Nitric Oxide, Endothelin-1, and Superoxide Production in Arterial Bypass Grafts

Abstract

In this study, basal and thrombin-stimulated release of nitric oxide and endothelin-1 in the internal mammary artery and the radial artery were measured, together with superoxide radicals generated after anoxia and reoxygenation. Arterial segments were obtained from patients undergoing coronary bypass operations. Quantification of nitric oxide was performed by measuring the stable oxidation products of nitric oxide. Endothelin levels were measured by an enzyme immunoassay kit, and the superoxides were measured by lucigenin-enhanced chemiluminescence. Basal and stimulated release of nitric oxide from the internal mammary artery is significantly higher than that in the radial artery. On the other hand, basal release of endothelin-1 is less in the internal mammary artery than in the radial artery, but similar after stimulation. In our study, the quantity of superoxide radicals produced by the internal mammary artery was greater than that produced by the radial artery. Our results show that there are differences between these 2 arteries in regard to production of nitric oxide, endothelin-1, and superoxide radicals. These differences may have a role in the process of atherogenesis and may contribute to long-term patency of arterial bypass grafts. These results may also explain the mechanism of radial artery graft spasm in coronary artery surgery and may constitute a basis for future pharmacological and clinical improvements for successful surgical application.

Various arteries and saphenous veins have been used as conduits for coronary artery bypass grafting (CABG). Clinical studies investigating the patency of conduits have shown that internal mammary artery (IMA) grafts are more resistant to atherosclerosis and more patent in the long term than are the other arteries and the saphenous vein. However, an increase in the number of patients who require multiple operations or multiple grafts for revascularization has resulted in the need for alternative conduits.

Use of the radial artery (RA) for CABG was first proposed by Carpentier and colleagues¹ in 1973, but that conduit was abandoned after reports of occlusion and spasm. In 1992, use of the RA as a conduit was re-proposed with some modifications of the technique and the adjunctive use of antispasmodic drugs.² Since then, many groups have reported encouraging results with this graft.^3,4 It has been thought that the implantation sites of RA and IMA grafts might account for differences in their patency rates, but the biological properties of grafts, which might contribute to such differences, have not been much explored.

One of the most important biological characteristics that may influence graft performance is endothelial function. It has been well established that the vascular endothelium synthesizes and releases potent vasoactive factors that play active roles in vascular biology and pathophysiology. Among the various compounds formed in the endothelium are nitric oxide (NO) and endothelin-1 (ET-1), vasoactive factors that strongly influence the modulation of vascular tone and the prevention and development of atherosclerosis.⁵ There is increasing evidence that NO has protective effects on the vasculature, whereas ET-1 affects the initiation and progression of atherosclerosis.⁵ In the intact blood vessel wall, there is continuous basal release of both NO and ET-1, which may indicate the capacity of the endothelium to release these substances. Mechanical forces such as shear stress and the activation of various receptors regulate the release of these vasoactive substances.⁶ However, the balance between the 2 antagonistic substances, together with other released factors and the reactivity of the smooth muscle cells, plays an important role in the determination of vascular tone and various other physiological processes.

In the vascular wall, endothelial cells and smooth muscle cells may also generate superoxide, which is involved in the pathogenesis of atherosclerosis through its effects on NO scavenging, on peroxynitrite generation, and on redox-sensitive cell-signaling pathways.⁷ During reperfusion after ischemia, reactive oxygen species are formed, and these are responsible for the free-radical–induced peroxidation and the consequent tissue injury.

In this study, we investigated the differences between the RA and IMA with regard to the formation of compounds that may have protective or detrimental roles in graft atherosclerosis. We measured the quantity of the vasoactive molecules NO and ET-1 released under basal and stimulated conditions and of superoxide radicals generated after anoxia and reoxygenation. Thrombin was used to stimulate the endothelium's production of both NO and endothelin.

Materials and Methods

Radial artery and IMA segments were collected from patients who underwent CABG. The study protocol was approved by the ethics committees of the Marmara University School of Medicine and the Istanbul Memorial Hospital. Informed consent was obtained from the participating patients. The same patients donated both RA and IMA samples. Patients with diabetes mellitus or with any known vasculopathy were excluded from the study. The vessels of participants were dissected from the surrounding connective tissue and cut into 4-mm-long rings, with great care to preserve the endothelium. From each artery, 3 segments of similar length and weight were prepared and used in the study.

Drugs. In the present study, thrombin was used as a stimulant because of its well-established simultaneous effects on NO and ET-1 production.⁷ In endothelial cell cultures, previous studies and our own preliminary experiments^8,9 had demonstrated NO release in response to thrombin during the first 10 to 15 minutes and ET-1 release between 4 and 12 hours. Thrombin was prepared in sterile distilled water and used in the experiments in a concentration of 4 U/mL. It has been suggested that, after stimulation, the physiologic dose of thrombin ranges between 0.025 and 5 U/mL.^10,11

Nitric Oxide Measurements. Arterial segments were washed and incubated for 1 hour in Tyrode solution, at 37 °C in 95% room air and 5% carbon dioxide (CO₂). Then the segments were incubated in 1 mL of buffer for 10 minutes with or without thrombin (4 U/mL). After the incubation, the medium and the vessels were frozen separately at –20 °C until evaluation of the nitrite levels.

The quantification of NO released from the segments was performed by measuring NO's stable oxidation products (NO_x) in the arterial incubation solution. The metabolic products of NO are nitrate (NO₃) and nitrite (NO₂). Nitrates were reduced to nitrites, and then nitrite concentration was measured calorimetrically by the Griess reaction.¹²

Endothelin-1 Measurements. Segments to be used for evaluation of ET-1 production were washed and kept in Krebs-Ringer solution. After 1 hour of equilibration, the segments were incubated for 6 more hours in solution with or without thrombin (4 U/mL). The incubation time was chosen according to the results of preliminary experiments, which showed that differences in ET-1 levels were detectable after 6 hours of incubation. The incubation solution and the arterial segments were kept at –20 °C until ET-1 levels were determined.

Immunoreactive ET-1 concentration in the medium was assayed using a commercially available RIA kit (RIK 6901, Peninsula Laboratories; Belmont, Calif). The medium was extracted with Sep-Pak®C₁₈ cartridges (Waters Corporation; Milford, Mass) and evaporated under nitrogen gas. The reconstituted solution was processed according to the manufacturer's instructions. The cross-reactivity of the kit with human big ET-1 is 17%, and with human endothelin-2 and -3 it is 7%. The intra-assay coefficient of variation is less than 5%, and the interassay coefficient of variation is less than 15%.

Measurement of Superoxide Radicals. Segments to be used for evaluation of superoxide generation were washed and kept in HEPES buffer during dissection and weight and length measurements. After 45 minutes of equilibration in fresh HEPES buffer at 37 °C in a CO₂ incubator (95% O₂ and 5% CO₂), the segments were incubated for 60 minutes under anaerobic conditions, with nitrogen gas. Then the segments were oxygenated by incubating with 95% O₂ and 5% CO₂ for 60 minutes.

Superoxide radical formation after anoxia and reoxygenation was estimated by lucigenin-enhanced chemiluminescence (CL) using a liquid scintillation counter (Tricarb 1500, Packard Instruments; Downer's Grove, Ill) in and out of coincidence mode with a single active photomultiplier tube.

Arterial segments were put into vials containing Krebs-HEPES buffer (NaCl, 99.01 mmol/L; KCl, 4.69 mmol/L; CaCl₂, 1.87 mmol/L; MgSO₄, 1.2 mmol/L; K₂HPO₄, 1.03 mmol/L; NaHCO₃, 25.0 mmol/L; Na-HEPES, 20.0 mmol/L; glucose, 11.1 mmol/L; pH, 7.4), and CL was recorded after the addition of lucigenin as an enhancer.¹³ Counts were obtained at 1-minute intervals, and the results were given as CL (cpm/mg tissue). For comparison of data, area under curve (AUC) values were calculated for total measurement time, and the results were given as AUC of cpm/mg tissue.

Data Analysis. The data were reported as mean ± SEM. Data were evaluated for statistical significance using the paired Student's t-test for dependent observations. A P value of less than 0.05 was considered to be statistically significant.

Results

NO_x. Data obtained from measurements of the stable oxidation products of NO in the arterial bathing solution after 10 minutes of incubation under basal conditions or after stimulation with thrombin are shown in Figure 1. The amount of NO released from IMA segments was significantly higher than the amount from RA (P <0.01). Under basal conditions, the quantities of NO_x accumulated in the IMA and RA incubation solutions were 159.09 ± 9.71 and 53.24 ± 2.52 pmol/mg tissue, respectively. After stimulation with thrombin, a rapid increase in NO_x production was observed for both arterial segments. After stimulation, the quantities of NO_x released from IMA and RA segments were 210 ± 11.94 and 72.64 ± 3.35 pmol/mg tissue, respectively. There were significant differences between the quantities of NO_x produced by the IMA and the RA, both under basal conditions and after stimulation (P <0.01).

Fig. 1 The quantity of nitrite accumulated in the incubation media of IMA and RA in 10 minutes under basal conditions and after stimulation with thrombin (4 U/mL). Nitrite is the stable oxidation product of NO. Data are mean ± SEM of 22 experiments. The basal and stimulated nitrite production of IMA is significantly greater than that of RA (^*P <0.01).

IMA = internal mammary artery; RA = radial artery

Endothelin-1. The quantities of ET-1 released from arterial segments under basal conditions or after stimulation with thrombin were assayed in the bathing solution after 6 hours of incubation and were expressed as fmol ET-1 per mg of arterial segment (Fig. 2). Under basal conditions, the amount of immunoreactive ET-1 released from the IMA was 104 ± 14 fmol/mg tissue, and the amount released from the RA was 144 ± 21 fmol/mg tissue. After stimulation with thrombin, ET-1 release from the arteries increased to 205 ± 39 fmol/mg tissue for the IMA and 201 ± 25 fmol/mg tissue for the RA. While the difference between basal ET-1 levels of the IMA and the RA was statistically significant (P <0.01), stimulated ET-1 responses were almost the same.

Fig. 2 The quantity of endothelin-1 (ET-1) released from the IMA and the RA under basal conditions and after stimulation with thrombin. The quantity in the medium after 6 hours of incubation is given as fmol/mg tissue. Data are mean ± SEM of 22 experiments. The differences between ET-1 concentrations in the incubation medium of the IMA and the RA were statistically significant (P <0.01) under basal conditions but not after stimulation with thrombin. When compared with basal values, thrombin-induced increases in ET-1 levels were significant in both the IMA (^*P <0.01) and the RA (^**P <0.05).

IMA = internal mammary artery; RA = radial artery

Superoxide Generation. Generation of superoxide was evidenced by increases in CL. After anoxia and reoxygenation, an increase in CL was observed in segments of IMA within 50 to 100 minutes, but this rapidly returned to basal levels. In segments of RA, we observed a gradual increase in CL (reaching peak values in 150 minutes), and then a gradual decrease (Fig. 3). The total quantity of superoxide radicals, expressed as AUC of cpm/mg tissue, was 1,365,819 ± 258,988 for IMA and 784,485 ± 119,810 for RA. The difference was statistically significant (P <0.01).

Fig. 3 Superoxide generation in RA and IMA after anoxia and reoxygenation estimated by lucigenin-enhanced chemiluminescence. Each point represents the mean ± SEM of 12 experiments. The graph demonstrates that the quantity of reactive oxygen species produced by the IMA is greater, although the duration of production is shorter. The difference between the total amount of superoxide radicals produced by the arteries was statistically significant (P <0.01).

IMA = internal mammary artery; RA = radial artery

Discussion

The results of the present study show that basal and thrombin-stimulated release of NO from the IMA is higher than from the RA, while the basal release of ET-1 from the IMA is less than from the RA, but similar after stimulation. The data also demonstrate that after ischemia/reperfusion, fewer superoxide radicals are formed in the RA, compared with the IMA. These findings indicate that there are important functional differences between these 2 arteries.

The NO formed in the endothelium is a potent vasodilator that plays an important role in the regulation of vascular tone and in the prevention of atherosclerosis and the development of cardiac graft atherosclerosis. There is increasing evidence that NO activity is impaired in atherosclerosis; decreased NO synthesis may be one of the earliest events that leads to endothelial dysfunction and subsequently to atherosclerosis.^5,14 Supplementation of the NO pathway has been shown to aid in the prevention of both native and transplant atherogenesis.¹⁴

Since endothelial NO production appears to be quite important in maintaining a healthy and patent conduit, higher basal and stimulated release of NO from the IMA might in part explain why the IMA is less prone to atherosclerosis than is the RA. Angiographically determined patency rates reported by various groups show that IMA grafts exhibit the highest patency after CABG. Acar and associates,³ for example, reported that the patency of the radial artery was 83% compared with 91% for the IMA. In addition, a histological study that examined the pre-existing disease in the RA, the IMA, and saphenous veins obtained from the same patients showed an increased prevalence of mild intimal thickening and medial sclerosis in the RA, compared with the IMA.¹⁵ Apart from us, other investigator's^16–18 have studied NO production and endothelium-dependent relaxation in the RA and the IMA, but they have reported conflicting results. However, those investigators who measured NO directly, with NO-sensitive electrodes, reported results similar to those of our present study: basal and stimulated releases of NO in the IMA are greater than those in the RA.¹⁷

Endothelin-1, another factor that may affect the performance of a graft, is produced both under basal conditions and after stimulation with agonists such as thrombin, arginine vasopressin, and angiotensin.⁶ It is a potent vasoconstrictor that also stimulates vascular smooth muscle cell proliferation and migration, and monocyte adhesion. Although the physiologic role of ET-1 has not been fully established, there is increasing evidence of its involvement in pathological conditions such as hypertension, atherosclerosis, myocardial infarction, congestive heart failure, and vasospasm related to cardiac operations.

Clinical studies have revealed that the RA is more spastic than other arterial grafts, and spasmolytic agents are required during graft preparation and in the postoperative period. When the response of various arterial grafts to vasoconstrictors was investigated, it was reported^19,20 that the RA had higher receptor-mediated contractility than did the IMA, in response to agonists such as ET-1, angiotensin II, serotonin, and norepinephrine. This finding was used to explain the higher propensity of the radial artery to spasm that is observed in clinical settings. However, the main cause of perioperative vasospasm of the RA remains unknown, as do the factors involved in long-term graft patency. The differences in endogenous production of vasoconstrictor compounds have not been explored previously.

The IMA might be expected to respond to circulating endothelin levels with more spasm than does the RA, since the IMA is generally exposed to lower amounts in its basal state. However, He and Yang¹⁹ have shown that the RA has higher receptor-mediated contractility as a response to ET-1 stimulation than does the IMA. Moreover, ET-1 production in the IMA after stimulation may be balanced by increased NO production. There are also some morphological differences between the IMA and the RA: the RA is a muscular artery, and the IMA is an elastic artery. Muscular arteries have thicker media than do elastic arteries, so muscular arteries may have a higher contractility response to vasoconstrictor stimulators.

The differences in both NO and endothelin production may be partly responsible for the clinically observed differences between the IMA and the RA. Higher NO but lower basal ET-1 production in the IMA may enhance vasodilator capacity and attenuation of the vasoconstrictor endothelin, whereas lower NO and higher basal endothelin production of the RA, together with its higher contractility, may contribute to the spastic character of that artery.

The differences in thrombin stimulation may be attributable to different receptor levels for thrombin in RA and IMA tissue beds.²¹ Moreover, in vascular injury, mitogen-activated protein kinases (MAPK) have been implicated in cell proliferation, migration, and apoptosis.²² Although the mechanisms of cell growth regulation are not completely understood, recent evidence indicates that the protein kinase cascade of c-Raf/MEK/MAPK—activated by tyrosine kinase receptors or G-protein–coupled receptors—is an important mechanism for transmitting extracellular growth signals into the cellular nucleus, which activates transcription factors and regulates downstream gene expression and cell cycle progression.^23,24 Therefore, the differences between RA and IMA that are evoked by thrombin stimulation can be further explained by possible differences in these vessels' receptor-mediated signal transduction mechanisms.

Our study also indicates that there are differences between the IMA and the RA in their capacities to generate superoxide. It has been shown that in both human arteries and veins, superoxide release is substantially modulated by interaction with endothelium-derived NO, which reduces the bioavailability of radicals and produces peroxynitrite.¹⁷ The present data indicate that after anoxia and reoxygenation, less superoxide is formed by the RA than by the IMA, which suggests that the RA may be more resistant to ischemic injury and to subsequent atherosclerosis. Higher superoxide production in the IMA might be balanced by increased NO production; studies have shown that superoxide or peroxynitrite in common with NO can act as a mediator of endothelium-dependent relaxation in human vessels by peroxynitrite-induced nitration of G proteins or other membrane components.^25,26 The radial artery is a muscular artery with a thicker media and a prominent tunica adventitia. Differences in morphology, together with the activity of antioxidant systems present on the vascular wall, might affect superoxide production and susceptibility to ischemic injury. Despite lower NO production, decreased superoxide levels in the RA must be investigated further through in vivo studies, especially in regard to the antioxidant systems.

The present study is ex vivo. The superoxide–NO balance may be different in arteries in vivo, because of the tonic regulation of endothelial NO production by shear stress and other activators and sources of NO. The balance between these 2 vasoactive substances varies under different biological conditions, which may account for the differences in superoxide and NO release from human arteries in previous studies.¹⁸

In conclusion, there are differences between the IMA and the RA in their release of important vasoactive mediators. Although the influence of other endothelial characteristics needs to be further defined, we conclude that these differences may have a role in the process of both native and graft atherosclerosis. This may also lend insight to the mechanism of RA graft spasm in coronary artery surgery and constitute a basis for pharmacological and clinical improvements in the future application of RA grafts.

Acknowledgments

We want to thank our colleagues for their contributions to this study: Ergun Demirsoy, MD; Harun Arbatli, MD; Goncagul Haklar, MD; and Ahmet Suha Yalcin, PhD.