Divergence in Regulation of Nitric-oxide Synthase and Its Cofactor Tetrahydrobiopterin by Tumor Necrosis Factor- a CERAMIDE POTENTIATES NITRIC OXIDE SYNTHESIS WITHOUT AFFECTING GTP CYCLOHYDROLASE I ACTIVITY*

Synthesis of 6( R )-5,6,7,8-tetrahydrobiopterin (BH 4 ), a required cofactor for inducible nitric-oxide synthase (iNOS) activity, is usually coordinately regulated with iNOS expression. In C6 glioma cells, tumor necrosis factor- a (TNF- a ) concomitantly potentiated the stimulation of nitric oxide (NO) and BH 4 production induced by IFN- g and interleukin-1 b . Expression of both iNOS and GTP cyclohydrolase I (GTPCH), the rate-limiting enzyme in the BH 4 biosynthetic pathway, was also mark- edly increased, as were their activities and protein levels. Ceramide, a sphingolipid metabolite, may mediate some of the actions of TNF- a . Indeed, we found that bacterial sphingomyelinase, which hydrolyzes sphingomyelin and increases endogenous ceramide, or the cell permeable ceramide analogue, C 2 -ceramide, but not C 2 - dihydroceramide ( N -acetylsphinganine), significantly mimicked the effects of TNF- a on NO production and iNOS expression and activity in C6 cells. Surprisingly, although TNF- a increased BH 4 synthesis and GTPCH activity, neither BH 4 nor GTPCH expression was af- fected by C 2 -ceramide or sphingomyelinase in IFN- g and interleukin-1 arbitrary density units of Health integration of ethidium bromide-stained gels, expression GTPCH were by RT-PCR. expression fold

Proinflammatory cytokines, such as IFN-␥, IL-1␤, 1 and TNF-␣, as well as a bacterial endotoxin (lipopolysaccharide (LPS)), stimulate the production of nitric oxide (NO) by increasing expression of the inducible form of nitric-oxide synthase (iNOS) in several types of cells, including macrophages (1), microglia, and astrocytes (2). Synthesis of this free radical gas is primarily a protective mechanism utilized by the host against invading organisms (reviewed in Ref. 3). On the other hand, it has been suggested that overproduction of NO in the central nervous system may mediate some of the pathological sequelae of neuroinflammatory disorders, such as multiple sclerosis (4) and neuronal death following acute injury (5).
iNOS is active as a homodimer of 130-kDa subunits and requires five cofactors to catalyze the conversion of L-arginine to L-citrulline, a reaction that liberates NO (reviewed in Ref. 6). Three of the cofactors, NADPH, FAD, and FMN, are usually present in cells at concentrations that are not limiting for enzyme activity. Calmodulin, the fourth cofactor, is constitutively bound to iNOS in a manner that, unlike its function with the two constitutive isoforms of NOS, makes iNOS activity calcium-independent (7). However, the intracellular level of the cofactor 6(R)-5,6,7,8-tetrahydrobiopterin (BH 4 ) is rate-limiting for NO generation, and its synthesis is usually co-induced by cytokines (8 -10). The exact role that BH 4 plays in iNOS catalysis is still equivocal, but it has been shown to bind to iNOS monomers, promoting their dimerization and subsequent activation (11), and recently has been proposed to play a role in the enzymatic reaction in a radical form (12).
The cellular level of BH 4 is largely regulated by the activity of GTP cyclohydrolase I (GTPCH), the first and rate-limiting enzyme in the BH 4 biosynthetic pathway (6). GTPCH, a homodecamer of 30-kDa subunits that are arranged as two pentamers facing one another (13), catalyzes the rearrangement of GTP to dihydroneopterin triphosphate. This intermediate is then converted to BH 4 in two subsequent reactions catalyzed by 6-pyruvoyltetrahydropterin synthase and sepiapterin reductase, respectively, neither of which is rate-limiting. GTPCH mRNA expression can be induced by the same proinflammatory stimuli that induce iNOS mRNA (14). Interestingly, in human umbilical vein endothelial cells, cytokine-stimulated NO production is predominantly regulated by increased GTPCH mRNA expression (15,16) and not by changes in expression of endothelial NOS.
Ceramide, formed by the sphingomyelinase (SMase)-mediated hydrolysis of sphingomyelin, is now emerging as a lipid second messenger that mediates some of the biological effects of TNF-␣, IL-1␤, and LPS in differentiation, apoptosis, and cell growth arrest (reviewed in Refs. [17][18][19]. Recently, LPS and SMase-mediated elevations in ceramide have been demonstrated to potentiate NO formation and iNOS expression in rat primary astrocytes and C6 gliablastoma cells (20). The signaling pathways involved in NO production have not yet been fully elucidated, although it appears that activation of the redoxsensitive transcription factor, NF-B, is essential for iNOS induction (21). To study the potential role of ceramide in cytokine-stimulated BH 4 production, we used C6 rat astroglioma cells, a convenient model astrocyte cell line. Furthermore, in this cell line, as in primary astrocytes, TNF-␣ has been shown to stimulate degradation of sphingomyelin to ceramide (22). Although ceramide generation, similar to TNF-␣ treatment potentiated NO and iNOS expression induced by IFN-␥ plus IL-1␤, surprisingly, we found that it did not mimic the effects of TNF-␣ on BH 4 or GTPCH expression. Our results thus suggest that BH 4 and NO biosynthesis can be differentially regulated in C6 cells.
Cell Culture-C6 cells were obtained from ATCC (Manassas, VA) and cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, penicillin (100 units/ml), and streptomycin (100 g/ml) (Sigma) at 37°C in a humidified atmosphere of 95% air, 5% CO 2 . All experiments were performed on a freshly thawed batch of cells that had been passaged only twice and then aliquoted and stored frozen in liquid nitrogen vapor. Prior to each experiment, cells were defrosted and grown in 175-cm 2 flasks until 80 -90% confluent, trypsinized, and then seeded at 100,000 cells/well in six-well plates (3 ml of medium). After 48 h, cells were serum-starved for 4 h prior to treatments. All cytokine solutions were prepared according to the manufacturers' instructions. Lipids were prepared as 10 mM solutions in methanol and stored at -70°C. C 2 -ceramide and C 2 -dihydroceramide were diluted in ethanol and added directly to the medium, maintaining a final ethanol concentration of Յ0.4% (v/v). Prior to use, an aliquot of sphingosine was dried under a gentle stream of nitrogen and then resuspended in 4 mg/ml fatty acid-free bovine serum albumin by probe sonication on ice. PDTC and SN-50 peptides were added 2 h prior to the addition of cytokines or lipids.
Nitrite Determination-Nitrite, which is the stable oxidation product of NO and an index of iNOS activity, accumulates in the medium and was measured essentially as described previously (9). In brief, after stimulation for the indicated times, 100 l of medium was mixed with 75 l of Griess reagent B (2% sulfanilamide in 1 M H 3 PO 4, w/v) and allowed to stand at room temperature for 5 min, after which 75 l of Griess reagent A (0.2% N-(1-naphthyl)ethylenediamine dihydrochloride in water (w/v)) was added. After a further 10 min, the absorbance was measured at 550 nm using an EL-340 Bio Kinetics microplate reader (Bio-Tek Instruments). Standard curves were generated with NaNO 2 added to the same medium.
Cell Lysates-Cells were washed with ice-cold phosphate-buffered saline (pH 7.4), detached by trypsinization, then pelleted in a microcentrifuge at 10,000 ϫ g for 3 min. Cell pellets were washed twice with 1 ml of ice-cold phosphate-buffered saline and resuspended in 250 l of extraction buffer containing 50 mM Tris, pH 7.4, 1 mM EDTA, 1 mM dithiothreitol. Cell suspensions were sonicated on ice with a fine-tipped probe sonicator for 15 s, lysates were cleared at 10,000 ϫ g for 3 min, supernatants were collected, and the protein concentrations were determined using Coomassie Plus reagent (Pierce).
BH 4 Determination-BH 4 was measured as described previously (9), with minor modifications. In brief, 50 l of cell lysate was diluted to 80 l with extraction buffer and then mixed with 20 l of 1 M H 3 PO 4 , 1.5 M HClO 4 (1:1, v/v). Approximately 10 mg of MnO 2 was added to oxidize reduced pterins to their fluorescent aromatic forms. After 20 min at room temperature, samples were centrifuged at maximum speed in a benchtop microcentrifuge for 5 min. Supernatants were removed and analyzed by reverse phase high performance liquid chromatography with fluorescence detection as described previously (23).
Determination of GTPCH Activity-GTPCH activity was measured essentially as described previously (24). In brief, to 30 l of lysate were added 5 l of 0.5 M Tris-HCl (pH 7.4), 5 l of 10 mM dithiothreitol, 5 l of 10 mg/ml bovine serum albumin, and 5 l of 10 mM GTP. Samples were incubated for 2 h at 37°C and placed on ice, and the reaction was terminated by the addition of 5 l of 1 M HCl, followed by 5 l of iodine reagent (1% I 2 /2% KI (1:1, w/v). After 20 min at room temperature in the dark, 5 l of 2% ascorbic acid (w/v) was added followed by 10 l of 2 M Tris base. Neopterin triphosphate was then dephosphorylated by incubation for 30 min at 37°C with 10 units of bovine intestinal alkaline phosphatase, followed by the addition of 50 l of 1 M H 3 PO 4 to terminate the reaction. Samples were cleared by centrifugation at maximum speed in a benchtop centrifuge for 10 min. Neopterin was measured by reverse phase high performance liquid chromatography with fluorometric detection (23). RT-PCR-Total RNA was isolated from confluent cultures with Trizol reagent (Life Technologies, Inc.) according to the manufacturer's directions. RNA (1 g) was converted to cDNA with random hexamers and Thermoscript reverse transcriptase according to the manufacturer's instructions (Life Technologies, Inc.). cDNA was amplified by PCR in a Perkin-Elmer 2400 thermal cycler using the following conditions: initial denaturation at 94°C for 3 min, followed by 30 cycles for iNOS and GTPCH feedback regulatory protein (GFRP), 32 cycles for GTPCH, and 25 cycles for actin (94°C for 45 s, 55°C for 45 s, and 72°C for 1 min). Final extension was at 72°C for 10 min. The following forward and reverse PCR primers were used (predicted product size): iNOS, 5Ј-CTGCAGGTCTTTGACGCTCGG-3Ј and 5Ј-GTGGAACACAGGGGT-GATGCT-3Ј (741 base pairs); GFRP, 5Ј-CAGATCCGTATGGAAGTGG-GTC-3Ј and 5Ј-CACCCCTGTCATGCTTAACAC-3Ј (195 base pairs); GTPCH, 5Ј-GGATACCAGGAGACCATCTCA-3Ј and 5Ј-TAGCATCCTG-CTAGTGACAGT-3Ј (372 base pairs); and actin, 5Ј-TTGTAACCAACT-GGGACGATATGG-3Ј and 5Ј-GATCTTGATCTTCATGGTGCTAGG-3Ј (743 base pairs). PCR products were resolved on 2% agarose gels containing ethidium bromide and visualized with UV fluorescence and a video camera, and bands were quantified with the National Institutes of Health Image program. Reaction conditions were optimized in preliminary experiments so that amplifications were within the logarithmic phase and yields were approximately linear with input cDNA concentration. To ensure that contaminating genomic DNA was not being amplified, PCR was also performed without reverse transcriptase treatment.

Determination of iNOS Activity-Aliquots
Western Analysis-Aliquots of cell lysates containing 20 g of protein were concentrated using chloroform:methanol:H 2 O phase partition. In brief, lysates were diluted to 450 l with H 2 O and mixed with 1 ml of chloroform:methanol (1:1) to give a final ratio of 1:1:0.9. The mixture was vortexed and then centrifuged in a benchtop microcentrifuge at maximum speed for 3 min to separate the phases. With this solvent combination, the proteins aggregate at the interphase. 700 l of the upper phase was removed without disturbing the interphase and discarded, and an equivalent volume of methanol was added back to the lower phase. The protein aggregates were pelleted at maximum speed for 10 min. Supernatants were carefully aspirated, and the pellets were dried at 50°C. Pellets were resuspended in 25 l of 1ϫ LDS NuPAGE sample buffer (NOVEX) containing 50 mM dithiothreitol and then heated at 100°C for 10 min. Proteins were resolved on 4 -12% Bis/Tris NuPAGE gels for 90 min at 175 V and transferred to polyvinylidene difluoride membranes (Millipore Corp., Bedford, MA) at 100 V for 60 min in NOVEX transfer buffer at 4°C. NuPAGE antioxidant (1:1000, v/v) was added to assist with transfer of high molecular weight proteins. Membranes were blocked with 10% (w/v) nonfat dry milk, 0.02% (v/v) azide, 0.05% (v/v) Tween 20 for at least 1 h at room temperature, or overnight at 4°C. After extensive washing with wash buffer (KPL), membranes were incubated with iNOS antibody (1:7500) for 2 h in milk diluent (KPL) and then with peroxidase-conjugated secondary antibody (1:3000) for 1 h. Membranes were extensively washed, and bands were visualized by enhanced chemiluminescence (NEN Life Science Products).

C 2 -ceramide and Bacterial SMase Potentiate IFN-␥/IL-1␤induced NO Production in C6
Cells-Combinations of proinflammatory cytokines, such as IFN-␥, IL-1␤, and TNF-␣, together with LPS, which induce NO production in astrocytes, also coordinately increase the synthesis of the NOS cofactor, BH 4 (25). Recently, it was shown that ceramide potentiated LPS and cytokine-induced NO and iNOS expression in rat primary astrocytes (20). Because TNF-␣ has been shown to stimulate hydrolysis of sphingomyelin to ceramide in many types of cells, including C6 cells (22), it was of interest to analyze the involvement of ceramide in TNF-␣-induced NO and BH 4 biosynthesis in these cells, as they express many of the properties of astrocytes.
In agreement with previous studies (26), the combination of IFN-␥, IL-1␤, and TNF-␣ evoked a marked stimulation of NO production as measured by nitrite accumulation to a level of 36 M in the medium after 24 h (Fig. 1A). TNF-␣ was unable to induce NO production by itself. Furthermore, in the absence of TNF-␣, the amount of nitrite produced by IFN-␥/IL-1␤ was dramatically lower (2.1 M). The cell permeable ceramide analogue, C 2 -ceramide, in a dose-dependent manner, or exogenous bacterial SMase, which hydrolyzes sphingomyelin to generate endogenous ceramide, mimicked the effect of TNF-␣, and potentiated IFN-␥/IL-1␤-induced NO production (Fig. 1). This appears to be a specific ceramide effect because other related sphingolipid metabolites, including sphingosine and the inactive ceramide analogue, C 2 -dihydroceramide, which has the same structure as C 2 -ceramide but lacks the double bond, did not replicate the effects of C 2 -ceramide or SMase (data not shown). In addition, C 2 -ceramide dose-dependently potentiated IFN-␥/IL-1␤-induced iNOS activity as measured in vitro with a maximum stimulation of more than 10-fold at a concentration of 7.5 M (Fig. 2A). Furthermore, neither C 2 -dihydroceramide nor sphingosine had any effect on IFN-␥/IL-1␤-induced iNOS activity (Fig. 2B). In agreement with its more potent ability to potentiate NO production, TNF-␣ notably enhanced IFN-␥/IL-1␤-induced iNOS activity by more than 40-fold. It should be noted that similar to TNF-␣ alone, treatment with C2-ceramide or SMase in the absence of IFN-␥/IL-1␤ had no significant effects on NO production or iNOS expression or activity. C 2 -ceramide Potentiates IFN-␥/IL-1␤-induced iNOS Expression and Protein-It was of interest to determine whether the stimulatory effect of TNF-␣ and ceramide on NO production and iNOS activity was due to an increase in iNOS expression. In agreement with previous reports, iNOS mRNA was not detectable by RT-PCR in untreated C6 cells (20,27) but was induced by IFN-␥/IL-1␤, and its expression was further enhanced by addition of TNF-␣ or C 2 -ceramide (Fig. 3A). The same pattern of responses was observed when iNOS protein levels were examined by immunoblotting (Fig. 3B). Furthermore, C 2 -ceramide dose-dependently increased iNOS mRNA and protein expression (Fig. 3, C and D). Thus, C 2 -ceramide is able to mimic the effects of TNF-␣ in potentiating IFN-␥/IL-1␤induced iNOS transcription, translation, and enzyme activity, albeit with less efficiency than TNF-␣.
As C 2 -ceramide potentiated IFN-␥/IL-1␤-induced iNOS activity and protein by 8 -10-fold yet increased NO production to a much smaller extent, it was possible that in vivo iNOS activity might be limited by the availability of its cofactor, BH 4 . In order to establish whether BH 4 levels were limiting for NO production, C6 cells were treated with 5 M sepiapterin, which is readily taken up and converted to BH 4 (28). Sepiapterin did not increase NO production in C6 cells treated with C 2 -ceramide and IFN-␥/IL-1␤ (data not shown). Thus, it is unlikely that iNOS activity in this case is limited by the intracellular concentration of BH 4 .
TNF-␣, but not Ceramide, Up-regulates BH 4 Synthesis and GTPCH Activity-Previously, we showed that cytokines and LPS induce both production of NO and de novo biosynthesis of BH 4 in astrocytes (25). TNF-␣ in the presence of IFN-␥/IL-1␤ markedly stimulated BH 4 biosynthesis in C6 cells by more than 5-fold over the effect of IFN-␥/IL-1␤ alone (Fig. 4). Changes in GTPCH activity mirrored the BH 4 increases (Fig. 4), in agreement with its role as the rate-limiting enzyme in BH 4 biosynthesis. Indeed, it is likely that the increased BH 4 results from increased de novo synthesis, rather than decreased catabolism, because the GTPCH inhibitor, 2,4-diamino-6-hydroxypyrimidine, blocked the cytokine-induced BH 4 increase (data not shown). Thus, it was of interest to determine whether ceramide, in a manner similar to its effects on NO production, mimicked the effects of TNF-␣ and mediated an increase in BH 4 synthesis in these cells. However, increasing ceramide levels by treatment of C6 cells with C 2 -ceramide or SMase did not potentiate the effects of IFN-␥/IL-1␤ on BH 4 biosynthesis or on GTPCH activity (Fig. 4).
It has previously been reported that proinflammatory cytokines in combination with LPS stimulate GTPCH mRNA expression in mouse osteoblasts (27) and in C6 cells (27,29). We next determined whether the stimulatory effect of TNF-␣ on BH 4 production and GTPCH activity in IFN-␥/IL-1␤-treated C6 cells was due to an increase in GTPCH mRNA expression. We found that GTPCH mRNA is constitutively expressed at low levels in C6 cells and that its expression is increased by IFN-␥/IL-1␤ and further enhanced by the addition of TNF-␣ (Fig.  5A). In agreement with their lack of effects on BH 4 and GTPCH activity (Fig. 4), neither C2-ceramide nor SMase had any significant effects on GTPCH mRNA expression (Fig. 5B). This is the first demonstration that GTPCH and iNOS expression can be differentially regulated.
As it was possible that there might have been a rapid and transient increase in GTPCH mRNA expression evoked by the addition of C 2 -ceramide or SMase to IFN-␥/IL-1␤-treated cells that might not be obvious after 16 h (Fig. 5), we examined a more complete time course for induction of both GTPCH and iNOS expression by semi-quantitative RT-PCR. TNF-␣ significantly increased GTPCH mRNA expression (normalized to actin expression) within 8 h in cells treated with IFN-␥/IL-1␤ (Fig. 6B). Expression then increased nearly linearly for at least another 8 h. Addition of TNF-␣ also increased the intracellular concentration of BH 4 in a time-dependent manner with a detectable increase as early as 8 h and increasing thereafter, whereas ceramide elevation did not result in BH 4 increases at any time point examined (data not shown). In agreement with their lack of effect on BH 4 levels and GTPCH activity (Fig. 4) D). A and C, RNA was isolated, and expression of iNOS and actin mRNAs was determined by RT-PCR. In B, total cell lysates (20 g) were subjected to SDS-PAGE, proteins were transferred to PVDF, and iNOS protein was measured by Western blotting using a specific iNOS antibody. Similar results were obtained in at least two additional independent experiments. treatment with SMase (Fig. 6B) or with C2-ceramide (data not shown) did not enhance GTPCH expression in cells treated with IFN-␥/IL-1␤ at any time point (Fig. 6A). In contrast, iNOS expression was rapidly increased by either TNF-␣, SMase or C2-ceramide (data not shown) in IFN-␥/IL-1␤-treated cells, and a near maximal stimulatory effect was observed within 8 h (Fig. 6A).
A potential mechanism of regulating de novo BH 4 biosynthesis independently of GTPCH expression that could result in increased BH 4 levels, is a decrease in BH 4 end product feedback inhibition. In some cell types, BH 4 inhibits GTPCH activity through the action of the GFRP, which forms a complex with GTPCH (24,30). Thus, cytokine-induced decreases in GFRP activity or expression might lead to an increase in GTPCH activity, and result in higher levels of BH 4 . However, although GFRP is expressed constitutively in C6 cells (Fig. 6C), its expression was not altered by cytokines, in the presence or absence of TNF-␣, throughout the 16 h time course, and it thus does not appear to be involved in the stimulation of GTPCH activity induced by TNF-␣.
TNF-␣ Regulates iNOS and GTPCH Expression by Distinct Signaling Pathways-Recently, antioxidants were shown to be potent inhibitors of cytokine-induced degradation of sphingomyelin to ceramide, suggesting that sphingomyelinase activa-tion is redox-sensitive (31). To examine the role of ceramide generation in iNOS expression, we utilized the antioxidant PDTC, which has been shown to inhibit ceramide generation in C6 cells induced by TNF-␣ (22). In agreement with previous results (32), PDTC completely blocked iNOS expression induced by TNF-␣. However, it had no effect on TNF-␣-stimulated GTPCH expression (Fig. 7B). PDTC also inhibits the release of the inhibitory IB subunit from the latent cytoplasmic form of NF-B, thereby blocking its transcriptional activity (33). Because expression of iNOS is regulated, at least in part, by NF-B (34), and because the stimulatory effect of ceramide on induction of iNOS is dependent on NF-B activation in astrocytes (20), we also examined the effects of the more specific NF-B inhibitor, SN-50. This peptide, which possesses a nuclear localization sequence that competes for the cellular machinery required for NF-B nuclear translocation (35), al- most completely inhibited TNF-␣-induced iNOS expression (Fig. 7A), whereas a control mutant SN-50 peptide had no effect. In sharp contrast, GTPCH mRNA levels in cytokinestimulated cells were unaffected by SN-50. Thus, TNF-␣ stimulates iNOS expression by a NF-B-dependent mechanism and GTPCH expression by a pathway that does not require activation of this transcription factor. DISCUSSION In this study, we have demonstrated that TNF-␣ stimulates iNOS and GTPCH expression, and therefore NO and BH 4 biosynthesis, by discrete pathways. The stimulatory effects on NO and iNOS levels by TNF-␣, which has previously been shown to increase ceramide levels in C6 cells (22), was mimicked by the short chain ceramide analogue, C 2 -ceramide, as well as bacterial SMase. Conversely, neither C 2 -ceramide nor bacterial SMase further enhanced IL-1␤/IFN-␥-induced BH 4 levels, even though TNF-␣ increased its levels by 5-fold. Furthermore, ceramide elevations, in contrast to TNF-␣, also had no effect on the GTPCH activity of IFN-␥/IL-1␤-stimulated C6 cells. Hence, ceramide can up-regulate iNOS without significantly affecting the levels of its cofactor BH 4 . To our knowledge, this is the first time that regulation of NO and BH 4 synthesis has been shown to diverge, as in many previous studies, elevations of BH 4 always mirrored induction of NO, suggesting common regulatory pathways (6).
The rat iNOS promoter contains consensus binding sites for numerous transcription factors (36). However, transcriptional regulation of iNOS is largely governed by the nuclear activity of the potent transcription factor NF-B, a DNA-binding protein that is activated by TNF-␣ and IL-1␤ in diverse types of cells (21,27,37). The iNOS promoter has two NF-B binding sites, a proximal site approximately 90 bases upstream of the initiation codon and a distal site located 980 bases upstream. The relative importance of these sites in regulating iNOS expression is still unclear because the NF-B-dependent pathways may vary depending on cell type and the particular combination of cytokines (38). Cytoplasmic NF-B can exist as either a p50/p65 heterodimer or as a p105/p65 heterodimer (39). The p50/p65 form is associated with a member of the inhibitory subunit family, IB, which is subject to cytokine-induced proteosomal degradation (40), allowing the p50/p65 heterodimer to translocate to the nucleus and bind to promoter target sequences. Alternatively, direct processing of a p105/p65 heterodimer to p50/p65 would also allow it to translocate into the nucleus.
In contrast, the GTPCH promoter has not been well charac-terized, although a portion of the mouse (41) and human (42) 5Ј-regulatory regions and, recently, 5.8 kilobases of the rat 5Ј flanking region (GenBank TM accession number AF131210 (57)) have been cloned. Interestingly, both murine and rat promoters have a conserved putative NF-B binding site located in a GC rich region 156 bases upstream of the initiation site. As this NF-B site has several overlapping potential transcription regulator binding sites, it is possible that the lack of effect of NF-B inhibitors on cytokine-stimulated GTPCH expression in C6 cells (Fig. 7) could be due to complex promoter effects. Indeed, simultaneous inhibition of activation of both NF-B and AP-1 transcription factors was required to block cytokineinduced GTPCH expression in mouse osteoblastic cells (27), suggesting that dual operation of both transcription factors is required. Further studies using GTPCH promoter-reporter constructs are underway in our laboratory and should help to clarify this issue. Previously, many studies have examined whether ceramide elevations mimic the effects of TNF-␣ on activation of NF-B. Ceramide analogues have been found to activate NF-B (43,44), to potentiate the activation in response to TNF-␣ (45), or to have no effect (46 -48). A recent study demonstrated that although ceramide does not appear to be involved in the pathway leading to IB phosphorylation and degradation, it signals p105 processing in response to TNF-␣ (49). In addition, SMase and ceramide analogues mimicked the stimulatory effects of TNF-␣ and IL-1␤ on iNOS transcription in vascular smooth muscle cells by promoting translocation of NF-B to the nucleus (21). Whereas cytokines induced degradation of the inhibitory IB subunit and maximally activated NF-B, SMase did not promote IB degradation but did enhance NF-B translocation to the nucleus and iNOS transcription, albeit with lower efficiency than cytokines (21). These results demonstrate an essential role of NF-B activation in mediation of neutral SMase-induced iNOS expression, distinct from proteosome-mediated IB-␣ degradation by TNF-␣, suggesting the possible involvement of an additional signaling pathway(s). Indeed, it is tempting to speculate that this accounts for the lower potency of C 2 -ceramide or SMase on stimulation of iNOS expression in C6 cells than observed with TNF-␣. However, it is also possible that the cellular location of ceramide could be crucial for the activation of downstream signaling pathways and for the ultimate biological response. Thus, ceramide formed from plasma membrane sphingomyelin may be targeted to different cellular compartments than short-chain ceramide analogs.
Our results confirm that NF-B plays an integral role in TNF-␣-induced expression of iNOS. On the other hand, GTP-CH expression appears not to be regulated in a NF-B-dependent manner, because the NF-B inhibitors PDTC and SN-50, which completely blocked iNOS expression, did not have a marked effect on GTPCH expression. However, LPS-induced GTPCH expression in rat vascular muscle cells (50) and IFN-␥/IL-1␤-induced BH 4 synthesis rat neonatal cardiac myocytes (32) have been shown to be NF-B-dependent. Thus, in different cell types, activation of different transcription factions might be important for regulation of GTPCH expression.
Discrete regulation of iNOS and GTPCH may be physiologically relevant, as BH 4 also functions as a cofactor for aromatic amino acid hydroxylations and may have other biochemical roles (51). We previously found that erythropoietic cells, despite the lack of any known BH 4 -dependent hydroxylation reactions, synthesized and contained high levels of BH 4 , which appears to play a regulatory role as a switch between growth and differentiation (52). Furthermore, we showed that proliferation of primary astrocytes was also regulated by endogenous BH 4 levels (53), results that were later confirmed in several other types of cells (54,55). In addition, ether lipid metabolism has been shown to be BH 4 -dependent (56). Thus, in some tissues and cells, parallel regulation of NO formation and BH 4 synthesis may not be required or desirable. Moreover, our results may have important implications for the development of novel, specific therapeutic approaches to specifically decrease aberrant levels of NO without affecting BH 4 .