INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

VASORELAXATION TO CARBACHOL and bradykinin in human radial arteries (RA) has nitric oxide-dependent and -independent components. The nitric oxide-independent component comprises, on average, ~50% of the total endothelium-dependent relaxation. It is inhibited by charybdotoxin and, to a greater extent, by charybdotoxin + apamin and has all the pharmacological characteristics of an endothelium-dependent hyperpolarizing factor (EDHF) (12). Its identity is unknown; however, epoxyeicosatrienoic acids (EETs) (5), K⁺ (7), cannabinoids (22), cytochrome P-450 epoxygenase products (8), nitric oxide (6), and H₂O₂ (18) have been suggested to act as EDHFs in different vascular beds.

H₂O₂, derived from nitric oxide synthase, was proposed to be an EDHF in mouse mesenteric arteries (18). It can induce vasorelaxation in porcine coronary arteries and activate Ca²⁺-dependent K⁺ (K_Ca) channels (1, 2). These relaxations are attenuated in the presence of charybdotoxin (13). K_Ca channels have also been shown to mediate H₂O₂-induced relaxations in canine trachealis (14). In human endothelial cells, Bychkov et al. (4) concluded that reactive oxygen species, generated locally, increase the K_Ca current amplitude.

Superoxide has been shown to be produced in endothelium and smooth muscle in human blood vessels, where it is converted to H₂O₂ by superoxide dismutases (SODs) (3). We therefore asked the question, Could H₂O₂ or another reactive oxygen species be contributing to the nitric oxide/prostanoid-independent relaxations in RA?

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

RA and internal mammary arteries (IMA) were surgically prepared as previously described (12). In the laboratory, they were cleaned and cut into 2- to 3-mm rings for organ chamber experiments (RA only) or 3- to 4-mm rings for superoxide measurement by lucigenin chemiluminescence (3). Organ chamber experiments were carried out as previously described (12). Briefly, vessels were cut into 2- to 3-mm rings and suspended on wires in 10-ml organ chambers filled with physiological salt solution maintained at 37°C and aerated with 95% O₂-5% CO₂. Indomethacin (0.02 mM) was added to the physiological salt solution to abolish contractile/relaxant effects of prostanoids. Rings were suspended under 2 g of resting tension and allowed to equilibrate for 1.5-2 h before any studies were performed.

Rings were contracted with 3 × 10⁶ M phenylephrine, and vasorelaxation to carbachol was studied, first, in the absence and, then, in the presence of SOD (50 U/ml) or polyethylene-glycolated cell-permeable SOD (PEG SOD; 10-100 U/ml) to dismutate superoxide or in the presence of catalase (1,000-2,000 U/ml) to scavenge H₂O₂ or vehicle. In further studies, relaxation to carbachol was studied in the absence of N^G-nitro-L-arginine methyl ester (L-NAME), again in the presence of 2 × 10⁴ M L-NAME, and, for a third time, in the presence of L-NAME and scavengers of reactive oxygen species or vehicle. The concentration of L-NAME was the same as that previously shown to produce maximal inhibition of relaxation to carbachol (12). To confirm this, vasorelaxation to carbachol after treatment with 400 µM L-NAME was also studied. SOD, catalase, and L-NAME were added to the organ chambers 30 min before construction of dose-response curves unless otherwise stated. The maximal relaxation (E_max) to carbachol and the concentration of agonist required to produce 50% relaxation (EC₅₀) were calculated for each ring using Microsoft Excel. Values before and after treatment with scavengers of reactive oxygen species were compared using paired t-tests; P = 0.05 was taken as significant. Values shown are means ± SE, with P values and 95% confidence intervals (CI) where appropriate.

Sections from six matched pairs of formalin-fixed, paraffin-embedded IMA and RA segments were incubated with a 1:1,000 dilution of sheep anti-human Cu/Zn SOD (Binding Site), followed by peroxidase-labeled streptavidin diluted 1:200, and then 0.03% diaminobenzidine to generate brown staining at the site of anti-SOD binding. In negative controls the anti-SOD antibody was omitted. For blinded semiquantitative grading of the immunostaining, a scale of 0-3 was used, with 0 representing no staining and 3 the maximum endothelial staining intensity corresponding to the intensity of staining with the endothelial marker Factor VIII-related antigen. Additional 4-µm-thick sections were stained using hematoxylin and eosin for each case and were examined immunohistochemically. These hematoxylin-and-eosin-stained sections were assessed for vessel wall integrity and endothelial continuity.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Levels of superoxide and SOD in IMA and RA. Superoxide was generated in IMA and RA; steady-state levels were 0.90 ± 0.19 nmol · mg wet wt¹ · min¹ in RA from 10 patients and 1.10 ± 0.19 nmol · mg wet wt¹ · min¹ in IMA from 12 patients. In the five patients from whom both RA and IMA were obtained, levels were 1.08 ± 0.34 and 1.04 ± 0.16 nmol · mg wet wt¹ · min¹, respectively.

View larger version (160K): [in this window] [in a new window]   Fig. 1.   Linear endothelial brown immunocytochemical positivity for Cu/Zn superoxide dismutase (SOD) in internal mammary artery (A) and radial artery (B). Intensity of endothelial positivity is greater in internal mammary artery than in radial artery. Magnification ×400.

Effect of inhibition of nitric oxide synthase on relaxation to carbachol in RA. Treatment of the arterial rings with 200 µM L-NAME attenuated but did not abolish relaxation to carbachol. In the absence of L-NAME, maximal relaxation to carbachol was 87 ± 10% in one series of experiments. After the addition of L-NAME to the organ baths, this was reduced to 43 ± 7%. Increasing the concentration of L-NAME and the duration of preincubation had no further effect on relaxation (40 ± 9% with 400 µM L-NAME and 1 h of preincubation).

Effect of catalase on vasorelaxation to carbachol in RA. Catalase had no significant effect on vasorelaxation in the absence or presence of L-NAME (Fig. 2, A and B). E_max values were 71 ± 10 and 65 ± 9% (n = 6) before and after treatment with catalase, respectively, in the absence of L-NAME. The reduction in E_max in the presence of catalase just failed to reach significance (P = 0.08, 95% CI = 1.88 to 19.08). Doubling the concentration of catalase added to the organ chambers had no additional effect on relaxation to carbachol. In the presence of L-NAME, relaxation curves to carbachol were virtually superimposable in the absence and presence of catalase: E_max was 54 ± 7 and 56 ± 7% (n = 6) before and after treatment with catalase, respectively. Catalase had no effect on EC₅₀ values for relaxation to carbachol in the absence or presence of L-NAME. Vehicle treatment did not modify E_max or EC₅₀ values for relaxation to carbachol.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Effect of catalase and SOD on nitric oxide-dependent and -independent relaxation to catalase and SOD in radial arteries. A and B: effect of catalase on relaxation to carbachol in the absence (n = 6) and presence (n = 6), respectively, of 200 µM L-NAME. C and D: effect of SOD (50 U/ml) on relaxation to carbachol in the absence (n = 14) and presence (n = 10), respectively, of L-NAME. , Absence of SOD or catalase; , presence of SOD or catalase.

Effect of superoxide scavenging on vasorelaxation to carbachol in RA. In the organ chamber studies, addition of SOD to the buffer bathing the tissues resulted in enhanced relaxation to carbachol in arteries not treated with L-NAME (Fig. 2C). E_max was 68 ± 5% before addition of SOD and 81 ± 6% in the presence of SOD (P = 0.002, 95% CI = 21.75 to 3.75, n = 14). The mean EC₅₀ was lower after SOD treatment (2.18 ± 0.65 vs. 3.13 ± 0.76 × 10⁷ M), but this difference was not significant. PEG SOD similarly increased vasorelaxation: E_max was 64 ± 8 and 74 ± 9% in the absence and presence of PEG SOD, respectively (P = 0.003, 95% CI = 11.27 to 3.45, n = 10). This SOD-induced increase in vasorelaxation to carbachol was dose related. PEG SOD at 10, 20, and 50 U/ml caused increases in E_max of 4 ± 2, 9 ± 1, and 10 ± 3%, respectively. Further increases of the SOD concentration in the organ chambers had no additional effect on vasorelaxation. In the presence of L-NAME, neither SOD nor PEG SOD affected relaxation to carbachol (Fig. 2D). E_max values were 64 ± 7 and 63 ± 8% in the absence and presence of SOD, respectively, and 58 ± 4 and 58 ± 4% in the absence and presence of PEG SOD, respectively.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study confirms previous work showing that ~50% of vasorelaxation to carbachol is independent of nitric oxide/prostanoid in RA. These nitric oxide/ prostanoid-independent relaxations have been shown to have the characteristics of an EDHF (12), but their identity is unknown. Here, we examined the hypothesis that H₂O₂ or another reactive oxygen species might act as an EDHF in RA.

The main source of H₂O₂ in vascular tissue is superoxide. Superoxide levels in RA were similar to those in IMA, whereas levels of Cu/Zn SOD, the enzyme that dismutates superoxide to H₂O₂, were lower in the endothelium of RA. Thus there is no evidence that H₂O₂ generation is higher in RA than in IMA, yet the contribution of non-nitric oxide/non-prostanoid mechanisms to endothelium-dependent vasorelaxation is much greater in RA than in IMA (12).

Catalase, which scavenges H₂O₂, converting it to water, has been reported to inhibit relaxation and hyperpolarization mediated by EDHF in mouse mesenteric arteries (18). However, in RA, catalase had no effect on nitric oxide/prostanoid-independent relaxation to carbachol. Furthermore, treatment with SOD or PEG SOD, which might be expected to enhance conversion of superoxide to H₂O₂, did not increase vasorelaxation to carbachol in RA pretreated with L-NAME to inhibit nitric oxide-mediated relaxations. In the absence of L-NAME, there was a trend toward a reduction in relaxation to carbachol in the presence of catalase, and it is possible that a small portion of the nitric oxide-dependent relaxation is due to H₂O₂, formed from nitric oxide, acting at K_Ca channels. Interestingly, there is evidence for a small contribution from K_Ca channels to nitric oxide-dependent vasorelaxation in IMA (10).

In the absence of L-NAME, a small but significant enhancement of vasorelaxation was observed when SOD was present in the baths. Previously, we showed a marked increase in vasorelaxation to carbachol in IMA treated with SOD (11). We believe that two reactive oxygen species interact. Superoxide scavenges nitric oxide, reducing nitric oxide bioavailability and vasodilation. The enhanced responses to carbachol in the presence of SOD under these conditions indicate that the concentrations of SOD were effective at dismutating superoxide to H₂O₂, and this resulted in an enhancement of the nitric oxide-dependent component of vasorelaxation to carbachol. In contrast, SOD did not attenuate the nitric oxide-independent component of carbachol-induced vasorelaxation, suggesting that the nitric oxide-independent component, unlike nitric oxide, is resistant to oxidative stress and is itself unlikely to be a reactive oxygen species. Consistent with these observations, resistance of EDHF-mediated vasorelaxation to oxidative stress has been reported in bovine coronary arteries and the perfused rat heart (9, 15).

If reactive oxygen species do not contribute to the nitric oxide/prostanoid-independent relaxations to carbachol in RA, what might? It was recently proposed (7) that K⁺ is an EDHF and that apamin and charybdotoxin block EDHF-dependent relaxation by inhibiting K⁺ release through endothelial small- and intermediate-conductance K_Ca channels. This hypothesis relies on a narrow myoendothelial space, in which released K⁺ accumulates to cause smooth muscle hyperpolarization by enhancing Ba²⁺-sensitive, inwardly rectifying K⁺ current and Na⁺-K⁺-ATPase. RA possess a prominent fibrointimal layer between the endothelium and smooth muscle measuring up to 58 µm thick (12). It seems unlikely that the endothelial cells could release sufficient K⁺ to affect the smooth muscle in these arteries.

A number of studies suggest that, in the coronary vasculature, EDHFs are cytochrome P-450-related products (8). EETs are a family of cytochrome P-450 epoxygenase metabolites of arachadonic acid. Campbell and colleagues (5) showed that EETs mediate the hyperpolarizing and vasodilator effects of acetylcholine, bradykinin, and arachidonic acid in bovine coronary arteries, and these substances are strong candidates for an EDHF in coronary arteries in a number of species. EDHF-dependent vasodilation has been observed in human coronary arteries (2). This EDHF-dependent vasodilation, similar to that in coronary arteries from other species, appears to be a cytochrome P-450-dependent metabolite (19), but whether EETs are involved remains to be determined. We previously showed that miconazole, an inhibitor of cytochrome P-450-dependent enzymes, attenuated the nitric oxide/prostanoid-independent relaxation to carbachol in RA. Inhibition by miconazole alone does not constitute firm evidence that the EDHF in RA is a cytochrome P-450-derived product, but it is consistent with the hypothesis. RA may be used as grafts in coronary artery bypass graft surgery, and it would be appropriate if these arteries possessed the same EDHF as native coronary vessels. However, the problem remains with RA as to how these endothelium-derived EETs might access smooth muscle to cause hyperpolarization.

Although we have not fully identified the nature of the nitric oxide/prostanoid-independent relaxations in RA, the observation that they are unlikely to be reactive oxygen species may have important pathophysiological consequences. The relative importance of nitric oxide-independent relaxation is increased in a number of animal models of cardiovascular disease in which levels of reactive oxygen species are increased (16, 17, 20). Although superoxide levels increase and nitric oxide-mediated vasodilation decreases in atherosclerotic disease, EDHF-mediated relaxation may remain unchanged or become upregulated, and its contribution to endothelium-dependent relaxation increases. This may be due, at least in part, to resistance of EDHF-mediated relaxation to oxidative stress. Thus there may be advantages in using vessels such as RA, in which an EDHF, independent of nitric oxide, makes a substantial contribution to endothelium-dependent vasorelaxation for revascularization surgery. However, much will depend on the final identification of EDHFs in different vascular beds.