Inhibition of thiol isomerase activity diminishes endothelial activation of plasminogen, but not of protein C.

INTRODUCTION: Cell surface thiol isomerase enzymes, principally protein disulphide isomerase (PDI), have emerged as important regulators of platelet function and tissue factor activation via their action on allosteric disulphide bonds. Allosteric disulphides are present in other haemostasis-related proteins, and we have therefore investigated whether thiol isomerase inhibition has any influence on two endothelial activities relevant to haemostatic regulation, namely activation of protein C and activation of plasminogen, with subsequent fibrinolysis. MATERIALS AND METHODS: The study was performed using the human microvascular endothelial cell line HMEC-1. Thiol isomerase gene expression was measured by RT-PCR and activation of protein C and plasminogen by cell-based assays using chromogenic substrates S2366 and S2251, respectively. Cell mediated fibrinolysis was measured by monitoring absorbance at 405 nm following fibrin clot formation on the surface of HMEC-1 monolayers. RESULTS AND CONCLUSIONS: A variety of thiol isomerase enzymes, including PDI, were expressed by HMEC-1 cells and thiol reductase activity detectable on the cell surface was inhibited by both RL90 anti-PDI antibody and by the PDI inhibitor quercetin-3-rutinoside (rutin). In cell-based assays, activation of plasminogen, but not of protein C, was inhibited by RL90 antibody and, to a lesser extent, by rutin. Fibrin clot lysis occurring on a HMEC-1 monolayer was also significantly slowed by RL90 antibody and by rutin, but RL90-mediated inhibition was abolished in the presence of exogenous tissue plasminogen activator (tPA). We conclude that thiol isomerases, including PDI, are involved in fibrinolytic regulation at the endothelial surface, although not via a direct action on tPA. These findings broaden understanding of haemostatic regulation by PDI, and may aid in development of novel anti-thrombotic therapeutic strategies targeted via the fibrinolysis system.

Inhibition of NADPH supply by 6-aminonicotinamide: effect on glutathione, nitric oxide and superoxide in J774 cells.

We have examined the integrity of J774 cell nitric oxide (NO) production and glutathione maintenance, whilst NADPH supply was compromised by inhibition of the pentose pathway with 6-aminonicotinamide. In resting cells 6-phosphogluconate accumulation began after 4 h and glutathione depletion after 24 h of 6-aminonicotinamide treatment. Cellular activation by lipopolysaccharide/interferon-lambda decreased glutathione by approximately 50% and synchronous 6-aminonicotinamide treatment exacerbated this to 31.2% of control (P < 0.05). In activated cells NO2- production was inhibited by 60% with 6-aminonicotinamide (P < 0.01), and superoxide production by 50% (P < 0.01) in zymosan-activated cells. NADPH production via the pentose pathway is therefore important to sustain macrophage NO production whilst maintaining protective levels of glutathione.

Mitochondrial superoxide production during oxalate-mediated oxidative stress in renal epithelial cells.

Crystals of calcium oxalate monohydrate (COM) in the renal tubule form the basis of most kidney stones. Tubular dysfunction resulting from COM-cell interactions occurs by mechanism(s) that are incompletely understood. We examined the production of reactive oxygen intermediates (ROI) by proximal (LLC-PK1) and distal (MDCK) tubular epithelial cells after treatment with COM (25-250 microg/ml) to determine whether ROI, specifically superoxide (O(2)(*-)), production was activated, and whether it was sufficient to induce oxidative stress. Employing inhibitors of cytosolic and mitochondrial systems, the source of ROI production was investigated. In addition, intracellular glutathione (total and oxidized), energy status (ATP), and NADH were measured. COM treatment for 1-24 h increased O(2)(*-) production 3-6-fold as measured by both lucigenin chemiluminescence in permeabilized cells and dihydrorhodamine fluorescence in intact cells. Using selective inhibitors we found no evidence of cytosolic production. The use of mitochondrial probes, substrates, and inhibitors indicated that increased O(2)(*-) production originated from mitochondria. Treatment with COM decreased glutathione (total and redox state), indicating a sustained oxidative insult. An increase in NADH in COM-treated cells suggested this cofactor could be responsible for elevating O(2)(*-) generation. In conclusion, COM increased mitochondrial O(2)(*-) production by epithelial cells, with a subsequent depletion of antioxidant status. These changes may contribute to the reported cellular transformations during the development of renal calculi.

Platelet function and the bleeding time in progressive renal failure.

Bleeding time and platelet function tests were performed on 31 patients with progressive chronic renal failure (CRF) due to non-immunological (urological) causes, and compared with 22 healthy controls. Patients were classified as mild (plasma creatinine less than 300 mumol/l), moderate (300-600 mumol/l) or severe renal failure (greater than 600 mumol/l). Bleeding time was rarely prolonged in mild and moderate CRF and mean bleeding time significantly elevated only in severe CRF (p less than 0.005). Haematocrit was the only index which correlated with bleeding time (r = -0.40). Platelet counts, collagen stimulated thromboxane generation, and platelet aggregation responses to ADP, collagen and ristocetin were all either normal or increased in all three CRF groups, but thromboxane production in clotting blood was reduced. Plasma fibrinogen, C reactive protein and von Willebrand factor (vWF) were elevated in proportion to CRF. We found no evidence that defects in platelet aggregation or platelet interaction with vWF prolong the bleeding time in patients with progressive CRF.

Plasma D dimer: a useful marker of fibrin breakdown in renal failure.

D dimer and other large fragments produced during the breakdown of crosslinked fibrin may be measured by enzyme immunoassay using monoclonal antibodies. In 91 patients with renal disease and varying degrees of renal dysfunction, plasma D dimer showed no correlation with renal function, whereas FgE antigen, a fibrinogen derivative which is known to be cleared in part by the kidney, showed a significant negative correlation with creatinine clearance. Plasma concentrations of D dimer were, however, increased in patients with chronic renal failure (244 +/- 31 ng/ml) (mean +/- SEM) and diabetic nephropathy (308 +/- 74 ng/ml), when compared with healthy controls (96 +/- 13 ng/ml), and grossly elevated in patients with acute renal failure (2,451 +/- 1,007 ng/ml). The results indicate an increase in fibrin formation and lysis, and not simply reduced elimination of D dimer by the kidneys, and are further evidence of activated coagulation in renal disease. D dimer appears to be a useful marker of fibrin breakdown in renal failure.

Evidence for a cyclic GMP-independent mechanism in the anti-platelet action of S-nitrosoglutathione.

1. We have measured the ability of a range of NO donor compounds to stimulate cyclic GMP accumulation and inhibit collagen-induced aggregation of human washed platelets. In addition, the rate of spontaneous release of NO from each donor has been measured spectrophotometrically by the oxidation of oxyhaemoglobin to methaemoglobin. The NO donors used were five s-nitrosothiol compounds: S-nitrosoglutathione (GSNO), S-nitrosocysteine (cysNO), S-nitroso-N-acetyl-DL-penicillamine (SNAP), S-nitroso-N-acetyl-cysteine (SNAC), S-nitrosohomocysteine (homocysNO), and two non-nitrosothiol compounds: diethylamine NONOate (DEANO) and sodium nitroprusside (SNP). 2. Using 10 microM of each donor compound, mean+/-s.e.mean rate of NO release ranged from 0.04+/-0.001 nmol min(-1) (for SNP) to 3.15+/-0.29 nmol min(-1) (for cysNO); cyclic GMP accumulation ranged from 0.43+/-0.05 pmol per 10(8) platelets (for SNP) to 2.67+/-0.31 pmol per 10(8) platelets (for cysNO), and inhibition of platelet aggregation ranged from 40+/-6.4% (for SNP) to 90+/-3.8% (for SNAC). 3. There was a significant positive correlation between the rate of NO release and the ability of the different NO donors to stimulate intra-platelet cyclic GMP accumulation (r = 0.83; P = 0.02). However, no significant correlation was observed between the rate of NO release and the inhibition of platelet aggregation by the different NO donors (r= -0.17), nor was there a significant correlation between cyclic GMP accumulation and inhibition of aggregation by the different NO donor compounds (r = 0.34). 4. Comparison of the dose-response curves obtained with GSNO, DEANO and 8-bromo cyclic GMP showed DEANO to be the most potent stimulator of intraplatelet cyclic GMP accumulation (P < 0.001 vs both GSNO and 8-bromo cyclic GMP), but GSNO to be the most potent inhibitor of platelet aggregation (P < 0.01 vs DEANO, and P < 0.001 vs 8-bromo cyclic GMP). 5. The rate of NO release from GSNO, and its ability both to stimulate intra-platelet cyclic GMP accumulation and to inhibit platelet aggregation, were all significantly diminished by the copper (I) (Cu+) chelating agent bathocuproine disulphonic acid (BCS). In contrast, BCS had no effect on either the rate of NO release, or the anti-platelet action of the non-nitrosothiol compound DEANO. 6. Cyclic GMP accumulation in response to GSNO (10(-9) 10(-5) M) was undetectable following treatment of platelets with ODQ (100 microM), a selective inhibitor of soluble guanylate cyclase. Despite this abolition of guanylate cyclase stimulation, GSNO retained some ability to inhibit aggregation, indicating the presence of a cyclic GMP-independent component in its anti-platelet action. However, this component was abolished following treatment of platelets with a combination of both ODQ and BCS, suggesting that Cu+ ions were required for the cyclic GMP-independent pathway to operate. 7. The cyclic GMP-independent action of GSNO, observed in ODQ-treated platelets, could not be explained by an increase in intra-platelet cyclic AMP. 8. The impermeable thiol modifying agent p-chloromercuriphenylsulphonic acid (CMPS) produced a concentration-dependent inhibition of aggregation of ODQ-treated platelets, accompanied by a progressive loss of detectable platelet surface thiol groups. Additional treatment with GSNO failed to increase the degree of aggregation inhibition, suggesting that a common pathway of thiol modification might be utilized by both GSNO and CMPS to elicit cyclic GMP-independent inhibition of platelet aggregation. 9. We conclude that NO donor compounds mediate inhibition of platelet aggregation by both cyclic GMP-dependent and -independent pathways. Cyclic GMP generation is related to the rate of spontaneous release of NO from the donor compound, but transfer of the NO signal to the cyclic GMP-independent pathway may depend upon a cellular system which involves both copper (I) (Cu+) ions and surface membrane thiol groups. The potent anti-platelet action of GSNO

Role of a copper (I)-dependent enzyme in the anti-platelet action of S-nitrosoglutathione.

1. S-nitrosoglutathione (GSNO) is a potent and selective anti-platelet agent, despite the fact that its spontaneous rate of release of nitric oxide (NO) is very slow. Our aim was to investigate the mechanism of the anti-aggregatory action of GSNO. 2. The biological action of GSNO could be mediated by NO released from S-nitrosocystylglycine, following enzymatic cleavage of GSNO by gamma-glutamyl transpeptidase. The anti-aggregatory potency of GSNO was not, however, altered by treatment of target platelets with the gamma-glutamyl transpeptidase inhibitor acivicin (1 mM). gamma-Glutamyl transpeptidase is not, therefore, involved in mediating the action of GSNO. 3. The rate of breakdown of S-nitrosoalbumin was increased from 0.19 +/- 0.086 nmol min-1 to 1.52 +/- 0.24 nmol min-1 (mean +/- s.e.mean) in the presence of cysteine (P < 0.05, n = 4). Inhibition of platelet aggregation by S-nitrosoalbumin was also significantly increased by cysteine (P < 0.05, n = 4), suggesting that the biological activity of S-nitrosoalbumin is mediated by exchange of NO from the protein carrier to form the unstable compound cysNO. Breakdown of GSNO showed a non-significant acceleration in the presence of cysteine, from 0.56 +/- 0.22 to 1.77 +/- 0.27 nmol min-1 (mean +/- s.e.mean) (P = 0.064, n = 4), and its ability to inhibit platelet aggregation was not enhanced by cysteine. This indicates that the anti-platelet action of GSNO is not dependent upon transnitrosation to form cysNO. 4. Platelets pretreated with the copper (I)-specific chelator bathocuproine disulphonic acid (BCS), then resuspended in BCS-free buffer, showed resistance to the inhibitory effect of GSNO. These findings suggest that BCS impedes the action of GSNO by binding to structures on the platelet, rather than by chelating free copper in solution. 5. Release of NO from GSNO was catalysed enzymatically by ultrasonicated platelet suspensions. This enzyme had an apparent K(m) for GSNO of 12.4 +/- 2.64 microM and a Vmax of 0.21 +/- 0.03 nmol min-1 per 10(8) platelets (mean +/- s.e.mean, n = 5). It was inhibited by BCS, but not by the iron chelator bathophenathroline disulphonic acid, nor by acivicin. 6. We conclude that the stable S-nitrosothiol compound GSNO may exert its anti-platelet action via enzymatic, rather than spontaneous release of NO. This is mediated by a copper-dependent mechanism. The potency and platelet-selectivity of GSNO may result from targeted NO release at the platelet surface.

Copper chelation-induced reduction of the biological activity of S-nitrosothiols.

1. The effect of copper on the activity of the S-nitrosothiol compounds S-nitrosocysteine (cysNO) and S-nitrosoglutathione (GSNO) was investigated, using the specific copper chelator bathocuproine sulphonate (BCS), and human washed platelets as target cells. 2. Chelation of trace copper with BCS (10 microM) in washed platelet suspensions reduced the inhibition of thrombin-induced platelet aggregation by GSNO; however, BCS had no significant effect on the anti-aggregatory action of cysNO. BCS inhibited cyclic GMP generation in response to both cysNO and GSNO. 3. The effect of BCS was rapid (within 30 s), and could be abolished by increasing the platelet concentration to 500 x 10(9) l-1. 4. In BCS-treated platelet suspensions, the addition of Cu2+ ions (0.37-2.37 microM) led to a restoration of both guanylate cyclase activation and platelet aggregation inhibition by GSNO. 5. The anti-aggregatory activity of GSNO was reduced in a concentration-dependent manner by the copper (I)-specific chelators BCS and neocuproine, and to a smaller extent by desferal. No effect was observed with the copper (II) specific chelator, cuprizone, the iron-specific chelator, bathophenanthroline sulphonate, or the broader-specificity copper chelator, D-penicillamine. 6. In both BCS-treated and -untreated platelet suspensions, cys NO was more potent than GSNO as a stimulator of guanylate cyclase. In BCS-treated platelet suspensions there was no significant difference between the anti-aggregatory potency of cysNO and GSNO; however, in untreated suspensions, GSNO was significantly more potent than cysNO. Thus, when copper was available, GSNO produced a greater inhibition of aggregation than cysNO, despite being a less potent activator of guanylate cyclase. 7. The breakdown of cysNO and GSNO was measured spectrophotometrically by decrease in absorbance at 334 nm. In Tyrode buffer, cysNO (10 microM) broke down at a rate of 3.3 microM min-1. BCS (10 microM)reduced this to 0.5 microM min-1. GSNO, however, was stable, showing no fall in absorbance over a period of 7 min even in the absence of BCS.8. We conclude that copper is required for the activity of both cysNO and GSNO, although its influence on anti-aggregatory activity is only evident with GSNO. The stimulatory effect of copper is unlikely to be explained solely by catalysis of S-nitrosothiol breakdown. The enhancement by copper of the anti-aggregatory activity of GSNO, relative to cysNO, suggests that copper may be required for biological activity of GSNO which is independent of guanylate cyclase stimulation.

Cell-mediated biotransformation of S-nitrosoglutathione.

Spontaneous release of nitric oxide (NO) from S-nitrosothiols cannot explain their bioactivity, suggesting a role for cellular metabolism or receptors. Using immortalised cells and human platelets, we have identified a cell-mediated mechanism for the biotransformation of the physiological S-nitrosothiol compound S-nitrosoglutathione (GSNO) into nitrite. We suggest the name "GSNO lyase" for this activity. GSNO lyase activity varied between cell types, being highest in a fibroblast cell line and lowest in platelets. In NRK 49F fibroblasts, GSNO lyase mediated a saturable, GSNO concentration-dependent accumulation of nitrite in conditioned medium, which was inhibited both by transition metal chelators, and by subjecting cells to oxidative stress using a combination of the thiol oxidant diamide and Zn2+, a glutathione reductase inhibitor. Activity was resistant, however, to both acivicin, an inhibitor of gamma-glutamyl transpeptidase (EC 2.3.2.2), and to ethacrynic acid, an inhibitor of Pi class glutathione-S-transferases (EC 2.5.1.18), thus neither of these enzymes could account for NO release. Although GSNO lyase does not explain the platelet-selective pharmacological properties of GSNO, cellular biotransformation suggests therapeutic avenues for targeted delivery of NO to other tissues.

Uremic concentrations of guanidino compounds inhibit neutrophil superoxide production.

BACKGROUND: In uremia, diminished reactive oxygen intermediate (ROI) production is an important consequence of impaired neutrophil function. We have studied the effect of guanidino compounds, known uremic toxins, on neutrophil ROI production in vitro. METHODS: Neutrophils from healthy volunteers were exposed for three hours to individual or mixed guanidino compounds (GCmix) at concentrations encountered in uremic plasma. After removal of guanidino compounds, neutrophils were activated by adhesion, N-formyl-methionyl-leucyl-phenyalanine (fMLP), phorbol 12-myristate 13-acetate (PMA), or opsonized zymosan, and superoxide production was measured by lucigenin chemiluminescence (CL). The direct effect of guanidino compounds on superoxide production in activated neutrophils was also measured. The energy status (ATP and creatine phosphate), antioxidant status (total glutathione), and glycolytic flux (lactate production) were measured. RESULTS: The GCmix pretreatment decreased the superoxide production in activated neutrophils (fMLP or zymosan) by 50% (P < 0.01) and the ATP concentration by 60% (P < 0.05), and it inhibited glycolytic flux (lactate production) by 45% (P < 0.01), but did not alter glutathione concentration. Simultaneous exposure to GCmix and activation did not inhibit nicotinamide adenine dinucleotide phosphate (NADPH) oxidase activity in cell lysates, but inhibited superoxide formation in zymosan-activated intact neutrophils, and this inhibition was reversed following removal of the guanidino compounds. CONCLUSION: Guanidino-succinate, -propionate, and -butyrate were individually as potent as the GCmix. Inhibition of neutrophil superoxide generation by guanidino compounds results from a depressed energy status. Uremic concentrations of guanidino compounds significantly inhibit neutrophil metabolism, and this has serious implications for their function in host defense.

An assessment of whole blood impedance aggregometry using blood from normal subjects and haemodialysis patients.

Impedance aggregometry allows the measurement of platelet responses in whole blood as well as in PRP. The variability of haematocrit values encountered when applying this technique to haemodialysis patients prompted an investigation of the effects of red cells on platelet aggregation in whole blood. Collagen induced aggregation was measured in both PRP and whole blood from haemodialysis patients and healthy controls. Platelets from haemodialysis patients were less aggregable than those from the controls when tested in PRP, but more aggregable when tested in whole blood. Blood samples with a range of haematocrit values were prepared by mixing PRP and autologous red cells, and used to study the effect of haematocrit on platelet aggregation. In blood from control subjects aggregation rate was reduced by rising haematocrit but no reduction of maximum aggregation occurred until haematocrit exceeded 40%. In contrast uraemic platelets showed increased responses in the presence of red cells. In a limited cross over study no significant difference was found in the effect on platelet aggregation of washed erythrocytes from uraemic and non-uraemic donors. It is concluded that red cell presence influences platelet aggregation by complex mechanisms during impedance aggregometry and that this effect must be considered when interpreting results.

Recombinant human erythropoietin shortens the uraemic bleeding time without causing intravascular haemostatic activation.

Blood rheology and haemostasis have been investigated in 8 haemodialysis patients during treatment with recombinant human erythropoietin (rHuEPO). The aim was to elucidate the mechanism by which rHuEPO improves haemostasis, and to determine whether rHuEPO promotes intravascular coagulation. Investigations were performed before, and after 3 months of treatment. Haemoglobin and haematocrit rose significantly after rHuEPO (p less than 0.001) and there was a concurrent shortening of the bleeding time. No significant changes were observed in platelet aggregation, thromboxane generation, or platelet nucleotide content during the treatment period. Whole blood viscosity increased following rHuEPO (p less than 0.01), but plasma viscosity and red cell deformability were unchanged, as were markers of intravascular platelet activation and plasma levels of cross-linked fibrin derivatives. No patient suffered from thrombosis during the study period, and elevation of the haematocrit in uraemic patients up to 0.35 with rHuEPO did not appear to lead to intravascular coagulation. Shortening of the prolonged bleeding time in haemodialyzed patients following rHuEPO appeared to be due to the increase in circulating red cells, rather than to changes in platelet reactivity.

An investigation of the protective effect of alpha+-thalassaemia against severe Plasmodium falciparum amongst children in Kumasi, Ghana.

INTRODUCTION: Several factors influence the severity of Plasmodium falciparum; here, we investigate the impact of alpha+-thalassaemia genotype on P. falciparum parasitemia and prevalence of severe anaemia amongst microcytic children from Kumasi, Ghana. METHODS: Seven hundred and thirty-two children (≤10 years) with P. falciparum were categorised into normocytic and microcytic (mean cell volume ≤76 fL). Microcytic individuals were genotyped for the -α(3.7) deletional thalassaemia mutation and parasite densities determined. RESULTS: Amongst microcytic patients both parasite densities and prevalence of severe malaria parasitemia (≥100 000/µL) were significantly lower (P < 0.001) in the presence of an alpha+-thalassaemia genotype compared with non-alpha+-thalassaemia genotype. There was no evidence that alpha+-thalassaemia protected against severe anaemia. The protection conferred by alpha-thalassaemia genotype against severe P. falciparum parasitemia did not change with increasing age. CONCLUSION: The severity of P. falciparum parasitemia was significantly lower in both the homozygous and heterozygous alpha+-thalassaemia groups compared with microcytic individuals with non-alpha+-thalassaemia genotype. The protective effect, from severe malaria, of the alpha+-thalassaemia allele does not alter with age.

S-nitrosothiols as selective antithrombotic agents - possible mechanisms.

S-nitrosothiols have a number of potential clinical applications, among which their use as antithrombotic agents has been emphasized. This is largely because of their well-documented platelet inhibitory effects, which show a degree of platelet selectivity, although the mechanism of this remains undefined. Recent progress in understanding how nitric oxide (NO)-related signalling is delivered into cells from stable S-nitrosothiol compounds has revealed a variety of pathways, in particular denitrosation by enzymes located at the cell surface, and transport of intact S-nitrosocysteine via the amino acid transporter system-L (L-AT). Differences in the role of these pathways in platelets and vascular cells may in part explain the reported platelet-selective action. In addition, emerging evidence that S-nitrosothiols regulate key targets on the exofacial surfaces of cells involved in the thrombotic process (for example, protein disulphide isomerase, integrins and tissue factor) suggests novel antithrombotic actions, which may not even require transmembrane delivery of NO.

Interactions between cell surface protein disulphide isomerase and S-nitrosoglutathione during nitric oxide delivery.

In this study, we investigated the role of protein disulphide isomerase (PDI) in rapid metabolism of S-nitrosoglutathione (GSNO) and S-nitrosoalbumin (albSNO) and in NO delivery from these compounds into cells. Incubation of GSNO or albSNO (1 microM) with the megakaryocyte cell line MEG-01 resulted in a cell-mediated removal of each compound which was inhibited by blocking cell surface thiols with 5,5'-dithiobis 2-nitrobenzoic acid (DTNB) (100 microM) or inhibiting PDI with bacitracin (5mM). GSNO, but not albSNO, rapidly inhibited platelet aggregation and stimulated cyclic GMP (cGMP) accumulation (used as a measure of intracellular NO entry). cGMP accumulation in response to GSNO (1 microM) was inhibited by MEG-01 treatment with bacitracin or DTNB, suggesting a role for PDI and surface thiols in NO delivery. PDI activity was present in MEG-01 conditioned medium, and was inhibited by high concentrations of GSNO (500 microM). A number of cell surface thiol-containing proteins were labelled using the impermeable thiol specific probe 3-(N-maleimido-propionyl) biocytin (MPB). Pretreatment of cells with GSNO resulted in a loss of thiol reactivity on some but not all proteins, suggesting selective cell surface thiol modification. Immunoprecipitation experiments showed that GSNO caused a concentration-dependent loss of thiol reactivity of PDI. Our data indicate that PDI is involved in both rapid metabolism of GSNO and intracellular NO delivery and that during this process PDI is itself altered by thiol modification. In contrast, the relevance of PDI-mediated albSNO metabolism to NO signalling is uncertain.

How cytotoxic is nitric oxide?

Nitric oxide (NO) shows an unusual divergence of action, being utilised both as a physiological signalling molecule, and as a toxic mediator. NO-mediated cellular injury may arise by a variety of mechanisms, including disruption of mitochondrial respiration, enzyme inhibition, lipid peroxidation and genetic mutation. Toxicity is largely mediated via intermediates such as N2O3 and peroxynitrite, arising from the reaction of NO with either molecular oxygen or reactive oxygen species. In general, such reactions become significant only when high concentrations of NO are generated by the induction of nitric oxide synthase.