Sphingosine 1-phosphate inhibits activation of caspases that cleave poly(ADP-ribose) polymerase and lamins during Fas- and ceramide-mediated apoptosis in Jurkat T lymphocytes.

Ceramide, a sphingolipid generated by the hydrolysis of membrane-associated sphingomyelin, appears to play a role as a gauge of apoptosis. A further metabolite of ceramide, sphingosine 1-phosphate (SPP), prevents ceramide-mediated apoptosis, and it has been suggested that the balance between intracellular ceramide and SPP levels may determine the cell fate (Cuvillier, O., Pirianov, G, Kleuser, B., Vanek, P. G., Coso, O. A., Gutkind, J. S., and Spiegel, S. (1996) Nature 381, 800-803). Here, we investigated the role of SPP and the protein kinase C activator, phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA), in the caspase cascade leading to the proteolysis of poly(ADP-ribose) polymerase (PARP) and lamins. In Jurkat T cells, Fas ligation or addition of exogenous C2-ceramide induced activations of caspase-3/CPP32 and caspase-7/Mch3 followed by PARP cleavage, effects that can be blocked either by SPP or TPA. Furthermore, both SPP and TPA inhibit the activation of caspase-6/Mch2 and subsequent lamin B cleavage. Ceramide, in contrast to Fas ligation, did not induce activation of caspase-8/FLICE and neither SPP nor TPA were able to prevent this activation. Thus, SPP, likely generated via protein kinase C-mediated activation of sphingosine kinase, suppresses the apoptotic pathway downstream of FLICE but upstream of the executioner caspases, caspase-3, -6, and -7.

Physiologic cell death occurs through an evolutionary conserved suicide process, termed apoptosis, which plays a considerable role in early development and homeostasis of adult tissues (1). Ceramide has recently emerged as a critical component of apoptosis (2,3). A variety of stress stimuli, such as tumor necrosis factor ␣ (TNF␣), 1 Fas ligand, growth factor withdrawal, anticancer drugs, oxidative stress, heat shock, ionizing radiation, and ultraviolet light increase cellular ceramide which, in turn, is capable of inducing apoptosis (3,4). This apoptotic effect is blocked by addition of the phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) or diacylglycerol, both activators of protein kinase C (PKC) (2,5), suggesting that PKC activation counteracts ceramide-mediated apoptosis. Activation of PKC in diverse cell types stimulates sphingosine kinase activity resulting in intracellular accumulation of sphingosine 1-phosphate (SPP) (6,7). Recently, we showed that SPP prevented the hallmarks of apoptosis resulting from elevated levels of ceramide induced by TNF␣, anti-Fas antibody, sphingomyelinase, or cell-permeable ceramide (7). Furthermore, inhibition of PKC, as well as inhibition of sphingosine kinase, induces apoptosis, which can be overcome by the addition of SPP (7). These results indicate that PKC may inhibit ceramide-induced apoptosis by activating sphingosine kinase.
Previous studies have identified a number of substrates for the caspases, particularly the DNA repair enzyme poly(ADPribose) polymerase (PARP) (10) and the nuclear lamins (11). Proteolysis of these substrates may account for many of the biochemical and morphological nuclear changes associated with apoptosis. The early cleavage of PARP can be catalyzed by several caspases belonging to the CED-3 subfamily (10,(12)(13)(14). Of these, caspase-3/CPP32 is thought to be the most efficient PARP protease (10,12), followed by caspase-7/Mch3 (13). The degradation of nuclear lamins is required for packaging of the condensed chromatin into apoptotic bodies, a typical morphology observed during the late stages of apoptosis (11). Caspase-6/Mch2 is the only known caspase capable of cleaving lamins (14,15).
Although the relative intracellular levels of ceramide and SPP have been proposed to be a critical gauge of cell fate (7), the molecular mechanism of actions of these sphingolipid me-* This work was supported in part by National Institutes of Health Grants GM43880 and CA61774 and by American Cancer Society Grant DE-275 (to S. S.). 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 1 The abbreviations used are: TNF␣, tumor necrosis factor ␣; Ac-DEVD-AMC, acetyl-Asp-Glu-Val-Asp-aminomethylcoumarin; DEVD, Asp-Glu-Val-Asp; ICE, interleukin1␤-converting enzyme; PARP, enzyme poly(ADP-ribose) polymerase; PKC, protein kinase C; SPP, sphingosine 1-phosphate; TPA, 12-O-tetradecanoylphorbol-13-acetate; PBS, phsophate-buffered saline; JNK, c-Jun N-terminal kinase; SAPK, stress-activated protein kinase; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid. tabolites are not well understood. Recently, ceramide has been implicated in PARP cleavage (16,17) and caspase-3/CPP32 activation (18). In addition, it appears that ceramide can be generated via a CrmA-inhibitable caspase, most likely caspase-8/FLICE, since overexpression of CrmA blocks both cell death and ceramide accumulation induced by TNF␣, while exogenous ceramide can bypass this sensitive step and induce apoptosis (16).
In this study, the relationship between sphingolipid metabolites and the caspases was examined. We demonstrated that both SPP and TPA decrease PARP cleavage by inhibiting activation of caspase-3/CPP32 and caspase-7/Mch3 induced by Fas ligation or cell-permeable ceramide. Furthermore, we also report that SPP and TPA inhibit activation of caspase-6/Mch2 and subsequent lamin B cleavage. Finally, ceramide, in contrast to Fas ligation, did not induce activation of caspase-8/ FLICE and neither SPP nor TPA were able to prevent this activation.
Cell Culture-Jurkat T cells (clone E6 -1, ATCC, Rockville, MD) were grown in RPMI 1640 supplemented with 10% fetal bovine serum. On the day of the experiment, cells were washed twice in RPMI 1640 containing 5 g/ml transferrin and 5 g/ml insulin in place of serum and resuspended in this serum-free medium (0.75-1 ϫ 10 6 cells/ml).
DNA Fragmentation Assay and Staining of Apoptotic Nuclei-DNA fragmentation was measured in cells prelabeled with [methyl-3 H]thymidine (0.5 Ci/ml) for 16 h (19). Cells were washed twice, resuspended in serum-free medium, plated in 12-well plates (3 ϫ 10 6 cells/well), and then exposed to the indicated agents. After 3 h, cells were lysed in TTE (10 mM Tris (pH 7.4), 10 mM EDTA, 0.2% (v/v) Triton X-100) and incubated for 15 min on ice. Fragmented DNA and intact chromatin were separated by centrifugation at 12,000 ϫ g for 10 min. Pellets were resuspended in TTE, and trichloroacetic acid was added to a final concentration of 12.5% (v/v). DNA fragmentation was calculated as follows: percent DNA fragmentation ϭ (fragmented/(fragmented ϩ intact chromatin)) ϫ 100. All results were determined in triplicate and expressed as means Ϯ standard deviations.
Apoptosis was also assessed by staining cells with bisbenzimide trihydrochloride (8 g/ml in 30% glycerol/PBS; Hoechst 33258, Calbiochem, San Diego, CA) for 10 min. Cells were then examined with a Zeiss Photoscope II fluorescent microscope (Petersburg, VA).
Western Blotting-Cytosolic proteins (20 g) were denatured by boiling in Laemmli buffer (Bio-Rad) for 3 min, then separated on 15% SDS-polyacrylamide gels, and blotted to nitrocellulose. Nonspecific binding was blocked by incubation of the membranes with 0.5% (v/v) Tween/PBS containing 5% (w/v) nonfat dry milk for 2 h at room temperature. Membranes were incubated for 2 h with rabbit antisera specific for the p17 subunit of caspase-3/CPP32 (gift of Dr. Donald Nicholson), goat antisera specific for the p21 subunit of caspase-6/Mch2 (Santa Cruz Biotechnology, Santa Cruz, CA), mouse monoclonal antilamin B (Calbiochem, San Diego, CA), rabbit antisera specific for the p20 subunit of caspase-7/Mch3, or rabbit antisera specific for the p20 subunit of caspase-8/FLICE (gift of Dr. Edward Gelmann). After washing with 0.5% (v/v) Tween/PBS, membranes were incubated for 2 h with horseradish peroxidase-conjugated secondary antibody. After three additional washes with 0.5% (v/v) Tween/PBS, SuperSignal enhanced chemiluminescent substrate (Pierce, Rockford, IL) was used to detect the proteins. Films were also scanned and analyzed by densitometry.
Fluorogenic DEVD Cleavage Enzyme Assays-Enzyme reactions were performed in 96-well plates and contained 20 g of cytosolic proteins in 100 l of Buffer A, diluted with 100 l of fresh Buffer A containing 40 M Ac-DEVD-AMC substrate. Fluorescent aminomethyl coumarin (AMC) product formation was measured over a 30-min period at an excitation wavelength of 360 nm and emission at 460 nm using a Cytofluor II fluorometer plate reader (PerSeptive Biosystems, Framingham, MA). Serial dilutions of AMC (Aldrich) were used as standards.

Sphingosine 1-Phosphate and TPA Inhibit
Fas-and C 2 -Ceramide-mediated Apoptosis-As described previously (7,18,20), Jurkat cells treated with Fas antibody underwent extensive cell death within 3 h as measured by the quantitative DNA fragmentation assay (Fig. 1A). Blebbing of cell membrane was observed and cells were fragmented into characteristic condensed nuclei and apoptotic bodies (Fig. 1C). In agreement with our previous study (7), when cells were treated simultaneously with TPA, DNA fragmentation was completely inhibited, while co-treatment with SPP reversed apoptosis by ϳ50% (Fig. 1, A, D, and E). Ceramide has been proposed to play a role in the Fas signaling pathway (20,21). Treatment of Jurkat T cells with the cell-permeable C 2 -ceramide (10 M) resulted in cell death that was overcome either by co-treatment with TPA or to a lesser extent by SPP (Fig. 1F). Stimulation of sphingosine kinase leading to intracellular accumulation of SPP has been shown to be one of the effects triggered by activation of PKC (6,7). Moreover, activation of PKC by phorbol esters limits ceramide production (7,22). In agreement, the PKC inhibitors, staurosporine, chelerythrine chloride, and calphostin C, induce sphingomyelin hydrolysis to generate ceramide (23,24). These studies suggest that PKC activation may oppose the Fas apoptotic pathway, both by activation of sphingosine kinase and by reducing the levels of ceramide.
Sphingosine 1-Phosphate and TPA Inhibit Activation of PARP Cleavage Induced by Fas Ligation and Ceramide-Because Fas ligation leads to proteolytic cleavage of PARP (12,(25)(26)(27), it was of interest to determine whether SPP and TPA could inhibit this cleavage. In agreement with previous results (25)(26)(27), extracts prepared from cells exposed to anti-Fas for 3 h were able to cleave in vitro translated [ 35 S]PARP substrate into the apoptotic 89-kDa fragment ( Fig. 2A). Densitometric analysis revealed that almost 70% of the 116-kDa full-length PARP was converted to the 89-kDa fragment, while only 38% was cleaved by extracts from control cells. SPP treatment reduced cleavage of PARP induced by Fas ligation by over 50% and as expected, treatment with TPA completely prevented Fas-induced PARP proteolysis. Ceramide generation has recently been shown to precede PARP cleavage in TNF␣-mediated apoptosis (16). In addition, treatment of Molt-4 or MCF-7 cells with exogenous C 6 -ceramide results in cleavage of PARP (16,17), suggesting that ceramide activates a caspase responsible for PARP cleavage. In accordance with these studies (16, 17), we found that C 2 -ceramide also induced PARP cleavage (Fig.  2B) and co-treatment with TPA fully prevented this cleavage (Fig. 2B), while SPP co-treatment inhibited PARP proteolysis by 60 -90% (Fig. 2B). These results clearly indicate that SPP can inhibit PARP cleavage induced either by Fas ligation or exogenous ceramide.
Sphingosine 1-Phosphate and TPA Prevent Activation of Caspase-3/CPP32 and Caspase-7/Mch3-Caspase-3/CPP32 is known as the most efficient PARP-cleaving caspase, with a K m of ϳ10 M (10, 12). Caspase-7/Mch3, which shows the highest homology to human caspase-3/CPP32, can also cleave PARP, albeit at a slightly lower efficiency (13). We used the fluorogenic substrate, Ac-DEVD-AMC, which corresponds to the motif that is cleaved in PARP (10), to measure the activity of these caspases. Extracts from cells exposed to anti-Fas showed a significant increase in caspase-3-like proteolytic activity (Fig.   3A), as described previously (18,27). In contrast, co-treatment with TPA completely blocked this activity, while SPP co-treatment markedly reduced Fas-induced caspase activity (Fig. 3A). Proteolytic processing of pro-caspase-3/CPP32 in response to Fas was also examined by Western blotting using antisera specific for the active p17 subunit. This caspase is synthesized as a 32-kDa precursor (p32) that is cleaved to generate the mature form composed of 17-kDa (p17) subunit, through an intermediary 20-kDa (p20) form, and 12-kDa (p12) subunit (10,12,28). In agreement with previous studies (25,27), following Fas activation, the p32 precursor was cleaved into active caspase-3 (Fig. 3B). Caspase-3 maturation was completely blocked by TPA, as active subunit was not detected in cells treated with both anti-Fas and TPA. SPP also prevented Fasmediated caspase-3/CPP32 processing by over 60% as measured by densitometry (Fig. 3B).
A similar approach was used to assess the protective effects of SPP and TPA on exogenous C 2 -ceramide-mediated caspase-3/CPP32 activation. In accordance with a previous study (18), extracts from cells treated with C 2 -ceramide displayed an increase in caspase-3-like activity (Fig. 4A), corresponding with the appearance of the active p20 and p17 subunits of caspase-3 (Fig. 4B). TPA effectively blocked both the caspase-3-like activity increase (Fig. 4A) and the processing of caspase-3/CPP32 into p20 and p17, triggered by ceramide (Fig. 4B). Cytosolic extracts from cells co-treated with SPP displayed significantly decreased CPP32-like activity (Fig. 4A). Western blotting also demonstrated the same strong inhibition (60 -90%) of p20 and p17 expression (Fig. 4B).
Caspase-7/Mch3, which is also capable of cleaving PARP (13), is expressed as a 35-kDa precursor, and upon activation is processed into 20-kDa (p20) and 12-kDa (p12) subunits (25,29). Fas ligation leads to the generation of the p20 and p12 subunits (25,29). Thus, it was of interest to examine whether ceramide was also capable of inducing activation of Mch3 and if SPP could affect this processing. As expected, Fas induced the ap- pearance of the activated p20 form, which was completely abolished by pretreatment with TPA and to a lesser extent with SPP (Fig. 5A). Similarly, processing of Mch3 to the p20 form induced by ceramide was markedly decreased by co-treatment with TPA, and SPP diminished C 2 -ceramide-induced Mch3 processing by over 50%. Thus, SPP is able to counteract proteolytic cleavage of caspase-3 and caspase-7, and subsequent cleavage of PARP triggered by Fas or exogenous ceramide.

SPP and TPA inhibit FAS-and Ceramide-mediated Caspase-6/Mch2 Activation and Subsequent Lamin B Degradation-In
addition to the breakdown of the nuclear enzyme PARP, another event that is common to apoptosis is the cleavage of lamins, which play major roles in nuclear envelope integrity (11). To date, caspase-6/Mch2 is the only known laminase (14,15). Engagement of Fas in Jurkat T cells has been shown to activate caspase-6 (14) and consequently trigger the cleavage of lamin B (26). Thus, we determined whether ceramide treatment could also induce caspase-6 activation and lamin B cleavage. Lamin B is cleaved into a characteristic 28-kDa fragment after treatment with Fas or exogenous C 2 -ceramide (Fig. 6, A  and B). Western blotting of extracts from cells co-treated with TPA and Fas antibody (Fig. 6A) or C 2 -ceramide (Fig. 6B) displayed only intact lamin B, indicative of inhibition of cleavage. When cells were treated with SPP during Fas- (Fig. 6A) or C 2 -ceramide-mediated apoptosis (Fig. 6B), lamin B proteolysis was markedly reduced. We next determined if caspase-6 was protected by SPP and TPA from activation caused by Fas-and C 2 -ceramide treatment. In accordance with previous studies (14), Fas antibody was able to trigger caspase-6 cleavage, generating 21-, 18-, and 14-kDa fragments (Fig. 7A). Treatment with C 2 -ceramide similarly led to processing of caspase-6 ( Fig.  7B). When cells were co-treated with TPA or SPP, expression of active forms of caspase-6, especially p18 and p14, was strongly reduced (Fig. 7, A and B). Therefore, SPP, like TPA, can attenuate proteolytic cleavage of caspase-6 and its target, nuclear lamins, during Fas-or C 2 -ceramide-induced apoptosis.
Fas-induced Caspase-8/FLICE Activation Is Not Inhibited by SPP or TPA-Apoptosis and ceramide accumulation induced by TNF-␣ or Fas are completely inhibited by CrmA, a product of the cowpox virus, while exogenous ceramide is able to bypass this block and induce apoptosis by activating downstream caspases (16). CrmA is a potent inhibitor of ICE (30). However, the importance of caspase-1/ICE itself in Fas-mediated apoptosis is controversial (Refs. 31 and 32 versus Refs. 27 and 33). Instead, caspase-8/FLICE, the most upstream caspase implicated in Fas-and TNF␣-mediated apoptosis, appears to be the target of CrmA in vivo (34). Thus, it was of interest to determine the involvement of sphingolipid metabolites in activation of caspase-8. Fas-ligation induced the appearance of a p20 band indicative of cleavage and corresponding to the mature form of FLICE (Fig. 8). In contrast, C 2 -ceramide treatment did not induce cleavage of caspase-8 (Fig. 8). This is in agreement with the concept that ceramide acts downstream of a CrmA-inhibitable caspase. Finally, neither SPP nor TPA have an inhibitory effect on Fas-triggered caspase-8/FLICE cleavage (Fig. 8). DISCUSSION Within the last few years, new studies have been reported, which have enhanced our understanding of how cell surface events are communicated to the cell suicide machinery (21,35,36). Protein-protein interactions link death domains of cell surface receptors for TNF-␣ (TNF receptor-1) or Fas ligand (CD95/Fas/APO-1) to a cascade of ICE/CED-3-homologous pro-  teases. Binding of the adaptor protein FADD, which also contains a death effector domain, either directly to CD95/Fas/ APO-1 or to TNF receptor-1 via another death domaincontaining protein TRADD (35), recruits caspase-8 to the plasma membrane, resulting in autocatalytic activation of the pro-apoptotic proteases (36). Among the caspases identified so far, caspase-3, the closest relative to C. elegans CED-3, is the pivotal protease involved in Fas-induced apoptosis (27,33). Caspase-3 not only cleaves the nuclear protein PARP (10,12), it is also capable of downstream activation of both laminase caspase-6 (37) and caspase-7 (13) (Fig. 9). Once activated, caspase-7 and caspase-6 can process their distinctive targets, PARP and lamins, respectively. Potential sites for interactions between sphingolipid metabolites and the caspase cascade were investigated in this study.
Although our understanding is not complete, several lines of evidence suggest that ceramide plays an important role in apoptosis induced by cytokines, chemotherapeutic agents, serum withdrawal, or ionizing radiation, and does not arise merely as a consequence of activation of the death machinery (reviewed in Ref. 3). First, overexpression of the proto-oncogene Bcl-2, a CED-9 homolog, or treatment with SPP, both of which prevent apoptosis, did not interfere with ceramide formation induced by various insults (7,16,38), yet did protect against apoptosis induced by ceramide (7,16,17,38,39). Second, ceramide production and apoptosis induced by Fas ligand or TNF-␣ can be blocked by overexpression of the viral serpin CrmA (16), whose in vivo target is most likely caspase-8 (34). In agreement, we show here that ceramide does not induce activation of caspase-8 (Fig. 8). Third, the cell-permeable analog C 2 -or C 6 -ceramide induce activation of caspase-3 (18, 40) (Fig.  4) and subsequent PARP cleavage (17) (Fig. 2). Furthermore, our study demonstrates that exogenous C 2 -ceramide also acti-vates PARP protease caspase-7 (Fig. 5), the closest homolog to caspase-3, acting through caspase-3 itself (13) or independently. Finally, we describe here for the first time that C 2ceramide induces lamin B cleavage (Fig. 6), likely mediated by caspase-6 activation (Fig. 7). The mechanism by which ceramide stimulates proteolytic processing of caspase-6 is still unclear, but it is tempting to speculate that this may also be mediated via caspase-3 as proposed by Alnemri and co-workers (41). Taken together, our observations highlight the relationship between ceramide and the caspase cascade, and demonstrate, in agreement with other studies (16), that ceramide can be placed downstream of the initiator caspase-8 and upstream of the executioner CPP32-like proteases (Fig. 9).
Although SPP was originally shown to rescue cells from ceramide-mediated cell death (7), it was not clear how it was linked to the deadly caspase cascade. Our results demonstrate that cleavage of the death substrates, PARP and lamins, induced by Fas antibody or exogenous ceramide, is inhibited by co-treatment with SPP. In addition, we found that activation of the caspases directly responsible for the breakdown of these substrates by Fas ligation or ceramide is also inhibited by SPP. Our results establish that SPP functions upstream of caspase-3/CPP32. Nevertheless, because the proximal molecular targets by which ceramide activates CPP32 and related CED-3 subfamily caspases remain largely unknown, further work is necessary to identify the targets for ceramide and SPP.
Several reports have provided evidence that some of the effects of SPP are mediated through cell surface receptors (3,4). However, the anti-apoptotic effects of SPP appear to be mediated via intracellular targets, as inhibition of sphingosine kinase by N,N-dimethylsphingosine not only eliminates formation of SPP induced by TPA, it also induces apoptosis and prevents the cytoprotective activity of TPA which can be restored by addition of SPP (7).
Recently, it has been suggested that the stress-activated protein kinase (SAPK/JNK) pathway is required for ceramidemediated apoptosis because overexpression of dominant-negative constituents of the JNK pathway abrogates ceramide-mediated apoptosis (42). It has also been well documented that Fas ligation induces SAPK/JNK activation in Jurkat T cells (43)(44)(45). Moreover, we have shown previously that SPP not only stimulates the extracellular signal-regulated kinases Erk1 and Erk2, it also prevents SAPK/JNK activation by ceramide and consequent apoptosis (7). Thus, we have proposed that the formation of distinct sphingolipid metabolites and consequent regulation of different family members of the mitogen-activated protein kinase family is an important factor determining the fate of cells (7). However, the relationships between activation of the caspase cascade and the mitogenactivated protein kinase signaling pathway leading to apoptosis have not yet been clarified. Recent evidence supports a parallel pathway model leading to apoptosis, in which the Fas-binding proteins, FADD and Daxx, activate the SAPK/JNK pathway and the caspase cascade independently (46). It is very intriguing that the sphingolipid metabolites, ceramide and SPP, have opposing effects on these two important apoptotic pathways downstream of Fas. Identification of the proximal intracellular targets of these sphingolipids should provide useful clues to the regulation cell survival.