Ceramide in Nitric Oxide Inhibition of Glioma Cell Growth EVIDENCE FOR THE INVOLVEMENT OF CERAMIDE TRAFFIC*

The treatment of C6 glioma cells with the nitric oxide donor, PAPANONOate (( Z )-[ N -(3-ammoniopropyl)- N -( n -propyl)amino]diazen-1-ium-1,2-diolate), resulted in a dose-dependent inhibition of cell proliferation. This was associated to a rapid and significant increase of ceramide levels and was mimicked by treatments that aug-ment cellular ceramide. Metabolic experiments with radioactive sphingosine, serine, and choline showed that nitric oxide strongly reduced the utilization of ceramide for the biosynthesis of both sphingomyelin and glucosyl-ceramide. Nevertheless, nitric oxide did not modify the activity of different enzymes of ceramide metabolism. The possibility that nitric oxide impairs the availability of ceramide for sphingolipid biosynthesis was then investigated. The metabolism of N -hexanoyl-[ 3 H]sphin-gosine demonstrated that nitric oxide did not affect the biosynthesis of N -hexanoyl-[ 3 H]sphingolipids but inhibited the metabolic utilization of long chain [ 3 H]ceram-ide, synthesized in the endoplasmic reticulum (ER) from the recycled [ 3 H]sphingosine. Moreover, results obtained with fluorescent ceramides, brefeldin A, ATP depletion, as well as in a ceramide

During the past 10 years, a great amount of evidence has demonstrated that ceramide (Cer), 1 a key molecule in sphingo-lipid metabolism, represents a crucial modulator of cell life, being involved in signal transduction mechanisms that control cell growth arrest, differentiation, and apoptosis (reviewed in Refs. [1][2][3]. Changes in Cer levels can be induced by a large number of stimuli (including growth factors, hormones, and cytokines), cytotoxic and infection agents, as well as different environmental conditions (1,3,4). The mechanisms through which stimuli control Cer intracellular levels involve the regulation of multiple enzymes with different subcellular localization: sphingomyelinase, sphingomyelin-synthase, glucosylceramide synthase, Cer-synthase, ceramidase, and serine palmitoyltransferase (3,(5)(6)(7). In addition, in the same cell type, different isoforms of some of these enzymes may be present, and a single stimulus may have more than one target enzyme. Moreover, it has been demonstrated that Cer interaction with multiple targets can result in different biological responses (1).
A key element in defining the role of Cer in cell signaling is its hydrophobic nature, and its consequent inability to spontaneously move among different subcellular sites where the enzymes of its metabolism and its molecular targets are located. As a consequence, the biological effects of Cer may also depend on the regulation of its intracellular traffic and the presence of specific signaling pools of this bioactive sphingoid in the cell (7)(8)(9).
The role of Cer as a mediator in signal transduction pathways governing life has also been demonstrated in cells of the nervous system (reviewed in Refs. 10 -12). In particular, recent studies indicate that, through the regulation of sphingomyelinase, sphingomyelin synthase, or serine palmitoyl transferase activities, the modulation of Cer levels in glial cells plays a fundamental role in the control of cell proliferation, differentiation, and apoptosis (13)(14)(15)(16). This crucial role of Cer is strongly supported by recent evidence about Cer mediation in the apoptotic and antitumoral activity of cannabinoids in gliomas (14) and Cer tumor levels inversely correlated with human astrocytoma malignancy and poor prognosis (17).
Among the molecules that can regulate growth, differentiation, and apoptosis in different cell types, nitric oxide (NO) has emerged as a key element in its role as an inter-and intracellular mediator of physiological and pathophysiological events in the nervous system (18 -21). The major enzymes responsible for NO production are the constitutive neuronal NO synthase in neurons and the inducible one in glial cells (20,21). Various stimuli are able to regulate the activity and/or expression of these enzymes. Noteworthy, in both astrocytes and glioma cells, Cer is involved in cytokine-mediated NO synthase induction (22,23).
The role of NO in the control of glial cell growth is documented by studies demonstrating the antiproliferative action of NO in both astrocytes and glioma cell lines (24,25). Notwith-standing this evidence, the molecular mechanisms underlying these effects on glial cell growth are still largely unknown. Furthermore, in some extraneural cells, recent evidence shows that NO may exert an apoptotic effect by increasing Cer level through the regulation of the metabolic pathways involved in its generation or removal (26,27).
In this study we investigate the possible involvement of Cer in the inhibitory role exerted by NO on C6 glioma cell proliferation. In particular, we focus on the effects of NO on the molecular mechanisms involved in the control of cellular Cer levels. Here, we present evidence that in C6 glioma cells NO promotes a Cer increase by inhibiting its transport, possibly vesicle-mediated, from ER to Golgi apparatus. As a consequence, newly synthesized Cer accumulates and appears to act as a mediator to the antiproliferative effect of NO.   (28). PA-PANONOate was from Alexis-Italia (Firenze, Italy). 6-((N-(7-Nitrobenz-2-oxa-1,3-diazol-4-yl)amino)hexanoyl) sphingosine (NBD-C 6 Cer) and N-(4,4-difluoro-5,7-dimethyl-bora-3a,4a-diaza-s-indacene-3-pentanoyl) sphingosine (BODIPY-C 5 Cer) were from Molecular Probes Europe (Leiden, The Netherlands). High performance thin layer chromatography (HPTLC) silica gel plates were from Merck (Darmstadt, Germany).
Cell Cultures-The C6 glioma cell line was obtained from Istituto Zooprofilattico Sperimentale della Lombardia e dell'Emilia (Brescia, Italy). Cells were routinely maintained in DMEM supplemented with 10% FCS (DMEM plus 10% FCS), 100 units/ml penicillin, and 100 g/ml streptomycin at 37°C in an atmosphere of 5% CO 2 and 95% humidified air. Cell viability was assessed with the trypan blue exclusion test.
Proliferation Assays-For proliferation assays, cells were plated in 35-mm dishes (10 4 cells/cm 2 ) and maintained in DMEM plus 10% FCS for 24 h. Then, the plates were washed twice with DMEM containing 0.25% FCS and incubated for 48 h in the same medium before use as quiescent cells. Cells were then incubated in DMEM plus 10% FCS for 24 h with or without the following molecules: 100 -400 M PA-PANONOate, 10 M-1 mM cGMP, 10 M C2-Cer or C6-Cer or C2-DHCer, and 0.25 unit/ml SMase. 1 Ci/ml [ 3 H]thymidine was added to each plate 4 h before cell harvesting, and the thymidine incorporation in trichloroacetic-insoluble material was determined.
Ceramide Quantification-Total sphingolipids of C6 glioma cells (6 ϫ 10 5 cells) were labeled at equilibrium by incubating cells with 1 Ci of [ 3 H]Sph (20 Ci/mmol) in DMEM plus 10% FCS for 24 h. At the end of incubation, the plates were washed twice with fresh medium and chased for 3 h in DMEM plus 10% FCS. [ 3 H]Sph-labeled cells were then stimulated with 400 M PAPANONOate in DMEM plus 10% FCS or incubated in DMEM plus 10% FCS. At different intervals, the cells were washed twice with PBS at 4°C and harvested, and total lipids were extracted as recently described (28). The total lipid extract was partitioned, and the organic phase was subjected to mild alkaline methanolysis, a 1-h treatment at 37°C with 0.1 M KOH in methanol. The methanolyzed organic phase and the aqueous phase were analyzed by HPTLC using as solvent systems chloroform/methanol/water (55:20:3, by volume) and butanol/acetic acid/water (3:1:1, by volume), respectively.
[ 3  In chase experiments, after a 1-h pulse with the radiolabeled compound (2 Ci/ml), the cells were submitted to a period of chase in DMEM plus 10%FCS with or without 400 M PAPANONOate. At appropriate pulse or chase times, the cells were washed twice with PBS at 4°C and harvested, and total lipids were extracted and processed as previously described (28).
[Sph- 3 (28); the same conditions were maintained during the [ 3 H]Sph pulse. At different pulse times the cells were washed twice with PBS at 4°C and harvested, and total lipids were extracted and processed as previously described (28).
In Vitro Enzyme Assays-SMases, SM synthase, and GlcCer synthase activities were assayed using as enzyme source cell homogenate (obtained by sonication in H 2 O three times, 10 s at 4°C) of control and PAPANONOate-treated cells. Mg 2ϩ -dependent neutral sphingomyelinase (N-SMase) and acidic sphingomyelinase (A-SMase) were assayed using [Sph-3 H]SM as substrate as previously described (29). GlcCer synthase activity was assayed as previously described (30) with minor modifications. The incubation mixture contained 50 mM Tris-Cl (pH 7.4), 25 mM KCl, 10 mM MnCl 2 , 5 mM UDP-Glc, 2 nmol (0.05 Ci) of [Sph-3 H]C6-Cer (as 1:1 complex with fatty acid-free BSA), and 15 g of cell protein in a final volume of 50 l. After 15-min incubation at 37°C, the reaction was stopped by adding 150 l of chloroform/ methanol (1:2, by volume) at 4°C. In all cases, after lipid extraction and phase separation, the [ 3 H]lipids were resolved by HPTLC as previously described (16).
Sphingosine kinase activity was assayed as described (31). Briefly, control and PAPANONOate-treated cells were washed with PBS and scraped in Sph kinase buffer (20 mM Tris-Cl (pH 7.4) containing 20% glycerol, 1 mM mercaptoethanol, 1 mM orthovanadate, 1 mM EDTA, 15 mM NaF, 0.5 mM 4-deoxypyridoxine, 0.4 mM phenylmethylsulfonyl fluoride, 40 mM ␤-glycerophosphate, and 10 g/ml each of leupeptin, aprotinin, and trypsin inhibitor). Cells were disrupted by freeze and thawing and centrifuged at 105,000 ϫ g for 90 min. An aliquot of the supernatant (20 -50 g of protein) was diluted in Sph kinase buffer, then Sph (50 M final concentration as a 1:1 complex with fatty acidfree BSA) was added. The reaction was started by the addition of 10 Ci of [␥-32 P]ATP (1 mM final concentration) containing MgCl 2 (10 mM final concentration) in a final volume of 200 l. After incubation at 37°C for 30 min, the reaction was stopped by the addition of 1 M HCl, and radiolabeled sphingosine 1-phosphate was resolved by HPTLC (31). After visualization by autoradiography, the radioactive spots were scraped off from the plate and counted for radioactivity.
Analysis of the Intracellular Distribution of Fluorescent Ceramides-C6 glioma cells plated at 1.5 ϫ 10 4 cell/cm 2 were grown on a glass coverslip and maintained 48 h in DMEM plus 10%FCS. Then cells were incubated with 5 M BODIPY-C 5 -Cer or NBD-C 6 -Cer (as 1:1 complex with fatty acid free BSA) in DMEM at 4°C for 30 min (32).
After washing (three times with DMEM plus 10%FCS and 0.34 mg/ml fatty acid-free BSA), cells were incubated 1 h at 37°C in DMEM plus 10%FCS with or without 400 M PAPANONOate. Cells were then washed (three times with PBS) and fixed with 0.5% glutaraldehyde solution in PBS for 10 min at 4°C. The specimens were immediately observed and analyzed with a fluorescence microscope (Olympus BX-50) equipped with a fast high resolution charge-coupled device camera (Colorview 12) and a image analytical software (Analysis from Soft Imaging System GmbH).
Preparation of Cytosolic Fractions-C6 glioma cells plated at 6 ϫ 10 4 cell/cm 2 in 100-mm dishes were maintained 48 h in DMEM plus 10%FCS and then incubated for 1 h in the same medium with or without 400 M PAPANONOate. At the end of incubation, the cells were washed twice with PBS at 4°C and gently harvested in 0.75 ml of PBS. Cytosolic fractions were prepared as previously described (33) with some modifications. All manipulations were carried out at 4°C or on ice. The cells were precipitated by centrifugation at 250 ϫ g for 10 min, suspended in 4 volumes of 10 mM Tris-Cl, pH 7.4, 0.25 M sucrose, and homogenized by 3 cycles of freezing in liquid nitrogen and thawing at 37°C. The homogenate was centrifuged at 900 ϫ g for 10 min to precipitate nuclei. The postnuclear supernatant was centrifuged twice at 100,000 ϫ g for 1 h to remove the particulate fraction completely. The supernatant obtained from the second centrifugation was rapidly frozen in liquid nitrogen and stored at Ϫ80°C until use as a cytosolic fraction. The protein concentration of the cytosolic fraction obtained was about 1.8 mg/ml. Hereafter, all manipulations were carried out at 4°C. Semi-intact C6 glioma cells were prepared as previously described (33) with several modifications. In particular, cells were washed twice in a hypotonic buffer (10 mM Hepes-KOH, 40 mM KCl, 0.1 mM MgCl 2 , pH 7.3) and incubated in the same buffer for 10 min. Then, the hypotonic buffer was replaced by an isotonic one (10 mM Hepes-KOH, 115 mM KCl, pH 7.3), and the cells were rapidly, gently scraped and collected by low speed centrifugation at 250 ϫ g for 5 min with the brake off. After washing with the isotonic buffer, the pellets were resuspended in the isotonic buffer (about 120 l/dish) and immediately used as [ 3 H]Cer-labeled semiintact cell preparations. With this procedure, more than 80% C6 glioma cells were trypan blue-positive. The enzymatic analysis of both precipitate and supernatant fractions showed that more than 85% lactate dehydrogenase activity was recovered in the supernatant. The protein concentration of the semi-intact cell preparations from both control and NO-treated cells was in the 1.2-1.4 mg of protein/ml range.

Preparation of [ 3 H]Cer-labeled Semi-intact Cells-Cells
In Vitro Assay of Cer Transport from ER to the Site of GlcCer and SM Synthesis-The Cer transport assay was performed (33) by incubating In some experiments, the transport assay was performed in the absence of the ATP-regenerating system or in 20 mM Hepes-KOH, pH 7.0, and 70 mM KCl, with or without 0.5 mM UDP-Glc. In the cytosol exchange experiments, semi-intact cells from control or NO-treated cells were incubated in the presence of the cytosol from NO-treated or control cells, respectively. Mixtures were incubated 30 min at 37°C and stopped by the addition of 780 l of chloroform/methanol (1:2, by volume) and partitioned as described (33). After counting the radioactivity, the organic phase was analyzed by HPTLC using as solvent system chloroform/methanol/water (55:20:3, by volume). To obtain quantitative data on the transport activities of semi-intact cells, [ 3 H]Cer-labeled semi-intact cells were also extracted prior to the Cer transport assay, and the radioactivity associated to Cer, SM, and GlcCer was used as the background.
Other Methods-Total protein was assayed with the Coomassie Bluebased Pierce Reagent, using BSA as standard. SM and GM3 contents were determined as previously reported (34,35). Lactate dehydrogenase activity was measured according to Strorrie and Maddie (36). Radioactivity was measured by liquid scintillation counting or radiochromatoscanning. Digital autoradiography of HPTLC plates was performed with a Beta-Imager 2000 (Biospace, France), and the radioactivity associated with individual lipids was determined with software provided with the instrument. The 3 H-labeled sphingolipids were recognized and identified as previously described (28). Statistical significance of differences was determined by Student's t test.

Effect of NO on C6 Glioma
Cell Proliferation-To investigate the effect of exogenously delivered NO on C6 glioma cell proliferation, cells were treated with different concentrations of the NO-releasing molecule PAPANONOate. In the 100 -400 M range, PAPANONOate induced a dose-dependent inhibition of cell proliferation (Fig. 1, left panel). The maximal inhibitory effect on [ 3 H]thymidine incorporation of about 80% was reached at 400 M PAPANONOate. In these experimental conditions, at both 24 and 48 h after treatment with the NO donor, cell viability, assessed by trypan blue exclusion, was not affected. This indicates that the effect of the NO-releasing compound on [ 3 H]thymidine incorporation results from the inhibition of cell growth and not from toxicity. Moreover, when PAPANONOate was used after being decayed for 24 h, no effect on [ 3 H]thymidine incorporation was observed, indicating that the effect of PAPANONOate on C6 glioma cell proliferation was due to NO generation. As shown in Fig. 1 (right panel), dibu- tyryl-cGMP (the membrane-permeant analog of cGMP) administered in the 0.01-1 mM range, did not affect [ 3 H]thymidine incorporation into proliferating glioma cells.
Involvement of Cer in the NO-induced Cell Growth Inhibition of C6 Glioma Cells-We next evaluated the possible involvement of Cer as a mediator for the antiproliferative activity of NO in C6 glioma cells. To this purpose, we first measured intracellular Cer levels at various intervals after the administration of PAPANONOate to cells labeled to the equilibrium with [ 3 H]Sph. In preliminary experiments, we set up the labeling conditions so as to obtain a [ 3 H]GM3/[ 3 H]SM ratio corresponding to that of the endogenous compounds as an index for a steady-state labeling of cell sphingolipids. As shown in Fig. 2, in these conditions, exogenously delivered NO caused a rapid and significant increase of the Cer cellular level. The maximum increase, corresponding to 167% of the control levels, was reached 4 h after NO treatment. Cer levels in NO-treated cells remained significantly higher than in controls, even after 24 h.
In addition, we obtained evidence that those treatments able to increase Cer cellular levels mimicked the NO-induced cell growth inhibition (Fig. 3). In particular, the incubation of C6 cells with cell-permeable analogs C2-or C6-Cer as well as the treatment with bacterial SMase resulted in a significant reduction of [ 3 H]thymidine incorporation into DNA. Under the same conditions, the C2-Cer dihydro-derivative had no effect.  (Fig. 4, upper panel). In these conditions, [ 3 H]Sph formed from [ 3 H]Cer hydrolysis was 3.3% total incorporated radioactivity in control and 3.1% in treated cells, indicating that NO did not modify Cer cleavage by ceramidases. We also evaluated the effect of NO on in vitro activity of N-SMase and A-SMase. Using homogenates obtained from control or NO-treated cells as the enzyme source, and adding PAPANONOate to the homogenate of control cells, no difference was observed in either N-SMase or A-SMase activities (Fig. 4, lower panel).

Effect of NO on [ 3 H]SM and [ 3 H]Choline Metabolism and on
The possible effect of NO on SM biosynthesis was then evaluated by pulsing cells with [ 3 H]choline with or without PA-PANONOate (Fig. 5, upper panel). At 2-h pulse, the incorporation of [ 3 H]choline into SM in NO-treated cells was about 40% of that measured in control cells, whereas the incorporation of the same precursor into phosphatidylcholine, the phosphorylcholine donor for the conversion of Cer to SM, was not affected by NO. On the basis of this evidence we assessed the effect of NO on in vitro activity of SM synthase. Here too, the homogenates obtained from control or NO-treated cells were used as the enzyme source, and, in the case of control cells, the enzymatic activity was also measured in the presence of PA-PANONOate. As shown in Fig. 5 (lower panel), neither NO cell treatment nor NO addition in the enzymatic assay affected SM synthase activity in C6 glioma cells.

Effect of NO on [ 3 H]Sph
Metabolism-For further information on the effect of NO on SM biosynthesis, taking into account that in glial cells exogenous Sph is rapidly incorporated first in Cer and then in SM and glycosphingolipids (37,38), we performed an additional metabolic study using [ 3 H]Sph. When C6 glioma cells were pulsed for 1 h with [ 3 H]Sph with or without PAPANONOate, the radioactive precursor was rapidly and efficiently incorporated in both control and treated cells. In both cases, [ 3 H]Sph was mainly metabolized to N-acylated compounds, most represented by Cer, SM, and, in lower amounts, GlcCer and GM3 (Fig. 6, upper panel). After 1-h pulse, the uptake of Sph and the incorporation of radioactivity into N-acylated compounds (as the sum of tritiated Cer, SM, GlcCer, and GM3) were very similar in control and NO-treated cells. However, treatment with NO strongly modified the distribution of radioactivity between the different Sph metabolites; in fact, [ 3 H]Cer was about 2-fold higher in NO-treated than in control cells (Fig. 6, upper panel). At the same time, in NO-treated cells, the radioactivity incorporated into SM, GlcCer, and GM3 was 40, 35, and 34% less, respectively, than in controls. In these conditions, treatment with NO donor did not modify the amount of radioactivity incorporated into sphingosine 1-phosphate (Fig. 6, upper panel). As in the case of SM synthase, the in vitro activity of GlcCer synthase was similar in control (0.27 Ϯ 0.04 nmol/mg of protein/min) and NO-treated cells (0.26 Ϯ 0.039 nmol/mg of protein/min), and this was also found after addition of PAPANONOate to the incubation mixture of control cell assay (0.28 Ϯ 0.045 nmol/mg of protein/min). In addition, the activity of Sph kinase assayed in vitro was very similar in control (0.12 Ϯ 0.013 nmol/min/mg of protein) and NO-treated (0.13 Ϯ 0.015 nmol/min/mg of protein) cells.

Effect of NO on Sphingolipid Metabolism from [ 3 H]Serine-
The possible effect of NO on the de novo biosynthesis and metabolic processing of Cer was investigated after administration of [ 3 H]serine. After a 1-h pulse, the incorporation of radioactivity into total lipids and the amount of radioactivity associated to sphingolipids were very similar in control and NO-treated cells indicating that NO did not affect the de novo biosynthesis of Cer. At this pulse time, NO treatment promoted an increase of the radioactivity associated to Cer with a concomitant decrease of that associated to SM (Fig. 6, lower left  panel). This effect was found to be more marked when NO was administered during chase. In particular, as shown in Fig. 6  (lower right panel), NO induced a 40% increase in [ 3 H]Cer levels, paralleled by a significant decrease of the radioactivity incorporated into SM. Effect of NO on the Intracellular Distribution of BODIPY-C 5 -Cer and NBD-C 6 -Cer-The transport of natural Cer from ER to the Golgi apparatus can be qualitatively evaluated from the analysis of BODIPY-C 5 -Cer redistribution into cells (32,40). We thus investigated the effect of NO on the behavior of this fluorescent Cer into intact C6 glioma cells. After labeling ER and other intracellular membranes, cells were chased, in the presence or absence of 400 M PAPANONOate to allow the redistribution of BODIPY-C 5 -Cer. In control cells most of fluorescence accumulated in the perinuclear region (Fig. 8A), representative of the Golgi apparatus (32), whereas in NO-treated cells the accumulation of fluorescence in the Golgi region was strongly reduced (Fig. 8B). In contrast, when cells were labeled with NBD-C 6 -Cer, which spontaneously moves between intracellular membranes, NO did not appreciably modify the accu-mulation of NBD fluorescence in the perinuclear Golgi region (Fig. 8, C and D). H]Sph in an ATP-depleting medium, the incorporation of radioactivity into SM and GlcCer was found to be about 70 and 30% less than controls (Fig. 10). This was paralleled by a 2-fold increase of [ 3 H]Cer levels. Moreover, ATP depletion did not reduce either [ 3 H]Sph uptake or its incorporation into N-acylated compounds (as the sum of tritiated Cer SM and GlcCer, Fig. 10). This ruled out the possibility that the decrease in [ 3 H]SM formation in ATP-depleted cells was due to a decrease in the cellular uptake of [ 3 H]Sph or its utilization for  (Fig. 11, upper left panel). In the transport mixture containing the cytosol obtained from control cells, the formation of both [ 3 H]SM and [ 3 H]GlcCer was increased by about 4-fold; the same increase in [ 3 H]GlcCer was observed after adding UDP-Glc to the buffer (Fig. 11, upper left  panel). The removal of the ATP-regenerating system from the transport mixture led to a strong reduction of SM but not GlcCer formation (Fig. 11, upper left panel).

Effect of NO on SM and GlcCer Formation from [ 3 H]Sph in BFA-treated Cells-To
As shown in Fig. 11 (upper right panel), incubation of NOtreated semi-intact cells with buffer alone led to [ 3 H]SM and [ 3 H]GlcCer biosynthesis similar to that observed in control semi-intact cells. The addition of NO-treated cytosol and the transport mixture promoted a nearly 3-fold increase in [ 3 H]GlcCer formation without any significant change in [ 3 H]SM; in the absence of the ATP-regenerating system, SM formation was not modified. Here too, the presence of UDP-Glc in the buffer was sufficient to promote [ 3 H]GlcCer biosynthesis. Thus, these results confirm that, in C6 glioma cells, NO impairs the ATP, cytosol-dependent, inter-membrane Cer traffic needed mainly for SM biosynthesis and, to a lesser extent, for GlcCer. To see if any specific cellular component is involved in the NO-induced defective Cer translocation, we carried out fraction exchange experiments in the transport assay (Fig. 11, lower  panel). The use of NO-treated cytosol with control membranes did not significantly modify the synthesis of [ 3 H]SM or [ 3 H]GlcCer. Moreover, the use of cytosol obtained from control cells in combination with NO-treated cells did not restore the capacity to synthesize [ 3 H]SM and [ 3 H]GlcCer (Fig. 11, lower panel) thus suggesting that NO exerts its effect on Cer transport at the level of the membranous compartment involved in Cer traffic. DISCUSSION The first evidence provided by this study is that NO exerts a dose-dependent antiproliferative effect on C6 glioma cells. As observed in other cell types (42,43), this effect was not mimicked by a membrane-permeant non-hydrolyzable analog of cGMP, supporting the evidence that also in these cells the antimitogenic effect of NO is independent from the activation of guanylate cyclase, and thus from cGMP. Instead, we found that NO induces an early and significant increase in Cer levels and that treatments resulting in the increase of cell Cer are able to mimic the antiproliferative effect of NO in C6 glioma cells. Altogether, these data strongly suggest that Cer may be a mediator of the antiproliferative effect of NO. In this context, it is noteworthy that Cer has been recently indicated as a mediator of the apoptotic response to NO in glomeruloendothelial as well as mesangial and HL60 cells (26,27). Furthermore, these results strengthen the general role of Cer in the control of cell proliferation (5) and, in particular, as a negative regulator of glial growth (13,29,44).
Evidence obtained from non-neural cells indicates that the modulation of Cer levels following NO treatment can depend on both the activation of N-and A-SMase and the inhibition of ceramidases (26,27). The studies here performed to single out the metabolic pathway responsible for NO-dependent Cer accumulation in C6 glioma cells indicate that the degradation of both SM and Cer is not involved in the NO-induced regulation of Cer levels. By using different experimental approaches, the major pathway responsible for the NO-dependent Cer increase was found to involve a reduced utilization of Cer for the biosynthesis of complex sphingolipids, mainly SM. Thus, unlike what was observed in other cell types (26,27) These results strongly support that Cer utilization for SM biosynthesis represents a key element/module in the Cer signaling involved in the control of glial cell growth (16,44) and more in general in cell proliferation (for a review see Ref. 5). Notwithstanding the evidence of a NO-dependent inhibition of both SM and GlcCer biosynthesis, no direct effect of NO on the enzymatic activity of either SM or GlcCer synthase was detected. Thus, differently from what occurs in cerebellar astrocytes and hippocampal neurons after stimulation with basic fibroblast growth factor (16,45), the control of Cer levels might be achieved by a mechanism other than the direct regulation of SM-and/or GlcCer-synthase. A reasonable possibility is that NO acts by reducing the availability of Cer as substrate for both enzymes, possibly inducing a defect in the translocation of Cer, synthesized in the ER, to the sites where SM-synthase and GlcCer-synthase are localized. Very convincing support to this hypothesis was obtained by treating C6 glioma cells with [Sph-3 H]C6-Cer, with or without the NO-releasing agent. In fact, in the presence of NO, the direct utilization of radiolabeled C6-Cer for the biosynthesis of C6-SM and C6-GlcCer, which does not require Cer exit from the ER, was unmodified. Moreover, treatment of C6 glioma cells with C6-Cer also resulted in the generation of endogenous long-chain Cer, as in A549 cells (39). This process most reasonably involves the recycling of Sph for the biosynthesis of long-chain Cer at the ER level. Opposite to what was observed for short-chain Cer, the utilization of longchain Cer for the biosynthesis of complex sphingolipids was strongly inhibited by NO. This finding further confirms that newly synthesized Cer in the ER is no longer available for complex sphingolipid biosynthesis in the presence of NO.
The analysis of intracellular distribution of BODIPY-C 5 -Cer, which mimics the intracellular movements of natural Cer, provides additional evidence that NO can influence the intracellular traffic of Cer from ER to the Golgi apparatus. Moreover, results obtained with BFA demonstrate that, in C6 glioma cells, the cis-Golgi represents the major subcellular site for both GlcCer and SM biosynthesis and supports that NO induces a defect in Cer translocation from ER to the Golgi apparatus. In fact, when cells were treated with BFA, which causes Golgi disassembly and redistribution to the ER (41), the conversion of Cer to both GlcCer and SM was strongly increased. The increase in GlcCer is in agreement with the generally observed cis-medial Golgi location of the glucosyltransferase involved in its biosynthesis (46,47). A different consideration must be made for the BFA-dependent SM increase in C6 glioma cells. As occurs in many cell types (48 -50), the results here obtained indicate that even in C6 glioma cells SM biosynthesis occurs mainly in the cis-medial Golgi. Nevertheless, in different cell types, SM synthesis can also occur at different subcellular sites (51)(52)(53), and, in rat cerebellar astrocytes, the activation of an SM synthase located in a compartment other than cis-medial Golgi is involved in the basic fibroblast growth factor signaling pathway involved in cell proliferation (16). The direct availability of Cer for SM and GlcCer synthases determined by BFA makes C6 glioma cells no more sensitive to NO-dependent inhibition of Cer utilization. This evidence further confirms that SM-and GlcCer-synthases are not the NO target and strengthens the notion that the NO-dependent Cer increase is mainly due to a defect in its translocation from ER to cis-Golgi.
The way Cer moves from the ER to Golgi is still unclear but, on the basis of many experimental results, two main mechanisms appear to be involved and include an ATP-dependent vesicle-mediated Cer transport, and a non-vesicular Cer trans- location that can involve the participation of transfer proteins as well as ER-Golgi membrane contacts (54 -56). Using a cell line with a specific defect in SM biosynthesis, evidence was recently presented that Cer may preferentially use an ATPindependent non-vesicular pathway for glycosphingolipid production and an ATP-dependent vesicle-mediated mechanism for the biosynthesis of SM (40). These same researchers also demonstrated that the transport of Cer for SM biosynthesis strongly depends on some unidentified cytosolic factors, whereas the biosynthesis of GlcCer is mainly cytosol-independent (33). The results obtained in our present study on ATPdepleted cells and on in vitro Cer transport assays are in agreement with this hypothesis. In fact, in C6 glioma cells, ATP depletion strongly affects SM biosynthesis and, to a lesser extent, that of GlcCer. On the other hand, the conversion of Cer into SM in semi-intact cells was strongly dependent on the presence of ATP and cytosol, whereas that of GlcCer appears to be mainly dependent on the availability of UDP-Glc. We also obtained evidence that ATP depletion mimics the inhibitory effect of NO on the biosynthesis of complex sphingolipids, in particular SM, and that NO is no longer able to exert any effect in ATP-depleted cells. Thus the ATP-dependent vesicle-medi-ated transport of Cer, primarily for SM biosynthesis, is involved in the NO-dependent Cer increase. Moreover, results obtained by using semi-intact cells and cytosol derived from NO-treated cells, led us to exclude the involvement of a cytosolic factor as a NO target and strongly suggest that the effect of NO on Cer traffic resides in the membrane component involved in this process. To explain Cer accumulation consequent to a NO-mediated impairment in its traffic, two major mechanisms can be hypothesized. First, in the presence of NO, newly synthesized Cer cannot move from the ER, thus accumulating in this compartment where it may exert its biological effects (57). Second, Cer-transferring structures, possibly vesicles, actually leave the ER but, in the presence of NO, their Golgispecific targeting is impaired, thus determining an accumulation of Cer-containing structures available for the interaction with Cer targets. Studies are in progress in our laboratory to investigate these possibilities.
In conclusion, to the best of our knowledge this represents the first evidence for an active role for Cer traffic in Cer signaling. The results provided here demonstrate that NO is able to modulate the intracellular traffic of Cer between the ER and the Golgi apparatus. As a consequence, newly synthesized Cer accumulates and appears to act as a mediator of the antiproliferative effect of NO in these cells. These results point out the relevance of Cer intracellular movements and of the role of specific signaling pools, in defining the biological effects of Cer (7,8). Altogether, this strengthens the paradigm that the understanding of the complex signaling role of Cer cannot leave out of consideration the metabolic origin, the topology of production, and the intracellular traffic of this bioactive lipid and its interplay with other signaling pathways.