Tetrahydrobiopterin, but Not l-Arginine, Decreases NO Synthase Uncoupling in Cells Expressing High Levels of Endothelial NO Synthase

In the vasculature, NO is generated by endothelial NO synthase (eNOS), where it regulates vascular tone (reviewed in Reference¹) and affects endothelial transcription.² Reactive oxygen species (ROS) play a role in signal transduction and are involved in the regulation of the biologically effective concentration of NO.³ In vascular disease states, excessive production of ROS may overwhelm the antioxidant defense mechanisms of cells, resulting in oxidative stress.⁴ Interestingly, eNOS itself can produce superoxide, a process referred to as “eNOS uncoupling.”⁵ Reduced levels of BH₄ or l-arginine lead to uncoupling of reduced nicotinamide-adenine dinucleotide phosphate (NADPH) oxidation and NO synthesis, with oxygen as terminal electron acceptor instead of l-arginine, resulting in the generation of superoxide (O₂^·−) by eNOS.^1,6,7

The antiatherogenic actions of NO on the vessel wall suggest that increasing eNOS expression may inhibit the development of atherosclerosis. Indeed, short-term studies on upregulation of eNOS by gene transfer of recombinant eNOS demonstrated beneficial effects in atherosclerotic arteries.^8–10 However, longer-term high expression of eNOS may not be as beneficial. In prehypertensive spontaneously hypertensive rats (SHR), increased eNOS expression has been observed,¹¹ but this was associated with decreased NO release and increased superoxide production in aortic tissue, suggesting the presence of eNOS uncoupling, which may contribute to the development of hypertension and its vascular complications in the SHR.¹² Moreover, studies in apolipoprotein (apoE)-deficient mice showed that chronic overexpression of eNOS accelerated atherosclerosis, which was associated with lower NO production relative to eNOS expression and enhanced superoxide production in the endothelium.¹³ Supplementation of BH₄ in these mice reduced atherosclerotic lesion size, suggesting that in these hypercholesterolemic mice, reduced BH₄ availability is involved in eNOS dysfunction during chronic eNOS overexpression.

We hypothesized that under conditions of sustained high expression of eNOS, acute stimulation of eNOS induces uncoupling of the eNOS enzyme because of a relative shortage of substrate and/or cofactors with superoxide production as a consequence. bEnd.3 cells¹⁴ chronically express high levels of eNOS protein and produce large amounts of NO in comparison with primary endothelial cells [eg, human umbilical vein endothelial cells (HUVECs)] or a human microvascular endothelial cell line (CDC.HMEC-1) while retaining the functional properties of endothelial cells. Use of bEnd.3 cells facilitates the detection of subtle differences in NO production as a consequence of treatment with agonists or antagonists. In this study, we addressed the following questions: (1) Does eNOS uncoupling occur in bEnd.3 cells?; (2) Does stimulation of bEnd.3 cells with vascular endothelial growth factor (VEGF) enhance eNOS uncoupling?; and (3) If this uncoupling occurs, is the uncoupling of eNOS because of a shortage of l-arginine, BH₄, or both?

Materials and Methods

Materials

All of the drugs were purchased from Sigma, except recombinant human VEGF₁₆₅ (Peprotech), diethylenetriamine NONOate (DETA/NO; Cayman Chemical), H₂O₂ (Merck), and BH₄ (Schircks Laboratories).

Cell Culture

An immortalized bEnd.3 cell line¹⁴ was generously provided by Dr Alan Schwartz (University of Washington, St Louis, MO). Cells were cultured at 37°C in humidified 95% air-5% CO₂ in DMEM supplemented with 10% FCS, 2 mmol/L glutamine, 100 IU/mL penicillin, and 100 IU/mL streptomycin (Life Technologies).

HUVECs were harvested from freshly obtained umbilical cords by use of the method described by Jaffe et al.¹⁵ The cells were cultured in fibronectin-coated T-flasks with EBM-2 (Bio-Whittaker) supplemented with 0.4% human fibroblast growth factor B, 0.1% human endothelial growth factor, 0.1% ascorbic acid, 0.1% gentamicin sulfate-amphotericin-B, 0.1% VEGF, 0.1% recombinant long R insulin–like growth factor, 0.1% heparin, 0.04% hydrocortisone (all supplements from Bio-Whittaker), 2% FCS (Life Technologies), 100 IU/mL penicillin, and 100 IU/mL streptomycin.

A human microvascular endothelial cell line (CDC.HMEC-1)¹⁶ was generously provided by Dr Edwin Ades, Francisco J. Candal (Centers for Disease Control and Prevention/National Center for infectious Diseases, Atlanta, GA), and Dr Thomas Lawley (Emory University, Atlanta, GA). CDC.HMEC-1 were cultured at 37°C in humidified 95% air-5% CO₂ in MCDB 131 (Life Technologies) supplemented with 10% FCS, 10 ng/mL human endothelial growth factor, 0.05 μmol/L hydrocortisone, 10 mmol/L glutamine, 100 IU/mL penicillin, and 100 IU/mL streptomycin.

Measurement of NO: 4,5-Diaminofluorescein Diacetate Assay

To measure intracellular NO production, the cell-permeable fluorescent NO indicator 4,5-diaminofluorescein diacetate (DAF-2DA; Calbiochem) was used. bEnd.3 cells were grown to confluence in a black clear-bottom 96-well plate and serum deprived for 16 hours in DMEM containing penicillin/streptomycin and 0.1% BSA. BH₄, tetrahydroneopterin (NH₄), and apocynin were incubated for 16 hours during starvation of the cells. All of the solutions were prepared in Tris buffer [200 mmol/L Tris-HCl, 10 mmol/L CaCl₂, 10 mmol/L MgCl₂, 1.33 mmol/L NaCl, 65 mmol/L KCl, 1% (wt/vol) d-glucose, 0.1% (wt/vol) BSA, and 50 μmol/L l-arginine (pH 7.4)]. Cells were washed and incubated with 5 μmol/L DAF-2DA for 40 minutes at room temperature in the dark. After incubation, cells were washed twice and incubated for 20 minutes at 37°C in the presence or absence of inhibitors or scavengers. VEGF was added, after which fluorescence was measured every 2 minutes for 70 minutes (excitation wavelength, 485 nm; emission wavelength, 538 nm; Fluoroskan Ascent, Labsystems).

To determine the reactivity of DAF-2DA toward NO, DAF-loaded bEnd.3 cells were exposed to the exogenous NO donor DETA/NO (1 to 10 μmol/L), and fluorescence was measured. At the highest concentrations of DETA/NO, the cells were still viable as determined by Trypan blue staining (data not shown). The fluorescent signal increased linearly over time (Figure 1A). DETA/NO concentration dependently increased the DAF signal as compared with basal NO production in bEnd.3 cells (Figure 1b), demonstrating time- and concentration-dependent linearity for the DAF assay of NO production.

• Download figure
• Download PowerPoint

Treating bEnd.3 cells with H₂O₂ (1 to 10 μmol/L) increased fluorescence (Figure 1C); however, N_ω-nitro-l-arginine methyl ester (l-NAME) attenuated the increase in DAF signal, implying H₂O₂-induced eNOS activation rather than direct oxidation of DAF-2 by H₂O₂.

Measurement of ROS: Diacetyldichlorofluorescein Diacetate Assay

Intracellular ROS were measured using diacetyldichlorofluorescein diacetate (CM-H₂DCFDA; Molecular Probes), a nonfluorescent cell–permeable indicator for ROS. Cells were seeded in a clear 96-well plate and treated as in the DAF-2DA assay. All of the solutions were made in PBS with additions [1 mmol/L CaCl₂, 0.5 mmol/L MgCl₂, 0.1% (wt/vol) d-glucose, and 50 μmol/L l-arginine]. Cells were washed and incubated with 10 μmol/L CM-H₂DCFDA for 30 minutes at 37°C in the dark. Interventions and measurements were done in a fashion comparable to the DAF-2DA assay.

Reactivity of CM-H₂DCFDA toward H₂O₂ was determined by exposing bEnd.3 cells to 1 to 10 μmol/L H₂O₂. H₂O₂ induced a time- and concentration-dependent increase in the CM-H₂DCFDA assay (Figure 1D). Treating bEnd.3 cells with DETA/NO (0.1 to 10 μmol/L) did not affect the DCF signal (Figure 1E), indicating that CM-H₂DCFDA is not reactive toward NO.

Measurement of Biopterin Levels in Cell Lysates

Biopterin levels in lysates of bEnd.3 cells or CDC.HMEC-1 were determined as described previously.¹⁷ Briefly, cell pellets from T-flasks were lysed in cold extract buffer [50 mmol/L Tris-HCl, 1 mmol/L DTT, 1 mmol/L EDTA, and 0.4 μmol/L 6,7-dimethylpterine (pH7.4)]. Protein concentration was measured using the Pierce BCA protein assay. The whole procedure was performed in the dark. Proteins were removed by adding 10 μL of a 1:1 mixture of 1.5 mol/L HClO₄ and 2 mol/L H₃PO₄ to 90 μL of extracts followed by centrifugation. Total biopterins [BH₄, 7,8-dihydro-l-biopterin (BH₂), and biopterin] were determined by acid oxidation. Therefore, 10 μL of 1% iodine in 2% KI solution was added to 90 μL protein-free supernatant. BH₂ and biopterin were determined by alkali oxidation by adding 10 μL of 1 mol/L NaOH to 80 μL of extract followed by 10 μL of iodine/KI solution. Samples were incubated at room temperature for 1 hour. Alkaline-oxidation samples were acidified with 20 μL of 1 mol/L H₃PO₄. Iodine was reduced by adding 5 μL of fresh ascorbic acid (20 mg/mL).

Pterines were measured by high-performance liquid chromatography (HPLC) on a Waters 600E HPLC (Etten-Leur). The HPLC was equipped with a Pontisil ODS 10 μm column (Alltech Associates Inc). A linear gradient was used for elution of the pterines [80% A [50% MeOH in H₂O (v/v)]] and 20% B (H₂O) in 10 minutes. After 10 minutes the column was washed with H₂O for 20 minutes. Fluorescence detection (360 nm excitation and 435 nm emission) was performed using a Waters 2475 Multi Lambda Fluorescence Detector. BH₄ concentration, expressed as picomoles per milligram of protein, was calculated by subtracting BH₂+biopterin from total biopterins.

Protein Measurements

To confirm high expression of eNOS and rule out expression of inducible NO synthase (iNOS) or neuronal NOS, bEnd.3 cells, HUVECs, and CDC.HMEC-1 were lysed [20 mmol/L Tris-HCl, 150 mmol/L NaCl, 1 mmol/L EDTA, 1 mmol/L EGTA, 1% Triton-X100, and protease inhibitors (MiniComplete, Roche Diagnostics Corporation; pH 7.4)]. Cell extracts and positive control protein supplied with antibodies were subjected to SDS-PAGE and transferred to polyvinylidene fluoride membrane. Blots were incubated with monoclonal antibodies against eNOS, nNOS (Transduction Laboratories), and iNOS (Alexis) and subsequently with a horseradish peroxidase–conjugated rabbit anti-mouse antibody (Jackson Immunochemicals). Membrane-bound antibodies were visualized using chemiluminescence reagent (Roche).

Expression of eNOS was also determined using an ELISA. A polyclonal rabbit antibody raised against a recombinant fragment spanning residues 1 to 67 of bovine eNOS (unpublished data) was bound to microtiter plates (NUNC-Immuno Plate Maxisorb Surface, NUNC) by incubation at 4°C. Plates were washed 3 times in Tris buffer [50 mmol/L Tris-HCl and 150 mmol/L NaCl (pH7.4)] with 0.1% Tween 20 and then incubated with 3% BSA in Tris buffer with 0.05% Tween 20 for 2 hours at room temperature. Samples were incubated in a total volume of 50 μL for 2 hours at room temperature. Plates were thoroughly washed 3 times and incubated for 2 hours at room temperature with 100 μL of 125 ng/mL eNOS antibody (Transduction Laboratories) in Tris buffer with 0.1% BSA and 0.05% Tween 20. Subsequently, plates were incubated with 650 μg/mL horseradish peroxidase–conjugated rabbit anti-mouse antibody (Jackson Immunochemicals) in Tris buffer with 0.1% BSA and 0.05% Tween 20 for 2 hours at room temperature. A color reaction was observed by incorporating 3,3′,5,5′-tetramethyl-benzidine into the reaction, which was stopped by adding 50 μL of 2 mol/L H₂SO₄. The optical density was measured at 450 nm in a microplate reader (Multiskan Ascent, Labsystems). The assay was validated using a dilution series of bEnd.3 cells. The interassay variability was 5±2%, whereas the intraassay variability was 2.0±0.8%.

Calculations and Statistical Analysis

NO and ROS production was calculated by determining the slopes of each line with linear regression. Results were expressed relative to control. Statistical comparisons were made by 1-way or 2-way ANOVA, as required. Subsequent post hoc testing was done with the Student–Newman–Keuls test. P values <0.05 were considered statistically significant.

Results

eNOS in bEnd.3 Cells Is, in Part, in an Uncoupled State

To identify the source of NO production in bEnd.3 cells, expression levels of eNOS, nNOS, and iNOS were determined. High expression of eNOS and the absence of iNOS or nNOS protein were confirmed by Western blot (Figure 2A). Nonstimulated bEnd.3 cells produced detectable levels of NO (Figure 2B). Incubation with l-NAME showed that 30 μmol/L was the minimal concentration to inhibit NO production (data not shown). Similar results were obtained with N^G-methyl-l-arginine acetate (minimal concentration, 100 μmol/L) and N_ω-nitro-l-arginine (minimal concentration 30 μmol/L; data not shown).

• Download figure
• Download PowerPoint

Incubation of bEnd.3 cells with the DCF probe resulted in a fluorescent signal, which was decreased 18% (P<0.05) by the addition of l-NAME. Because CM-H₂DCFDA is not reactive toward NO (Figure 1E), these data show that, other than NO production, there is constitutive eNOS-dependent formation of ROS, that is, uncoupling, in the basal state.

VEGF Enhances eNOS-Dependent NO Production but Also ROS Production in bEnd.3 Cells

VEGF induced a concentration-dependent increase in eNOS-dependent NO production (Figure 2B). VEGF also significantly increased eNOS-dependent ROS production (Figure 3). In VEGF-stimulated cells, the addition of l-NAME resulted in a 35% (P<0.05) decrease in the DCF signal versus an 18% (P<0.05) decrease in DCF signal in nonstimulated cells.

• Download figure
• Download PowerPoint

eNOS Uncoupling Is Because of Relative Shortage of BH₄ but Not Because of Shortage of l-Arginine

Shortage of BH₄ can lead to eNOS uncoupling. Determination of biopterin levels in bEnd.3 cells and CDC.HMEC-1 revealed that BH₄ levels are &8 times increased in bEnd.3 cells as compared with CDC.HMEC-1 (Figure 4A). The BH₄/BH₂+biopterin ratio in bEnd.3 cells and HMECs was not significantly different. In addition, eNOS expression in bEnd.3 cells was 250-fold increased in comparison with CDC.HMEC-1 as determined with Western blot (Figure 4B) and ELISA (OD450 bEnd.3 cells, 2.162±0.3; CDC.HMEC-1, 0.007±0.001). The addition of BH₄ (1 to 10 μmol/L) to bEnd.3 cells resulted in a significant (P<0.01) increase in NO production at 3 and 10 μmol/L (Figure 5A). A similar result was obtained with VEGF. ROS levels showed a tendency to decrease in the presence of BH₄ (1 to 10 μmol/L; Figure 5B). A similar result was found in the presence of VEGF. To exclude any antioxidant effects of BH₄, NH₄ was used as a negative control. The addition of 10 μmol/L NH₄ had no effects on NO production and decreased ROS production in bEnd.3 cells (Figure 6).

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Because eNOS uncoupling can also be caused by a shortage of the substrate l-arginine, the effect of additional l-arginine on NO and ROS production was determined. In the presence of high l-arginine levels (500 μmol/L), the addition of BH₄ did not significantly increase basal NO production and also had no effect on VEGF-induced NO production (Figure 5A). Increasing l-arginine concentrations up to 500 μmol/L did not affect the basal ROS production but did lead to a significant increase in eNOS-dependent ROS production in the presence of VEGF (Figure 3). In the presence of high l-arginine levels, ROS production was significantly decreased after the addition of 3 and 10 μmol/L BH₄, both in the absence and presence of VEGF (Figure 5B).

Other Potential ROS Sources

Diphenyleneiodonium (DPI) was used to identify flavin-containing enzymes as possible ROS sources. Incubation of bEnd.3 cells with DPI lead to a significant 20% (P<0.05) decrease in basal ROS production (Figure 7) and a 25% (P<0.05) decrease in the presence of VEGF (data not shown). To determine whether NADPH oxidase, xanthine oxidase, and mitochondria are possible ROS sources, bEnd.3 cells were incubated with apocynin, oxypurinol, and thenoyltrifluoroacetone (TTFA), respective inhibitors of these oxidase systems. Apocynin and TTFA had no effect on the DCF signal (Figure 7). Oxypurinol decreased ROS production slightly in the absence (Figure 7) but not in the presence of VEGF (data not shown). DPI and oxypurinol were also combined with l-NAME, resulting in an additional decrease in ROS production as compared with the effect of the separate inhibitors (Figure 7).

• Download figure
• Download PowerPoint

To exclude the involvement of peroxynitrite (ONOO⁻) in eNOS uncoupling, cells were treated with the ONOO⁻ scavenger Ebselen. Incubation with 10 μmol/L Ebselen did not significantly affect the DCF signal (control, 1.00±0.02; 10 μmol/L Ebselen, 1.05±0.05; 50 ng/mL VEGF, 1.12±0.06; 50 ng/mL VEGF +10 μmol/L Ebselen, 0.96±0.05).

Discussion

It has been proposed that long-term elevation of eNOS expression may require adjustments in tetrahydrobiopterin metabolism,⁶ which could explain the controversial results between studies on prolonged^12,13 and short-term eNOS overexpression.^8–10 The present study demonstrates that bEnd.3 cells chronically express high eNOS levels but do not express the other NOS isoforms. With respect to eNOS, bEnd.3 cells are relatively deficient in BH₄ as compared with other endothelial cells. As a consequence, eNOS is partially in the uncoupled state. Stimulation of eNOS activity in these cells by VEGF enhanced ROS production, particularly at high l-arginine concentrations, which was corrected to a substantial degree by BH₄. Our data suggest that under conditions of chronically increased eNOS expression and particularly if eNOS activity is stimulated, a relative shortage of BH₄ may lead to eNOS uncoupling, resulting in superoxide production. Our results are in line with observations by Ozaki et al,¹³ who demonstrated that overexpression of eNOS in the endothelium promoted atherosclerosis in apoE-knockout mice. Dysfunctional eNOS and subsequent increased superoxide production seems to be responsible for the progression of atherosclerosis in these animals, mainly caused by a deficiency of the cofactor BH₄. In agreement with this study, Cosentino et al¹² showed that increased eNOS expression in SHR is also associated with dysfunctional eNOS, which may contribute to the development of hypertension and its vascular complications in these rats.

bEnd.3 cells are particularly useful for studying eNOS uncoupling, because NADPH oxidase, which is a major source of ROS in endothelial cells,¹⁸ does not contribute to ROS production in these cells. The combined results of this study show that the ROS production in bEnd.3 cells is eNOS dependent, implying that eNOS in bEnd.3 cells is partly in an uncoupled state.

We hypothesized that activation of the enzyme would enhance ROS production. Indeed, stimulation with VEGF increased eNOS-dependent ROS production. This is in line with studies in other endothelial cells that show an eNOS-dependent increase in ROS production after VEGF stimulation.^19,20

We and others found a decrease in eNOS-dependent ROS production after addition of l-NAME, whereas Colavitti et al¹⁹ reported the opposite. The latter observation is consistent with the common perspective that inhibition of functional NOS and, thus, blockage of NO production, results in the loss of the NO-scavenging effect and, thus, increased O₂^·− levels. On the other hand, inhibition of dysfunctional NOS will lead to a decrease in ROS production because of the coupling state of eNOS.^21–24

eNOS dysfunction and subsequent O₂^·− production is commonly thought to result from a shortage of BH₄.⁶ In the present study, the addition of BH₄ resulted in a significant increase in NO production in bEnd.3 cells, whereas ROS levels showed a tendency to decrease. NH₄ was used as a negative control, because it exerts antioxidant effects but has no influence on eNOS uncoupling. The addition of NH₄ had no effect on NO production and decreased ROS production in bEnd.3 cells. From these results, it can be concluded that the decreased ROS production in the presence of BH₄ is because of a decrease in eNOS uncoupling and not caused by antioxidant effects of BH₄. In line with these results, it has been shown that overexpression of GTP-cyclohydrolase I, the rate-limiting enzyme in BH₄ synthesis, reduced endothelial dysfunction in apoE-knockout mice.²⁵ In addition, GTP-cyclohydrolase I–knockout mice, which are deficient in BH₄, showed hypertension and increased vascular superoxide production. The latter was inhibited by N^G-methyl-l-arginine acetate, indicating that ROS was generated by uncoupled eNOS.²¹

Other than BH₄ deficiency, shortage of the substrate l-arginine could theoretically lead to uncoupling of eNOS.¹ However, there is limited evidence on the role of l-arginine in eNOS dysfunction.^26,27 Moreover, there is no consensus on whether supplementation of l-arginine is beneficial under conditions of endothelial dysfunction. Some studies in animals and humans show restored endothelial function,^28,29 whereas others fail to show a beneficial effect of l-arginine.^30–32 Chen et al³³ found that chronic treatment with l-arginine negated the positive effect of iNOS deficiency in apoE/iNOS double-knockout mice. Furthermore, diabetic rats showed increased ROS production after the addition of l-arginine, which was partly reduced by the addition of l-NAME and DPI and the addition of BH₄.²⁴ Our results indicate that a shortage of l-arginine is not causing eNOS-dependent ROS production in bEnd.3 cells. In fact, increasing the l-arginine concentration enhanced VEGF-induced ROS production. There are 2 possible explanations for these findings. First, increased l-arginine levels have been associated with an increase in total biopterin levels, but unaltered BH₄ levels, suggesting that l-arginine supplementation leads to oxidation of biopterin.³³ Second, increased O₂^·− production in the presence of l;arginine might be because of increased calmodulin binding and a resulting increased electron flow through the enzyme,^34,35 leading to a relatively augmented shortage of BH₄. The mechanism through which increasing levels of l-arginine enhances ROS production by increasing l-arginine levels in bEnd.3 cells is unclear.

Although the addition of BH₄ led to a reduction of ROS production in bEnd.3 cells, it did not completely abolish ROS production. These results correspond with the in vitro study with the purified enzyme,⁷ in which BH₄ only partly inhibited uncoupling of the purified eNOS enzyme. We identified xanthine oxidase as an additional source of ROS in bEnd.3 cells under basal conditions. ONOO⁻, NADPH oxidase, and mitochondria were excluded as contributors to ROS production in bEnd.3 cells.

Several reports have been published on intrinsic superoxide production by NO synthases. For nNOS, it has been reported that ROS are produced by the haem group.³⁶ For iNOS and eNOS, however, the source of ROS production is unclear. Some reports suggest that both the haem and the flavin domain in eNOS produce superoxide.^27,37 In bEnd.3 cells, the addition of DPI led to a decrease in but not complete inhibition of ROS production, indicating that flavins are not involved in the intrinsic ROS production by eNOS in these cells. These data suggest that the source of the remaining eNOS-dependent ROS production in bEnd.3 cells, even in the presence of additional BH₄, is probably the haem group in eNOS.^27,7,37

Perspectives

Although eNOS gene transfer has been proposed to be useful in the treatment of endothelial dysfunction,^8–10 we now show that sustained high expression of eNOS protein results in ROS production because of uncoupling. eNOS-dependent ROS production in bEnd.3 cells is partly corrected by BH₄ supplementation. These findings suggest that attempts to improve endothelial function, by sustained increases of eNOS expression and activity, for instance by gene transfer of eNOS or VEGF, should be accompanied by concomitant increases of BH₄ to prevent eNOS uncoupling.

Footnotes