INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

ENDOTHELIAL CELLS PLAY A KEY ROLE in the regulation of vascular tone (7). Previous studies (7, 10, 17) have demonstrated that the venous endothelium produces and releases lower amounts of vasodilator substances than endothelial cells from corresponding arteries. This phenomenon is believed to be one of the major contributors to the high failure rates of venous grafts (8).

Several investigators (6, 11-13) have suggested that the presence of valves in the venous wall may contribute to graft failure. Although the exact mechanism by which venous valves may participate in the pathogenesis of the graft failure is still unknown, it is likely that the endothelial cells at the valvular site of the vein may have different functional characteristics. The function of valvular endothelial cells has not been studied. Therefore, the present study was designed to characterize mechanisms of endothelium-dependent relaxations in valvular segments of the isolated canine saphenous vein.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

In vitro studies. Saphenous veins were taken from mongrel dogs (18-27 kg) anesthetized with 30 mg/kg of intravenous Pentothal. All procedures were in accordance with the Institutional Animal Care and Use Committee guidelines of Mayo Clinic. To remove blood, vessels were gently rinsed with cold modified Krebs-Ringer bicarbonate solution [control solution; containing (in mM) 118.3 NaCl, 4.7 KCl, 2.5 CaCl₂, 1.2 MgSO₄, 1.2 KH₂PO₄, 25.0 NaHCO₃, 0.026 calcium-ethylenediamine-tetraacetic acid, and 11.1 glucose]. Loose perivascular tissue and the adventitia were carefully removed. The experiments were performed on rings with and without valves. The presence of valves was identified by examination of the vein under a dissecting microscope and verified by examination of the intraluminal surface of the rings after an organ chamber study. Each ring (5 mm long) was connected to an isometric force-displacement transducer (Grass FT03; Grass Instrument, Quincy, MA) and suspended in an organ chamber filled with 25 ml of control solution (at 37°C, pH 7.4) aerated with 94% O₂-6% CO₂. Isometric force was recorded continuously. The rings were allowed to stabilize at a resting tension of 1-1.5 g for 30-45 min. Each ring was then gradually stretched to the optimal point of its length-tension curve (~2.0 g) as determined by repeated exposure to 30 mM KCl. In certain rings of saphenous veins, the endothelium was removed mechanically by gentle rubbing of the intimal surface with a stainless steel wire (31-gauge diameter). The successful removal of the endothelium was verified by the absence of relaxation induced by 10⁶ M of A-23187.

Measurement of cGMP. A radioimmunoassay technique was used to determine the levels of cGMP, as reported previously (22). Rings were initially incubated in minimal essential media in an CO incubator at 37°C for 30 min. After this 30-min period, rings were incubated for another 30 min in 3-isobutyl-1-methylxanthine (IBMX; 10³ M) to inhibit the degradation of cyclic nucleotides by phosphodiestarases. During the last 1 min of the 30-min incubation, certain rings were stimulated with 10⁶ M A-23187. The rings were then removed from the medium and quickly frozen in liquid nitrogen. After homogenization, cGMP levels were measured by a cGMP radioimmunoassay kit (Amersham, Arlington Heights, IL). Protein assay was conducted by a detergent compatible (DC) Protein Assay Kit (Bio-Rad, Hercules, CA). Venous rings taken from the same dogs were studied in parallel.

Drugs. The following pharmacological agents were used: U-46619, A-23187, forskolin (FK), N^G-nitro-L-arginine methyl ester (L-NAME), indomethacin, IBMX, papaverine hydrochloride (Sigma, St. Louis, MO), diethylaminodiazen-1-ium-1,2-dioate (DEA-NONOate; Cayman Chemical, Ann Arbor, MI), and 1H-[1,2,4]-oxadiazolo-[4,3-a]quinozalin-1-one (ODQ; BioMol Research Laboratory, Natick, PA). Drugs were dissolved in distilled H₂O, and volumes of <0.15 ml were added to the organ chambers. Solutions of DEA-NONOate and FK in the highest concentration were prepared in 1.5 M Tris (pH 8.8) and DMSO (Sigma), respectively. Concentrations of all drugs are expressed as the final molar concentration in the control solution. The rings were contracted with 3 × 10⁷ M U-46619 before the addition of vasodilating agents. This concentration of U-46619 causes ~80% of maximal contraction. Concentration-response curves were obtained in a cumulative fashion. Several rings prepared from the same vein were studied in parallel, and a concentration-response curve was established for each preparation. The relaxation response was expressed as the percent decrease in force from contraction induced by U-46619. Either L-NAME or ODQ was added for 15 min, indomethacin was added 30 min before the concentration-response curve to vasodilating agent was obtained, respectively.

Statistical analysis. The results of this study are expressed as means ± SE; n refers to the number of dogs. Contractions to KCl (30 mM), U-46619 (3 × 10⁷ M), and cGMP levels between the with and without valve groups were analyzed by unpaired t-test. cGMP levels in each group were compared by ANOVA. The concentration-response curves were analyzed by repeated measured ANOVA followed by Bonferroni/Dunn's post hoc test. Statistical significance was accepted at a probability level of <0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Endothelium-dependent relaxation to A-23187. Contractions to KCl (30 mM) and U-46619 (3×10⁷ M) were not significantly different between rings with and without valves (Table 1). During contraction to U-46619 (3×10⁷ M), A-23187 (10⁹-10⁶ M) caused endothelium-dependent relaxations (Fig. 1A). In rings with valves, endothelium-dependent relaxation was significantly attenuated compared with the relaxation in the rings without valves (Fig. 1B; maximal relaxation = 22.4 ± 3.47 and 48.5 ± 5.01%, respectively, n = 18, P < 0.0001). Removal of the endothelium abolished these relaxations to A-23187 (data not shown).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Concentration-response curves to A-23187 in canine saphenous vein rings with and without valves. A: original trace; B: dose-response plot. Relaxations were obtained during contractions induced by U-46619 (3×107 M). Data are shown as means ± SE and are expressed as percent decrease in force from contraction to U-46619; 100% = 21.5 ± 1.53 (n = 18) and 23.4 ± 1.12 g (n = 18) with and without valves, respectively. Endothelium-dependent relaxation in rings with valves was significantly attenuated compared with that in rings without valves; P < 0.0001 by repeated measures ANOVA. PPV, papaverine.

Endothelium-independent relaxations to FK and DEA-NONOate. During contractions to U-46619 (3 × 10⁷ M), FK (10⁸-10⁵ M) and DEA-NONOate (10⁹-10⁵ M) caused concentration-dependent relaxations (Fig. 2 and 3). Relaxation responses to these agonists were almost identical in rings with and without valves (n = 8).

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Concentration-response curves to forskolin (FK) in canine saphenous vein rings with and without valves. A: with endothelium; B: without endothelium. Relaxations were obtained during contractions induced by U-46619 (3×107 M). Data are shown as means ± SE and are expressed as percent decrease in force from contraction to U-46619; 100% = 19.9 ± 1.71 and 22.1 ± 3.45 g (A; n = 8), 18.9 ± 1.68 and 20.7 ± 2.17 g (B; n = 8) with and without valves, respectively. Endothelium-independent relaxation to FK was identical in rings with and without valves.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Concentration-response curves to diethylaminodiazen-1-ium-1,2-dioate (DEA-NONOate) in canine saphenous vein rings with and without valves. A: with endothelium; B: without endothelium. Relaxations were obtained during contractions induced by U-46619 (3×107 M). Data are shown as means ± SE and are expressed as percent decrease in force from contraction to U-46619; 100% = 21.6 ± 1.49 and 20.6 ± 3.04 g (A; n = 8), 23.7 ± 1.61 and 26.6 ± 3.15 g (B, n = 8) with and without valve, respectively. Endothelium-independent relaxation to DEA-NONOate were identical in rings with and without valves.

Effects of indomethacin and L-NAME on A-23187-induced relaxation. Indomethacin (10⁵ M) attenuated A-23187-induced relaxation in rings with and without valves (Fig. 4, A and B; P < 0.05, n = 18 in control and n = 8 in treated group, respectively). On the other hand, L-NAME (3×10⁴ M) attenuated the relaxation only in veins without valves (Fig. 4B; P < 0.05). L-NAME did not affect relaxations to A-23187 in rings with valves (Fig. 4A; P = 0.52). In both groups, indomethacin and L-NAME together almost abolished relaxations to A23187 (Fig. 4, A and B; P < 0.001)

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Effects of indomethacin (105 M) and NG-nitro-L-arginine methyl ester (L-NAME; 3×104 M) on concentration-response curves to A-23187 in canine saphenous veins with (A) and without valves (B). Relaxations were obtained during contractions induced by U-46619 (3×107 M). In the rings with valves, indomethacin almost completely inhibited the relaxation to A-23187 (P < 0.05 by repeated measures ANOVA), although L-NAME had no effect on the relaxation. On the other hand, in the rings without valves, both indomethacin and L-NAME partially inhibited the relaxation to A-23187 (P < 0.05 by repeated measures ANOVA), and indomethacin and L-NAME together abolished the relaxation (P < 0.001 by repeated measures ANOVA). Data are shown as means ± SE and are expressed as percent decrease in tension from contraction to U-46619; 100% = 21.5 ± 1.53 (n = 18), 20.0 ± 1.67 (n = 8), 22.6 ± 2.63 (n = 8), and 21.9 ± 3.27 g (A; n = 8), 23.4 ± 1.12 (n = 18), 24.1 ± 2.47 (n = 8), 26.0 ± 1.64 (n = 8), and 20.9 ± 2.95 g (B; n = 8) for the control, indomethacin, L-NAME, and indomethacin + L-NAME-treated groups, respectively.

Effects of ODQ on relaxations in response to A-23187 and DEA-NONOate. ODQ (3×10⁶ M) had no effect on relaxation induced by A-23187 in rings with and without valves (Fig. 5A; n = 5, P = 0.74 and 0.64, respectively). On the other hand, ODQ almost completely inhibited relaxations in response to DEA-NONOate (Fig. 5B; n = 4, P < 0.001).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Effects of 1H-[1,2,4]-oxadiazolo-[4,3-a]quinalin-1-one (ODQ; 3×106 M) on concentration-response curves to A-23187 (A) and DEA-NONOate (B) in canine saphenous veins with and without valves. Relaxations were obtained during contractions induced by U-46619 (3×107 M). ODQ had no effect on the relaxation to A-23187; however, ODQ significantly inhibited the relaxation to DEA-NONOate (P < 0.001 by repeated measures ANOVA). Data are shown as means ± SE and are expressed as percent decrease in tension from contraction to U-46619; 100% = 19.1 ± 1.99 vs. 23.9 ± 1.88 g (n = 5), 20.9 ± 2.90 vs. 18.6 ± 2.36 g (n = 5) with vs. without valve for control rings and rings treated with ODQ, respectively (A); 22.8 ± 2.07 vs. 28.4 ± 2.25 g (n = 4), 22.5 ± 2.60 vs. 26.6 ± 1.56 g (n = 4) with vs. without valve for control rings and rings treated with ODQ, respectively (B).

Basal and stimulated cGMP productions. Under basal conditions and during stimulation with A-23187, cGMP production in rings with valves was significantly lower than that in rings without valves (Fig. 6A; n = 7, P < 0.05). Pretreatment with L-NAME (3×10⁴ M), ODQ (3×10⁶ M) and removal of the endothelium abolished cGMP production in both groups (Fig. 6A; n = 7, P < 0.0001). Exposure of venous rings to A-23187 (10⁶ M) for 60 s resulted in a significant increase over basal cGMP levels in both groups of veins (P < 0.001). DEA-NONOate (10⁵ M) produced a significant elevation of cGMP in rings with and without valves. Pretreatment with ODQ (3×10⁶ M) significantly inhibited the DEA-NONOate-induced increase in cGMP (Fig. 6B; P < 0.001 , n = 5).

View larger version (15K): [in this window] [in a new window]   Fig. 6.   The levels of cGMP production in response to A-23187 (A) and DEA-NONOate (B) and the effects of L-NAME or ODQ and denudation of endothelium. Columns and bars represent means ± SE. Stimulation with A-23187 significantly increased the cGMP production in both groups (P < 0.001), but production of cGMP was significantly higher in veins without valves (P < 0.001). L-NAME or ODQ and the removal of endothelial cells significantly reduced formation of cGMP (P < 0.001). DEA-NONOate significantly increased production of cGMP (P < 0.001), and ODQ significantly reduced this effect (P < 0.001).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present study demonstrates that endothelium-dependent relaxation to A-23187 is significantly reduced in segments of saphenous vein with valves. A lower production of nitric oxide (NO) in endothelial cells of valvular segments as well as uncoupling of cGMP elevation from relaxation appear to be the mechanisms responsible for reduced reactivity to A-23187.

In contrast to endothelium-dependent relaxations, reactivity of smooth muscle cells to NO and the adenylate cyclase activator FK were identical in valvular and nonvalvular venous segments. These findings indicate that the endothelium rather than smooth muscle is responsible for the differential reactivity to A-23187. This conclusion was further supported by the fact that the contractile effects of potassium and a thromboxane A₂ analog, U-46619, were also identical in valvular and nonvalvular venous segments. Several previous studies (7, 10, 17) demonstrated that venous endothelial cells produce and release lower amounts of relaxing factors than the arterial endothelial cells. In the present study, the maximal endothelium-dependent relaxation to A-23187 was 48%. This is consistent with results reported by several different laboratories (9, 17, 19). However, in valvular segments, A-23187 produced a maximal relaxation of only 22%, suggesting that the valvular endothelium has a more limited capacity to release vasodilator substances. It is important to keep in mind that, in our study, the relaxation of smooth muscle cells reflects mostly abluminal release of endothelium-derived vasodilators. Whether the venous endothelium can release higher amount of relaxing factors into the venous lumen remains to be determined. However, our findings are consistent with the concept that reduced release of relaxing factors toward underlying smooth muscle cells in valvular segments is designed to decrease the risk of excessive venodilatation and subsequent valvular incompetence (2, 13a, 15, 16).

To elucidate the mechanisms responsible for endothelium-dependent relaxation to A-23187 in the canine saphenous vein, we used pharmacological tools and examined the effects of a NO synthase inhibitor, L-NAME, and a cyclooxygenase inhibitor, indomethacin. Our results obtained in nonvalvular segments are consistent with previous reports (1, 14) demonstrating that both NO and prostacyclin are released from canine saphenous vein endothelial cells activated by A-23187. It is interesting that, in valvular venous segments, relaxation to A-23187 was exclusively dependent on activation of arachidonic acid metabolism via the cyclooxygenase pathway. Neither L-NAME nor the guanylate cyclase inhibitor ODQ had any effect on endothelium-dependent relaxation in valvular segments, suggesting that NO does not contribute to the effect of A-23187. Whether this is due to low expression of endothelial NO synthase or some other mechanism (e.g., formation of superoxide anions) that may inhibit the biosynthesis or activity of NO remains to be determined.

Measurements of cGMP production demonstrated that NO is produced in segments both with and without valves. Consistent with the results of vasomotor studies, a significantly higher level of cGMP was detected in veins without valves under basal conditions and during activation of endothelial cells with A-23187. The removal of endothelial cells, inhibition of NO synthase with L-NAME, or inhibition of guanylate cyclase with ODQ significantly reduced production of cGMP. These results strongly suggest that the cGMP levels reflect NO biosynthesis in endothelial cells of the saphenous vein. The exact reason why, in veins with valves, A-23187-induced production of NO does not contribute to endothelium-dependent relaxation is unclear. It is possible that lower production of NO in valvular segments is insufficient to affect vascular tone.

ODQ almost abolished the production of cGMP in response to A-23187 or the NO donor DEA-NONOate in both valvular and nonvalvular venous segments. In contrast, ODQ did not affect endothelium-dependent relaxation to A-23187, suggesting that the effect of endogenously released NO is mediated by a mechanism independent of cGMP elevation. This conclusion is further supported by the fact that ODQ inhibited both relaxation and cGMP production in response to exogenous NO generated by DEA-NONOate, demonstrating that we used a sufficiently high concentration of ODQ to achieve inhibition of soluble guanylate cyclase. Our conclusion is also in agreement with the reported dissociation between endothelium-dependent relaxations to A-23187 and increases in cGMP detected in the canine femoral vein (23). Indeed, several groups (5, 21) reported that NO can induce smooth muscle relaxation by mechanisms independent of cGMP. A recent study (4) suggested that activation of both potassium channels as well as soluble guanylate cyclase could be involved in mediation of NO-induced relaxation in canine femoral veins. Bolotina et al. (3) demonstrated that NO can directly activate Ca²⁺-dependent K⁺ channels, which could cause hyperpolarization and relaxation of vascular smooth muscle independently of cGMP. Although we did not use K⁺ channel inhibitors in the present study, our results are consistent with the idea that endogenous NO causes relaxation by mechanism other than activation of guanylate cyclase. However, the results of the present study do not allow any conclusion concerning the exact mechanisms responsible for NO-induced relaxation in saphenous veins without valves.

Our study is the first to examine endothelium-dependent relaxation in valvular segments of venous blood vessels. The obtained results demonstrate that endothelium-dependent relaxation to A-23187 is reduced in the venous segments with valves. This observation has two important implications. First, future studies designed to characterize endothelial function in isolated veins should take into consideration differences in endothelial function between valvular and nonvalvular segments. Second, the apparent reduced ability of valvular endothelial cells to produce and release NO suggests that venous segments with valves, commonly used as coronary bypass grafts, may favor platelet aggregation, smooth muscle cells proliferation, and white blood cells adhesion, leading to graft failure (20, 24).