Sphingosine 1-phosphate protects human umbilical vein endothelial cells from serum-deprived apoptosis by nitric oxide production.

Sphingosine 1-phosphate (S1P) can prevent endothelial cell apoptosis. We investigated the molecular mechanisms and signaling pathways by which S1P protects endothelial cells from serum deprivation-induced apoptosis. We show here that human umbilical vein endothelial cells (HUVECs) undergo apoptosis associated with increased DEVDase activity, caspase-3 activation, cytochrome c release, and DNA fragmentation after 24 h of serum deprivation. These apoptotic markers were suppressed by the addition of S1P, the NO donor S-nitroso-N-acetylpenicillamine (100 micrometer), or caspase-3 inhibitor z-VAD-fmk. The protective effects of S1P were reversed by the nitric-oxide synthase (NOS) inhibitor N-monomethyl-l-arginine, but not by the soluble guanylyl cyclase inhibitor 1H-(1,2,4)oxadiazolo[4,3-a]-quanoxaline-1-one, suggesting that NO, but not cGMP, is responsible for S1P protection from apoptosis. Furthermore, S1P increased NO production by enhancing Ca(2+)-sensitive NOS activity without changes in the eNOS protein level. S1P-mediated cell survival and NO production were suppressed significantly by pretreatment with antisense oligonucleotide of EDG-1 and partially by EDG-3 antisense. S1P-mediated NO production was suppressed by the addition of pertussis toxin, an inhibitor of G(i) proteins, the specific inhibitor of phospholipase C (PLC), and the Ca(2+) chelator BAPTA-AM. These findings indicate that S1P protects HUVECs from apoptosis through the activation of eNOS activity mainly through an EDG-1 and -3/G(i)/PLC/Ca(2+) signaling pathway.

Sphingosine 1-phosphate (S1P), 1 generated from sphingo-sine by sphingosine kinase, is a recently described lipid mediator regulating cellular responses, including cell proliferation (1), cell motility (2,3), and morphological changes (2,4) in endothelial cells. In addition, S1P was also demonstrated to protect endothelial cells from apoptosis induced by C 2 -ceramide, TNF␣, and anti-Fas antibody (1,4,5). All these effects of S1P on endothelial functions are thought to correlate with the intracellular signaling pathway linked to the EDG-1 (endothelial differentiation gene-1), EDG-3, and EDG-5 subtypes of G protein-coupled receptors (1)(2)(3)(4)6). Of these receptors, EDG-1 is exclusively expressed in HUVECs (2,4), and S1P induces a robust calcium response mainly via the G i -coupled S1P receptors EDG-1 and -3 in HUVECs (4). Consistent with this view, we have recently observed that intracellular Ca 2ϩ mobilization evoked by the G i -PLC system is responsible for the signaling pathway of mitogen-activated protein kinases, focal adhesion kinase (p125 FAK ), and chemotaxis in response to S1P in HUVECs (6). It is likely that intracellular signaling events of S1P in endothelial cells may be directly associated with increase in [Ca 2ϩ ] i .
Endogenous nitric oxide (NO) is synthesized from L-arginine by catalytic reaction of three isotypes of NO synthases (NOS), the neuronal or type I isoform (nNOS), the inducible or type II isoform (iNOS), and the endothelial or type III isoform (eNOS) (7,8). Endothelium-derived NO is formed by eNOS, which is constitutively expressed in endothelial cells and localized in plasmalemmal caveolae through association with the caveolae integral membrane structural protein caveolin-1 (9). Binding of eNOS with caveolin-1 inhibits the eNOS catalytic activity, and the inhibitory effect of caveolin-1 on the eNOS activity can be reversed by Ca 2ϩ -calmodulin (10). Thus, increase in [Ca 2ϩ ] i elicited by diverse extracellular signals (stimuli), including shear stress (11), bradykinin (12), and Na ϩ /Ca 2ϩ exchange (12,13), leads to activation of eNOS (8), resulting in increased NO production.
Numerous previous studies have demonstrated the role of eNOS in controlling blood pressure, vascular remodeling, angiogenesis, and apoptosis (14 -17). In particular, NO is implicated as a cytoprotective effector molecule, protecting some cell types from apoptotic cell death induced by TNF␣, anti-Fas antibody, LPS, and trophic factor withdrawal. These apogenic stimuli activate a series of tightly controlled intracellular sig- ). 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.
naling events that induced mitochondrial cytochrome c release into cytosol and the activation of cysteine proteases known as caspases (18). Inappropriate endothelial cell apoptosis may be linked to several cardiovascular diseases. For example, eNOSdeficient mice exhibit delayed angiogenesis thought to be due to decreases in endothelial cell migration, proliferation, and differentiation, compared with wild type controls (19). In addition, we have shown that transduction with NOS gene into aortic allografts suppresses the development of allograft arteriosclerosis (20), possibly in part via suppression of endothelial apoptosis (21). The antiapoptotic effects of NO are associated with direct inhibition of caspase activity by S-nitrosylation of the catalytic cysteine residue of the enzymes, as well as through cGMP-dependent mechanism (22). Therefore, eNOSactivating agents are capable of protecting endothelial cells from apoptosis. We hypothesized that S1P would protect endothelial cells by Ca 2ϩ -dependent activation of eNOS. Here we show that S1P increases NO production in HUVECs by elevating [Ca 2ϩ ] i through the EDG-1 and -3/G i -PLC pathway. This increased NO production prevents serum-deprived apoptosis by inhibiting mitochondrial cytochrome c and release caspase-3 activation/activity.
Cell Culture-HUVECs were isolated as described previously (6). The cells were grown onto a gelatin-coated 75-cm 2 flask in M199 with 20% fetal bovine serum, 100 units/ml penicillin, 100 mg/ml streptomycin, 3 ng/ml basic fibroblast growth factor, and 5 units/ml heparin at 37°C under 5% CO 2 , 95% air. To induce apoptosis, serum-free medium without basic fibroblast growth factor was used. The cells used in this study were between passages 2 and 7.
Nitrite and Nitrate Measurement-Production of nitrite plus nitrate (NOx) was measured by ozone-chemiluminescence method. Culture media from HUVECs were collected and assayed for NOx in the chemiluminescent NO analyzer NO 2000 analyzer (Ki-Woo Biotech, Seoul, Korea) and quantified with sodium nitrate as a standard.
Cell Viability-HUVECs (1 ϫ 10 5 cells/well) were plated onto six-well plates in 1 ml of M199 containing 20% fetal bovine serum. The next day, the cells were switched to serum-free M199 with or without 20% fetal bovine serum and S1P. After 24 h, cell viability was assessed by trypan blue exclusion. Counts were performed on triplicate wells.
Western Blot Analysis-Cell pellets were suspended in ice-cold sterilized water and kept on ice for 1 min. The suspension was mixed with an equal volume of buffer A (20 mM HEPES, pH 7.5, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EGTA, 1 mM EDTA, 0.5 mM phenylmethanesulfonyl fluoride, 5 g/ml aprotinin, 5 g/ml pepstatin, and 10 g/ml leupeptin) containing 500 mM sucrose solution and carefully homogenized in a Dounce tissue grinder with a loose pestle. Cytosols were obtained for measuring cytochrome c release by centrifugation at 100,000 ϫ g for 1 h. For Western blot of caspase-3, cells (2 ϫ 10 5 cells) were suspended in 100 mM Tris-HCl buffer (pH 7.4) and lysed by three freeze-thaw cycles. Cytosolic proteins were obtained by centrifugation at 12,000 ϫ g for 20 min at 4°C. Proteins (40 g) were separated on 14% SDS-PAGE for caspase-3 and then transferred to nitrocellulose. The membranes were hybridized with antibodies against caspase-3 and cytochrome c, and protein bands were visualized by exposing to x-ray film, as described previously (23).
Assay for DEVDase Activity-The cell pellets were washed with ice-cold phosphate-buffered saline and resuspended in 100 mM HEPES buffer, pH 7.4, containing protease inhibitors (5 g/ml aprotinin and pepstatin, 10 g/ml leupeptin, and 0.5 mM phenylmethanesulfonyl fluoride). The cell suspension was lysed by three freeze-thaw cycles, and the cytosolic fraction was obtained by centrifugation at 12,000 ϫ g for 20 min at 4°C. DEVDase activity was assayed in the presence or absence of 20 mM DTT by measuring the increased absorbency at 405 nm after cleavage of 150 M Ac-DEVD-pNA (22).
In Vitro Cleavage of PARP-[ 35 S]Methionine-labeled PARP was synthesized using a transcription/translation-coupled transcription and translation system (Promega). Aliquots (4 l) of in vitro translated 35 S-labeled PARP were incubated with 4 g of cytosolic protein in 10 l of the total reaction volume at 37°C for 1 h. The reaction was stopped by mixing with an equal volume of 2 ϫ SDS sample buffer and heating the mixture for 2 min. Cleavage profiles of PARP were examined by electrophoresis on 10% SDS-PAGE and protein visualized by fluorography.
Determination of NOS Activity-Cells were cultured in 60-mm culture dishes with S1P for different time periods, washed twice, and equilibrated for 20 min with HEPES buffer, pH 7.4, containing: 10 mM HEPES, 145 mM NaCl, 5 mM KCl, 1 mM MgSO 4 , 1 mM CaCl 2 , and 1 mM glucose. Cells were incubated for 40 min with 1 Ci of L-[ 3 H]arginine plus 8 M L-arginine. To assess the calcium dependence of NOS activation, experiments were performed in calcium-free HEPES buffer containing 20 M BAPTA-AM. The reaction mixture was removed, and ice-cold ethanol (0.5 ml) was added. After allowing for evaporation of ethanol, 2 ml of 10 mM HEPES-Na, pH 5.5, was added and kept for 20 min. The supernatant was collected and applied onto a 2-ml cationexchange column (50W-X8 resin (converted to Na ϩ form), Bio-Rad).
[ 3 H]Citrulline product was collected in the eluate by washing the column with 4 ml of H 2 O and then quantified by scintillation counting. Enzyme activity was normalized to protein concentration. NOS activity was expressed as pmol/mg protein.
DNA Fragmentation-DNA fragmentation was demonstrated by harvesting total cellular DNA and agarose gel electrophoresis. After the indicated treatment, both adherent and detached cells were harvested, washed with phosphate-buffered saline, and lysed in 50 mM Tris-EDTA, pH 7.4, 1% SDS, and 0.5 mg/ml proteinase K for 3 h at 50°C. Samples were then extracted with phenol/chloroform and precipitated with ethanol. The pellet was resuspended in 20 mM Tris-EDTA, pH 8.0. After digesting RNA with RNase (0.1 mg/ml) at 37°C for 1 h, DNA was separated by electrophoresis on a 1.2% agarose gel stained with ethidium bromide.
Other Analysis-Protein concentration was determined with the BCA assay (Pierce). Data are presented as mean Ϯ S.D. of at least three separate experiments, except where results of blots are shown in which case a representative experiment is depicted in the figures. Comparisons between two values analyzed using Student's t test. Differences were considered significant when p Ͻ 0.05.

S1P Protects HUVECs from Serum-deprived Apoptosis-Pre-
vious studies showed that trophic factor deprivation induces apoptotic cell death in several cell types, including endothelial cells (24), and that S1P protects endothelial cells from C 2ceramide-and TNF␣-induced apoptotic cell death (25,26). To investigate whether S1P protects HUVECs from serum deprivation-induced cell death, HUVECs were cultured in serumfree M199 medium with or without S1P, and cell viability was assayed by trypan blue exclusion after 24 h. Serum deprivation decreased cell viability to ϳ30% of control cells grown in the presence of serum. The addition of S1P (0 -5 M) reversed the serum-deprived effects associated with cell death in a dose-dependent manner (Fig. 1). Cells can die by either necrosis or apoptosis, and the inhibition of caspase activation/activity can protect cells from apoptotic cell death, but not from necrotic cell death (27). HUVECs were cultured in serum-free medium with or without the caspase inhibitor z-VAD-fmk. z-VAD-fmk prevented serum-deprived cell death (Fig. 1), indicating that serum deprivation causes caspase-dependent apoptosis.
NMA Suppresses the Inhibitory Effects of S1P on Serum-deprived Apoptosis-It has been shown that NO production from eNOS or exogenous NO donor protects endothelial cells from apoptosis induced by TNF␣ (22) and LPS (21). To investigate the involvement of NO in the protective effect of S1P, HUVECs were treated with S1P in the presence of the specific inhibitor of NOS NMA, and cell viability was measured. The protective effect of S1P was significantly reduced by the addition of NMA ( Fig. 2A). DNA fragmentation typical of apoptosis was identified in serum-deprived HUVECs (Fig. 2B). S1P prevented DNA fragmentation, and the protective effect of S1P was blocked by the addition of NMA. These results suggest that the cytoprotective effect of S1P on serum-deprived apoptosis is mediated by NO production from HUVECs. Antisense Oligonucleotides of EDG-1 and -3 Suppress S1Pmediated Survival and NO Production-Exogenous S1P regulates cellular responses such as angiogenesis (2), morphogenesis (2, 4), cell migration (2, 3), and cell survival (28) through the activation of EDG-1 and -3 receptors or through the activation of sphingosine kinase, which synthesizes endogenous formation of S1P. We next examined the involvement of these signaling pathways in the S1P-mediated increase in cell survival and NO production by treating the HUVECs with antisense oligonucleotides for EDG-1 and -3 or the sphingosine kinase inhibitor dihydrosphingosine. Pretreatment of HUVECs with the antisense, but not sense, oligonucleotide to EDG-1 significantly suppressed S1P-mediated cell survival and S1P-induced NO production (Fig. 3, A and B). In contrast, pretreatment with antisense, but not sense, oligonucleotide to EDG-3 only partially suppressed the effects of S1P (Fig. 3, C and D). A combination of antisense oligonucleotides to EDG-1 and -3 suppressed the effects of S1P to about 90% (data not shown). Treatment with the sphingosine kinase inhibitor dihydrosphingosine did not inhibit S1P-mediated cell survival or NO production (data not shown). These results suggest that S1P requires both EDG-1 and -3 for full activation of the cellular signal needed to promote cell survival and NO production in HUVECs. S1P Inhibits Caspase Activity/Activation and Cytochrome c Release in a NO-dependent Manner-Since serum deprivation induces apoptosis by activating the caspase signal cascade and mitochondrial cytochrome c release (29), we examined the effect of S1P on caspase activation/activity and mitochondrial cytochrome c release in serum-deprived HUVECs. HUVECs were treated with or without S1P in serum-free medium, and DEVDase (caspase-3-like protease) activity was measured in the cytosol by a colorimetric assay using the tetrapeptide sub-strate Ac-DEVD-pNA and a PARP cleavage assay. DEVDase activity was significantly increased in the serum-deprived HUVECs compared with that of the untreated control cells (Fig. 4A). Treatment with S1P or the caspase inhibitor z-VADfmk blocked the increase in DEVDase activity. The effect of S1P was almost completely reversed by NMA. NO produced by endothelial cells and interleukin-1␤ plus interferon ␥-stimulated hepatocytes inhibits DEVDase activity by S-nitrosylation of the caspase enzymes. This inhibition can be reversed by the reducing agent DTT, which removed the NO from the active site cysteine (17,22). Preincubation of the cytosol with DTT partially reversed the suppressed DEVDase activity by S1P. PARP, a well known biosubstrate of caspase-3, was cleaved to an 85-kDa fragment by the cytosol from serum-deprived HUVECs, while cytosols from z-VAD-fmk and S1P-treated cells did not cleave PARP (Fig. 4B). This effect of S1P was also reversed by NMA.
To further determine whether the release of mitochondrial cytochrome c into cytosol and the activation of caspase-3 by serum deprivation are affected by treatment with S1P, we measured appearance of the cytosolic cytochrome c and the active fragment (p17) of caspase-3 by Western blot analyses. As shown in Fig. 4C, serum deprivation induced the release of cytochrome c to cytosol and the cleavage of pro-caspase-3 (p32) into a 17-kDa fragment. These events were suppressed by the addition of S1P or z-VAD-fmk. The S1P effects were reversed by NMA. However, cytochrome c oxidase as an indicator of mitochondrial contamination was not detected in the cell lysates. Studies in several cell-free systems of apoptosis have revealed that cytosolic cytochrome c interacts with Apaf-1 and procaspase-9 and forms an apoptosome in the presence of dATP. This results in the activation of caspase-9, which leads to activation of caspase-3 (18). Therefore, we next determined whether mitochondrial cytochrome c release precedes increases in DEVDase activity (Fig. 4D). Mitochondrial cytochrome c release into cytosol and DEVDase activity increased 12 and 16 h following serum withdrawal, respectively. Cytochrome c release was maximal at 16 h, and DEVDase activity peaked at 20 h. This is consistent with the view that cytochrome c release is upstream of caspase-3 activation in the apoptotic signal cascade of serum-deprived HUVECs. These results suggest that the antiapoptotic effect of S1P in serum-deprived HUVECs includes the suppression of cytochrome c release and the inhibition of caspase-3 activation/activity. NO-mediated Protective Effect of S1P-The antiapoptotic action of NO is mediated at least in part through cGMP in hepatocytes (22), PC12 cells (29), and cultured neuronal cells (31). To examine the role of NO/cGMP pathway in S1P-mediated protection, the effect of S1P on cell viability was examined in the presence of the specific inhibitor of soluble guanylate cyclase ODQ (Fig. 5A). ODQ did not inhibit the protective effect of S1P, while NMA did. Furthermore, treatment with an exogenous NO donor SNAP (100 M) protected cells from serumdeprived cell death, and this protection would not be blocked by ODQ. As expected a NO scavenger hemoglobin (400 M as heme concentration) prevent the protective effects of SNAP (Fig. 5B). In addition, the membrane-permeable cGMP analog 8-Br-cGMP did not inhibit serum-deprived apoptotic cell death. These results indicate that the protective effect of S1P is NOdependent, but cGMP-independent. S1P Increases NO Synthesis in HUVECs-If S1P protects via NO then it should be possible to measure increases in NO formation following S1P exposure. The formation of NOx as stable oxidized products of NO was measured in culture medium. S1P increased NO production in a dose-dependent manner reaching about 2-fold over controls at a concentration of 5 M S1P (Fig. 6A). The accumulation of NOx in the medium was detectable 4 h following S1P treatment (Fig. 6B). To examine mechanism possible for S1P-induced NO production, we examined the NOS activity in response to S1P. The NOS activity was increased to about 1.5-and 2-fold at 4 and 12 h, respectively, following treatment with S1P (Fig. 6C). The increased NOS activity was suppressed by addition of the Ca 2ϩ chelator BAPTA-AM. Western blot analysis revealed no significant difference of eNOS protein levels between 0 and 4 h (Fig. 6D). These data suggest that increased NO production by S1P is due to the activation of Ca 2ϩ -sensitive eNOS. S1P Induces NO via the G i Protein-PLC Signaling Pathway-To examine the signaling mechanism by which S1P stimulates NO production in HUVECs, the effects of various inhibitors on S1P signaling were evaluated (Fig. 7). We (2, 6) and others (4) have observed that S1P activates intracellular signaling cascade by a G i -coupled receptor-dependent manner. Treatment of HUVECs with pertussis toxin, a blocker of G i protein, significantly reduced the S1P-mediated NO production, suggesting that S1P mediated NO production in endothelial cells through G i -coupled endothelial cell surface EDG receptors (Fig. 7A). To examine the possible involvement of PLC and Ca 2ϩ in the downstream signaling pathway of G i -coupled receptor(s), HUVECs were cultured with S1P in the presence of the specific inhibitor of PLC U73122 and the Ca 2ϩ chelator BAPTA-AM. Treatment of HUVECs with U73122 suppressed NOx accumulation in culture medium by S1P, while U73343, which is a structurally similar derivative of U73122, failed to exert a similar effect (data not shown). The calcium chelator BAPTA-AM significantly blocked NOx accumulation by S1P. Furthermore, these signal blockers reversed S1P-mediated protection of HUVECs from serum-deprived apoptotic cell death (Fig. 7B). Together, these results indicate that S1Pmediated increase in NO production from HUVECs may be associated with increased [Ca 2ϩ ] i by EDG-1 and -3/G i protein-PLC signaling pathway. DISCUSSION This study was undertaken to determine the molecular mechanism by which S1P protects endothelial cells from apoptosis. We provide evidence that S1P protects HUVECs from serum-deprived apoptosis by increasing NO production. S1P increases NO production by Ca 2ϩ -sensitive eNOS activation without changing eNOS protein expression. NO production is associated with the inhibitory effect of S1P on apoptotic cell death, DEVDase activity, and mitochondrial cytochrome c release. Increases in cell survival and NO production by S1P were significantly suppressed by pretreatment with antisense oligonucleotides of EDG-1, and to a lesser extent, by EDG-3 antisense, indicating that the cellular effects of S1P involve the activation of EDG-1 and -3. The antiapoptotic effects of S1P could not be inhibited by ODQ, indicating that the protective effect does not involve cGMP production. This study also provides evidence that the increase in NO production by S1P occurred through the involvement of EDG-1 and -3, G i protein, PLC, and Ca 2ϩ . Thus, S1P increases eNOS activity via intracellular Ca 2ϩ mobilization, and the resulting increases in NO protect endothelial cells from apoptosis by suppression of apoptotic signaling cascade.
Apoptotic cell death by cytotoxic stimuli and serum or growth factor withdrawal is induced by tightly controlled intracellular signaling events, which require serial activation of caspase family proteases and mitochondrial cytochrome c release. Caspase-deficient animals or cells are resistant to apoptosis induced by anti-Fas antibody (32) and DNA-damaging agents (33). Treatment with caspase inhibitors reduces fulminant liver injury and animal mortality induced by administration of anti-Fas antibody (34) or LPS (35). These studies indicate that inhibition of caspase activation/activity is a potential approach for the prevention of apoptosis. Our data show that serum deprivation induced HUVEC apoptosis, as indicated by increased DNA fragmentation, DEVDase activity, caspase-3 activation, and cytochrome c release. Treatment with S1P and the caspase inhibitor z-VAD-fmk protected cells from serumdeprived apoptosis by suppressing these effects. These results raise the possibility that S1P is an endogenous protectant against endothelial cell apoptosis through a mechanism of suppressed cytochrome c release and caspase-mediated signaling.
In this study, our major question was how S1P prevents serum-deprived apoptotic cell death in HUVECs. We (21) and others (17) have previously shown that NO protects endothelial cells from LPS-and TNF␣-induced apoptosis. These studies directed our attention toward NO as a potential mediator of the S1P effect. Importantly, the present study revealed that S1P increased NO production from HUVECs by enhancing Ca 2ϩsensitive eNOS activity without significant increase in the eNOS protein. eNOS is normally located at the cell membrane within caveolae in an inactive state (36). Increases in intracellular Ca 2ϩ promote calmodulin binding to the enzyme, leading to transient eNOS activation and NO production. Other interaction and modification can alter eNOS activity, including binding to caveolin-1, which inhibits eNOS activity (37), and phosphorylation (38) and binding of heat shock protein 90 (39) that enhance NO production. Our studies suggest that the dominant action of S1P is to increase Ca 2ϩ levels; however influence on these other regulatory processes cannot be excluded.
NO has been shown to regulate either induction of apoptosis in some cells (40) or prevention of apoptosis in others (22,29). High levels of NO have been shown to be a cytotoxic through the inhibition of ATP synthetic enzymes, including mitochondrial aconitase and electron transfer complexes I and II (41) or through DNA damage (42). NO also suppresses cell proliferation through the inhibition of ribonucleotide reductase (43) and eukaryotic initiation factor-2␣ (44). NO can also prevent apoptosis in several cell types, including hepatocytes (22,45,46), human B lymphocytes (47), PC12 cells (29), splenocytes (48), eosinophils (49), ovarian follicles (50) teins such as heat shock protein 70 (23), heme oxygenase (53), or Bcl-X L (54). NO-mediated increases in cGMP levels account for the protective effects in some cells (22,29,49,50). In our studies, S1P did not increase many of proteins listed above, as measured by Western blot analysis (data not shown), suggesting that the antiapoptotic effect of S1P is unlikely linked to the expression of these cytoprotective genes. We also excluded cGMP as a mediator of the S1P-NO protective pathway using the soluble guanylate cyclase inhibitor ODQ. Another mechanism involves the directed inhibition of caspases by NO via S-nitrosylation of either the inactive (55) or active forms of the protease (17,22). NO suppresses the proteolytic activation and activity of multiple caspases in intact cells, including caspase-3 and caspase-8. This can prevent the downstream amplification of the signal by suppressing mitochondrial cytochrome c release or activation of other downstream caspases. Our data show that S1P suppressed mitochondrial cytochrome c release and caspase-3 activation/activity and that this suppression was reversed by NMA. Also, the inhibition of DEVDase activity was partially reversed by preincubating the cytosol with the reducing agent dithiothreitol, suggesting that NO produced by S1P inhibited in part DEVDase activity by redox-sensitive S-nitrosylation of the catalytic cysteine residue. This is consistent with the observation that S-nitrosylation of caspase-3 by NO inhibits TNF␣-induced apoptosis in endothelial cells (17). It is likely that S-nitrosylated caspases cannot proteolytically activate other caspases or induce mitochondrial cytochrome c release. NO can also activate other survival signals, including Akt/PKB (56), which prevents caspase activation and mitochondrial cytochrome c release by phosphorylation of Bad (57) and procaspase-9 (58) as well as activation of NF-B (59, 60). Our studies did not exclude a role for Akt. Others have shown that Akt activation can also protect endothelial cells through increasing eNOS activity (38).
We also addressed the question of how S1P increases NO production in HUVECs. Endothelial cells produce measurable levels of NO from eNOS at basal levels of [Ca 2ϩ ] i (50 -140 nmol/liter). Activity of eNOS can be enhanced by receptor-dependent stimuli such as bradykinin (12) and by receptor-independent stimuli such as Ca 2ϩ ionophore (62) and shear stress (63). In these cases, an acute increase in [Ca 2ϩ ] i has been observed before the elevation in NO production. Recent studies have shown that S1P induces a rapid increase in [Ca 2ϩ ] i in endothelial cells, CHO, and HL-60, mainly via the G i -coupled receptors (4,64,65). HUVECs express EDG-1 and -3, which transduce signals to increase [Ca 2ϩ ] i in response to Ͻ10 nM of S1P (4). Our results show that S1P increased cell survival and NO production in serum-deprived HUVECs at the concentration of M levels. We (2, 6) and others (1, 66, 67) have also shown that M levels of S1P regulate ERK activation, FAK phosphorylation, and apoptosis. Antisense oligonucleotides of EDG-1 suppressed much of the S1P-mediated cell survival and NO production, whereas EDG-3 antisense oligonucleotide partially inhibited the effects of S1P (Fig. 3). Furthermore, HUVECs were cotreated with S1P and the sphingosine kinase inhibitor dihydrosphingosine, and then the S1P-mediated cell survival and NO production were measured. Dihydrosphingosine did not inhibit the effects of S1P (data not shown), suggesting that exogenous S1P does not act through intracellular generation of S1P by activation of sphingosine kinase. Recently, it has been shown that S1P treatment targets its receptor EDG-1 to plasmalemmal caveolae and facilitates the activation of eNOS in COS cells cotransfected with EDG-1 and eNOS (67), suggesting the involvement of EDG-1 in NO production. Therefore, S1P-induced NO production is most likely to be occurred by the S1P receptors EDG-1 and-3 in HUVECs. Binding of eNOS with caveolin-1 inhibits the eNOS catalytic activity, and the inhibitory effect of caveolin-1 on the eNOS activity can be reversed by Ca 2ϩ -calmodulin (10). This evidence suggests that S1P may increase NO production by enhancing [Ca 2ϩ ] i through the activation of S1P receptor-mediated signaling cascade in HUVECs. Indeed, our data reveal that S1Pinduced NO production in HUVECs was significantly reduced by the treatment with the G i protein inhibitor pertussis toxin, indicating that S1P induced NO production in endothelial cells through the G i -coupled S1P receptors, EDG-1 and -3. G i protein dissociates into G␣ i and ␤␥ subunits upon receptor activation, and the released ␤␥ subunit is known to stimulate PLC (61). S1P-induced NO production was markedly suppressed by the specific inhibitor of PLC U73122, suggesting that a PLC pathway may be downstream mediator of the G i -coupled S1P receptors. PLC is known to activate intracellular Ca 2ϩ mobilization by generation of inositol triphosphate. The calcium chelator BAPTA-AM almost completely blocked NOx accumulation in culture media of HUVECs by S1P. These observations suggest that the increased NO production from HUVECs by S1P may be associated with the eNOS activation by the elevation of [Ca 2ϩ ] i through the EDG-1 and -3/G i -PLC pathway. However, the delayed, but sustained, increase in eNOS activity measured in our cells suggest that the up-regulation of eNOS activity is influenced not only by rapid increases in [Ca 2ϩ ] i levels, but also by other mechanisms such as Ca 2ϩ -independent phosphorylation of eNOS (30). FIG. 7. S1P-induced NO production is dependent on the G i protein-PLC signaling pathway. HUVECs in serumfree M199 were pretreated with or without pertussis toxin (PTX, 100 ng/ml), U73122 (U, 5 M), and BAPTA-AM (BATA, 5 M) for 1 h and then incubated with or without S1P (5 M) for additional 12 h for NOx measurement or 24 h for cell viability. NMA (1 mM) was cotreated with S1P. A, NOx was measured in the culture medium by chemiluminescence. B, cell viability was determined by trypan blue exclusion. Data represent mean Ϯ S.D. (n ϭ 4).