Intracellular Acidification Triggered by Mitochondrial-derived Hydrogen Peroxide Is an Effector Mechanism for Drug-induced Apoptosis in Tumor Cells*

We recently showed that two photoproducts of merocyanine 540, C2 and C5, triggered cytochrome C release; however, C5 was inefficient in inducing caspase activity and apoptosis in leukemia cells, unlike C2. Here we show that HL60 cells acidified upon exposure to C2 but not C5. The intracellular drop in pH and caspase activation were dependent upon hydrogen peroxide production, and were inhibited by scavengers of hydrogen peroxide. On the contrary, caspase inhibitors did not block hydrogen peroxide production. In turn, increased intracellular hydrogen peroxide concentration was downstream of superoxide anion produced within 2 h of exposure to C2. Inhibitor of NADPH oxidase diphenyleneiodonium neither inhibited superoxide production nor caspase activation triggered by C2. However, exposure of purified mitochondria to C2 resulted in significantly increased superoxide production. Furthermore, cytochrome C release from isolated mitochondria induced by C2 was completely inhibited in the presence of scavengers of hydrogen peroxide. Contrarily, scavenging hydrogen peroxide had no effect on the cyclosporin A-sensitive mitochondrial permeability transition induced by C5. Our data suggest a scenario where drug-induced hydrogen peroxide production induces intracellular acidification and release of cytochrome C, independent of the inner membrane pore, thereby creating an intracellular environment permissive for caspase activation.

It is now well established that anti-cancer drugs kill tumor cells by activating essential components of the apoptotic pathway (1)(2)(3)(4)(5)(6)(7). In this regard the pivotal role of intracellular caspases (8) and mitochondrial-derived pro-apoptotic factors, such as cytochrome c and apoptosis-inducing factor (9), has been highlighted in recent communications (6, 10 -15). As a result, the current model supports a cross-talk between caspases and mitochondria during the commitment and execution of the apoptotic signal (16 -25); however, the early events responsible for triggering the recruitment and activation of caspases and mitochondrial factors are still sketchy. An important observation reported with CD95-and drug-induced apoptosis is the marked acidification of the intracellular milieu of cells that exhibit morphological and biochemical changes con-sistent with apoptosis (26 -30). Prevention of acidification blocks subsequent apoptosis in some systems (31), and, conversely, generation of a rapid intracellular pH decrease induces caspase activation and apoptosis (32,33). However, the mechanism of intracellular acidification and its temporal relationship to caspase activation and mitochondrial dysfunction during apoptosis are debatable. Contradictory findings have been reported with the general caspase inhibitor benzyloxycarbonylvalinyl-alaninyl-aspartyl-(O-methyl)-fluoromethylketone as to whether intracellular acidification induces caspases or is a downstream consequence of caspase activation (32,34). Similarly, there is a fair amount of controversy with respect to the role of the mitochondrial permeability transition (MPT) 1 pore in cytosolic acidification during chemically induced apoptosis (35,36). In a recent report, we demonstrated that hydrogen peroxide (H 2 O 2 )-mediated apoptosis was accompanied by intracellular acidification (37). These findings, together with the fact that an increase in intracellular H 2 O 2 is observed in tumor cells upon exposure to anticancer agents (38 -40), stimulated our interest in investigating the temporal relationship between drug-induced H 2 O 2 production, intracellular acidification, and caspase activation.
We have recently demonstrated that two novel anticancer agents generated upon photo-oxidation of merocyanine 540 (MC540) triggered cytosolic translocation of cytochrome c in tumor cells (6,41). Despite efficient cytochrome c release and induction of MPT, one of the agents, namely C5, was ineffective in triggering apoptosis due to insufficient caspase activation (41). We concluded that the mere release of cytochrome c in the absence of caspase activation was not sufficient for effective drug-induced apoptosis.
In the present report, the role of intracellular acidification in the apoptotic activity of the two photoproducts was investigated. We show that HL60 cells acidify upon exposure to C2, and that the intracellular acidification and caspase activation is dependent upon H 2 O 2 production. We also provide evidence to support that oxidant-dependent release of cytochrome c from purified mitochondria occurs independent of the MPT, but could be inhibited by scavengers of H 2 O 2 . These findings provide a mechanistic explanation for the differential response of tumor cells to two drugs that activate cytochrome c translocation but differ in their abilities to trigger caspase activation and effective apoptosis, and highlight the critical role of H 2 O 2 in acidifying the intracellular milieu, thus creating a permissive environment for caspase activation and apoptosis.

EXPERIMENTAL PROCEDURES
Cells and Culture Conditions-The human promyelocytic leukemia cell line HL60 was obtained from ATCC (Rockville, MD) and maintained in culture in a 37°C incubator with 5% CO 2 in RPMI 1640 supplemented with 10% fetal bovine serum (Life Technologies, Inc.).
Purification of Photo-oxidation Products C2 and C5-Photoactivation of MC540 (Sigma) and purification of photoproducts C2 and C5 were performed as described recently (6). Purified compounds were resuspended in Me 2 SO at 100 mg/ml and stored at Ϫ20°C.
Measurement of Intracellular pH with 2Ј,7Ј-Bis(2-carboxyethyl)-5,6carboxyfluorescein (BCECF)-Intracellular pH (pH i ) was measured by loading cells with membrane-impermeant dye BCECF (Sigma) as described elsewhere (42). Briefly, HL60 cells (1 ϫ 10 6 ) before or after exposure (2-12 h at 37°C) to C2 (50 g/ml), C5 (150 g/ml), or H 2 O 2 (35-138 M) were washed once with HBSS, resuspended in 0.1 ml of HBSS, and loaded with 10 l of 1 mM BCECF at 37°C for 30 min in the dark. Cells were resuspended in 0.5 ml of HBSS and analyzed using a Coulter Epics Elite ESP (Coulter, Hialeah, FL) flow cytometer with the excitation set at 488 nm. A minimum of 10,000 events was analyzed, and the ratio of BCECF fluorescence at 525 and 610 was used to obtain intracellular pH from a pH calibration curve. In order to generate a pH calibration curve, HL60 cells were loaded with BCECF as above, washed once with HBSS, and then resuspended in high K ϩ buffer (135 mM KH 2 PO 4 , 20 mM NaCl, and 110 mM KH 2 PO 4 , and 20 mM NaCl with a range of pH between 6.0 and 8.0). Immediately before flow cytometry, cells were loaded with 20 M nigericin (1 mM stock in absolute alcohol; Sigma), and fluorescence ratio measurements (525 nm/610 nm) of cells in nigericin-containing buffers of a range of pH were then used to relate histogram channel numbers to pH i . To evaluate the effect of scavenging intracellular H 2 O 2 on pH i , cells were incubated with either catalase (1000 units/ml) or the synthetic non-protein catalase/SOD mimetic EUK-8 (30 M) for 1 h before treatment with C2 or C5. Where indicated, pH i was clamped by incubating cells in medium at the required pH in the presence of 1 g/ml nigericin. 1 g/ml nigericin did not affect cell viability for up to 8 h.
Flow Determination of Caspases 3, 8, and 9 Activities-Caspases 3, 8, and 9 activities were assayed by using AFC-conjugated substrates supplied by Bio-Rad. HL60 cells (1 ϫ 10 6 cells/ml) were incubated with C2 (50 -100g/ml) or C5 (150g/ml) for 8 h, washed twice with 1ϫ phos-phate-buffered saline, resuspended in 50l of chilled cell lysis buffer (provided by the supplier), and incubated on ice for 10 min. 50 l of 2ϫ reaction buffer (10 mM HEPES, 2 mM EDTA, 10 mM KCl, 1.5 mM MgCl 2 , 10 mM dithiothreitol) and 6 l of the fluorogenic caspase-specific substrate (DEVD-AFC for caspase 3, IETD-AFC for caspase 8, and LEHD-AFC for caspase 9) were added to each sample and incubated at 37°C for 30 min. Protease activity was determined by the relative fluorescence intensity at 505 nm following excitation at 400 nm using a spectrofluorimeter (Luminescence Spectrometer LS50B, PerkinElmer Life Sciences, Buckinghamshire, United Kingdom). Where indicated, cell lysates were analyzed for poly(ADP-ribose) polymerase (PARP) cleavage by SDS-PAGE and Western blotting with a monoclonal anti-PARP antibody (clone C2-10 from PharMingen, San Diego, CA). Detection of the 116-kDa and/or the cleaved 85-kDa bands was carried out using horseradish peroxidase-conjugated anti-mouse IgG. In addition, the effect of H 2 O 2 on caspase 3 activity was assessed by exposing HL60 cells to increasing concentrations of H 2 O 2 (35-276 M) for 8 h. In order to evaluate the effect of H 2 O 2 on caspase 3 activation induced by C2 or C5, HL60 cells were incubated (1 h) with catalase (1000 units/ml) or EUK-8 (30 M) prior to the addition of 50 g/ml C2 or 150 g/ml C5.
Propidium Iodide Staining for DNA Fragmentation and Apoptosis-Cell death was assessed by propidium iodide (PI) staining for DNA fragmentation as described elsewhere (5). Briefly, HL60 cells (1 ϫ 10 6 cells/ml) were treated with 50 g/ml C2 or 150 g/ml C5 for 12-18 h, fixed with 70% ethanol, and stained with PI for DNA content analysis. At least 10,000 events were analyzed by flow cytometry with the excitation set at 488 nm and emission at 610 nm. Data are shown as percentage of cells with sub-diploid DNA and are mean Ϯ S.D. of three experiments.
Detection of Cytochrome c in Mitochondrial Supernatants-Mitochondria were isolated from mouse liver (BALB/c) as described elsewhere (44). 0.5 mg of mitochondria were incubated with C2 (100 g) or C5 (150 g) for 1 h at 30°C in the presence and absence of 1000 units/ml catalase or 30 M EUK-8, followed by centrifugation at 4000 ϫ g for 5 min. Fifty g of the supernatants were subjected to 12% SDS-PAGE, and transferred to polyvinylidene difluoride membranes using a Trans-blot SD semi-dry system (Bio-Rad). Membranes were then blocked overnight with 5% dry milk in TBST (50 mM Tris/HCl, pH 7.4, 150 mM NaCl, 0.1% Tween 20). Incubation with anti-cytochrome c antibody (1:5000) was performed at room temperature for 1 h in Tris-buffered saline ϩ 0.05% Tween 20 and FIG. 1. Cytosolic translocation of cytochrome c induced upon incubation of tumor cells with C2 or C5. a, HL60 cells (30 ϫ 10 6 ) were exposed to 50 g/ml C2 or 150 g/ml C5 for 12 h and cytosolic fractions were subjected to SDS-PAGE and Western blotting for cytochrome c as described under "Experimental Procedures." b, M14 cells (1 ϫ 10 3 ) were grown on coverslips, treated with C2 or C5 for 12 h, and cytochrome c localization was determined by confocal microscopy using anti-cytochrome c as described under "Experimental Procedures." c, HL60 (1 ϫ 10 6 ) cells were treated with C2 or C5 for 12 h followed by flow cytometric analysis of caspase 3 processing using a PE-conjugated anti-active caspase 3 that recognizes the processed form of caspase 3 (p17). c, lysates obtained from C2-and C5-treated HL60 cells (18 h) were also analyzed for cleavage of PARP by Western blotting using a monoclonal anti-PARP antibody (clone C2-10).
1% bovine serum albumin. Following three washes with TBST, the membranes were exposed to goat anti-mouse IgG-horseradish peroxidase conjugate (1:5000) supplied as 0.8 mg/ml (Pierce) for 1 h and washed thrice with TBST. Chemiluminescence was detected by the SuperSignal substrate Western blotting kit (Pierce).
Intracellular O 2 Ϫ Measurement-A lucigenin-based chemiluminescence assay (45) was used for measuring intracellular O 2 Ϫ as described previously (46). Chemiluminescence was monitored for 20, 40, and 60 s in a TD-20/20 luminometer (Turner Designs, Sunnyvale, CA). Data are shown as relative light units (RLU) per 2 ϫ 10 6 cells Ϯ S.D. of three to six independent measurements. For detection of mitochondrial O 2 Ϫ , purified mitochondria (0.5 mg) were exposed to 50 g of C2 or 150 g of C5 for 0 -20 min and O 2 Ϫ was measured immediately as described above. Determination of Mitochondrial Matrix Swelling-Large amplitude mitochondrial swelling was determined spectroscopically by the loss of absorbance at 540 nm as described elsewhere (47). Where indicated, 10 M cyclosporin A (CsA; Sigma), 1000 units of catalase, or 30 M EUK-8 were added to the mitochondria prior to the addition of C5 (150 g).

FIG. 2. Drug-induced intracellular acidification triggers caspase activation and apoptosis in HL60 cells.
a, 1 ϫ 10 6 cells were incubated with C2 (50 g/ml) for 4 h or C5 (150 g/ml) for 4 -18 h, loaded with pH-sensitive probe BCECF, and the fluorescence ratio (525/610 nm) was used to measure pH i as described under "Experimental Procedures." In a separate experiment, cells were first incubated with C5 for 12 h followed by clamping of pH i at 7.0 with 1 g/ml nigericin for 6 h. Data shown are mean Ϯ S.D. of six independent experiments. b, HL60 cells were exposed to 50 g/ml C2 for 4 -16 h, followed by fluorimetric detection of caspase 3 activity as described under "Experimental Procedures." Data are mean Ϯ S.D. of four independent experiments and are shown as -fold increase of caspase 3 activity (ϫ increase) compared with the untreated cells (1ϫ). c, lysates obtained from cells treated with C5 (12 or 18 h) or C5 (12 h) ϩ nigericin (6 h) were analyzed for caspase 3 activity as described under "Experimental Procedures." Data are mean Ϯ S.D. of four independent experiments and are shown as -fold increase of caspase 3 activity (ϫ increase) compared with the untreated cells (1ϫ). d, PARP cleavage was analyzed by Western blotting of lysates from cells treated with C5 (18 h), nigericin (8 h), or C5ϩnigericin using anti-PARP antibody (clone C2-10). e, DNA fragmentation induced upon exposure to C2, C5, or C5 ϩ nigericin, for 18 h was assessed by PI staining and is expressed as subdiploid DNA (%).

Intracellular Acidification Is Required for Efficient Caspase 3 Activation in HL60 Cells-
We have recently shown that the release of cytochrome c in the absence of efficient caspase activation was not sufficient for effective drug-induced apoptosis in HL60 cells (41). Although both C2 and C5, recently described photoproducts of MC540, induced cytosolic translocation of cytochrome c as shown by Western blotting of HL60 cell cytosols and immunofluorescence analyses in an adherent cell line M14 ( Fig. 1 (a and b) and Ref. 41), only C2 triggered efficient processing of caspase 3 and PARP cleavage (Fig. 1, c  and d). Intrigued by these results, we hypothesized that the intracellular environment may not be conducive for effective caspase activation in C5-treated cells, unlike C2. Therefore, we wondered if the ability of C2 to effectively trigger caspase activation and apoptosis could be explained by the induction of intracellular acidification. Our results showed that, within 4 h of treatment with C2 (50 g/ml), the pH i of HL60 cells dropped from 7.4 to 7.0 (Fig. 2a). Interestingly, kinetic analysis of caspase 3 activity clearly showed that a significant increase in caspase 3 activity occurred between 4 and 8 h of exposure to C2 (Fig. 2b), suggesting that caspase 3 activation was downstream of intracellular acidification. Similar treatment with C5 (150 g/ml) for 4 -18 h had no effect on pH i or caspase 3 activation (Fig. 2, a and c). We next asked if the activity of C5 could be enhanced by acidification of the intracellular milieu. To do so, we first incubated the cells for 12 h with C5 to allow intracellular events triggered by C5, and then clamped pH i by culturing cells for the next 6 h in pH 7.0 medium in the presence of 1 g/ml nigericin (Fig. 2a). Whereas clamping pH i at 7.0 for 6 h in cells cultured in medium alone or by exposure to C5 for 12 or 18 h had no effect on caspase 3 activation, exposure to C5 followed by clamping pH i at 7.0 resulted in a significant increase in caspase 3 activation (Fig. 2c). This was further confirmed by Western blot analysis of cleavage of caspase 3 substrate PARP (48). PARP was not cleaved by exposure to either C5 (150 g/ml for 18 h) or nigericin (1 g/ml for 8 h), but C5 followed by intracellular acidification with nigericin (pH 7.0) efficiently induced PARP cleavage as shown in Fig. 2d. Furthermore, the percentage of cells with sub-diploid DNA (indicative of apoptotic fraction) significantly increased (Ͼ50% as compared with 10% with C5 alone) upon reducing the pH i to 7.0 for 6 h following C5 exposure (Fig. 2e).
C2-induced Acidification Is Accompanied by H 2 O 2 Production-In our recent report, we demonstrated that H 2 O 2 -triggered apoptosis in M14 melanoma cells was accompanied by acidification of intracellular milieu (37). Similar results were obtained by incubating HL60 cells with 138 M H 2 O 2 . HL60 cells exposed to H 2 O 2 acidified within 4 h of treatment, which was inhibited by prior incubation of cells with H 2 O 2 scavenger catalase (Fig. 3a). In addition, efficient caspase 3 activation, PARP cleavage, and DNA fragmentation were induced by exposure to 138 M H 2 O 2 , which was inhibited by preincubation with catalase (Fig. 3, b-d).
Having shown that C2 induced intracellular acidification in HL60 cells, we next asked if the acidification was dependent upon intracellular production of H 2 O 2 . Indeed, HL60 cells exposed to C2 (50 g/ml) for 4 h showed an increase in dichlorofluorescein fluorescence by flow cytometry (Fig. 4a), indicative of intracellular H 2 O 2 production. Pre-incubation of cells with catalase (1000 units/ml) completely inhibited the production of H 2 O 2 (Fig. 4b) and the intracellular acidification induced by C2 (Fig.  4c). These findings were confirmed by using a cell permeable catalase mimetic EUK-8 (30 M) (49) as shown in Fig. 4c. On the contrary, similar to its effect on pH i , treatment of cells with 150 g/ml C5 for 4 or 8 h had no effect on intracellular H 2 O 2 production (Fig. 5a). However, addition of H 2 O 2 (69 M) for 4 h following 12 h incubation with C5 resulted in a drop in pH i to 7.0 (data not shown), significant activation of caspases 9 and 3, and increase in percentage of DNA fragmentation as compared with C5 (18 h) or H 2 O 2 (4 h) alone (Fig. 5b) shown to involve activation of the nuclear repair enzyme PARP, and inhibition of the ATP-dependent Na ϩ /H ϩ antiporter (50). In order to assess if acidification induced by C2 was dependent upon PARP activation and the subsequent inhibition of the Na ϩ /H ϩ antiporter, cells were preincubated with the PARP inhibitor 3-aminobenzamide, which does not itself alter pH i . Subsequent addition of either C2 (50 g/ml) for 4 h or H 2 O 2 (138 M) for 30 min completely inhibited the drop in pH induced by either C2 or H 2 O 2 (Fig. 6a). Similar results were obtained with another inhibitor phenanthridinone (51), suggesting the involvement of PARP activation in the acidifying activity of C2. Next we questioned if the drop in intracellular pH could be due, in part, to inhibition of the Na ϩ /H ϩ antiporter. In order to accomplish that the effect of a known inhibitor of the Na ϩ /H ϩ exchanger, amiloride, on C2 and H 2 O 2 induced acidification was investigated. Following preincubation with C2 (50 g/ml) or H 2 O 2 (138 M), cells were incubated with 10 M amiloride for 30 min. Results showed no subsequent effect of amiloride on intracellular pH following prior incubation with either C2 or H 2 O 2 (Fig. 6b). Conversely, prior inhibition of the Na ϩ /H ϩ exchanger with amiloride blocked the subsequent effect of C2 or H 2 O 2 on pH i (data not shown). Thus, the absence of a significant ⌬pH i upon incubation with C2 following amiloride treatment, or vice versa, suggests the presence of a common intracellular target. These results support earlier findings that H 2 O 2 -mediated intracellular acidification is due, at least in part, to inhibition of the Na ϩ /H ϩ exchanger (50).
H 2 O 2 Production and Acidification by C2 Is Upstream of Caspase Activation-Having shown that C2 triggered H 2 O 2 production and intracellular acidification, we next assessed the temporal relationship between caspase activation and H 2 O 2 production/acidification. In a previous report, we have shown that apoptosis induced by C2 in HL60 cells was caspase 8-dependent (6); therefore, we asked if incubation of cells with inhibitors of caspases 8, or executioner caspase 3 could block H 2 O 2 production by C2. Flow cytometric analysis of cells treated with 50 g/ml C2 in the presence of cell permeable caspase inhibitors (IETD-CHO for caspase 8 or DEVD-CHO for caspase 3) did not inhibit intracellular H 2 O 2 production (Fig.  7a). In addition, preincubation of cells with inhibitors of caspases 8, 3, and 9 (LEHD-CHO) did not significantly inhibit intracellular acidification induced by C2 (Table I). On the contrary, incubation of cells with catalase (1000 units/ml) or EUK-8 (30 M) resulted in almost complete inhibition of both caspase 8 and 3 activities, and mitochondrial-dependent caspase 9 activity triggered in HL60 cells by 50 g/ml C2 (Fig.  7b). These results indicate that intracellular H 2 O 2 production and acidification induced upon exposure of HL60 cells to C2 occurred upstream of caspase activation.
C2 5. C5 is ineffective as an apoptotic trigger due to lack of H 2 O 2 production. a, HL60 cells (1 ϫ 10 6 /ml) were incubated with C5 (150 g/ml) for 4 -8 h, loaded with DCHF-DA, and analyzed by flow cytometry for H 2 O 2 production. b, HL60 cells were exposed to C5 (150 g/ml) for 18 h or C5 for 12 h followed by H 2 O 2 (69 M) for 6 h or H 2 O 2 (69 M) for 6 h, and caspases 3 and 9 activities were determined as described under "Experimental Procedures." Data are shown as RLU, following excitation at 410 nm and emission at 505 nm. Cell death was determined by PI staining for DNA fragmentation. Ϫ with the maximum concentration attained at 2 h following C2 treatment, which dropped to base-line levels by 4 h (Fig. 8a).

FIG. 6. Acidification induced by C2 is PARP-dependent and involves inhibition of Na
Intracellular sources of O 2 Ϫ include the NADPH oxidase and the mitochondria (52,53). Whereas inhibition of NADPH oxidase by incubation of cells with 10 M diphenyleneiodonium (DPI) blocked phorbol 12-myristate 13-acetate-induced intracellular production of O 2 Ϫ , there was no significant inhibition of intracellular O 2 Ϫ production triggered by C2 (Fig. 8b), which correlated with the inability of DPI to inhibit caspase 3 activation (Fig. 8c), and percentage of cell death assessed by PI staining (data not shown) induced by C2. We then resorted to the mitochondria, as the potential source of intracellular O 2 Ϫ in response to C2. Purified mouse liver mitochondria were incubated with 50 g/ml C2 or 150 g/ml C5 for 0 -20 min and O 2 Ϫ was measured as described before. Results showed that, within 5 min of incubation with C2, the mitochondrial O 2 Ϫ was significantly increased, peaked by 10 min, and returned to base-line levels by 20 min (Fig. 9a). In contrast, treatment with C5 had no effect on mitochondrial O 2 Ϫ concentration (Fig. 9a). H 2 O 2 -mediated Release of Cytochrome c Is Independent of the MPT Pore-We recently showed that cytochrome c release triggered by C5 was dependent upon induction of MPT and opening of the CsA-inhibitable inner membrane pore, whereas the mechanism of C2-induced cytochrome c release was independent of the mitochondrial pore (41). Having shown that mitochondria were the source of O 2 Ϫ generated upon exposure of HL60 cells to C2, we next evaluated the effect of scavenging intracellular H 2 O 2 on the release of cytochrome c, and the induction of permeability transition. Our results showed that preincubation of mitochondria with catalase or EUK-8 completely inhibited the release of cytochrome c triggered by exposure to 50 g of C2 (Fig. 9b). In contrast, although CsA inhibited C5-induced mitochondrial swelling and MPT pore opening, catalase had no effect on the induction of MPT and matrix swelling triggered by C5 (Fig. 9c). Similarly, the CsA-inhibitable release of cytochrome c from C5-treated mitochondria was unaffected by prior incubation with the radical scavengers (data not shown). DISCUSSION Intracellular acidification has been described in cells undergoing apoptosis through the cell surface receptor CD-95 or in response to anti-cancer drug treatment (26,27,29,32). Thus, intracellular acidification appears to represent a common event in the pathway of apoptotic cell death. Indeed, our data show that acidification of the intracellular milieu is critical to induce efficient caspase activation and cell death following drug-induced apoptosis. Moreover, we provide evidence that acidification of the intracellular milieu is due to mitochondrial production of radical oxygen species, in particular H 2 O 2 , upstream of caspase activation.
We used two novel compounds derived from photo-oxidation of MC540, namely C2 and C5, because in a recent report we had shown that, despite their similar ability to induce release of mitochondrial cytochrome c in tumor cells, C5 was unable to trigger efficient apoptotic cell death, which was due to inefficient processing/activation of cytosolic caspase (41). Hence, we hypothesized that the mere release of cytochrome c and induction of permeability transition was insufficient for effective apoptosis in HL60 cells, and that an additional factor with cytochrome c could explain for the efficient apoptotic signaling by C2. We show here that the intracellular milieu of HL60 cells acidified upon exposure to C2. On the contrary, cells incubated with C5 failed to show any significant change in the pH i with minimal caspase activation and apoptosis; however, the apoptotic activity of C5 was significantly enhanced upon acidification of the intracellular milieu. These data strongly support that cytochrome c release without acidification is not efficient for effective apoptotic cell death, and that acidification creates an intracellular environment permissive for caspase activation.
Although intracellular acidification during apoptosis has been attributed to a selective loss of pH regulation, the factor(s) that trigger acidification during induction of apoptosis are not well understood. A recent report (37) and data shown in Fig. 3 demonstrate that apoptotic signaling by H 2 O 2 is mediated by    (56), and another supports caspase 3-dependent production of H 2 O 2 in drug-induced apoptosis (39). Furthermore, a recent report has demonstrated that intracellular acidification occurs down-stream of caspase 8 activation (57). In our model, inhibiting caspase activity did not block H 2 O 2 production triggered by exposure to C2; however, scavenging intracellular H 2 O 2 inhibited C2-induced caspase 8, 9, and 3 activities. In addition, the intracellular acidification induced by C2 was completely inhibited by blocking H 2 O 2 . These data suggest a sequence of events whereby generation of intracellular H 2 O 2 induced by C2 dysregulates pH i and the resulting acidification facilitates the downstream activation of the caspase cascade. Intracellular acidification induced by H 2 O 2 has been linked to the inhibition of the membrane Na ϩ /H ϩ exchanger in a variety of cell types (50,54,58,59), which appears to be mediated by the activation of the nuclear repair enzyme PARP (50,60). H 2 O 2 -mediated PARP activation could rapidly decrease cellular NAD ϩ and, subsequently, ATP levels (60). Depletion of ATP could then inhibit the ATP-dependent Na ϩ /H ϩ exchanger (61, 62) and induce acidification. Indeed, prior inhibition of PARP by 3-aminobenzamide or phenanthridinone completely blocked acidification induced by C2 or H 2 O 2 . In addition, we provide evidence that the pH dysregulation induced by C2 and H 2 O 2 , at least in part, involved inhibition of the Na ϩ /H ϩ exchanger. Interestingly, it has been previously reported that photoproducts of MC540 induce rapid reduction in cellular ATP levels in tumor cells (63). These results suggest that intracellular acidification induced by C2 is dependent upon H 2 O 2 -mediated PARP activation, and may involve inhibition of the Na ϩ /H ϩ exchanger.
Intracellular H 2 O 2 is produced via the enzymatic dismuta- Ϫ was measured by a lucigenin-based chemiluminescence assay as described under "Experimental Procedures." Data shown are mean Ϯ S.D. of three independent experiments done in duplicates and are expressed as RLU/60 s. b, HL60 cells were preincubated with the NADPH oxidase inhibitor DPI (10 M) for 1 h followed by 2 h of treatment with 50 g/ml C2 or phorbol 12-myristate 13-acetate (1 M) and intracellular O 2 Ϫ was measured as described. Data are shown as RLU/60 s and expressed as percentage of control. c, cells (1 ϫ 10 6 /ml) were pretreated with DPI (10 M) for 1 h, followed by incubation for 8 h with 50 g/ml C2, and lysates were assayed for caspase 3 activity.

FIG. 9. Drug-induced mitochondrial generation of O 2
؊ triggers cytochrome c release, without the induction of MPT. a, purified mouse liver mitochondria (0.1 mg) were incubated with 50 g of C2 or 150 g of C5 for 5, 10, or 20 min at 30°C and O 2 Ϫ was determined by the lucigenin chemiluminescence assay. Data shown are mean Ϯ S.D. of three experiments. b, 0.5 mg purified mitochondria were incubated with 100 g of C2 in the absence or presence of H 2 O 2 inhibitors catalase (1000 units) or EUK-8 (30 M) at 30°C for 1 h and centrifuged at 2500 ϫ g. Supernatants were then subjected to SDS-PAGE and Western blot analysis using anti-cytochrome c antibody. c, induction of the MPT pore was assessed by monitoring mitochondrial matrix swelling spectrophotometrically at 540 nm after the addition of 100 g of C2 or 150 g of C5. Where indicated CsA (10 M) or catalase (1000 units) were added before the addition of the drugs.

tion of O 2
Ϫ by Cu/ZnSOD or the mitochondrial MnSOD. Indeed, our results showed that intracellular H 2 O 2 production in HL60 cells was preceded by a surge in intracellular O 2 Ϫ , which returned to base-line levels by 4 h following drug exposure. This is consistent with our results on H 2 O 2 production in HL60 cells, which could be detected as early as 2 h (data not shown) and peaked at 4 h after the addition of C2. The two major intracellular sources of O 2 Ϫ are the membrane-bound NADPH oxidase system, and the mitochondrial electron transport chain. Whereas inhibitor of NADPH oxidase DPI (64) had no effect on O 2 Ϫ production, caspase activation, and apoptosis triggered in HL60 cells by C2, incubation of isolated mitochondria with C2 showed a significant increase in O 2 Ϫ . Contrarily, C5 had no effect on mitochondrial O 2 Ϫ production. The O 2 Ϫ thus generated could then be converted by the mitochondrial MnSOD to H 2 O 2 , which can readily diffuse through mitochondrial and other cellular membranes and become distributed throughout the cell. These data are supported by published evidence of H 2 O 2 release from the mitochondria (53).
One of the critical steps in apoptosis is the cytosolic translocation of cytochrome c (12), which is an amplification factor for the caspase cascade (24). Here we provide evidence for the involvement of mitochondrial-derived oxygen radicals in C2induced cytosolic translocation of cytochrome c. Whereas scavenging H 2 O 2 completely inhibited cytochrome c release induced by C2, there was no effect on mitochondrial swelling and MPT pore-dependent cytochrome c release triggered by C5. Thus, mitochondrial generation of H 2 O 2 might induce the dissociation of cytochrome c from the inner membrane, independent of the MPT pore and matrix swelling. In this respect, a previous report has demonstrated that exogenous H 2 O 2 triggered procaspase activation via cytosolic translocation of cytochrome c before any loss in ⌿ m in Jurkat T cells (65).
Taken together, these data suggest that acidification triggered by mitochondrial-derived H 2 O 2 , coupled with the release of cytochrome c, could be a strong amplification signal for the efficient activation of the caspase cascade during drug-induced apoptosis. However, with respect to caspase 9 activation, our results do not rule out the possibility that, although dependent upon H 2 O 2 production and cytochrome c release, activation could occur independent of acidification. This is corroborated by our published findings that slight caspase 9 activation, in the absence of caspase 3 and 8 activation, is induced with MPT pore-dependent release of cytochrome c by the other photoproduct C5 (41).