Hydrogen Peroxide-Mediated Cytosolic Acidification Is a Signal for Mitochondrial Translocation of Bax during Drug-Induced Apoptosis of Tumor Cells

Apoptotic execution is orchestrated by intricate networking between caspases and apoptogenic factors released from the mitochondria (1). A growing body of evidence seems to favor the involvement of intracellular reactive oxygen species at some point during apoptotic execution (2, 3, 4, 5, 6). These observations become more important considering the critical role of the mitochondria during apoptosis and the fact that mitochondria have been implicated directly or indirectly as the prime source of reactive oxygen species during drug-induced apoptosis (2, 5, 7, 8). However, it is still unclear whether reactive oxygen species generation is a critical initial trigger or a downstream effect of caspase-mediated mitochondrial damage. To that end, our findings have highlighted the regulatory role of intracellular reactive oxygen species in the apoptotic pathway (6, 9, 10, 11). We showed that intracellular increase in H₂O₂ was a critical effector mechanism during drug-induced apoptosis of human tumor cells (5). This increase in H₂O₂ was responsible for early cytosolic acidification, thus creating an environment conducive for caspase activation. Although our data and a number of other reports describing H₂O₂-mediated apoptosis, e.g., the inhibitory effect of Bcl-2 overexpression on H₂O₂-induced death signaling (12, 13, 14) and the up-regulation of the proapoptotic protein Bax in some systems (15), point to the mitochondria as the target organelle, the upstream events leading to the engagement of the mitochondria remain clouded.

The role of the proapoptotic protein Bax in the recruitment of the mitochondria has been well established (16, 17, 18). A well-accepted model is the death receptor-mediated activation of caspase 8 that triggers Bid processing (19, 20, 21). Truncated Bid can then signal translocation of cytosolic Bax to the mitochondria (19), where it can form homo- or heterodimers with other Bcl-2 family members. The resultant conformational change in Bax can result in channel formation that could mediate egress of proteins, such as cytochrome c, from the mitochondrial intermembrane space (22, 23, 24, 25), thereby activating downstream execution. Hence, translocation of Bax is a critical signal for the involvement of the mitochondrial death machinery, and redistribution of Bax has been reported with a variety of apoptotic stimuli (22, 23, 26, 27, 28). The critical involvement of Bax during drug-induced apoptosis is further supported by the relative insensitivity of cells lacking Bax expression to drug treatment (29) and observation that Bax−/− cells when xenografted in mice are deficient in apoptotic signaling.

In the light of these findings, the objectives of this study were 2-fold: (1) to decipher the role of Bax in H₂O₂-mediated apoptosis, and (2) to investigate whether the intracellular increase in H₂O₂ was a signal for Bax translocation during drug-induced apoptosis of tumor cells. Here, we report that H₂O₂-induced apoptosis is inhibited in colorectal carcinoma cells lacking Bax (HCT116 Bax−/−), unlike HCT116 Bax+/− cells. Exposure of HCT 116 Bax+/− or HL60 human leukemia cells to H₂O₂ or its intracellular generation during drug-induced apoptosis signals translocation of Bax to the mitochondria, which is mediated by cytosolic acidification.

The human leukemia HL60 and CEM cell lines were purchased from American Type Culture Collection (Rockville, MD) and maintained in RPMI 1640, supplemented with 10% fetal bovine serum, 1% l-glutamine, and 1% S-Penicillin. HCT116 Bax+/− and Bax−/− cell lines were generous gifts from Dr. Vogelstein at Johns Hopkins University (Baltimore, MD). HCT116 cell lines were maintained in McCoy’s 5A (Invitrogen Life Technologies, Inc., Carlsbad, CA, USA), supplemented with 10% fetal bovine serum, 1% l-glutamine, and 1% S-Penicillin. All cell lines were maintained in a 37°C incubator with 5% CO₂. Apoptosis was induced by exposure of cells (1 × 10⁶/mL) to H₂O₂ (100–500 μmol/L) or the anticancer agent merodantoin (C1; 50 μg/mL; ref. 30) for 4 to 18 hours. Cell survival was determined by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay as described previously (31).

Subcellular fractions were obtained as described previously (5). Briefly, 20 × 10⁶ cells were washed with ice-chilled 1 × PBS at 1200 × g. Cell pellets were resuspended in 500 μL of extraction buffer [200 mmol/L mannitol, 68 mmol/L sucrose, 50 mmol/L Pipes (pH 7.4), 50 mmol/L KCl, 5 mmol/L EGTA, 2 mmol/L MgCl₂, 1 mmol/L DTT, and protease inhibitor mixture in double distilled H₂O] and incubated at 4°C for 20 minutes, followed by Dounce homogenization. The homogenate was centrifuged at 150 × g for 5 minutes at 4°C. The supernatant was additionally centrifuged at 14,000 × g for 10 minutes (fraction enriched with intact mitochondria). The supernatant from the last centrifugation was used as the cytosolic fraction. Purity of the mitochondrial fractions was confirmed by Western blot analysis using a monoclonal antibody that recognizes the mitochondrial-specific protein MnSOD.

Caspase 3, 8, and 9 activities were assayed by using AFC-conjugated substrates supplied by Bio-Rad Laboratories (Hercules, CA). Cells (1 × 10⁶/mL) were exposed to H₂O₂ (100–500 μmol/L) or C1 (50 μg/mL) 4 to 24 hours, washed twice with 1 × PBS, resuspended in 50 μL of chilled cell lysis buffer (provided by the supplier), and incubated on ice for 10 minutes. Fifty μL of 2 × reaction buffer (10 mmol/L HEPES, 2 mmol/L EDTA, 10 mmol/L KCl, 1.5 mmol/L MgCl₂, and 10 mmol/L DTT) and 6 μL of the fluorogenic caspase-specific substrate (DEVD-AFC for caspase 3, LETD-AFC for caspase 8, and LEHD-AFC for caspase 9) were added to each sample and incubated at 37°C for 1 hour. Protease activity was determined by the relative fluorescence intensity at 505 nm after excitation at 400 nm using a spectrofluorimeter (TECAN Spectrofluor Plus, Maennedorf, Switzerland). Results are shown as fold increase (× increase) in activity relative to untreated cells (1×). Caspase 3 activation was also assessed by staining cells with an affinity-purified antibody that recognizes the processed (active) form of caspase 3 (tagged with phycoerythrin; BD PharMingen) and analyzed by flow cytometry.

Briefly, 1 × 10⁶ cells/mL were triggered with H₂O₂ or C1 for 24 hours, fixed with 70% EtOH, and stained with propidium iodide for DNA content analysis as described elsewhere (5). Events (≥10,000) were analyzed by flow cytometry with the excitation set at 488 nm and emission at 610 nm. Data are shown as a percentage of cells with subdiploid DNA and are mean ± SD of three independent observations.

In addition, apoptosis on exposure of cells to H₂O₂ or C1 was also verified by assaying exposure of phosphatidylserine using the Apoalert Annexin-V kit (Clontech) and analyzed by flow cytometry using excitation and emission wavelengths at 488 and 525 nm, respectively.

Cells were exposed to the apoptotic trigger for 15 minutes to 12 hours, loaded with 5-(and-6)-chloromethyl-2′,7′-dichlorofluorescin diacetate (5 μmol/L) at 37°C for 15 minutes, and analyzed by flow cytometry (Coulter EPICS Elite ESP) using an excitation wavelength of 488 nm, as described previously (32).

In addition, intracellular reactive oxygen species production was assessed by a lucigenin-based chemiluminescence assay as described previously (16) or flow cytometry using the sensitive probe hydroethidine (2 μmol/L). Chemiluminescence was monitored for 60 seconds in a TD-20/20 Luminometer (Turner Designs, Sunnyvale, CA). Lucigenin is a widely used and validated chemiluminescent detector of intracellular O₂⁻ in biological systems (33, 34). Data are shown as relative light units (RLU) per microgram of protein (RLU μg⁻¹ protein) ± SD from three to six independent measurements. Protein concentration was determined using the Coomassie Plus protein assay reagent from Pierce (Pierce Chemical Co., Rockford, IL).

Cytochrome c release was assessed by Western blot analysis of cytosolic extracts from 30 × 10⁶ cells as described previously (5) using anti-cytochrome c (7H8.2C12, PharMingen, San Diego, CA). Signal was detected by the Super Signal Substrate Western Blotting kit (Pierce).

Potential-sensitive probe 3, 3′ dihexyloxacarbocyanine iodide (DiOC₆) was used to measure mitochondrial Δψ_m as described previously (31). Briefly, 1 × 10⁶ cells were incubated with 3,3′DiOC₆ (40 nmol/L) for 15 minutes at 37°C and immediately analyzed in Epic Profile flow cytometer with excitation set at 488 nm. Data were analyzed for 10,000 events using the WinMDI software.

For analysis of poly(ADP-ribose) polymerase cleavage, lysates from 2 × 10⁶ cells were prepared in sample buffer [50 mmol/L Tris/HCl (pH 6.8), 6 mol/L urea, 3% SDS, 0.003% bromphenol blue, and 6% β-mercaptoethanol] and subjected to Western blot analysis using anti-poly(ADP-ribose) polymerase (clone C-2–10, PharMingen) as described (30). For Western blot analysis of Bax, cells (2 × 10⁶) were lysed by adding 100 μL of chilled 1 × radioimmunoprecipitation assay buffer lysis buffer, and 50 μg of protein were subjected to 15% PAGE and transferred to polyvinylidene difluoride as above. Alternatively for analysis of Bax dimerization/multimerization, cell lysates were subjected to 10% native gel electrophoresis. Membranes were exposed to 1:2,000 dilution of mouse monoclonal anti-Bax antibody (clone 6A7, BD Pharmigen, San Diego, CA) at 25°C for 2 hours, followed by 1:5,000 dilution of goat antimouse IgG-horseradish peroxidase. The anti-Bax antibody (6A7) recognizes epitopes that are in the vicinity of the dimerization domains of Bax (AA 12–24; ref. 35). Western blot analysis for Bid cleavage was performed on whole cell lysates using a rabbit polyclonal anti-Bid IgG (Biovision Research Products, Paolo Alto, CA) that recognizes the M_r 22,000 full-length Bid. Chemiluminescence was detected as described above.

Transient transfection of HCT116 Bax−/− cells with Bax was performed using pcDNA3-Bax (generously provided by Prof. Stanley Korsmeyer, Boston, MA). HCT116 Bax−/− (2 × 10⁵) cells were seeded per six-well plate in 2 mL of 10% McCoy’s Medium and incubated at 37°C for 24 hours. The medium was then removed, and cells were resuspended in 1.5 mL of normal growth medium. Cells were cotransfected with a β- galactosidase-containing plasmid (p-cytomegalovirus-β galactosidase) at a ratio of 3.5 μg of pCDNA3-Bax:0.5 μg of p-cytomegalovirus-β-galactosidase as described previously (36). Luminescence was recorded with a Luminometer (TD20/20; Turner Designs). The relative light units were calculated per microgram of protein.

For measurement of cytosolic pH, cells were loaded with 10 μmol/L BCECF-AM [2′,7′-bis(2-carboxyethyl)-5,6-carboxyfluorescein; Sigma, St. Louis, MO], and the fluorescence ratio of 525:610 nm was used to derive cytosolic pH using a standard pH calibration curve as described previously (5).

To manipulate cytosolic pH, cells were incubated with a known inhibitor of the Na+-H+ antiporter, methylamiloride (10–40 μmol/L), for 4 hours before the measurements. Alternatively, cells were preincubated with 4 to 8 μmol/L resveratrol for 2 hours before exposure to the apoptotic stimuli and pH. In our recent study (37), we have demonstrated the inhibitory effect of low doses of resveratrol on drug-induced acidification.

The sensitivity of human colon carcinoma HCT116 Bax+/− and HCT116 Bax−/− cells (Fig. 1,A) to H₂O₂ (100 μmol/L) for 4 to 24 hours was determined. Unlike HCT116 Bax+/− cells, Bax−/− cells were resistant to the effect of H₂O₂ (Fig. 1,B). H₂O₂ triggered significant increases in caspases 9 and 3 activities (but no caspase 8 activation or cleavage of Bid; Fig. 1,C), increase in membrane phosphatidylserine exposure and processing of caspase 3 (Fig. 1,D), cleavage of caspase 3 substrate poly(ADP-ribose) polymerase (Fig. 1,E), and increase in subdiploid DNA fraction (Fig. 1,E) in Bax+/− cells, whereas little effects on caspase activation (Fig. 1, C and E), and DNA fragmentation (Fig. 1,F) were observed in Bax−/− cells. H₂O₂ also induced cytosolic translocation of cytochrome c in HCT 116 Bax+/− cells (inhibited by catalase) but not the Bax−/− variant (Fig. 1,G). Furthermore, transient transfection of Bax−/− cells with a vector containing full-length Bax (pcDNA3-Bax) restored their sensitivity to H₂O₂ (Fig. 1 H), thus highlighting the importance of Bax in H₂O₂-induced apoptosis.

H₂O₂ has been shown to induce up-regulation of Bax and a shift in the intracellular Bax:Bcl-2 ratio in some systems (38). Although exposure of HCT116 Bax+/− cells to H₂O₂ did not significantly change the Bax:Bcl-2 ratio (Fig. 2,A), analysis of subcellular fractions clearly showed translocation of Bax from the cytosol to the mitochondria (Fig. 2,B). Mitochondrial localization of Bax could be detected 6 hours after exposure to H₂O₂ (data not shown) and was maximal at 12 hours after treatment (Fig. 2,B). Importantly, Bax translocation was inhibited by the H₂O₂ scavenger catalase (1000 units/mL), additionally supporting the role of H₂O₂ in mitochondrial translocation of Bax. Contrarily, preincubation with the general caspase inhibitor ZVAD-fmk (50 μmol/L) did not inhibit H₂O₂-induced Bax translocation (Fig. 2,C). Similar to HCT116 cells, HL60 and CEM leukemia cells underwent apoptosis on exposure to H₂O₂, as indicated by the appearance of subdiploid DNA, increases in caspase 3 and 9 activities, processing of caspase 3, and decrease in cell survival (Fig. 3, A–D). In addition, the involvement of the mitochondrial death pathway was evident by the drop in ΔΨ_m of cells exposed to H₂O₂ (Fig. 4,A) and cytosolic translocation of cytochrome c from the mitochondria (Fig. 4,B) that was significantly inhibited by catalase. More importantly, cytosolic Bax was redistributed to the mitochondrial fractions in HL60 and CEM cells, which could be inhibited by catalase (Fig. 4, C and D) but not ZVAD-fmk (Fig. 4 C). These data suggest a critical role for H₂O₂ in signaling Bax to the mitochondria but argue against the involvement of the caspase family in this pathway.

Having demonstrated that exogenously added H₂O₂ triggered translocation of Bax in tumor cells, we next asked whether this was also the mechanism used by anticancer drugs that trigger apoptosis via intracellular H₂O₂ production. HL60 cells were exposed to the novel experimental anticancer agent C1, a small (M_r = 242) synthetic compound, first isolated and purified on photo-oxidation of merocyanine 540. The chemical structure of this imidazole compound is N,N′-Dibutyl-thio-4,5-imidazolindion, and its apoptosis-inducing activity was reported previously by our group (30).

Exposure of HL60 cells to 50 μg/mL C1 for 4 hours resulted in a significant increase in intracellular H₂O₂ (Fig. 5,A, panel a). The intracellular generation of reactive oxygen species was additionally confirmed by a lucigenin-based chemiluminescence assay and flow cytometry on loading cells with hydroethidine. Results clearly indicate that exposure of cells to C1 for 2 to 8 hours resulted in a significant increase (1.4–2×) in intracellular reactive oxygen species compared with the cells treated with the carrier solvent (Fig. 5, panel b). Similarly, a right shift in the fluorescence of C1-treated cells loaded with hydroethidine (Fig. 5, panel c; mean fluorescence intensity shifted from 2.9 to 3.25) provided additional proof for drug-induced increases in intracellular reactive oxygen species generation. In addition, exposure to C1 resulted in a significant decrease in cell survival (Fig. 5,B), H₂O₂-dependent increase in subdiploid DNA fraction, increased processing of caspase 3, and cleavage of the caspase 3 substrate poly(ADP-ribose) polymerase (Fig. 5, C–E). In addition, our data clearly provide evidence that C1 induced apoptosis by targeting the mitochondria, as demonstrated by the drop in ΔΨ_m of cells, cytosolic translocation of cytochrome c, and increase in caspase 9 activity (Fig. 6, A–C). More importantly, Bax translocated from the cytosol to the mitochondria on exposure to C1, which could be significantly inhibited by catalase but not the general caspase inhibitor ZVAD-fmk (Fig. 6,D). In addition, native gel electrophoresis revealed that C1, similar to pure H₂O₂, triggered dimerization and multimerization of Bax (Fig. 6,E). A careful look at the blots in Fig. 6, D and E, showed that the intensity of the bands was much weaker in the catalase pretreated lanes in both the cases. This was additionally reinforced by a stronger inhibition of the oligomeric form by catalase (Fig. 6,E), thus corroborating the involvement of H₂O₂ in C1-induced Bax activation. The fact that preincubation of cells with H₂O₂ scavenger catalase significantly blocked all mitochondrial changes triggered by C1 in tumor cells (Fig. 6) underscores the critical role of H₂O₂ as the mediator of mitochondrial recruitment and apoptosis triggered by C1.

We have demonstrated previously that H₂O₂ added exogenously or triggered endogenously during drug-induced apoptosis is a stimulus for cytosolic acidification, thereby creating a permissive intracellular milieu for death execution (5, 6). In agreement with our earlier studies, here we show that C1 or H₂O₂ induced a significant drop in cytosolic pH, which could be inhibited by catalase (Fig. 7,A). Therefore, we questioned whether H₂O₂-dependent cytosolic acidification could be the stimulus for Bax translocation during drug-induced apoptosis. To prove that, we exploited the ability of low doses of the polyphenolic compound resveratrol to block H₂O₂ and drug-induced pH drops in human leukemia cells (37). Corroborating our recent findings, preincubation of HL60 cells for 2 hours with 4 to 8 μmol/L resveratrol inhibited acidification induced by C1 or H₂O₂ (Fig. 7,B). More interestingly, resveratrol exposure blocked mitochondrial translocation of Bax induced by either of the stimuli (Fig. 7, C and D). The effect of resveratrol on C1-induced Bax activation appears to be a function of its inhibitory effect on Bax dimerization (data not shown). It should be pointed out that resveratrol treatment for 0 to 24 hours did not alter the overall expression of either Bax or Bcl-2 (Fig. 7 E). Furthermore, we have reported previously that resveratrol did not elicit H₂O₂ scavenging activity at these low concentrations.

Stimulated by these findings and to verify that drop in cytosolic pH was a signal for Bax translocation, we assessed the effect of enforced intracellular acidification on Bax translocation. To that end, we made use of the pharmacological inhibitor of the pH regulator Na+/H+ antiporter, methylamiloride. As expected, methylamiloride (10–40 μmol/L) induced a significant drop in cytosolic pH (Fig. 8,A). Most interestingly, exposure of tumor cells to methylamiloride alone resulted in significant translocation of Bax to the mitochondria (Fig. 8,B), which was followed by cell death (Fig. 8,C). It should be pointed out that exposure of cells to methylamiloride for 0 to 12 hours did not alter the intracellular expression of Bax or Bcl-2 (Fig. 8,D). Indeed, enforced acidification of the milieu was a much stronger stimulus for translocation of Bax as shown in Fig. 8 B (virtually all Bax translocated as compared with ∼60% in case of C1), additionally consolidating the direct effect of a pH drop on mitochondrial recruitment of Bax. These data provide strong evidence to support cytosolic acidification, downstream of H₂O₂ production, as an effector mechanism for Bax translocation during drug-induced apoptosis of tumor cells.

Current opinion holds that the principal source of reactive oxygen species, including H₂O₂, during apoptotic signaling is the mitochondria (2, 5, 12). How does one reconcile with these findings given our observations that H₂O₂ acts upstream of the mitochondria to signal mitochondrial localization of Bax? To address this, we asked whether mitochondrial translocation of Bax resulted in a second wave of intracellular H₂O₂. HCT116 Bax+/− and Bax−/− cells were exposed to 100 μmol/L H₂O₂, and the intracellular level of H₂O₂ was determined at 30 minutes, 1, 6, and 12 hours. The initial increase in fluorescence observed in both Bax+/− and Bax−/− cells is indicative of cellular uptake of exogenously added H₂O₂ (Fig. 9,A, panel a). However, by 1 hour, intracellular H₂O₂ returned to the baseline in Bax−/− cells and remained unchanged for ≤12 hours (Fig. 9,A, panels b and c). On the contrary, Bax+/− cells showed a subsequent increase in intracellular H₂O₂ starting at 6 hours and significantly detectable even after 12 hours of the initial exposure to H₂O₂ (Fig. 9,A, panels b and c). Indeed, a summary of three independent experiments showing change (percentage of control cells) in mean fluorescence intensity of dichlorofluorescin clearly demonstrates a significant increase in mean fluorescence intensity at 6 and 12 hours after exposure to H₂O₂ in Bax+/− cells, whereas no detectable difference is seen in Bax−/− cells (Fig. 9,B). Intracellular reactive oxygen species production was further verified in Bax+/− and Bax−/− cell lines by hydroethidine staining and flow cytometry (Fig. 9,C). It is important to point out that the translocation of Bax to the mitochondria is also detected first at 6 hours after incubation with H₂O₂ with a maximum shift observed at 12-hour postincubation (Fig. 9,D). Furthermore, to confirm that intracellular reactive oxygen species generation on exposure to the apoptosis-inducing agent C1 was, indeed, upstream of Bax, HCT116Bax+/− and Bax−/− cells were exposed to C1, and reactive oxygen species production was assessed by loading cells with DCHF-DA or hydroethidine and analyzed by flow cytometry. Results clearly indicate that C1 triggered reactive oxygen species production in both cell lines as early as 15 minutes after exposure (Fig. 10); however, similar to H₂O₂, the second increase in intracellular reactive oxygen species was only observed in Bax+/− cells (data not shown). These results suggest a scenario where exogenous H₂O₂ or intracellular H₂O₂ produced by drug exposure triggers Bax translocation (via inducing cytosolic acidification), which then acts as a signal for the second wave of H₂O₂ production from the mitochondria, thereby reinforcing the death signal.

Our recent findings and many related studies strongly argue in favor of a signaling role for reactive oxygen species, such as H₂O₂ (5, 10, 11, 38, 39). Aside from the up-regulation of Bax on exposure of some cell types to H₂O₂, the inhibitory effect of Bcl-2 overexpression (12) supports the involvement of the mitochondrial death pathway in H₂O₂-induced apoptosis. However, as the mitochondria are a major source of intracellular reactive oxygen species, it is tempting to speculate that reactive oxygen species, such as H₂O₂, may function both upstream and downstream of the mitochondria.

Tumor cells lacking Bax (Bax−/−) are resistant to the effect of some anti-cancer drugs (29). In agreement with that, we show here that HCT116 Bax(−/−) cells are resistant to death induced by apoptotic concentrations of H₂O₂. Furthermore, transient transfection of Bax in HCT116 Bax(−/−) cells restores their sensitivity to H₂O₂, thus underscoring the critical role of Bax in H₂O₂-induced apoptosis. One probable mechanism underlying the differential sensitivity of Bax+/− and Bax−/− cells to H₂O₂ could be that H₂O₂ increases the expression of Bax or alters the cellular Bax:Bcl-2 ratio in Bax+/− cells, as suggested by some studies. However, our results show that the Bax:Bcl-2 ratio is not significantly altered on 4- to 24-hour exposure to H₂O₂. Interestingly, analysis of subcellular distribution of Bax (in HCT116, HL60, and CEM cells) revealed that Bax redistributed to the mitochondrial fraction from the cytosol on exposure to H₂O₂, which could be significantly blocked by the H₂O₂ scavenger catalase.

One could always argue that the changes elicited on exposure of cells to exogenous H₂O₂ may not have physiologic relevance, because such high levels of intracellular H₂O₂ are rarely observed in living cells. Therefore, to present a more real-life situation, we exploited the ability of certain anticancer drugs to increase intracellular production of reactive oxygen species, specifically H₂O₂ (5). Indeed, exposure of HCT116 Bax+/− or HL60 cells to a novel anticancer compound C1 resulted in an increase in intracellular H₂O₂ and translocation of Bax to the mitochondria. This translocation of Bax was inhibited by catalase, thus establishing the critical role of intracellular H₂O₂ in mitochondrial recruitment during drug-induced apoptosis of tumor cells.

Recruitment of Bax to the mitochondria during apoptotic signaling has been linked to the activation of upstream caspase 8 and caspase 8-mediated cleavage of the proapoptotic protein Bid (40). This is particularly true on ligation of death receptors, such as CD95 (Apo1/Fas). Incidentally, H₂O₂ and anticancer drugs have been shown to up-regulate the expression of the CD95 receptor or its ligand (CD95L) in some systems (41, 42). In these systems, blocking receptor signaling or downstream caspase activation could abrogate Bax translocation and, consequently, the death signaling circuitry downstream of the mitochondria. Interestingly, in our study, neither of the stimuli triggered up-regulation of the CD95 receptor (data not shown), and inhibition of caspase activation had no effect on the mitochondrial translocation of Bax. In addition, the relative lack of caspase 8 activation and absence of downstream Bid cleavage provide additional evidence in favor of a mechanism for signaling Bax to the mitochondria that is either parallel to or independent of death receptor and/or caspase activation. A similar mechanism of caspase-independent conformational change of Bax on triggering apoptosis has been reported previously (43).

Recent findings also implicate ceramide production in response to activation of the membrane sphingomyelinase in signaling Bax to the mitochondria (44, 45). Ceramide production is observed in a variety of apoptotic models, including drug-induced apoptosis (45, 46), and in response to H₂O₂ in tumor cells (3). However, inhibiting ceramide synthesis in our system did not affect the subcellular localization of Bax triggered by C1 or H₂O₂ (data not shown). Collectively, these data indicate that Bax translocation triggered in tumor cells during drug (C1)-induced apoptosis was a direct result of intracellular H₂O₂ production, independent of the upstream caspase 8 or ceramide pathways. Corroborating these findings is a recent study suggesting glutathione-dependent activation of Bax by oxidative stress in HeLa cells stably transfected with the cystic fibrosis transmembrane conductance regulator (47).

Cytosolic acidification is an early event in apoptosis and provides an intracellular milieu permissive for efficient death execution. In this regard, exposure of cells to H₂O₂ or drugs that trigger intracellular increase in H₂O₂ results in a significant drop in cytosolic pH (5). Accordingly, signals that inhibit apoptotic acidification impede death signaling as demonstrated in our recent study (37). Our results provide strong evidence that the link between H₂O₂ and Bax translocation could be the drop in cytosolic pH brought about by exposure of cells to exogenous H₂O₂ or endogenous production of H₂O₂ on drug exposure. Not only did inhibition of acidification triggered by H₂O₂ or C1 result in a significant reduction in mitochondrial localization of Bax, but also, most interestingly, clamping cytosolic pH to a more acidic range (∼7.1) directly induced Bax translocation in tumor cells. Because caspase inhibitors had no effect in our system, these results point to a direct effect of an acidic intracellular milieu in trafficking Bax from the cytosol to the mitochondria. It is interesting to note that one of the first studies to report a relationship between cytosolic pH and Bax translocation had concluded to the contrary, i.e., pH increase was shown to facilitate Bax translocation (48). However, in our system, the absence of any initial increase in cytosolic pH downstream of H₂O₂ or C1 even as early as 30 minutes after treatment (data not shown), fact that inhibition of pH drop abrogated Bax translocation, and direct effect of acidic intracellular milieu on Bax translocation provide strong evidence that acidic pH is a direct signal for Bax translocation even in the absence of a trigger. Furthermore, this phenomenon of H₂O₂-mediated, acidification-induced Bax translocation may not be exclusive to C1, because in our earlier study, we demonstrated the ability of commonly used chemotherapeutic drugs vincristine and daunorubicin to trigger an early increase in intracellular H₂O₂ (37). Interestingly, these compounds also induce cytosolic acidification and Bax-dependent apoptosis in tumor cells.⁵ Given these observations, it is highly likely that acidification-induced Bax activation may be a common mechanism during drug-induced apoptosis in tumor cells. Indeed, our findings linking an acidic milieu to Bax translocation are in agreement with a recent study demonstrating a relationship between cytosolic acidification and Bax translocation during staurosporine and tumor necrosis factor-induced apoptosis (49). Changes in pH affect the conformation of Bax, and a change in conformation facilitates Bax translocation to the mitochondria. In this regard, both positively and negatively charged residues contribute to the pH dependence of Bax conformation (48). It is plausible that similar to the effect of an alkaline pH, a considerable acidic shift in the pH could also induce a conformational change in Bax, thus making it more amenable for membrane insertion. Indeed, dimerization/multimerization of Bax in cells treated with C1 or H₂O₂ provide evidence to support conformational change in Bax in response to these apoptotic stimuli. Furthermore, considering the ability of drugs such as C1 to trigger early increases in intracellular H₂O₂, it is possible that the presence of scavenger catalase may have a more longer and sustained inhibitory effect on drug-induced Bax translocation.

Our data provide a novel mechanism for drug-induced translocation of Bax and argue in favor of an upstream role for H₂O₂-mediated cytosolic acidification. Earlier studies have implicated superoxide anion and H₂O₂ as probable species responsible for mitochondrial damage, such as lipid peroxidation of the mitochondrial membranes and induction of mitochondrial permeability transition (MPT) (50), changes that could explain the increase in the outer membrane permeability and leakage of pro-apoptotic proteins to the cytosol. However, most of these studies seem to imply that the mitochondrial burst of H₂O₂ is likely to be a downstream effector mechanism for the execution signal. Our data argue strongly in favor of a two-hit mechanism for drug-induced increases in intracellular H₂O₂. The initial increase (first hit) in intracellular H₂O₂ on drug treatment of tumor cells leads to the targeting of Bax to the mitochondria in a caspase-independent manner, thus resulting in the recruitment of the mitochondrial death pathway. Bax translocation can then bring about mitochondrial changes, such as induction of MPT, formation of a functional channel in the outer membrane, and as our data suggest, serve as a stimulus for the next burst of H₂O₂ (second hit) from the mitochondria. This could lead to peroxidative damage of mitochondrial lipids, such as cardiolipin, and facilitate the egress of cytochrome c, Smac/Diablo, or apoptosis inducing factor (AIF). This phenomenon of reactive oxygen species-induced reactive oxygen species production has been described before as a probable mechanism for the induction of MPT, and more recently, in a model of nerve growth factor-deprived neuronal cell death, Bax insertion has been proposed to induce reactive oxygen species burst, strong enough to trigger the release of cytochrome c (51). It should be pointed out that the initial early increase (within 15 minutes of exposure) in intracellular H₂O₂ occurs independent of Bax (both Bax+/− and Bax−/− cells); however, more importantly, the subsequent (second) increase is entirely dependent on Bax and its translocation to the mitochondria. One probable mechanism for Bax-induced H₂O₂ production could be that Bax translocation leads to cytochrome c release, which depletes the mitochondrial electron transport chain with the resultant leakage of electrons to molecular oxygen to form O₂⁻, which is then dismutated by the mitochondrial superoxide dismutase to H₂O₂.

Taken together, our data demonstrate a novel mechanism for mitochondrial recruitment of the proapoptotic protein Bax and provide a mechanistic explanation to connect extramitochondrial death circuitry to the mitochondria, the point of convergence for effective drug-induced death signals.