Pteridine Biosynthesis in Human Endothelial Cells IMPACT ON NITRIC OXIDE-MEDIATED FORMATION OF

Stimulation of nitric oxide (NO) synthase in endothe- lial cells by Ca2+ influx leads to increased intracellular levels of cGMP. NO synthase from various sources is known to use tetrahydrobiopterin, flavins, and NADPH as cofactors. We studied the effect of inter-feron-?, tumor necrosis factor-a, and lipopolysaccha- ride on tetrahydrobiopterin biosynthetic activities in human umbilical vein endothelial cells (HUVEC). These stimuli led to an up to 40-fold increase of GTP cyclohydrolase I (EC 3.5.4.16) activity and to increased accumulation of neopterin and tetrahydrobi- opterin in HUVEC. Further enzyme activities of tetrahydrobiopterin biosynthesis, Le. 6-pyruvoyl tetrahy- dropterin synthase and sepiapterin reductase (EC 1.1.1.153), remained unchanged. NO synthase activity in protein fractions from homogenates of cells treated with interferon-y plus tumor necrosis factor-a was not influenced as compared with untreated controls. How- ever, interferon-y alone or in combination with tumor

Stimulation of nitric oxide (NO) synthase in endothelial cells by Ca2+ influx leads to increased intracellular levels of cGMP. NO synthase from various sources is known to use tetrahydrobiopterin, flavins, and NADPH as cofactors. We studied the effect of interferon-?, tumor necrosis factor-a, and lipopolysaccharide on tetrahydrobiopterin biosynthetic activities in human umbilical vein endothelial cells (HUVEC). These stimuli led to an up to 40-fold increase of GTP cyclohydrolase I (EC 3.5.4.16) activity and to increased accumulation of neopterin and tetrahydrobiopterin in HUVEC. Further enzyme activities of tetrahydrobiopterin biosynthesis, Le. 6-pyruvoyl tetrahydropterin synthase and sepiapterin reductase (EC 1.1.1.153), remained unchanged. NO synthase activity in protein fractions from homogenates of cells treated with interferon-y plus tumor necrosis factor-a was not influenced as compared with untreated controls. However, interferon-y alone or in combination with tumor necrosis factor-a significantly increased intracellular cGMP formation in intact HUVEC by 50 and 80%, respectively. These stimuli increased intracellular tetrahydrobiopterin concentrations up to 14-fold. NOtriggered cGMP formation was similarly increased by incubation of otherwise untreated cells with sepiapterin, leading to elevated intracellular tetrahydrobiopterin levels. Thus, cytokines indirectly stimulate the activity of constitutive NO synthase in HUVEC by upregulating production of the cofactor tetrahydrobiopterin.
Nitric oxide (NO)' synthase converts L-arginine to L-citrulline and NO in an NADPH-dependent reaction (1). This enzyme occurs in different isoforms, and the most prominent effects of NO are relaxation of smooth muscle, inhibition of platelet aggregation, cytotoxicity, and neurotransmission (for review, see Ref. 2). NO released from endothelial cells has been shown to account for the biological activity of endothelium-derived relaxing factor (3,4), which leads to smooth *This work was supported by Austrian Research Funds "Zur Forderung der wissenschaftlichen Forschung," Project P 8231. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
One of the metabolic activities strongly stimulated by cytokines like interferon-y or tumor necrosis factor-a is the de novo synthesis of tetrahydrobiopterin in human (27-29) and murine cells (30). In a previous paper we have shown that the increased amount of tetrahydrobiopterin produced due to cytokine action is required for the high and long term synthesis of NO by cytokine-treated murine fibroblasts (24). The possible impact of cytokine-induced pteridine formation on NO synthesis in human cells remained open, however, since NO synthesis is not inducible in cultured human cells by protocols successful in murine cells and cell lines.
In the present study we show that human umbilical vein endothelial cells (HUVEC) express high activities of GTP cyclohydrolase I (EC 3.5.4.16), the first enzyme of the tetrahydrobiopterin biosynthetic pathway, when treated by stimuli such as interferon-y, tumor necrosis factor-a, or lipopolysaccharide. Due to the constitutive presence of 6-pyruvoyl tetrahydropterin synthase and sepiapterin reductase (EC 1.1.1.153), the two subsequent biosynthetic enzymes (for review, see Ref. 31), this stimulation leads to the accumulation of neopterin and tetrahydrobiopterin in the cells. We present evidence that increased intracellular tetrahydrobiopterin levels resulting from cytokine treatment significantly increase NO-mediated generation of cGMP in HUVEC.
HPLC System-For quantitation of pteridines, an HPLC apparatus consisting of a liquid chromatograph (LC 5500) (Varian Associates, Inc., Palo Alto, CA), an LS 4 fluorescence detector (Perkin-Elmer, Beaconsfield, United Kingdom), and an AASP module (Varian) for direct insertion of solid-phase cation-exchange cartridges (Varian) was used. Fluid connections of the AASP instrument were modified as described (33). The reversed phase columns used were Lichrosorb R P 18 columns with 7 pm particle size (Merck).
Cell Culture-HUVEC were obtained from Technoclone (Vienna, Austria) in the first passage. Cells were cultured on polystyrene surfaces coated with 0.1% (w/v) gelatin, using Medium 199 with Earle's salts, supplemented with 20% (v/v) heat-inactivated fetal calf serum, 2 mM L-glutamine, 100 units/ml penicillin, 100 pg/ml streptomycin, and 50 pg/ml endothelial cell growth supplement. For experiments, cells of passages 3-8 were used, and stimulations of confluent monolayers (180,000 ? 20,000 cells/well, mean & S.D. of 6 wells) were carried out in the absence of endothelial cell growth supplement.
Cytokine Treatment-Various doses of interferon-y (50-1000 units/ml), tumor necrosis factor-a (250 and 1000 units/ml), and a high dose of lipopolysaccharide (1 pg/ml), which is highly effective in stimulation of GTP cyclohydrolase I in a human monocytic cell line (34), were tested for their capacity to induce neopterin and biopterin release from HUVEC into the supernatant as well as for their effect on cell viability after 24, 48, and 72 h (see below). Interferon-y alone and in combination with a second stimulus reduced cell numbers by u p to 50% after 72 h of treatment. Based on these pilot experiments and on previous work with other human and murine cell types (28, 30, 34), we used interferon-y and tumor necrosis factor-a a t 250 units/ml and an incubation time of 24 h. Under these conditions a strong and significant increase of pteridine synthesis together with a tolerable loss of adherent cells (see "Results") was observed.
Determination of Pteridines and Tetrahydrobiopterin Biosynthetic Enzyme Activities in Cell Extracts-Confluent monolayers grown in 75-cmZ culture flasks were treated with cytokines or lipopolysaccharide, applied either as single stimuli or as a combination of interferony with tumor necrosis factor and lipopolysaccharide, respectively. After treatment for 24 h, cells were harvested by trypsinization, washed in phosphate-buffered saline, pH 7.4, containing 130 mM NaC1,2 mM KCl, 6 mM NazHPO,, and 1 mM KH2P04, and disrupted by rapid freezing and thawing in distilled water. Supernatants were collected after centrifugation a t 10,000 X g for 10 min and immediately subjected to determination of intracellular pteridines and enzyme activities.
Intracellular concentrations of pteridines were measured as described previously (33), using HPLC with fluorescence detection after oxidation with iodine in acidic or basic media according to Fukushima and Nixon (35). In some experiments, cells were treated for 24 h with increasing concentrations of sepiapterin or with DAHP, and intracellular pteridine concentrations were determined. Concentrations are given as pmol/106 cells, corresponding to about 300 pg of total protein.
Assays for determination of GTP cyclohydrolase I, 6-pyruvoyl tetrahydropterin synthase, and sepiapterin reductase activities were performed as detailed previously (28), using protein fractions purified over Sephadex G-25. Briefly, GTP cyclohydrolase I was determined according to Viveros et al. (36), using 2 mM GTP as a substrate in the presence of 5 mM EDTA. Enzyme activity is given as pmol of neopterin formed per mg of protein/min. 6-Pyruvoyl tetrahydropterin synthase was measured after incubation of 7&dihydroneopterin triphosphate, freshly prepared by G T P cyclohydrolase I purified from E. coli, in the presence of excess sepiapterin reductase, NADPH, and M e . Thus, the resulting 6-pyruvoyl tetrahydropterin is converted into tetrahydrobiopterin, which is then determined as the fluorescent biopterin by HPLC after iodine oxidation a t acidic pH. 6-Pyruvoyl tetrahydropterin synthase activity is given as pmol of biopterin/mg of protein/min. Sepiapterin was used as substrate for sepiapterin reductase determination, and the resulting tetrahydrobiopterin was oxidized to biopterin. Sepiapterin reductase activity is given as pmol of biopterin formed per mg of protein/min.
Protein determination was done according to Bradford (37), using bovine serum albumin as a standard.
Determination of Pteridines and Nitrite in Supernatants-HUVEC were grown in 24-well plates and treated with interferon-y, tumor necrosis factor-a, or lipopolysaccharide for 72 h. Supernatants were then collected and assayed for total neopterin and biopterin (after iodine oxidation a t acidic pH), using HPLC reversed phase chromatography (33). Concentrations given are nmol/liter accumulated in the supernatant after 24 h. Due to supplementation with fetal calf serum, culture medium contained 17.3 f 2.9 nmol/liter biopterin (mean of six determinations k S.D.). This blank value was subtracted for calculation of biopterin accumulation in the supernatants.
Nitrite was determined in these supernatants by the Griess reaction (38) using the stable Griess-Ilosvay's reagent from Merck. The detection limit is 1 pmol/liter.
Determination of Intracellular cGMP Levels-Determination of NO-triggered intracellular cGMP was modified from Schmidt et al. (39). Briefly, confluent monolayers of HUVEC, grown in 6-well plates and in some cases pretreated for 24 h with various concentrations of sepiapterin, of DAHP, or with cytokines and/or lipopolysaccharide, were washed with 10 mM Hepes buffer, pH 7.4, containing 2.5 mM CaC12, 1 mM MgC12, 5 mM KC1, and 145 mM NaC1. Monolayers were then incubated for 15 min with 1 mM 3-isobutyl-1-methylxanthine at 37 "C and then treated with 1 p~ A23187 or vehicle in a final volume of 1 ml for 6 min. S N P (1 mM) was used for directly stimulating guanylyl cyclase without involvement of endogenous NO synthase. The reaction was stopped by removing the supernatant and adding 1 ml of 0.01 M HCl. Intracellular cGMP was extracted for 1 h at 4 "C and then quantified by radioimmunoassay with acetylation of the sample for maximal sensitivity. The detection limit is 20 pmol/liter. Measurement of NO Synthase Activity in Cell Extracts-NO synthase was assayed as modified from Stuehr et al. (14) and Mayer et al. (16). Cell extracts from HUVEC (untreated or treated with interferon-y) plus tumor necrosis factor-a, 250 units/ml each, for 24 h) were prepared by rapid freezing of cells in distilled water. After thawing and centrifugation a t 10,000 X g for 10 min, the supernatant was collected and Sephadex G-25 eluates were prepared using 40 mM Tris-HC1, pH 8.0, containing 100 p~ phenylmethylsulfonyl fluoride. The pellet was washed twice with phosphate buffer, pH 7.4 (see above), and then treated with 20 mM CHAPS in 40 mM Tris-HC1, pH 8.0, with phenylmethylsulfonyl fluoride for 20 min on ice with gentle shaking in order to solubilize the particulate NO synthase from endothelial cells (18). After centrifugation at 10,000 X g for 10 min, the supernatant was purified from low molecular mass compounds by Sephadex G-25 and eluted with the above described Tris buffer. Incubation was then carried out in 40 mM Tris-HC1, pH 8.0, containing 100 p M L-arginine, 25 p M FAD, 25 p M FMN, 2 mM NADPH, 5 p~ (6R)-tetrahydrobiopterin, 100 p~ phenylmethylsulfonyl fluoride, 60,000-80,000 cpm of purified L- [2,3,4,5-3H]arginine (see "Materials"), and 100 pl of cell extract or CHAPS extract, containing about 300 pg of total protein. The free Ca2+ concentration was adjusted in these incubation mixtures using 0.15 mM EGTA, 0.9 mM EDTA, 2.05 mM MgC12 without CaC12 for Ca2+-free conditions, and 0.15 mM EGTA, 0.9 mM EDTA, 1.78 mM MgC12, and 0.27 mM CaClz for a free Ca2+ concentration of 3 p M (40). The final incubation volume was 200 pl, and incubation was carried out a t 37 "C for 30 min. The reaction was then stopped by 800 pl of 20 mM sodium acetate, pH 5.0, containing 200 p~ EDTA and 1 mM L-citrulline. ['HI Citrulline was quantified after separation from [3H]arginine by the cation exchanger Dowex 50W. NO synthase activity is given as pmol of [3H]citrulline formed per mg of protein/min.
Cell Viability Assay-This assay is based on staining adherent cells after incubation with various stimuli. Since anchorage-dependent cells do not survive or proliferate after they have detached from the surface, it is generally agreed that such an assay reasonably estimates the amount of viable cells. Briefly, HUVEC were grown in 96-well plates and treated with sepiapterin, DAHP, cytokines, and lipopolysaccharide for 24 h. Supernatants were then removed, and cells were fixed with 5% (v/v) formaldehyde for 10 min. After washing with tap water, cells were stained with 1% (w/v) of crystal violet for 5 min according to standard procedures. After vigorous washing with tap water, the dye was solubilized with 30% (v/v) acetic acid (100 pl/ well), and the optical density at 560 nm was read. Table I summarizes the activities of the three enzymes involved in the biosynthesis of tetrahydrobiopterin from GTP, as well as the intracellular concentrations of pteridines following the treatment of HUVEC with different activators. As shown in Table I, 250 units/ml interferon-y led to about a 20-fold increase of GTP cyclohydrolase I activity, when applied as a single stimulus, and to a 46-fold increase when applied in combination with 250 units/ml tumor necrosis factor-a. This was not further enhanced by higher doses (up to 1000 units/ml) of interferon-y. Tumor necrosis factor-a alone (250 units/ml) induced GTP cyclohydrolase activity 5fold ( p < 0.0001, Student's t test). Lipopolysaccharide (1 pg/ ml) stimulated GTP cyclohydroase I activity about 13-fold. Lipopolysaccharide applied in combination with interferon-y had only an additive effect but did not potentiate the interferon-y stimulus like tumor necrosis factor-a. Further enzyme activities of tetrahydrobiopterin synthesis, i.e. 6-pyruvoyl tetrahydropterin synthase and sepiapterin reductase, were not significantly (Student's t test) changed upon treatment with these stimuli. In parallel to the stimulation of G T P cyclohydrolase I activity, neopterin and biopterin derivatives accumulated in the cells. At least 80% of biopterin occurred in its tetrahydro form, as was calculated from the difference of biopterin detected after oxidation with iodine in alkaline or acidic media (35).

RESULTS
Neopterin and biopterin were also released from the cells but were not detectable in supernatants of untreated controls or of cells treated with tumor necrosis factor-a alone (see Table I). No nitrite accumulation (<1 pmollliter) could be observed in supernatants of these long term treated cultures indicating the lack of a macrophage type, Ca'+-independent, high output NO synthase activity.
We then investigated the effect of manipulating intracellular levels of tetrahydrobiopterin by incubating otherwise untreated cells with various concentrations of sepiapterin or DAHP for 24 h. Sepiapterin is intracellularly converted into tetrahydrobiopterin via a salvage pathway (31). The GTP cyclohydrolase I inhibitor DAHP inhibits de nouo synthesis of tetrahydrobiopterin by acting as an analogue of the first pyrimidine intermediate formed in the GTP cyclohydrolase I reaction subsequent to opening the GTP ring structure in position 8 (41). The correlation between cGMP levels accumulated after stimulation of Ca2+ influx by A23187 and intracellular tetrahydrobiopterin levels is shown in Fig. 1. The amount of tetrahydrobiopterin found in untreated control cells (C in Fig. 1) was 3.9 f 0.5 pmol/106 cells. No biopterin was detectable with 5 mM DAHP, and in this case cGMP formation was also reduced to undetectable levels. A 50% inhibition of tetrahydrobiopterin biosynthesis was achieved by about 0.3 mM DAHP. Intracellular tetrahydrobiopterin concentrations were strictly correlated to the sepiapterin dose applied (linear correlation coefficient 0.999). As little as 0.96 f 0.15 pmol of tetrahydrobiopterin/106 cells, achieved by 1 mM DAHP, were still sufficient to detect 0.57 f 0.08 pmol of cGMP/106 cells. Like the Ca'+ ionophore-stimulated cGMP formation (Fig. l), basal NO-triggered cGMP formation was also increased from 0.52 f 0.02 to 0.96 f 0.09 pmol/106 cells (mean f S.D. from triplicate cultures, p < 0.002, Student's t test) upon treatment with 250 pM sepiapterin and decreased to 0.06 f 0.02 pmol/106 cells by 0.5 mM DAHP. Table I1 shows that DAHP and sepiapterin specifically influenced NO synthesis rather than guanylyl cyclase itself, since SNP-stimulated guanylyl cyclase activity was not affected by these components. Further, the effect of DAHP could be restored by concurrent application of sepiapterin. The arginine analogue NMMA decreased NO-triggered cGMP formation, and this inhibition could not be overcome by sepiapterin. Interferon-? increased basal and A23187-stimulated NO-triggered cGMP formation by 50% ( p < 0.01, Student's t test) but did not alter guanylyl cyclase itself. The action of DAHP, sepiapterin, and NMMA was not influenced by interferon-y treatment. Interferon-y in combination with tumor necrosis factor-a increased NO-triggered cGMP formation by about 80% ( p < 0.002 as compared with untreated controls, p < 0.05 as compared with interferon-y alone, Student's t test) ( Table 11).
As compared with untreated control cells, a 24-h treatment of HUVEC with 100 p~ sepiapterin, 5 mM DAHP, 100 pM sepiapterin plus 5 mM DAHP, 250 units/ml interferon-y, 250 units/ml tumor necrosis factor-a, or 1 pglml lipopolysaccharide did not affect cell viability. In contrast, a combination of FIG. 1. Dependence of cGMP formation on intracellular tetrahydrobiopterin levels. Confluent monolayers of HUVEC were grown in 6-well plates for cGMP determination and pteridine determination in the case of sepiapterin treatment and in 75-cm2 culture flasks for pteridine determination in the case of DAHP treatment. Cells were treated with various doses of either DAHP or sepiapterin for 24 h. Formation of cGMP was determined upon stimulation of NO synthase with 1 p~ A23187 for 6 min (see "Experimental Procedures"). Cells of the same passage treated in parallel in the same experiment were collected by trypsinization, and intracellular biopterin levels were assayed in cell extracts by HPLC after oxidation in acidic or alkaline media as described under "Experimental Procedures." Levels of intracellular cGMP were plotted against the logarithm of intracellular tetrahydrobiopterin levels (means f S.D. of triplicate cultures). DAHP: A, 1 mM; B, 0.5 mM; C, untreated control. sepiapterin: D, 1 pM; E, 5 p~; F, 25 pM; G, 100 pM; H, 250 pM.

TABLE I1
Intracellular cGMP levels in HUVEC: the effect of interferon-y, DAHP, sepiapterin, and NMMA Confluent monolayers of HUVEC grown in 6-well plates were treated with various additives for 24 h. Supernatants were then removed, and basal cGMP levels as well as cGMP accumulated after treatment with 1 p M A23187 or 1 mM SNP for 6 min were determined as described under "Experimental Procedures." Values are means f S.D. from triplicate cultures from one of two similar experiments. IFN-y, interferon-y (250 units/ml); TNF-a, tumor necrosis factor-a (250 unitdml): nt. not tested. interferon-y with tumor necrosis factor-a or lipopolysaccharide reduced cell viability by 25 f 5% and by 32 k 7%, respectively (mean k S.D. of 8 wells from two experiments). Unlike in murine fibroblasts (24), concurrent treatment with 5 mM DAHP, 100 pM sepiapterin, or 250 pM NMMA did not influence the cytotoxic effect of these combined stimuli, thus excluding involvement of NO synthase in the observed cytotoxicity. NO synthase activity in protein fractions from homogenates of cells treated with interferon-y plus tumor necrosis factor-LY (250 units/ml each) was not higher than in those of untreated control cells. NO synthase activities detected were

DISCUSSION
In the present investigation, we describe the effects of interferon-y, tumor necrosis factor-a, and lipopolysaccharide on pteridine biosynthesis and the resulting accumulation of neopterin and tetrahydrobiopterin in HUVEC with respect to NO-mediated cGMP formation in intact cells. We found that interferon-y strongly stimulated the biosynthesis of tetrahydrobiopterin in HUVEC by increasing the activity of the key enzyme GTP cyclohydrolase I. Parallel accumulation of neopterin can be explained by the comparatively low 6-pyruvoyl tetrahydropterin synthase activity typical for human cells (28,30). Pteridine synthesis is also stimulated by tumor necrosis factor-a and lipopolysaccharide, and tumor necrosis factor-a potentiates the effect of interferon-y. Both neopterin and biopterin derivatives leak from cells treated with those cytokines or lipopolysaccharide and are detectable in supernatants. Thus, the previously reported increased supernatant concentrations of neopterin (42) in HUVEC cultures treated by cytokines are an indicator of increased intracellular de nouo synthesis of tetrahydrobiopterin in these cells.
Regarding endothelial cells, high output NO synthase leading to accumulation of nitrogen oxides in supernatants has been demonstrated thus far only for murine endothelial cells from brain (43). A cytokine-inducible, Ca2+-independent NO synthase was also reported for porcine aortic endothelial cells, as determined by inhibition of platelet aggregation and by spectrophotometric detection of NO (44). In tumor necrosis factor-a-treated bovine aortic endothelial cells, an up to 5fold increase of intracellular cGMP levels was observed, suggesting induction of NO synthesis (45). As expected from work with other human cells in uitro (for discussion of this phenomenon, see Refs. 46 and 47), high output NO synthase leading to accumulation of nitrite in supernatants could not be induced in HUVEC. NO-mediated cGMP formation in intact HUVEC, however, is significantly increased by interferon-y alone or in combination with tumor necrosis factora. Furthermore, our results obtained by manipulating otherwise untreated cells with drugs altering the intracellular tetrahydrobiopterin concentrations (Fig. 1) show that the amount of cGMP formed can be varied about 10-fold by decreasing and increasing the intracellular tetrahydrobiopterin concentration. In particular, increasing intracellular tetrahydrobiopterin levels with sepiapterin to levels comparable with cells treated with interferon-y alone or together with tumor necrosis factor-a also leads to a comparable increase in cGMP formation. Under these conditions, activity, Ca2+ dependence, and intracellular localization of NO synthase in cell homogenates remain unchanged. We conclude, therefore, that the increase of cGMP formation in HUVEC following cytokine treatment is due to the increased endogenous synthesis of tetrahydrobiopterin, one of the cofactors of NO synthase.
Using N-15-labeled arginine, an up to 10-fold increase in turnover from arginine to nitrogen oxides, indicating induction of NO synthase activity, has been demonstrated in humans receiving interleukin 2 therapy (47). In these patients interleukin-2 causes NO production most likely due to induction of cytokines such as interferon-y or tumor necrosis Pteridines, Nitric Oxide, and c( factor-a (for discussion, see Ref. 47). Interleukin-2 therapy is also known to lead to increased pteridine synthesis in humans, as measured by increased concentrations of neopterin in body fluids (49). Highly elevated formation of pteridines in humans has also been observed in a number of clinical situations (for review, see Ref. 49) including septic complications (50). Our study demonstrates that manipulation of intracellular tetrahydrobiopterin levels by certain drugs and by cytokines such as interferon-? and tumor necrosis factor-a can stimulate an NO-mediated cGMP increase in HUVEC. Thus, mechanisms regulating tetrahydrobiopterin biosynthesis may contribute to some extent to the observed stimulation of NO synthesis by cytokines in humans in uiuo (47).