Introduction

Top Abstract Introduction Methods Results Discussion References

Endothelium-dependent vasodilation has generally been attributed to the release of NO. However, several agonists including acetylcholine and BK exhibit endothelium-dependent but NO-independent vasodilation (Quilley et al., 1994) and release of an EDHF has been invoked, a term first coined by Taylor and Weston (1988). Our studies in the rat heart and kidney are consistent with a CYP-derived metabolite of arachidonic acid in the NO-independent response, a concept supported by several recent reports (Bauersachs et al., 1994; Campbell et al., 1996; Popp et al., 1996). Accordingly, the vasodilator effect of BK is attenuated by inhibitors of phospholipases, CYP and K⁺ channels (Fulton et al., 1994a, 1995, 1996). CYP-derived HETEs were excluded as potential mediators of the coronary vasodilator response to BK as GC-MS analysis of coronary perfusates failed to detect release under basal conditions or after challenge with either BK or arachidonic acid. In contrast, EETs, detected in the coronary perfusates under basal conditions, were increased in response to arachidonic acid and bradykinin (Fulton et al., 1994b) and were considered as potential mediators of BK-induced vasodilation. Indirect evidence was also gained from the use of CYP inhibitors that exhibit differential activity against -hydroxylation and epoxygenation. Clotrimazole, considered to be more selective for CYP enzymes catalyzing epoxygenation, was more effective in inhibiting vasodilator responses to BK than 17-ODYA that exhibits no differential inhibitory activity (Fulton et al., 1995). Moreover, EETs have been shown to be produced by the endothelium, to elicit vasodilation and to activate K⁺ channels, thereby meeting some of the criteria for an EDHF (Bauersachs et al., 1994; Campbell et al., 1996; Popp et al., 1996). However, there are four distinct EET regioisomers, each of which may exist in two stereoisomeric forms. The formation of EET regioisomers is tissue and cell selective and the stereoselective formation is indicative of enzymic production (McGiff, 1991). Although the studies of Campbell et al. (1996) show that all four EETs are equipotent in relaxing the bovine coronary artery, it is unlikely that all would subserve the role of hyperpolarizing factors. Further, most ligand-receptor interactions exhibit strict stereochemical specificity and, in this regard, 11R,12S EET is a far more potent vasodilator agent than the corresponding 11S,12R isomer which is essentially inactive (Zou et al., 1996).

To evaluate which, if any, of the EET regioisomers could fulfill the requirements for a mediator of BK-induced vasodilation in the rat, we applied several pharmacological criteria based on our previous experiments with BK in the isolated kidney and heart. Thus, the vasodilator effect of the putative endothelium-derived mediator should not be affected by agents or interventions that prevent its synthesis but should be inhibited by agents that prevent its action on vascular smooth muscle. Consequently, we determined the vasodilator activity of the EET regioisomers in the presence of nifedipine (an inhibitor of voltage-dependent calcium channels) which was used as a probe for vasodilator mechanisms dependent on closure of voltage-gated Ca⁺⁺ channels and which has been shown to inhibit coronary vasodilator responses to BK (Fulton et al., 1994). Based on the results with nifedipine, only 5,6 EET was considered as a potential effector for BK. Subsequently, we compared vasodilator responses to 5,6 EET and BK in the presence of charybdotoxin, an inhibitor of Ca⁺⁺-activated K⁺ channels, which attenuates the coronary and renal vasodilator effect of BK, and iberiotoxin, a specific inhibitor of large conductance Ca⁺⁺-activated K⁺ channels, which does not affect the renal response to BK (Rapacon et al., 1996). In addition, we determined the vasodilator activity of 5,6 EET and BK in hearts denuded of endothelium and those treated with the CYP inhibitor, clotrimazole; these interventions should inhibit vasodilator responses to BK but not of its putative mediator. The results are consistent with 5,6 EET as a putative mediator of BK-induced vasodilation in the rat perfused heart.

	Methods

Top Abstract Introduction Methods Results Discussion References

Isolated hearts from male Wistar rats (350-450 g) were perfused according to the method of Langendorff (1895), as modified by Broadley (1979) and as described previously (Fulton et al., 1994a). Anesthesia was induced with pentobarbitone (65 mg/kg i.p.) and, after administration of heparin (1000 U/kg i.v.) and thoracotomy, the heart with attached aorta was excised and immersed in ice-cold Krebs' buffer. The aorta was cannulated and the heart perfused at constant flow (8-10 ml/min) with oxygenated Krebs' buffer at 37°C to obtain an initial perfusion pressure of 30 to 40 mmHg. The perfusate contained indomethacin (2.8 µM) to inhibit prostaglandin synthesis and nitroarginine (50 µM) to inhibit NO formation and elevate perfusion pressure to 130 to 140 mmHg to amplify vasodilator responses as well as to reproduce the experimental conditions under which our previous studies with BK were conducted (Fulton et al., 1992, 1994a, 1995, 1996; Rapacon et al., 1996). Once a stable, elevated perfusion pressure was attained, responses to bolus doses of the EET isomers and BK were determined.

In the first series of experiments, the vasodilator effects of the four EETs and BK were examined in the presence (n = 4) and absence (n = 4) of nifedipine (5 nM) which was added to the coronary perfusate 15 min beforehand, having first determined the effectiveness of nifedipine in inhibiting dilation to hyperpolarizing agents using cromakalim, an agonist for ATP-sensitive K⁺ channels. Nifedipine reduced vascular tone, associated with a decrease in perfusion pressure from approximately 130 to 140 mmHg to approximately 70 mmHg. Consequently, perfusion pressure was restored to its pretreatment level (130-140 mmHg) with the addition of the endoperoxide analogue, U46619 (10 ng/ml), to the perfusate. Based on the results obtained with nifedipine, only 5,6 EET was considered a potential mediator of the vasodilator effect of BK and, therefore, in all subsequent experiments only responses to 5,6 EET and BK were compared.

The effect of inhibition of Ca⁺⁺-activated K⁺ channels with charybdotoxin (10 nM) where n = 4 for the control and experimental groups, or iberiotoxin (20 nM) was determined on vasodilator responses to 5,6 EET and BK. Thus, we have previously shown that charybdotoxin inhibits the vasodilator effect of BK. Iberiotoxin was included because it is a specific inhibitor of large conductance Ca⁺⁺-activated K⁺ channels whereas charybdotoxin also inhibits intermediate conductance Ca⁺⁺-activated K⁺ channels and voltage-dependent K⁺ channels (Kuriyama et al., 1995). These agents were added to the perfusate at least 10 min before obtaining dilator responses. Next, we examined the effects of denudation of the endothelium on responses to 5,6 EET and BK. The endothelium was removed by brief (5-10 sec) perfusion of the heart with distilled water and introduction of 2 to 3 ml air into the perfusion line. For those experiments with iberiotoxin (n = 3) and endothelial denudation (n = 3), the same control group (n = 4) was used, as a control preparation was sandwiched between test preparations. In these experiments, we addressed nonspecific effects of the interventions by also testing vasodilator responses to bolus doses of nitroprusside (1 µg) and cromakalim (5 µg). Dilator responses to a bolus dose of NS 1619 (1 µg), an opener of Ca⁺⁺-activated K⁺ channels (Holland et al., 1996), were also determined in this group of experiments and those in which charybdotoxin was used, ostensibly to assess the effectiveness of the inhibitors.

Finally, we examined the effects of inhibition of CYP with clotrimazole (1 µM; n = 4) vs. ethanol vehicle (n = 4) on vasodilator responses to bolus doses of BK and 5,6 EET, the premise being that inhibition of synthesis of the putative mediator should prevent the response to the endothelium-dependent vasodilator agent but not that to the effector. Clotrimazole or vehicle (ethanol) was added to the perfusate at least 15 min before any vasodilator responses were determined. We also determined responses to nitroprusside to assess any potential non-specific actions of clotrimazole.

Statistical analysis. Results are presented as mean ± S.E.M. Vasodilator responses in control and experimental groups were compared by analysis of variance and differences between means evaluated by Newman-Keuls modified t test. P < .05 was considered statistically significant.

Materials. BK, indomethacin, U46619, nifedipine and clotrimazole were purchased from Sigma Chemical Co., St. Louis, MO. BK was dissolved in distilled water, indomethacin in 4.2% NaHCO₃, U46619 (11,9 epoxy methano prostaglandin H₂) in ethanol before dilution with water and nifedipine and clotrimazole in ethanol. Charybdotoxin and iberiotoxin were obtained from Peptides International, Louisville, KY and dissolved in distilled water. NS1619 (1-(2'-hydroxy-5'-trifluoromethylphenyl)-5-trifluoromethyl-2(3H)benzimidazolone) was obtained from Research Biochemicals International, Natick, MA and was stored in ethanol before diluting with water. The EETs were supplied by J. R. Falck, Dept. Molecular Genetics, University of Texas, Southwestern Medical Center at Dallas or purchased from Cayman Chemical Co., MI and administered in ethanol. The 5,6 EET used in the experiments with clotrimazole was synthesized by Dr. Bernd Spur of the Dept. Cell Biology at UMDNJ and administered in a solution of acetone:water (1:2).

	Results

Top Abstract Introduction Methods Results Discussion References

Mean basal perfusion pressures in the various groups ranged between 30 and 40 mmHg. Elevated perfusion pressures, induced by nitroarginine, were comparable in all groups (approximately 140 mmHg) except those treated with charybdotoxin or in which the endothelium was removed. Thus, charybdotoxin increased perfusion pressure over that seen with nitroarginine to 170 ± 6 mmHg compared to 141 ± 5 mmHg for the respective control group. Similarly, perfusion pressure following removal of the endothelium was 165 ± 2 mmHg compared to 140 ± 8 mmHg for the respective control group and 156 ± 7 mmHg for the iberiotoxin-treated group. Perfusion pressure in the nifedipine-treated group was restored to 144 ± 6 mmHg with U46619 after the nifedipine-induced reduction and this was comparable to the perfusion pressure of the control group, 141 ± 4 mmHg. Perfusion pressure in the clotrimazole-treated group was 135 ± 1 mmHg compared to 135 ± 4 mmHg for the respective control group.

The vehicles (ethanol, acetone:water or water) for the various dilator agents were without effect on coronary perfusion pressure when given as bolus injections of 10 µl.

Voltage-gated Ca⁺⁺ channels. Figure 1 shows coronary vasodilator responses to cromakalim, an ATP-sensitive K⁺ channel opener, in the absence and presence of nifedipine to inhibit voltage-gated Ca⁺⁺ channels. Cromakalim elicited dose-dependent falls in perfusion pressure that were almost abolished by nifedipine, showing that vasodilation resulting from hyperpolarization is dependent on closure of voltage-dependent Ca⁺⁺ channels as depicted at the bottom of figure 1. Figure 2 compares the vasodilator effects of the EET regioisomers under control conditions, i.e., in the presence of indomethacin and nitroarginine, and after exposure of the heart to nifedipine. Bradykinin was the most effective vasodilator agent, reducing perfusion pressure by 94 ± 2 mmHg. All four of the EETs dilated the coronary vasculature, 5,6 EET being the most effective and reducing coronary perfusion pressure by 74 ± 2 mmHg compared to less than 20 mmHg for equivalent doses of the other regioisomers. As we had shown before (Fulton et al., 1994a), the vasodilator activity of BK was markedly reduced (approximately 65%) by nifedipine which, at higher concentrations (10 and 20 nM), abolishes vasodilator responses to both BK and cromakalim (Fulton et al., 1994a). Nifedipine also decreased the response to 5,6 EET by 43% but the vasodilator effects of the other EET isomers were not reduced by inhibition of voltage-gated Ca⁺⁺ channels. Based on these results, the 8,9-, 11,12- and 14,15 EETs were excluded from consideration as potential mediators of the coronary vasodilator effect of BK and were not addressed further.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Effect of nifedipine (open bars) on changes in perfusion pressure (PP) in response to cromakalim in isolated hearts treated with indomethacin (2.8 µM) and nitroarginine (50 µM) to elevate PP to approximately 130 to 140 mmHg. Nifedipine reduced coronary perfusion pressure that was elevated to the pretreatment level with U46619 (10 ng/ml). *P < .05, **P < .01. The schematic depicts closure of voltage-gated calcium channels in response to hyperpolarization which results in reduced intracellular calcium and vasodilation.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Effect of nifedipine (5 nM) on vasodilator responses to bradykinin (BK) and the four epoxides (EET) in isolated hearts treated with indomethacin (2.8 µM) and nitroarginine (50 µM) which raised perfusion pressure (PP) to approximately 130 to 140 mmHg. Nifedipine reduced PP which was restored to its original level with U46619 (10 ng/ml). N = 4 for nifedipine (open bars) and control (solid bars). **P < .01.

Ca⁺⁺-activated K⁺ channels. We next assessed the effect of inhibition of Ca⁺⁺-activated K⁺ channels with charybdotoxin on coronary vasodilator responses to 5,6 EET and BK, as we have previously shown that the coronary and renal vasodilator effect of BK is dependent on a charybdotoxin-sensitive K⁺ channel (Fulton et al., 1994a; Rapacon et al., 1996). Charybdotoxin markedly reduced the vasodilator activity of BK and 5,6 EET by 73 and 61%, respectively (fig. 3), indicating the involvement of a charybdotoxin-sensitive K⁺ channel in the responses to both of these agents. In contrast, charybdotoxin did not influence vasodilator responses to NS1619, an opener of Ca⁺⁺-activated K⁺ channels, 45 ± 12 vs. 37 ± 6 mmHg for control.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Vasodilator responses to bradykinin (BK), 5,6 EET and NS1619 and the effect of charybdotoxin (10 nM) in isolated hearts constricted with nitroarginine (50 µM) to raise perfusion pressure to approximately 130 to 140 mmHg and treated with indomethacin (2.8 µM). Charybdotoxin caused a further elevation of PP to approximately 170 mmHg. N = 4 for charybdotoxin (open bars) and control (solid bars). *P < .05, **P < .01.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Effect of iberiotoxin (20 nM) on vasodilator responses to bradykinin (BK), 5,6 EET, nitroprusside (NP) and cromakalim in isolated hearts treated with indomethacin (2.8 µM) and constricted with nitroarginine (50 µM) to raise perfusion pressure (PP) to approximately 130 to 140 mmHg. Control (solid bars; n = 4), iberiotoxin (open bars; n = 3).

Endothelial denudation. The effects of endothelial denudation on coronary vasodilator responses to BK and 5,6 EET are shown in figure 5. The vasodilator action of BK was virtually eliminated by removal of the endothelium, 2 ± 2 vs. 72 ± 4 mmHg, confirming the obligatory role of an endothelium-derived mediator. In contrast, the vasodilator effect of 5,6 EET was unaffected by removal of the endothelium, 35 ± 3 vs. 32 ± 5 mmHg. Denudation of the endothelium was also without effect on the vasodilator response to cromakalim (68 ± 17 vs. 69 ± 12 mmHg) or nitroprusside (56 ± 2 vs. 52 ± 10 mmHg), showing that the capacity for vasodilation was not impaired.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Effect of removal of the endothelium on vasodilator responses to bradykinin (BK), 5,6 EET, nitroprusside (NP) and cromakalim in isolated hearts treated with indomethacin (2.8 µM) and constricted with nitroarginine (50 µM) to elevate perfusion pressure (PP) to approximately 130 to 140 mmHg. Removal of the endothelium resulted in a further elevation of PP to approximately 165 mmHg. Control (solid bars; n = 4); endothelium denuded (open bars; n = 3). **P < .01.

Inhibition of CYP. The effects of inhibition of CYP with clotrimazole on response to 5,6 EET and BK are shown in figure 6. As expected, the vasodilator response to BK was greatly reduced by clotrimazole, 7 ± 1 mmHg compared to 49 ± 3 mmHg for the control. In contrast, the coronary vasodilator effect of 5,6 EET was unaffected by clotrimazole as was the vasodilator response to nitroprusside.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Effect of the CYP inhibitor, clotrimazole (1 µM), on vasodilator responses to bradykinin (BK), 5,6 EET and nitroprusside in isolated hearts treated with indomethacin (2.8 µM) and constricted with nitroarginine (50 µM) to raise perfusion pressure (PP) to approximately 130 to 140 mmHg. N = 4 for clotrimazole (open bars) and control (solid bars).

	Discussion

Top Abstract Introduction Methods Results Discussion References

Several studies support a CYP-derived eicosanoid as an EDHF-mediating NO-independent vasodilator responses to BK and acetylcholine (Bauersachs et al., 1994; Campbell et al., 1996; Popp et al., 1996). Of the arachidonic acid metabolites, an EET has been considered the most likely candidate as EETs are produced by the endothelium, activate K⁺ channels and elicit vasodilation (Campbell et al., 1996). As there are four EET regioisomers, our study applied pharmacological criteria, based on previous observations with BK, to evaluate which of the EETs could be considered potential mediators of the NO-independent, but endothelium-dependent, vasodilator effect of BK in the rat isolated heart. Our approach was based on the premise that under identical experimental conditions to those used to demonstrate an NO-independent effect of BK, the vasodilator effect of the mediator should be independent of interventions aimed at preventing synthesis but should be susceptible to interventions aimed at preventing the effect of the putative mediator. However, caution is required in extrapolating findings obtained in a perfused system subjected to multiple pharmacological interventions to the physiological condition.

Our results are consistent with 5,6 EET as a potential mediator of BK-induced vasodilation in the rat heart in which the other EETs were excluded based on the initial experiments with nifedipine which reduced vasodilator responses to cromakalim, BK and 5,6 EET but failed to decrease the responses to the 8,9-, 11,12- and 14,15 EETs. The results with nifedipine, when viewed in isolation, could be explained by more than one mechanism particularly when considering that U46619 was required to elevate perfusion pressure after nifedipine treatment. However, when these observations are coupled with those implicating K⁺ channels (Fulton et al., 1994a; this study), they are consistent with the vasodilator effects of BK and 5,6 EET resulting from hyperpolarization of vascular smooth muscle. This, in turn, leads to closure of voltage-dependent Ca⁺⁺ channels resulting in reduced influx of extracellular Ca⁺⁺ and promotion of vasorelaxation. As voltage-dependent Ca⁺⁺ channels are reportly confined to smooth muscle rather than the endothelium (Revest and Abbott, 1992), these results indicate that nifedipine is interfering with the effects of BK and 5,6 EET at the level of the vascular smooth muscle. In contrast, the failure of nifedipine to reduce responses to the other EET isomers suggests that the vasodilator actions of these agents involves a different mechanism that is independent of closure of voltage-gated Ca⁺⁺ channels and, therefore, possibly of hyperpolarization. If this is the case, then these results seemingly contradict those of Campbell et al. (1996) who reported that 8,9 EET activates K⁺ channels, and those of Zou et al. (1996) who reported that 11R,12S EET increased the open probability of K⁺ channels although it should be noted that a racemic 11,12 EET was used in our study. In addition, we found 5,6 EET to be the most efficacious of the EETs in dilating the coronary vasculature of the rat in contrast to Rosolowsky and Campbell (1993) who reported that all of the EETs were equally effective in relaxing the bovine coronary artery. The difference in these results probably relates to experimental conditions, vessel size and species. Rosolowsky and Campbell (1993) used rings of bovine coronary artery whereas we used the rat perfused heart that will reflect responses of the microvasculature.

Further evidence that hyperpolarization is the mechanism underlying vasodilation in response to 5,6 EET derives from the experiments with charybdotoxin, an inhibitor of large conductance Ca⁺⁺-activated K⁺ channels that also affects channels of intermediate conductance as well as voltage-dependent K⁺ channels (Kuriyama et al., 1995). Thus, charybdotoxin was an effective inhibitor of vasodilator responses to 5,6 EET, a prerequisite if 5,6 EET is to be considered a mediator for BK because BK-induced vasodilation was also markedly impaired by charybdotoxin. These effects of charybdotoxin could not be attributed to nonspecific actions on vasodilator mechanisms as responses to NS1619 were unaffected. However, it is possible that the inhibitory effect of charybdotoxin on responses to BK results from an effect on endothelial K⁺ channel to prevent hyperpolarization, influx of Ca⁺⁺ and generation of a vasodilator mediator. This possibility can be disregarded for 5,6 EET that produces vasodilation in the absence of the endothelium. Nonetheless, the results with charybdotoxin are consistent with 5,6 EET as a potential mediator for BK. Similarly, the lack of an effect of iberiotoxin on responses to 5,6 EET, as well as those to BK, fail to eliminate 5,6 EET as a candidate and indicate that the vasodilator action of either agent is independent of large conductance Ca⁺⁺-activated K⁺ channels. However, these observations raise some concerns with respect to other studies that suggest that EETs activate large conductance Ca⁺⁺-activated K⁺ channels blocked by iberiotoxin. It is unlikely that the concentration of iberiotoxin (20 nM) that we used was insufficient to inhibit, even partially, large conductance K⁺ channels as iberiotoxin is reportedly more potent than charybdotoxin that caused substantial inhibition of vasodilator responses to BK and 5,6 EET at a concentration of 10 nM. Moreover, in studies of the NO-independent renal vasodilator action of acetylcholine which is also charybdotoxin sensitive, we failed to detect any inhibitory effect of iberiotoxin in concentrations up to 50 nM (Mieyal et al., 1998). We included NS1619 as a positive control as this agent has been reported to activate Ca⁺⁺-activated K⁺ channels (Holland et al., 1996). However, neither iberiotoxin nor charybdotoxin modified the vasodilator response to NS1619, suggesting that, in the rat heart, the response to this compound was independent of opening of K⁺ channels, a finding supported by Edwards et al. (1994).

Removal of the endothelium did not affect the vasodilator activity of 5,6 EET but abolished that elicited by BK, consistent with 5,6 EET as a putative mediator of the BK response. Thus, removal of the source of the mediator would be expected to eliminate the activity of an endothelium-dependent agonist such as BK but not affect the activity of the mediator applied directly, thereby, by-passing the site of synthesis. Similar results were obtained when an alternate approach was used to prevent the synthesis of the mediator. Thus, the CYP inhibitor, clotrimazole, virtually abolished the vasodilator effect of BK as expected from our previous studies (Fulton et al., 1992, 1994a) but had no effect on the vasodilator activity of 5,6 EET. These results support the concept of an endothelium-derived CYP product mediating the BK response that is inhibited by clotrimazole. For 5,6 EET to be considered as a candidate, interventions that prevent synthesis of the mediator should be without effect on responses to 5,6 EET or should actually enhance the vasodilator response by removing endogenous background levels. Clotrimazole is reported to exert effects in addition to inhibition of CYP, including inhibition of K⁺ channels (Alvarez et al., 1992; Edwards et al., 1996). However, we found no evidence of this in our study as clotrimazole had no effect on vasodilator responses to nitroprusside or on those to 5,6 EET which are dependent on K⁺ channels as they were inhibited by charybdotoxin (it should be noted that the concentration of clotrimazole used by Edwards et al. (1996) was 30 times greater than that used by us). Moreover, in previous studies, vasodilator responses to cromakalim, diazoxide or SCA 40, the last which is reported to open Ca⁺⁺-activated K⁺ channels (Laurent et al., 1993), were unaffected by 1 µM clotrimazole (Fulton et al., 1994a; Oyekan et al., 1994).

Of note from these studies, is that removal of the endothelium or inhibition of K⁺ channels by charybdotoxin further elevates coronary perfusion pressure over that seen with inhibition of NO synthesis with nitroarginine. This suggests that, in addition to NO, the endothelium produces another vasodilator that activates K⁺ channels and contributes to the maintenance of vascular tone. Whether this factor is a CYP-dependent eicosanoid remains to be determined. The inability of clotrimazole to further elevate perfusion pressure suggests otherwise although an effect of clotrimazole on Ca⁺⁺ entry could explain this (Villabos et al., 1992).

Based on this study, of the four EET regioisomers, 5,6 EET is the most likely potential candidate for the mediator of BK-induced coronary vasodilation. Comparisons between responses to 5,6 EET and BK show that both agents activate a charybdotoxin-sensitive, but iberiotoxin-insensitive, K⁺ channel for vasodilation. A recent study using a bioassay system for EDHF also supports a role for 5,6 EET as the half-life of the mediator was estimated to be about 90 sec (Popp et al., 1996) and only 5,6 EET of the EET regioisomers shows this degree of lability, the other EETs being relatively stable. Moreover, in our study 5,6 EET was the most efficacious of the EETs as a vasodilator agent. However, the dose of 5,6 EET (micrograms) required to elicit vasodilation compared to that of BK (nanograms) raises major questions that are not easily answered. The relative lack of activity of 5,6 EET given into an arterial line can be explained in several ways. Administration of 5,6 EET in this fashion does not reflect endogenous synthesis and release which might be abluminal and, consequently, to reach equivalent concentrations in the subendothelial space, large amounts of exogenous 5,6 EET would have to be administered, analogous to the situation with NO. This problem would be magnified for an unstable compound such as 5,6 EET that is subject to rapid uptake, metabolism by epoxide hydrolase and spontaneous degradation (McGiff, 1991). In addition, the endothelium could be expected to provide a barrier to exogenous 5,6 EET, preventing easy access to its site of action. However, this does not appear to be the case as removal of the endothelium did not enhance coronary vasodilator responses to 5,6 EET. Further, it should be noted that metabolism of 5,6 EET by endothelial cell cyclooxygenase was not required for 5,6 EET to induce vasodilation as reported for the rat caudal artery (Carroll et al., 1990). Thus, in our experiments removal of the endothelium did not modify the response to 5,6 EET and hearts were perfused with indomethacin to inhibit cyclooxygenase.

In summary, our results suggest that 5,6 EET can be considered a putative mediator for NO-independent BK-induced vasodilation, a phenomenon attributed to release of an EDHF. The vasodilator activity of 5,6 EET, like that of BK, appears to depend on a K⁺ channel that is charybdotoxin sensitive.