PS-341 (Bortezomib) Induces Lysosomal Cathepsin B Release and a Caspase-2-dependent Mitochondrial Permeabilization and Apoptosis in Human Pancreatic Cancer Cells*

PS-341 (bortezomib) is a potent and reversible proteosome inhibitor that functions to degrade intracellular polyubiquitinated proteins. PS-341 induces apoptosis and has shown broad antitumor activity with selectivity for transformed cells. We studied the effect of PS-341 on lysosomal and mitochondrial permeabilization, including the role of caspase-2 activation in apoptosis induction in the BxPC-3 human pancreatic carcinoma cell line. PS-341 induced a dose-dependent apoptosis in association with reactive oxygen species generation and cleavage of caspase-2 to its 33- and 14-kDa fragments. PS-341 disrupted lysosomes with redistribution of cathepsin B to the cytosol, as shown using fluorescence confocal microscopy, that was blocked by the free radical scavenger tiron but not by a caspase-2 inhibitor (benzyloxycarbonyl (Z)-VDVAD-fluoromethyl ketone (FMK)). PS-341-induced caspase-2 activation was attenuated by a selective pharmacological inhibitor of cathepsin B (R-3032), suggesting that cathepsin B release occurs upstream of caspase-2. PS-341-induced mitochondrial depolarization was attenuated by Z-VDVAD-FMK, tiron, and an inhibitor of the mitochondrial permeability transition pore (bongkrekic acid). Regulation of mitochondrial permeability by caspase-2 was confirmed using caspase-2 small interfering RNA. PS-341-induced cytochrome c release and phosphatidylserine externalization were attenuated by Z-VDVAD-FMK and partially by R-3032. PS-341 activated the BH3-only proteins Bik and Bim and down-regulated Bcl-2 and Bcl-xL mRNA and protein expression. Taken together, PS-341 induces lysosomal cathepsin B redistribution upstream of caspase-2. Caspase-2 activation regulates PS-341-induced mitochondrial depolarization and apoptosis, suggesting that caspase-2 can serve as a link between lysosomal and mitochondrial permeabilization.


PS-341 (bortezomib) is a potent and reversible proteosome inhib-
itor that functions to degrade intracellular polyubiquitinated proteins. PS-341 induces apoptosis and has shown broad antitumor activity with selectivity for transformed cells. We studied the effect of PS-341 on lysosomal and mitochondrial permeabilization, including the role of caspase-2 activation in apoptosis induction in the BxPC-3 human pancreatic carcinoma cell line. PS-341 induced a dose-dependent apoptosis in association with reactive oxygen species generation and cleavage of caspase-2 to its 33-and 14-kDa fragments. PS-341 disrupted lysosomes with redistribution of cathepsin B to the cytosol, as shown using fluorescence confocal microscopy, that was blocked by the free radical scavenger tiron but not by a caspase-2 inhibitor (benzyloxycarbonyl (Z)-VDVAD-fluoromethyl ketone (FMK)). PS-341-induced caspase-2 activation was attenuated by a selective pharmacological inhibitor of cathepsin B (R-3032), suggesting that cathepsin B release occurs upstream of caspase-2. PS-341-induced mitochondrial depolarization was attenuated by Z-VDVAD-FMK, tiron, and an inhibitor of the mitochondrial permeability transition pore (bongkrekic acid). Regulation of mitochondrial permeability by caspase-2 was confirmed using caspase-2 small interfering RNA. PS-341-induced cytochrome c release and phosphatidylserine externalization were attenuated by Z-VDVAD-FMK and partially by R-3032. PS-341 activated the BH3-only proteins Bik and Bim and down-regulated Bcl-2 and Bcl-x L mRNA and protein expression. Taken together, PS-341 induces lysosomal cathepsin B redistribution upstream of caspase-2. Caspase-2 activation regulates PS-341-induced mitochondrial depolarization and apoptosis, suggesting that caspase-2 can serve as a link between lysosomal and mitochondrial permeabilization.
Adenocarcinoma of the pancreas is the fifth most incident cancer in the United States (1), and fewer than 5% of affected patients survive 5 years after diagnosis. Pancreatic cancers are among the most intrinsically resistant tumors to chemotherapeutic drugs and irradiation. A novel agent targeting the proteosome has recently been developed and shown to display broad antitumor activity (2) and to overcome resistance to chemotherapy (3). Based upon highly favorable results in patients with refractory or relapsed multiple myeloma (4), PS-341 (bortezomib or Velcade) was approved by the United States Food and Drug Administration for this indication, and this drug is currently undergoing evaluation in several other cancers. PS-341, a dipeptidyl boronic acid, is a highly selective and reversible inhibitor of the 26 S proteosome, which is a multicatalytic enzyme that degrades damaged or misfolded/ unfolded cellular proteins targeted by ubiquitin conjugation (5). The proteosome contributes to cellular homeostasis by directly or indirectly regulating intracellular protein levels, including those regulating cell cycle progression, apoptosis, and transcription factor activation (5,6). PS-341 has shown activity against several malignant cell types with selectivity for transformed cells compared with normal cells (5). PS-341 has been shown to promote apoptosis by diverse mechanisms, including the membrane death receptor pathway, as evidenced by the ability of PS-341 to act cooperatively with tumor necrosis factor-related apoptosis-inducing ligand (e.g. TRAIL) to induce apoptosis (7). PS-341 can also stabilize proapoptotic Bax by inhibiting its degradation (8) and engage the mitochondrial apoptotic pathway (9). These findings suggest that PS-341-induced apoptosis may, in part, be regulated by the Bcl-2 family (6,8). The Bcl-2 family includes the BH3-only proteins, so-called because they contain a BH3 protein interaction domain (e.g. Bid, Bik, and Bim) and serve as triggers for the apoptotic signal, whereas Bax and Bak act downstream to regulate apoptosis purportedly through permeabilization of mitochondria (10). Furthermore, PS-341 can stabilize IB, thereby decreasing the antiapoptotic effects of nuclear factor B (11).
The membrane death receptor and mitochondrial apoptotic pathways are dependent upon activation of a caspase cascade. Recent data suggest that a caspase-independent mechanism involving lysosomes may function in the initiation of cell death (12). Lysosomes release a family of proteases known as cathepsins that translocate to the cytosol in response to signals including oxidative stress, increased sphingosine levels, and Fas and tumor necrosis factor-␣ ligation (13). Of the 11 known cathepsins in mammalian lysosomes, cathepsin B and D are the most prominent and stable at physiological pH, and cathepsin B is ubiquitously expressed (13). In hepatocytes treated with tumor necrosis factor-R1, lysosomal permeabilization with release of cathepsin B into the cytosol has been associated with mitochondrial dysfunction and caspasedependent cell death (14). The ability of cathepsin B to induce cell death was shown in a cell-free system whereby cathepsin B produced chromatin condensation and apoptotic morphology (15). Although cathepsin B has been shown to play a role in apoptosis in vitro and in vivo (14,16), cellular mechanisms responsible for ligand-or drug-induced lysosomal permeabilization are poorly understood. Whereas evidence supports participation of the mitochondria in transmitting a death signal initiated at the lysosomes (17), the mechanism by which lysosomal disruption leads to mitochondrial permeabilization remains unknown. In this regard, in vitro cleavage experiments have shown that the major human caspases are poor substrates for lysosomal extracts or cathepsins (18).
The placement of caspase-2 within the known apoptotic signaling pathways remains incompletely understood. Recent evidence suggests that caspase-2 can serve as a proximal caspase that can be activated by cytotoxic stress and may be involved in the mitochondria-mediated apoptosis (19,20). In this regard, sequential caspase-2 and caspase-8 activation were shown to occur upstream of mitochondria during ceramide-and etoposide-induced apoptosis (20). Chemotherapy-induced apoptosis has been shown to involve oxidative stress (21,22). Caspase-2 has also been reported to be cleaved by effector caspase-3, indicating that it can function downstream of mitochondria during apoptotic signaling (23). Caspase-2 is recruited to a large protein complex whose formation occurs independently of an Apaf1/apoptosome pathway and is sufficient to mediate its activation (24). However, the mechanisms of caspase-2 activation remain unclear. After an apoptotic stimulus, caspase-2 processing occurs by two proteolytic steps. A first cleavage at aspartic acid 316 generates two fragments: one of 32-33 kDa, the large subunit, and a second fragment of 14 kDa, the small subunit. The appearance of the 32-33-kDa fragment has been generally used as marker of caspase-2 activation (25,26). Of note, oocytes from caspase-2-deficient mice display resistance to chemotherapy-induced apoptosis (27). Whereas data suggest that caspase-2 regulates apoptosis, an siRNA 2 was shown to inhibit caspase-2 expression but not to suppress apoptosis in transformed cells (19,28). Additionally, the absence of severe phenotypic abnormalities in caspase-2-deficient mice has cast doubt as to the importance of this caspase in apoptosis (27).
In this study, we determined the relationship of lysosomal disruption and caspase-2 activation to mitochondrial apoptotic signaling induced by PS-341 in BxPC-3 human pancreatic cancer cells. Our results indicate that PS-341 triggers reactive oxygen species (ROS) generation and lysosomal permeabilization with cathepsin B redistribution to the cytosol. Oxidative stress and cathepsin B activate caspase-2, which is required for mitochondrial permeabilization and apoptosis, suggesting that caspase-2 can serve as a link between the lysosomes and the mitochondria. PS-341 also down-regulates antiapoptotic Bcl-2 family members and activates proapototic BH-3-only proteins.

EXPERIMENTAL PROCEDURES
Cell Culture and Reagents-BxPC-3 and CFPAC-1 human pancreatic carcinoma cell lines were obtained from the American Type Culture Collection (Manassas, VA). BxPC-3 cells were grown in RPMI 1640 supplemented with 10% fetal bovine serum, 2 mM glutamine, 1 mM sodium pyruvate, and 10 mM HEPES. CFPAC-1 cells were grown in Iscove's Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 4 mM glutamine, and for both cell lines, 100 units/ml penicillin and 100 g/ml streptomycin in an atmosphere of 95% air and 5% CO 2 at 37°C. PS-341 was generously provided by Millennium Pharmaceuticals Inc. (Cambridge, MA). The superoxide scavenger tiron (Sigma) was utilized as outlined below.
Caspase-2, Caspase-8, and Cathepsin B Inhibitor-An irreversible caspase-2 inhibitor I, Z-VDVAD-FMK, was purchased from Calbiochem. A cell-permeable and irreversible caspase-8 inhibitor, Z-IETD-FMK, was obtained (R&D Systems Inc., Minneapolis, MN). A selective, reversible cathepsin B inhibitor, R-3032, was obtained from Celera Genomics (South San Francisco, CA) (29). The R-3032 dose was based on pharmacokinetic data demonstrating a half-life of 4.2 h and preliminary data demonstrating that extracellular concentrations of 10 M were required to maximally inhibit tumor necrosis factor-␣/actinomycin D-mediated apoptosis in cultured murine hepatocytes (29). The K I for cathepsin B is 0.02 M, and the drug does not inhibit caspases.
Annexin V Assay-Cells were seeded at a density of 4 ϫ 10 5 cells/well in a 6-well plate, grown overnight, and then treated with PS-341. For experiments employing inhibitors, cells were preincubated with 5 mM tiron or 50 M Z-VDVAD-FMK, both for 3 h, or 20 M R-3032 for 24 h followed by PS-341 treatment. After treatments, adherent and floating cells were collected by trypsinization and centrifuged at 1300 rpm for 5 min. Cell pellets were resuspended and incubated in complete medium for 10 min. After centrifugation, cell pellets were washed with PBS twice and resuspended in 500 l of 1ϫ annexin binding buffer (BD Biosciences) with 5 l of annexin V-fluorescein isothiocyanate (BD Biosciences) and 0.5 l of propidium iodide (Sigma). The percentage of cells with annexin V ϩ /propidium iodide (PI) Ϫ was measured using fluorescence-activated cell sorting.
Western Blotting-Cells were seeded at a density of 2 ϫ 10 6 cells/plate overnight. At the indicated time periods (see figure legends), cells were harvested by RIPA B lysis buffer (20 mM sodium phosphate (pH 7.4), 150 mM NaCl, 1% Triton X-100, 5 mM EDTA, 5 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, and 250 g/ml sodium vanadate) on ice for 15 min. Total cellular proteins were collected and centrifuged at 10,000 ϫ g for 15 min at 4°C to remove cellular debris, and the supernatant was collected. Protein concentration was measured by the DC protein assay kit (Bio-Rad). Total cell lysates (50 g) were subjected to 10% SDS-PAGE, and following electrophoresis, proteins were electroblotted onto Immobilon-P Transfer Membrane (Millipore, Bedford, MA). After blocking nonspecific binding sites with I-Block (Tropix, Foster City, CA) at room temperature for 1 h, membranes were incubated with the primary antibodies overnight at 4°C. After washes in PBST (PBS with 0.1% Tween 20), membranes were incubated with the corresponding alkaline phosphatase-conjugated secondary antibodies for 1 h at room temperature. Signal was detected by chemiluminescence using CDP-Star reagent (PerkinElmer Life Sciences).
Detection of ROS Generation-Dihydroethidium (HE) (Molecular Probes, Inc., Eugene, OR) was utilized to measure ROS generation in BxPC-3 cells. Cells (2.0 ϫ 10 6 ) were preincubated with 5 mM tiron or 50 M Z-VDVAD-FMK for 3 h or with 20 M R-3032 for 24 h, followed by PS-341 treatment. After drug exposure, cells were collected by trypsinization and incubated in complete medium for 10 min. Then cells were incubated in 2 M of HE (excitation wavelength 488 nm; emission 605 nm) at 37°C for 15 min and washed with PBS. Finally, cells were resuspended in 400 l of PBS and then subjected to fluorescenceactivated cell sorting analysis.
RNA Interference-BxPC-3 cells (1.7 ϫ 10 6 cells/plate) were mocktransfected or transfected with 10 nM siCONTROL nontargeting siRNA duplex or transfected with human caspase-2-specific siRNA duplex (Dharmacon, Lafayette, CO) using siLectFect TM , according to the manufacturer's instructions (Bio-Rad). Twenty-four hours after cells were transfected, media were replaced with fresh media with or without 100 nM PS-341 and incubated for 24 h. The cells were then analyzed for mitochondrial depolarization using JC-1 staining as described above.
Transfection of Cathepsin B-Green Fluorescent Protein-The rat cathepsin B-green fluorescent protein (GFP) expression plasmid was a generous gift of Dr. Gregory Gores (Mayo Clinic). Cells (0.4 ϫ 10 6 ) were seeded overnight before transfection. Transfection was performed with 1 ml of Opti-MEM I containing 6 l of Plus reagent, 1 g of plasmid, and 4 l of Lipofectamine reagent (Invitrogen). After a 24-h transfection, cells were preincubated with 5 mM tiron or 50 M Z-VDVAD-FMK for 3 h with 20 M R-3032 for 24 h, followed by PS-341 treatment for 24 h. Confocal microscopy was performed with an inverted Zeiss laser-scanning confocal microscope (Zeiss LSM S10) using excitation of 488 nm and emission of 507 nm.
Measurement of Caspase-2 Activity-ICH-1/caspase-2 protease activity was measured in cell lysates obtained from cultured cells with colorimetric substrates (Chemicon, Temecula, CA). Cells (1.0 ϫ 10 6 ) were preincubated with 20 or 50 M R-3032 for 24 h, followed by PS-341 treatment. After drug exposure, cells were collected by trypsinization. 2-5 ϫ 10 6 cells were pelleted, and lysed in 50 l of lysis buffer on ice for 10 min. The cytosolic extract was obtained by centrifugation at 10,000 ϫ g for 1 min, and protein concentration was assayed. Protease activity was measured by adding 50 l of cytosolic extract (200 g), 50 l of 2ϫ reaction buffer, and 200 M VDVAD-p-nitroanilide substrate and then incubated at 37°C for 1 h. The colorimetric signal was detected using a VERSAmax Tunable Microplate Reader (Molecular Devices, Inc., Sunnyvale, CA) at wavelength of 405 nm. -Fold increase in caspase-2 activity was then determined relative to untreated control.
Reverse Transcription-PCR-Cells were seeded at a density of 4 ϫ 10 5 cells/well of a 6-well plate overnight. After drug treatments, cells were harvested and dissolved in 1 ml of TRIZOL reagent (Invitrogen). Total RNA was extracted according to the manufacturer's instruction. The RNA A 260 /A 280 ratios were between 1.6 and 1.8. cDNA was made from 1 g of total RNA with the iScript cDNA Synthesis Kit (Bio-Rad) according to the manufacturer's instructions. mRNA expression level was detected by real time PCR using the iQ SYBR Green Supermix (Bio-Rad). Gene expression level was quantified using the threshold PCR cycle number (Ct) method. In brief, the relative expression level of the target gene compared with that of the housekeeping gene, GAPDH, was calculated as 2 Ϫ⌬Ct , where ⌬Ct ϭ Ct target gene Ϫ Ct GAPDH . The ratio of relative expression of the target gene to that of untreated cells was then calculated as 2 ϪD⌬Ct , where ⌬Ct ϭ ⌬Ct treated Ϫ ⌬Ct untreated (31). The primers were as follows: sequence of GAPDH (GAGTCAACG-GATTTGGTCGT (forward) and TTGATTTTGGAGGGATCTCG (reverse)), Bcl-2 (GGATGCCTTTGTGGAACTGT (forward) and AGCCTGCAGCTTTGTTTCAT (reverse)), and Bcl-x L (GGAGCTG-GTGGTTGACTTTC (forward) and CTCCGATTCAGTCCCTTCTG (reverse)). The PCR was run for 35 cycles with a 58°C annealing cycle (30 s), a 72°C extension cycle (30 s), and a 95°C denaturing cycle (50 s) plus final incubation at 72°C for 10 min.
Statistical Analysis-All experiments, except immunoblots, were performed in triplicate, and the results were expressed as the mean Ϯ S.D. values. Statistical significance was determined utilizing the Student's t test or a one-way analysis of variance. Statistical significance was defined as p value of Ͻ0.05.

PS-341 Induces Cytotoxic Stress and Activates Caspase-2-mediated
Apoptosis-Induction of oxidative stress by cytotoxic drugs has been shown to directly engage the mitochondrial apoptotic pathway (22). We determined the potential for PS-341 to generate ROS in BxPC-3 cells. Cells were incubated in dihydroethidium (HE), and the fluorescence signal was detected by flow cytometry. PS-341 treatment produced a 5-6-fold increase in ROS generation relative to control that was significantly attenuated by preincubation of cells with the superoxide scavenger, tiron (Fig. 1A). Induction of oxidative stress by PS-341 was confirmed by measuring ␥-glutamylcysteine synthase mRNA levels. ␥-Glutamylcysteine synthase catalyzes the rate-limiting step in the de novo biosynthesis of GSH, an abundant physiological antioxidant (32). PS-341 (50 and 100 nM doses at 24 h) produced a 1.8-fold induction in ␥-glutamylcysteine synthase mRNA expression compared with control (data not shown). PS-341 treatment of BxPC-3 cells produced a dosedependent induction of apoptosis, as detected by phosphatidylserine externalization using an annexin V binding assay (Fig. 1B). PS-341 treatment cleaved caspase-8 to its active 43-and 18-kDa fragments and reduced the level of full-length Bid (Fig. 1C). To confirm that PS-341 can activate caspase-8 and engage the extrinsic apoptotic pathway, we utilized a specific inhibitor of caspase-8 (Z-IEHD-FMK). Z-IETD-FMK was found to block PS-341-induced cleavage of caspase-8 and to attenuate PS-341-induced apoptosis (Fig. 1D). After Bid cleavage, its truncated form is translocated to the mitochondria, where it mediates crosstalk with the mitochondrial pathway (33). PS-341 cleaved downstream effector caspase-9 and caspase-3 and triggered the release of mitochondrial cytochrome c and Smac (second mitochondria-derived caspase activator) proteins (Fig. 1C). Smac has been shown to promote caspase-9 activation via Apaf-1 and to antagonize the inhibitory effect of XIAP on caspase-9 (34). Preincubation of cells with increasing doses of the superoxide scavenger tiron (3 h) inhibited PS-341-induced apoptosis in a dose-dependent manner (Fig. 1E) and suppressed PS-341-induced cytochrome c release (Fig. 1F).
We then determined whether PS-341-induced ROS generation could activate caspase-2 in BxPC-3 cells. In a recent study, cytotoxic stress was shown to activate caspase-2 in human tumor cell lines, although its role as an initiator caspase and mediator of stress-induced apoptosis remains controversial (19). PS-341 treatment was associated with a reduction in procaspase-2 and the appearance of its first cleavage product (33-kDa fragment), as detected using an N terminus-specific monoclonal antibody ( Fig. 2A). Consistent with these data, paclitaxel was shown to increase the ratio of the first cleavage product (33 kDa) to procaspase-2 in a time-dependent manner (35). Caspase-2 was also cleaved by PS-341
PS-341-induced Caspase-2 Activation Regulates Mitochondrial Depolarization-PS-341 was shown to increase mitochondrial membrane permeabilization in BxPC-3 cells, as measured using the vital mitochondrial dye JC-1 and shown as an increase in depolarization (Fig.  3A). JC-1 undergoes a reversible change in fluorescence emission from red to green during mitochondrial depolarization. Cells with high membrane potential promote the formation of dye aggregates that fluoresce red, and cells with low potential contain monomeric JC-1 that produces green fluorescence. Therefore, we can accurately determine mitochondrial depolarization using the ratio of red to green fluorescence, since it is not affected by mitochondrial size, shape, or density (37). We then determined whether caspase-2 can regulate mitochondrial permeabilization by PS-341. PS-341-induced mitochondrial depolarization (⌬m loss) was blocked by the caspase-2 inhibitor, Z-VDVAD-FMK, indicating that caspase-2 activation occurs upstream of the mitochondria and mediates ⌬m loss (Fig. 3A). PS-341-induced mitochondrial depolarization was also blocked by bongkrekic acid, a potent inhibitor of the mitochondrial permeability transition pore, and by tiron (Fig. 3A).
To confirm that caspase-2 regulates mitochondrial permeability in BxPC-3 cells, we transfected these cells with an siRNA against caspase-2. Caspase-2 siRNA was shown to reduce caspase-2 expression relative to control cells, and its selectivity for caspase-2 is shown by a lack of effect upon caspase-8 (Fig. 3B). In cells transfected with caspase-2 siRNA, PS-341-induced mitochondrial depolarization was attenuated by ϳ40% compared with PS-341-treated mock-transfected cells (Fig. 3B). The greater inhibitory effect of Z-VDVAD-FMK (Fig. 3A) compared with caspase-2 siRNA supports studies demonstrating that Z-VDVAD-FMK can also inhibit caspase-3 and -7 (38). However, incomplete suppression of caspase-2 by siRNA may be contributory, as can potential caspase-2-independent mechanisms affecting mitochondrial permeabilization.

PS-341 Induces Lysosomal Disruption and Cathepsin B Release-
Most organelle-specific death responses lead to mitochondrial depolarization and caspase activation. Evidence indicates, however, that some models of apoptosis are dependent upon release of lysosomal proteases known as cathepsins and are independent of caspases (12), whereas others rely on both (39,40). We determined the effect of PS-341 upon lysosomal permeabilization and determined its relationship to caspase-2. BxPC-3 cells were transfected with the cathepsin B-GFP expression plasmid prior to PS-341 exposure. The cellular localization of cathepsin B was then examined using confocal microscopy. Cathepsin B-GFP has been characterized and identified as a native protein for assessing lysosomal integrity during apoptosis (14). As shown in Fig. 4A, cathepsin B-GFP displayed a punctate fluorescent pattern consistent with its localization to a vesicular compartment (i.e. the lysosomes). After PS-341 treatment (24 h), the cathepsin B-GFP fluorescent signal became diffuse, indicating a redistribution to the cytosol. PS-341-induced ROS generation triggered lysosomal permeabilization, as shown by the ability of tiron to block cathepsin B movement. Cathepsin B redistribution was also blocked by the highly specific cathepsin B inhibitor, R-3032 (Fig. 4A). No inhibitory effect on cathepsin B redistribution was observed when cells were incubated with Z-VDVAD-FMK, indicating that lysosomal permeabilization occurred upstream of caspase-2. Taken together, PS-341-induced ROS triggers lysosomal disruption and a caspase-2-independent release of cathepsin B into the cytosol.
We then determined the effect of R-3032 upon caspase-2 activation by PS-341 using immunoblotting and also measured its effect upon caspase-2 enzymatic activity. R-3032 attenuated the reduction in procaspase-2 expression and cytochrome c release induced by PS-341 (Fig.  4B), suggesting that cathepsin B is upstream of caspase-2 and can also activate caspase-2. As further evidence, PS-341 was found to induce

. PS-341-induced ROS generation activates caspase-2 and caspase-2-mediated apoptosis. A, BxPC-3 cells were incubated with PS-341
for 24 h, and caspase-2 expression was analyzed in total cell lysates using SDS-PAGE and immunoblotting. PS-341 treatment cleaved procaspase-2 (48 kDa) to its 33-and 14-kDa active fragments that was blocked by preincubation (3 h) of cells with tiron. ␤-Actin was used as a control for protein loading. B, BxPC-3 cells were treated with PS-341 for 24 h with or without preincubation with the caspase-2 inhibitor (Z-VDVAD-FMK) for 3 h. ROS generation was detected by HE staining as described under "Experimental Procedures." ROS levels are shown relative to control cells. Columns, mean of triplicate determinations; bars, ϮS.D. C, PS-341-induced apoptosis was analyzed in the presence or absence of Z-VDVAD-FMK. Apoptosis was determined using the annexin V-fluorescein isothiocyanate assay and PI staining. The percentage of apoptotic cells (annexin V ϩ , PI Ϫ ) was analyzed by flow cytometry. *, p Ͻ 0.05 versus PS-341 alone at corresponding dosages.
caspase-2 enzymatic activity 1.8-fold, and this activity was attenuated (ϳ50% reduction) by both 20 and 50 M doses of R-3032 (Fig. 4C). Given that the mechanisms of interaction between the lysosomal and mitochondria-mediated apoptotic pathways remain unknown, we sought to determine the role of cathepsin B in PS-341-induced apoptosis. In BxPC-3 cells treated with PS-341, R-3032 attenuated PS-341induced apoptosis in a dose-dependent manner. Specifically, R-3032 at 20 and 50 M inhibited PS-341-induced phosphatidylserine externalization by 16 and 38%, respectively (Fig. 4D). These data indicate that cathepsin B contributes to PS-341-induced caspase-2 activation and apoptosis induction.
PS-341-induced Apoptosis Is Regulated by Caspase-2 in CFPAC-1 Human Pancreatic Carcinoma Cells-To determine whether our results can be generalized, we performed experiments in the CFPAC-1

PS-341 Modulates Expression of Bcl-2 Family
Members-Recent data indicate that PS-341 may promote apoptosis, in part, by effects upon Bcl-2 family members (6,8,9,41). The Bcl-2 family includes proapoptotic BH3-only proteins (Bik, Bim, Bid, Puma, etc.) that contain a small BH3 interaction domain (10). A recent report suggests the importance of Bik and Bim in PS-341-induced apoptosis (41). We determined the effect of PS-341 on Bik and Bim in BxPC-3 cells. PS-341 was found to cleave/activate Bim and to increase the level of Bik protein expression at 24 h (Fig. 6A). The importance of Bik and Bim activation by PS-341 is further suggested by data indicating that mouse embryonic fibroblasts in which genes for both of these BH3-only proteins have been disrupted display resistance to cytotoxic therapy (41). We also determined the effect of PS-341 upon Bcl-2 and Bcl-x L expression levels. Treatment of BxPC-3 cells with PS-341 resulted in a dose-and time-dependent downregulation of Bcl-2 and Bcl-x L protein (Fig. 6B) and mRNA (Fig. 6, C and  D) levels. Bcl-2 and Bcl-x L mRNA levels were reduced after a 12-h incubation with PS-341, whereas protein expression was decreased markedly at 48 h. These results indicate that Bik and Bim activation precede Bcl-2 and Bcl-x L down-regulation, and together, these events may enhance mitochondrial apoptotic signaling by PS-341. The postulated relationship between the lysosomes, caspase-2, and mitochondrial signaling, including the role of Bcl-2 family members, is schematically depicted (Fig. 7).

DISCUSSION
To elucidate the apoptotic mechanisms utilized by PS-341, we studied the effects of this drug upon lysosomal and mitochondrial perme- ability in the initiation of cell death in human pancreatic carcinoma cells. We also examined the role and placement of caspase-2 in the sequence of events during PS-341-induced apoptotic signaling. Our findings provide the first evidence that PS-341 disrupts lysosomes to release cathepsin B, a noncaspase protease, in pancreatic cancer cells. Using fluorescence confocal microscopy, PS-341 treatment resulted in a redistribution of cathepsin B from the lysosomes to the cytosol. Importantly, cathepsin B translocation was blocked by the superoxide scavenger tiron but not by a pharmacological inhibitor of caspase-2 (Z-VD-VAD-FMK), indicating that ROS generation by PS-341 contributes to lysosomal disruption. Treatment of BxPC-3 cells with PS-341 was shown to cleave procaspase-2 to its 33-and 14-kDa fragments. Caspase-2 cleavage was partially inhibited by a selective cathepsin B inhibitor, R-3032, and R-3032 attenuated the induction of caspase-2 enzymatic activity by PS-341. Together, these data suggest that lysosomal permeabilization and cathepsin B release occur upstream of caspase-2 and partially regulate its activation. Additionally, these results indicate that lysosomal permeabilization and cathepsin B release are initiator events in the cellular commitment to apoptosis (12).
To date, the mechanism by which lysosomal disruption leads to mitochondrial membrane permeabilization remains unknown, and a direct link between cathepsin B and caspases has not been demonstrated. We found that caspase-2 is required for PS-341-mediated mitochondrial permeabilization as shown using a caspase-2 inhibitor (Z-VDVAD-FMK) and confirmed by caspase-2 siRNA. Mitochondrial permeabilization by PS-341 was suppressed to a greater extent by Z-VDVAD-FMK than with caspase-2 siRNA, suggesting that the inhibitor may suppress other caspases, such as caspase-3 and -7 (39), or is related to incomplete suppression of caspase-2 by siRNA. Recent evidence indicates that caspase-2 can be activated downstream of the mitochondria by caspase-3 in a feedback amplification loop (23). The cathepsin B inhib-itor R-3032 was shown to suppress PS-341-mediated caspase-2 activation and enzymatic activity, suggesting that caspase-2 may serve as a link between lysosomal and mitochondrial permeabilization. However, R-3032 showed only a partial suppression of caspase-2 activity, suggesting that events independent of cathepsin B and/or a feedback amplification loop via caspase-3 are involved in caspase-2 activation. To demonstrate that our findings can be generalized, we treated CFPAC-1 human pancreatic cancer cells with PS-341 and observed caspase-2 activation and apoptosis induction in these cells. PS-341-induced apoptosis was regulated by caspase-2 and to a lesser extent by cathepsin B compared with BxPC-3 cells.
We found that PS-341 treatment induced oxidative stress in BxPC-3 cells as evidenced by production of ROS and up-regulation of ␥-glutamylcysteine synthase expression. PS-341-induced ROS activated caspase-2 and resulted in mitochondrial depolarization and apoptosis induction; these events were suppressed by the superoxide scavenger tiron. Evidence indicates that ROS-related hydroxide production in mitochondria damages lysosomal membranes with leakage of cathepsin D that can then permeabilize mitochondrial membranes to trigger apoptosis (42). Mitochondria are both a major endogenous source of ROS and a target of ROS, since oxidative stress has been shown to induce apoptosis by targeting the mitochondria directly (43). In addition to mitochondria, endoplasmic reticulum stress can increase mitochondrial Ca 2ϩ resulting in altered mitochondrial permeability (44). In this regard, recent evidence indicates that PS-341 can induce endoplasmic reticulum stress to generate ROS (45). Activation of caspase-2 by c-Jun N-terminal kinase upstream of mitochondria has been shown to occur in response to ROS (46). In addition to c-Jun N-terminal kinase, experiments have shown that STAT1 (signal transducer and activator of transcription 1)-mediated apoptosis also involves specific activation of caspase-2 (47).
Recent data have cast doubt as to the requirement for caspase-2 in cytotoxic drug-induced apoptosis. In this regard, Lassus et al. (28) identified an siRNA that suppressed caspase-2 expression but failed to prevent apoptosis, whereas other siRNAs did both. As shown here, caspase-2 is an initiator protease that can be activated prior to mitochondrial depolarization. PS-341 triggers a signaling cascade, resulting in lysosomal disruption with downstream activation of caspase-2 to produce mitochondrial permeabilization. Placement of caspase-2 downstream of lysosomal permeabilization and upstream of the mitochondria is supported by our finding that Z-VDVAD-FMK failed to block PS-341-induced ROS generation and cathepsin B redistribution, but did block its effect on mitochondrial depolarization and apoptosis. Specifically, we found that mitochondrial release of cytochrome c and phosphatidylserine externalization by PS-341 were regulated by caspase-2. In contrast to results for the caspase-2 inhibitor, a cathepsin B inhibitor only modestly inhibited PS-341-induced cytochrome c release and partially suppressed apoptosis induction, suggesting that cathepsin B contributes to but is not required for apoptosis propagation after PS-341. A role for cathepsin B in apoptotic signaling has been shown in cathepsin B knock-out cells that display resistance to tumor necrosis factor-␣-induced apoptosis (16). Our findings were not limited to BxPC-3 cells in that PS-341 also cleaved caspase-2 in CFPAC-1 cells that showed greater apoptotic susceptibility to PS-341 than did BxPC-3 cells. Caspase-2 inhibition by Z-VDVAD-FMK reduced PS-341-induced phosphatidylserine externalization, indicating that caspase-2 also regulates PS-341-induced apoptosis in CFPAC-1 cells. However, pretreatment with R-3032 produced only a modest reduction in apoptosis, indicating a lesser dependence upon lysosomal permeabilization by PS-341 in CFPAC-1 as compared with BxPC-3 cells.

PS-341 Induces a Caspase-2-dependent Apoptosis
PS-341 treatment was shown to cleave caspase-8, and a specific inhibitor of caspase-8 blocked its activation and attenuated PS-341-induced apoptosis, consistent with its engagement of the extrinsic apoptotic pathway. Recent studies have shown that PS-341 acts synergistically with TRAIL to induce apoptosis in solid tumor cell lines (7). Furthermore, PS-341 reduced the level of full-length Bid, a BH3-only protein, enabling cross-talk with the mitochondria (33) as shown by release of cytochrome c, Smac/DIABLO, and downstream caspase-9 activation. Because apoptosis is often triggered by BH3-only proteins of the Bcl-2 family (10), we determined the effect of PS-341 on Bik and Bim levels. Bik and Bim are activating BH3-only proteins that initiate apoptosis by inducing Bax and Bak oligomerization and subsequent cytochrome c release (10). Bik is a BH3-sensitizing protein that exerts its prodeath function by binding antiapoptotic Bcl-2 family members, thereby preventing them from sequestering and activating BH3 proteins (10). Bim is known to be required for participation in apoptosis in several cell types (48). Bim knock-out mice show a severe phenotype, whereas Bik knock-outs appear unaltered (48,49), perhaps reflecting overlapping functions or indicating that Bik is not essential for embryonic development. We found that PS-341 treatment resulted in accumulation of Bik and cleavage of Bim, suggesting that their activation contributes to apoptosis induction by this drug. Bik has been shown to be ubiquitinated, and its accumulation in cell lines in response to PS-341 suggests that it is stabilized by proteosome inhibition (10). Furthermore, suppression of Bik induction has been shown to confer resistance to PS-341-induced apoptosis, and a greater degree of resistance was observed when expression of both Bim and Bik were suppressed in tumor cell lines (41). We also determined the effect of PS-341 on Bcl-2 and Bcl-x L expression, FIGURE 7. Schematic diagram of PS-341-induced apoptosis. During PS-341-induced apoptotic signaling, lysosomal disruption with cathepsin B redistribution is triggered by PS-341-induced ROS generation that is followed by caspase-2 activation and mitochondrial depolarization. Additionally, PS-341 activates BH3-only proteins, Bik and Bim, and downregulates the expression of antiapoptotic proteins, Bcl-2 and Bcl-x L . Mitochondria act as amplifiers, with resultant release of cytochrome c and effector caspase activation. Activation of caspase-3 can serve as a feedback loop to further activate initiator caspase-2 (23). FIGURE 6. PS-341 down-regulates antiapoptotic proteins, Bcl-2 and Bcl-X L , and activates pro-apoptotic BH-3-only proteins, Bik and Bim. BxPC-3 cells were incubated with PS-341 for the specified time periods. A, in cells incubated with PS-341, the BH-3-only proteins Bik and Bim were analyzed, and their activation is shown in total cell lysates using SDS-PAGE and immunoblotting. ␤-Actin was used as a loading control. B, the effect of PS-341 on Bcl-2 and Bcl-x L protein expression was also determined in total cell lysates. C and D, Bcl-2 (C) and Bcl-x L (D) mRNA levels were determined using 1 g of total RNA subjected to reverse transcription and real time PCR, as described under "Experimental Procedures." Relative mRNA levels were quantified using the threshold PCR cycle number method (31). Mean Ϯ S.D. values are shown for triplicate determinations.
given their known role in regulating mitochondria-mediated apoptosis and their potential to protect against ROS generation (50). We found that PS-341 treatment decreased the expression of Bcl-2 and Bcl-x L mRNA and protein levels in a dose-and time-dependent manner. Similar results were reported for PS-341 in human glioblastoma multiforme cells (51). In non-small cell lung cancer cells, PS-341 was shown to induce Bcl-2 phosphorylation and cleavage in association with its apoptosis induction (52). However, and in contrast to Bik and Bim, the delayed reduction in Bcl-2 or Bcl-x L by PS-341 seen at 48 but not 24 h suggests that alterations in their levels did not significantly impact PS-341-induced apoptosis in BxPC-3 cells.
In conclusion, the proteosome inhibitor PS-341 induces oxidative stress and initiates apoptotic signaling by triggering lysosomal disruption and cathepsin B redistribution to the cytosol. Activated cathepsin B can then activate caspase-2 that is required for mitochondrial permeabilization and apoptosis induction by PS-341. Caspase-2 appears to provide a link between the lysosomal and mitochondrial permeabilization pathways (schematically depicted in Fig. 7). In contrast to caspase-2-dependent mitochondrial signaling, cathepsin B release was contributory to but not required for apoptosis induction by PS-341. PS-341 activated the proapoptotic BH3-only proteins, Bik and Bim, and reduced the level of the antiapoptotic proteins Bcl-2 and Bcl-x L , which would be expected to enhance mitochondrial apoptotic signaling. These novel findings provide mechanistic insights into apoptosis induction by proteosome inhibitors and may facilitate strategies to enhance therapeutic efficacy in pancreatic and other solid tumors.