Apamin-sensitive, non-nitric oxide (NO) endothelium-dependent relaxations to bradykinin in the bovine isolated coronary artery: no role for cytochrome P₄₅₀ and K⁺

Abstract

1. Since cytochrome P₄₅₀-derived metabolites of arachidonic acid and K⁺ have been implicated in endothelium-derived hyperpolarizing factor (EDHF)-dependent responses, the aim of this study was to determine whether such factors contribute to non-nitric oxide (NO), endothelium-dependent relaxation to bradykinin (BK) in bovine isolated coronary artery.

2. In rings of artery contracted with U46619 and treated with indomethacin (3 μM) and N^G-nitro-L-arginine (L-NOARG; 100 μM), relaxation to BK (0.01 nM-0.3 μM) was blocked by ∼60% after inhibition of K⁺ channels with either high extracellular K⁺ (high [K⁺]_o; 15–67 mM) or apamin (0.3 μM).

3. Ouabain (1 μM), an inhibitor of Na⁺/K⁺-ATPase, decreased the sensitivity to BK without affecting the maximum response. In L-NOARG-treated rings, ouabain had no further effect on the relaxation to BK. An inhibitor of inward-rectifying K⁺ channels, Ba²⁺ (30 μM), had no effect on relaxations to BK in the absence or presence of either L-NOARG or ouabain.

4. KCl (2.5–10 mM) elicited small relaxations (∼20%) that were abolished by nifedipine (0.3 μM) and ouabain.

5. Both the high [K⁺]_o/apamin-sensitive relaxation to BK, and the relaxation to the K_ATP channel-opener, levcromakalim (0.6 μM), were unaffected by the cytochrome P₄₅₀ inhibitor, 7-ethoxyresorufin (10 μM), or by co-treatment with a phospholipase A₂ inhibitor, arachidonyl trifluoromethyl ketone (AACOCF₃; 3 μM) and a diacylglycerol (DAG)-lipase inhibitor, 1,6-bis-(cyclohexyloximinocarbonylamino)-hexane (RHC 80267; 30 μM). The non-NO/high [K⁺]_o-insensitive, ∼40% relaxation to BK was, however, abolished by these treatments.

6. Therefore, neither cytochrome P₄₅₀-derived metabolites of arachidonic acid nor K⁺ appear to mediate the EDHF-like relaxation to BK (i.e the non-NO, high [K⁺]_o/apamin-sensitive component) in bovine coronary arteries. Cytochrome P₄₅₀-derived metabolites may be released at higher BK concentrations to act in parallel with NO and the high [K⁺]_o/apamin-sensitive mechanism.

Introduction

Non-prostanoid, endothelium-dependent relaxation in many blood vessels is mediated not only by nitric oxide (NO) but also by a non-NO, high K⁺-sensitive mechanism which may be attributable to the release of an endothelium-derived hyperpolarizing factor (EDHF; Taylor & Weston, 1988; Kilpatrick & Cocks, 1994; Garland et al., 1995; Drummond & Cocks, 1996; Feletou & Vanhoutte, 1999). The mechanism of EDHF-mediated relaxation of vascular smooth muscle remains controversial. Two recent proposals are that EDHF is either K⁺ (Edwards et al., 1998) or a cytochrome P₄₅₀-derived metabolite of arachidonic acid (Komori & Vanhoutte, 1990; Mombouli & Vanhoutte, 1997; Feletou & Vanhoutte, 1999; Fisslthaler et al., 1999).

In support of their claim that EDHF is K⁺, Edwards et al. (1998) showed that non-prostanoid, non-NO, endothelium-dependent hyperpolarization to acetylcholine (ACh) in rat resistance arteries was blocked by a combination of the small (SK_Ca) and large (BK_Ca) conductance Ca²⁺-activated K⁺ channel inhibitors, apamin and charybdotoxin, respectively. The response was also blocked by a combination of ouabain and Ba²⁺, inhibitors of Na⁺/K⁺-ATPase and inward-rectifying K⁺ channels (K_IR), respectively. Edwards et al. (1998) found that low concentrations (<20 mM) of K⁺ applied exogenously caused ouabain- and Ba²⁺-sensitive smooth muscle hyperpolarization and relaxation, mimicking the non-NO, endothelium-dependent responses to ACh. Furthermore, using a K⁺ measuring electrode strategically positioned between the endothelium and smooth muscle, they were able to show that in response to ACh, the K⁺ concentration increased by approximately 5 mM in the myoendothelial space in an apamin- and charybdotoxin-sensitive manner. Thus, the authors concluded that EDHF is likely to be universally K⁺, which after being released from endothelial cells via SK_Ca and BK_Ca channels activates smooth muscle Na⁺/K⁺-ATPase and K_IR to initiate hyperpolarization and relaxation (Edwards et al., 1998).

The conclusion by Edwards et al. (1998) that EDHF is K⁺ may not hold true for all arteries. For example, it is well established that in large epicardial arteries of the pig, an EDHF-like mechanism compensates for up to 80% of the relaxation to BK after inhibition of NO synthesis (Kilpatrick & Cocks, 1994). This is despite the fact that these large arteries express few functional K_IR (Quayle et al., 1996). Also, Fisslthaler et al. (1999) have shown that while combined application of Ba²⁺ and ouabain had no effect on BK-induced, EDHF-mediated vascular responses in porcine large epicardial arteries, transfection with antisense oligonucleotides against cytochrome P₄₅₀ 2C inhibited the response. Thus, these findings not only argue against a role for K_IR and Na⁺/K⁺-ATPase but also suggest that in porcine coronary arteries at least, EDHF represents a product of cytochrome P₄₅₀ 2C metabolism.

Earlier observations also support a role for cytochrome P₄₅₀ in EDHF-dependent vasorelaxation. For instance, a variety of inhibitors of arachidonic acid formation and cytochrome P₄₅₀ have been shown to partially inhibit endothelium-dependent relaxations in many isolated arteries (Singer et al., 1984; Pinto et al., 1987; Hecker et al., 1994; Bauersachs et al., 1997; Satake et al., 1997) and perfused vascular beds (Fulton et al., 1992; 1996; Bauersachs et al., 1994; Adeagbo & Henzel, 1998). Also, products of cytochrome P₄₅₀-dependent metabolism of arachidonic acid including epoxyeicosatrienoic acids (EETs) and hydroxyeicosatetraenoic acids (HETEs), have been shown to be released by cultured bovine coronary artery endothelial cells (Rosolowsky et al., 1996; Rosolowsky & Campbell, 1996; Gebremedhin et al., 1998) and to relax vascular smooth muscle (Proctor et al., 1987; Rosolowsky et al., 1990, 1996; Gebremedhin et al., 1992; Hecker et al., 1994; Campbell et al., 1996; Zou et al., 1996; Pratt et al., 1998). Also, EETs can activate smooth muscle calcium-dependent K⁺ channels to initiate hyperpolarization (Gebremedhin et al., 1992; Hu & Kim, 1993; Campbell et al., 1996; Zou et al., 1996).

In the bovine coronary artery, the non-prostanoid relaxation to BK, which is sensitive to high extracellular K⁺ and therefore likely attributable to an EDHF-like mechanism, compensates for up to 95% of the overall response after inhibition of NO (Drummond & Cocks, 1996). Also, a third mechanism independent of both NO and the EDHF-like mechanism is present in this tissue and acts as a second level backup behind NO and EDHF, accounting for ∼40% of the overall response to BK. Therefore, the aim of this study was to determine whether endothelium-derived K⁺ or cytochrome P₄₅₀-derived metabolites of arachidonic acid mediate non-NO, endothelium-dependent relaxation to BK in the bovine coronary artery. Our data show that while cytochrome P₄₅₀-derived metabolites of arachidonic acid may be involved, they do not mediate the EDHF-like response. Also, our results indicate that endothelium-derived K⁺ does not appear to have any role in non-prostanoid, non-NO endothelium-dependent relaxation in the bovine coronary artery.

Methods

Tissue preparation

Sections of bovine myocardium containing the left anterior descending coronary artery were obtained from a local abattoir and transported to the laboratory in ice cold Krebs solution (composition in mM; Na⁺ 143.1, K⁺ 5.9, Ca²⁺ 2.5, Mg²⁺ 1.2, Cl⁻ 127.8, HCO₃⁻ 25.0, SO₄²⁻ 1.2, H₂PO₄⁻ 1.2 and glucose 11.0). Ring segments (3 mm) of coronary artery were prepared as previously described (Drummond & Cocks, 1996) and then suspended between two stainless steel wire hooks, one of which was connected to a force-displacement transducer (model FT03C, Grass, Quincy, MA, U.S.A.) and the other to a micrometer-adjustable support leg. Preparations were immersed in water-jacketed 30 ml organ baths containing warm (37°C), carbogen-bubbled (95% O₂, 5% CO₂) Krebs solution (pH 7.4). Changes in isometric force were amplified and displayed on dual-channel, flat-bed recorders (W & W Scientific Instruments, Basel, Switzerland).

Tissue equilibration

Rings of artery were allowed to equilibrate under zero tension for a period of 25 min, after which time they were passively stretched to a resting tension of 5 g. After a further 25 min, rings were stretched again to 5 g passive tension, allowed to equilibrate and then maximally contracted with an isotonic, high potassium physiological salt solution (KPSS; composition in mM; K⁺ 124.9, Na⁺ 25.0, Ca²⁺ 2.5, Mg²⁺ 1.2, Cl⁻ 128.7, HCO₃⁻ 25.0, SO₄²⁻ 1.2, H₂PO₄⁻ 1.2 and glucose 6.1). Once a stable plateau of active force to KPSS was obtained (KPSS_max), tissues were washed with normal Krebs solution and allowed to return to resting levels of passive force. All tissues were then treated with the cyclo-oxygenase inhibitor, indomethacin (3 μM), for the remainder of the experiment. Some tissues were also treated with the L-type voltage-operated Ca²⁺ channel (VOCC) inhibitor, nifedipine (0.3 μM), to prevent excessive levels of pre-contraction in experiments where an isotonic 67 mM KCl bathing solution (high [K⁺]_o) was used to inhibit K⁺ channels. Previously, we have shown that while nifedipine abolished contractions to a maximum depolarizing concentration of extracellular K⁺, it had no effect on either the ability of U46619 to cause contraction or any component of the relaxation to BK (Drummond & Cocks, 1996).

Experimental protocol

Twenty-five minutes after the addition of indomethacin and nifedipine, rings were either left untreated (control) or treated with either the NO synthase inhibitor, N^G-nitro-L-arginine (L-NOARG; 100 μM), high [K⁺]_o, the soluble guanylatecyclase inhibitor, 1H-[1,2,4]oxadiazolo[4,3-α]quinoxaline-1-one (ODQ; 10 μM), the SK_Ca inhibitor, apamin (0.3 μM), the Na⁺/K⁺-ATPase inhibitor, ouabain (1 μM or 1 mM) or the K_IR inhibitor, Ba²⁺ (30 μM). Some L-NOARG-treated tissues were further treated with either high [K⁺]_o, ODQ, apamin, ouabain or Ba²⁺ or ouabain plus Ba²⁺. Other L-NOARG-treated rings were treated with the combination of apamin (0.3 μM), charybdotoxin (0.1 μM) and the K_ATP channel inhibitor, glibenclamide (10 μM). Also, separate rings were treated with either inhibitors of arachidonic acid formation including the phospholipase A₂ inhibitors, quinacrine (30 μM; Bauersachs et al., 1994) and arachidonyl trifluoromethyl ketone (AACOCF₃; 3 μM; Street et al., 1993; Riendeau et al., 1994) and the diacylglycerol (DAG) lipase inhibitor, 1,6-bis-(cyclohexyloximinocarbonylamino)-hexane (RHC 80267; 30 μM; Sutherland & Amin, 1982), or blockers of subsequent metabolism via the cytochrome P₄₅₀ pathway such as N,N-diethylaminoethyl-2,2-diphenylvalerate (SKF-525a; 100 μM; Hecker et al., 1994), clotrimazole (100 μM; Hecker et al., 1994) or 7-ethoxyresorufin (10 μM; Tassaneeyakul et al., 1993). All tissues were then contracted with titrated concentrations of the thromboxane A₂-mimetic, U46619, to ∼40% of their KPSS_max. Upon reaching a stable plateau, rings were exposed to cumulatively increasing half-log molar concentrations of either BK or NO-donors, S-nitroso-N-acetylpenicillamine (SNAP) and sodium nitroprusside (SNP), or with either single concentrations of levcromakalim (0.6 μM) or KCl (2.5–10 mM).

Statistics

All cumulative concentration-relaxation curves were normalized as percentages of relaxation from the initial U46619-induced contraction level. Each normalized curve was then computer-fitted (Graphpad Prism, version 1.00) with a sigmoidal regression curve of the following equation,

where X is the logarithm of the agonist concentration and Y is the response. BOTTOM is the lower response plateau, TOP is the upper response plateau and pD₂ is the X value when the response is halfway between BOTTOM and TOP. The variable HILLSLOPE controls the slope of the curve. Mean sensitivity (pEC₅₀ values), maximum relaxations (R_max) and their standard errors (s.e.) were then calculated for each response curve. U46619-induced levels of contraction were expressed as percentages of the respective KPSS_max values for each tissue. Values of n represent number of rings of artery, each from different animals. Differences in mean pEC₅₀ and R_max values were tested for significance by means of one way analysis of variance (ANOVA) with multiple comparisons via Dunnett's test (when treatment groups were being compared with controls only) or Tukey-Kramer's test (when treatment groups were being compared amongst themselves as well as with controls). All differences were accepted as significant at the P<0.05 level.

Drugs and their sources

Ba²⁺ chloride, bradykinin triacetate, clotrimazole, indomethacin, N^G-nitro-L-arginine (L-NOARG), quinacrine (Sigma, MO, U.S.A.); apamin, arachidonyl trifluoromethyl ketone (AACOCF₃), 7-ethoxyresorufin, nifedipine, 1H-[1,2,4]oxadizolo[4,3-α]quinoxaline-1-one (ODQ), N,N-diethylaminoethyl-2,2-diphenylvalerate hydrochloride (SKF-525a), 1,5,5-hydroxy-11,9-(epoxymethano) prosta-5Z,13E-dienoic acid (U46619), S-nitroso-N-acetylpenicillamine (SNAP), 1,6-bis-(cyclohexyloximinocarbonylamino)-hexane (RHC 80267; Sapphire Bioscience, N.S.W., Australia); ouabain (Calbiochem); sodium nitroprusside (SNP; David Bull Laboratories, Australia) and levcromakalim (kind gift from Dr Grant McPherson).

Stock solutions of nifedipine (10 mM), ouabain (100 mM) and U46619 (1 mM) were made up in absolute ethanol, while those of indomethacin (100 mM) and L-NOARG (100 mM) were made up in Na₂CO₃ (1 M) and NaHCO₃ (1 M), respectively. Stock solutions of AACOCF₃ (1 mM), clotrimazole (100 mM), 7-ethoxyresorufin (10 mM), ODQ (10 mM), RHC 80267 (10 mM) and levcromakalim (10 mM) were made up in dimethyl sulphoxide. All subsequent dilutions of these drugs were in distilled water. All other drug stocks were made up in distilled water. Note that none of the vehicles used in this study had an effect on any component of the response to BK or on the responses to SNAP, SNP, levcromakalim and KCl.

Results

Confirmation that an EDHF-like mechanism acts as a backup relaxation mechanism for NO

BK caused concentration-dependent (pEC₅₀, 9.74±0.16) and maximum (R_max, 103.6±0.9%) relaxation in endothelium-intact rings of bovine isolated coronary artery contracted to ∼40% KPSS_max with U46619 (Figure 1). L-NOARG (100 μM) caused a small, significant reduction in R_max to BK (95.8±1.6%) and although there was a trend for L-NOARG to cause a reduction in sensitivity (pEC₅₀, 9.37±0.13), it did not reach significance (Figure 1). Also, a near-identical pattern of inhibition of responsiveness to BK was observed with ODQ (10 μM; pEC₅₀, 9.25±0.23; R_max, 96.2±2.1%; Figure 1). The combination of L-NOARG and ODQ was no more effective at inhibiting the response to BK than was either compound alone (pEC₅₀, 9.29±0.12; R_max, 97.8±2.3%; Figure 1). ODQ caused an ∼100-fold increase in the concentration of the NO donor, SNAP, required to elicit a threshold response (Figure 1).

Figure 1.

Effect of (a) the NO synthase inhibitor, L-NOARG (100 μM) and guanylate cyclase inhibitor, ODQ (10 μM), on endothelium-dependent relaxations to BK in rings of bovine coronary artery contracted with U46619. (b) Effect of ODQ (10 μM) on endothelium-independent relaxations to the NO-donor, SNAP in similarly contracted rings of bovine coronary artery. Values (mean±s.e.mean from n=6 experiments) are expressed as a percentage reversal of the initial U46619-induced contraction. (+) indicates that R_max values for all treatments significantly different from those obtained in control tissues (P<0.05 for Tukey-Kramer's t-statistic after one way ANOVA).

In the presence of L-NOARG, the response to BK was markedly inhibited by high [K⁺]_o (67 mM; pEC₅₀, 8.07±0.23; R_max, 38.6±8.6%; Figure 2). By contrast, high [K⁺]_o had no effect on the response to BK in the absence of L-NOARG (Figure 2). A similar degree of inhibition of L-NOARG-resistant relaxation to BK was observed with a lower concentration of [K⁺]_o (15 mM) (n=5; data not shown). Like high [K⁺]_o, apamin (0.3 μM) inhibited the L-NOARG-resistant response to BK (pEC₅₀, 8.35±0.13; R_max, 51.5±10.0%) and had no effect on the control response (Figure 2). A combination of apamin (0.3 μM), glibenclamide (10 μM) and charybdotoxin (0.1 μM) was no more effective at inhibiting the L-NOARG-resistant response to BK than was apamin on its own (n=4; data not shown).

Figure 2.

Effect of (a) 67 mM extracellular K⁺ (high [K⁺]_o), (b) apamin (0.3 μM) and (c) ouabain (1 μM) on endothelium-dependent relaxations to BK in the absence and presence of L-NOARG (100 μM) in rings of bovine coronary artery contracted with U46619. Values (mean±s.e.mean from n=5–6 experiments) are expressed as a percentage reversal of the initial U46619-induced contraction. (^*P<0.05 and ^**P<0.001) and (⁺P<0.001) indicate that pEC₅₀ and R_max values, respectively were significantly different from those obtained in control tissues (for Tukey-Kramer's t-statistic after one way ANOVA).

Effect of ouabain and Ba²⁺

The pEC₅₀ (8.4±0.15; n=6) and R_max (92.3±1.7%) to BK in tissues treated with a combination of ouabain (1 mM) and Ba²⁺ (30 μM) were significantly (P<0.05) reduced compared to controls. The relaxation to BK was further reduced by ouabain and Ba²⁺ (30 μM) in L-NOARG-treated tissues (pEC₅₀, 7.9±0.15; R_max, 62.1±5.5%; n=6). Ouabain alone caused the same degree of inhibition of BK-induced relaxations in the absence and presence of L-NOARG as that observed with the combination of ouabain and Ba²⁺ (n=6; data not shown). Ba²⁺ (30 μM) had no effect on control or L-NOARG-resistant responses to BK (n=6; data not shown). Ouabain (1 μM) caused a significant inhibition of the pEC₅₀ (8.58±0.22) but unlike the higher concentration, had no effect on R_max (88.9±8.3%) to BK (Figure 2). The subsequent addition of L-NOARG to tissues treated with 1 μM ouabain, caused no further inhibition of relaxations to BK (pEC₅₀, 8.30±0.21; R_max, 88.46±2.5%; Figure 2).

Combined treatment with ouabain (1 mM) and Ba²⁺ (30 μM) had no effect on relaxation to the NO donor, SNP (control: pEC₅₀, 7.65±0.2; R_max, 91.5±4.6; ouabain and Ba²⁺: pEC₅₀, 7.28±0.12; R_max, 76.3±9.3%; n=6).

Response to KCl

Increasing the extracellular concentration of K⁺ from 5.3 to 10.3 mM with KCl caused relaxation of 18.0±3.7% (n=5) followed by contraction (Figure 3). No further relaxation to K⁺ was observed when the concentration was raised to 15.3 mM (data not shown). Both relaxation (R_max, 4.8±1.2%) and contraction to K⁺ were inhibited by nifedipine (0.3 μM; Figure 3). Also, the relaxation to K⁺ was abolished by ouabain (1 μM; Figure 3).

Figure 3.

Digitized recordings showing responses to exogenous administration of 2.5–10 mM KCl in rings of bovine coronary artery contracted with U46619 in the (a) absence or presence of either (b) nifedipine (0.3 μM) or (c) ouabain (1 μM). Also shown are maximum relaxations to bradykinin (BK; 100 nM).

Effect of inhibitors of arachidonic acid formation and cytochrome P₄₅₀ activity on NO-independent relaxations to BK and responses to levcromakalim

Quinacrine (30 μM), SKF 525a (100 μM) and clotrimazole (100 μM), each significantly reduced both the R_max (63.0±8.6%, 45.8±7.0%, 42.1±15.2%, respectively) and sensitivity (pEC₅₀, 8.21±0.18, 8.09±0.14, 8.64±0.17, respectively) of the L-NOARG-resistant response to BK (pEC₅₀, 8.98±0.08; R_max, 94.7±2.1%; Figure 4). By contrast, AACOCF₃ (3 μM) and RHC 80267 (30 μM), either alone (n=6; data not shown) or in combination (pEC₅₀, 8.77±0.23; R_max, 88.9±2.2%) and 7-ethoxyresorufin (10 μM; pEC₅₀, 9.15 ±0.17; R_max, 81.8±4.2%) had no significant effect on the L-NOARG-resistant response to BK (Figure 4).

Figure 4.

Effect of inhibitors of free arachidonic acid formation [quinacrine (QUIN; 30 μM); AACOCF₃ (3 μM); RHC 80267 (30 μM)] and cytochrome P₄₅₀ activity [7-ethoxyresorufin (7-ER; 10 μM); clotrimazole (CLOT; 100 μM); SKF 525a (100 μM)] on L-NOARG-resistant responses to BK (a and b) and on relaxations to levcromakalim (0.6 μM; c) in rings of bovine coronary artery contracted with U46619. All values (mean±s.e.mean from n=4–6 experiments) are expressed as a percentage reversal of the initial U46619-induced contraction. ^* (P<0.01) and (+) (P<0.01) indicate R_max and pEC₅₀ values significantly different from control (Dunnett's modified t-statistic after a single one-way ANOVA).

The ATP-sensitive K⁺ channel opener, levcromakalim (0.6 μM), caused a near-maximum relaxation (R_max, 95.3±2.5%) which was unaffected by either 7-ethoxyresorufin (10 μM) or the combination of RHC 80267 (30 μM) and AACOCF₃ (3 μM; Figure 4). By contrast, quinacrine (30 μM), clotrimazole (100 μM) and SKF-525a (100 μM) significantly reduced the R_max to levcromakalim to 13.8±11.5% 45.0±13.7% and 21.8±5.1%, respectively (Figure 4).

Effect of inhibitors of arachidonic acid formation and cytochrome P₄₅₀ activity on non-NO, high K⁺-resistant relaxations to BK

While both AACOCF₃ (n=5) and RHC 80267 (n=6) alone had no effect on non-NO, K⁺-resistant relaxation to BK, in combination they virtually abolished this response (R_max, 6.1±1.7%; Figure 5). Similarly, 7-ethoxyresorufin (R_max, 5.4±4.0%) abolished the L-NOARG/high [K⁺]_o-resistant relaxations to BK (Figure 5), despite the fact that the latter of these compounds had no effect on the L-NOARG-resistant responses to BK.

Figure 5.

Effect of 7-ethoxyresorufin (7-ER; 10 μM) and a combination of AACOCF₃ (3 μM) and RHC 80267 (30 μM) on L-NOARG/high [K⁺]_o-resistant responses to BK in rings of bovine coronary artery contracted with U46619. All responses were obtained in the presence of L-NOARG (100 μM) and 67 mM [K⁺]_o. Values (mean±s.e.mean from n=6 experiments) are expressed as a percentage reversal of the initial U46619-induced contraction.

Effect of the combination of AACOCF₃ and RHC 80267 and 7-ethoxyresorufin on control and high [K⁺]_o-resistant relaxation to BK

In the absence of any treatments, the combination of AACOCF₃ (3 μM) and RHC 80267 (30 μM) had no effect on the control response to BK (pEC₅₀, 9.3±0.2; R_max, 99.9±2.0%; n=6). Under the same conditions, however, 7-ethoxyresorufin (10 μM) caused a significant (P<0.01) reduction in R_max to BK (82.8±3.9%; n=5) without affecting the pEC₅₀.

Similar to its effects on the control response, the combination of AACOCF₃ (3 μM) and RHC 80267 (30 μM) also had no effect on the high [K⁺]_o-resistant response to BK (pEC₅₀, 9.2±0.09; R_max, 95.7±2.3%; n=6). By contrast, 7-ethoxyresorufin (10 μM) abolished this response (R_max, 4.8±3.7%; n=5; P<0.01). Concentration-dependent relaxations to SNAP (pEC₅₀, 7.80±0.11; R_max, 98.4±0.7%; n=12) were unaffected by the combination of AACOCF₃ and RHC 80267 (n=5) but were markedly inhibited by 7-ethoxyresorufin (pEC₅₀, 5.65±0.64; R_max, 20.3±6.6; n=6, P<0.01 for both values).

Discussion

The results of this study suggest that a cytochrome P₄₅₀-derived metabolite of arachidonic acid is involved in endothelium-dependent relaxation in the bovine coronary artery, but that it does not mediate the component of the response attributable to an EDHF-like mechanism. The cytochrome P₄₅₀-dependent mechanism appeared to be activated only by high concentrations of BK, to function as a backup with the EDHF-like mechanism for NO, the predominant non-prostanoid, endothelium-derived relaxing factor (EDRF) in this tissue. Also, our findings do not support a role for K⁺ as an EDRF in the bovine coronary artery.

Recently, Fisslthaler et al. (1999) reported that cytochrome P450 2C is an EDHF synthase in the porcine coronary artery. Our findings that both the PLA₂ inhibitor, quinacrine, and the cytochrome P₄₅₀ inhibitors, SKF-525a and clotrimazole, all attenuated the L-NOARG-resistant response to BK in the bovine coronary artery appear to support the notion that a cytochrome P₄₅₀-derived metabolite of arachidonic acid is involved in a smooth muscle relaxation mechanism attributable to EDHF (Komori & Vanhoutte, 1990; Fulton et al., 1992; Hecker et al., 1994; Bauersachs et al., 1994; Mombouli & Vanhoutte, 1997; Feletou & Vanhoutte, 1999). However, the inhibitory effects of these compounds most likely resulted from non-specific actions on smooth muscle K⁺ conductance since each compound also blocked relaxation to the K⁺ channel opener, levcromakalim. Similar inhibition of relaxations to the K_ATP channel openers, levcromakalim and pinacidil, has been reported in rat hepatic (Zygmunt et al., 1996) and mesenteric (Fukao et al., 1997) arteries by quinacrine and SKF-525a. By contrast, neither AACOCF₃, an inhibitor of PLA₂ which is structurally and mechanistically distinct from quinacrine (Gelb et al., 1994), nor 7-ethoxyresorufin, a xenobiotic cytochrome P₄₅₀ inhibitor (Tassaneeyakul et al., 1993) which inhibits endothelium-dependent relaxations to arachidonic acid in the perfused rat kidney (Oyeken et al., 1991), had any effect on relaxations to either levcromakalim or BK after NO inhibition. Also, combined PLA₂ and DAG lipase inhibition with AACOCF₃ and RHC 80267, respectively, failed to have any effect on the response to BK after NO inhibition. Therefore, relaxations to the EDHF-like mediator do not appear to involve cytochrome P₄₅₀-derived metabolites of arachidonic acid, at least in bovine coronary arteries.

A cytochrome P₄₅₀-derived metabolite of arachidonic acid may, however, have mediated the component of the response to BK not attributable to either NO or EDHF, since inhibition of the two main sources of arachidonic acid with AACOCF₃ and RHC 80267, as well as inhibition of cytochrome P₄₅₀ with 7-ethoxyresorufin abolished this L-NOARG/high [K⁺]_o-resistant response. 7-ethoxyresorufin is a relatively specific competitive substrate inhibitor of 1A isoforms of cytochrome P₄₅₀ (Tassaneeyakul et al., 1993). Also, cytochrome P₄₅₀1A isozymes have been immunohistochemically localized in both mammalian and non-mammalian endothelial cells. Thus, the block by 7-ethoxyresorufin of L-NOARG/high [K⁺]_o-resistant relaxations to BK may implicate a role for members of the cytochrome P₄₅₀1A family in this low efficacy component of the response.

While 7-ethoxyresorufin inhibited relaxations to BK in the presence of L-NOARG and high K⁺, it also blocked the NO-dependent component of the response. This effect most likely occurred downstream of NO synthase since relaxations to the NO-donor, SNAP, were also markedly attenuated by 7-ethoxyresorufin. These findings are consistent with previous studies which have shown 7-ethoxyresorufin inhibits endothelium-dependent relaxation known to be due solely to NO (Rees et al., 1990) as well as relaxation to exogenous NO in the rat isolated aorta (Bennett et al., 1992; Oyeken et al., 1994; Li & Rand, 1996) and guinea-pig taenia coli (Selemidis et al., 1997). One possible mechanism underlying these inhibitory actions of 7-ethoxyresorufin involves the generation of superoxide anions which are known to inactivate NO (Ignarro et al., 1988). Irrespective of the mechanism involved, however, the findings here that 7-ethoxyresorufin abolished all the response to BK in the presence of high K⁺ indicates that it may prove to be a useful compound to study responses due to EDHF-like mechanisms in isolation from other non-prostanoid relaxing factors.

Although the nature of the cytochrome P₄₅₀-derived metabolite of arachidonic acid which appeared to be involved in the response to BK is unknown, two possible classes of compounds are EETs and HETEs. Such compounds are not only released from bovine endothelial cells in culture, but also have been shown to relax isolated segments of bovine coronary arteries (Rosolowsky et al., 1996; Rosolowsky & Campbell, 1996; Gebremedhin et al., 1998; Pratt et al., 1998). If these metabolites did play a role in mediating the non-NO/[K⁺]-resistant component, they need not have acted in a paracrine manner. For example, inhibitors of cytochrome P₄₅₀ are known to prevent the influx of Ca²⁺ that follows agonist-induced depletion of intracellular Ca²⁺ stores in many cell types including bovine coronary artery endothelial cells (Alvarez et al., 1992; Graier et al., 1995). Furthermore, the exogenous application of 5,6-EET to bovine coronary and human umbilical artery endothelial cells was shown to cause a sustained increase in intracellular Ca²⁺ which was similar in time course and magnitude to the second phase increase in Ca²⁺ observed in response to BK (Graier et al., 1995). Taken together, these observations suggest that EETs can act as intracellular second messengers in endothelial cells, signalling transmembrane Ca²⁺ influx in response to agonist-induced depletion of internal Ca²⁺ stores (Baron et al., 1997). In the bovine coronary artery, non-NO/high [K⁺]_o-resistant relaxations occurred over a higher concentration range of BK than that which stimulated NO and the EDHF-like mechanism. This may indicate that activation of the remaining non-prostanoid mechanism requires higher levels of intracellular endothelial cell Ca²⁺. Thus, cytochrome P₄₅₀-derived metabolites of arachidonic acid like EETs and HETEs may act intracellularly to raise endothelial cell Ca²⁺ levels high enough to activate an as yet unidentified smooth muscle relaxing mechanism.

The present study extends our earlier findings in the same tissue (Drummond & Cocks, 1996) that NO is the predominant mediator of endothelium-dependent relaxation with the EDHF-like and the low-efficacy, cytochrome-P₄₅₀-dependent mechanisms, both acting as backup vasodilator systems for NO. Thus, the NO synthase inhibitor, L-NOARG, and the guanylate cyclase inhibitor, ODQ (Garthwaite et al., 1995), when used alone or in combination with each other, inhibited relaxations to BK to the same small but significant degree. Also, the NO-independent relaxation to BK was blocked by ∼60% by either high [K⁺]_o (Drummond & Cocks, 1996) or the SK_Ca channel inhibitor, apamin, neither of which had any effect in the absence of L-NOARG. Furthermore, the observation that combined treatment with AACOCF₃ and RHC 80267, as well as 7-ethoxyresorufin, all had no effect in L-NOARG-treated tissues when the EDHF-like mechanism was active (i.e. in the absence of high [K⁺]_o) indicates that the cytochrome P₄₅₀-dependent mechanism also acts as backup for the EDHF-like mechanism.

This study also shows that EDHF-like relaxation to BK in the bovine coronary artery is unlikely to be mediated by endothelial cell K⁺ (Edwards et al., 1998). Thus, while the EDHF-like mechanism mediated near maximum relaxation (∼95%), 5–10 mM extracellular K⁺ caused only small relaxations (∼20%). Also, relaxations to K⁺ appeared to be mediated by activation of Na⁺/K⁺-ATPase since they were abolished by low concentrations of ouabain. By contrast, ouabain had no effect on the EDHF-like component of the response to BK. This finding supports those of Quignard et al. (1999) who showed that endothelium-dependent hyperpolarization to BK in the pig coronary artery was unaffected by ouabain. Finally, we have previously demonstrated that EDHF-like relaxations to BK in bovine and pig coronary arteries are unaffected by nifedipine in concentrations (>0.3 μM) that abolished contractions to KCl (Kilpatrick & Cocks, 1994; Drummond & Cocks, 1996). Thus, EDHF-like relaxations to BK in these tissues can occur independently of the need to close L-type VOCCs. This was not the case for the small relaxations to K⁺ since they were abolished by nifedipine.

Since our studies appear to exclude a role for cytochrome P₄₅₀ and K⁺, it remains a possibility that direct communication between endothelial cells and smooth muscles cells, via heterocellular gap junctions, was involved in the EDHF-like relaxations to BK (Kúhberger et al., 1994; Chaytor et al., 1998; Dora et al., 1999; Yamamoto et al., 1999). As described for similar EDHF-like responses in the guinea-pig mesenteric artery (Yamamoto et al., 1999), we have preliminary data which shows that the proposed gap junctional uncoupling agent, 18-β-glycyrrhetinic acid (Goldberg et al., 1996), appears to inhibit EDHF-like relaxation in the bovine isolated coronary artery (Selemidis & Cocks, unpublished observations). The pattern of inhibition we observed with 18-β-glycyrrhetinic acid was similar to that with the high (1 mM) concentration of ouabain used here. That is, 18-β-glycyrrhetinic acid not only inhibited the response to BK, but it also significantly improved the degree of block by L-NOARG. High concentrations of ouabain (0.1–1 mM) are known to prevent the formation of gap junctions and conversely 18-β-glycyrrhetinic acid is known to inhibit Na⁺/K-ATPase (Rabito et al., 1987; Watsky et al., 1990). Since our findings with lower concentrations of ouabain suggest that Na⁺/K⁺-ATPase is not involved in the EDHF-like response in the bovine coronary artery, the shared inhibitory effects of 18-β-glycyrrhetinic and high concentrations of ouabain may implicate a role for gap junctions. It remains to be determined whether such gap junctions are involved in transfer of the EDHF-like response from endothelium to smooth muscle (i.e. myoendothelial gap junctions) or if they co-ordinate the spread of hyperpolarization within a single layer of the vessel wall (i.e. endothelial-endothelial or muscle-muscle).

In conclusion, our studies show that three non-prostanoid factors operate in parallel to mediate endothelium-dependent relaxations to BK in the bovine coronary artery. NO is the dominant factor and is responsible for most of the response. After block of NO, an EDHF-like mechanism which does not involve either endothelial K⁺ or a cytochrome P₄₅₀-derived metabolite of arachidonic acid, compensates for nearly all of the maximum relaxation, whilst inhibition of both NO and the EDHF-like mechanism reveals a second level of backup perhaps mediated by a cytochrome P₄₅₀-dependent mechanism.

Abbreviations

AACOCF₃

arachidonyl trifluoromethyl ketone

ACh

acetylcholine

Ba²⁺

barium

BK

bradykinin

BK_Ca

large conductance Ca²⁺-activated K⁺ channel

DAG

diacylglycerol

EDHF

endothelium-derived hyperpolarizing factor

EDRF

endothelium-derived relaxing factor

EET

epoxyeicosatrienoic acid

HETE

hydroxyeicosatetranoic acid

high [K⁺]_o

high extracellular K⁺

K_IR

inward-rectifying K⁺ channel

L-NOARG

N^G-nitro-L-arginine

NO

nitric oxide

ODQ

1H-[1,2,4]-oxadiazolo[4,3-α]quinoxaline 1-one

PLA₂

phospholipase A₂

RHC 80267

1,6-bis-(cyclohexyloximinocarbonylamino)-hexane

SK_Ca

small conductance Ca²⁺-activated K⁺ channel

SKF-525a

N,N-diethylaminoethyl-2,2-diphenylvalerate

SNAP

S-nitroso-N-acetylpenicillamine

SNP

sodium nitroprusside

U46619

1,5,5-hydroxy-11,9-(epoxymethano)prosta-5Z,13E-dienoic acid