Introduction

Top Abstract Introduction Methods Results Discussion References

Acetylcholine induces vasodilation by releasing several factors from endothelial cells. One of these, NO, is produced from L-arginine by NO synthase and evokes relaxation by activating soluble guanylate cyclase in vascular smooth muscle cells (Furchgott and Vanhoutte, 1988). Prostacyclin, a metabolite of arachidonic acid produced by cyclooxygenase, also is released from endothelial cells in several vascular beds and acts via the activation of adenylate cyclase (Torphy, 1994). ACh also has been reported to hyperpolarize vascular smooth muscle cells via an endothelium-dependent mechanism. It is suggested that endothelium-dependent hyperpolarization underlies the NO/prostaglandin-independent relaxation in several types of blood vessels (Félétou and Vanhoutte, 1988; Huang et al., 1988) and is produced by a diffusible factor, EDHF (Chen et al., 1991).

Although EDHF is believed to activate K⁺ channels in vascular smooth muscle cells, no consensus of view has been reached on the identity of the K⁺ channels activated by EDHF. Studies with K⁺ channel blockers have produced conflicting results. In the rabbit aorta and carotid artery and the rat mesenteric resistance artery, EDHF-mediated relaxation was inhibited by ChTX (Cowan et al., 1993; Hwa et al., 1994; Lischke et al., 1995) which is an inhibitor of BK_ca channels (Strong, 1990). In the rabbit middle cerebral arteries, ACh-induced smooth muscle hyperpolarization was abolished by glybenclamide, an inhibitor of K_ATP channels (Standen et al., 1989). In the rat and rabbit mesenteric arteries, the effects of EDHF were prevented by apamin, an inhibitor of SK_Ca channels (Adeagbo and Triggle, 1993; Murphy and Brayden, 1995; Chen and Cheung, 1997). Furthermore, in the rat hepatic and mesenteric arteries and guinea pig carotid and basilar arteries, the action of EDHF was abolished by a combination of ChTX plus apamin, but unaffected by either apamin or ChTX alone (Corriu et al., 1996b; Zygmunt and Högestätt, 1996; Petersson et al., 1997).

Several recent studies suggested that cytochrome P450 epoxygenase in endothelial cells contributes to the production of EDHF (Hecker et al., 1994; Chen and Cheung, 1996; Campbell et al., 1996). The metabolites of arachidonic acid produced by cytochrome P450 epoxygenase, i.e., 5,6-, 8,9-, 11,12- and 14,15-EETs, are synthesized in endothelial cells (Rosolowsky and Campbell, 1996) and induce hyperpolarization and vasodilation in bovine (Rosolowsky and Campbell, 1993), porcine (Hecker et al., 1994; Popp et al., 1996) and canine (Rosolowsky et al., 1990) coronary arteries. EETs also have been shown to activate BK_Ca channels (Hu and Kim, 1993; Campbell et al., 1996; Popp et al., 1996). Furthermore, cytochrome P450 inhibitors such as proadifen and clotrimazole have been shown to inhibit EDHF-mediated hyperpolarization and relaxation (Hecker et al., 1994; Chen and Cheung, 1996; Campbell et al., 1996; Popp et al., 1996). These observations have led to the conclusion that EDHF is an EET. However, more recent evidence casts doubt on this contention, because 17-ODYA, a suicide-substrate inhibitor of cytochrome P450 epoxygenase responsible for the production of EETs, has no effect on EDHF-mediated responses in the rat hepatic and mesenteric arteries (Zygmunt et al., 1996; Fukao et al., 1997a).

ACh has produced significant endothelium-dependent membrane hyperpolarization in guinea pig coronary artery (Keef and Bowen, 1989; Eckman et al., 1992; Hammarstrom et al., 1995), and EDHF has played an important role in eliciting relaxation (Parkington et al., 1995). The present study investigated the effects of inhibitors of the enzymes participating in synthesis of EETs, i.e., PLA₂ and cytochrome P450 epoxygenase, on the response to ACh in the guinea pig coronary artery. We also examined the effects of several K⁺ channel inhibitors on ACh-induced NO/prostaglandin-independent relaxation in this tissue.

	Methods

Top Abstract Introduction Methods Results Discussion References

Contraction measurement. Male Hartley guinea pigs (450-600 g) were anesthetized with pentobarbital (50 mg/kg i.p.) and exsanguinated. The heart was removed and placed in cold Krebs' solution with the following composition (mM): NaCl, 113; KCl, 4.8; NaHCO₃, 25; KH₂PO₄, 1.2; CaCl₂, 1.8; MgSO₄, 1.2; and glucose, 5.5 aerated with a mixture of 95% O₂ and 5% CO₂. Coronary arteries were dissected out and cleaned of all adhering fat and myocardial tissue.

Current measurement. Single vascular smooth muscle cells were isolated from guinea pig coronary arteries by a dispersal procedure similar to that described previously (Ishikawa et al., 1993). The dissected coronary arteries were cut into small segments and placed in Ca⁺⁺-free Hanks' solution (mM: NaCl, 125; KCl, 5.4; NaHCO₃, 15.5; KH₂PO₄, 0.44; Na₂HPO₄, 0.34; sucrose, 2.9; glucose, 10, and aerated with 95% O₂-5% CO₂) for 90 min at room temperature. The segments were then placed in Ca⁺⁺-free Hanks' solution containing 1 mg/ml collagenase (cls2, Worthington Biochemical, Freehold, NJ), 2 mg/ml trypsin inhibitor (type II-S, Sigma Chemical, St. Louis, MO), 5 mg/ml bovine serum albumin, 0.2 mg/ml protease (type XIV, Sigma) and 0.2 mg/ml Na₂-ATP, and incubated for approximately 20 min with gentle agitation at 35°C. After the completion of digestion, single cells were dispersed by gentle agitation in Ca⁺⁺-free Hanks' solution. The cells obtained were stored in Ca⁺⁺-free Hanks' solution containing 1 mg/ml bovine serum albumin and 0.5 mg/ml trypsin inhibitor at 4°C.

Ca⁺⁺ measurement. The dissected coronary artery of guinea pig was cut open longitudinally and divided into approximately 2-mm-length strips. The strips were incubated with 5 µM fluo 3-AM with 0.03% cremophor EL in Krebs' solution aerated with a mixture of 95% O₂ and 5% CO₂ for 2 hr at room temperature, and then rinsed several times with Krebs' solution. Each strip was pinned to a silicone plate (2 × 15 × 1 mm in width × length × height) with the endothelium facing upward. The silicone plate was mounted on a glass-coverslip-bottomed chamber containing 3 ml Krebs' solution aerated with a mixture of 95% O₂ and 5% CO₂, with the endothelium facing the objective of an inverted microscope (Olympus IX-70, Tokyo, Japan), and with space of approximately 0.5 mm between the endothelial surface and the coverslip so that the bath solution freely passed between them. Fluorescence images were collected with use of a confocal laser-scanner (Yokogawa CSU10, Tokyo, Japan) connected to an image analyzer system (Applied Imaging Quanticell 700, Wunderland, UK). An excitation wavelength of 488 nm was provided by an argon laser. For measurement of intracellular free Ca⁺⁺ concentration ([Ca⁺⁺]_i), xy-scanning images composed of 256 × 256 pixels were acquired every 0.5 sec. The mean intensity of fluo 3 fluorescence in the selected area of the image was divided by resting fluorescence intensity, and the relative intensity was used as an indicator for [Ca⁺⁺]_i.

Drugs. The following drugs were used: 4-AP, apamin, ChTX, cromakalim, dexamethasone, U-46619, IbTX, indomethacin, L-NNA, 17-ODYA and proadifen (all from Sigma); acetylcholine chloride (Daiichi Seiyaku, Tokyo, Japan); TEA (Wako, Osaka, Japan); AACOCF3 (Calbiochem, La Jolla, CA); and 5,6-,11,12- and 14,15-EETs (Cascade Biochem, Berkshire, UK).

Data analysis. Values are presented as mean ± S.E. and n indicates the number of experiments (animals). EC₅₀ values for vasorelaxant responses were obtained from individual concentration-response curves as the concentration at which the half-maximal relaxant response occurred. EC₅₀ and maximum response (R_max) were determined by fitting the data to the logistic equation: R = R_max/(1 + (pC/pEC₅₀)^f), where R is the response, pC the negative logarithm of drug concentration, pEC₅₀ the negative logarithm of EC₅₀ and f the slope factor. Curve fitting was carried out by the use of SigmaPlot software (Jandel Scientific, San Rafael, CA). Statistical analysis of EC₅₀ and R_max values was performed by analysis of variance, with significant differences between groups being identified by Bonferroni's post hoc test. P values less than .05 were considered to be significant.

	Results

Top Abstract Introduction Methods Results Discussion References

ACh-induced endothelium-dependent relaxation resistant to L-NNA and indomethacin. In the guinea pig coronary artery preconstricted with U-46619 (30 nM), ACh elicited concentration-dependent relaxation. After the addition of L-NNA (0.1 mM), an NO synthase inhibitor, there was a significant right shift in the concentration-response curves for ACh, with no significant change in the maximum relaxation; pEC₅₀ was 7.84 ± 0.22 before and 7.04 ± 0.18 after (P < .05), and R_max was 87.6 ± 2.0 before and 87.2 ± 2.5% after the addition of L-NNA (n = 6). The response to ACh in the presence of L-NNA was not affected by indomethacin (10 µM), a cyclooxygenase inhibitor (data not shown), but was abolished by atropine (1 µM), a muscarinic receptor antagonist (data not shown) or by endothelium removal, which was accomplished by treating the inner aspect of vessels with collagenase (1 mg/ml) for 5 min (fig. 1B), which suggests that the response was mediated by muscarinic receptors and was endothelium-dependent. The L-NNA/indomethacin-resistant relaxation was attenuated profoundly when the K⁺ concentration in the bath solution was raised to 20 mM (fig. 1B). The response was also inhibited by TEA (10 mM), a nonselective K⁺ channel inhibitor (n = 5, data not shown). These properties coincide with those of EDHF reported previously (Chen et al., 1991), which suggests that ACh-induced L-NNA/indomethacin-resistant relaxation in the guinea pig coronary artery is mediated by EDHF.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Endothelium-dependent, NO/prostaglandin-independent relaxation induced by ACh in coronary arteries constricted by U46619 (30 nM). (A) ACh caused concentration-dependent relaxation in the absence () and presence () of L-NNA (0.1 mM). (B) In the presence of L-NNA (0.1 mM) and indomethacin (10 µM), the relaxant response to ACh () was abolished in the presence of 20 mM KCl () and when the endothelium was denuded (). Relaxation is expressed as percent of contraction before the addition of ACh. Data are presented as mean ± S.E. of four to nine experiments.

Effects of PLA₂ inhibitors. Arachidonic acid is released from the membrane phospholipid bilayer by activation of cytosolic PLA₂. If EDHF is a metabolite of arachidonic acid produced by cytochrome P450 epoxygenase, then the response mediated by EDHF should be attenuated by the inhibition of cytosolic PLA₂. We explored the effects of two different PLA₂ inhibitors, dexamethasone and AACOCF3, on EDHF-mediated relaxation. Dexamethasone has been shown to decrease arachidonic acid release and the subsequent production of prostaglandins in cultured endothelial cells (Rosenbaum et al., 1986), and AACOCF3, a trifluoromethyl ketone analog of arachidonic acid, also inhibits cytosolic PLA₂ activity by directly binding to the enzyme (Street et al., 1993). Pretreatment with dexamethasone (10 µM) or AACOCF3 (10 µM) had no effect on ACh-induced L-NNA/indomethacin-resistant relaxation (fig. 2, table 1).

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Effects of PLA2 inhibitors on NO/prostaglandin-independent relaxation induced by ACh. Graphs show concentration-response relationship for ACh-induced relaxation in the absence () and presence () of 10 µM dexamethasone (A) or 10 µM AACOCF3 (B) in coronary arteries constricted by U46619 (30 nM). All experiments were performed in the presence of L-NNA (0.1 mM) and indomethacin (10 µM). Relaxation is expressed as percent of contraction before addition of ACh. Data are presented as mean ± S.E. of four experiments.

Effects of cytochrome P450 inhibitors. In the presence of L-NNA (0.1 mM) and indomethacin (10 µM), proadifen (10 µM), an inhibitor of a wide range of cytochrome P450-dependent enzymes, almost abolished the vasodilator response to ACh (fig. 3) in the guinea pig coronary artery. Proadifen also significantly suppressed contraction induced by U-46619 (30 nM) to 78.0 ± 3.4% (n = 5) of the control. In contrast to proadifen, 17-ODYA (10 µM), a suicide-substrate inhibitor of the cytochrome P450 pathway responsible for the formation of EETs, had no effect on the concentration-response curve for ACh in the presence of L-NNA (0.1 mM) and indomethacin (10 µM) (fig. 3, table 1).

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Effects of cytochrome P450 inhibitors on ACh- and cromakalim-induced relaxation. Graph shows concentration-response relationship for ACh-induced (A, B) and cromakalim-induced (C, D) relaxation in the absence () and presence () of 10 µM proadifen (A, C) or 10 µM 17-ODYA (B, D) in coronary arteries constricted by U46619 (30 nM). All experiments were performed in the presence of L-NNA (0.1 mM) and indomethacin (10 µM). Relaxation is expressed as percent of contraction before addition of ACh. Data are presented as mean ± S.E. of four experiments. The concentration-response curves for cromakalim did not fit the logistic equation described under "Methods."

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Effects of cytochrome P450 inhibitors on whole-cell K+ currents in isolated smooth muscle cells of the guinea pig coronary artery. Time courses of peak outward currents elicited by voltage step to 0 mV from a holding potential of 70 mV every 15 sec are shown (A). The cell was superfused with 10 µM proadifen () or 10 µM 17-ODYA (). Current traces, a through d, were obtained at the time correspondingly indicated in the left panel. Graphs show mean values (mean ± S.E.; n = 4) of the currents obtained with ramp pulses from 90 to +90 mV for 900 msec at a holding potential of 50 mV before () and 2 min after exposure to 10 µM proadifen (; B) or 5 min after exposure to 10 µM 17-ODYA (; C). The amplitude of currents was measured in 10-mV increments.

Response to EETs. As shown in figure 5, 14,15-EET (1 and 3 µM) did not cause vasodilation in arteries preconstricted with U-46619 (30 nM) (n = 5). Application of 5,6- and 11,12-EETs at concentrations up to 3 µM also elicited no response (n = 5 and 6, respectively; data not shown).

View larger version (8K): [in this window] [in a new window]   Fig. 5.   Typical trace showing effect of 14,15-EET in coronary arteries constricted with U46619 (30 nM). Numbers in the figure indicate log molar concentrations of 14,15-EET. The experiment was performed in the presence of L-NNA (0.1 mM) and indomethacin (10 µM).

Effects of K⁺ channel blockers. 4-AP, which is a blocker of K_dr and K_ATP channels, at 1 mM suppressed and at 10 mM completely abolished the relaxation induced by ACh in the presence of L-NNA (0.1 mM) and indomethacin (10 µM) (fig. 6A, table 1). In agreement with previous work in guinea pig coronary artery (Eckman et al., 1992), pretreatment with glybenclamide (10 µM), a K_ATP channel blocker, did not affect the response to ACh (n = 3, data not shown). Apamin (0.1 µM) and IbTX (0.1 µM), SK_Ca and BK_Ca channel blockers, respectively, also had no effect on this relaxation (fig. 6, B and D; table 1). In contrast, ChTX (0.1 µM), which also is known to be a BK_Ca channel blocker, significantly shifted the concentration-response curve for ACh to the right, and suppressed the maximum response (fig. 6C, table 1).

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Effects of K+ channel blockers on NO/prostaglandin-independent relaxation induced by ACh. Graphs show concentration-response relationship for ACh-induced relaxation in the absence () and presence of 1 mM () or 10 mM () 4-AP (A), 0.1 µM apamin (; B), 0.1 µM ChTX () or 0.1 µM ChTX plus 0.1 µM apamin (; C), or 0.1 µM IbTX () or 0.1 µM IbTX plus 0.1 µM apamin (; D). All experiments were performed in the presence of L-NNA (0.1 mM) and indomethacin (10 µM). Relaxation is expressed as percent of contraction before addition of ACh. Data are presented as mean ± S.E. of four to six experiments.

View larger version (11K): [in this window] [in a new window]   Fig. 7.   Effect of apamin on NO/prostaglandin-independent relaxation induced by ACh in the presence of ChTX or IbTX. After 10 min incubation with ChTX (0.1 µM; A) or IbTX (0.1 µM; B), ACh (10 µM) was applied to coronary arteries constricted by U46619 (30 nM). When maximal response to ACh was attained, apamin (0.1 µM) was administered. The results of statistical analysis are shown in C; open and closed bars show responses in the presence of ChTX and IbTX, respectively. Data are presented as mean ± S.E. of four experiments.

View larger version (17K): [in this window] [in a new window]   Fig. 8.   ACh-induced increase in intracellular Ca++ concentration ([Ca++]i) in endothelial cells in coronary artery strips. ACh (1 µM)-induced changes in [Ca++]i in endothelial cells were measured with a confocal laser scanning microscopy in regular () and Ca++-free () Krebs' solutions (A), and before () and after () the treatment with ChTX (0.1 µM) plus apamin (0.1 µM) in regular Krebs' solution (B). Tracings in each panel show the time course of changes in relative intensity of fluo 3 fluorescence in the selected area containing three to five cells in the same preparation. ACh was added at Time 0. Panel C shows the results of statistical analysis of ACh (1 µM)-induced changes in [Ca++]i in endothelial cells in the absence (Control) and presence of ChTX (0.1 µM) plus apamin (0.1 µM) in regular Krebs' solution. Closed and open bars show the relative intensity of fluo 3 fluorescence at the first peak and 1 min after the addition of ACh, respectively. Data are presented as mean ± S.E. of seven experiments.

	Discussion

Top Abstract Introduction Methods Results Discussion References

EDHF recently has been proposed to be an EET, a metabolite of arachidonic acid produced by cytochrome P450 epoxygenase, in porcine, bovine and rat coronary arteries (Hecker et al., 1994; Bauersachs et al., 1994; Campbell et al., 1996; Popp et al., 1996). In the guinea pig coronary artery, ACh elicits endothelium-dependent relaxation and hyperpolarization which are independent of NO and prostaglandins (Keef and Bowen, 1989; Eckman et al., 1992; Hammarstrom et al., 1995; Parkington et al., 1995); however, whether EDHF in this tissue is an EET has not yet been elucidated. The results obtained in the present study suggest that EDHF in the guinea pig coronary artery is unlikely to be an EET, because 17-ODYA, which is a selective inhibitor of cytochrome P450 epoxygenase responsible for the generation of EETs (Zou et al., 1994), had no effect on ACh-induced L-NNA/indomethacin-resistant relaxation, and exogenously applied EETs caused no relaxation in this tissue. The lack of effect of IbTX, a specific BK_Ca channel inhibitor, on EDHF-mediated relaxation also can rule out the possible involvement of EETs in the response in guinea pig coronary artery, because EETs cause membrane hyperpolarization by activating BK_Ca channels (Hu and Kim, 1993; Campbell et al., 1996; Popp et al., 1996).

In the cytosol of endothelial cells, various substances such as prostaglandins and EETs can be synthesized from free arachidonic acids (Proctor et al., 1987; Rosolowsky and Campbell, 1993). Although these reactions are achieved via several enzymes, these cascade responses are initiated by the activation of cytosolic PLA₂, which selectively releases free arachidonic acids from membrane phospholipids (Suga et al., 1990). In the porcine and rat coronary arteries, PLA₂ inhibition with quinacrine has been shown to attenuate the EDHF-mediated relaxant response to bradykinin (Bauersachs et al., 1994; Hecker et al., 1994). However, Fukao et al. (1997a) reported that although quinacrine nearly completely eliminated ACh-induced hyperpolarization in the rat mesenteric artery, the drug also abolished the hyperpolarizing effects of pinacidil, a K_ATP channel opener, which suggests a nonspecific effect of quinacrine. We explored the possible involvement of PLA₂ in ACh-induced, L-NNA/indomethacin-resistant relaxation with more specific PLA₂ inhibitors, dexamethasone and AACOCF3. The present results showed that neither dexamethasone nor AACOCF3 caused any effect on the relaxation, which suggests that PLA₂-induced release of arachidonic acid is not involved in the generation of EDHF in the guinea pig coronary artery. This finding supports the view that EDHF is unlikely to be an EET that is a cytochrome P450 epoxygenase metabolite of arachidonic acid. This result also rules out the proposal by Randall et al. (1996) that anandamide, which is derived from arachidonic acid and ethanolamine, represents EDHF.

Since Hecker et al. (1994) proposed that EDHF might be a cytochrome P450 metabolite, several studies have shown that EDHF-mediated relaxation and hyperpolarization are inhibited by cytochrome P450 inhibitors such as proadifen and clotrimazole (Bauersachs et al., 1994; Lischke et al., 1995; Chen and Cheung, 1996; Campbell et al., 1996). However, these agents apparently cause inhibitory effects on Ca⁺⁺-activated K⁺ channels in the human red blood cell (Alvarez et al., 1992) and K_ATP channels in several blood vessels (Edwards et al., 1996; Graier et al., 1996; Zygmunt et al., 1996; Fukao et al., 1997a). This is an important point, because EDHF elicits hyperpolarization of smooth muscle cells by activating K⁺ channels (Nagao and Vanhoutte, 1993). That is, the inhibitory effects of these agents on EDHF-mediated relaxation might be caused by direct inhibition by the agents of K⁺ channels on smooth muscle cells. In good agreement with previous reports (Edwards et al., 1996; Graier et al., 1996; Zygmunt et al., 1996; Fukao et al., 1997a), the present study demonstrated that proadifen abolished not only EDHF-mediated relaxation but also the relaxation induced by cromakalim, a K_ATP channel opener, in the guinea pig coronary artery. The patch-clamp experiment also showed that proadifen almost abolished K_dr current in isolated smooth muscle cells of the guinea pig coronary artery. The inhibition by proadifen of K_dr current also was shown in the rat portal vein, in which, however, there are no reports of EDHF as an endogenous factor (Edwards et al., 1996). Although the type of K⁺ channels activated by EDHF has not been identified, these results suggest that the inhibition by proadifen of the relaxant response to EDHF might result from the inhibition of K⁺ channels on smooth muscle cells rather than inhibition of the pathway involved in the formation of EDHF in endothelial cells. In the light of the K⁺ channel inhibitory properties of cytochrome P450 inhibitors such as proadifen, the inhibition by these drugs of EDHF-mediated responses is not firm evidence for the conclusion that EDHF is a product of a cytochrome P450-dependent pathway.

EETs have been reported to elicit relaxation and hyperpolarization in several vessels (Proctor et al., 1987; Rosolowsky and Campbell, 1993; Hecker et al., 1994; Campbell et al., 1996), whereas EETs do not cause relaxation in the rat hepatic (Zygmunt et al., 1996) and guinea pig coronary (the present study) arteries. Furthermore, 17-ODYA has been shown to inhibit the bradykinin-induced, EDHF-mediated vasodilation in the rat (Fulton et al., 1995) and porcine (Popp et al., 1996) coronary and rabbit carotid (Dong et al., 1997) arteries, whereas the drug fails to inhibit EDHF-mediated relaxation or hyperpolarization in the rat hepatic (Zygmunt et al., 1996) and mesenteric (Fukao et al., 1997a) and guinea pig coronary (the present study) arteries. In the guinea pig carotid artery, endothelium-dependent hyperpolarization produced by ACh is not affected by several cytochrome P450 inhibitors including proadifen and 17-ODYA (Corriu et al., 1996a). Graier et al. (1996) reported that NO-independent relaxation in bovine and porcine coronary arteries contributes to two distinct pathways; cytochrome P450 epoxygenase-derived compounds are involved in that in the bovine artery, but not in the porcine artery. Although the reasons for these discrepancies are not certain, they may be explained by species and/or tissue differences or suggest that EDHF is not the same in all tissues.

The present patch-clamp study showed that 17-ODYA had no effect on K_dr current in isolated smooth muscle cells of the guinea pig coronary artery. In contrast to this result, Edwards et al. (1996) reported that not only proadifen but also 17-ODYA inhibits K_dr current in the rat portal vein cells. In their investigation, the inhibitory effect of 17-ODYA on K_dr currents was recorded 20 min after the application of 17-ODYA. In the present study, we observed K⁺ current only for 5 min after the addition of 17-ODYA. This difference may be the reason for the discrepancy. In the conventional dialyzed whole-cell configuration, significant "run-down" of K_dr current usually is observed when the configuration is kept for more than 15 min (Ishikawa et al., 1993). We therefore performed preliminary experiments in which cells were incubated with 17-ODYA for more than 30 min; then the access to cells was gained. K_dr current similar to that without 17-ODYA in other cells was observed (n = 3, data not shown). Edwards et al. (1996) showed discrepant results that 17-ODYA had no effect on the spontaneous mechanical activity of the portal vein, in contrast to ciclazindol, a K_dr channel inhibitor, which increased the magnitude, prolonged the duration and disrupted the regular pattern of spontaneous contractions. It also was reported that, whereas proadifen depolarizes the resting membrane potential probably by inhibiting K_dr channels, 17-ODYA produces no detectable change in it in the rat mesenteric artery (Fukao et al., 1997a). Although we cannot rule out the possible contribution of cytochrome P450 epoxygenase to the gating of K_dr channels, the inhibition of cytochrome P450 epoxygenase is unlikely to be responsible for the profound inhibition by proadifen of the channels.

The present study showed that ChTX inhibited the NO/prostaglandin-independent relaxation induced by ACh in the guinea pig coronary artery. However, the effect of ChTX apparently is independent of BK_Ca channels, because IbTX, a more selective inhibitor of BK_Ca channels (Galvez et al., 1990), had no effect on EDHF-mediated relaxation, which is consistent with the results in the rat hepatic (Zygmunt and Högestätt, 1996) and guinea pig basilar (Petersson et al., 1997) arteries. ChTX blocks not only BK_Ca but also IK_Ca and SK_Ca channels (Strong, 1990). Furthermore, ChTX recently was demonstrated to inhibit cloned voltage-dependent K⁺ channels, Kv1.2 and Kv1.3 (Grissmer et al., 1994). Therefore, it is highly possible that K⁺ channels other than BK_Ca channel might contribute to the inhibition by ChTX of the EDHF-mediated relaxation.

Several studies have shown that EDHF-mediated hyperpolarization and relaxation are inhibited by a combination of ChTX plus apamin (Corriu et al., 1996b; Zygmunt and Högestätt, 1996; Chen and Cheung, 1997; Petersson et al., 1997). The present study also demonstrated that ACh-induced L-NNA/indomethacin-resistant relaxation in the guinea pig coronary artery was abolished completely by the combination of ChTX plus apamin. Apamin caused no inhibition of the response to ACh when it was applied alone or in the presence of IbTX. Furthermore, when the maximal response to ACh was achieved in the presence of ChTX, the following application of apamin completely abolished the relaxation. These results suggest that the presence of ChTX is indispensable in the inhibitory effect of apamin. It is hypothesized, therefore, that ChTX and apamin inhibit a single type of K⁺ channel in a synergistic manner. In a recent report which supports this hypothesis, Zygmunt et al. (1997) showed that in an homogenate of rat brain, in which the binding of [¹²⁵I]ChTX is displaced by agitoxin-2, an inhibitor of voltage-dependent K⁺ channels, but not by IbTX, apamin increases [¹²⁵I]ChTX binding in a concentration-dependent manner. Alternatively, the K⁺ conductances regulated by two distinct types of K⁺ channel, one sensitive to apamin and the other to ChTX, may act in a synergistic manner to produce vasodilation. Further investigations are required to clarify the mechanism.

In the present study, the NO/prostaglandin-independent relaxation induced by ACh was abolished completely by the combination of ChTX plus apamin. K⁺ channels sensitive to these inhibitors appear to be involved in the NO/prostaglandin-independent relaxation induced by ACh. Here, it must have been considered that these K⁺ channel inhibitors might interact with K⁺ channels on endothelial cells. The elevation of [Ca⁺⁺]_i in endothelial cells is essential for the release of EDHF (Chen and Suzuki, 1990). Fukao et al. (1997b) recently proposed that ACh-induced release of EDHF from endothelial cells of the rat mesenteric artery is initiated by Ca⁺⁺ release from intracellular stores and maintained by Ca⁺⁺ influx via nonselective cation channels coupled to depletion of intracellular stores. It is likely, therefore, that depolarization of endothelial cells by the inhibition of K⁺ channels reduces the Ca⁺⁺ influx as a consequence of a decreased electrical driving force. Rusko et al. (1992) reported that endothelial cells have BK_Ca and IK_Ca channels. K⁺ channels sensitive to ChTX and apamin, conductance of which is 18 and 9.1 pS in symmetric K⁺ solutions, respectively, also were identified in endothelium of intact rat aorta (Marchenko and Sage, 1996). However, the present study ruled out the possibility that the combination of toxins reduces Ca⁺⁺ influx by inhibiting K⁺ channels on endothelial cells. We demonstrated that the plateau phase of ACh-induced increase in [Ca⁺⁺]_i in endothelial cells, which was produced by Ca⁺⁺ influx, was not affected by the treatment with the combination of ChTX plus apamin. The result suggests that the toxins affect channels on smooth muscle cells.

In conclusion, the failure of the selective cytochrome P450 inhibitor and PLA₂ inhibitors to inhibit relaxation mediated by EDHF and of externally applied EETs to elicit relaxation suggests that EDHF is unlikely to be an EET in the guinea pig coronary artery. It is very likely that the inhibitory effect of proadifen on EDHF-mediated relaxation results from the inhibition by the drug of K⁺ channels on smooth muscle cells or endothelial cells, but not the reduced synthesis of EDHF by cytochrome P450 epoxygenase in endothelial cells. The present study also provided no evidence that BK_Ca channels contribute to the NO/prostaglandin-independent relaxation induced by ACh in the guinea pig coronary artery. K⁺ channels on smooth muscle cells, which both apamin and ChTX must affect for inhibition to occur, appears to be the target for EDHF.