Development of Endothelium-Dependent Relaxation in Canine Coronary Collateral Arteries
Background—Little information exists regarding development of vasomotor control mechanisms during coronary collateral artery maturation. Therefore, we studied endothelium-dependent relaxation of canine collateral arteries isolated 2, 4, and 9 months after placement of an ameroid occluder around the proximal left circumflex coronary artery.
Results—Collateral arteries isolated after 2 months exhibited markedly reduced endothelium-dependent relaxation in response to acetylcholine (ACh; 10−10 to 10−4 mol/L) and bradykinin (BK; 10−11 to 10−6 mol/L) compared with relaxation of noncollateral coronary arteries (P<0.01). In contrast, endothelium-independent relaxation of collateral arteries to nitroprusside was only slightly reduced compared with relaxation of noncollateral arteries (P<0.05). Endothelium-dependent relaxation of collateral arteries isolated after 4 and 9 months was increased significantly, to the extent that relaxation to ACh and BK was not significantly different between collateral and noncollateral arteries at these periods. Inhibition of nitric oxide synthesis with NΤ-nitro-l-arginine methyl ester (L-NAME; 100 μmol/L) markedly inhibited ACh-induced relaxation in all noncollateral arteries and in collateral arteries isolated after 9 months. However, neither L-NAME nor indomethacin (5 μmol/L) alone inhibited ACh-mediated relaxation of collateral arteries isolated after 4 months. ACh-induced relaxation of these collateral arteries was only inhibited when arteries were preconstricted with 30 mmol/L K+ and pretreated with L-NAME and indomethacin (ie, when synthesis/effects of nitric oxide, prostaglandins, and endothelium-derived hyperpolarizing factor were inhibited).
Conclusions—Development of endothelium-dependent relaxation in canine coronary collateral arteries is not complete after 2 months. After 4 months, endothelium-dependent relaxation of collateral arteries is similar to relaxation of noncollateral arteries, but the relaxation exhibits decreased dependence on synthesis of nitric oxide and increased involvement of prostaglandins and endothelium-derived hyperpolarizing factor(s). After 9 months of development, collateral arteries exhibit normal nitric oxide–dependent relaxation, similar to noncollateral arteries.
Viability of myocardium located distal to an occlusion or severe stenosis depends on blood flow through coronary collateral arteries. Several studies1 2 3 have shown that mature collateral arteries exhibit vasomotor activity in response to a variety of agonists and thus play an active role in regulating blood flow to the dependent myocardium. Pharmacological reactivity of collateral arteries differs from that of normal coronary arteries. For example, collateral arteries are more responsive to vasopressin than normal coronary arteries.1 4 5 In contrast, collateral arteries are less responsive than normal coronary arteries to other vasoconstrictors, including endothelin, prostaglandin F2α, and the thromboxane A2 mimetic U46619.3 5
Mature collateral arteries have been reported to exhibit functional endothelium-dependent relaxation in response to acetylcholine (ACh), bradykinin (BK), and substance P3,6,7; however, the time course of development of endothelial function in maturing collateral arteries after coronary occlusion has not been studied. Direct evaluation of endothelial function of collateral arteries is difficult in vivo because of complicating influences (myocardial metabolites, circulating hormones, and sympathetic nerves) and the complex system of series and parallel resistances that contribute to regulation of blood flow to collateral-dependent myocardium. Therefore, we performed in vitro evaluations of endothelium-dependent relaxation of collateral and noncollateral coronary arteries 2, 4, and 9 months after induction of coronary occlusion with an ameroid occluder. These studies enabled us to compare endothelium-dependent relaxation responses of collateral arteries directly with responses of similar-size normal coronary arteries and collateral-dependent arteries (located distal to the occlusion). We also investigated the relative roles of nitric oxide, prostaglandins, and endothelium-derived hyperpolarizing factor(s) (EDHF) as mediators of relaxation of collateral arteries at progressive stages of collateral development.
Induction of Collateral Artery Growth
Adult mongrel male dogs (23 to 37 kg) were anesthetized with acepromazine maleate (0.8 mg/kg SC) and sodium pentobarbital (25 mg/kg IV) and ventilated mechanically. Using sterile techniques, we placed an ameroid constrictor (2.75 to 4.0 mm inner diameter, Research Instruments and Manufacturing) around the proximal left circumflex coronary artery (LCx). During surgery and recovery, dogs received buprenorphine hydrochloride (0.3 mg IV or IM) as needed for pain relief. Antibiotics were given immediately before surgery (900 000 U penicillin IM) and for 5 days after surgery (800 mg sulfamethoxazole and 160 mg trimethoprim). All experimental procedures were in accordance with the “Position of the American Heart Association on Research Animal Use,” adopted November 11, 1984, and approved by the Animal Care and Use Committee of the University of Missouri.
Preparation of Coronary Artery Rings
We studied the function of coronary arteries isolated from dogs 2, 4, and 9 months after implantation of the ameroid. Dogs were anesthetized with sodium pentobarbital (40 mg/kg), and hearts were rapidly removed and placed in aerated cold Krebs bicarbonate buffer. Collateral arteries were easily identified as tortuous, epicardial vessels extending from branches of the left anterior descending coronary artery (LAD) to branches of the LCx. Midportions of collateral arteries were excised and cleaned of fat and connective tissue. Size-matched branches from the LAD (normal) and LCx (collateral-dependent) were isolated from the same hearts. Each artery was cut into rings. Care was taken to avoid damage to the intimal surface. Using a thin section cut from the end of each ring, we measured vessel dimensions (outer diameter, inner diameter, and vessel wall thickness) with a Filar microscope eyepiece (Hitschfel Instruments, Inc).
Coronary rings were mounted on two stainless steel wires (Rocky Mountain Orthodontics) in individual tissue baths as described previously.5 Arteries were bathed in Krebs bicarbonate buffer containing (in mmol/L) NaCl 131.5, KCl 5.0, NaH2PO4 1.2, MgCl2 1.2, CaCl2 2.5, NaHCO3, and glucose 10.1 (37°C, bubbled with 95% O2/5% CO2). This solution also contained 3 μmol/L propranolol and 25 μmol/L EDTA.
Coronary arteries were allowed to stabilize for 1 hour before the experiment. Then, each coronary ring was systematically stretched to the optimum of its length-active tension relation by increments equal to 10% of the initial vessel diameter. After each stretch, a contraction was induced with 30 mmol/L K+. Arteries were studied at their optimal length, which was defined as the length at which the active tension produced was <5% greater than the tension produced at the previous length.
In Vitro Evaluation of Relaxation Responses
Relaxation responses were studied with arteries preconstricted with 30 to 40 mmol/L K+ or 30 nmol/L endothelin. Some rings were incubated continuously with enzyme inhibitors beginning at least 30 minutes before evaluation of relaxation responses. Concentration-response relationships to various agonists were determined by cumulative additions of small aliquots (20 to 150 μL; total volume <5% of bath volume) of concentrated stock solutions directly to the tissue bath. Unless otherwise indicated, drugs were purchased from Sigma Chemical Co. We purchased endothelin-1 from Peninsula Laboratories, Inc and BK from Bachem.
Relaxation responses were expressed as percent decrease of the precontraction. The concentration of vasodilator causing 50% of the maximal relaxation response was designated as the EC50 and was determined by nonlinear regression analysis of the concentration-response data for each vessel. Concentration-response curves were compared by two-way ANOVA for repeated measures followed by Fisher’s test for least significant difference. EC50 values and maximal relaxation values were analyzed by two-way ANOVA, and pairwise multiple comparisons between groups were performed with the Student-Newman-Keuls test. A P value <0.05 was considered significant. Data are presented as mean±SEM, and n values reflect the number of animals.
Dimensions of the collateral, LAD, and LCx studied 2, 4, and 9 months after ameroid placement are presented in Table 1⇓. All dimensions of LAD and LCx arterial rings were not significantly different. Outer diameters of collateral arteries isolated after all three periods were not significantly different from outer diameters of LAD and LCx. However, collateral arteries had thicker walls than noncollateral arteries, resulting from greater thickness of the intimal layer of collateral vessels. Inner diameters of all collateral arteries were smaller than inner diameters of LAD and LCx isolated at the same periods.
Development of Endothelium-Dependent Relaxation
We evaluated endothelium-dependent relaxation of collateral and noncollateral coronary arteries 2, 4, and 9 months (60±2, 120±2, and 279±2 days) after placement of an ameroid occluder around the proximal LCx. K+-induced precontractions of collateral arteries isolated after all periods were significantly smaller than contractions of noncollateral vessels, averaging 34%, 41%, and 47%, respectively, of contractions of the LAD and LCx. Relaxation responses of all arteries were expressed as a percentage of the stable K+ precontraction.
Noncollateral LAD and LCx isolated 2 months after ameroid placement relaxed 70% to 80% in response to ACh and BK (Figures 1⇓ and 2⇓). In contrast, maximal relaxation in response to ACh and BK averaged only 35% and 25%, respectively, in collateral arteries isolated after 2 months (P<0.01 versus noncollateral arteries). Relaxation of collateral arteries isolated after 4 and 9 months was increased significantly compared with relaxation of collateral arteries isolated after 2 months. Relaxation in response to ACh and BK was not significantly different between collateral and noncollateral arteries at these later periods.
ACh- and BK-induced relaxation was not significantly different between collateral-dependent LCx and normal LAD isolated 2 and 4 months after ameroid placement. However, ACh-induced relaxation of collateral-dependent LCx isolated after 9 months was reduced significantly compared with relaxation of LAD isolated from the same hearts. In contrast, relaxation in response to BK was not significantly different between LCx and LAD isolated after 9 months.
Development of Endothelium-Independent Relaxation
Maximal relaxation to the endothelium-independent vasodilator averaged >80% in collateral and noncollateral arteries isolated at all periods (Figure 3⇓). However, sodium nitroprusside (NP)-induced relaxation of collateral arteries studied at all periods was slightly but significantly attenuated compared with relaxation of noncollateral arteries.
Role of Nitric Oxide
We determined the effect of inhibition of nitric oxide synthesis with L-NAME (100 μmol/L) on endothelium-dependent relaxation of collateral arteries isolated after 4 and 9 months. In studies performed 4 months after ameroid placement, arteries were preconstricted either with 30 to 40 mmol/L K+ (Figure 4A⇓) or with a maximal concentration (30 nmol/L) of endothelin (Figure 4B⇓). As observed with K+, contractions produced by endothelin were significantly smaller in collateral than in noncollateral arteries; endothelin-induced precontractions averaged 2.5±0.7, 8.1±0.7, and 7.7±0.8 g in collateral, LAD, and LCx vessels, respectively. Neither K+- nor endothelin-induced precontractions were altered significantly by L-NAME pretreatment.
L-NAME markedly inhibited ACh-induced relaxation of all noncollateral LAD and LCx (P<0.01, Figures 4⇑ and 5⇓). L-NAME significantly increased the EC50 for ACh in noncollateral arteries (Tables 2 and 3). In contrast, pretreatment with L-NAME did not significantly alter ACh relaxation of collateral arteries isolated after 4 months and preconstricted with either high K+ (Figure 4A⇑) or endothelin (Figure 4B⇑). However, 9 months after ameroid placement, relaxation of collateral arteries in response to ACh was markedly inhibited by L-NAME (P<0.01, Figure 5⇓).
The relative role of nitric oxide in mediating relaxation of noncollateral coronary arteries to BK appeared to depend on the type of precontraction (K+ versus endothelin). L-NAME markedly attenuated BK-mediated relaxation of noncollateral arteries preconstricted with K+ (P<0.01, Figure 6A⇓). However, when noncollateral arteries were preconstricted with endothelin, L-NAME produced only a slight decrease (≤ 15%) in maximal BK-induced relaxation of LAD and LCx (P<0.05; Figure 6B⇓) and a small increase in EC50 (P<0.05, Table 3⇓).
The effects of L-NAME on relaxation to BK were similar in collateral arteries preconstricted with either K+ or endothelin. L-NAME did not significantly decrease maximal BK-induced relaxation of collateral arteries isolated after 4 months, which was in contrast to marked inhibition of relaxation of noncollateral arteries preconstricted with K+ (Figure 6A⇑). However, L-NAME significantly increased the EC50 (decreased pEC50) for BK in collateral arteries (Tables 2⇓ and 3⇓). BK-induced relaxation of collateral arteries isolated after 9 months was markedly inhibited by L-NAME, similar to the relaxation of noncollateral arteries (P<0.05, Figure 7⇓).
L-NAME did not significantly alter the relaxation of LAD or LCx to NP (data not shown). L-NAME slightly enhanced the relaxation of collateral arteries to higher concentrations of NP (>10−7 mol/L; P<0.05), but did not significantly alter the EC50 value for NP (Table 2⇑).
Role of Prostaglandins
The cyclooxygenase inhibitor indomethacin (5 μmol/L) did not significantly alter ACh-induced relaxation of either collateral or noncollateral arteries isolated after 4 months (Figure 8A⇓).
BK-mediated relaxation of collateral arteries isolated after 4 months was slightly inhibited by indomethacin (Figure 8B⇑). The EC50 for BK in collateral arteries was increased in the presence of indomethacin from 32±26 to 138±113 nmol/L (P<0.05). Relaxation of noncollateral arteries in response to BK was not altered by indomethacin.
Combined Effects of L-NAME and Indomethacin
We evaluated the effects of combined inhibition of nitric oxide synthase and cyclooxygenase on relaxation of arteries preconstricted with either endothelin or high K+ to investigate the role of EDHF as a mediator of ACh relaxation of collateral arteries isolated after 4 months. ACh-induced relaxation of LAD and LCx preconstricted with either endothelin or K+ was markedly inhibited by combined pretreatment with L-NAME and indomethacin (P<0.01, Figures 9⇓ and 10⇓). In contrast, ACh-induced relaxation of collateral arteries preconstricted with endothelin was not altered by combined inhibition of nitric oxide synthase and cyclooxygenase with these drugs (P>0.05, Figure 9⇓). However, when collateral arteries were preconstricted with K+, ACh-mediated relaxation was markedly inhibited in the presence of L-NAME and indomethacin (P<0.01, Figure 10⇓).
The present study is the first systematic evaluation of the chronologic development of endothelium-dependent relaxation in maturing coronary collateral arteries. Our results indicate that development of agonist-mediated endothelium-dependent relaxation is not complete in collateral arteries 2 months after ameroid implantation. After 4 months of development, endothelium-dependent relaxation of collateral arteries in response to ACh and BK appears to be similar to endothelium-dependent relaxation of noncollateral coronary arteries; however, this relaxation exhibits decreased dependence on synthesis of nitric oxide compared with relaxation of noncollateral arteries. In contrast, 9 months after ameroid implantation, endothelium-dependent relaxation of collateral arteries appears to be mediated primarily by synthesis of nitric oxide, similar to noncollateral arteries. These observations may have important implications relative to the critical role of collateral arteries in providing blood flow to collateral-dependent myocardium.
Time Course of Development of Endothelium-Dependent Relaxation
This is the first study to evaluate endothelial function of collateral arteries isolated as early as 2 months (60 days) after ameroid placement. Previous studies have demonstrated intact endothelium-dependent responses to ACh and BK in well-developed coronary collateral arteries.3 7 8 However, the present study has provided a new finding, that collateral arteries in early stages of development exhibit reduced relaxation in response to endothelium-dependent vasodilator agonists. Potential mechanisms responsible for the marked attenuation of agonist-mediated endothelium-dependent relaxation in immature collateral arteries could include a paucity of muscarinic and kininergic receptors on endothelial cells, deficiencies of signal transduction elements coupling endothelial receptors to synthesis of mediators of relaxation, low intrinsic activity of nitric oxide synthase, and/or dysfunctional guanylyl cyclase–cGMP cascade in smooth muscle. Impairment of guanylyl cyclase–cGMP cascade in smooth muscle of collateral arteries appears to be unlikely as a primary mechanism of impairment because relaxation in response to the nitric oxide donor NP was only mildly decreased in collateral arteries isolated after 2 months compared with noncollateral arteries.
Findings from studies of endothelial function during regrowth after vascular injury are consistent with our results. For example, agonist-mediated endothelium-dependent relaxation has been shown to be impaired in canine coronary arteries with newly developed endothelial cells 5 weeks after balloon injury.9 Thus, it is plausible that receptor/signal transduction coupling mechanisms in proliferating endothelial cells may require extended periods (>2 months) to develop in collateral arteries as well.
Mechanisms of Endothelium-Dependent Relaxation in Developing Coronary Collateral Arteries
Our results indicate that endothelium-dependent relaxation of collateral arteries isolated after 4 months is not dependent on synthesis of nitric oxide. These findings are in distinct contrast to those from noncollateral coronary arteries isolated from the same hearts. Results of our studies suggest that all three known endothelium-derived mediators (nitric oxide, prostaglandins, and EDHF) are involved in ACh-induced relaxation of collateral arteries isolated after 4 months. Inhibition of the synthesis or action of any one or two of the known mediators of endothelium-dependent relaxation did not significantly attenuate ACh relaxation of collateral arteries. ACh relaxation of collateral arteries isolated after 4 months was attenuated significantly only when the synthesis and/or action of nitric oxide, prostaglandins, and EDHF were inhibited simultaneously. These results indicate that prostaglandins and EDHF have important roles as mediators of endothelium-dependent relaxation in collateral arteries isolated after 4 months and that endothelium-dependent relaxation of collateral arteries at this period involves a redundancy or reserve of mediators of relaxation. However, endothelium-dependent relaxation of collateral arteries isolated after 9 months was primarily dependent on synthesis of nitric oxide, which indicates that the roles of prostaglandins and EDHF as redundant mediators of endothelium-dependent relaxation diminish with time during collateral development.
In contrast to our findings, Flynn and colleagues7 reported that NG-monomethyl-l-arginine (L-NMMA) produced similar inhibition of ACh- and BK-mediated relaxation of collateral and noncollateral coronary arteries. The inconsistency in our results may be explained by the differences in the experimental design or in the time allowed for collateral development. Flynn et al studied collateral arteries isolated over a broad, unspecified range of time (“a minimum of 12 weeks”), whereas we studied collateral arteries isolated at specific time intervals after ameroid implantation.
Dulas and coworkers8 studied endothelium-dependent relaxation of collateral arteries isolated 6 months after coronary occlusion. In contrast to the report of Flynn et al,7 Dulas and colleagues found that L-NMMA produced less inhibition of ACh relaxation of collateral arteries than of noncollateral coronary arteries, indicating decreased dependence on nitric oxide synthesis. These investigators concluded that the greater residual relaxation in collateral arteries induced by ACh after nitric oxide synthase and cyclooxygenase blockade could result from enhanced EDHF production in collateral vessels. The results of Dulas et al correlate well with our findings. Taken in concert, these results indicate that endothelium-dependent relaxation of maturing collateral arteries progressively develops increased dependence on nitric oxide synthesis between 4 and 9 months after coronary occlusion.
Implications and Conclusions
Our findings have potentially important implications relative to understanding collateral artery function and regulation of perfusion of collateral-dependent myocardium. Reduced endothelium-dependent vasodilation in collateral arteries early in development suggests the possibility of an increased propensity for vasospasm of collateral arteries and myocardial ischemia soon after occlusion of a major coronary artery. Endothelium-derived relaxing factors have been implicated as important mediators of flow-mediated vasodilation of coronary vessels10 11 and coronary metabolic vasodilation.12 Endothelium-derived relaxing factors (nitric oxide, prostacyclin) also exert protective antithrombotic effects.13 14 Thus, the decreased capability to produce endothelium-derived relaxing factors in developing collateral arteries may result in reduced vasodilation and increased thrombogenesis, potentially limiting blood flow to collateral-dependent myocardium.
In contrast to risks posed by delayed development of endothelium-dependent relaxation in collateral arteries early after coronary occlusion, the mechanisms underlying relaxation of collateral arteries studied after 4 months may enhance the ability of the collateral circulation to provide adequate perfusion of the collateral-dependent myocardium. In contrast to the marked dependence of noncollateral canine coronary arteries on nitric oxide synthesis, these collateral arteries exhibit an increased capacity for production of multiple endothelium-derived relaxing factors in response to receptor-dependent agonists. This redundancy or reserve of endothelium-derived mediators of relaxation may promote and maintain adequate perfusion of the collateral-dependent myocardium under diverse conditions and thereby prevent myocardial ischemia and infarction.
The authors greatly appreciate the technical contributions made by M. L. Mattox and M. A. Tanner. These studies were supported by funds from the American Heart Association and National Institutes of Health grant HL-47812, training grant HL-07094, and program project P01 HL-52490. Dr Rapps was supported by a predoctoral fellowship from the American Heart Association, Missouri Affiliate.
- Received March 23, 1998.
- Revision received May 12, 1998.
- Accepted May 20, 1998.
- Copyright © 1998 by American Heart Association
Harrison DG, Chilian WM, Marcus ML. Absence of functioning α-adrenergic receptors in mature canine coronary collaterals. Circ Res. 1986;59:133–142.
Hautamaa PV, Dai X, Homans DC, Bache RJ. Vasomotor activity of moderately well-developed canine coronary collateral circulation. Am J Physiol. 1989;256:H890–H897.
Angus JA, Ward JE, Smolich JJ, McPherson GA. Reactivity of canine isolated epicardial collateral coronary arteries. Circ Res. 1991;69:1340–1352.
Peters KG, Marcus ML, Harrison DG. Vasopressin and the mature coronary collateral circulation. Circulation. 1989;79:1324–1331.
Rapps JA, Jones AW, Sturek M, Magliola L, Parker JL. Mechanisms of altered contractile responses to vasopressin and endothelin in canine coronary collateral arteries. Circulation. 1997;95:231–239.
Altman J, Dulas D, Pavek T, Laxson DD, Homans DC, Bache RJ. Endothelial function in well-developed canine coronary collateral vessels. Am J Physiol. 1993;264:H567–H572.
Flynn NM, Kenny D, Pelc LR, Warltier DC, Bosnjak J, Kampine JP. Endothelium-dependent vasodilation of canine coronary collateral vessels. Am J Physiol. 1991;30:H1797–H1801.
Dulas D, Altman J, Hirata-Dulas C, Bache RJ. Endothelium-dependent vasodilation in well-developed coronary collateral vessels. J Cardiovasc Pharmacol. 1997;28:488–493.
Cox RH, Haas KS, Moisey DM, Tulenko TN. Effects of endothelium regeneration on canine coronary artery function. Am J Physiol. 1989;257:H1681–H1692.
Rubanyi GM, Romero JC, Vanhoutte PM. Flow-induced release of endothelium-derived relaxing factor. Am J Physiol. 1986;250:H1145– H1149.
Cox DA, Vita JA, Treasure CB, Fish RD, Alexander RW, Ganz OP, Selwyn AP. Atherosclerosis impairs flow-mediated dilation of coronary arteries in humans. Circulation. 1989;80:458–465.
Quyyumi AA, Dakak N, Andrews NP, Gilligan DM, Panza JA, Cannon RO. Contribution of nitric oxide to metabolic coronary vasodilation in the human heart. Circulation. 1995;92:320–326.