800 g for 10 min. Nuclei were lysed in sample and 4 h, harvested, lysed in KPM buer and assayed as described buer for assessment of PARP. Twenty mg of nuclear or cytosolic in `Materials and methods'. Measurements were calibrated against proteins were separated on 4 ± 20% gradient SDS ± PAGE and a standard curve of 7-amino-4-methylcoumarine (AMC), and expression of PARP and DFF45 was examined by Western blot data were expressed in nM of released AMC per mg of proteins analysis

Hygro was four times higher than in PC3-Bcl-2. These data suggest that the higher molecular weight form of processed caspase-3 is inactive and only the fully processed form is active.

Caspases -2, -7, and -9 reside both in the cytosol and mitochondria and their activation is inhibited by Bcl-2 overexpression We next investigated activation of caspases -2, -7, and -9 under TRAIL treatment of PC3-Hygro and PC3-Bcl-2 (Figure 5) and found distinct patterns of expression and activation of these caspases compared to caspases -8 and -3. First, caspases -2, -7, and -9 were found both in the cytosol and in mitochondria. An additional dierence between these two groups of caspases (caspases -8 and -3 versus caspases -2, -7, and -9) is the time course of activation. Caspase-2 was activated in the cytosol of PC3-Hygro after 1 h of treatment and activated in mitochondria after 4 h of treatment, although the level of proenzyme was decreased in mitochondria after 1 h of treatment with TRAIL Figure 3 Bcl-2 overexpression does not aect the activation of caspase-8 but results in underprocessing of caspase-3. Cells were (Figure 5). At the same time, caspase-2 activation in treated with TRAIL (500 ng/ml) for indicated times and PC3-Bcl-2 was found to be dramatically lower in the expression of caspases -8 and -3 was examined by Western blot cytosol and was not activated in mitochondria. The analysis as described in legend to Figure 1. Both caspases were activation pattern between PC3-Hygro and PC3-Bcl-2 not found in mitochondria (data not shown). The arrow indicates the activated p20 subunit of caspases -8 and -3 with caspase-7 showed only qualitative dierences. Caspase-7 was activated in the cytosol of PC3-Hygro after 4 h of treatment with TRAIL and activated after 1 h of treatment in mitochondria. However, we did not found that these caspases did not reside in mitochon- ®nd caspase-7 activation in PC3-Bcl-2 either in the dria (data not shown). We then estimated caspase cytosol or mitochondria. The dierences between PC3- activity using Ac-DEVD-AMC as a substrate that is Hygro and PC3-Bcl-2 of caspase-9 activation resembled cleaved by dierent caspases. As shown in Figure 4, caspase-2 expression: there were quantitative dierences caspase activity was found to be slightly lower in PC3- in the cytosol and qualitative dierences in mitochon- Bcl-2 compared to PC3-Hygro after 1 h of treatment dria. Taken together, these data clearly indicate that with TRAIL, but after 4 h caspase activity in PC3- activation of caspases -2, -7, and -9 are Bcl-2-dependent.

Oncogene Bcl-2 dependent apoptosis in prostate cancer OW Rokhlin et al 2839

Figure 5 Bcl-2 inhibits the activation of caspases -9, -7, and -2 both in the cytosol and mitochondria. Cells were treated with TRAIL (500 ng/ml) for indicated times; 20 mg of cytosolic and 5 mg of mitochondrial proteins were separated on 5 ± 20% SDS ± PAGE and expression of caspases -9, -7 and -2 was examined by Western blot analysis

TRAIL-induced mitochondrial events are Bcl-2-dependent FADD-DN inhibits TRAIL-mediated apoptosis We subsequently investigated whether Bcl-2 regulates Based on the results described above, it became evident TRAIL-induced apoptosis by blocking dissipation of that Bcl-2 can protect PC3 cell death under TRAIL DCm and cytochrome c release from mitochondria. treatment without inhibiting caspase-8. This observation Dissipation of DCm as measure by DiOC6 ¯uorescence may be interpreted to mean that activation of caspase-8 was prevented by Bcl-2 (Figure 6c). In addition, Bcl-2 is not sucient for apoptosis following TRAIL completely blocked cytochrome c release in PC3-Bcl-2 treatment. Alternatively, it may indicate that the (Figure 6a). It is noteworthy that cytochrome c release downstream eect(s) of this caspase could be mediated was already observed after 20 min of TRAIL treat- (ampli®ed) by mitochondrial pathways and that an ment in sensitive cells. This indicates that cytochrome c ampli®cation loop and cross-talk between cytosolic release from mitochondria into the cytosol is one of caspases (-8 and -3) and mitochondrial caspases (-2, -7, the early events in TRAIL-induced apoptosis in PC3. -9) can occur via Bid. To test this hypothesis, we In separate experiments, we have found that the pan- overexpressed FADD-DN in PC3. Figure 8 shows that caspase inhibitor Z-VAD-fmk inhibited cytochrome c PC3-FADD-DN was resistant to TRAIL-mediated cell release (data not shown) and therefore conclude that death. We then examined the activation of caspases and TRAIL-induced cytochrome c release is caspase- proteolysis of Bid, DFF45 and PARP. As can be seen dependent. We have also investigated Bid processing, from Figure 9, FADD-DN prevented processing of the Bcl-2 family member that has been described as caspase-8 and all other caspases tested as well as being cleaved by caspase-8 and then triggering cleavage of Bid, DFF45 and PARP. These data indicate cytochrome c release (Li et al., 1998; Luo et al., that activation of caspase-8 is necessary but not 1998; Yamada et al., 1999). As shown in Figure 7, Bid sucient in TRAIL-induced apoptosis, and cell death processing is caspase-dependent and occurred at can occur only when the mitochondrial pathway similar levels in PC3-Hygro and PC3-Bcl-2. These remains intact. Overexpression of Bcl-2 blocks the data are consistent with the pattern of caspase-8 mitochondria presumably by overriding the eect of activation which has been found to be the same in Bid and prevents cell death under TRAIL treatment. both cell lines. It appears that Bcl-2 acts downstream In summary, TRAIL can not bypass the mitochon- of Bid and Bcl-2 overexpression blocks the activity of drial pathway in the human prostatic carcinoma cell processed Bid and prevents disruption of DCm and line PC3 as has been observed in lymphoid cells cytochrome c release. (Walczak et al., 2000; Keogh et al., 2000).

Oncogene Bcl-2 dependent apoptosis in prostate cancer OW Rokhlin et al 2840

Figure 7 Bid processing is similar in PC3-Hygro and PC3-Bcl-2. Cells were treated with TRAIL (500 ng/ml) for indicated times, and mitochondria-free cytosol was prepared as described in `Materials and methods'. The eect of the pan-caspase inhibitor Z-VAD-fmk (100 mM) on the processing of Bid was also examined. Twenty mg of cytosolic protein were separated by 4 ± 20% SDS ± PAGE, and Bid expression was examined by Western blot analysis

Figure 6 (a,b) Bcl-2 prevents TRAIL-dependent cytochrome c release from mitochondria into the cytosol. PC3-Hygro and PC3- Bcl-2 were treated with TRAIL (500 ng/ml) for 20 min, 1 and 4 h. Cells were gently lysed in a Dounce homogenizer with a Te¯on Figure 8 PC3-FADD-DN is resistant to treatment with TRAIL. Cells were plated at a density of 7000 cells/well in 96-well tissue pestle in ice-cold Mito-buer containing 250 mM of sucrose, and homogenates were spun at 800 g for 20 min at 48C. The culture ¯at-bottom plates; TRAIL at indicated concentrations was supernatants were then spun at 16 000 g for 20 min and used as added at the time of plating and cell death was estimated 48 h mitochondria-free cytosol. Pellets containing mitochondria were later by calcein assay as described in `Materials and methods'. washed twice with Mito-buer and were lysed by TNC buer. Each point represents mean values of six replicates in one of three Twenty mg of cytosolic and 5 mg of mitochondrial proteins were separate experiments, which gave similar results separated on 4 ± 20% SDS ± PAGE and expression of cytochrome c (a) and cytochrome oxidase (subunit II) (b) was examined by Western blot analysis. The absence of cytochrome oxidase in activation, and apoptosis is dependent of mitochon- cytosol indicated that cytosol was not contaminated by mitochondria. (c) Disruption of mitochondrial transmembrane drial activity. Thus, only in type II cells does activation potential (DCm) after treatment with TRAIL occurs in PC3- of caspases -8 and -3 occur at a level downstream of Hygro but not in PC3-Bcl-2. Cells were treated for indicated mitochondria, depend on mitochondrial release of periods of times with TRAIL (100 ng/ml), then incubated cytochrome c and apoptosis is inhibited by Bcl-2 for 45 min at 378C with 3,3'-dihexyloxacarbocyanine iodide overexpression. Since it has been shown that TRAIL (DiOC6(3)) (40 nM), washed twice with PBS and live cells were analysed in a FACS Vantage cyto¯uorometer (Becton Dickinson). recruits FADD and caspase-8 to TRAIL-R1 and -R2 it The percentage of cells with low DCm is shown is presumed that it utilizes a pathway similar to the one used by Fas receptor (Sprick et al., 2000; Kischkel et al., 2000), and it is reasonable therefore to apply the Discussion idea of type I and II cells to TRAIL-mediated apoptosis. Two dierent cell types have been identi®ed by We noted caspase-8 activation after 20 min of investigating CD95 (Fas receptor) signaling pathways treatment with TRAIL and caspase-3 activation after (Scadi et al., 1998; Walczak and Krammer, 2000). In 1 h of treatment in PC3 cells. We also noted caspase- type I cells, activation of caspases -8 and -3 occurs dependent cytochrome c release from mitochondria prior to the loss of DCm, Bcl-2 overexpression does not after 20 min of treatment. At the same time, the loss of block caspases -8 and -3 activation, and apoptosis is DCm was observed after 2 h of treatment with TRAIL. independent of mitochondrial activity. In type II cells, These data indicate that activation of caspases -8 and activation of caspases -8 and -3 occurs after the loss of -3 occurs prior to the loss of DCm which is DCm, Bcl-2 overexpression blocks caspases -8 and -3 characteristic of type I cells. We also compared the

Oncogene Bcl-2 dependent apoptosis in prostate cancer OW Rokhlin et al 2841

Figure 9 FADD-DN prevents activation of caspases -2, -3, -7, -8, -9, inhibited Bid processing and proteolysis of DFF45 and PARP. PC3-mock (#1) and PC3-FADD-DN (#2) were treated with TRAIL (500 ng/ml) for 2 h. Expression of proteins was examined by Western blot analysis as described in legend to Figure 1 activation pattern of caspase-8 between control PC3 Apaf-1 and procaspase-9, resulting in the activation (PC3-Hygro) and PC3 overexpressing Bcl-2 and did of caspase-9 (Green and Reed, 1998). Activated not ®nd any dierences between them indicating that caspase-9 can in turn activate executioner caspases activation of caspase-8 is Bcl-2 independent; this is (-3, -6, -7) which result in a positive feedback between another feature of type I cells. However, in the case of caspase activation and cytochrome c release from PC3-Bcl-2, caspase-3 was found to be underprocessed mitochondria (Green and Reed, 1998; Green and after 4 h of treatment with TRAIL. The pattern of Kroemer, 1998). Bcl-2 levels are known to be elevated activation of caspases -2, -7, and -9 cannot be in a broad range of tumors, indicating that Bcl-2 likely attributed to type I cells. First, in contrast to plays a role in resistance to chemotherapeutic treat- caspase-8, these caspases were found not only in the ment and provides a rationale for TRAIL to be used cytosol but also in mitochondria. Second, overexpres- for treatment of chemoresistant tumors (Nicholson, sion of Bcl-2 blocked activation of caspases -2, -7, and 2000). -9 as well as the loss of DCm and cytochrome c. Since In contrast to ®ndings in lymphoid cells, our study in PC3-Bcl-2 activation of caspases -2, -7, and -9 was clearly shows that the human prostatic carcinoma cell blocked, it is conceivable that one of these caspases or line PC3 was eectively protected from TRAIL- a combination of some of them is required to fully mediated apoptosis by overexpression of Bcl-2. The process caspase-3. These data suggest that caspase-8 importance of mitochondrial localization of caspases alone cannot activate caspase-3 to full extent and as a was observed during TRAIL-mediated apoptosis in result PARP and DFF45 were not cleaved in PC3-Bcl- PC3 and it is conceivable that control of apoptosis by 2. Therefore, TRAIL-mediated apoptosis in PC3 is Bcl-2 can occur by preventing activation of mitochon- dependent on the disruption of mitochondria with drial localized caspases -2, -7, and -9. We observed that release of cytochrome c and activation of caspase-9. overexpression of Bcl-2 almost completely inhibited the This pathway is blocked by Bcl-2 which appears to act activation of these caspases and resulted in under- biochemically downstream of the processing of cas- processing of caspase-3 whereas activation of caspase-8 pase-8. Consistent with this interpretation, Bid was was unaected. These data indicate that activation of found to be cleaved in PC3 and this could not be caspase-8 is insucient to complete apoptosis without blocked by Bcl-2. Thus, apoptosis in PC3 following involvement of mitochondrial localized caspases. We TRAIL treatment has characteristics of both type I also found caspases -2, -7, and -9 in mitochondria both and type II pathways. in untreated and apoptotic cells and therefore we Mitochondrial localization of caspases has pre- cannot estimate whether translocation of caspases from viously been found in dierent cell types (Susin et al., the cytosol to mitochondria or vice versa occurs during 1999; Zhivotovsky et al., 1999; Mancini et al., 1998, apoptosis. Using PC3-FADD-DN, we have also 2000) and it was concluded that caspases with over- observed that activation of caspase-8 is necessary but lapping substrate speci®city may cleave speci®c sub- not sucient for TRAIL-mediated apoptosis since Bcl- strates in dierent cellular compartments (Chandler et 2 blocks the mitochondria by overriding the eect of al., 1998). The role of Bcl-2 in TRAIL-mediated Bid and prevents cell death under TRAIL treatment. apoptosis has been mainly investigated in lymphoid Therefore, TRAIL cannot bypass the mitochondrial cells (Walczak et al., 2000; Keogh et al., 2000). Bcl-2, a pathway in PC3 as it has been observed in lymphoid mammalian homologue of CED-9 in C. elegans, cells. prevents cell death induced by a number of dierent While it has been suggested (Walczak et al., 2000; agents (reviewed by Yang and Korsmeyer, 1996). Bcl-2 Keogh et al., 2000; Bonavida et al., 1999; Walczak and was found to be associated with the outer (Bay et al., Krammer, 2000) that TRAIL can be used as an 1993) as well as with the inner (Gotow et al., 2000) antitumor agent against Bcl-2-overexpressing lymphoid mitochondrial membranes. Bcl-2 is considered to be a tumors, this conclusion might only be valid for multifunctional protein (Reed, 1997) and one of its lymphoid tumors. It does not appear to be a general important anti-apoptotic functions is to block the loss rule. Our study clearly shows that the epithelial cell line of mitochondrial membrane potential (Shimizu et al., PC3 was eectively protected from TRAIL-mediated 1996) and the release of cytochrome c into the cytosol apoptosis by Bcl-2 overexpression. Other appropriate (Zamzami et al., 1996). Cytochrome c released from therapeutic treatment need to be established for solid mitochondria forms a complex (apoptosome) with tumors that overexpress Bcl-2.

Oncogene Bcl-2 dependent apoptosis in prostate cancer OW Rokhlin et al 2842 Materials and methods freezing and thawing of cells in KPM buer (50 mM PIPES, pH 7.0, 50 mM KCl, 1 mM EGTA, 2 mM MgCl2,1mM DTT, Cell lines and transfection 10 mg/ml cytochalasin B, 0.1 mM PMSF, and 2 mg/ml pepstatin, leupeptin, and antipain). Forty mg of protein The human prostatic carcinoma cell line PC3 was cultured in lysate was incubated at room temperature for 30 min in assay RPMI 1640 as has been previously described (Rokhlin et al., buer (20 m PIPES, pH 7.2, 100 mM NaCl, 10 mM DTT, 1997). PC3 was transfected with the plasmid pSFFV-Hygro M 1m EDTA, 0.1% CHAPS and 10% sucrose) with 20 mM with or without cDNA encoding Bcl-2 followed by M ¯uorescent substrate-Ac-DEVD-AMC. Fluorescence at 360/ hygromycin selection. A dominant-negative form of FADD 460 nm was measured with a FL600 ¯uorimeter. Measure- (FADD-DN) was made with PCR primers (5'-CCAAGCT- ments were calibrated against a standard curve of 7-amino-4- TATGGACGACTTCGAGGCGGGG and 5'-AAGGATC- methilcoumarin (AMC) (Sigma) and data were expressed in CACCAGCGCAAAGCAGCGGCC) ¯anking the region n of released AMC per mg of cytoplasmic extract proteins. encoding 117 ± 208 aa and using the FADD cDNA as M Western blot detection of proteins was performed as template. This region contains the DD but does not contain described previously (Rokhlin et al., 1997). Brie¯y, 5 ± 20 mg the death-eector domain (DED) (Chinnaiyan et al., 1995; of proteins were separated on 4 ± 20% gradient SDS ± PAGE, Peter et al., 1996). The HindIII/BamHI restriction sites were and blotted to nitrocellulose membrane (Novex, San Diego, used for cloning FADD-DN into the expression vector CA, USA). Equal loading was controlled routinely by pcDNA3.1/Hygro (Invitrogen, Carlsbad, CA, USA). Gene- reversible staining of the membrane with Ponceau S solution PORTER reagent kit (Gene Therapy Systems) was used for (Sigma, St. Louis, MO, USA). Membranes were blocked with transfection and cells were selected in the presence of 5% nonfat dry milk in PBS containing 0.1% Tween-20 and hygromycin. Empty vector was used to obtain control cells then incubated with the corresponding mouse monoclonal or resistant to hygromycin (PC3-mock). Human TRAIL was rabbit polyclonal antibodies: anti-cytochrome c, anti-PARP, purchased from PeptoTech (Rocky Hill, NJ, USA). anti-Bcl-2 (Pharmingen, San Diego, CA, USA), cytochrome oxidase (subunit II) (Molecular probes), anti-caspase 3 Estimation of cells viability (Transduction Laboratories, San Diego, CA, USA), anti- caspase-2, anti-Bid (R&D Systems, Minneapolis, MN, USA), To measure cell viability, we used calcein AM (Molecular anti-DFF45, anti-caspase-8, anti-FADD (Upstate, Lake Probes, Eugene, OR, USA). Calcein AM is a ¯uorogenic Placid, NY, USA), anti-caspase-7 and anti-caspase-9 (Onco- esterase substrate that is hydrolyzed in live cells to a green gene, Uniondale, NY, USA). The blots were counterstained ¯uorescent product (calcein) and ¯uorescence at 485/530 nm with goat antimouse or antirabbit IgG conjugated with HRP was measured with a FL600 ¯uorimeter (Bio-Tek Instru- (Pierce, Rockford, IL, USA). The immunoreactive bands ments, Inc., Burlington, VT, USA). Cells were incubated and were visualized by incubation of the membrane with treated in 96-well ¯at-bottomed plates. After incubation, enhanced chemiluminescence reagent (Pierce). medium was removed, the plates were washed with PBS, and incubated with 100 mlof2mg/ml calcein AM solution for 30 min at room temperature. Fluorescence, which is propor- Detection of cytochrome c release and disruption of the tional to cell viability, was then measured with FL600 mitochondrial transmembrane potential (DCm) ¯uorimeter. Cell viability was also measured by crystal violet To detect cytochrome c release, mitochondria-free cytosol staining. In this case, cells (30 000/well/2 ml) were incubated was prepared as described (Yang et al., 1997). Brie¯y, cells and treated in 24-well plates. After treatment, 1 ml medium were lysed in ice cold Mito-buer (20 m HEPES, pH 7.5, was removed and 750 ml of crystal violet solution (0.1% M 10 m KCl, 1.5 mM MgCl ,1mM EGTA, 1 mM EDTA, crystal violet, 2.1% citric acid in water) was added for 60 min M 2 1mM DTT, 250 mM sucrose, 0.1 mM PMSF, 2 mg/ml at room temperature. Cells were washed three times with pepstatin, leupeptin, and aprotinin) by homogenization in a water and SDS solution (1%) was added to each well. small glass homogenizer with te¯on pestle (50 strokes on ice). Absorbance was measured using an ELISA plate reader at The homogenates were ®rst spun at 800 g to remove nuclei 630 nm. In both assays, TRAIL in various concentrations and cell debris, then spun at 16 000 g for 20 min at 48C twice was added at the time of plating. to pellet the mitochondria, and the supernatants were used for anti-cytochrome c Western blot analysis. Pelleted Assessment of apoptosis mitochondria were washed twice with Mito-buer and lysed in TNC buer (10 m Tris-acetate, pH 8.0, 0.5% NP-40, Apoptosis was estimated by several methods: DNA ladder- M 5m CaCl ,1mM DTT, 0.1 mM PMSF, 2 mg/ml pepstatin, ing, caspase activation using both ¯uorogenic substrates and M 2 leupeptin, and aprotinin) on ice for 15 min and then used Western blot analysis of activated caspases, and proteolysis for Western blot analysis. To investigate the disruption of of PARP and DFF-45 (Rokhlin et al., 1997, 2000). For DNA DC , we used 40 nM 3,3'-dihexyloxacarbocyanine iodide laddering, cells were ®xed in 70% ethanol for 24 h at 7208C m (DiOC (3)) (Molecular Probes). Cells were treated for 1, 2, and DNA was extracted with 0.2 phosphate-citrate buer 6 M 3 and 4 h with TRAIL (500 ng/ml), then incubated for at pH 7.8 as described (Gewies et al., 2000). Under these 30 min at 378C in complete media with DiOC6, washed twice conditions, the partially degraded, oligonucleosomal DNA is with PBS and analysed in FACS Vantage cyto¯uorometer extracted quite selectively from the cells whereas the higher (Becton Dickinson). molecular weight DNA remains associated with the nuclei. The extracts were then treated with RNase A (1 mg/ml) in the presence of 0.25% NP-40 for 30 min at 378C, and then with proteinase K (1 mg/ml) for 30 min at 378C. The DNA Acknowledgments was analysed on 1.5% agarose gel (at 4 V/cm) containing The authors thank R Glover for technical assistance and A 0.5 mg/ml ethidium bromide. In order to measure caspase Gewies for assistance with Bcl-2 cDNA recloning. This activity with ¯uorogenic substrates, cytosolic extracts were research was supported by National Institutes of Health prepared as described (Rokhlin et al., 1998) by repeated grant CA 76673.

Oncogene Bcl-2 dependent apoptosis in prostate cancer OW Rokhlin et al 2843 References

Bay G, Miyashita T, Williamson JR and Reed JC. (1993). J. Rokhlin OW, Glover RA and Cohen MB. (1998). Cancer Biol. Chem., 268, 6511 ± 6519. Res., 58, 5870 ± 5875. Bonavida B, Ng CP, Jazirehi A, Schiller G and Mizutani Y. Rokhlin OW, Gudkov AV, Kwek S, Glover RA, Gewies AS (1999). Int. J. Oncol., 15, 793 ± 802. and Cohen MB. (2000). Oncogene, 19, 1959 ± 1968. Chandler JM, Cohen GM and MacFarlane M. (1998). J. Scadi C, Fulda S, Srinivasan A, Friesen C, Li F, Tomaselli Biol. Chem., 273, 10815 ± 10818. KJ, Debatin K-M, Krammer PH and Peter ME. (1998). Chinnaiyan AM, O'Rourke K, Tewari M and Dixit VM. EMBO J., 17, 1675 ± 1687. (1995). Cell, 81, 505 ± 512. Screaton GR, Mongkolsapaya J, Xu XN, Cowper AE, Gewies A, Rokhlin OW and Cohen MB. (2000). Cancer Res., McMichael AJ and Bell JI. (1997). Curr. Biol., 7, 693 ± 696. 60, 2163 ± 2168. SheridanJP,MarstersSA,PittiRM,GurneyA,Skubatch Gotow T, Shibata M, Kanamori S, Tokuno O, Ohsawa Y, M, Baldwin D, Ramakrishnan L, Gray CL, Baker K, Sato N, Isahara K, Yayoi Y, Watanabe T, Leterrier JF, Wood WI, Goddard AD, Godowski P and Ashkenazi A. Linden M, Kominami E and Uchiyama Y. (2000). Cell (1997). Science, 277, 818 ± 821. Death Dier., 7, 666 ± 674. Shimizu S, Eguchi Y, Kamiike W, Waguri S, Uchiyama Y, Green DR and Reed JC. (1998). Science, 281, 1309 ± 1312. Matsuda H and Tsujimoto Y. (1996). Oncogene, 13, 21 ± Green DR and Kroemer G. (1998). Trends Cell Biol., 8, 267 ± 29. 271. Sprick MR, Weigand MA, Rieser E, Rauch CT, Juo P, Blenis KeoghSA,WalczakH,Bouchier-HayesLandMartinSJ. J, Krammer PH and Walczak H. (2000). Immunity, 12, (2000). FEBS Lett., 471, 93 ± 98. 599 ± 609. Kischkel FC, Lawrence DA, Chuntharapai A, Schow P, Kim Susin SA, Lorenzo HK, Zamzami N, Marzo I, Brenner C, KJ and Ashkenazi A. (2000). Immunity, 12, 611 ± 620. Larochette N, Prevost M-C, Alzari P and Kroemer G. Li H, Zhu H, Xu CJ and Yuan J. (1998). Cell, 94, 491 ± 501. (1999). J. Exp. Med., 189, 381 ± 393. Luo X, Budihardjo I, Zou H, Slaughter C and Wang X. Utz PJ and Anderson P. (2000). Cell Death Dier., 7, 589 ± (1998). Cell, 94, 481 ± 490. 602. ManciniM,MachamerCE,RoyS,NicholsonDW, WalczakH,BouchonA,StahlHandKrammerPH.(2000). Thornberry NA, Casciola-Rosen LA and Rosen A. Cancer Res., 60, 3051 ± 3057. (2000). J. Cell Biol., 149, 603 ± 612. Walczak H, Degli-Esposti WA, Johnson RS, Smolak PJ, Mancini M, Nicholson DW, Roy S, Thornberry NA, Waugh JY, Boiani N, Timour MS, Gerhart MJ, Schooley Peterson EP, Casciola-Rosen LA and Rosen A. (1998). KA and Smith CA. (2000). EMBO J., 16, 5386 ± 5397. J. Cell Biol., 140, 1485 ± 1495. Walczak H and Krammer PH. (2000). Exp. Cell Res., 256, Nicholson DW. (2000). Nature, 407, 810 ± 816. 58 ± 66. Pan G, O'Rourke K, Chinnaiyan AM, Gentz R, Ebner R, Ni Willey SR, Schooley K, Smolak PJ, Din WS, Huang CP, J and Dixit VM. (1997a). Science, 276, 111 ± 113. Nicholl JK, Sutherland GR, Smith TD, Rauch C and PanG,NiJ,WeiYF,YuG,GentzRandDixitVM.(1997b). Smith CA. (1995). Immunity, 3, 673 ± 682. Science, 266, 815 ± 818. Yamada H, Tada-Oikawa S, Uchida A and Kawanishi S. Peter M, Kischkel FC, Hellbardt S, Chinnaiyan AM, (1999). Biochem. Biophys. Res. Commun., 265, 130 ± 133. Krammer PH and Dixit VM. (1996). Cell Death Dier., Yang E and Korsmeyer SJ. (1996). Blood, 88, 386 ± 401. 3, 161 ± 170. Yang J, Liu X, Bhalla K, Kim CN, Ibrado AM, Cai J, Peng Pitti RM, Marsters SA, Ruppert S, Donahue CJ, Moore A TI, Jones DP and Wang X. (1997). Science, 275, 1129 ± and Ashkenazi A. (1996). J. Biol. Chem., 271, 12687 ± 1132. 12690. ZamzamiN,SusinSA,MarchettiP,HirschT,Gomez- Reed JC. (1997). Nature, 387, 773 ± 776. Monterrey I, Castedo M and Kroemer G. (1996). J. Exp. Rokhlin OW, Bishop GA, Hostager BS, Waldschmidt TJ, Med., 183, 1533 ± 1544. Sidorenko SP, Pavlo N, Kiefer MC, Umansky SR, Zhivotovsky B, Samali A, Gahm A and Orrenius S. (1999). Glover RA and Cohen MB. (1997). Cancer Res., 57, Cell Death Dier., 6, 644 ± 651. 1758 ± 1768.

Oncogene