Activation of Stress-activated Protein Kinase/c-Jun NH2-terminal Kinase and p38 Kinase in Calphostin C-induced Apoptosis Requires Caspase-3-like Proteases but Is Dispensable for Cell Death*

Apoptosis was induced in human glioma cell lines by exposure to 100 nm calphostin C, a specific inhibitor of protein kinase C. Calphostin C-induced apoptosis was associated with synchronous down-regulation of Bcl-2 and Bcl-xL as well as activation of caspase-3 but not caspase-1. The exposure to calphostin C led to activation of stress-activated protein kinase/c-Jun NH2-terminal kinase (SAPK/JNK) and p38 kinase and concurrent inhibition of extracellular signal-regulated kinase (ERK). Upstream of ERK, Shc was shown to be activated, but its downstream Raf1 and ERK were inhibited. The pretreatment with acetyl-Tyr-Val-Ala-Asp-aldehyde, a relatively selective inhibitor of caspase-3, or benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (z-VAD.fmk), a broad spectrum caspase inhibitor, similarly inhibited calphostin C-induced activation of SAPK/JNK and p38 kinase as well as apoptotic nuclear damages (chromatin condensation and DNA fragmentation) and cell shrinkage, suggesting that caspase-3 functions upstream of SAPK/JNK and p38 kinase, but did not block calphostin C-induced surface blebbing and cell death. On the other hand, the inhibition of SAPK/JNK by transfection of dominant negative SAPK/JNK and that of p38 kinase by SB203580 induced similar effects on the calphostin C-induced apoptotic phenotypes and cell death as did z-VAD.fmk and acetyl-Tyr-Val-Ala-Asp-aldehyde, but the calphostin C-induced PARP cleavage was not changed, suggesting that SAPK/JNK and p38 kinase are involved in the DNA fragmentation pathway downstream of caspase-3. The present findings suggest, therefore, that the activation of SAPK/JNK and p38 kinase is dispensable for calphostin C-mediated and z-VAD.fmk-resistant cell death.

It has been expected that signal transduction pathways involving specific protein kinases are involved in mediating apoptosis. Extracellular signal-regulated kinase (ERK) 1 is most strongly stimulated by activation of protein-tyrosine kinase receptors (1) and also activated by both Ras-dependent (2)(3)(4) and Ras-independent signalings (5) in response to activation of G protein-coupled receptors, and common intermediates in intracellular signaling cascades are involved in diverse cellular functions including growth and differentiation (6). Activation of PKC by 12-O-tetradecanoylphorbol-13-acetate results in the activation of Raf1 (7) and ERK (8 -11) within minutes, suggesting the involvement of PKC in the signaling pathway leading to ERK activation. In addition, Ras functions as an essential transducer of various physiological signals leading to cell growth and proliferation in a PKC-dependent or PKC-independent manner (12), and activated Ras also renders cells susceptible to apoptosis after depression of PKC activity (13,14).
SAPK/JNK and p38 kinase have been proposed to mediate apoptosis, but a number of reports have challenged the notion that the activation of SAPK/JNK and/or p38 kinase is sufficient to induce apoptosis (15)(16)(17)(18)(19)(20)(21), and the integration and balance of SAPK/JNK and p38 pathways probably contribute to commitment to apoptosis (22,23). In addition, the inhibition of ERK activity has been reported to correlate with the activation of SAPK/JNK and p38 kinase as well as with the induction of apoptosis in nerve growth factor-deprived PC12 pheochromocytoma cells (22), Fas-induced Jurkat cells (15,24), and UVirradiated mouse fibroblasts (25), and the balance between the activity of the stress kinases to that of ERK has been proposed to determine cell fate (26), because the Raf1-ERK signaling pathway plays a pivotal role in suppressing apoptotic death (27,28).
Apoptosis is induced in human gliomas by exposure to calphostin C, a specific inhibitor of PKC, and calphostin C-induced apoptosis was preceded by down-regulation of bcl-2 mRNA and protein (29). Overproduction of Bcl-2 can render cells more resistant to induction of apoptosis by a wide variety of stimuli (30), and Bcl-2 appears to work both upstream and downstream of the caspase cascade to prevent effects of caspases (31). To characterize the molecular mechanism that regulates calphostin C-induced apoptosis of gliomas, the contributions of mitogen-activated protein kinase family members to cell death caused by calphostin C-induced down-regulation of Bcl-2 are explored using various inhibitors of caspases, SAPK/JNK, and p38 kinase in this work.

MATERIALS AND METHODS
Glioma Cell Lines-Two glioma cell lines (wild type p53-positive U-87MG and mutant p53-positive T98G obtained from the American Type Culture Collection, Rockville, MD) were maintained in DMEM supplemented with 10% fetal calf serum and antibiotics (100 units/ml penicillin and 100 g/ml streptomycin) in a humidified atmosphere of 5% CO 2 and 95% air at 37°C. Test exposures to 100 nM calphostin C * 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.
were conducted in the continuous presence of light to promote photoactivation of this compound.
Microscopic Analysis-Cells were fixed with 4% paraformaldehyde for 30 min after washing with ice-cold phosphate-buffered saline and then stained with 2.5 g/ml Hoechst 33258. The number of apoptotic cells was assessed by nuclei staining, and nuclei that were fragmented or condensed were scored as apoptotic. The viability of cells was confirmed by the trypan blue or propidium iodide dye exclusion method.
Agarose Gel Electrophoresis-DNA was prepared from cells as described previously (29). The resulting DNA preparation was analyzed by 1.5% agarose gel electrophoresis in TBE buffer (89 mM Tris borate buffer, pH 8.0, 89 mM boric acid, 22 mM EDTA) containing 0.5 g/ml ethidium bromide at 100 V/cm for 30 min. The DNA fragmentation pattern was examined on photographs taken under UV illumination.
Loading of Cells with Caspase-1 and -3 Inhibitor Peptides-The osmotic lysis of pinosome method as described by Moore et al. (32) was used to enhance uptake of the inhibitor peptides into the cells. Briefly, cells were incubated for 8 -10 min at 37°C with 500 l of a prewarmed hypertonic solution of 0.5 M sucrose, 10% (v/v) polyethylene glycol 1000, 10 mM HEPES (pH 7.0) in DMEM containing 100 M Ac-YVAD.cho or 100 M Ac-DEVD.cho. Hypotonic medium (60% DMEM, 40% distilled water) was added to a final volume of 15 ml, and incubation at 37°C continued for 10 min. The hypotonic medium was changed into DMEM, 10% fetal calf serum. Calphostin C was added to the culture medium, and the mixture was incubated for the indicated times, followed by agarose electrophoresis.
Shc Phosphorylation-Cells were lysed on ice in lysis buffer (50 mM HEPES at pH 7.5, 0.5% Triton X-100, 100 mM NaF, 10 mM sodium pyrophosphate, 4 mM EDTA, 2 mM Na 3 VO 4 , 2 mM Na 2 MoO 4 , 2 g/ml aprotinin, 2 g/ml leupeptin, 2 g/ml phenylmethylsulfonyl fluoride). After centrifugation at 15,000 ϫ g for 10 min at 4°C, supernatants were absorbed with Affi-Gel protein A (Bio-Rad)-normal rabbit serum complex for 90 min at 4°C and then incubated with Affi-Gel protein Apolyclonal anti-human Shc antibody (Upstate Biotechnology, Inc., Lake Placid, NY) complex, which had been treated in advance with 20 mM dimethyl pimelimidate in 0.2 M sodium borate for cross-linking. The immunoprecipitates were boiled for 5 min in 2ϫ sample buffer, subjected to 10% SDS-polyacrylamide gel electrophoresis, and transferred to polyvinylidene difluoride membrane, which was then blocked with 1% gelatin in phosphate-buffered saline containing 0.05% Tween 20 overnight and probed with mouse monoclonal anti-phosphotyrosine antibody (Upstate Biotechnology) for 1 h. The secondary antibody was horseradish peroxidase-conjugated anti-mouse IgG (Dako, Glostrup, Denmark), and proteins were detected using the ECL system (Amersham Pharmacia Biotech). The membranes were also immunologically reprobed with anti-human Shc antibody after stripping each of the primary antibodies off the membranes.
Inhibition of SAPK/JNK and p38 Kinase-Cells were plated 24 h before transfection at a density of 10 5 /18-mm cover glass. The cells were co-transfected with pcDL-SR␣-SAPK-VPF plasmid expressing dominant-negative SAPK/JNK or pcDL-SR␣-WT-SAPK plasmid expressing wild-type SAPK/JNK as a control, together with pEGFP-Vector (CLON-TECH) using FuGENE TM 6 transfection reagent (Boehringer Mannheim) by following the manufacturer's instructions and/or were then treated with 100 M SB203580 (33), a specific inhibitor of p38 kinase, 15 min prior to exposure to calphostin C for 8 h in U-87MG cells and for 6 h in T98G cells. The cells, after washing three times with phosphatebuffered saline, were incubated for 3 min with 10 M Hoechst 33258 or 10 M propidium iodide and analyzed under an Olympus AX-80 fluorescence microscope. Dominant-negative and wild-type SAPK/JNK expression plasmids were obtained from Prof. E. Nishida (Kyoto University, Kyoto, Japan).

Calphostin C-induced Cellular
Responses and DNA Fragmentation-Glioma cells exposed to 100 nM calphostin C showed an apoptotic morphology: cell shrinkage and surface blebbing by phase contrast images, and condensation or fragmentation of nuclei by Hoechst 33258 stain (Figs. 1 and 11). Mutant p53-positive T98G cells appeared to become more quickly apoptotic than wild type p53-positive U-87MG cells, suggesting that calphostin C-induced apoptosis is p53-independent. Unambiguous electrophoretic patterns of laddered oligonucleosomal DNA fragments were observed as early as 8 h in U-87MG cells and 6 h in T98G cells following exposure to 100 nM calphostin C, as shown representatively in Fig. 2, lanes 2 and 7, respectively.
Calphostin C-induced Down-regulation of Bcl-2 and Bcl-x L and Activation of Caspase-3-Bcl-2 was shown as a single 26-kDa peptide by immunoblot analysis, Bax as a single 21-kDa peptide, and Bcl-x L as a doublet of peptides with an apparent molecular mass of 29 -31 kDa. The 19-kDa Bcl-x S was not detected. Bcl-2 and Bcl-x L were down-regulated nearly synchronously as early as 2-4 h after treatment with 100 nM calphostin C in both U-87MG and T98G cells. In contrast, Bax levels in both U-87MG and T98G cells were not changed during treatment with calphostin C (Fig. 3).
There was no evidence of caspase-1 activation following 100 nM calphostin C treatment, since immunoreactive active 10-kDa caspase-1 fragment was not detected in both U-87MG and T98G cells during treatment with calphostin C for 12 h (Fig. 4). The levels of 45-kDa procaspase-1 remained unchanged during calphostin C treatment for the first 6 h and thereafter were decreased in both U-87MG and T98G cells, but active caspase-1 fragment was not formed, probably due to nonspecific proteolysis. Immunoreactive 32-kDa procaspase-3 was degraded into active 12-kDa caspase-3 fragment as early as 8 h in U-87MG cells and 4 h in T98G cells after treatment with calphostin C, persisting for at least 12 h in both U-87MG and T98G cells (Fig.  4). The generation of a COOH-terminal 85-kDa PARP apoptotic fragment (34,35) was shown to be associated with the generation of active 12-kDa caspase-3 fragment (Fig. 4). These data are consistent with activation of caspase-3 in calphostin Ctreated gliomas.
Two tetrapeptide derivatives, Ac-YVAD.cho and Ac-DEVD-.cho, which are relatively selective inhibitors of caspases-1 and -3, respectively (36,37), were added to the culture medium 30 min prior to the exposure to calphostin C to preferentially block one or the other group of cysteine proteases. The induction of oligonucleosomal DNA laddering by calphostin C could be completely blocked by treatment with 100 M Ac-DEVD.cho in both U-87MG and T98G cells but not by treatment with 100 M Ac-YVAD.cho (Fig. 2), confirming the activation of caspase-3 but not caspase-1. These findings again suggest that calphostin C-induced apoptosis proceeds independently of caspase-1 and may depend on caspase-3 or similar proteases. The measure- Formation of active 10-kDa caspase-1 fragment from 45-kDa procaspase-1 was not detected during treatment with 100 nM calphostin C for 12 h, suggesting no activation of caspase-1, whereas 32-kDa procaspase-3 was gradually degraded into a 12-kDa active caspase-3 fragment in both U-87MG and T98G cells, indicating activation of caspase-3. Cleavage of 116-kDa PARP into an 85-kDa apoptotic fragment was monitored by immunoblot analysis. Similar results were achieved in three separate experiments. ments of activities of caspase-1 and -3 using fluorescent substrates also indicate similar results. 2 Calphostin C-induced Shc Phosphorylation, Raf-1 Activation, and ERK Phosphorylation-The anti-human Shc antibody specifically recognized 46-, 52-, and 66-kDa isoforms of Shc as shown in Fig. 5. When the Shc immunoprecipitates were immunoblotted with anti-phosphotyrosine antibody, 46-, 52-, and 66-kDa Shc isoforms were shown to be tyrosine-phosphorylated, and the 52-kDa isoform was the major phosphorylated species (Fig. 5). The phosphorylation levels of Shc were increased as early as 2 h after treatment with 100 nM calphostin C and persisting for at least 6 h in both U-87MG and T98G cells, whereas the levels of immunoreactive Shc were not changed during the increase in the Shc phosphorylation, suggesting Shc activation. The activation of Raf1 was determined by the presence of Raf1 with reduced gel electrophoretic mobility due to phosphorylation (27,38,39). As shown in Fig. 6, immunoreactive Raf1 doublet in control cells became a singlet 2 to 8 h after treatment with 100 nM calphostin C in both U-87MG and T98G cells, suggesting a marked decrease in phosphorylation of Raf1. Phosphorylation of ERK was estimated by an in gel MBP kinase assay. Two isoforms of ERK1 and ERK2 were shown to be phosphorylated to a similar degree prior to treatment with calphostin C, and the phosphorylation of ERK1 and ERK2 begins to rapidly diminish as early as 2 h after treatment with calphostin C in both U-87MG and T98G cells and thereafter continuously decreased, as shown representatively in Fig. 7. On the contrary, the levels of immunoreactive ERK1 and ERK2 were not changed during treatment with calphostin C for 8 h and were slightly decreased by treatment for 12 h in both U-87 MG and T98G cells as shown in Fig.  7.
Calphostin C-induced Activation of SAPK/JNK and p38 Kinase-The activities of SAPK/JNK and p38 kinase were analyzed by Western blotting using anti-phosphospecific SAPK/ JNK (Thr 183 /Tyr 185 ) antibody and anti-phosphospecific p38 kinase (Tyr 182 ) antibody, respectively. SAPK/JNK was shown as a doublet, and p38 kinase was shown as a singlet. Phospho-SAPK/JNK1 and -SAPK/JNK2 (Thr 183 /Tyr 185 ) as well as phospho-p38 kinase (Tyr 182 ) were increased gradually as early as 2 h after treatment with calphostin C in both U-87MG and T98G cells. The phosphorylation of SAPK/JNK1 and SAPK/ JNK2 as well as p38 kinase reached a peak at 8 or 12 h in U-87MG cells and at 6 or 8 h in T98G cells (Figs. 8 and 9). SAPK/JNK2 was sometimes slightly more phosphorylated than SAPK/JNK1. On the other hand, the levels of immunoreactive SAPK/JNK1 and SAPK/JNK2 as well as p38 kinase were not changed during the increase in phosphorylation of SAPK/JNK and p38 kinase in both U-87MG and T98G cells. Therefore, SAPK/JNK and p38 kinase are activated during treatment with calphostin C for 12 h in U-87MG and T98G cells.
Calphostin C-induced Activation of SAPK/JNK and p38 Kinase Is Inhibited by z-VAD.fmk or Ac-DEVD.cmk-To analyze the relationship between caspases and SAPK/JNK or p38 kinase, we used two caspase inhibitors with different specificities: 100 M z-VAD.fmk (a broad spectrum caspase inhibitor) and 100 M Ac-DEVD.cho. They were added to the culture medium 30 min prior to the addition of 100 nM calphostin C. The COOH-terminal 85-kDa PARP fragment generated by calphostin C treatment for 8 h in U-87MG and for 6 h in T98G cells was not detected in the presence of z-VAD.fmk or Ac-DEVD.cho (Fig. 10A), suggesting a marked inhibition of caspase-3-like proteases by z-VAD.fmk or Ac-DEVD.cho (34,35,37,40). In addition, oligonucleosomal DNA fragmentation  6. Analysis of Raf1 activation shows that immunoreactive Raf1 with reduced mobility (upper arrow) is barely detectable as early as 2 h after treatment with 100 nM calphostin C and persisting during the treatment for at least 8 h in both U-87MG and T98G cells, suggesting the decrease in phosphorylation of Raf1 by calphostin C. Cell lysates were subjected to 12% SDS-polyacrylamide gel electrophoresis and probed with anti-Raf1 antibody by immunoblotting. Similar results were achieved in three separate experiments with comparable outcomes. Ϫ, no treatment; ϩ, treatment.

FIG. 7. Analysis of activities of ERK in U-87MG and T98G cells by in gel MBP kinase assay, showing a rapid decline in levels of phosphorylation of ERK1 and ERK2 as early as 2 h after stimulation with 100 nM calphostin C and a continuous decrease during the treatment for 4 -12 h in U-87MG (A) and T98G (B) cells.
Cell lysates were electrophoresed in 11% SDS-polyacrylamide gel containing 0.5 mg/ml MBP. The gel was incubated with a kinase buffer containing 25 Ci of [␥-32 P]ATP to measure kinase activity. Cell lysates were also immunologically probed with anti-ERK1/ERK2 antibody, showing no changes in levels of ERK1 and ERK2 during the treatment with calphostin C for 8 h and thereafter slight decreases in levels of ERK1 and ERK2 in U-87MG (AЈ) and T98G (BЈ) cells. Results are representative of three independent experiments. P, phosphorylation. induced by 100 nM calphostin C was not visualized in the presence of z-VAD.fmk or Ac-DEVD.cho (Fig. 3, lanes 4, 5, 9, and 10). The pretreatment with z-VAD.fmk or Ac-DEVD.cho had no stimulatory effect on calphostin C-induced decrease in phosphorylation of ERK1 and ERK2 (Fig. 10B), whereas calphostin C-induced activation of SAPK/JNK and p38 kinase was almost completely blocked by the pretreatment with z-VAD-.fmk or Ac-DEVD.cho (Fig. 10, C and D).  (Fig. 1, e, eЈ, g, and gЈ). They also blocked chromatin condensation and fragmentation as revealed by Hoechst 33258 stain (Fig. 1, f, fЈ, h, and hЈ), and this result was consistent with the inhibition of oligonucleosomal DNA fragmentation by z-VAD.fmk or Ac-DEVD.cho as shown by agarose gel electrophoresis (Fig. 2). In contrast, the pretreatment with Ac-YVAD.cho did not inhibit calphostin C-induced apoptotic morphological features (Fig. 1, i and iЈ) and oligonucleosomal DNA fragmentation (Fig. 2). To examine the viability of calphostin C-treated cells, we used the trypan blue dye exclusion method. Two caspase inhibitors (100 M z-VAD-.fmk or 100 M Ac-DEVD.cho), which inhibited calphostin Cinduced activation of SAPK/JNK and p38 kinase as well as cell shrinkage and apoptotic nuclear changes but did not block calphostin C-induced surface blebbing, could not prevent calphostin C-induced cell death in both U-87MG and T98G cells (Fig. 10E). It is suggested, therefore, that the activation of SAPK/JNK and p38 kinase is dispensable for z-VAD.fmk-resistant surface blebbing and cell death and that both glioma cell lines succumb not to the nuclear damages but to the surface blebbing events in the presence of z-VAD.fmk or Ac-DEVD.cho. Ac-YVAD.cmk did not inhibit the calphostin Cmediated apoptotic morphological features and cell death in both cell lines.
Transfection of dominant negative SAPK/JNK was achieved in about 75% of U-87MG and T98G cells. Transfection of dominant negative SAPK/JNK or treatment with SB203580 induced similar effects on calphostin C-mediated cellular responses in both U-87MG and T98G cells as did z-VAD.fmk or Ac-DEVD.cho: marked decrease in cell shrinkage and chroma- tin condensation and fragmentation but maintenance of surface blebbing (Fig. 11, A and B). The marked decrease in calphostin C-induced apoptotic nuclear damages by dominant negative SAPK/JNK or SB203580, as revealed by Hoechst 33258 stain, suggested the inactivation of caspase-3, but the calphostin C-induced PARP cleavage was unchanged by the inhibition of SAPK/JNK or p38 kinase (Fig. 11C), although about 25% of untransfected U-87MG and T98G cells still demonstrated the apoptotic nuclear damage. However, calphostin C-induced cell death shown by the propidium iodide dye exclusion method in Fig. 11, A and B, was not inhibited by transfection of dominant negative SAPK/JNK or SB203580.

DISCUSSION
Since treatment of cells with 12-O-tetradecanoylphorbol-13acetate results in activation of Raf1 (7) and ERK (8 -11), the effect of calphostin C on the signaling pathway leading to ERK activation was studied. In general, ERK1 and ERK2 are most strongly stimulated by activation of protein-tyrosine kinase receptors (13), and one established mechanism of receptor protein-tyrosine kinase coupling involves activation of Ras by translocation of the Grb2-Sos complex to the plasma membrane, as a consequence of its binding to an autophosphorylation site in the receptor protein-tyrosine kinase itself or to a substrate or adaptor protein, such as Shc, phosphorylated by the receptor protein-tyrosine kinase (41). In addition, ERK is activated by both Ras-dependent (2)(3)(4) and Ras-independent (5) signaling pathways in response to activation of G proteincoupled receptors. The release of ␤␥ subunits of heterotrimeric G proteins following ligand binding to G i -coupled receptors leads to tyrosine phosphorylation of Shc and increased functional association of Shc, Grb2, and Sos in COS-7 cells (4). These data suggest that G␤␥ subunits can activate Ras using mechanisms that are similar to those employed by receptor tyrosine kinases and that Shc is implicated as a go-between that relays message from G␤␥-coupled receptors to proteintyrosine kinase receptors. On the other hand, Ras-independent signaling pathways involve the G␣ subunits of G o and G q -coupled receptors and are PKC-dependent (5).
In calphostin C-induced apoptosis, Raf1 and ERK in both U-87MG and T98G cells were inactivated despite Shc activation prior to the onset of apoptosis. In general, signal transduction from Shc leads to Ras activation, which recruits Raf1 to the plasma membrane, where Raf1 is activated by an intricate multistep process (42). PKC has been reported to activate Raf1 (43)(44)(45), but whether PKC directly phosphorylates and regulates Raf1 activity requires further investigation. A dominant inhibitory form of Raf1 abrogates the protective effect of Bcl-2 on cell survival (46). Bcl-2 is reported to target Raf1 to the outer mitochondrial membrane (46), where Raf1 mediates the phosphorylation of Bad. The phosphorylated Bad is sequestered in the cytosol bound to 14-3-3, and only unphosphorylated Bad heterodimerizes with Bcl-2 or Bcl-x L to promote cell death (47). However, the cross-talk between Bcl-2 and Raf1 may be little developed in calphostin C-induced apoptosis because of the down-regulation of Bcl-2 and the inactivation of Raf1, but abrogation of c-Raf expression is reported to induce apoptosis in tumor cells (48). Direct and selective activation of the ERK pathway may actually suppress the effects of factors that mediate apoptosis (22,23), and Raf1-ERK signaling pathway plays a pivotal role in suppressing apoptotic death probably due to rapid up-regulation of bcl-2 and bcl-x L (27,28), both suggesting that Raf1-ERK signaling is a survival pathway.
In addition to the inhibition of ERK, the activation of SAPK/ JNK and p38 kinase coincided with the onset of calphostin C-induced apoptosis, as determined by apoptotic morphological features, suggesting that activation of SAPK/JNK and p38 kinase appears to be required for induction of calphostin Cinduced apoptosis in glioma cells. The cleavage of mitogenactivated protein kinase kinase kinase 1 into the 91-kDa fragment induced by caspase-3 (50) increases the kinase activity of mitogen-activated protein kinase kinase kinase 1 to enhance SAPK/JNK activation (51,52). The inhibition of ERK below a basal threshold level is reported to trigger apoptosis in HeLa cells with activation of p38 kinase but not SAPK/JNK (49). The inhibition of ERK activity and the concomitant activation of SAPK/JNK and p38 kinase has been reported in apoptosis of nerve growth factor-deprived PC12 cells (22), Fas-stimulated Jurkat cells (15,24), and UV-irradiated mouse fibroblasts (25). Thus, it seems that the ability of a cell to die or survive may be dictated by a critical balance between the ERK and the SAPK/ JNK and p38 pathway (26). However, the results as to the involvement of SAPK/JNK and p38 kinase in apoptosis are conflicting so far, depending upon variety of stimuli and cell types.
In the calphostin C-mediated apoptosis, the down-regulation of Bcl-2 and Bcl-x L induces cytochrome c release from mitochondria 2 to activate caspase-3. Here, we try to shed light on the molecular details of the contribution of caspase-3 to the activation of the SAPK/JNK and p38 signaling pathway using three caspase inhibitors (z-VAD.fmk, Ac-DEVD.cho, and Ac-YVAD.cho). Calphostin C-induced inhibition of ERK in glioma cells remained unchanged by z-VAD.fmk or Ac-DEVD.cho. However, both of these inhibitors inhibited calphostin C-induced activation of SAPK/JNK and p38 kinase as well as apoptotic nuclear damage (chromatin condensation and DNA fragmentation) and cell shrinkage but did not block calphostin C-induced surface blebbing and cell death. In addition, the induction of cell death by calphostin C was not changed by transfection of dominant negative SAPK/JNK or treatment with SB203580, suggesting that the activation of SAPK/JNK or p38 kinase is dispensable for cell death induced by calphostin C. Calphostin C-induced cell death in U937 cells was reported to be unaffected by genetic manipulation of SAPK/JNK (53). Furthermore, the hypothesis that stress-activated protein kinase/ERK kinase (SEK1)-mediated activation of SAPK/JNK is required for the induction of cell death in response to apoptotic inducers is invalidated by experiments involving cell lines or animals in which the SEK1 gene is inactivated. Apoptosis does occur in SEK Ϫ/Ϫ ES cells, SEK1 Ϫ/Ϫ thymocytes, and SEK1 Ϫ/Ϫ splenic T cells in response to anisomycin, serum depletion, UV and ␥-irradiation, sorbitol-mediated changes in osmolarity, heat shock, anti-cancer drugs (etoposide, adriamycin, and cisplatinum), CD3/CD28 ligation, and phorbol myristate acetate/ Ca 2ϩ ionophore with similar kinetics and at similar doses as in SEK ϩ/ϩ and SEK1 ϩ/Ϫ cells (54).
In the calphostin C-mediated apoptosis, z-VAD.fmk and Ac-DEVD.cho inhibit not only caspase-3 but also SAPK/JNK and p38 kinase, resulting in inhibition of apoptotic nuclear damage and cell shrinkage but maintenance of cell death and surface blebbing. Ac-YVAD.cho did not block any calphostin C-mediated responses. It is suggested that caspase-3-like proteases function upstream of SAPK/JNK and p38 kinase in calphostin C-induced apoptosis, in which cells die without nuclear damages. In caspase-independent commitment and death systems, inhibition of caspase activation prevents acquisition of the apoptotic phenotype but does not inhibit cell death. This applies to a number of different models of apoptosis induced by overexpression of bax (55), c-myc overexpression, and serum withdrawal or overexpression of bak (56), protonophore, protoporphyrine IX, dexamethasone, ceramide, etoposide, or nitric oxide (57)(58)(59)(60)(61)(62) and cytotoxic T lymphocyte granule exocytosis (63). However, Bcl-2 maintains survival of cells that have been induced to die in a caspase-independent manner with glucocorticoids, growth factor withdrawal, c-Myc, Bax, or Bak (55,56,61,62). Therefore, the down-regulation of Bcl-2 in calphostin C-induced apoptosis is an important commitment event that determines whether or not a cell dies even when it is independent of caspase activation. On the other hand, the inhibition of SAPK/JNK and p38 kinase in calphostin C-induced apoptosis did not inhibit PARP cleavage, namely activation of caspase-3, but induced similar effects on the apoptotic phenotypes and the cell death as did z-VAD.fmk and Ac-DEVD.cho. Since caspase-3 cleaves DNA fragmentation factor 45 (DFF45)/inhibitor of caspase-activated DNase and inactivates its DFF40/caspaseactivated DNase-inhibitory effect, allowing DFF40/caspase-activated DNase to enter the nucleus and degrade chromosomal DNA (64 -66), the absence of calphostin C-induced chromatin condensation or fragmentation by the inhibition of SAPK/JNK or p38 kinase suggests that SAPK/JNK or p38 kinase is involved downstream of caspase-3 in the inactivation of DFF45/ inhibitor of caspase-activated DNase and/or in the translocation of DFF40/caspase-activated DNase into the nucleus.