[Need for blood transfusion in premature infants in 2 Dutch perinatology centres particularly determined by blood sampling for diagnosis]. / Bloedtransfusiebehoefte bij prematuren in 2 Nederlandse perinatologische centra sterk bepaald door bloedafname voor diagnostiek.

OBJECTIVE: Determination of factors related to the need for transfusion in premature infants. DESIGN: Descriptive. METHOD: The need for transfusion in premature infants was determined in 2 academic centres: University Medical Center Utrecht and Leiden University Medical Center, The Netherlands. The data had been acquired in another study. The factors under study were: hospital, pregnancy duration, birth weight, gender, time of clamping of the umbilical cord, total volume of blood sampled for diagnostic purposes, number of days of mechanical ventilation, total duration of admission and duration of the admission to the Neonatal Intensive care unit. Both hospitals followed the national interdisciplinary practice guideline 'Blood transfusion'. RESULTS: The total volume ofsampled blood for diagnosis, the duration of the mechanical ventilation and the admission period were related to a greater need for transfusion. On the other hand, the chance of transfusions diminished with longer pregnancy duration or increased birth weight. The difference in need for blood transfusion between both centres was significant. The total volume of transfused erythrocytes showed a strong correlation with the volume sampled for diagnostic procedures. CONCLUSION: Anaemia in neonates is strongly related to the amount of blood taken for diagnostic procedures. Alternatives for blood transfusions in premature infants, and consequently for the reduction of the number of donors per child, are to be sought in delayed clamping of the umbilical cord, use of erythropoietin and use ofautologous umbilical cord blood.

Vitamin K deficiency and bleeding after long-term use of cholestyramine.

Although it has been long known that in theory the use of cholestyramine can cause coagulopathy due to reduced absorption of vitamin K, only a few cases have been reported. In those cases the coagulopathy occurred within a few weeks to months after the start of therapy. We report a patient with severe pruritus due to intrahepatic cholestasis, who was on cholestyramine therapy for over 25 years before haemorrhage occurred. This case demonstrates that one should be aware of the possibility of depletion of fat-soluble vitamins during the long-term use of cholestyramine.

Platelet activation by the apoB/E receptor-binding domain of LDL.

Low density lipoprotein (LDL) increases the sensitivity of human platelets for agonists by activating p38MAPK. Antibody 4G3 disturbs apoB100 binding to the classical apoB/E receptor and inhibits LDL-induced p38MAPK activation, whereas an antibody against a distal domain on apoB 100 has no effect. Peptide RLTRKRGLKLA mimics the binding domain of apoB 100 called the B-site and activates platelet p38MAPK. Activation by B-site peptide is dose-dependent, transient and followed by desensitization, in accordance with receptor-mediated signalling. A scrambled peptide and a partially homologous peptide RKLRKRLLRDA mimicking the apoB/E receptor binding site of apoE in high density lipoprotein (HDL) also activate p38MAPK albeit 40% weaker, but an uncharged peptide lacks p38MAPK activating capacity. LDL and B-site peptide bind to the same binding sites and initiate similar signalling to p38MAPK and cytosolic phospholipase A2. Thus, LDL and to a lesser extent HDL activate platelets via specific domains in the protein moiety that recognize receptors of the LDL receptor family.

Lysophosphatidic acid-independent platelet activation by low-density lipoprotein.

Mildly oxidized low-density lipoprotein activates platelets through lysophosphatidic acid (LPA). Hence, the platelet-activating properties attributed to native low-density lipoprotein (nLDL) might be caused by LPA contamination. We show that nLDL enhances thrombin receptor-activating peptide (TRAP)-induced fibrinogen binding to alpha(IIb)beta(3). The LPA receptor blocker N-palmitoyl-L-serine-phosphoric acid did not affect nLDL-enhanced fibrinogen binding induced by TRAP, but reduced TRAP-induced binding. cAMP and inhibitors of protein kinase C and Ca(2+) rises completely blocked ligand binding by TRAP and nLDL/TRAP. Inhibitors of p38(MAPK) and ADP secretion interfered only partially. Blockade of Rho-kinase increased ligand binding 2-3-fold. We conclude that nLDL enhances TRAP-induced fibrinogen binding independent of LPA.

The identification of polymorphisms in the coding region of the apolipoprotein (a) gene--association with earlier identified polymorphic sites and influence on the lipoprotein (a) concentration.

Lipoprotein (a) [Lp(a)] is a quantitative genetic trait in human plasma and elevated levels represent a major inherited risk factor for the development of atherosclerotic disease. In our search for sequence polymorphisms in the coding region of the apolipoprotein(a) [apo(a)] gene that may affect the Lp(a) concentration, four new polymorphic sites were identified. These include two coinciding polymorphisms with an allele frequency of 38% located at amino acid positions 87 and 101 (Leu87,101-->Val) in the interkringle region of kringle IV (K.IV) type 7 and two polymorphisms located in K.IV type 7 (Arg60-->Ser) and in K.IV type 10 (Tyr2-->Phe) both with estimated allele frequencies of about 1%. The linkage between the newly identified K.IV type 7 Leu87,101 -->Val polymorphism and earlier described polymorphic sites in the non-coding and coding regions of the apo(a) gene, its distribution over the apo(a) isoform sizes and its possible influence on the Lp(a) concentration was analysed in 201 healthy unrelated Caucasians. The earlier described polymorphic sites included in this study were the variable number of a TTTTA pentanucleotide repeat (7-11 PNR) starting at -1231 bp, the -772 bp G/A polymorphism, the +93 bp C/T polymorphism and the +121 bp G/A polymorphism in the non-coding region, and the K.IV type 8 Thr12/Pro polymorphism and the K.IV type 10 Thr66/Met polymorphism in the coding region of the apo(a) gene. Linkage disequilibria were observed between the polymorphic sites in the 5' non-coding region and the sites in K.IV type 7 and 8 in the coding region of the apo(a) gene, confirming that the expansion of the variable number of K.IV type 2 repeats results from intrachromosomal recombinational events. The distribution over the apo(a) isoform sizes of the K.IV type 7 Val87,101 subtype was not significantly different from that of the K.IV type 7 Leu87,101 wild-type, suggesting a relative ancient mutational event. No influence of the K.IV type 7 Leu87,101-->Val polymorphism on the Lp(a) level was observed. In fact, of all the polymorphic sites studied, only the +121 A subtype could be associated with an increased, and the K.IV type 8 Pro12 and the 10 PNR subtypes with a reduced, Lp(a) concentration corrected for apo(a) isoform size (p <0.05).

Early platelet activation by low density lipoprotein via p38MAP kinase.

Low Density Lipoprotein (LDL) is known to sensitize platelets for physiological agonists. To clarify the basis of this sensitization, we investigated the involvement of p38MAP Kinase (p38MAPK). As dual phosphorylation on Thr180 and Tyr182 of p38MAPK is the trigger for activation of the kinase, p38MAPK-activity was measured with an antibody that recognizes the dual-phosphorylated sequence. LDL induced a rapid and dose dependent activation of p38MAPK. The activation was not inhibited by a wide variety of inhibitors of platelet signalling, including TxA2-formation, Phospholipase C-activation, Ca2+-mobilization and ERK 1/2-activation. Only a slight reduction in p38MAPK-activation was observed when protein kinase C was inhibited. Activation of p38MAPK was strongly inhibited by a rise in cAMP. Thus, p38MAPK-activation was upstream of most signalling pathways and close to the LDL-receptor. A number of platelet receptors was screened with the use of antibodies. Integrins alphaIIbbeta3 and alpha2beta1, as well as the FcgammaRII-receptor, CD36 (platelet glycoprotein IV), CD68 (gp110) and Low Density Lipoprotein-receptor related protein (LRP) were not implicated in LDL-induced p38MAPK-activation. Inhibition of LDL binding by modification of apo B100 lysines reduced p38MAPK-activation by 80%. Activation of p38MAPK resulted in an increase in release of arachidonic acid, the precursor for thromboxane A2 synthesis. In conclusion, activation of p38MAPK might be the first step in platelet sensitization by LDL, leading to formation of arachidonate metabolites and increased aggregation and secretion responses to physiological agonists.

Low-density lipoprotein enhances platelet secretion via integrin-alphaIIbbeta3-mediated signaling.

LDL is known to increase the sensitivity of human platelets for agonists and to induce aggregation and secretion independently at high concentrations, but its mechanism of action is largely obscure. To clarify how LDL increases platelet sensitivity, cells were incubated in lipoprotein-poor plasma and treated with collagen at a concentration that induced approximately 20% secretion of 14C-serotonin. Preincubation with LDL (30 minutes at 37 degreesC) enhanced secretion in a dose-dependent manner to 60+/-14% at a concentration of 2 g LDL protein/L. Similar stimulation by LDL was seen when secretion was induced by the thrombin receptor-activating peptide. This enhancement was strongly reduced (1) in the presence of monoclonal antibody PAC1 against activated alphaIIbbeta3, a polyclonal antibody against alphaIIb, and in the presence of the fibrinogen peptides GRGDS and HHLGGAKQAGDV; (2) in alphaIIbbeta3-deficient platelets; and (3) after dissociation of alphaIIbbeta3. In contrast, binding of 125I-LDL to normal platelets in the presence of PAC1, anti-alphaIIb, GRGDS, and HHLGGAKQAGDV, and to alphaIIbbeta3-deficient platelets was normal. LDL increased the surface expression of fibrinogen in lipoprotein-poor plasma and fibrinogen-free medium, suggesting that extracellular and granular fibrinogen bind to alphaIIbbeta3 after platelet-LDL interaction. Platelets deficient in fibrinogen (<0.5% of normal) or von Willebrand Factor (<1% of normal) but containing normal amounts of other ligands for alphaIIbbeta3 preserved responsiveness to LDL, indicating that occupancy of alphaIIbbeta3 was not restricted to fibrinogen. Inhibition of protein kinase C (bisindolylmaleimide) diminished fibrinogen binding and sensitization by LDL; inhibition of tyrosine kinases (herbimycin A) left fibrinogen binding unchanged but diminished sensitization by LDL. We conclude that an increased concentration of LDL, such as observed in homozygous familial hypercholesterolemia, sensitizes platelets to stimulation by collagen and thrombin receptor-activating peptide via ligand-induced outside-in signaling through integrin-alphaIIbbeta3.

Low density lipoprotein phosphorylates the focal adhesion-associated kinase p125(FAK) in human platelets independent of integrin alphaIIb beta3.

Low density lipoprotein (LDL) is known to sensitize platelets to agonists via integrin mediated outside-in signaling (Hackeng, C. M., Huigsloot, M., Pladet, M. W., Nieuwenhuis, H. K., Rijn, H. J. M. v., and Akkerman, J. W. N. (1999) Arterioscler. Thromb. Vasc. Biol., in press). As outside in signaling is associated with phosphorylation of p125(FAK), the effect of LDL on p125(FAK) phosphorylation in platelets was investigated. LDL induced p125(FAK) phosphorylation in a dose- and time- dependent manner. The phosphorylation was independent of ligand binding to integrin alphaIIbbeta3 and aggregation, such in contrast to alpha-thrombin-induced p125(FAK) phosphorylation, that critically depended on platelet aggregation. Platelets from patients with Glanzmann's thrombastenia showed the same LDL- induced phos- phorylation of p125(FAK) as control platelets, whereas alpha-thrombin completely failed to phosphorylate the kinase in the patients platelets. LDL signaling to p125(FAK) was independent of integrin alpha2 beta1, the FcgammaRII receptor, and the lysophosphatidic acid receptor and not affected by inhibitors of cyclooxygenase, protein kinase C, ERK1/2 or p38(MAPK). Phosphorylation of p125(FAK) by LDL was strongly inhibited by cyclic AMP. These observations indicate that LDL is a unique platelet agonist, as it phosphorylates p125(FAK) in platelet suspensions, under unstirred conditions and independent of integrin alphaIIb beta3.

Efficacy and muscle safety of fluvastatin in cyclosporine-treated cardiac and renal transplant recipients: an exercise provocation test.

BACKGROUND: Dyslipidemia is found in the majority of renal and cardiac transplant recipients. Although 3-hydroxy-3-methyl-glutaryl coenzyme A reductase inhibitors significantly lower low-density lipoprotein cholesterol (LDL-C) levels, such treatment has been associated with muscle toxicity, especially when used in combination with cyclosporine (CsA). We investigated the efficacy and muscle safety of fluvastatin, a new 3-hydroxy-3-methyl-glutaryl coenzyme A reductase inhibitor, in CsA-treated transplant recipients. METHODS: The efficacy was determined by measuring the lipid profile before and after 8 weeks of fluvastatin therapy. As parameter for possible muscle damage, the rise in serum levels of the muscle proteins creatine kinase and myoglobin was measured after an exercise provocation test (30 min on a bicycle ergometer at 60% of their maximal work load) before and during fluvastatin therapy. Nineteen CsA-treated renal and cardiac transplant recipients with hypercholesterolemia were selected. RESULTS: After 8 weeks of treatment with a dose of fluvastatin necessary to reduce LDL-C below 3.5 mmol/L (20 mg for 3 and 40 mg for 16 patients), total cholesterol was lowered by 20% and LDL-C by 30%, and HDL2-C was increased by 35% (all P<0.01). The rise in creatine kinase after exercise before and during fluvastatin therapy was, respectively, 40% and 51%, and the rise in myoglobin was 64% and 50%. These rises were not significantly different. Hence, there was no indication for subclinical muscle pathology by fluvastatin use. Fluvastatin was well tolerated, and no adverse effects on liver or kidney function were found. CONCLUSIONS: Fluvastatin can effectively lower LDL-C in CsA-treated renal and cardiac transplant recipients, without demonstrable adverse effects.

The functional and clinical significance of the Met-->Thr substitution in Kringle IV type 10 of apolipoprotein(a).

Lipoprotein(a) [Lp(a)], an independent risk factor for the development of atherosclerosis, contains an apolipoprotein(a) [apo(a)] moiety covalently linked to a LDL moiety. Apo(a) is a glycoprotein homologous to plasminogen as it contains multiple repeats of a lysine binding domain resembling plasminogen kringle IV (K.IV). The multiple K.IV repeats can be differentiated in ten types that show a variation in their lysine binding capacity. Since K.IV type 10 shows the highest conservation of the amino acids postulated to form the lysine binding pocket, this kringle is suggested to be the main lysine binding site of apo(a). Recently, a T-->C polymorphism in the apo(a)-gene was reported, leading to a Met-->Thr substitution at amino acid position 66 of K.IV type 10, in the vicinity of the postulated lysine binding pocket. To investigate the significance of this substitution on some in vitro characteristics of Lp(a), the affinity for lysine-Sepharose and the binding affinity for limited plasmin digested des AA fibrin (Desafib-X) of the two subtypes was determined using plasma of donors homozygous for the polymorphism. These studies revealed a large heterogeneity in the binding characteristics, irrespective of the subtype. The comparison of the allele frequencies of this polymorphism in 155 patients having symptomatic atherosclerosis versus 153 normolipidemic controls revealed no significant differences. In conclusion, this study suggests that the presence of either a Met66 or a Thr66 residue in K.IV type 10 of apo(a) has no consequences for the binding characteristics of Lp(a) toward lysine-Sepharose or Desafib-X, nor is it associated with the presence of symptomatic atherosclerosis.

Oral magnesium supplementation in insulin-requiring Type 2 diabetic patients.

Oral magnesium (Mg) supplementation can improve insulin sensitivity and secretion in patients with Type 2 diabetes mellitus (DM). We studied the effect of Mg supplementation on glycaemic control, blood pressure, and plasma lipids in insulin-requiring patients with Type 2 DM. Fifty moderately controlled patients were randomized to 15 mmol Mg or placebo daily for 3 months. Plasma Mg, glucose, HbA1c, lipids, erythrocyte Mg, Mg and glucose concentrations in 24-h urine, and systolic and diastolic pressure were measured before and after 3 months treatment. Plasma Mg concentration was higher after supplementation than after placebo (0.82 +/- 0.07 vs 0.78 +/- 0.08 mmol l(-1), p < 0.05), as was Mg excretion (5.5 +/- 1.9 vs 3.7 +/- 1.4 mmol 24 h(-1), p = 0.004) but erythrocyte Mg concentrations were similar. No significant differences were found in glycaemic control (glucose: 10.7 +/- 3.8 vs 11.6 +/- 6.2 mmol l(-1), p = 0.8; HbA1c: 8.9 +/- 1.6 vs 9.1 +/- 1.2%, p = 0.8), lipids or blood pressure. On-treatment analysis (34 patients: 18 on Mg, 16 on placebo) yielded similar results. An increase in plasma Mg concentration irrespective of medication was associated with a tendency to a decrease in diastolic pressure (increased plasma Mg vs no increase: -4.0 +/- 10.1 vs +2.5 +/- 12.0 mmHg, p = 0.059). Three months' oral Mg supplementation of insulin-requiring patients with Type 2 DM increased plasma Mg concentration and urinary Mg excretion but had no effect on glycaemic control or plasma lipid concentrations.

Acute exogenous elevation of plasma free fatty acids does not influence the plasma magnesium concentration.

OBJECTIVE: Plasma non-esterified (free) fatty acid concentrations rise as a consequence of stimulated endogenous lipolysis and are inversely related to the plasma magnesium concentration when plasma adrenaline concentration is increased. The aim of the study was to test whether high plasma non-esterified fatty acid concentration after infusion of non-esterified fatty acids decreases plasma magnesium concentration. METHODS: Twelve healthy subjects received 500 ml Intralipid or saline in a randomised, cross-over, double-blind design. Infusion of Intralipid results in an isolated elevation of plasma non-esterified fatty acid concentration. Plasma magnesium concentration was determined at baseline and every 30 minutes; plasma non-esterified fatty acid and triglyceride concentrations at baseline and after 120 minutes. RESULTS: Initial plasma magnesium, non-esterified fatty acid, and triglyceride concentrations were similar in both groups. A significant increase in plasma non-esterified fatty acids (2.42 +/- 0.96 mmol/l vs 0.58 +/- 0.23 mmol/l, p = 0.00013) and triglyceride (median and 95th percentile 5.36 (7.35) mmol/l vs 1.18 (1.92) mmol/l, p = 0.003) concentrations was seen with Intralipid. Plasma magnesium concentration increased significantly after Intralipid (0.89 +/- 0.09 mmol/l vs 0.81 +/- 0.06 mmol/l, p = 0.007). No significant changes were seen with saline. A positive association was found between the change in plasma magnesium and triglyceride concentrations (r = 0.85, p = 0.001). CONCLUSION: Acute infusion of non-esterified fatty acids from an exogenous source does not result in a fall in plasma magnesium concentration, indicating that the circulating non-esterified fatty acids play no part in a decrease in plasma magnesium concentration. The high circulating non-esterified fatty acid levels and the fall in plasma magnesium concentration are both a consequence of intracellular lipolysis.

Effect of low-dose prednisone (with calcium and calcitriol supplementation) on calcium and bone metabolism in healthy volunteers.

The administration of moderate to high doses of corticosteroids is associated with bone loss. This probably results from the uncoupling of bone formation (decreased) and bone resorption (unchanged or increased). We examined the effect of low-dose (10 mg/day) prednisone (LDP) and the possible mitigating effects of calcium and 1.25 (OH)2 vitamin D (calcitriol) on calcium and bone metabolism in eight healthy, young male volunteers. The study consisted of four observation periods: in the first period, LDP was prescribed during 1 week; in the second, third and fourth periods, calcium (500 mg/day), calcitriol (0.5 micrograms b.i.d.) and calcium in combination with calcitriol, respectively, were added to LDP. Bone formation was measured by means of serum osteocalcin, carboxy-terminal propeptide of type 1 procollagen (P1CP) and alkaline phosphatase, bone resorption by means of urinary excretion of calcium, hydroxyproline, (free and total) pyridinoline, (free and total) deoxypyridinoline and serum carboxy-terminal cross-linked telopeptide of type 1 collagen (1CTP). Dietary calcium and sodium intake were maintained at a stable level during the entire study period. Treatment with LDP led to a decrease in osteocalcin, P1CP and alkaline phosphatase (all P < 0.01). Urinary excretion of pyridinolines, hydroxyproline and serum 1CTP did not increase, but remained unchanged or slightly reduced (P < 0.05), depending on the time of measurement and the marker of bone resorption. Parathyroid hormone (PTH) (insignificantly) increased during LDP (+19%) and LDP plus calcium (+14%), but decreased during supplementation with calcitriol (-16%) and calcium/calcitriol (-44%; P < 0.01). Urinary excretion of calcium increased during treatment with LDP and calcitriol (P < 0.05) and calcium/calcitriol (P < 0.05). It is concluded that LDP has a negative effect on bone metabolism, since bone formation decreased while bone resorption remained unchanged or decreased slightly. The increase in PTH during LDP could be prevented by calcitriol combined with calcium supplementation.

Effect of an increase in the plasma potassium concentration on renal magnesium handling in healthy volunteers.

BACKGROUND: Lower plasma magnesium concentrations are associated with clinical problems such as arrhythmias and hypertension. Plasma magnesium concentration is tightly controlled by the kidney. Modifying renal magnesium threshold may provide a means to increase the plasma magnesium concentration. Since evidence has been presented that potassium deficiency by itself may increase renal magnesium loss, the hypothesis that elevating plasma potassium would result in an increase in plasma magnesium concentration was tested in healthy volunteers. METHODS: Plasma potassium was raised in nine healthy volunteers by oral administration of 20 mg amiloride daily during 3 weeks. Magnesium metabolism was assessed before and after this period by plasma levels, urinary magnesium excretion and fractional magnesium excretion, and magnesium loading test (MLT). This MLT allows calculation of renal retention of a magnesium load. RESULTS: Basal plasma magnesium levels (0.84 +/- 0.07 vs 0.84 +/- 0.05 mmol/l) as well as urinary magnesium excretion (4.37 +/- 1.73 vs 3.67 +/- 1.37 mmol/day) and erythrocyte magnesium levels (1.72 +/- 0.16 vs 1.76 +/- 0.14 mmol Mg/l red blood cells) were similar before and on amiloride. Plasma potassium rose significantly on amiloride (3.64 +/- 0.24 vs 4.07 +/- 0.54 mmol/l, P < 0.05). No change was observed in magnesium retention with the MLT: 22.7 +/- 26.7 vs 29.2 +/- 20.6% (P = 0.5). CONCLUSIONS: Despite an increased plasma potassium concentration, no change was observed in plasma magnesium levels, urinary magnesium excretion or renal magnesium retention of an intravenously administered magnesium load. This indicates that increasing plasma potassium within the normal range does not modify the renal magnesium threshold.

Assessment of nitric oxide production by measurement of [15N]citrulline enrichment in human plasma using high-performance liquid chromatography-mass spectrometry.

Nitric oxide (NO) is formed by a class of NO synthases (NOS), which convert arginine into citrulline. A decreased in vivo NO availability can be the result of an increased NO inactivation or a decreased NO production. The latter can be assessed by measurement of isotopic enrichment of plasma citrulline during infusion of isotopically labeled arginine. The potential of high-performance liquid chromatography (HPLC) coupled to mass spectrometry (MS) to determine enrichments of [15N2]arginine and [15N]-citrulline in plasma during infusion of [15N2]arginine in humans was investigated. Two types of MS instruments were evaluated: a sector-type mass spectrometer equipped with a frit fast-atom bombardment (FAB) interface and a quadrupole instrument with electrospray ionization (ESI). FAB-MS appeared to be unsuitable for determination of isotope ratios, because background ions influenced the observed isotope ratio in an unpredictable way. In combination with either off- or on-line reversed-phase HPLC, ESI-MS proved to be a more reliable technique. However, the amount of material that is introduced in the mass spectrometer is critical and should be carefully controlled. During infusion of [15N2]arginine in 14 healthy subjects, a mean arginine-to-citrulline conversion rate of 0.22 +/- 0.07 (SD) mumol.kg-1.h-1 was found. In 4 subjects who received an intravenous infusion with the NOS antagonist L-NMMA, the conversion rate decreased from 0.30 +/- 0.14 to 0.10 +/- 0.06 mumol.kg-1.h-1. It is concluded that ESI-MS in combination with HPLC can be successfully applied for determination of arginine and citrulline enrichments in plasma, thus providing a useful tool for assessment of in vivo NO production.

Tetrahydrobiopterin regulates superoxide and nitric oxide generation by recombinant endothelial nitric oxide synthase.

Nitric oxide (NO) produced by the endothelial isoform of nitric oxide synthase (NOS III) is a key determinant of the anti-atherosclerotic properties of the endothelium. Recent in vivo studies suggest that NOS III may also be a source of superoxide production, which would limit its role as a NO-producing enzyme. In the current study we examined both the NO and the superoxide generating potential of recombinant NOS III obtained from a baculovirus/Sf9 expression system. Using lucigenin chemiluminesence we could indeed demonstrate (superoxide dismutase inhibitable) superoxide production by NOS III. This superoxide production was not affected by administration of L-arginine, but could be inhibited dose-dependently by the co-factor tetrahydrobiopterin (BH4). BH4 also dose dependently decreased superoxide generation by hypoxanthine/xantine oxidase suggesting a direct antioxidant effect. Superoxide generation by NOS III could be completely inhibited by diphenyleneiodonium (DPI), an inhibitor of the flavin moiety of the enzyme, indicating that this group is a main source of superoxide production by the enzyme. Using measurement of [3H-L-arginine] conversion to [3H-L-citrulline], it appeared that BH4 directly increased the production of NO by NOS III. In addition, we observed that BH4 stablized the NOS III in its dimeric form, suggesting that an effect on allosteric conformation could be involved in this effect on NO production. NOS III thus appears to be a superoxide generating enzyme probably through its flavin moiety, as well as a BH4-dependent NO producing enzyme.