Mitochondrial Phospholipid Hydroperoxide Glutathione Peroxidase Suppresses Apoptosis Mediated by a Mitochondrial Death Pathway*

Phospholipid hydroperoxide glutathione peroxidase (PHGPx) is a key enzyme in the protection of biomembranes exposed to oxidative stress. We investigated the role of mitochondrial PHGPx in apoptosis using RBL2H3 cells that overexpressed mitochondrial PHGPx (M15 cells), cells that overexpressed non-mitochondrial PHGPx (L9 cells), and control cells (S1 cells). The morphological changes and fragmentation of DNA associated with apoptosis occurred within 15 h in S1 and L9 cells upon exposure of cells to 2-deoxyglucose (2DG). The release of cytochrome c from mitochondria was observed in S1 cells after 4 h and was followed by the activation of caspase-3 within 6 h. Overexpression of mitochondrial PHGPx prevented the release of cytochrome c, the activation of caspase-3, and apoptosis, but non-mitochondrial PHGPx lacked the ability to prevent the induction of apoptosis by 2DG. An ability to protect cells from 2DG-induced apoptosis was abolished when the PHGPx activity of M15 cells was inhibited by diethylmalate, indicating that the resistance of M15 cells to apoptosis was indeed due to the overexpression of PHGPx in the mitochondria. The expression of members of the Bcl-2 family of proteins, such as Bcl-2, Bcl-xL, Bax, and Bad, was unchanged by the overexpression of PHGPx in cells. The levels of hydroperoxides, including hydrogen and lipid peroxide, in mitochondria isolated from S1 and L9 cells were significantly increased after the exposure to 2DG for 2 h, while the level of hydroperoxide in mitochondria isolated from M15 cells was lower than that in S1 and L9 cells. M15 cells were also resistant to apoptosis induced by etoposide, staurosporine, UV irradiation, cycloheximide, and actinomycin D, but not to apoptosis induced by Fas-specific antibodies, which induces apoptosis via a pathway distinct from the pathway initiated by 2DG. Our results suggest that hydroperoxide, produced in mitochondria, is a major factor in apoptosis and that mitochondrial PHGPx might play a critical role as an anti-apoptotic agent in mitochondrial death pathways.

gen peroxide, and organic peroxide, are toxic by-products of various metabolic reactions and are produced in response to various stimuli. Moreover, it has recently been revealed that ROS modulate the physiological state of cells and influence cell death (1). A relationship between ROS and apoptosis has been suggested by many experimental findings. Apoptosis is induced by pro-oxidant agents, such as hydrogen peroxide, diamide, etoposide, and semiquinones (2,3). Other apoptotic stimuli, such as treatment with tumor necrosis factor ␣ (TNF␣) and ceramide, also elevate intracellular levels of ROS (4,5). Antioxidants, such as N-acetylcysteine, suppress apoptosis by acting as scavengers of ROS, and their actions provide additional evidence that ROS act as signaling molecules to initiate apoptosis (6).
Mitochondria are a major physiological source of ROS, which are generated during mitochondrial respiration (7). ROS that are generated in excess in mitochondria act as mediators of the apoptotic signaling pathway (8). TNF␣-and ceramide-induced production of ROS in mitochondria have been proposed as early events in the induction of apoptosis (9,10). Activators of apoptosis, such as caspase-2, caspase-9, cytochrome c (as an activator of caspases), and apoptosis-inducing factor, are all found in mitochondria (11,12). The liberation of cytochrome c and that of apoptosis-inducing factor from mitochondria are irreversible execution in the process that leads to apoptosis. Mitochondria also contain proteins that regulate apoptosis, such as members of the Bcl-2 family, namely Bcl-2, Bcl-xL, Bax, and Bad, which can prevent or accelerate apoptosis (13,14). The evidence in the cited reports indicates that mitochondria are major participants in apoptosis and that ROS produced in mitochondria contribute to cell death by acting as apoptotic signaling molecules.
The production of ROS in mitochondria is regulated by a number of antioxidant enzymes within mitochondria, which include phospholipid hydroperoxide glutathione peroxidase (PHGPx), classical glutathione peroxidase (cGPx), and Mnsuperoxide dismutase (Mn-SOD). The proposal of a role for mitochondrial antioxidant enzymes in preventing apoptosis is based on the observation that Bcl-2, a general inhibitor of apoptosis in mammalian cells, has an apparent antioxidant function (15). Bcl-2 is localized predominantly in mitochondrial membranes. Cells that overexpress Bcl-2 are resistant to apoptosis that is normally caused by agents such as TNF␣, staurosporine, and etoposide (16 -18). Recently, Manna et al. (19) demonstrated that apoptosis of MCF-7 cells upon treatment with TNF␣ was completely suppressed by the overexpression of Mn-SOD. Such reports suggest that antioxidant enzymes localized in mitochondria might be linked to apoptosis and might contribute to modulation of apoptotic signals. PHGPx is a unique intracellular antioxidant enzyme that directly reduces peroxidized lipids that have been produced in cell membranes (20,21). Although PHGPx might participate in defense systems as an anti-oxidant enzyme that protects lipids in mitochondria from damage (22), little is known about the biological significance of mitochondrial PHGPx in apoptosis that is mediated by mitochondrial pathways.
We conducted a series of experiments to clarify the role of PHGPx in mammalian cells. Two types of PHGPx (of 20 and 23 kDa, respectively) were translated in vitro from a cDNA that included two sites for initiation of transcription (23,24). We demonstrated that the long form of PHGPx (23 kDa) was the mitochondrial PHGPx and included a signal peptide for transport to mitochondria, while the short form of PHGPx (20 kDa) was the non-mitochondrial PHGPx. Recently, we obtained evidence to indicate that PHGPx plays a role as an antioxidant enzyme in mammalian cells. Stably transformed rat basophile leukemia 2H3 (RBL2H3) cells harboring the gene for nonmitochondrial PHGPx were resistant to cell death caused by a radical initiator (25). Overexpression of non-mitochondrial PHGPx suppresses the activity of 5-lipoxygenase by reducing lipid hydroperoxides in nuclei (26). Overexpression of mitochondrial PHGPx also protects cell from necrotic death caused by chemical hypoxia. However, non-mitochondrial PHGPx does not protect cells from necrotic death. Cells undergo necrosis when exposed to severe oxidative stress due to exogenously added ROS (27).
In this study, we demonstrated that mitochondrial PHGPx suppressed apoptotic cell death via the mitochondrial death pathway that was activated by 2DG, deprivation of glucose, etoposide, staurosporine, UV irradiation, actinomycin D, or cycloheximide, but it was ineffective in cases of Fas-mediated or A23187-induced apoptosis. Non-mitochondrial PHGPx suppressed neither Fas-mediated apoptosis nor mitochondrial apoptosis. It seems possible that hydroperoxide produced in mitochondria might play a crucial role in the liberation of cytochrome c from mitochondria.
Cell Culture-We used the previously established strains of PHGPxoverexpressing RBL2H3 cells, namely L9 cells that overexpressed nonmitochondrial PHGPx and M15 cells that overexpressed mitochondrial PHGPx. The control cells (S1) have been stably transfected with the expression vector without inserts (27). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) that contained 5% fetal calf serum (FCS). Glucose depletion was achieved by incubating cells in the presence of 100 mM 2DG in DMEM that contained 5% FCS for 16 h or in glucose-free DMEM (Life Technologies, Inc.) supplemented with 1 mM sodium pyruvate, 10 mM HEPES (pH 7.4), 100 units/ml penicillin, 100 g/ml streptomycin, and 5% dialyzed FCS (Life Technologies, Inc.) for 16 h.
Assessment of Cell Viability-S1, L9, and M15 cells were plated at 0.5 ϫ 10 5 cells/well in flat-bottomed 96-well culture plates and cultured for 24 h. Apoptotic cell death was induced by treating cells with indicated doses of antibodies against rat Fas for 40 h, of etoposide for 16 h, of the calcium ionophore A23187 with 1 mM CaCl 2 for 16 h, of actinomycin D for 16 h, of cycloheximide for 16 h, and of staurosporine for 16 h. Some cells were exposed to UV-B light at the indicated dose and then incubated for 16 h. An assay of the release of lactate dehydrogenase (LDH) was used for the determination of cell viability, as described previously (27).
In one series of experiments, cells were incubated for 1 h with 2 mM DEM for depletion of glutathione (GSH) prior to exposure to 2DG, and they were then incubated for 16 h with 100 mM 2DG in addition to glucose at the indicated dose to examine the requirement for glucose of cell viability.
Cytochemical Staining-Apoptotic cell death was evaluated by fluorescence microscopy after staining with Hoechst 33258. Cells were collected, washed twice with phosphate-buffered saline (PBS), (137 mM NaCl, 8.1 mM Na 2 HPO 4 , 2.68 mM KCl, 1.47 mM KH 2 PO 4 , pH 7.4), and then fixed for 20 min in 3.6% paraformaldehyde in PBS. After washing with PBS, cells were stained with 1.6 M Hoechst 33258 for 10 min. After three washes with PBS, the samples were treated with Aqua-Poly/ Mount (Polysciences, Inc., Warrington, PA) before mounting. Samples were observed with a fluorescence microscope (Nikon, Tokyo, Japan) that was equipped with a 100ϫ objective, with excitation at 360 nm.
DNA Fragmentation-At the indicated times, 2 ϫ 10 6 cells were pelleted and resuspended in 250 l of Tris-EDTA buffer (10 mM Tris-HCl, pH 7.6, 1 mM EDTA). An equal amount of ice-cold lysis buffer (0.5% Triton X-100 and 2 mM EDTA in 5 mM Tris-HCl buffer, pH 8.0) was then added, and cells were lysed for 30 min. Then the samples were centrifuged at 12,000 rpm for 20 min. DNA in supernatants was precipitated with ethanol, treated with RNase (Roche Molecular Biochemicals, Almere, The Netherlands) at 37°C for 30 min, and then extracted with phenol and chloroform. Recovered fragments of DNA were separated by electrophoresis in a 1.5% agarose gel and visualized by staining with ethidium bromide.
Measurement of the Activity of Caspase-3-Cells were treated with 100 mM 2DG, 50 M etoposide, 0.1 M staurosporine, 1 M calcium ionophore (A23187), plus 1 mM CaCl 2 or 300 ng/ml antibodies against rat Fas. At the times indicated, cells were washed twice with PBS and incubated in 400 l of PBS with 10 g/ml digitonin at 37°C for 5 min. Lysates were collected and centrifuged at 12,000 rpm for 20 min. Supernatants were diluted with 400 l of reaction buffer (1 mM dithiothreitol, 2 mM EDTA, 0.1% CHAPS, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml pepstatin, and 10 g/ml leupeptin in 10 mM Tris-HCl buffer, pH 7.4) and incubated with 50 mol of Ac-DEVD-MCA, as the substrate for caspase-3, at 37°C for 30 min. Fluorescence of 7-amino-4-methylcoumarin that had been cleaved from Ac-DEVD-MCA by caspase-3 was measured with a CytoFluor system (model 4000; PerSeptive Biosystems, Framingham, MA) with excitation at 380 nm and emission at 460 nm.
Release of Cytochrome c-After incubations, supernatants of lysates of digitonin-treated cells were collected, as described above, for analysis of the release of cytochrome c from mitochondria into the cytosol. Proteins in supernatants were precipitated by addition of trichloroacetic acid and precipitated proteins were fractionated by SDS-PAGE (15.0% polyacrylamide) under non-reducing conditions. Bands of proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore Corp., Bedford, MA) as described previously (26). The PVDF membrane was blocked by incubation with Block Ace (Dainippon Pharmaceutical Co., Osaka, Japan) for 1 h and incubated with antibodies against cytochrome c for 2 h. The PVDF membrane was then incubated for 1 h with horseradish peroxidase-conjugated antibodies raised in goat anti-mouse IgG (Zymed Laboratories Inc.). The binding of antibodies to the PVDF membrane was detected with an enhanced chemiluminescence Western blotting analysis system (Amersham Pharmacia Biotech, Buckinghamshire, United Kingdom).
Immunoblot Analysis-Cell homogenates (50 g of protein) were fractionated by SDS-PAGE on a 15.0% acrylamide gel, and bands of protein were transferred to a PVDF membrane. The expression of Bcl-2, Bcl-xL, Bad, and Bax was examined by immunoblotting with the corresponding specific polyclonal or monoclonal antibodies, as described above.
Analysis of Cellular Levels of ATP-Cellular levels of ATP were determined by the luciferin-luciferase method (28) with a kit from Sigma. After incubation in medium that contained 100 mM 2DG for indicated times, cells were washed twice with PBS and then scraped into 0.04 M Tris borate buffer (pH 9.2). The suspensions of cells were then placed in boiling water for 5 min to inactive cellular ATPases. The samples were then chilled and centrifuged at 15,000 rpm for 3 min at 4°C. Then 15-l aliquots of samples were added to 100-l aliquots of the luciferin-luciferase reagent. After addition of the luciferin-luciferase, luminescence was immediately monitored over a 10-s period with a CytoFluor plate reader. Measurement of the Generation of Intracellular Hydroperoxide and Superoxide-Levels of intracellular hydroperoxide and superoxide were monitored by measuring changes in fluorescence that resulted from oxidation of an intracellular probe. To assess levels of intracellular hydroperoxide, we used 2 g/ml DHR. DHR is oxidized by hydrogen peroxide and lipid hydroperoxide to yield fluorescent rhodamine 123 (Rh123). To assess levels of superoxide, we used 2 g/ml DHE. DHE is oxidized by superoxide to yield fluorescent ethidium. Cells were plated at 10 5 cells/well in a 12-well plate, washed three times with PBS, and then incubated in DMEM that contained DHR (2 g/ml) or DHE (2 g/ml). 2DG was added to give a final concentration of 100 mM in a final total volume of 1 ml. Fluorescence was monitored with a CytoFluor plate reader at the indicated times. To assess levels of mitochondrial hydroperoxide, mitochondria of RBL2H3 cells treated with or without 2DG for 4 h were prepared by differential centrifugation of cell homogenates as described previously (27). Aliquots of mitochondria were incubated with 2 g/ml DHR for 20 min, and fluorescence was monitored with a CytoFluor plate reader.
Quantitation of Proteins-Concentrations of proteins were determined with the BCA protein assay reagent (Pierce), with bovine serum albumin as the standard.

Induction of Apoptosis by Glucose Deprivation or 2-Deoxyglu-
cose-In a previous study, we prepared three lines of RBL2H3 cells as follows: cells that overexpressed mitochondrial PHGPx (M15 cells), cells that overexpressed non-mitochondrial PHGPx (L9 cells), and cells that expressed the vector only (S1 cells). We studied the resistance of these three lines to apoptotic cell death in response to glucose deprivation. Cells were cultured in glucose-free medium, and viability was monitored in terms of the release of LDH (Fig. 1A). Numbers of dead S1 cells increased gradually after incubation for 8 h and reached approximately 80% of the total within 12 h. By contrast to S1 cells, M15 cells were resistant to the cytotoxic effects of glucose deprivation and more than 90% remained viable at 12 h. L9 cells, which overexpressed non-mitochondrial PHGPx, were quite sensitive to the glucose deprivation, resembling S1 cells in this respect. Cells were exposed to 2DG, which competitively inhibits the cellular uptake and utilization of glucose. The cytotoxic effect of 2DG was observed, and the effect was both dose-and time-dependent (Fig. 1, B and C). S1 and L9 cells were particularly susceptible to treatment with 2DG, and cell deaths were ob-served after 15 h. Overexpression of mitochondrial PHGPx markedly protected cells from the toxic effects of 2DG. By contrast, non-mitochondrial PHGPx did not prevent cell death. The difference in the effects of 2DG on M15 and L9 cells was not due to a difference between rates of uptake of 2DG since no significant differences in rates of uptake of tritiated 2DG were observed among the three lines of cells during treatment with 2DG (data not shown).
The nature of the cell death caused by 2DG was examined by monitoring the morphology of nuclei by fluorescence microscopy and the pattern of DNA fragmentation by electrophoresis on an agarose gel (Fig. 2, A and B). The cleavage of DNA into a "ladder" of fragments was clearly detected in S1 and L9 cells after incubation with 2DG for 14 h (Fig. 2A). Condensation of nuclei was also observed after staining with Hoechst 33258 of nuclei in S1 and L9 cells, when treatment with 2DG had been continued for 16 h (Fig. 2B). These results showed clearly that 2DG had induced apoptotic cellular death in S1 and L9 cells. Neither fragmentation of DNA nor the condensation of nuclei was observed in M15 cells.
We examined the effects of exogenously added glucose on apoptosis in order to determine whether the apoptosis caused by 2DG was initiated by the depletion of utilizable glucose in cells (Fig. 3). Exogenously added glucose prevented the killing of cells by 2DG, and the viability of S1 and L9 cells were restored to 80 -90% after the addition of glucose at 100 mM to the growth medium.
Release of Cytochrome c and Activation of Caspase-3 by 2DG-The release of cytochrome c and the activation of caspase-3 were monitored in order to examine the signaling pathway of 2DG-induced apoptosis (Fig. 4). Cytochrome c, released from mitochondria, was detectable in S1 and L9 cells 4 h after the start of exposure to 2DG, while no detectable cytochrome c was found in the cytosol of 2DG-treated M15 cells (Fig. 4A). The proteolytic activity of caspases was measured in terms of the ability to cleave Ac-DEVD-MCA, which is a substrate of caspase-3, a common effector of apoptosis (Fig. 4B). No evidence for the activation of caspase-1 was found during 2DGinduced apoptosis in RBL2H3 cells (data not shown). No activation of caspase-3 in S1 and L9 cells was detected for 3 h after the start of treatment with 2DG, but activation was evident at 6 h, after the release of cytochrome c. The activation of caspase-3 reached a maximum at 15 h, and apoptotic cell death occurred at this time. No activation of caspases was induced by 2DG in M15 cells. Activation of caspase-3 appears to be associated with the process of apoptosis since acetyl-DEVD-Chinese hamster ovary, an irreversible inhibitor of caspase-3, effectively protected cells from apoptosis induced by 2DG (data not shown). These results indicate that 2DG induces the liberation FIG. 1. Glucose deprivation and a competitive analog of glucose, 2DG, induce cell death in rat basophile leukemia cells (RBL2H3 cells). Control cells (S1 cells; closed circles), cells that overexpressed non-mitochondrial PHGPx (L9 cells; closed triangles), and cells that overexpressed mitochondrial PHGPx (M15 cells; closed squares) were plated at 0.5 ϫ 10 5 cells/well in DMEM plus 5% FCS. Cells were incubated in glucose-free DMEM for the indicated times (A). Cells were also exposed to the indicated dose of 2DG for 16 h (B) and treated with 2DG at 100 mM for the indicated times (C). Cell viability was estimated from the release of LDH as summarized in the text. Data are means Ϯ S.D. of results from three replicates in each case. of cytochrome c from mitochondria and subsequent apoptosis and, moreover, that mitochondrial PHGPx prevents apoptosis by blocking the release of cytochrome c from mitochondria.
Effects of Diethylmalate on the Apoptosis of M15 Cells-Since the overexpression of mitochondrial PHGPx had a considerable protective effect against the toxicity of 2DG, we examined whether the resistance to 2DG of M15 cells might have resulted from the overexpression of mitochondrial PHGPx. DEM, which is a glutathione-depleting (GSH-depleting) agent, inhibits the activity of PHGPx by lowering the level of intracellular glutathione. The level of intracellular GSH fell to approximately 3% of the control level after treatment of cells with DEM, as described previously (29). We examined the effects of DEM on cell viability, the release of cytochrome c into the cytosol, and the activity of caspase-3 after the exposure to 2DG of DEMpretreated cells (Fig. 5). When M15 cells had been pretreated with DEM, they lost their resistance to 2DG (Fig. 5A) and exhibited typical symptoms of apoptotic death, such as DNA fragmentation and the condensation of nuclei (data not shown). Inhibition of the release of cytochrome c from mitochondria and the activation of caspase-3 in 2DG-treated M15 cells were prevented when the PHGPx activity was inhibited (Fig. 5, B  and C). These results confirm that the resistance of M15 cells to 2DG-induced apoptotic death was indeed due to the overexpression of PHGPx in mitochondria.
A Decline of the Intracellular Level of ATP after Treatment with 2DG-Levels of intracellular ATP were measured in S1, L9, and M15 cells when the utilization of glucose was impaired by 2DG (Fig. 6). The level of ATP fell markedly to 25% of the control level within 1 h. No significant differences in the extent of the reduction in the level of ATP were observed among the three lines of cells at the early stage of apoptosis. Levels of ATP FIG. 2. Nuclear fragmentation and changes in the morphology of cells during exposure to 2DG. S1 cells, L9 cells, and M15 cells were treated with 100 mM 2DG for the indicated times. A, to detect nuclear fragmentation, low molecular weight DNA was recovered from S1 cells, L9 cells, and M15 cells at the indicated times, fractionated by gel electrophoresis, and stained with ethidium bromide. B, morphological changes in cells treated with 2DG for 16 h (b, d, and f) and untreated cells (a, c, and e) were examined by staining with Hoechst 33258 for 10 min and fluorescence microscopy. Bar, 10 m. in S1 and L9 cells fell still further at the late phase (16 h), at which time most S1 and L9 cells were no longer viable.
Expression of Members of the Family of Bcl-2-related Proteins-Members of the family of Bcl-2-related proteins, which are apoptosis-regulating proteins, include both antagonists (Bcl-2 and Bcl-xL) and agonists (Bax and Bad) of apoptosis. We evaluated the expression of these proteins by immunoblotting analysis in order to characterize the participation of each protein in apoptosis (Fig. 7). We found no significant differences in the respective levels of expression of Bcl-2, Bcl-xL, Bax, and Bad between cells that were resistant and cells that were sensitive to 2DG-induced apoptosis.
Production of Superoxide and Hydroperoxide in Response to 2DG-Our finding that protection from apoptosis was associated with the overexpression of mitochondrial PHGPx suggested that mitochondrial injury caused by 2DG might increase levels of ROS. Therefore, we determined levels of superoxide and hydroperoxide using specific fluorescent indicators. Marked increases in the levels of intracellular superoxide and hydroperoxide including hydrogen and lipid peroxides were observed after the exposure of cells to 2DG (Fig. 8). Superoxide was detectable in S1, L9, and M15 cells within 1 h after the start of exposure to 2DG, and its level increased at a steady rate for up to 4 h. There were no differences in rates of production of superoxide among the three lines of cells (Fig. 8A). Hydroperoxide also accumulated in cells after exposure to 2DG. It accumulated more slowly than superoxide and was first detectable in S1 cells at 2 h. Overexpression of PHGPx suppressed the production of hydroperoxide, and the production of hydroperoxide was more effectively suppressed in M15 cells than in L9 cells (Fig. 8B). The levels of hydroperoxide were determined in mitochondria isolated from S1, L9, and M15 cells (Fig. 8C). Hydroperoxides were detected in mitochondria of S1, L9, and M15 cells without treatment with 2DG, and the levels of hydroperoxides were increased significantly in mitochondria isolated from S1 and L9 cells treated with 2DG for 4 h. However, productions of hydroperoxides were completely prevented in PHGPx-overexpressing mitochondria of 2DG-treated M15 cells.
Effects of Various Apoptotic Agonists on the Apoptosis of M15 Cells-Apoptosis is induced by a variety of apoptotic agonists, and each triggers its own specific apoptotic pathway (30). We examined the effects of several kinds of apoptotic agonists on the viability of PHGPx-overexpressing cells (Fig. 9). Staurosporine, etoposide, actinomycin D, cycloheximide, and UV irradiation killed S1 and L9 cells, while M15 cells were resistant to these apoptotic agents (Fig. 9, A-E). By contrast, overexpression of mitochondrial PHGPx failed to protect cells from apoptosis caused by Fas-specific antibodies. (Fig. 9G). A23187 showed a biphasic effect on the apoptosis of M15 cells (Fig. 9F). Apoptosis caused by low dose of A23187 (0.25 M) was suppressed in M15 cells, while apoptosis was induced in M15 cells by the relatively high dose of A23187 (more than 0.5 M). These results demonstrate that protection from apoptosis by mitochondrial PHGPx depends on the apoptotic agonists to which cells are exposed.
Activation of caspase-3 was recognized in S1 and L9 cells after their exposure to staurosporine and to etoposide, but mitochondrial PHGPx significantly inhibited the activation of caspase in M15 cells exposed similarly to etoposide or staurosporine (Fig. 10A). Considerable cytochrome c was released from mitochondria to the cytosol in S1 and L9 cells, but not in M15 cells, after their exposure to staurosporine or to etoposide (Fig. 10B). The activation of caspase-3 and the release of cytochrome c were inhibited in M15 cells after treatment with etoposide or with staurosporine, but these phenomena were not suppressed when Fas-specific antibodies and A23187 were used to induce apoptosis (Fig. 10). These results demonstrated that mitochondrial PHGPx was able to suppress apoptosis that was induced by the mitochondrial death pathways that are activated by etoposide, staurosporine, actinomycin D, cycloheximide, and UV irradiation. By contrast, mitochondrial PHGPx was unable to suppress apoptosis, the activation of caspase-3, and the release to the cytosol of cytochrome c in M15 cells exposed to Fas-specific antibodies. DISCUSSION Apoptosis of RBL2H3 cells is initiated after the elimination of glucose from the medium or by the chemical anoxia induced by 2DG (31)(32)(33). Overexpression of mitochondrial PHGPx abolished the 2DG-induced apoptosis of RBL2H3 cells (M15 cells), but overexpression of non-mitochondrial PHGPx failed to prevent apoptosis. Activity of cGPx, which is another major antioxidant enzyme in mitochondria, was not changed by the overexpression of mitochondrial PHGPx (27). This indicates that cGPx in mitochondria would not participate the prevention of apoptosis in M15 cells. Resistance of M15 cells to apoptosis was abolished upon inhibition of PHGPx activity by DEM. Thus, overexpression of PHGPx in mitochondria appeared to contribute to protection from 2DG-induced apoptosis.
Apoptosis induced by 2DG was prevented by the addition of glucose to the medium (Fig. 3), and 2DG induced the depletion of intracellular ATP via an ATP-consuming reaction by which 2DG is actively metabolized to 2-deoxyglucose 6-phosphate (34). Marton et al. (35) reported similarly that inhibitors of energy conservation, such as 2DG, rotenone and oligomycin, decrease levels of intracellular ATP in FL5.12 cells and induce apoptosis. By contrast, apoptosis induced by, for example, Fasspecific antibodies, a Ca 2ϩ ionophore, etoposide, dexamethasone and staurosporine, requires ATP (36 -38) and is abolished when levels of ATP are very low. Lelli et al. (39) demonstrated that intracellular levels of ATP determine the mode of cell death, either apoptosis or necrosis, and that the threshold level of ATP must be 25% of the basal level in endothelial cells for apoptosis to proceed after exposure to hydrogen peroxide. Thus, intracellular levels of ATP are critically important in the control of apoptotic cell death. In RBL2H3 cells, levels of ATP fell to 25% of basal levels within 1 h of exposure to 2DG, and this level was sufficient for induction of apoptosis in S1 and L9 cells (Fig. 6). However, M15 cells were resistant to 2DG-induced apoptosis even though the level of ATP fell to the same level as in S1 cells. This result suggests that the change in ATP level is an independent event in the development of resistance of M15 to 2DG-induced apoptosis.
Mitochondrial membrane potentials (⌬) in 2DG-induced cells were examined by flow cytometric analysis with Rh123. No significant decreases in ⌬ were observed in all three lines of cells until 12 h after the start of treatment with 2DG (data not shown). Thus, release of cytochrome c from mitochondria was unrelated to the mitochondrial membrane potential in 2DG-induced apoptosis.
Chemical anoxia induces oxidative stress via enhanced mitochondrial generation of ROS, as well as via depletion of ATP (40 -43). 2DG induced the rapid generation of superoxide in S1 cells within 1 h, and the production of hydroperoxide was detected about 2 h later (Fig. 8). Superoxide was released at the same rate in three cell lines, while production of hydroperoxide was suppressed in M15 cells as compared with that in S1 and L9 cells. These results suggest that ATP depletion might induce the production of ROS, probably via acceleration of respiratory chain reactions and the accumulation of hydroperoxide in mitochondria. Considerable evidence has accumulated to suggest that ROS might act as mediators of apoptosis (44,45). For example, hydrogen peroxide can induce apoptosis in a variety of cell types (46,47). Moreover, NO-generating agents cause apoptosis in macrophages (48). ROS are formed during cell death triggered by ceramide in B cells (49) and might be a mediator in such apoptosis. The intracellular sources of ROS include mitochondrial oxidation, the microsomal cytochrome P450 system, and plasma membrane NADPH oxidase. The ROS from mitochondria might be responsible for a close association between the activities of mitochondria and cell death (50). This possibility is supported by the observation that TNF␣-, Fe 2ϩ -, amyloid ␤-peptide-, and alkaline-mediated apoptosis are blocked by Mn-SOD, which scavenges superoxide that has leaked from the mitochondrial respiratory chain (19,51,52). Higuchi et al. (53) demonstrated that apoptosis induced by TNF␣ does not occur in ML-1a cells that are rendered respiration-deficient by treatment with ethidium bromide. In addition, ceramide-induced apoptosis of U937 cells is suppressed by rotenone, a specific inhibitor of the mitochondrial respiratory chain that inhibits the production of ROS from mitochondria. By contrast, apoptosis is enhanced by antimycin, which stimulates the production of ROS via inhibition of complex III (54). We found that 2DG-induced apoptosis was blocked when the accumulation of hydroperoxides in mitochondria was prevented by overexpression of mitochondrial PHGPx. This phenomenon suggests that production of hydroperoxide in mitochondria might be an important early trigger of apoptosis.
The release of cytochrome c and the activation of caspase-3 were inhibited in M15 cells (Fig. 4). Thus, the interruption of the apoptotic signaling pathway in M15 cells seems to occur prior to the release of cytochrome c. Release of cytochrome c from mitochondria is generally accepted as a crucial step in the activation of apoptosis in various model systems (55). Bcl-2 is a well characterized anti-apoptotic factor that prevents release of cytochrome c (56), but the absence of any changes in the expression of Bcl-2-related proteins in M15 cells indicates that these proteins are not involved in the resistance of M15 to 2DG-induced apoptosis (Fig. 7). The release of cytochrome c from mitochondria is poorly understood. Yang and Cortopassi (57) demonstrated that canonical inducers of a mitochondrial permeability transition, such as t-butylhydroperoxide (100 M), induce the swelling-dependent release of cytochrome c from isolated mitochondria. ROS might induce the dissociation of cytochrome c from inner membranes of mitochondria that contain significant amounts of highly unsaturated fatty acid (58). Furthermore, oxidation of mitochondrial membrane lipids might trigger the opening of mitochondrial permeability transition pores, such as the adenine nucleotide translocator, with release of cytochrome c from mitochondria (59). We demonstrated previously that mitochondrial PHGPx effectively suppresses the production of lipid hydroperoxide in response to inhibitors of the mitochondrial respiratory chain, such as KCN (27). The present study demonstrates that the generation of ROS is an early and critical event in apoptosis. Excessive production of hydroperoxide and the resultant damage to mitochondria might induce the liberation of cytochrome c in the 2DG-induced apoptosis of RBL2H3 cells.
Studies of Apaf-1 and caspase-9 knockout mouse revealed that two independent apoptotic pathways operate upstream of the critical activation of caspase-3 (60,61). In the first pathway, involved in Fas-induced apoptotic death, caspase-8 is activated prior to activation of caspase-3. Such Fas-induced activation of caspase-8 is independent of any proapoptotic mitochondrial activity. In the second pathway (mitochondrial death pathway), various forms of cellular stress, due to apoptotic agents such as etoposide, staurosporine and UV irradiation, induce the release from mitochondria of cytochrome c, which activates caspase-9 and finally caspase-3. This pathway is independent of the activation of caspase-8. The activation of caspase-8 was not observed prior to the caspase-3 activation in S1, L9, and M15 cells treated with 2DG, etoposide, and staurosporine (data not shown). Overexpression of mitochondrial PHGPx prevented apoptosis via the mitochondrial death pathway upon treatment with etoposide, UV irradiation, staurosporine, or glucose deprivation. By contrast, the Fas-mediated apoptotic pathway was intact in control cells and in cells that did or did not overexpress mitochondrial PHGPx. Thus, mitochondrial PHGPx is irrelevant to Fas-induced apoptosis, and it failed to hinder release of cytochrome c in such apoptosis. Mitochondrial PHGPx also failed to prevent the activation of caspase-8 by Fas-specific antibodies (data not shown). With respect to the involvement of cytochrome c in Fas-induced apoptosis, Li et al. (62) demonstrated that the cleavage of Bid by caspase-8 induced the release of cytochrome c in Fas-induced apoptosis of Jurkat cells. Marzo et al. (63) showed that recombinant caspases directly disrupted barrier functions of mitochondrial membranes, with consequent release of cytochrome c. Thus, mitochondrial PHGPx failed to prevent the release of cytochrome c in response to Bid or caspase. However, it appears that mitochondrial PHGPx selectively prevented apoptosis via the mitochondrial death pathway, in which production of hydroperoxide is apparently one of the key apoptotic signals. Recent reports indicate that etoposide and staurosporine cause apoptosis via the generation of hydroperoxide (64 -66), and the present study indicates that M15 cells might be useful for further analysis of apoptotic signaling pathways.
M15 cells were resistant to the induction, by an as yet unidentified mechanism, of apoptosis by actinomycin D or cycloheximide. Our results suggest that such apoptosis might be induced via the mitochondrial death pathway. In A23187-induced apoptosis, mitochondrial PHGPx protected apoptosis caused by low dose of A23187 (0.25 M); however, PHGPx could not prevent apoptosis by high dose of A23187 (0.5 M). These biphasic effects of A23187 on apoptosis would suggest that A23187 could activate simultaneously several independent apoptotic pathways including Ca 2ϩ mobilization.