Involvement of Sphingosine in Mitochondria-dependent Fas-induced Apoptosis of Type II Jurkat T Cells*

Exposure to anti-Fas antibody in Jurkat cells (type II cells), which are characterized by a weak caspase-8 activation at the death-inducing signaling complex (DISC), induced a biphasic increase in ceramide levels. The early generation of ceramide preceded transient activation of acidic ceramidase and subsequent production of sphingosine, followed by cytochrome c release, activation of caspases-2, -3, -6, -7, -8, and -9, Bid cleavage, and a later sustained ceramide accumulation. The caspase inhibitor benzyloxycarbonyl-Val-Ala-Asp-fluoromethyl ketone inhibited early increases of ceramide and sphingosine, whereas overexpression of Bcl-xL had no effect, and both prevented the later sustained ceramide accumulation. Exogenous sphingosine, as well as cell-permeable C2-ceramide, induced cytochromec release from mitochondria in a caspase-independent fashion leading to activation of caspase-9 and executioner caspases and, surprisingly, activation of the initiator caspase-8 and processing of its substrate Bid. These effects were also completely abolished by Bcl-xL overexpression. Our results suggest that sphingosine might also be involved in the mitochondria-mediated pathway of Fas-induced cell death in type II cells.

Exposure to anti-Fas antibody in Jurkat cells (type II cells), which are characterized by a weak caspase-8 activation at the death-inducing signaling complex (DISC), induced a biphasic increase in ceramide levels. The early generation of ceramide preceded transient activation of acidic ceramidase and subsequent production of sphingosine, followed by cytochrome c release, activation of caspases-2, -3, -6, -7, -8, and -9, Bid cleavage, and a later sustained ceramide accumulation. The caspase inhibitor benzyloxycarbonyl-Val-Ala-Asp-fluoromethyl ketone inhibited early increases of ceramide and sphingosine, whereas overexpression of Bcl-x L had no effect, and both prevented the later sustained ceramide accumulation. Exogenous sphingosine, as well as cell-permeable C 2 -ceramide, induced cytochrome c release from mitochondria in a caspase-independent fashion leading to activation of caspase-9 and executioner caspases and, surprisingly, activation of the initiator caspase-8 and processing of its substrate Bid. These effects were also completely abolished by Bcl-x L overexpression. Our results suggest that sphingosine might also be involved in the mitochondria-mediated pathway of Fas-induced cell death in type II cells.
Fas induces apoptosis when cross-linked with either agonistic antibodies or its natural ligand via formation of the deathinducing signal complex (DISC) 1 by recruiting Fas-associated death domain protein adapter proteins and caspase-8 (caspase-8) proenzymes, leading to activation of caspase-8 by autocatalytic cleavage (1). Recent studies have described the existence of two types of Fas-signaling pathways that depend on the amount of active caspase-8 produced at the DISC (2). Type I cells are characterized by abundant recruitment of procaspase-8 to the DISC following Fas ligation, resulting in sub-sequent direct activation of executioner caspases in a Bcl-2/Bclx L -insensitive fashion (2). In contrast, in type II cells (e.g. Jurkat cells), amounts of active caspase-8 generated at the DISC are very low, and apoptosis in these cells is believed to depend on Bcl-2 or Bcl-x L -regulated translocation of cytochrome c (cyt c) from the mitochondria to the cytosol (2). Once released, cyt c binds to Apaf-1 (3), inducing it to associate with procaspase-9 (4), thereby triggering its autoactivation into a mature form, which in turn can directly cleave procaspase-3 and -7 (5). It has recently been proposed that in type II cells, Bid, a pro-apoptotic member of the Bcl-2 family, is activated by caspase-8 and, following cleavage, is translocated from the cytosol to the mitochondria, where it can mediate cyt c release (6 -8).
Ceramide generated from sphingomyelin by activation of a sphingomyelinase has been implicated in apoptosis induced by diverse apoptotic stimuli, in particular by TNF␣ or Fas ligand (9). In contrast to numerous studies demonstrating potential involvement of ceramide in Fas signaling (10 -21), others have challenged the concept of a role for ceramide (22)(23)(24)(25)(26)(27). Furthermore, it is possible that a further metabolite of ceramide may be the active component produced as a result of sphingomyelin turnover. One metabolite, sphingosine, has been shown to be rapidly produced during TNF␣-mediated apoptosis in human neutrophils (28), and sphingosine is capable of inducing apoptosis when added exogenously to many cell types (28 -34).
Herein, we report that treatment of Jurkat cells with anti-Fas antibody induces rapid generation of ceramide immediately followed by transient production of sphingosine, both preceding cyt c release and activation of executioner caspase-3, -6, and -7 as well as caspase-8 and Bid activation. In addition, exogenous C 2 -ceramide and sphingosine recapitulate Fas-induced apoptosis, cyt c redistribution, and activation of caspase-2, -3, -6, -7, -8, -9, and Bid, in a Bcl-x L -sensitive fashion. Our findings suggest the potential involvement of sphingosine in mitochondria-dependent apoptosis in type II cells.
Mass Measurements of Ceramide and Sphingosine-Ceramide and sphingosine levels were measured and normalized to total cellular phospholipids as described previously (35).
Apoptosis Assay-Apoptosis was assessed by Hoechst staining as described previously (37). The percent apoptotic cells was calculated as number of apoptotic cells per number of total cells counted.
Preparation of Mitochondria and Western Blot Analysis of Cyt c-Mitochondrial preparations were performed as described previously (38). Cells were harvested by centrifugation at 1,000 ϫ g for 5 min at 4°C. After washing with ice-cold phosphate-buffered saline, mitochondrial and cytosolic fractions were prepared by resuspending cell pellets in 5 volumes of ice-cold buffer A (20 mM Hepes-KOH, pH 7.5, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM sodium EDTA, 1 mM sodium EGTA, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, 20 g/ml leupeptin, 10 g/ml aprotinin, and 10 g/ml pepstatin A) containing 250 mM sucrose. Cells were lysed by 15-20 passages through a 25-gauge needle, and homogenates were centrifuged at 1,000 ϫ g for 5 min at 4°C. The supernatants were then centrifuged at 10,000 ϫ g for 15 min at 4°C, and the resulting mitochondria pellets were resuspended in cold buffer A. For Western blot analysis, equal amounts of mitochondrial and cytosolic proteins were separated on 15% SDS-PAGE and then transblotted to nitrocellulose. Anti-cyt c mAb (PharMingen) and anti-cyt c oxidase subunit II mAb (Molecular Probes, Eugene, OR) were used as primary antibodies.
Fluorogenic DEVD Cleavage Enzyme Assays-Enzyme reactions were performed in 96-well plates with 20 g of cytosolic proteins and a final concentration of 20 M Ac-DEVD-AMC substrate as described previously (37).

Early Transient Increases in Ceramide and Sphingosine during Fas-induced Apoptosis Precede Executioner Caspases Activation and Cell Death and Are Followed by a Sustained
Accumulation of Ceramide-As previously reported (10,14,20,21), we found that treatment of Jurkat cells with 50 ng/ml anti-Fas antibody resulted in acute changes in ceramide levels within 30 min from a basal level of ϳ10 pmol/nmol total phospholipids (Fig. 1A). Moreover, in agreement with recent studies (15,27), we also detected a late and sustained increase in ceramide beginning within 2 h of treatment and increasing by 1.5-2-fold within 6 h (Fig. 1A). The early generation of ceramide preceded activation of the executioner caspases as demonstrated by DEVDase activity measurements (Fig. 1C) and the appearance of nuclear fragmentation (Fig. 1D).
Sphingosine, the product of ceramide hydrolysis catalyzed by ceramidases, has been shown to be rapidly produced during TNF␣-mediated apoptosis in human neutrophils (28) and rat cardiomyocytes (32). Hence, it was of interest to determine whether cross-linking of Fas could also increase sphingosine levels. The basal level of sphingosine in Jurkat cells is ϳ0.1 pmol/nmol total phospholipids. Treatment with anti-Fas mAb caused a rapid 2-fold increase in sphingosine with a peak at 30 -45 min (Fig. 1A), which proceeded the rapid ceramide increase, and returned to basal levels by 60 -90 min (Fig. 1A). Treatment with exogenous C 2 -ceramide also resulted in increased sphingosine levels (Fig. 1B), which preceded the increased DEVDase activity (Fig. 1C) and apoptosis (Fig. 1D) induced by C 2 -ceramide. Similarly, treatment with exogenous long chain C 16 -ceramide (25 M) or with bacterial sphingomyelinase (100 milliunits/ml), which hydrolyzes membrane sphingomyelin to increase long chain ceramide levels, markedly induced activation of caspases and formation of apoptotic nuclei, in agreement with a previous report (15).
In agreement with previous studies in other cell lines (28 -34), addition of exogenous sphingosine also induced apoptosis in Jurkat cells in a time-and concentration-dependent manner ( Fig. 1D). Nuclear fragmentation was clearly evident within 3 h after the addition of sphingosine (data not shown). After treatment for 6 h with 10 M sphingosine, almost 75% of the cells showed apoptotic changes ( Fig. 1D), whereas a concentration as low as 2.5 M was sufficient to induce fragmented nuclei in ϳ30% of the cells (data not shown).
To examine the possibility that sphingosine could induce apoptosis due to its conversion to ceramide, we investigated the effect of the mycotoxin fumonisin B1, a known inhibitor of ceramide synthase (39). Similar to previous results in HL-60 cells (29 -31) and Hep3B hepatoma cells (34), fumonisin B1 itself did not induce apoptosis nor did it affect the extent of apoptosis induced by sphingosine ( Fig. 2A). Although this ceramide synthase inhibitor almost completely inhibited the late peak of ceramide induced by Fas ligation (Fig. 2B), it had no significant effect on the extent of apoptosis ( Fig. 2A), suggesting that the late increase in ceramide is not required for Fasinduced apoptosis. Moreover, a higher concentration of fumonisin B1 (100 M) not only inhibited elevation of ceramide induced by 5 M sphingosine by more than 70% after 5 h of treatment, it also markedly enhanced apoptosis induced by sphingosine. It should be noted that similar to addition of sphingosine, treatment with C 2 -ceramide also resulted in the appearance of a late peak of endogenous ceramide (Fig. 2B), which could be due in part to conversion of sphingosine ( Fig.  1B) to ceramide by ceramide synthase, inasmuch as fumonisin B1 was capable of inhibiting the late increase.
Sphingosine Induces Apoptosis by Activation of Executioner Caspases in a Bcl-x L -inhibitable Fashion-Because current evidence suggests that ceramide functions proximal to the executioner caspases (15, 37), we examined whether addition of sphingosine resulted in activation of caspase-3 and -7, which drive the effector phase of apoptosis. To determine caspase activity, we used the fluorogenic substrate, Ac-DEVD-AMC, which corresponds to the cleavage site found in numerous caspase-3 and -7 targets (40). Sphingosine treatment resulted in a time-dependent increase in DEVDase activity which correlated with the onset of apoptosis and preceded the appearance of fragmented nuclei by approximately 1 h (Fig. 1, C and D). A similar correlation was observed with ceramide or anti-Fas treatment, which increased DEVDase activity prior to detection of the first morphological signs of apoptosis. Proteolytic processing of procaspase-3 and -7 was also examined by Western blotting. In agreement with our previous study (37), procaspase-3 and -7 were cleaved into their respective active forms after anti-Fas (Fig. 3, A and B) treatment in a time-dependent fashion similar to the increase in DEVDase activity (Fig. 1C). Extracts from cells treated with sphingosine also showed significantly increased caspase-3 and -7 processing (Fig. 3, E and F), again with a similar time course to the increase in DEVDase activity (Fig. 1C).
Because overexpression of Bcl-2 or Bcl-x L protects against diverse apoptotic stimuli, including ceramide-induced apoptosis (41)(42)(43)(44)(45)(46), it was important to determine whether the effects of sphingosine were also blocked by these proteins. Jurkat cells stably transfected with bcl-x L expressed high levels of Bcl-x L protein ( Fig. 4A) and were completely resistant to apoptosis induced by C 2 -ceramide as well as sphingosine (Fig. 4B). Bcl-x L similarly prevented C 2 -ceramide-and sphingosine-induced increases in DEVDase activity (Fig. 4C). In agreement with other studies (27,42), Bcl-x L also blocked apoptosis induced by C 16ceramide and bacterial sphingomyelinase which hydrolyzes sphingomyelin to generate natural ceramide. Processing of executioner procaspase-3 (Fig. 5A), -7 (Fig. 5B), and -6 ( Fig. 5C) into their mature forms and cleavage of their respective substrates, PARP and lamin B (data not shown), were also completely prevented in Jurkat cells overexpressing Bcl-x L . Caspase-2 shares the optimal peptide recognition motif DEXD with executioner caspase-3 and -7 (47). Activation and processing of procaspase-2 into its mature form occurred in a time-dependent fashion, similar to procaspase-3 and -7 processing (data not shown), and was also abolished by Bcl-x L overexpression in cells treated with ceramide or sphingosine (Fig. 5D).
Effects of Bcl-x L and ZVAD-fmk on Sphingolipid Levels-A wealth of reports indicates that ceramide formation in response to various death triggers is not affected by overexpression of Bcl-2 or Bcl-x L (41,43,46,48,49). In agreement, we found that overexpression of Bcl-x L did not inhibit the early increase in ceramide or sphingosine generation triggered by Fas ligation. Nevertheless, in agreement with Tepper et al. (27), we noted that the late and robust increase in ceramide levels was markedly reduced by Bcl-x L (Fig. 6A). To assess whether caspase activation was required for early ceramide and sphingosine generation triggered by anti-Fas mAb, we utilized the pharmacological peptide ZVAD-fmk, a competitive inhibitor of all caspases (50). Pretreatment of Jurkat cells with ZVAD-fmk (25 M) inhibited early accumulation of ceramide (Fig. 6B), similar to a previous report (20), and sphingosine (Fig. 6B) induced by anti-Fas mAb. Furthermore, the late elevation of ceramide was nearly totally blocked by ZVAD-fmk (5 Ϯ 3% increase versus 55 Ϯ 5% after 5 h). In agreement with these results, it has been shown that Ac-YVAD-chloromethylketone or transient CrmA transfection prevented stimulation of acidic sphingomyelinase, release of ceramide, and Fas-triggered apoptosis in Jurkat T cells (20). Collectively, these results suggest that the early increases in ceramide and sphingosine are upstream of Bcl-x Lregulated events, whereas the second peak in ceramide results from downstream mitochondrial events.
The initial ceramide increase preceded that of sphingosine, suggesting that the increase in sphingosine might result from degradation of the ceramide formed upon Fas ligation. This is a likely possibility because sphingosine is not synthesized de novo and can only be produced by metabolism of ceramide (39). Thus it was of interest to determine whether Fas triggered activation of ceramidase. Two major isoforms of ceramidase, acidic and alkaline forms, exist in most tissues and cells. Although no changes in alkaline or neutral ceramidase activity could be detected after Fas ligation, a rapid and transient increase in acidic ceramidase activity was found, peaking within 20 min (Fig. 6C), in close correlation with the time course for increases in sphingosine levels (Fig. 1A). However, the ceramidase inhibitors (1S,2R)-D-erythro-2-(N-myristoylamino)-1-phenyl-1-propanol, which inhibits alkaline ceramidase (51), and N-oleoylethanolamine (NOE), a proposed acidic ceramidase inhibitor (52), did not affect Fas-induced apoptosis in Jurkat cells. It should be pointed out that NOE may not be a specific acidic ceramidase because it did not inhibit acidic ceramidase activity in an in vitro assay.
To gain further insight into the relative function of endogenous sphingosine, we also used N,N-dimethylsphingosine (DMS), a competitive inhibitor of sphingosine kinase that induced apoptosis in a dose-dependent manner (Fig. 7A). Both nuclear DNA fragmentation and DEVDase activity induced by DMS treatment were blocked by overexpression of Bcl-x L (Fig.  7, B and C) supporting the notion that elevated sphingosine acts upstream of the mitochondrial mediated apoptotic pathway.
Sphingosine Induces Cyt c Release and Caspase-9 Activation in a Bcl-x L -inhibitable Manner-Translocation of mitochondrial cyt c to the cytosol has been shown to occur in response to anti-Fas treatment and is blocked by Bcl-2 or Bcl-x L (2, 53-58). Cyt c then functions to activate Apaf-1 and procaspase-9 proc- essing, which in turn processes executioner caspase-3 and -7 (3)(4)(5). Because Bcl-x L overexpression totally blocks activation of executioner caspases induced by ceramide and sphingosine elevations, it was of interest to determine whether these sphingolipid metabolites could also induce cyt c release. In agreement with recent studies (2,55,59,60), treatment with anti-Fas resulted in a significant increase in cytosolic cyt c levels that occurred within 60 min, with a corresponding decrease in mitochondrial cyt c (Fig. 8A). Similarly, treatment with sphingosine (Fig. 8A) also caused rapid cyt c accumulation in the cytosol and decreased mitochondrial content. Cyt c release induced by anti-Fas, C 2 -ceramide, or sphingosine was prevented by Bcl-x L expression (Fig. 8B). The absence of detectable cyt c oxidase in cytosolic extracts confirmed that our preparations were free of mitochondrial contamination (Fig. 8C). Since sphingosines were capable of inducing cyt c release, we next determined whether caspase-9 was also activated. C 2 -ceramide and sphingosine triggered processing of caspase-9 which was reversed by Bcl-x L overexpression (Fig. 8D). A similar result was obtained in response to anti-Fas mAb treatment (Fig. 8D).
Sphingosine Induces Cyt c Release in a Caspase-independent Manner-Because apoptosis induced by C 2 -ceramide, sphingosine, or anti-Fas was completely blocked by the broad caspase inhibitor ZVAD-fmk (Fig. 9A), we determined whether translocation of cyt c from mitochondria to the cytosol also required caspase activation. However, as shown in Fig. 9B, pretreatment of cells with ZVAD-fmk had no effect on cytosolic accumulation of cyt c in C 2 -ceramide or sphingosine-treated cells, yet the inhibitor effectively suppressed DEVDase activity (data not shown). These results demonstrate that sphingosine and ceramide induce translocation of cyt c from mitochondria to cytosol in a caspase-independent manner. In contrast, ZVADfmk pretreatment led to inhibition of cyt c release induced by anti-Fas (Fig. 9B), as previously reported (61).

Caspase-8 Activation and Bid Processing by Sphingosine Are
Inhibited by Bcl-x L -In Fas-induced apoptosis, caspase-8 plays a pivotal role inasmuch as it is the most apical caspase, directly activated at the DISC (62,63). Therefore, caspase-8 cleavage often represents the first detectable apoptotic event, followed by direct processing of executioner caspases (64 -66). However, in type II cells such as Jurkat, recruitment of procaspase-8 to the DISC is strongly reduced, and caspase-8 activation also requires involvement of mitochondria, as Bcl-2 or Bcl-x L overexpression inhibits its processing (2). Because ceramide and sphingosine induce apoptosis in a Bcl-x L -dependent manner, we determined whether they were also capable of mediating caspase-8 activation. In agreement with recent reports (2,61,67), we found that anti-Fas treatment of Jurkat cells resulted in only slow cleavage of caspase-8; the intermediate cleavage products p43/41 were not readily apparent until after 60 min, and significant processing was clearly evident after this time (Fig. 10A). When comparing the time course for processing of caspase-8 ( Fig. 10A) with caspase-3 and -7 (Fig. 3, A and B), it was difficult to determine which was activated first, since caspase-3 and -7 were activated with similar kinetics (Fig. 3, A  and B). Treatment with exogenous ceramides or sphingosine also induced caspase-8 processing (Fig. 10, D and G). Moreover, the time courses for activation of caspase-8 and caspase-3 and -7 by sphingosine (Fig. 10G versus Fig. 3, E and F) were similar.
Recently, it was reported that Bid, a BH3 domain-containing proapoptotic member of the Bcl-2 family, is cleaved by caspase-8 in Fas-and TNF␣-induced apoptosis into a C-termi-nal fragment that can directly act on mitochondria to trigger cyt c release (6,7,61,68). Because ceramide and sphingosine are capable of inducing cyt c release, we investigated the relationship between these sphingolipid metabolites and Bid activation. In agreement with previous studies (6, 7), Fas ligation resulted in activation of Bid (Fig. 10B). Surprisingly, Bid was also processed in cells treated with C 2 -ceramide (Fig. 10E) or sphingosine (Fig. 10H). Only small amounts of truncated Bid (p15), resulting from processing by caspase-8, are found in the cytosol as the p15 Bid fragment translocates rapidly to the mitochondria. Overexposing films allowed us to visualize the truncated p15 Bid. After treatment with anti-Fas mAb for 1 h, cleavage of Bid was clearly visible (Fig. 11A), which is in accordance with the time course of caspase-8 activation (Fig.  10A). Similarly, sphingosine also induced Bid cleavage that was detected as early as 2 h (Fig. 11B), reminiscent of the kinetics of caspase-8 activation by sphingosine (Fig. 10G).
Bcl-x L overexpression partially inhibited caspase-8 activation and Bid processing in cells treated with anti-Fas, whereas sphingosine-induced activation of caspase-8 and Bid processing was totally blocked (Fig. 12, A and B). These results suggest that Fas-induced activation of caspase-8 and Bid processing proceeds by at least two pathways as follows: one that is blocked by Bcl-x L and one that is Bcl-x L -insensitive and occurs by direct recruitment of caspase-8 to the DISC. As the release of cyt c by exogenous sphingosine occurred earlier than activation of caspase-8 and subsequent processing of Bid, it is unlikely that Bid directly initiates sphingolipid-induced cyt c release in Jurkat cells.
FasL/Fas Interaction Is Not Required for Apoptosis Induced by Sphingosine-Up-regulation of Fas ligand (FasL) has been proposed as a mechanism of ceramide-mediated apoptosis induced by ␥-irradiation or doxorubicin (69). Levels of FasL protein were determined by Western blotting to examine whether sphingosine treatment affected its expression. There were no detectable changes in FasL protein expression discernible after sphingosine or anti-Fas treatment (Fig. 10). To examine further the requirement for Fas signaling in apoptosis induced by sphingosine, we also utilized an antagonistic anti-Fas antibody (clone ZB4), which neutralizes the Fas receptor without inducing apoptosis. Pretreatment for 1 h with 300 ng/ml ZB4 completely blocked apoptosis induced by Fas ligation in agreement with previous studies (70, 71), without affecting apoptosis triggered by treatment with sphingosine (Fig. 13A). Consistent with this result, we found that ZB4 antibody completely inhibited DEVDase activity induced by Fas but not that induced by sphingosine (data not shown). In agreement, blockade of the Fas receptor by ZB4 totally antagonized caspase-8 and -3 processing induced by Fas but not that induced by sphingosine (Fig.  13, B and C). Taken together, these results suggest that sphingosine-mediated activation of caspase-8 does not require Fas signaling.

Potential Involvement of Ceramide and Sphingosine in Fasinduced Apoptosis of Type II Cells-Abundant reports have
suggested that ceramide may play a role in apoptotic signals initiated through the Fas system in various types of cells (10 -21, 27, 72). Notwithstanding, there is disagreement as to both the extent of the increase in ceramide levels as well as the timing, resulting in controversy regarding its role in apoptosis (22)(23)(24)(25)(26)(27). The results presented here may reconcile many findings, as we observed a biphasic ceramide response to Fas ligation in type II Jurkat cells as follows: a rapid and transient low amplitude increase in the cellular concentration of ceramide followed by a delayed, persistent, and robust elevation. Thus, these data are consistent with the reports in which a rapid rise in ceramide has been implicated as an early event in the Fas-signaling pathway in Jurkat cells (10,14,20,21), as well as association solely with a late increase in ceramide (15,27). It is important to note that in many other cell types in which apoptosis occurs more slowly than in Jurkat T cells, such as MCF-7 human breast cancer cells in response to TNF␣ (46,73), the "early" rise in ceramide occurs at a later time, albeit still prior to the effector phase of apoptosis in these cells.
Our study also suggests a potential role for sphingosine, a further metabolite of ceramide, in Fas-induced apoptosis of type II cells, and demonstrates that similar to ceramide it also mediates cyt c release. Moreover, sphingosine induced apoptosis in an executioner caspases-dependent manner, in agreement with previous studies demonstrating that caspase inhibitors block sphingosine-induced apoptosis in HL-60 cells (31) and Hep3B hepatoma cells (34). Ceramide and sphingosine are interconvertable sphingolipid metabolites; however, we have shown that sphingosine induces apoptosis without being converted to ceramide inasmuch as the ceramide synthase inhibitor fumonisin B1 did not affect sphingosine-or Fas-induced apoptosis. However, it should be noted that fumonisin B1 markedly decreased the late increase of ceramide induced by Fas ligation, suggesting that the late increase in ceramide could play a role in amplification rather than initiation of apoptosis. Similar observations have recently been reported in U937 cells where fumonisin B1 inhibited the late increase in ceramide, without affecting cell death induced by C 6 -ceramide or the anticancer drug daunorubicin (48). Although no changes in alkaline or neutral ceramidase activity could be detected, we found that acidic ceramidase activity was rapidly and transiently increased in Jurkat cells after Fas ligation, which correlated closely with the time course for sphingosine formation. It seems unlikely that this acidic ceramidase activity is the previously described lysosomal form (74), because fibroblasts from patients with Farber disease, which accumulate ceramide as the result of genetic deficiency of acid ceramidase, do not show defects in apoptosis (75). Moreover, endogenous long chain ceramide cannot exit from lysosomes, making the ceramides produced in the lysosomes unlikely signaling molecules (76).
Short chain ceramide analogues are commonly used to mimic the effects of endogenous ceramide species. Treatment with bacterial sphingomyelinase or with C 16 -ceramide has often been used to confirm the effects of short chain ceramide analogues. However, we have found that exogenous C 2 -ceramide also increases levels of sphingosine as well as endogenous ceramide in Jurkat cells. Rapid generation of endogenous ceramide in cells treated with C 2 -ceramide has also been previously reported in Jurkat cells (21) and U937 cells (48). The early increases of ceramide and sphingosine after treatment with C 2 -ceramide are similar to those achieved after Fas ligation in Jurkat cells. Moreover, the late ceramide increases after treatment with anti-Fas antibody, C 2 -ceramide, or sphingosine are also comparable. Despite these correlations, it should be emphasized that because the uptake and distribution within the cell of exogenously added sphingolipids is complex, direct correlations between effects of exogenous sphingolipids and increases in endogenous levels cannot be made.
Sphingosine-and Ceramide-mediated Apoptosis Is Mitochondria-dependent-Mitochondrial dysfunction appears to be important in ceramide signaling of apoptosis (77). In vitro studies have shown that ceramide itself is not an efficient inducer of nuclear apoptosis, unless mitochondria are present (78). Bcl-2 and Bcl-x L protect against cell-permeable ceramide-induced apoptosis (41)(42)(43)(44)(45)(46), cyt c release (54), and apoptosis-inducing factor release (78). Moreover, exogenous ceramide can induce mitochondrial permeability transitions, thought to be involved in the mitochondrial inner transmembrane potential (⌬⌿ m ) collapse observed during the effector phase of apoptosis (77), which can be prevented by Bcl-2 overexpression (78). Finally, mitochondria are believed to be a major site of reactive oxygen species production, and cell-permeable ceramide has been reported to increase production of reactive oxygen species (79 -81), resulting in depletion of cellular glutathione (81). Herein, we suggest that sphingosine, as well as ceramide, induces apoptosis in a mitochondria-dependent fashion, as Bcl-x L completely abolishes apoptosis, DEVDase activity, and release of cyt c induced by C 2 -ceramide or by sphingosine. Furthermore, caspase-9, positioned at the apex in the apoptotic signaling cascade activated by cyt c through Apaf-1 binding, was activated by either C 2 -ceramide or sphingosine treatment, and its activation was also inhibited by Bcl-x L . Recent gene targeting experiments have revealed that Apaf-1 mutant embryonic fibroblasts, in contrast to wild type cells, were not susceptible to apoptosis induced by ceramide (82). At present, it is still a matter of debate whether ceramide acts directly or indirectly on mitochondria. Ceramide levels are increased in hepatocyte mitochondria following TNF␣ treatment, with a direct effect on the mitochondrial electron transport chain leading to generation of reactive oxygen species (81). Kroemer and co-workers (78,83) have reported that addition of exogenous ceramide to isolated mitochondria does not generate a permeability transitions-inducing activity, and they suggested that, in intact cells, ceramide may trigger the generation of another signal which then is responsible for induction of permeability transitions. Recent data suggest that ceramide could signal mitochondrial apoptosis by inhibiting the protein kinase Akt, responsible for Bad phosphorylation, hence leading to inhibition of anti-apoptotic protein Bcl-2 by Bad (84 -86). Although little is known about the potential function of sphingosine in the mitochondria, in preliminary experiments we have detected sphingosine in the mitochondria (0.09 Ϯ 0.02 pmol/nmol phospholipids), and we found that its concentration increases by 1.7-fold within 15 min after treatment of Jurkat cells with anti-Fas antibody.
Sphingosine Activates Caspase-8 Downstream of Mitochondria-We have confirmed the observations of Scaffidi et al. (2) that caspase-8 activation could occur downstream of mitochondria in type II cells, as Bcl-x L partially inhibits its processing induced by Fas stimulation. However, sphingosine-mediated caspase-8 activation was totally blocked by Bcl-x L overexpression, yet these effects were independent of Fas receptor/ligand signaling. Importantly, we also found that subsequent activa- tion of Bid by anti-Fas or sphingosine was inhibited by Bcl-x L to the same extent as inhibition of caspase-8 activation. Hence, we infer that caspase-8 can be activated both upstream and downstream of mitochondria in Fas-induced apoptosis but only downstream of mitochondria in ceramide-and sphingosinemediated cell death.
Proposed Role of Sphingosine and Ceramide in Fas-induced Apoptosis of Type II Cells-Consistent with other studies on inhibition of Fas-induced apoptosis by Bcl-2 or Bcl-x L in type II cells (2,55,67,(87)(88)(89), we found that Bcl-x L inhibited Fasinduced apoptosis in Jurkat cells, albeit to a lesser extent than its effect on ceramide-and sphingosine-induced apoptosis. Hence, we suggest the existence of distinct pathways originating from the Fas receptor, one involving direct activation of caspase-3 via caspase-8, functioning independently of mitochondria, and a second, which is under the control of Bcl-x L , that may involve sphingolipid metabolites as well as Bid in regulating mitochondrial events. Thus, cross-linking of Fas could rapidly trigger generation of ceramide, in a caspase-dependent manner, which is then converted to sphingosine, and both sphingolipid metabolites may be involved in the translocation of cyt c from mitochondria to the cytosol (under the control of Bcl-x L ), where it can stimulate proteolytic activation of executioner caspases through caspase-9. These sphingolipid metabolites can also trigger activation of caspase-8 and Bid in a mitochondria-sensitive fashion, although most likely not via cyt c, but rather perhaps through the action of another apoptogenic molecule. Nonetheless, activation of caspase-8 can occur independently of Bcl-x L (and sphingolipid metabolites), culminating in processing of caspase-3 and/or activation of Bid, which is then capable of inducing cyt c release once translocated to the mitochondria.
A recent study has established that only type II cells, such as Jurkat or CEM cells, are sensitive to exogenous C 2 -ceramide, whereas type I cells, such as SKW6 or H9 cells, are not responsive (90). This result supports the view that sphingolipid metabolites act at the level of the mitochondria and suggests that type I cells lack a component downstream of the mitochondria that is necessary for the processing of caspase-3. Moreover, activation of protein kinase C inhibits Fas-induced apoptosis in type II cells but not in type I cells (90). This is in agreement with our previous study in which activation of protein kinase C was shown to block totally apoptosis and executioner caspases activation induced by anti-Fas or C 2 -ceramide treatment in Jurkat cells (37). Interestingly, activation of protein kinase C can lead to the activation of sphingosine kinase (91), the enzyme responsible for the conversion of sphingosine into sphingosine 1-phosphate, a sphingolipid metabolite known to antagonize ceramide-mediated apoptosis (35,37,(91)(92)(93). Treatment of type II Jurkat cells with exogenous sphingosine 1-phosphate or overexpression of sphingosine kinase prevented both apoptosis and executioner caspases activation induced by Fas or C 2 -ceramide (37,94). Notably, inhibition of sphingosine kinase by the competitive inhibitor, DMS (95), causes apoptosis in many cell types, including Jurkat T cells. Therefore, the modulation of both ceramidase and sphingosine kinase activities may play a pivotal role in the regulation of apoptosis by regulating the intracellular ratio between sphingosine and sphingosine 1-phosphate.