Ceramide formation leads to caspase-3 activation during hypoxic PC12 cell death. Inhibitory effects of Bcl-2 on ceramide formation and caspase-3 activation.

PC12 cells undergo apoptosis as well as necrosis following exposure to hypoxia. Following a 6-h hypoxic treatment, a time-dependent increase in intracellular ceramide level was observed with a concurrent decrease in sphingomyelin. It was also shown that the hypoxia-induced ceramide accumulation resulted from activation of neutral magnesium-dependent sphingomyelinase. Comparative kinetic analyses of the neutral sphingomyelinase in the cells under normoxia and hypoxia showed that hypoxia increased Vmax but did not affect Km of the enzyme. In PC12 cells overexpressing Bcl-2 which show strong resistance to hypoxia, sphingomyelin hydrolysis was decreased and activation of neutral sphingomyelinase was reduced. Addition of exogenous C2-ceramide induced cell death and activated caspase-3 as markedly as the hypoxia treatment. On the other hand, in PC12 cells overexpressing Bcl-2, significant decreases in cell death and inhibition of caspase-3 activation were observed after exogenous addition of C2-ceramide. The inhibitors of caspase-3 prevented cell death by either hypoxia or C2-ceramide. These results suggest that ceramide generated by activation of neutral magnesium-dependent sphingomyelinase mediates hypoxic cell death and that Bcl-2 has inhibitory effects on ceramide formation and caspase activation.

PC12 cells undergo apoptosis as well as necrosis following exposure to hypoxia. Following a 6-h hypoxic treatment, a time-dependent increase in intracellular ceramide level was observed with a concurrent decrease in sphingomyelin. It was also shown that the hypoxiainduced ceramide accumulation resulted from activation of neutral magnesium-dependent sphingomyelinase. Comparative kinetic analyses of the neutral sphingomyelinase in the cells under normoxia and hypoxia showed that hypoxia increased V max but did not affect K m of the enzyme. In PC12 cells overexpressing Bcl-2 which show strong resistance to hypoxia, sphingomyelin hydrolysis was decreased and activation of neutral sphingomyelinase was reduced. Addition of exogenous C 2 -ceramide induced cell death and activated caspase-3 as markedly as the hypoxia treatment. On the other hand, in PC12 cells overexpressing Bcl-2, significant decreases in cell death and inhibition of caspase-3 activation were observed after exogenous addition of C 2 -ceramide. The inhibitors of caspase-3 prevented cell death by either hypoxia or C 2 -ceramide. These results suggest that ceramide generated by activation of neutral magnesium-dependent sphingomyelinase mediates hypoxic cell death and that Bcl-2 has inhibitory effects on ceramide formation and caspase activation.
Cell death due to hypoxia is a major concern in various clinical entities such as cerebral ischemia and other diseases. However, the mechanisms of hypoxic cell death have not yet been fully understood. Cell death by hypoxia has long been generally believed to be represented as necrosis (1,2), based on various ultrastructural findings in hypoxic cells. In contrast, recent biochemical observations have suggested the possibility of hypoxia-induced cell death including apoptosis (3,4). These findings are supported by recent reports that hypoxic or ischemic cell death is prevented by the anti-apoptotic proteins, Bcl-2 (5)(6)(7)(8)(9)(10)(11), Bcl-X L (12), and by the inhibitors of the interleukin-1␤ converting enzyme and CPP32 caspases (13,14). It has also been described that Bcl-2 functions upstream of interleukin-1␤ converting enzyme-like caspases during hypoxic death of PC12 cells (13). The family of cysteine proteases has been thought to be implicated in apoptotic cell death on the basis of observations that their inhibitors or negative mutants inhibit apoptosis induced by various stimuli (15)(16)(17)(18)(19)(20)(21). Ten proteins homologous to interleukin-1␤ converting enzyme (caspase-1) have been identified in mammals and are classified into three subfamilies, caspase-1-, caspase-2 (NEDD2/ICH-1)-, and caspase-3 (CPP32/Yama/apopain)-like proteases, based on their structures (22). Among these caspases, caspase-3 has been deemed an attractive candidate as a putative mediator of apoptosis.
On the other hand, recently the sphingomyelin (SM) 1 cycle, with regulated conversion of SM to ceramide by sphingomyelinase (SMase), has been extensively studied as a key pathway involved in apoptosis as well as differentiation (23)(24)(25). Ceramide is postulated to be a second messenger of apoptosis, because it appears to induce typical morphological changes of apoptosis following the inhibition of cell growth in many types of cells, including neuronal cells (26 -28). SM hydrolysis is shown to be induced not only by a variety of cytokines, including tumor necrosis factor-␣, interferon-␥ (29,30), interleukin-1␤ (31), nerve growth factor (32), and cross-linking of Fas (33), but also by physical stresses such as radiation, heat shock, and chemotherapeutic drugs (26,34,35).
In the current study, to elucidate the involvement of SM metabolism in hypoxia-induced cell death, we have examined the intracellular levels of ceramide and SM in vector-and Bcl-2-transfected PC12 cells after exposure to hypoxia. N-Acetylsphingosine (C 2 -ceramide), a membrane-permeable analog, which is known to induce apoptosis in many types of cells including neuronal cells (28,36,37), was also used to determine direct action of ceramide. We show here that hypoxia induces ceramide formation in PC12 cells through activation of neutral magnesium-dependent SMase, which was inhibited by highly overexpressed Bcl-2. These results suggest that ceramide formation contributes to hypoxic death of PC12 cell and that Bcl-2 has inhibitory effects on ceramide formation and caspase activation.

EXPERIMENTAL PROCEDURES
Materials-The tetrapeptide substrate for caspase-3, Ac-DEVD-MCA, was purchased from Peptide Institute (Osaka, Japan). The relatively nonselective inhibitor of caspase-1-like proteases, z-VAD.FMK, the selective inhibitor of caspase-1, z-YVAD.AFC, and the selective inhibitor of caspase-3, z-DEVD.FMK, were obtained from Enzyme Sys-* This work was supported in part by research grants from the Ministry of Education, Culture, Sports, and Science of Japan and from the Uehara Memorial Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom correspondence should be addressed: Dept. of Neurosurgery, Gifu University School of Medicine, Tsukasamachi-40, Gifu 500-8705, Japan. Cell Culture-A PC12 cell line (38) was maintained in RPMI 1640 medium supplemented with 10% (v/v) FBS, 5% (v/v) horse serum, 100 units/ml penicillin, and 100 g/ml streptomycin in a humidified atmosphere containing 5% CO 2 at 37°C. Hypoxia was induced with a multigas incubator (SANYO MCO-175M, Osaka, Japan). Briefly, after the cells were plated at a low density (5 ϫ 10 4 /cm 2 ) in RPMI 1640 medium supplemented with 10% FBS and 5% horse serum for 24 h, cells were incubated in the medium containing 2% FBS for 24 h, and then placed into the multigas incubator (set to 1.0% O 2 , with 5% CO 2 , and the balance N 2 ) at 37°C for indicated periods according to the methods previously described (39,40). For treatment with exogenous C 2 -ceramide, the cells plated at a low density (5 ϫ 10 4 /cm 2 ) were maintained in RPMI 1640 medium supplemented with 2% FBS for 24 h. After the cells were rinsed twice with serum-free RPMI medium, 20 M C 2 -ceramide was delivered into the RPMI medium containing 2% FBS. To verify the effects of caspase inhibitors, the cells were preincubated for 1 h with medium containing 100 M z-VAD.FMK, z-DEVD.FMK, or z-YVAD.AFC. For experiments using FB1, the cells were preincubated for indicated periods with medium containing 100 M FB1, and then exposed to hypoxia. Optimal concentration of each agent was determined from dose-response curves versus the extent of cell death (data not shown).
Cell Transfection-Human bcl-2 cDNA (41) was transfected using the pUC-CAGGS vector (42) bearing a ␤-actin promoter and cytomegalovirus enhancer as described previously (10,13). Production of Bcl-2 proteins was confirmed by Western blot analysis using anti-human Bcl-2 monoclonal antibody, and the amount of Bcl-2 expressed using the pUC-CAGGS system was much higher than those using the pBC140 system (10).
Fluorescence Microscopy-Cells were stained with Hoechst 33258 (10 M) and PI (10 M) for 10 min and analyzed under a non-confocal fluorescence microscope (Olympus BX60) with excitation at 360 nm, as previously reported (13). Because Hoechst 33258 stains all nuclei and PI stains nuclei of cells with a disrupted plasma membrane, nuclei of viable, necrotic, and apoptotic cells were observed as blue intact nuclei, red round nuclei, and fragmented (or condensed) nuclei, respectively, under a fluorescence microscopy.
LDH Assay-Extent of cell death was assessed using a kit to measure released LDH activity from dead cells, because loss of cell membrane integrity was observed in both necrotic and apoptotic cells (10). LDH activity in the culture medium was determined using a commercially available kit (Kyokuto, Tokyo) exactly as described by the manufacturer.
Measurement of Cellular Ceramide Level by Diacylglycerol Kinase Assay-Cellular ceramide level was measured according to the previously described method (43) with a slight modification. Lipids were extracted by a Folch partition (44), dried, and incubated in 0.1 M KOH in chloroform:methanol (1:2, v/v) at 37°C for 1 h. Ceramide was converted to ceramide 1-[ 32 P]phosphate by E. coli diacylglycerol kinase in the presence of [␥-32 P]ATP. Labeled lipids (ceramide 1-phosphate) were separated by high performance thin-layer chromatography in chloroform:acetone:methanol:acetic acid:water (50:20:15:10:5, v/v). Following autoradiography, spots corresponding to ceramide 1-phosphate were scraped and the radioactivity was counted in a scintillation counter. Quantitation of ceramide was based on a standard curve of known amounts of ceramide. The changes in ceramide content were normalized based on total protein.

Measurement of [ 14 C]Ceramide and [ 14 C]SM in [ 14 C]
Serine-labeled Cells-To label sphingolipids, PC12 cells (5 ϫ 10 5 cells/120 ml) were cultured for 72 h in the medium containing 25 Ci of [ 14 C]serine. Lipids were extracted as described previously (44) and the lower phase was collected and evaporated under a flow of nitrogen gas. The extracted lipids were incubated in 0.1 M KOH in chloroform:methanol (1:2, v/v) at 37°C for 1 h. Samples were then separated on high performance thinlayer chromatography plates using the solvent system of chloroform: methanol:water (70:30:5, v/v) for SM or chloroform:methanol (95:5, v/v) for ceramide. Ceramide was observed as doublet; the upper band consisted mostly of a mixture of C 22:0 , C 24:0 , and C 24:1 fatty acids and the lower band consisted primarily of C 16:0 , C 16:1 , and C 18:0 fatty acids (45). Following autoradiography, spots were scraped and the radioactivity was determined by liquid scintillation counter (Beckman LS-6500). The changes in ceramide content were normalized based on total protein.
SMase Assay-Membrane and cytosolic fractions were prepared from the cells after exposure to hypoxia. Cells were washed in cold phosphate-buffered saline and homogenized in lysis buffer (25 mM Tris/HCl, pH 7.5, 2 mM EDTA, 2 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, and 20 g/ml E-64) using 20 strokes in a homogenizer (Iuchiseieido, Tokyo, Japan) with a Teflon pestle. The post-nuclear homogenate was centrifuged at 100,000 ϫ g for 1 h, and the pellet was resuspended in lysis buffer.
The activities of both neutral and acid SMases were determined using a mixed micelle assay system as described (24,46,47). For determining neutral SMase activity, membrane and cytosolic fractions (20 g) were mixed with [methyl-14 C]SM (40,000 cpm in 1 nmol of bovine brain SM in 0.25% Triton X-100 solubilized by sonication). The activities of magnesium-dependent and -independent neutral SMase were measured separately using 0.1 M Tris/HCl buffer (pH 7.4) with or without 6 mM MgCl 2 , respectively. After incubation for 30 min at 37°C, the reaction was stopped by the addition of 1.5 ml of chloroform:methanol (2:1, v/v) followed by 0.2 ml of H 2 O. After phase separation, a portion of the upper phase was transferred to scintillation vials and the radioactivity was determined by liquid scintillation counting. Negative controls containing no enzyme was run concomitantly. Acid SMase activity in membrane and cytosol was measured as above except that 0.1 M sodium acetate buffer (pH 5.5) containing 5 mM EDTA replaced the Tris/HCl buffer.
Activity of Caspase-3(-like) Proteases-Cells were harvested after exposure to hypoxia or ceramide for the indicated periods of time and washed three times with phosphate-buffered saline, and then suspended in buffer containing 50 mM Tris/HCl (pH 7.4), 1 mM EDTA, and 10 mM EGTA. After addition of 10 M digitonin, cells were incubated at 37°C for 10 min. Complete cell lysis was verified by the amount of released LDH activity. Lysates were centrifuged at 900 ϫ g for 3 min, and the resulting supernatant (40 g of protein) were incubated with 50 M of the enzyme substrate Ac-DEVD-MCA at 37°C for 1 h. Levels of released 7-amino-4-methylcoumarin were measured using spectrofluorometers (Hitachi F-3000 and F-2000) with excitation at 380 nm and emission at 460 nm. Excitation and emission slit width were adjusted to 10 and 20 mm, respectively. One unit was defined as the amount of enzyme required to release 0.22 nmol of 7-amino-4-methylcoumarin per min at 37°C.

RESULTS
Induction of Cell Death by Hypoxia-Cytotoxicity was assayed by the morphological features of the cells using double staining with Hoechst 33258 and PI under fluorescence microscopy as described by Shimizu et al. (13). Viable, necrotic, early apoptotic, terminal apoptotic cells under hypoxia could be easily distinguished and quantified by this fluorescence microscope analysis with Hoechst 33258 (blue) which stains all nuclei and PI (red) which stains only nuclei in cells with disrupted membrane integrity. Nuclei of viable, necrotic, and apoptotic cells were observed as blue intact nuclei, red intact nuclei, and fragmented or condensed blue or red nuclei, respectively (Fig.  1A). Cells with fragmented or condensed nuclei were classified further based on PI staining as "early" apoptotic with membrane integrity and "terminal" apoptotic without membrane integrity. These nuclear morphological changes under fluorescence microscopy are well corresponded to necrosis and apoptosis defined by electron microscopy in hypoxic PC12 cells and also in other cell lines.
Vector-transfected PC12 cells (designated as PC12-V) were used as control cells. PC12-V cells underwent a time-dependent decrease in cell viability including apoptosis and necrosis after exposure to 1% hypoxia (Fig. 1, A and B). On the other hand, PC12 cells with highly overexpressed Bcl-2 using the pUC-CAGGS system (designated as PC12-Bcl-2) showed strong resistance to hypoxia throughout the time course (Fig. 1, A and  C). To confirm cell viability during hypoxia, the LDH assay was performed because loss of cell membrane integrity was observed in both necrotic and apoptotic PC12 cells after exposure to hypoxia (13). The LDH assay in each group was well correlated with the morphological changes under fluorescent microscopy (data not shown).
Ceramide Production with Concomitant Decrease in SM during Hypoxic Cell Death-Changes in the levels of ceramide and SM during hypoxia were measured by the enzymatic analysis with E. coli diacylglycerol kinase and also by the metabolic labeling of cells with [ 14 C]serine. Up to 3 h after exposure to hypoxia, no significant increase in ceramide level was observed in PC12-V nor PC12-Bcl-2 cells by the diacylglycerol kinase assay. However, a marked increase in the ceramide content was observed at 6 h in PC12-V cells (Fig. 2A). The maximal level was obtained at 24 h (3-fold increase over the control level) and sustained to 48 h. In contrast, significant increase was not observed in PC12-Bcl-2 cells up to 24 h after exposure to hypoxia, but a slight increase was observed at 48 h. Similar profile of ceramide formation was observed in cells labeled with [ 14 C]serine. An increase in [ 14 C]ceramide was observed at 12 and 24 h after exposure to hypoxia in PC12-V cells, but not PC12-Bcl-2 cells (Fig. 2B). There is a small difference in ceramide generation between Fig. 2, A and B, although temporal profiles of ceramide generation were similar. This difference could be due to the different assays used.
In addition, changes in SM level were also analyzed by the metabolic [ 14 C]serine labeling method (Fig. 2C). In PC12-V cells, a decrease in [ 14 C]SM was observed as early as 6 h and its level was declined (approximately 40% of control) at 24 h and sustained thereafter. In contrast, only a small decrease was observed in PC12-Bcl-2 cells even at 48 h.
To confirm the possibility that ceramide formation could mediate hypoxic cell death, we used an inhibitor of sphinganine N-acyltransferase, FB1, which is known to inhibit de novo synthesis of ceramide (12,48,49). Preincubation of PC12-V cells for 1 h with 100 M FB1 before hypoxia was unable to prevent cell death (data not shown). In contrast, prolonged (for 96 h) incubation of cells with 100 M FB1 decreased the intracellular level of both ceramide and SM, and decreased hypoxic cell death by about 40% compared with the control culture (data not shown). These results suggest that ceramide is produced from SM via SMase during hypoxia, but not from de novo synthesis, and that FB1 decreased intracellular ceramide and SM by prolonged incubation, which prevented hypoxic cell death. Thus, these findings also support the hypothesis that ceramide mediates cell death caused by hypoxia.
Activation of Magnesium-dependent Neutral SMase in

PC12-V Cells but Not in PC12-Bcl-2 Cells Exposed to Hypoxia-
Recently neutral and acid SMases were shown to be implicated in the production of ceramide in response to apoptosis inducers (for review, see Ref. 25). Therefore, we have measured activities of three types of SMases (neutral magnesium-dependent, neutral magnesium-independent, and acid) in both membrane and cytosol fractions. It was observed that hypoxia treatment of PC12-V cells for 24 h produced an increase in membrane-bound neutral magnesium-dependent SMase activity (pH 7.5) (Fig.  3A). Interestingly, such an increase was not observed in PC12-Bcl-2 cells. No significant changes were observed in neutral magnesium-independent and acid (pH 5.5) SMase activities by hypoxia treatment, although the activity of acid SMase was higher than that of neutral SMase in both PC12-V and PC12-Bcl-2 cell membranes (the ratio of neutral to acid SMase is approximately 1:3). Activities of neutral and acid SMases in cytosol were very low and no significant changes were observed during hypoxia (data not shown).
A kinetic analysis was performed to know whether hypoxia affects V max or K m of the magnesium-dependent membranebound neutral SMase. Double-reciprocal plots showed that hypoxic treatment did not significantly affect K m , but increased V max (1.5-fold increase over the control level) of the neutral SMase in PC12 cells cultured for 24 h under hypoxic condition (Fig. 3B).

Activation of Caspase-3(-like) Proteases by Hypoxia and Exogenous C 2 -Ceramide-
The above results suggest that ceramide produced during hypoxia is closely involved in hypoxic death signaling. Recently, caspases, especially caspase-3, are regarded as important regulators of apoptosis. However, the relationship between ceramide and caspases remains unknown. If the caspase-3 acts downstream of ceramide, this caspase should be activated by exogenous C 2 -ceramide. To test this possibility, the proteolytic activity of the caspase was measured using fluorogenic tetrapeptide substrates. First, the activities of caspase-3(-like) proteases increased and reached a peak at 12 h after exposure to hypoxia in PC12-V cells (Fig. 4A). C 2 -ceramide also gave rise to the sharp increase in the activities of caspase-3(-like) proteases, which peaked at 6 h and rapidly decreased thereafter (Fig. 4B). However, dihydro-C 2ceramide, often used as an inactive analog of C 2 -ceramide, failed to activate this caspase.
To know the effect of Bcl-2 on the increase in this caspase activity by hypoxia and C 2 -ceramide, the activity of this caspase was also measured in PC12-Bcl-2. As shown in Fig. 4, overexpression of Bcl-2 markedly inhibited activation of caspase-3(-like) proteases during both hypoxia and ceramide treatment.
Prevention of Cell Death Induced by Hypoxia or Ceramide by Bcl-2 and Caspase-3 Inhibitors-To further assess the involvement of caspases in cell death induced by hypoxia and C 2ceramide, the cell viability was assessed by the LDH assay after hypoxia or the ceramide treatment in the presence of a caspase-3 inhibitor, z-DEVD.FMK, and a relatively nonselective caspase inhibitor, z-VAD.FMK in PC12-V cells. The cell death induced by hypoxia at 48 h was found to be less than 20% in the cells treated with 100 M z-VAD.FMK or z-DEVD.FMK, whereas cell death was approximately 40% in the untreated cells (Fig. 5A). In PC12-Bcl-2 cells, the percentage of cell death was about 15 and 25% at 48 and 72 h, respectively. To know the relationship between activation of caspase-3(-like) proteases and ceramide formation, the cytotoxicity was also measured in the cells treated with exogenous ceramide. The cell death induced by C 2 -ceramide at 24 h was less than 50% in the cells treated with 100 M z-DEVD.FMK or 100 M z-VAD.FMK, although more than 90% of the non-treated cells underwent cell death 24 h after addition of 20 M C 2 -ceramide (Fig. 5B). Dihydro-C 2 -ceramide, an inactive analog of ceramide, did not induce cell death (data not shown). In PC12-Bcl-2 cells, the percentage of dead cells after 24 h was approximately 25%. These data imply that caspases-3(-like) proteases and Bcl-2 act downstream of ceramide because both Bcl-2 and caspase-3 inhibitors effectively inhibited the cell death induced by exogenous ceramide.
Effects of Caspase Inhibitors on Ceramide Production during Hypoxia-Recently, it was reported that ceramide generation in response to tumor necrosis factor-␣ or Fas activation was inhibited by an inhibitor of caspase-1, YVAD.CMK, or CrmA (50,51). The results obtained in the present study suggest that activation of neutral SMase is diminished by Bcl-2. Therefore, certain caspases may also act upstream of neutral SMase. To verify this possibility, the ceramide levels during hypoxia were measured in cells treated with various caspase inhibitors (z-VAD.FMK, z-YVAD.FMK, and z-DEVD.FMK). Although these inhibitors effectively decreased cell death induced by hypoxia as shown in Fig. 5A, the level of ceramide was not changed (Table I). Thus, formation of ceramide via SMase during hypoxia was thought to be independent of caspase-1 and caspase-3.

DISCUSSION
Ceramide, a product of SM hydrolysis by SMase, is postulated as a second messenger of apoptosis in many types of cells including neuronal cells. A number of extracellular agents and insults such as tumor necrosis factor, Fas ligands, and chemotherapeutic agents cause the activation of SMase which acts on membrane SM and generates ceramide. Ischemia can be caused by a decrease in tissue perfusion that results in an inadequate supply of oxygen, glucose, and other metabolites. Prolonged ischemia results in the death of neurons or other cells. Recent biochemical observations have suggested that the hypoxia-or ischemia-induced cell death includes apoptosis (10,52). Interestingly, it is reported that the level of SM changes in rat cerebral cortex during focal ischemia (53), postulating that SM hydrolysis occurs in vivo. However, the injury caused by ischemia is multifactorial, including severe hypoxia, substrate deprivation, and failure to remove toxic metabolic products. It is not clear whether SM hydrolysis is a cause or a result of ischemic damage. FIG. 3. Activity of neutral magnesium-dependent sphingomyelinase in PC12-V cells and PC12-Bcl-2 cells exposed to hypoxia. A, PC12-V and PC12-Bcl-2 cells were exposed to hypoxia for 24 h, washed in cold phosphate-buffered saline, and pelleted in microcentrifuge tubes. The cellular pellet was homogenized in lysis buffer. The post-nuclear homogenate was centrifuged at 100,000 ϫ g for 1 h, and the resulting pellet was dissolved in lysis buffer. The activities of neutral magnesium-dependent, magnesium-independent, and acid SMases were determined using a mixed micelle assay system with [methyl- 14   To our knowledge, this study is the first demonstration that exposure of PC12 cells to hypoxia induced a time-dependent increase in the intracellular ceramide with a concurrent decrease in SM. It was also shown that ceramide accumulation was due to activation of neutral magnesium-dependent SMase. No significant changes were observed in neutral magnesiumindependent or acid SMase activities, although activity of acid SMase is higher than that of neutral ones. This may imply that hypoxia causes activation of neutral magnesium-dependent SMase, leading to ceramide formation. Moreover, comparative kinetic analyses of neutral SMase in the cells under normoxia and hypoxia showed that hypoxia increased V max but did not affect K m of the enzyme. This may imply that neutral SMase is up-regulated during hypoxia by increasing activating factor(s) or the enzyme protein itself, or by decreasing inhibitory factor(s).
Although many laboratories have shown that ceramide (synthetic and natural) is able to induce apoptosis (for review, see Ref. 25), it has not been clear whether ceramide accumulates nonspecifically in dying cells or acts as a messenger for cell death. Several observations presented in this study provide possible evidence for an important role of ceramide as a mediator for hypoxic cell death. First, we have modulated the intracellular ceramide level by use of FB1, an inhibitor of de novo synthesis of ceramide. Preincubation with FB1 for 1 h did not affect cell death by hypoxia, suggesting that ceramide is derived from SM hydrolysis, but not de novo synthesis. This finding is in accordance with other reports (47,54). Second, to know whether ceramide formation is directly implicated in hypoxic cell death, the cells were incubated with 100 M FB1 for 96 h. It has been reported that FB1 treatment for 96 h decreased SM and ceramide by about 50% in rat sympathetic neurons (49). Under the same condition, hypoxic death of PC12 cells was suppressed by about 40% compared with the control culture. Third, exogenously added C 2 -ceramide caused apoptotic cell death, whereas dihydro-C 2 -ceramide was unable to cause cell death. Thus, it is likely that ceramide acts as an important mediator in hypoxic cell death.
It is of interest to note that little changes in the levels of ceramide, SM and SMase activity were seen in PC12-Bcl-2 during hypoxia. Zhang et al. (55) reported that ceramide formation in response to chemotherapeutic drugs is not changed by overexpression of Bcl-2 in Molt-4 cells, indicating that Bcl-2 does not affect ceramide formation. Differences in the extent of Bcl-2 expression, cell death inducer, and cell type may explain this discrepancy in ceramide production, since it is reported that Bcl-2 acts in a dose-dependent manner (13). The alternate explanation may be that certain caspases which are inhibitable by Bcl-2 (25, 51) act upstream of SMase. To test this possibility, we measured the ceramide levels in PC12 cells treated with various caspase inhibitors. Although these inhibitors effectively suppressed cell death induced by hypoxia, they had no effect on ceramide formation, suggesting that neither caspase-1(-like) nor caspase-3(-like) proteases are involved in ceramide formation. Accordingly, Bcl-2 is likely to act upstream of SMase, but we have no definite evidence for the precise site(s) of action.
It has already been described that caspase-3(-like) proteases are involved in hypoxic cell death in PC12 cells (10). To examine the functional site of caspase-3 in relation to ceramide formation, we have measured the activities of caspase-3(-like) proteases using fluorogenic tetrapeptide substrates during exogenous C 2 -ceramide treatment or hypoxia. As expected, caspase-3(-like) proteases were activated in PC12-V but not PC12-Bcl-2 cells by exogenous C 2 -ceramide. C 2 -ceramide caused a marked increase in their activities, peaked at 6 h and then rapidly declined. These results indicate that caspase-3 (-like) proteases act downstream of ceramide and its activation is prevented by Bcl-2 (55,56).
It is well known that minute order ischemia causes cerebral infarction in experimental in vivo ischemia models of rats. In the present study, the time course of PC12 cell death induction under hypoxia was rather long, 48 -72 h. Similar results were reported that 48 h hypoxia caused 70% of cell death in primary cultured cortical neurons of rats (57). In other types of cells, hypoxia-induced cell death becomes evident after 24 -48 h. Therefore, these results, in turn, support the notion that neuronal cell death in the brain due to ischemia would result from a variety of pathological events, including severe hypoxia, substrate deprivation, and failure to remove toxic metabolic products.
It has recently been reported that cells deficient in p53 tumor suppresser gene are resistant to hypoxia-induced cell death (58). The results from acid SMase knockout mice and p53 knockout mice revealed that acid SMase-mediated and p53mediated events are distinct and independent in radiationinduced apoptosis (34). However, it is well known that acid SMase functions differently from neutral SMase (for review, see Ref. 25). In addition, the hypoxic treatment activated neutral, but not acid SMase in PC12 cells, as shown in the present study. Relationship of the ceramide signaling pathway with p53 is currently unknown. A study to explore hypoxia-induced ceramide formation in p53-knockout cells will provide a clue to elucidate their relationship. Further investigation of the signaling pathway(s), especially identifying the upstream and downstream components of SMase-ceramide, will lead to better understanding of the molecular mechanism of hypoxic cell death.
In summary, the results obtained in the present study led us to propose a hypothetical scheme for ceramide signaling of hypoxic cell death in PC12 cells (Fig. 6). Neutral SMase is activated during hypoxia, which in turn causes SM hydrolysis and ceramide production. Increased ceramide activates caspase-3(-like) proteases, thereby leading to cell death. Bcl-2 is considered to function not only downstream of ceramide, FIG. 6. A hypothetical scheme of ceramide signaling in hypoxic death of PC12 cells. Neutral SMase is activated during hypoxia, which causes SM hydrolysis and production of ceramide. Increased ceramide activates caspase-3(-like) proteases, thereby leading to cell death. Bcl-2 is considered to function not only downstream of ceramide, leading to inhibition of caspase-3(-like) proteases, but also upstream of SMase to prevent ceramide production. leading to inhibition of caspase-3(-like) proteases, but also upstream of SMase to prevent ceramide production. Better understanding of signal transduction pathways leading to cell death induced by hypoxia in this system may provide a basis for development of the new therapeutic approach to control hypoxic neuronal cell death.