c-Myc and Caspase-2 Are Involved in Activating Bax during Cytotoxic Drug-induced Apoptosis^*

EXPERIMENTAL PROCEDURES

Cell Lines—Human lung adenocarcinoma A-549 cells and osteosarcoma U2-OS cells (ATCC) were grown and maintained respectively in F-12K and McCoy's 5A medium (ATCC) containing 10% fetal bovine serum (Invitrogen) and 1% penicillin-streptomycin (Cellgro) in 5% CO₂ and 95% humidified air at 37 °C.

Transfection with siRNAs—Apaf-1 siRNA duplex with the target sequence 5′-AAT TGG TGC ACT TTT ACG TGA-3′ (nucleotides 978–998) and caspase-2 siRNA duplex with the target sequence 5′-AAA CAG CTG TTG TTG AGC GAA-3′ (nucleotides 94–114) were purchased from Dharmacon. Double-stranded c-Myc siRNA and negative control siRNA were purchased from Ambion (silencer™ c-Myc siRNA control, catalog number 4604). A-549 cells were cultured in 100-mm dishes (Falcon) in 12 ml of F-12K medium complemented with 10% fetal bovine serum and transfected at 30% of confluency by adding 90 μl of siPORT™ amine transfection agent (Ambion) and 60 μl of 20 μm Apaf-1 or caspase-2 siRNA or 16 μl of 75 μm c-Myc or negative control siRNA (final concentration 100 nm). Control cells were treated with 90 μl of siPORT™ amine transfection agent alone. Cells were rinsed with the medium after 20 h of incubation and then maintained in culture for an additional 28 h. U2-OS cells were cultured and transfected similarly as A-549 cells except that 60 μl of siPORT™ amine transfection agent was added to each 100-mm dish.

Analysis of Protein Expression—Following siRNA transfection for 48 h, cells were harvested after trypsinizing for 10 min and lysed, and proteins were extracted and quantitated as described (25, 26). Equal amounts of lysate protein (50 μg) were subjected to SDS-PAGE followed by immunoblotting with Apaf-1, caspase-2, or Bax antibodies (Pharmingen) and c-Myc, caspase-9, or actin antibodies (Santa Cruz Biotechnology).

Drug-induced Apoptosis—Following siRNA transfection for 48 h, A549 cells and U2-OS cells were treated with 40 or 30 μm cisplatin, respectively, or 50 μm etoposide (Sigma) for 24 h. Cell viability was measured using the trypan blue exclusion method as described (27, 28).

Caspase-3 Activity Assay—Following siRNA transfection for 48 h, cells were treated with 40 μm cisplatin or 50 μm etoposide for 20 h. Apo-ONE™ homogeneous caspase-3 assay was performed following the manufacturer's manual (Promega) as described (29).

Cytochrome c Release—Following siRNA transfection for 48 h, cells were treated with cisplatin for 18 h and then harvested after trypsinizing for 10 min. To assess cytochrome c release, the cytosolic fraction was obtained using a digitonin-based subcellular fractionation technique as described (30–32). Equal amounts of cytosolic fractions (80 μg) were subjected to SDS-PAGE followed by immunoblotting with an anti-cytochrome c antibody (Pharmingen).

Alkali Extraction of Proteins Peripherally Associated with the Outer Mitochondrial Membranes—Following siRNA transfection for 48 h, cells were treated with cisplatin for 16 h and then harvested after trypsinizing for 10 min. Isotonic subcellular fractionation was performed to collect the heavy membrane fraction as described (13, 14). The heavy membrane fraction (0.5 mg of protein) was resuspended in freshly prepared 0.1 m Na₂CO₃, pH 11.5, and incubated on ice for 30 min (33). The alkali-extracted membrane was then centrifuged at 200,000 × g for 30 min, and the pellet was lysed. Equal amounts of lysate protein (10 μg) were subjected to SDS-PAGE followed by immunoblotting with Bax antibody (Pharmingen) and prohibitin antibody.

Cross-linking of Bax in the Outer Mitochondrial Membranes— Following siRNA transfection for 48 h, cells were treated with cisplatin for 18 h and then harvested after trypsinizing for 10 min. Isotonic subcellular fractionation was performed to collect the heavy membrane fraction as described (13, 14). The heavy membranes (0.25 mg of protein) were resuspended in 200 μl of mitochondrial buffer. The cross-linker disuccinimidyl suberate (Pierce) was added to a final concentration of 5 mm (13, 14). Incubation was carried out at room temperature for 30 min, after which the cross-linking reaction was quenched by the addition of Tris-HCl, pH 8.0, to a final concentration of 20 mm. The heavy membrane fraction was then lysed, and 40 μg of lysate protein was subjected to SDS-PAGE and Western blotting with an anti-Bax antibody (Cell Signaling).

RESULTS

Transfection of siRNA Specifically Knocks Down the Expression of c-Myc, Caspase-2, and Apaf-1 in A549 and U2-OS Cells—Cytotoxic stress, such as DNA damage, induces apoptosis through the mitochondrial pathway, and proapoptotic Bax (or Bak) is required to disrupt mitochondrial membrane integrity to initiate stress-induced apoptosis (1–3). Recent reports show that caspase-2 and c-Myc may be involved in destabilizing mitochondrial integrity (16, 17, 23, 24). To test whether c-Myc and caspase-2 are involved in activating Bax during stress-induced apoptosis, an siRNA strategy was utilized to knock down the expression of c-Myc, caspase-2, and Apaf-1 in human lung carcinoma A-549 and osteosarcoma U2-OS cells. After 48 h of transient transfection, endogenous c-Myc, caspase-2, and Apaf-1 levels were reduced 80–90% (Fig. 1). Demonstrating specificity of this system, a negative control siRNA (nRNA), with no sequence homology to any known gene, had no effect on the expression of the proteins measured, and none of the siRNAs used affected the expression of caspase-9. Although it was reported that Bax may be a transcriptional target of c-Myc and mediator of c-Myc-induced apoptosis (22), Bax expression is not affected by knockdown of c-Myc, caspase-2, or Apaf-1 in this experimental system (Fig. 1). These data indicate that c-Myc does not affect steady-state levels of Bax under these experimental conditions.

Expression of c-Myc, Caspase-2, and Apaf-1 Is Essential for Drug-induced Apoptosis—Transient transfection with siRNAs alone has no effect on the viability of A-549 and U2-OS cells (data not shown). Cisplatin and etoposide, DNA-damaging drugs that are used clinically in cancer chemotherapy, were used to treat these cells. When treated with cisplatin for 24 h, about 50% of the cells died, as measured by trypan blue exclusion (Fig. 2A). However, when c-Myc expression is knocked down, cisplatin-induced cell death is significantly inhibited since only 10–15% cells die. Similarly, when Apaf-1 or caspase-2 is knocked down, cisplatin-induced cell death is reduced to about 20%. Importantly, the negative control siRNA (nRNA) has no effect on cisplatin-induced apoptosis (Fig. 2A). Therefore, we can conclude that optimal cell death requires c-Myc, caspase-2, and Apaf-1. Activation of procaspases-9 and -3 is a key event in activating the execution phase of intrinsic apoptosis. To test whether these cells underwent apoptosis by this route, caspase-3 activity was measured following siRNA transfection and cisplatin treatment. Results reveal that caspase-3 activation is significantly inhibited when c-Myc, caspase-2, or Apaf-1 is knocked down (Fig. 2B), suggesting that c-Myc, caspase-2, and Apaf-1 are three important mediators of drug-induced caspase-3 activation and apoptosis. Similar results are observed for single agent treatment with etoposide (data not shown).

c-Myc and Caspase-2, but Not Apaf-1, Are Important for Cytochrome c Release—During stress-induced activation of the intrinsic apoptotic pathway, caspase-9/-3 activation is initiated by cytochrome c release from the mitochondria into the cytosol (7, 34). Cytochrome c facilitates the formation the apoptosome that contains Apaf-1 that recruits and activates procaspase-9, which in turn cleaves and activates its downstream executioner, procaspase-3 (35, 36). Using a digitonin-based subcellular fractionation technique (Fig. 3), results indicate that cisplatin-induced cytochrome c release is significantly decreased when either c-Myc or caspase-2 is knocked down, whereas, as expected, knockdown of Apaf-1 has no effect on cytochrome c release (Fig. 3). These results demonstrate that both c-Myc and caspase-2 are essential for permeabilization and dysregulation of mitochondria in this experimental system.

Recent reports proposed cytoplasmic proteins 14-3-3θ and Ku-70 as Bax inhibitors that normally sequester Bax in the cytosol, thereby preventing Bax-induced cytochrome c release, and that 14-3-3θ and Ku-70 may be directly cleaved by some caspases to release Bax upon apoptotic stimulation (37, 38). However, we have not observed any significant change of 14-3-3θ and Ku-70 expression levels that may suggest specific cleavage or degradation (Fig. 3).

Caspase-2, but Neither c-Myc nor Apaf-1, Is Important for Bax Integration into the Outer Mitochondrial Membranes—In non-stressed cells, Bax is located in the cytosol, where it may be peripherally but not integrally associated with the OMM (8). Following a death stimulus, Bax is translocated and integrated into the OMM through conformational exposure of its C-terminal transmembrane domain (8, 39). This process has been shown to precede mitochondrial dysfunction resulting in cytochrome c release into the cytosol (33, 40). It has been shown that alkali extraction of mitochondrial membranes will denature any peripherally associated (i.e. non-integrated) proteins that may otherwise pellet with the isolated membranes and thereby appear to be integral to the OMM. Importantly, this alkali extraction strategy can distinguish whether Bax is peripherally associated with or integrally inserted into the OMM (33). Studies were performed using this technique to determine whether caspase-2 or c-Myc affects the membrane integration of Bax during drug treatment. Results reveal that when caspase-2 is knocked down, the membrane-integrated Bax is decreased by 60–70% (Fig. 4), whereas knockdown of either c-Myc or Apaf-1 has no effect on Bax integration into the OMM. As a control, prohibitin is normally located in the inner mitochondrial membrane and is not affected in this experimental system (29).

c-Myc Is Critical for Bax Oligomerization after Integration into the Outer Mitochondrial Membranes—It has been reported that c-Myc can promote apoptosis by destabilizing mitochondrial integrity with cytochrome c release (23, 24). However, the mechanism is not clear. In particular, it remains controversial whether c-Myc potentiates apoptosis by stimulating Bax activity at the mitochondria or whether c-Myc functionally cooperates with Bax but works independently of Bax activation to induce apoptosis (23, 24). It is also controversial whether c-Myc activation can stimulate a detectable conformational change in Bax in mitochondria (23, 24). Interestingly, it has been consistently demonstrated that Bax is present as a dimer or oligomer following insertion into the OMM and that Bax oligomerization may be required for any channel-forming activity in artificial liposomes (13–15). Thus, the Bax oligomers are likely the functionally active form of Bax that facilitates dysfunction of the OMM with cytochrome c release. To test whether c-Myc or caspase-2 affects Bax oligomerization, mitochondria were isolated and subjected to cross-linking. Following cisplatin treatment for 18 h, Bax dimers, trimers, and tetramers are clearly detected in isolated mitochondrial membranes (Fig. 5). When c-Myc is knocked down, Bax oligomerization is significantly reduced. In contrast, knockdown of Apaf-1 has no effect on Bax oligomerization, although apoptosis is inhibited. These results suggest that c-Myc, either directly or through an as yet unidentified protein or cofactor, may indirectly facilitate Bax oligomerization in the OMM. These results are consistent with the report that c-Myc is required for Bax activation at the mitochondria, as detected using the conformation-specific Bax 6A7 antibody (23). Furthermore, Bax oligomerization is also decreased when caspase-2 is knocked down (Fig. 5). Since caspase-2, but not c-Myc, is critical for Bax integration (Fig. 4), these results suggest that Bax oligomerization occurs following or during Bax membrane insertion. Previously, we demonstrated that the N-terminal cleavage of full-length p21-Bax to p18-Bax also facilitates Bax oligomerization and enhances cytochrome c release (28). In this study, potential dimers between p21-Bax and p18-Bax are detected (Fig. 5). Cross-linking studies also reveal that the only antibody that can successfully detect a Bax dimer or oligomer is the one developed against the Bax N-terminal domain (Cell Signaling). Since the N-terminal domain of p21Bax is cleaved during the generation of p18-Bax, no homodimer of p18-Bax will be detected, although it may actually exist.

DISCUSSION

Multiple signals influence mitochondria during apoptosis, and activation of proapoptotic Bax (or Bak) has been shown to be an essential gateway to mitochondrial dysfunction that is required for activation of the intrinsic apoptotic pathway (3). Mounting evidence also indicates that cytosolic Bax can integrate into and oligomerize in the OMM with the potential formation of death pores that facilitate cytochrome c release (4–7, 13–15). However, it was unclear how the proapoptotic activity of Bax is activated to initiate mitochondrial dysfunction. Recent reports propose that cytotoxic stress, such as DNA damage, may cause activation of caspase-2 and that caspase-2 is required for the permeabilization of mitochondria, leading to cytochrome c release (16, 17). It has also been reported that caspase-2 can be activated by the formation of a multimeric protein complex that is independent of cytochrome c and Apaf-1 and that recruitment of caspase-2 to such a complex would be sufficient to mediate its activation (41). In that system, caspase-2 was thought to play a role similar to that of caspase-8 in the extrinsic or death receptor-induced apoptotic pathway, as immediately upstream of mitochondrial dysregulation. It was argued that stress-induced activation of the intrinsic apoptotic pathway in some cells might even proceed in a manner that is conceptually similar to that of cytokine-induced, receptor-mediated apoptosis (17). That is, cytokines such as tumor necrosis factor-α may activate initiator caspases (e.g. caspase-8) by assembling a death receptor complex. In this case, mitochondrial permeabilization as a result of caspase-8 activation could be seen to accelerate cell death by amplifying the caspase activity (42, 43). In this study, we examined the role of caspase-2 in drug-induced apoptosis in two human cancer cell lines. Consistent with the reports mentioned above, our results indicate that caspase-2 is necessary for efficient cytochrome c release from mitochondria and subsequent caspase-3 activation following application of chemotherapy. These results support the notion that caspase-2 functionally links or senses genotoxic DNA damage that leads to the compensatory engagement of the mitochondrial apoptotic pathway (16). Importantly, when Apaf-1 is knocked down by siRNA, caspase-3 activation is largely blocked (Fig. 2B), confirming the findings by others that Apaf-1 and the apoptosome are required for activation of caspase-3. These results extend those and suggest that caspase-2 is not sufficient to directly activate downstream executioner caspases like caspase-3. Therefore, activation of the mitochondrial pathway is necessary.

The characteristics of drug-induced apoptosis in this system are reminiscent of the death receptor-mediated apoptosis in so-called type II cells where death receptor signaling to effector caspases amplifies the death process by recruiting mitochondria into an amplification loop that involves caspase-8 cleavage of Bid, a Bax activator that can induce Bax insertion into the OMM and lead to cytochrome c release (42, 43). Since Bax has been shown to be essential for mitochondrial dysfunction and activation of the intrinsic apoptotic pathway (3), we tested the role of caspase-2 in Bax activation. Our results demonstrate that caspase-2 is necessary for efficient Bax integration into the mitochondrial membrane (Fig. 4), findings consistent with those reported for caspase-2 being required to translocate Bax to mitochondria (17). Taken together, these data suggest that Bax activation may be the critical linkage point between caspase-2 activation and mitochondrial dysfunction. However, how caspase-2 mediates Bax translocation and/or integration into the mitochondrial membrane is not yet clear. It was recently reported that cytoplasmic protein 14-3-3θ or Ku-70 may act to inhibit Bax by directly binding and sequestering Bax in the cytosol and that 14-3-3θ or Ku-70 may be directly cleaved by some protease/caspase to release Bax upon apoptotic stimulation, now allowing it to become integral to the OMM and disrupt mitochondrial integrity (37, 38). Therefore, we tested whether caspase-2 can activate Bax by degrading 14-3-3θ or Ku-70 but did not observe any significant degradation of 14-3-3θ or Ku-70 to indicate their role in Bax activation, regardless of whether or not caspase-2 is knocked down (Fig. 3).

It has been reported that c-Myc oncoprotein promotes apoptosis by destabilizing mitochondrial integrity (18). Interestingly, Bax-deficient MEF cells have been found to be markedly resistant to c-Myc-induced apoptosis, and c-Myc is unable to induce cytochrome c release in Bax-deficient MEFs (24). In contrast, MEFs deficient in either Bak or Bid remain sensitive to c-Myc-induced apoptosis (24), excluding their role in the c-Myc apoptosis mechanism. Importantly for the mechanism, ectopic expression of Bax in Bax-deficient MEFs restores sensitivity to c-Myc-induced apoptosis, indicating a requisite role for Bax (24). However, Bax does not appear to be a transcriptional target of c-Myc, at least in this system, because c-Myc knockdown has no effect on steady-state Bax expression levels in the cells used or in cells tested by others including Rat-1 fibroblasts and MEFs (23, 24, 44, 45). Thus, we propose that c-Myc may somehow indirectly trigger activation of existing Bax. Several findings support this notion. First, we find that neither Bax expression nor the integration of Bax into the mitochondrial membranes is influenced by c-Myc knockdown (Figs. 1 and 4). Second, c-Myc activation does not affect Bax expression or its subcellular localization (23, 24, 44). Third, when endogenous c-Myc is knocked down by 80–90%, Bax oligomerization is significantly reduced, and cytochrome c is not released (Figs. 3 and 5). Importantly, in this experimental circumstance, cytotoxic drug-induced apoptosis is blocked (Fig. 2).

It has been reported that etoposide or taxol-induced Bax activation, as detected by the conformation-specific 6A7 antibody, does not occur in c-Myc null Rat-1 fibroblasts, whereas reintroduction of c-Myc into the c-Myc null cells restores Bax activation, cytochrome c release, and apoptosis (23). Consistent with this report, our data support a role for c-Myc in activating Bax because knockdown of c-Myc significantly reduces Bax oligomerization following its integration into the mitochondrial membranes (Fig. 5). This novel experimental finding supports a model whereby Bax is an effector molecule cooperating with c-Myc during cytotoxic drug-induced apoptosis. However, since our co-immunoprecipitation studies did not detect any direct interaction between c-Myc and Bax, or between c-Myc and other Bcl-2 family members tested, including Bcl-X_L, Bak, Bid, and Bad (data not shown), we suspect that it is unlikely that c-Myc directly activates Bax. Alternatively, although it is formally possible that c-Myc may indirectly regulate Bax activity through transcriptional modulation of a Bax activator that functions at the mitochondria to facilitate Bax oligomerization (Fig. 6), it is also possible that another as yet unidentified protein may be involved in such a process. Thus, further studies are required to identify such a potential c-Myc regulated activator of Bax.

In sum, this study has examined several key steps required for mitochondrial dysfunction and activation of the intrinsic apoptotic pathway. Our findings provide novel biochemical evidence that c-Myc and caspase-2 can functionally cooperate to promote drug-induced apoptosis by activating Bax at the level of the mitochondria. That is, caspase-2 is important for Bax integration into the outer mitochondrial membrane, and c-Myc is critical, but apparently acts indirectly, to facilitate Bax oligomerization. Furthermore, neither caspase-2 nor c-Myc is sufficient alone to fully activate Bax to initiate apoptosis. Thus, the blockade of these two rate-limiting steps, either by RNA interference or by small molecule inhibitor, may efficiently inhibit the apoptotic process and thus might have merit to suggest a novel therapeutic approach in diseases where apoptosis plays a pathological role.