Ceramide-induced Intracellular Oxidant Formation, Iron Signaling, and Apoptosis in Endothelial Cells

Sphingolipid ceramide (N-acetylsphingosine), a bioactive second messenger lipid, was shown to activate reactive oxygen species (ROS), mitochondrial oxidative damage, and apoptosis in neuronal and vascular cells. The proapoptotic effects of tumor necrosis factor-α, hypoxia, and chemotherapeutic drugs were attributed to increased ceramide formation. Here we investigated the protective role of nitric oxide (·NO) during hydrogen peroxide (H2O2)-mediated transferrin receptor (TfR)-dependent iron signaling and apoptosis in C2-ceramide (C2-cer)-treated bovine aortic endothelial cells (BAECs). Addition of C2-cer (5–20 μm) to BAECs enhanced ·NO generation. However, at higher concentrations of C2-cer (≥20 μm), ·NO generation did not increase proportionately. C2-cer (20–50 μm) also resulted in H2O2-mediated dichlorodihydrofluorescein oxidation, reduced glutathione depletion, aconitase inactivation, TfR overexpression, TfR-dependent uptake of 55Fe, release of cytochrome c from mitochondria into cytosol, caspase-3 activation, and DNA fragmentation. Nw-Nitro-l-arginine methyl ester (l-NAME), a nonspecific inhibitor of nitricoxide synthases, augmented these effects in BAECs at much lower (i.e. nonapoptotic) concentrations of C2-cer. The 26 S proteasomal activity in BAECs was slightly elevated at lower concentrations of C2-cer (≤10 μm) but was greatly suppressed at higher concentrations (>10 μm). Intracellular scavengers of H2O2, cell-permeable iron chelators, anti-TfR receptor antibody, or mitochondria-targeted antioxidant greatly abrogated C2-cer- and/or l-NAME-induced oxidative damage, iron signaling, and apoptosis. We conclude that C2-cer-induced H2O2 and TfR-dependent iron signaling are responsible for its prooxidant and proapoptotic effects and that ·NO exerts an antioxidative and cytoprotective role.

Ceramide belongs to a group of naturally occurring sphingolipid second messenger molecules that is formed by sphingomyelinase-catalyzed hydrolysis of sphingomyelin (1)(2)(3). There is growing interest on the potential role of ceramide-mediated proapoptotic cell signaling in response to treatment with reac-tive oxygen species (ROS) 1 (e.g. superoxide and hydrogen peroxide) and other proapoptotic stress factors, including inflammatory cytokines such as tumor necrosis factor-␣ and lipopolysaccharide, hypoxia, and chemotherapeutic drugs (4 -6). Exogenous treatment of endothelial cells and neuronal cells with ceramide also caused oxidative stress and activation of caspase-3 leading to apoptosis (7)(8)(9)(10)(11)(12). Ceramide treatment has been shown to trigger both nitric oxide ( ⅐ NO) and superoxide generation in endothelial cells (13)(14)(15). The relative ratio between superoxide and ⅐ NO determine the ultimate cytotoxicity in ceramide-treated cells (14). Exposure of endothelial cells to lower concentrations of ceramide (ϳ5 M) causes an increase in ⅐ NO formation due to Ca 2ϩ activation and translocation of endothelial nitric-oxide synthase (eNOS) (14,15). At higher concentrations (Ͼ20 M) ceramide treatment induced ROS formation in cells (13,14).
Endothelial Cell Culture-Bovine aortic endothelial cells (BAECs) were obtained from the Clonetics Corp. Cells were obtained at the third passage, transferred to 75-cm 2 filter vent flasks (Costar, Cambridge, MA), and grown to confluence (5.2 ϫ 10 6 cells/75 cm 2 ) in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS), L-glutamine (4 mM), penicillin (100 units/ml), and streptomycin (100 g/ml), incubated at 37°C in a humidified atmosphere of 5% CO 2 and 95% air. Cells were passaged as described by Balla et al. (23) and used between passages 4 and 12. On the day of the treatment, the medium was replaced with DMEM containing 2% FBS, which contains ϳ25-30 g of transferrin/ml. The above experimental conditions were used in all the experiments performed in this study.
Measurement of Oxidative Stress-The level of intracellular oxidant production was estimated by oxidations of DHE and DCFH. Following treatment of BAECs with C 2 -ceramide, the medium was aspirated, and cells were washed with DPBS and incubated in 2 ml of fresh culture medium without FBS. DHE and DCFH-DA were added at a final concentration (10 M) and incubated for 20 min, respectively. The cells were then washed twice with DPBS and maintained in 1 ml of culture medium. Fluorescence was monitored using a Nikon fluorescence microscope equipped with rhodamine and FITC filters. The intensity values were calculated using Metamorph software.
Measurement of 55 Fe Uptake-55 Fe uptake into the cells was measured as described previously (16,24). Briefly, 0.2 Ci/ml 55 Fe (ferric chloride) was added to the medium for 0 -8 h, and its levels were measured as a function of time. Cells were washed twice with DPBS and lysed with PBS containing 0.1% Triton X-100, and the cell lysate was counted in a beta counter.
Western Blotting of TfR, PARP, Hsp-70, and Bcl-2-After the treatment with C 2 -ceramide, the cells were washed with ice-cold DPBS and resuspended in 150 l of radioimmune precipitation assay buffer (20 mM Tris-HCl, pH 7.4, 2.5 mM EDTA, 1% Triton X-100, 1% sodium deoxycholate, 1% SDS, 100 mM NaCl, 100 mM sodium fluoride) containing 1 mM sodium vanadate, 10 g/ml aprotinin, 10 g/ml leupeptin, and 10 g/ml pepstatin inhibitors. Cells were homogenized by passing the suspension through a 25-gauge needle (10 strokes). The lysate was centrifuged at 750 ϫ g for 10 min at 4°C to pellet out the nuclei. The remaining supernatant was centrifuged for 30 min at 12,000 ϫ g. Protein was determined by Lowry method, and 20 g of the lysate was used for the Western blot analysis. Proteins were resolved on SDSpolyacrylamide gels and blotted onto nitrocellulose membranes. Membranes were washed with TBS (140 mM NaCl, 50 mM Tris-HCl, pH 7.2) containing 0.1% Tween 20 (TBST) and 5% skim milk to block the nonspecific protein binding. Membranes were incubated with 1 g/ml mouse anti-human transferrin receptor monoclonal antibody (Zymed Laboratories Inc., San Francisco, CA), mouse anti-bovine poly(ADPribose) polymerase (PARP) monoclonal antibody (Zymed Laboratories Inc.), mouse anti-human Hsp-70 antibody (Zymed Laboratories Inc.) or hamster anti-human Bcl-2 monoclonal antibody (BD Pharmingen) in TBST for 2 h at room temperature, washed 5 times, and then incubated with goat anti-mouse IgG-horseradish peroxidase-conjugated secondary antibody for TfR, PARP, Hsp-70, and mouse anti-hamster (1:5,000) for Bcl-2 for 1.5 h at room temperature. The band was detected using the ECL method (Amersham Biosciences).
Mitochondrial Cytochrome c Release-The release of mitochondrial cytochrome c into the cytosol in C 2 -ceramide-treated BAECs was measured according to the methods as described previously (17,25). Briefly, BAECs were washed with DPBS and homogenized in PBS supplemented with 40 g/ml saponin. Lysate was centrifuged at 750 ϫ g for 10 min and followed by 12,000 ϫ g for 20 min. The supernatant was used as the cytosolic fraction to measure the released cytochrome c into the cytosol by Western blot analysis using a mouse anti-cytochrome c antibody (BD Pharmingen). Detection was by horseradish peroxidaseconjugated goat anti-mouse antibody using the ECL method.
Measurement of Apoptosis by TUNEL Assay-The terminal deoxynucleotidyl transferase-mediated nick-end labeling (TUNEL) assay was used for microscopic detection of apoptosis (26). This assay is based on labeling of 3Ј free hydroxyl ends of the fragmented DNA with fluorescein-dUTP catalyzed by terminal deoxynucleotidyl transferase. Procedures were followed according to a commercially available kit (ApoAlert) from Clontech. Apoptotic cells exhibit a strong nuclear green fluorescence that can be detected using a standard fluorescein filter (520 nm). All cells stained with propidium iodide exhibit a strong red cytoplasmic fluorescence at 620 nm. The areas of apoptotic cells were detected by fluorescence microscopy equipped with rhodamine and FITC filters. The quantification of apoptosis was performed using the Metamorph image analysis package.
Measurement of Intracellular ⅐ NO-Intracellular ⅐ NO levels were monitored using a DAF-2 fluorescence probe (27,28). The treated cells were washed with DPBS and incubated in 2 ml of fresh culture medium without FBS. DAF-2 was added at a final concentration of 10 M, and the cells were incubated for 20 min. The cells were washed twice with DPBS and maintained in 1 ml of culture medium for monitoring the fluorescence using a Nikon fluorescence microscope (excitation, 488 nm; emission, 610 nm) equipped with an FITC filter. The values of fluorescent intensity were calculated using the Metamorph software.
Measurement of Aconitase Activity-BAECs were washed twice with cold DPBS and lysed with buffer containing 0.2% Triton X-100, 100 M diethylenetriaminepentaacetic acid, and 5 mM citrate in PBS. The activity of aconitase was measured in 100 mM Tris-HCl (pH 8.0) containing 20 mM DL-trisodium isocitrate. An extinction coefficient for cisaconitate of 3.6 mM Ϫ1 at 240 nm was used (29).
Measurement of Glutathione-The level of glutathione (GSH) was measured by HPLC as the o-phthalaldehyde (OPA) adduct at pH 8.0 (30). BAECs were washed twice with DPBS, suspended in 250 l of PBS, and lysed by sonication. After centrifugation at 10,000 ϫ g for 2 min, 200 l of the clear supernatant was derivatized by incubation for 30 min at room temperature with OPA. An aliquot of sample was injected onto a Kromasil C-18 column and eluted isocratically with a mobile phase consisting of 150 mM sodium acetate:methanol (91.5:8.5, v/v). The OPA-GSH adduct was monitored using a fluorescence detector operating at excitation and emission wavelengths at 250 and 410 nm, respectively. The levels of intracellular GSH were quantified using a GSH solution as a standard.

Ceramide Induces Intracellular Superoxide and Hydrogen
Peroxide-Intracellular ROS levels were measured in BAECs treated with different concentrations of C 2 -cer for different time periods. The oxidation of DCFH, a nonfluorescent probe, to a fluorescent dichlorofluorescein (DCF) was used to measure intracellular H 2 O 2 -derived oxidants. Although H 2 O 2 itself does not react with DCFH to form DCF, it was proposed that intracellular peroxidases or redox-active metal ions could catalyze the oxidation of DCFH to DCF in the presence of H 2 O 2 (17,33). Results show that C 2 -ceramide induced a dose-and time-de-pendent increase in DCF staining (Fig. 2, A-C). DCF fluorescence was noticeable in cells treated with 20 M ceramide and reached a maximum in cells exposed to 50 M C 2 -ceramide for 8 h (Fig. 2, B and C). The oxidation of the probe gradually increased over a 2-to 8-h time period in cells treated with 50 M C 2 -ceramide. Incubation with an inactive form of ceramide, C 2dihydroceramide, lacking a double bond did not induce any appreciable DCF green fluorescence (Fig. 2, A and B). These results suggest that the active C 2 -ceramide induces intracellular oxidant generation as detected by DCF fluorescence.
Next, we determined the effect of C 2 -cer in cells treated with dihydroethidium (DHE), a fluorescent probe that reacts with superoxide to form a characteristic red fluorescence. Recent reports indicate that superoxide anion reacts with dihydroethidium to form a product that is distinctly different from ethidium (34). As shown in Fig. 2 (D-F), there was a dose-and time-dependent increase in the intensity of red fluorescence with ceramide treatment, indicative of enhanced superoxide generation. Again the inactive form of C 2 -cer, C 2 -dihydroceramide, did not induce red fluorescence in cells treated with HE (Fig. 2, D and E).
Ceramide-induced 55 Fe Uptake and TfR Expression-We determined the role of transferrin iron in ceramide-induced apoptosis and oxidant formation in endothelial cells by measuring TfR levels and 55 Fe uptake. BAECs were treated with different concentrations of C 2 -cer for various time periods as described in Fig.  3C, and TfR levels were measured under these conditions. Results show that ceramide treatment increased TfR levels in a dose-and time-dependent manner (Fig. 3, A and B). TfR levels were found to be significantly increased by 2 h with 50 M C 2 -ceramide (Fig. 3B). The inactive analog of C 2 -ceramide did not appreciably affect TfR levels in BAECs. Consistent with the increase in TfR levels, the uptake of 55 Fe was considerably increased in C 2 -ceramide-treated cells (Fig. 3C). Pretreatment with anti-TfR antibody inhibited ceramide-induced 55 Fe uptake (not shown). These results clearly indicate a definite role for Tf-iron in ceramide-induced oxidative stress in endothelial cells.
Ceramide-induced Apoptosis in Endothelial Cells-To investigate the mitochondrial involvement in ceramide-induced apoptosis, we measured the release of cytochrome c from the mitochondria into the cytosol. Incubation of BAECs with C 2ceramide for 8 h caused a dose-dependent increase in the release of cytochrome c into the cytosol (Fig. 4A). BAECs treated with 50 M of C 2 -ceramide induced a decrease in the antiapoptotic protein Bcl-2 located on the outer membrane of the mitochondria (Fig. 4B) as well as in the Hsp-70 protein levels (Fig. 4B). We then identified the actual caspase family (caspases 3, 6, 8, and 9) that was activated during ceramideinduced apoptosis. As shown in Fig. 4, only the activity of the effector caspase-3 began to increase as early as 4 h in ceramidetreated cells, reaching a peak value by 8 h. In addition, the increase in caspase-3 activity was detected in cells only when the concentration of C 2 -ceramide exceeded 20 M (Fig. 4C). In contrast, the caspase-3 activation was not evident in BAECs treated with an inactive analog of C 2 -ceramide, namely C 2dihydroceramide (Fig. 1). Although there was a marginal increase in the activation of caspase-6 and caspase-8 (Fig. 4, E and F), their significance in C 2 -ceramide-induced apoptosis was not apparent, because their activities became noticeable only at 16 h; at this time point, however, the execution of DNA fragmentation had already occurred (Fig. 4H). The marginal increase observed in the caspase-9 activity (Fig. 4G) by 4-h treatment with C 2 -ceramide would likely activate the caspase-3 activity, initiating a possible feedback loop that further increased their individual activities.
The time-dependent increase in caspase-3 activity was further correlated with its downstream target, namely, poly(ADPribose) polymerase (PARP) cleavage (Fig. 4H), which is responsible for DNA fragmentation. The PARP cleavage started to appear at 4 h of treatment with C 2 -ceramide and significantly increased at 8 and 16 h of incubation with C 2 -ceramide. The inactive PARP precursor protein (ϳ116 kDa) was cleaved to form a ϳ85-kDa active fragment (Fig. 4H). This active fragment of PARP translocates into the nucleus and cleaves the DNA. Finally, a dose-dependent increase in DNA fragmentation was shown by TUNEL staining (Fig. 4I) wherein cells were treated with different concentrations of C 2 -ceramide (0 -50 M) for a period of 8 h. As shown, C 2 -ceramide treatment increased the TUNEL-positive staining in cells, from 1.5% (0 M) to 48.4% (50 M). Treatment with C 2 -dihydroceramide, the inactive ceramide, did not yield DNA fragmentation as detected by TUNEL staining (Fig. 4C).
Antioxidants and Iron Chelators Mitigate Ceramide-induced Oxidative Stress and Apoptosis-The effects of different antioxidants and iron chelators in C 2 -ceramide-treated cells were investigated. BAECs were pretreated for 2 h with GSH ester (5 mM), FeTBAP (25 M), desferal (20 M), or anti-TfR antibody (12 g/ml, 42/6, IgA class, which specifically binds to the extracellular domain of the transferrin receptor and inhibits receptor endocytosis) prior to treating cells with 50 M C 2 -ceramide for 8 h. These agents inhibited ceramide-induced DCFH oxidation (Fig. 5, A and B) and caspase-3 activation (Fig. 5D). Pretreatment with a mitochondria-targeted antioxidant (e.g. Mito-Q) showed similar inhibitory effects, suggesting that mitochondrial generation of ROS is responsible for ceramidemediated oxidative stress and apoptosis (Fig. 5C). Under these conditions, these antioxidants and iron chelators also inhibited ceramide-induced DNA fragmentation as shown by TUNEL staining (Fig. 5D).
Effect of C 2 -ceramide on Intracellular ⅐ NO Generation in Endothelial Cells-At low concentrations (5-20 M), ceramide enhanced ⅐ NO release in BAECs (15). Thus, intracellular ⅐ NO levels were monitored using DAF-2 fluorescence. It has been shown that DAF-2 forms a fluorescent triazolo-type product in the presence of an oxidant derived from ⅐ NO and molecular oxygen (28). We observed a dose-and time-dependent increase in DAF-2 fluorescence in endothelial cells treated with C 2ceramide, but during longer exposure time there was a disproportionate increase in DAF-2 fluorescence that started to plateau in cells treated with 50 M C 2 -cer (Fig. 6, A and B). Ceramide-induced intracellular generation of ⅐ NO was further confirmed using a nitric-oxide synthase inhibitor, L-NAME. Pretreatment of L-NAME for 2 h before the addition of 50 M C 2 -ceramide for 8 h inhibited ceramide-induced DAF-2 fluorescence in a dose-and time-dependent manner (Fig. 6, C-F).

NOS Inhibitor Exacerbates Ceramide-induced Intracellular
ROS-Ceramide, at lower concentrations (5-20 M), induced little or no ROS (Fig. 6A) (15). However, in the presence of L-NAME, ROS generation increased in cells treated with lower concentrations of ceramide, as measured by DCF and hydroethidium fluorescence (Fig. 7, A and B). This suggests that ceramide-induced ⅐ NO counteracts the effects of ROS or ROS generation induced by ceramide; at higher concentrations of C 2 -ceramide, ROS genera-tion overwhelms the protective effects of ⅐ NO. NOS inhibition sensitized these cells to ceramide-induced oxidative stress. This is in agreement with the data showing a disproportionate increase in ⅐ NO in cells treated with C 2 -ceramide during prolonged incubation (Fig. 6, A and B), as compared with shorter period of incubation. These results show that L-NAME exacerbates C 2ceramide-induced ROS.
The next step was to investigate whether Tf-iron could play a potential role in mediating ceramide toxicity. To this end, GSH levels were measured in the presence or absence of L-NAME in C 2 -ceramide treated cells. Results show that L-NAME significantly enhanced the depletion of GSH levels caused by C 2 -ceramide (Յ20 M) (Fig. 8A). Under these conditions, the total aconitase activity was significantly lower in cells treated with both L-NAME and ceramide (Fig. 8B). The TfR levels were increased at much lower levels of ceramide in cells pretreated with L-NAME (Fig. 8C). This result prompted us to measure the iron uptake under similar conditions. We found that L-NAME treatment synergistically enhanced ceramide-mediated uptake of radiolabeled iron (Fig. 8D). At lower concentrations of C 2 -ceramide (5 M), L-NAME treatment increased iron uptake compared with C 2 -ceramide alone (Fig.  8D). Thus, intracellular depletion of ⅐ NO augments ceramideinduced iron signaling.
In addition to enhancing oxidant generation, L-NAME treatment exacerbated C 2 -ceramide-mediated apoptotic effects (Fig.  9A). The release of cytochrome c from the mitochondria was increased in the presence of L-NAME and C 2 -ceramide, as compared with C 2 -ceramide or L-NAME alone. Similar results were observed with caspase-3 activity, i.e. lower concentrations of C 2ceramide (10 -20 M) in the presence of L-NAME increased the caspase-3 activity (Fig. 9B). The DNA fragmentation (Fig. 9C) as measured by the TUNEL staining of cells increased in cells treated with 5 M and 20 M C 2 -ceramide for 16 h in the presence of L-NAME as compared with C 2 -ceramide alone.
In the presence of antioxidants and iron chelator, C 2 -ceramide/L-NAME-induced ROS generation was inhibited. As shown in Fig. 10, C 2 -ceramide/L-NAME-mediated DCF fluorescence (Fig. 10, A and B) and caspase-3 activity (Fig. 10C) in cells treated with either GSH ester (5 mM), FeTBAP (25 M), Mito-Q (1 M), desferal (20 M), or anti-TfR antibody (12 g/ml, IgA class) was considerably decreased. These results indicate that C 2 -ceramide induces the formation of ROS to a greater extent in L-NAME-treated cells and that Tf-iron plays a predominant role in ceramide-mediated toxicity.
The Biphasic Effect of Ceramide on Proteasomal Activities-Next, we investigated whether C 2 -cer treatment modulates the proteasomal activity in BAECs. Fig. 11 shows the trypsin-like and chymotrypsin-like activities of the 26 S proteasome in BAECs treated with different concentrations of C 2 -cer. In the presence of C 2 -cer (10 M), there was initially a slight increase in the proteasomal activity that drastically decreased at higher C 2 -cer concentrations (Fig. 11, A and B). This biphasic effect is attributed to the fact that lower ceramide concentration stimulates ⅐ NO generation, whereas ROS formation is favored at higher ceramide concentrations (14,15). This is in agreement with the previous report indicating that ⅐ NO, both exogenous and endogenous, enhanced the proteasomal activity, whereas ROS suppressed the proteasomal function in BAECs (18). Pretreatment of cells with antioxidants (Mito-Q and desferal) considerably increase the proteasomal activity in ceramide-treated cells (Fig. 11, C and D). This is attributed to a decrease in ROS and increase in ⅐ NO in Mito-Q-and desferal-treated cells (data not shown). DISCUSSION In this study, we demonstrated that a cell-permeable bioactive ceramide analog induces ⅐ NO at lower concentrations (Յ20 M) and ROS at higher concentrations (Ͼ20 M) in bovine aortic endothelial cells. Results showed that ceramide-induced TfR-dependent iron uptake was responsible for its prooxidant and proapoptotic effects, because pretreatment with TfR antibody or cell-permeable iron chelators greatly mitigated these effects. Depletion of intracellular ⅐ NO augmented ceramideinduced iron signaling, oxidative stress, and apoptosis. ⅐ NO suppressed the prooxidant and proapoptotic effects of ceramide by maintaining intracellular iron homeostasis.
DCF Fluorescence as an Indicator of Oxidant-induced Transferrin Iron Signaling-Ceramide-induced oxidative stress was assessed by monitoring the increase in DCF fluorescence ( Fig.  2A). Because ceramide-induced DCF fluorescence was inhibited by anti-TfR antibody, we concluded that TfR-transported

FIG. 5. Effect of antioxidants and iron chelators on ceramide-induced H 2 O 2 generation and DNA fragmentation.
A, BAECs were pre-treated with either GSH ester (5 mM), FeTBAP (25 M), desferal (20 M), or anti-TfR antibody (IgA class, 12 g/ml) for 2 h before the addition of 50 M C 2 -ceramide for 8 h, and DCF staining was carried out as described in Fig. 1 and under "Experimental Procedures." B, fluorescence intensity of data shown in A and C, same as A, except that after the treatments caspase-3 activity was measured spectrophotometrically at 405 nm by following the release pnitroanilide. D, same as A and C except that cells were treated for 16 h with C 2ceramide and other antioxidants as indicated and DNA fragmentation was measured by TUNEL staining. The data represent mean Ϯ S.D. of three independent experiments. Tf-iron was responsible for catalyzing intracellular oxidation of DCFH to DCF. We have recently shown that oxidative inactivation of iron-sensing iron-sulfur proteins (e.g. aconitase) was responsible for cellular iron uptake through increased ironregulatory protein (IRP) activity (16,17). Inactivation of aconitase by oxidative disassembly of the [4Fe-4S] cluster is accompanied by enhanced IRP activation through increased mRNA-binding activity associated with the iron-responsive element (35,36). Increased IRP1/iron-responsive element binding stabilizes the TfR mRNA, leading to enhanced mRNA FIG. 8. Effect of NOS inhibitor on ceramide-induced GSH depletion, aconitase activity, TfR expression, and iron uptake in endothelial cells. A, BAECs were treated with different concentrations of C 2 -ceramide in the presence or absence of 2 mM L-NAME (pretreatment for 2 h), and the GSH levels were determined by measuring the GSH-o-phthalaldehyde adduct using HPLC. B, same as A except that the total aconitase activity was measured spectrophotometrically at 240 nm. C and D, same as A except that the transferrin receptor levels and 55 Fe uptake were measured with and without L-NAME. Cells were treated with different concentrations of C 2 -ceramide in the presence or absence of 2 mM L-NAME for 8 h, and the uptake of labeled iron was measured as described under "Experimental Procedures." Data represent mean Ϯ S.D. of at least three separate experiments. *, significantly different (p Ͻ 0.05) compared with untreated conditions.
FIG. 9. NOS inhibitor exacerbates ceramide-induced apoptosis in endothelial cells. A, BAECs were pretreated with 2 mM L-NAME for 2 h and subsequently treated with 20 M C 2 -ceramide for 8 h, and the release of cytochrome c from the mitochondria was measured by Western analysis using an anti-cytochrome c antibody. B, cells were treated with low concentrations of C 2 -ceramide (5-20 M) for 8 h in the presence or absence of 2 mM L-NAME (2-h pretreatment), and caspase-3 activity was measured spectrophotometrically. C, same as in B except that cells were treated for 16 h with 5 and 20 M of C 2 -ceramide either in the presence or absence of L-NAME, and DNA fragmentation was measured by TUNEL-positive staining as described under "Experimental Procedures." translation, TfR synthesis, and Tf-iron uptake (37). The exogenous addition of bolus or continuously generated H 2 O 2 to endothelial cells caused enhanced oxidation of DCFH to DCF that was regulated by TfR-mediated uptake of Tf iron (17). Pretreatment of cells with anti-TfR antibody that specifically binds to the extracellular domain of TfR inhibited H 2 O 2 -induced iron signaling and ROS-mediated DCF fluorescence. The present data indicate that intracellular oxidation of DCFH is catalyzed by ceramide-induced H 2 O 2 and Tf-iron transported through TfR.
Superoxide/Hydroethidine-induced Intracellular Red Fluorescence-Hydroethidine or dihydroethidium has been widely used to detect intracellular superoxide (16,34). This assay is based on the fact that the product formed from the reaction between superoxide and dihydroethidium exhibits a red fluorescence. This product has long been thought to be ethidium (38). We recently reported that the fluorescence characteristics of the superoxide/hydroethidine product are distinctly different from those of ethidium (34). In addition, the HPLC retention time of ethidium is different from that of the product formed from the reaction between superoxide and hydroethidine (34). Although the exact structure of this product is not known, HE is still a viable fluorescent probe for detecting intracellular superoxide by monitoring the "red fluorescence" formed from HE. Preliminary experiments show that cells treated with hydroethidine and C 2 -cer exhibit a product whose fluorescence

FIG. 10. Effect of co-incubation of NOS inhibitor and antioxidants and iron chelators on ceramide-induced H 2 O 2 generation and caspase-3 activation in endothelial cells.
A, BAECs were treated with 20 M C 2 -ceramide alone or ceramide plus L-NAME or ceramide plus L-NAME plus antioxidants or iron chelators for 8 h, and H 2 O 2 generation was measured as an index of DCF fluorescence. Note that L-NAME or other compounds were pretreated for 2 h before the addition of C 2 -ceramide. In the case of antioxidant or iron chelator plus L-NAME plus ceramide groups, antioxidant or iron chelator was added 2 h prior to the addition of L-NAME, which was added 2 h prior to the addition of ceramide. B, fluorescence intensity of data shown in A; C, same as A except that caspase-3 activity was measured spectrophotometrically at 405 nm by following the release of p-nitroanilide. ϩϩ, cells were pretreated with both L-NAME (2 mM) and different antioxidants before they were treated with yield was enhanced in the presence of DNA and whose HPLC profile was distinctly different from that of ethidium (not shown). Clearly, elucidating the reaction mechanism between hydroethidine and superoxide is pivotal to our understanding of the oxidative reactions induced by ceramides and other bioactive lipid mediators.
Mitochondria as a Source of Superoxide Generation in Ceramide-treated Cells-Based on the published data (4, 39 -41), we propose that ceramide-induced ROS generation inhibits mitochondrial enzymes (aconitase and complex activity) associated with the electron transport chain, which in turn leads to more superoxide and H 2 O 2 . Inhibition of complex-1 activity stimulated superoxide formation through increased auto-oxidation of ubisemiquinone (43,44). Ceramide-induced oxidative stress was thought to result from inhibition of mitochondrial complex III activity (45). The molecular signaling events induced by the lipid second messenger, C 2 -ceramide, and their role in inhibiting the respiratory enzymes are not fully known. The activation of rac 1, the protooncogene family member and a regulatory component of NADPH oxidase, in ceramidetreated endothelial cells was reported to induce mitochondrial oxidative stress (39). The role of mitochondria in ceramideinduced ROS generation is implicated, because pretreatment with mitochondria-targeted antioxidants (Mito-quinone or Mito-vitamin E) (22) greatly inhibited C 2 -cer-induced DCF and HE fluorescence (data not shown).
Nitric Oxide, Iron Homeostasis, and Proteasomal Activation-It has been previously shown that addition of low concentrations of C 2 -ceramide (5 M) to BAECs significantly increases the eNOS activity due to translocation of eNOS from the endothelial cell membrane to intracellular sites (15). However, at higher concentrations of ceramide, even though there was an increased eNOS mRNA and protein expression, there was a decrease in the levels of bioactive ⅐ NO (13). These findings are in accordance with the present data in that at low concentrations of C 2 -ceramide (Ͻ 20 M), there was an increase in the generation of ⅐ NO and a decrease in ROS generation, whereas, at higher concentrations of ceramide (Ͼ 20 M), ROS generation becomes dominant. Although the reasons for the shift in mechanism are not completely understood, it is likely that ROS can rapidly react with ⅐ NO, thereby reducing its bioactivity. The antioxidative and cytoprotective effects of ⅐ NO were recently attributed to ⅐ NO-induced proteasomal activation (18). Proteasomal inhibitors abrogate ⅐ NO-mediated cytoprotection and antioxidative effects (18). ⅐ NO as a stimulator of proteasomal function is still a nascent concept (18). It was suggested that ⅐ NO mitigated peroxide-induced transferrin iron uptake, DCFH oxidation, and apoptosis by enhancing the proteolytic activity in endothelial cells (17,18). Depletion of endogenous ⅐ NO with L-NAME decreased the trypsin-like activity of 26 S proteasome in endothelial cells (18). Previous results also indicated that ⅐ NO controls peroxide-induced iron signaling, intracellular iron homeostasis, and oxidative stress through increased proteasomal activation (18). In the present study, C 2 -cer-induced Tf-iron uptake, DCFH oxidation, and caspase-3 activation were greatly enhanced in the presence of L-NAME. Treatment with C 2 -cer activated the proteasomal function at lower concentrations (which induced ⅐ NO), and higher concentrations of C 2 -cer (which induced more ROS and less ⅐ NO) inhibited the intracellular proteasomal function (Fig.  11). C 2 -ceramide-induced proteolytic activation was inhibited by L-NAME, suggesting a role for ⅐ NO. In contrast, L-NAME did not inhibit the proteasomal inactivation observed at higher concentrations of C 2 -cer (Fig. 11). If any, L-NAME enhanced the proteasomal inactivation induced by C 2 -cer at higher concentrations (Fig. 11). This is consistent with the enhanced iron uptake and oxidant generation observed in BAECs treated with C 2 -cer (5-20 M) and L-NAME (Fig. 3C).
Ceramides: Prooxidant Lipid Mediators in Cardiovascular and Neurodegenerative Diseases?-Ceramide-induced oxidative stress was proposed to play a key role in the pathogenesis of age-related diseases with lipid abnormalities, leading to the accumulation of sphingomyelin, ceramide, and cholesterol esters (46 -48). Altered lipid metabolism resulting in elevated ceramide levels and increased oxidative stress had been implicated in atherosclerosis, amyotrophic lateral sclerosis, and Alzheimer's disease (46,49,50). The ceramide-induced ROSmediated apoptotic signaling pathway has been suggested to play a key role in the degeneration of dopaminergic neurons in patients with Parkinson's disease (7,12). In these pathologies, perturbed iron metabolism or elevated iron levels were shown to be prominent. Thus, ceramide-induced oxidative stress and iron signaling reported in the present work explain in part the mechanism of initiation and propagation of lipid peroxidative processes in the pathogeneses of these diseases. ⅐ NO levels are impaired in age-related cardiovascular and neurodegenerative diseases (42,51). The present data suggest that there exists an inverse relationship between endogenous ⅐ NO and cellular iron levels. Clearly, future studies should be directed at investigating more thoroughly the intriguing connection between ceramide, ⅐ NO, and iron in neurovascular and cardiovascular pathologies.