Coronary vasodilator responses to bradykinin in euglycemic and diabetic rats

Abstract

Diabetes mellitus is associated with endothelial dysfunction that is believed to result in impaired release of vasoconstrictor and vasodilator substances from the endothelium and thereby diminished reactivity of many vascular beds. This study was designed to characterize bradykinin (BK)-induced coronary vasodilation in normal and diabetic rats. Bradykinin-stimulated vasodilation of the rat coronary vasculature is mediated by a cytochrome P450-1A (CYP-1A)- inhibitable metabolite that activates K_Ca, but not K_ATP, channels on the coronary vascular smooth muscle. Although BK stimulates the release of nitric oxide from the vascular endothelium, the released nitric oxide and its ability to stimulate guanylate cyclase only modulates the duration of, rather than the magnitude of, BK-induced coronary vasodilation. Twelve weeks of streptozotocin-induced diabetes did not affect the coronary vascular responses to BK or the components that mediate BK-induced vasodilation (ie, K-channel activation, nitric oxide-guanylate cyclase). The data support the conclusions that the coronary vasodilator response of the rat to BK is CYP-1A and K_Ca-channel mediated, that coreleased nitric oxide only modulates the duration of BK-induced vasodilation, and that these mechanisms are unaffected by moderate diabetes. Am J Hypertens 2001;14:446–454 © 2001 American Journal of Hypertension, Ltd.

Coronary, vasculature, diabetes mellitus, bradykinin, endothelium-derived factors, potassium channels

The vascular endothelium modulates and regulates vascular tone through the production and release of vasoconstrictor and vasodilator substances. The impaired synthesis/release of these vasoactive substances by a dysfunctional endothelium has been implicated in the pathogenesis of the vascular complications of diabetes.¹ Kinins, and in particular bradykinin (BK), generated by vascular tissues stimulate the synthesis/release of endothelial-derived vasoactive substances through receptor-dependent mechanisms. In the presence of an intact coronary endothelium BK stimulates kinin B₂ receptors located on the microvascular arteriolar endothelium membrane to stimulate the synthesis/release of vasoactive mediators, which subsequently dilate the underlying vascular smooth muscle.² It is generally accepted that the vasodilator response to BK in many vascular beds is mediated by vasodilator prostaglandins and nitric oxide (NO).^3,4 In the coronary circulation, however, BK-induced vasodilation appears independent of prostaglandins, whereas the importance of NO as a mediator of BK-induced coronary vasodilation remains unresolved.^5,6 This is due in part to the existence of an endothelium-derived cytochrome P450 (CYP)-generated metabolite that contributes to endothelium-dependent relaxation. This mediator does not appear to be NO and appears responsible for endothelium-dependent hyperpolarization of vascular smooth muscle.^7,8 BK-induced vasodilation of the rat coronary circulation appeared dependent on a CYP-generated metabolite of arachidonic acid, although the CYP isozyme responsible was not defined.⁹ Thus, the role of NO as a mediator of BK-induced coronary vascular vasodilation remains undefined.

The aim of the present study was to assess the relative contribution of NO and the CYP-derived hyperpolarizing factor to the coronary vasodilator responses to BK in the isolated perfused rat heart and to evaluate the changes to these systems that occur in the coronary circulation of streptozotocin (STZ)-induced diabetic rats. This preparation was chosen for study because the changes in perfusion pressure with constant flow result in measured responses that reflect the events occurring in the coronary microcirculation at which level the coronary vascular resistance is regulated.

Methods

Drugs and Chemicals

N-imino-ethyl-L-ornithine (L-NIO) and troleandomycin were obtained from Alexis Corporation (San Diego, CA). Ibuprofen, bradykinin (acetate salt), adenosine triphosphate (ATP), methylene blue, glibenclamide, iberiotoxin, clotrimazole, 7-ethoxyresorufin (7-ER), and STZ were obtained from Sigma Chemical Co. (St. Louis, MO). Pinacidil was obtained from ICN Pharmaceuticals (Costa Mesa, CA). Glyceryl trinitrate (GTN) was obtained from Parke-Davis (Morris Plains, NJ). U46619 was obtained from Upjohn Co (Kalamazoo, MI). The sodium salt of 2-(N-N-diethylamino)-diazenolate-2-oxide (DEA-O) was provided by Dr. P.J. Kadowitz of Tulane University Medical Center, New Orleans, LA.

Animals

Male Sprague-Dawley SPF rats (Charles River Laboratories, Wilmington, MA) weighing between 250 and 275 g were used in these studies. Rats were housed two per cage in a controlled-temperature (20°C to 22°C) room where they were exposed to a 12-h light/dark cycle. All groups of animals received food and water ad libitum. The study was approved by the Louisiana State University Medical Center Institutional Animal Care and Use Committee.

Experimental Diabetes Mellitus

Rats were injected with STZ (45 mg/kg) prepared in a 0.02 mol/L sodium citrate solution (pH 4.5) through a tail vein. Age-matched controls were injected with a similar volume of sodium citrate solution. The rats were maintained for 12 weeks after administration of STZ or sodium citrate. Six-hour fasting glucose levels were measured in tail vein blood by a modified glucose oxidase method (Accucheck III monitor, Boehringer Mannheim, Indianapolis, IN) in both STZ and buffer-treated rats 4 days after the administration of drug or buffer and again at the time of sacrifice. Animals were considered diabetic when their blood glucose levels exceeded 350 mg/dL. Body weights were recorded on a weekly basis. All diabetic rats exhibited the clinical signs of polydipsia, polyuria, and polyphagia and continued to gain weight, albeit at a much reduced rate, when compared with control rats.

Coronary Perfusion

After induction of anesthesia (50 to 70 mg/kg ketamine, 5 to 7 mg/kg xylazine, intramuscularly) the abdominal cavity of the rat was opened and sodium heparin (100 U/kg) was injected into the inferior vena cava. The thoracic cavity was opened and the heart, lungs, and aorta removed and placed in ice-cold saline. Thymic tissue was then removed by blunt dissection, the hilar structures ligated with 4-0 silk, and the lung tissue removed distal to the hilar ligation. With the tissue maintained in ice-cold physiologic saline, the heart was connected to an aortic cannula, and an incision was made in the pulmonary artery. The heart was then perfused in a standard Langendorff apparatus (Radnoti Glass Technology, Monrovia, CA) with a Gilson Minipuls 3 pump at a flow rate of 12 mL/min. This rate of flow resulted in baseline coronary perfusion pressures of 42 + 3 mm Hg (n = 8).

The perfusate, Krebs-Henseleit buffer, contained the following: sodium chloride, 118 mmol/L; glucose, 5 mmol/L; potassium phosphate, monobasic, 1.2 mmol/L; magnesium sulfate, 1.2 mmol/L; potassium chloride, 4.7 mmol/L; calcium chloride, 2.5 mmol/L; sodium bicarbonate, 25 mmol/L; pyruvate, 2 mmol/L; Ca-EDTA, 0.5 mmol/L. Albumin (500 mg/L) was also included in the buffer. The buffer was aerated with 95% O₂-5% CO₂ mixture to maintain a pH of 7.40 to 7.48 at 37°C, filtered in line using a Whatman Polycap HD filter (10-μm pore size) and was not recirculated. The stable thromboxane A₂ mimetic U46619 (0.02 μg/mL) was added to the buffer to increase coronary vascular perfusion pressure to 100 to 120 mm Hg. Ibuprofen (10 μmol/L) was also included in the perfusate to block formation of prostacyclin, which has been reported to activate ATP-sensitive potassium (K_ATP) channels.¹⁰

Experiments were initiated when the preparation reached a stable coronary pressure (20 to 30 min). At that time all hearts were paced at 250 beats/min at 1.5 V (threshold for capture plus 10%) using a Grass Instruments S88 stimulator and allowed to stabilize before the experiment began. Coronary perfusion pressure was measured with a pressure transducer positioned at the level of the aortic valve. Because the coronary flow rate was maintained constant, changes in coronary perfusion pressure reflected changes in coronary vascular resistance. Isovolumic left ventricular pressure was measured with a pressure transducer connected to a fluid-filled latex balloon (size 3) inserted through the left atrium into the left ventricle. The left ventricular pressure transducer was positioned at a level corresponding to mid-ventricle, and the fluid volume was adjusted to produce a diastolic pressure of approximately 10 mm Hg. Left ventricular dP/dt was obtained by differentiating the left ventricular pressure signal. Bolus doses of saline (volume control) and agonists were injected through an injection port situated in-line immediately distal to the aortic cannula. Pressures and derived indices were continuously recorded on a polygraph. A computer sampled the data from the polygraph and converted it into digital form for storage and analysis.

Agonists and Antagonists

The dose-response to a particular agonist was determined both in the absence and presence of an antagonist. This was accomplished by the addition of an antagonist to the buffer reservoir after the initial dose-response determination, and treating the coronary vasculature for 10 to 15 min before redetermination of the dose-response to the agonist. Control hearts were perfused with buffer for 10 to 15 min if the antagonist was in aqueous solutions. Dimethylsulfoxide (DMSO) or ethanol was required in the cytochrome-P450 experiments to solubilize the inhibitors of these isozymes. Control hearts were, therefore, perfused with buffer containing 0.5% ethanol in the clotrimazole experiments or 0.03% DMSO in the 7-ER experiments (the highest concentrations to which the tissues were exposed). In all experiments the coronary vascular response to nitroglycerin (GTN) was determined both before and after treatment of antagonists (or solvent) to determine the specificity of any inhibitory effect. The response to an agonist was determined before and after treatment of the coronary vascular bed with inhibitor allowing each preparation to serve as its own control. Additional control experiments were performed to determine the degree to which preparation reactivity decreased over time.

Reactive Nitrogen Intermediates

The assay for reactive nitrogen intermediates (RNI) has been described in detail.¹¹ Vials (2 mL) were completely filled with coronary effluent (to displace atmospheric gases) and sealed with an O-ring cap. Samples (100 μL) were injected directly into a purge chamber containing 2.28% vanadium chloride in 4 N HCl at 100°C under nitrogen. The generated vapor stream of ultra-pure N₂ gas flowed through two refrigerated fractionating tubes (0°C) into a Dasibi 821 nitric oxide analyzer (Dasibi Environmental, Glendale, CA). The heated vanadium/HCl solution converted NO₂⁻ and NO₃⁻ into nitric oxide, which reacted with instrument-generated ozone to form excited nitric oxide. The latter released light in the red region (6500 to 8000). The amount of light generated was concentration dependent and was measured with a photomultiplier tube. The sensitivity of the instrument was 1 ppb to 1 ppm of free NO. The lower limit of sensitivity of the assay was 0.03 to 0.05 nmol in 10 μL of solution. Each sample was assayed in triplicate and the data expressed as free NO parts per billion.

Statistical Analyses

All data are expressed as mean ± standard deviation. The vasodilator responses to the agonists were normalized to the baseline coronary perfusion pressure and expressed as a percent change from basal pressure. Comparisons of body weights, blood glucose levels, and baseline left ventricular pressure and ±dP/dt, between diabetic and control groups were made using Student's t test (two-tailed). The RNI content of the perfusate (pre- and postinfusion of BK) was compared using a paired Student's t test (one-tailed). Dose-response data were analyzed using one-way analysis of variance with Bonferroni's correction for repeated measures analysis of variance. Dunnett's test (two-tailed) was used to identify significant differences between group means and control. Statistical significance was accepted at the P ≤ .05 level.

Results

Hemodynamic and Cardiac Characteristics of Euglycemic and Diabetic Rats

Nonfasting blood glucose levels were 135 ± 27 mg/dL in buffer-treated rats and 467 ± 32 mg/dL in diabetic rats (P ≤ .01). Body weight increased slightly in diabetic animals from 311 ± 20 g to 342 ± 48 g (10% ± 11% change) over the course of the study, whereas the body weights of the age-matched control rats increased from 316 ± 21 g to 528 ± 105 g (66% ± 23% change) (P ≤ .01).

Baseline left ventricular systolic pressures were similar in diabetic and age-matched euglycemic rats (104 ± 10 mm Hg and 112 ± 19 mm Hg, respectively; P ≥ .45), as was +dP/dt (2793 ± 459 mm Hg/sec and 2952 ± 376 mm Hg/sec, respectively; P ≥ .40). However, −dP/dt was significantly reduced in diabetic animals when compared with age-matched euglycemic rats (−1539 ± 240 mm Hg/sec and −2007 ± 463 mm Hg/sec, respectively; P ≤ .03).

BK Responses in Euglycemic Rats

The contribution of endothelium-derived NO to BK-mediated coronary arterial vasodilation was examined. Pretreatment of the coronary vasculature with L-NIO, a potent semi-irreversible inhibitor of constitutive and inducible NO synthase¹² produced dose-dependent increases in coronary resistance in the rat coronary vasculature (Fig. 1, top). However, infusion of L-NIO (100 μg/min or 50 μmol/L) did not affect the magnitude of BK-induced coronary vasodilation (Fig. 1, bottom). Although the magnitude of the vasodilator response to BK was not affected by treatment of the heart and coronary vasculature with L-NIO, this NO synthase inhibitor significantly reduced the duration of the vasodilator response to BK from 5.6 ± 1.3 min to 2.7 ± 1.3 min (P < .02; n = 5).

Effect of N-imino-ethyl-L-ornithine (L-NIO) on coronary perfusion pressure (top) and bradykinin (BK)-induced decreases in coronary perfusion pressure (bottom) in the rat Langendorff heart perfused at constant flow. The ordinate for the top panel is the peak percent change in perfusion pressure from basal perfusion pressure (110 mm Hg) in response to increasing concentrations of L-NIO (abscissa). The ordinate for the bottom panel is the change in coronary perfusion pressure (percent of basal perfusion pressure) to BK. The abscissa is the dose of BK. Vertical lines are the SD. The responses to BK in the presence and absence of L-NIO (100 μg/min) do not differ (P < .05). Each mean represents five rats/treatment.

Figure 1.

The L-NIO-mediated reduction of the duration of BK-induced coronary vasodilation may have been mediated by inhibition of NO release or resulted from the differences in basal pressure. We evaluated the effect of BK on the release of NO from the endothelium and the effect of L-NIO on this process. Samples of coronary effluent were collected for 10 sec (2 mL) immediately after the peak of BK-induced vasodilation. These samples exhibited a small but significant increase in the concentration of RNI when compared with effluent collected over an identical time frame after intracoronary injection of sterile saline (baseline, 28 ± 8 ppb; saline, 25 ± 9 ppb; BK, 36 ± 16 ppb; n = 14; (P = .02). Levels of RNI did not increase in the coronary perfusate of hearts pretreated with L-NIO and given BK (pre-BK 27 ± 11 ppb; post-BK 22 ± 7 ppb; (P = .24, n = 6). Although basal levels of RNI decreased after treatment with L-NIO, the decrease failed to reach statistical significance. The reason for this is unclear but may indicate incomplete blockade so these results should be viewed with some caution.

If BK-stimulated NO only modulates the duration of the coronary vasodilator response to this kinin, then inhibition of guanylate cyclase and inhibition of cyclic GMP, which mediates the responses to endothelium-derived NO, should mimic the effects of L-NIO. Pretreatment of euglycemic rat hearts with methylene blue (10 μmol/L), an inhibitor of guanylate cyclase, failed to significantly alter the magnitude of the vasodilator response to BK (Fig. 2, top), despite its significant attenuation of ATP-induced coronary vasodilation (16 to 165 ng) (Fig. 2, middle). However, similar to L-NIO, methylene blue attenuated the duration of the vasodilator response to a bolus dose of BK (16 ng) by 62% (from 5.6 ± 1.5 min to 2.1 ± 0.5 min, P < .001; n = 7) (Fig. 2, bottom).

Effect of methylene blue (MB, 10 μmol/L) on bradykinin (BK)-induced (top) and ATP-induced (middle) coronary vasodilation in the euglycemic rat. The effect of N-imino-ethyl-L (L-NIO, 100 μg/min) and methylene blue (10 μmol/L) on the duration of the BK-induced coronary vasodilation in the euglycemic rat (bottom). Ordinates for the top and middle panels are the change in coronary perfusion pressure (percent of basal perfusion pressure) to BK and ATP. The abscissa represents doses of BK and ATP given. The ordinate for the bottom panel is the time (in minutes) required for coronary perfusion pressure to return to 75% of its basal value from the nadir of the BK-induced vasodilation. Vertical lines are the SD. An * indicates the responses to ATP in the presence and absence of methylene blue differs (P < .05). Each mean represents five rats/treatment.

Figure 2.

The NO and agonists that release NO from the vascular endothelium have been reported to directly activate calcium-sensitive potassium (K_Ca) channels to produce relaxation of vascular smooth muscle.^12,–14 The hypothesis that the magnitude of BK-induced coronary vasodilation was dependent on activation of K_Ca channels was tested by measuring BK-induced (6 to 8 ng) changes in perfusion pressure before and after treatment of the coronary vascular bed with the K_ATP channel inhibitor glibenclamide (2 μmol/L) or the K_Ca channel inhibitor iberiotoxin (10 nmol/L). Ibuprofen was present in all experiments to inhibit prostanoid synthesis. Treatment of the hearts with glibenclamide did not affect perfusion pressure (119 ± 23 mm Hg and 122 ± 16 mm Hg, before and during perfusion with glibenclamide, respectively) or the magnitude of the vasodilator response to BK (Fig.. 3, top). However, treatment of the rat coronary circulation with iberiotoxin (10 nmol/L) for 20 min increased perfusion pressure from 116 + 18 mm Hg to 134 + 16 mm Hg (15.5% increase) (n = 5) and inhibited the magnitude of BK-induced coronary vasodilation (Fig. 3, bottom).

Effect of glibenclamide (2 μmol/L; top) or iberiotoxin (10 nmol/L; bottom) on bradykinin (BK)-induced coronary vasodilation in the euglycemic rat. The ordinates are the change in coronary perfusion pressure (percent of basal perfusion pressure) to BK. The abscissa are the doses of BK administered. Vertical lines are the SD. An * indicates the responses to BK in the presence and absence of antagonist differs (P < .05). Each mean represents five rats/treatment.

Figure 3.

Cytochrome P450 Inhibitors

The previous experiments demonstrated that BK-induced relaxation was associated with activation of K_Ca channels but that NO was not the mediator of BK-induced relaxation of the coronary microvasculature. The role of a CYP-derived metabolite as a mediator of BK-induced relaxation was examined. Treatment of the rat coronary circulation with clotrimazole, a nonspecific CYP isozyme inhibitor significantly inhibited (60% to 73%, P < .001, n = 5) BK-induced decreases in coronary vascular perfusion pressure (Fig. 4, top). Perfusion of the isolated rat heart with 7-ER (300 nmol/L), a selective inhibitor of CYP-1A, also significantly attenuated BK-induced coronary vasodilation (Fig. 4, middle), whereas a lower concentration of this inhibitor was without effect. Finally, perfusion of the rat coronary circulation with the mixed CYP-3A and heme-oxygenase inhibitor troleandomycin (200 μmol/L), a concentration that inhibits CYP-3A in the rat,¹⁵ failed to inhibit BK-induced vasodilation (P = .56) (Fig. 4, bottom).

Effect of clortrimazole (10 mmol/L; top), 7-ethoxyresorufin (7-ER, 300 nmol/L; middle), and troleandomycin (200 μmol/L; bottom) or their vehicles (control) on BK-mediated coronary vasodilation in the euglycemic rat. The ordinates are the change in coronary perfusion pressure (percent of basal perfusion pressure) to BK. The abscissa are the doses of BK administered. Vertical lines are the SD. An * indicates the responses to BK in the presence and absence of antagonist differs (P < .05). Each mean represents five rats/treatment.

Figure 4.

Comparison of BK Responses in Ruglycemic and Diabetic Rats

Fig. 5 compares the responses of the perfused rat coronary vascular bed to BK and the two components of the response to BK, NO, and K-channel activation. Intracoronary administration of BK (0.8 to 16 ng) to both STZ diabetic and age- and sex-matched euglycemic rats produced dose-dependent decreases in coronary vascular perfusion pressure, and thus coronary vascular resistance, in both groups of rats (Fig. 5, top). The magnitude and duration of BK-induced relaxation did not differ between the euglycemic and diabetic rats (Fig. 5, top). The DEA-O is a NO nucleophile adduct that releases NO spontaneously in aqueous buffer. Its half-life in buffer at 37°C and pH 7.4 is 2 min. Intracoronary administration of DEA-O (0.5 to 100 ng) produced dose-dependent coronary vasodilation that did not differ in the euglycemic and diabetic coronary vasculature (Fig. 5, middle). Finally, activation of K-channels with pinacidil (0.05 to 3 μg) also produced dose-dependent decreases in coronary perfusion pressure that did not differ in the coronary microvasculature of the euglycemic and diabetic rats (Fig. 5, bottom).

Effect of bradykinin (top), 2-(N-N-diethylamino)-diazenolate-2-oxide (DEA-O; middle), and pinacidil (bottom) on coronary perfusion pressure in the euglycemic rat and diabetic rat. The ordinates are the change in coronary perfusion pressure (percent of basal perfusion pressure). The abscissa represents doses of drug administered. Vertical lines are the SD. The responses to each of the agonists did not differ between the two experimental groups. Each mean represents five to eight rats/group.

Figure 5.

Discussion

Bradykinin-mediated relaxation of rings and strips of vascular smooth muscle from many species, as well as BK-mediated vasodilation, has been attributed to its ability to act on kinin B2 receptors on the vascular endothelium wherein it stimulates a small K_Ca channel conductance. This, in turn, increases intracellular calcium ion and upregulates the synthesis/release of NO synthase and prostacyclin.^2,3,16–19 However, anomalous results were obtained when BK was studied in vivo.^2,20–22 It was subsequently found that hyperpolarizing vasodilators increased K_ATP channel conductance in smooth muscle and vascular endothelium and that an endothelium-derived hyperpolarizing factor (EDHF) existed differing from that of NO and it appeared to be an arachidonic acid metabolite produced by CYP isozymes.^21–25 Subsequent studies demonstrated that the EDHF exhibited the characteristics of an arachidonic acid metabolite, appeared to be derived from a cytochrome P450 enzyme system, and hyperpolarized the smooth muscle cell membrane.^21–27 Moreover, studies in coronary vascular smooth muscle suggested that coronary vasodilation in response to BK was mediated by EDHF as well as NO in animal and human vessels.^28–30

Although BK-induced relaxation of rings and strips of conduit and small arteries are mediated by endothelium-derived prostacyclin and NO,^16,31 these endothelium-derived vasoactive substances do not appear to mediate BK-induced coronary vasodilation at the level of the resistance vessels. Previous studies demonstrated that administration of the NO synthase inhibitor L-NAME to the rat coronary circulation decreased the duration, but not the magnitude, of the vasodilator response to BK.^2,20,21 Similar results were also obtained in this study using L-NIO, the semi-irreversible inhibitor of nitric oxide synthases NOS II and NOS III.^32,33 Moreover, our studies demonstrated that BK-mediated coronary vasodilation was associated with measurable increases of NO in the coronary effluent. However, the concentrations produced were either insufficient to influence the magnitude of the vasodilator response to BK, as L-NIO inhibited the increased production of NO yet failed to affect the magnitude of BK-mediated coronary vasodilation, or occurred at a time after the peak vasodilator response. The ability of BK-mediated NO to modulate the duration of the coronary vasodilator response to BK appears to have resulted from NO-mediated increases of vascular smooth muscle cyclic GMP. This conclusion is speculative but based on the observation that methylene blue, which inhibits guanylate cyclase and may inactivate cyclic GMP,^34,35 mimicked the effect of inhibition of NO synthesis on the duration of BK-mediated coronary vasodilation. Because BK increased the NO content of the coronary effluent, it may be argued that the inability to affect the magnitude of BK-induced coronary vasodilation resulted from the inclusion of albumin in the perfusate. NO released by the vascular endothelium in vivo circulates in mammalian plasma primarily as an S-nitroso adduct of serum albumin.³⁶ Because NO is a lipid-soluble gas that reaches equilibrium with second order kinetics, the existence of S-nitrosolated albumin would not affect the diffusion of free NO into plasma.³⁷ Moreover, similar results were obtained in previous studies in which albumin was omitted from the perfusate.^20,30 Thus, the data fail to support a significant role for NO as a mediator of BK-induced coronary vasodilation in the rat.

The BK-mediated vasodilation of the perfused rat coronary vasculature appears to be mediated by BK-stimulated hyperpolarization of the vascular endothelium that increases the entry of calcium ions into the endothelium.³⁸ This is believed to activate phospholipase A2 and liberate arachidonic acid, which is acted on by cytochrome P450 to produce the endothelium-derived arachidonic acid metabolite that then hyperpolarizes the vascular smooth muscle membrane resulting in vasodilation.^29–39 The data presented herein support a role for a CYP-generated metabolite as a mediator of BK-induced coronary vasodilation as clotrimazole, which does not differentiate between the CYP isoforms, attenuated BK-induced relaxation. Although previous studies demonstrated the dependence of BK-induced coronary vasodilation on CYP, the isozyme involved was not defined.^6,9,20,30 The data presented herein demonstrate that inhibition of CYP-1A with 7-ER inhibited BK-induced coronary vasodilation, but that inhibition of heme-oxygenase and CYP-3A with troleandomycin was without effect. CYP-3A was selected for comparison because it has been shown to modulate cyclosporine-induced changes in blood pressure in spontaneously hypertensive rats.⁴⁰ Thus, BK-mediated coronary vasodilation in the rat is mediated by a CYP-1A derived metabolite and K_Ca channels are involved in the response.

Because we did not measure vascular smooth muscle membrane potential or channel activity, we cannot distinguish between a hyperpolarizing effect of the CYP-generated metabolite on the vascular smooth muscle membrane to produce vascular relaxation and BK-mediated hyperpolarization of the microvascular endothelium to stimulate the synthesis/release of the CYP-derived metabolite. It has been suggested that hyperpolarizing vasodilators, such as BK, act by stimulating K_ATP channels in some vascular smooth muscle membranes to produce relaxation.^21–26 Other studies suggest that the activation of K_ATP channels and hyperpolarization occur on the endothelial membrane to mediate the synthesis/release of the vasodilator metabolite.^27–29 However, most of these studies were performed with isolated vessels that may differ from the microvascular smooth muscle and endothelium. Glibenclamide, the K_ATP channel blocker, did not increase coronary vascular resistance or inhibit the coronary vasodilator response to BK in our study. In contrast, iberiotoxin, a highly selective inhibitor for K_Ca channels, increased coronary resistance in the perfused hearts and attenuated BK-induced coronary vasodilation. These data support the conclusion that K_Ca channels contribute to both the intrinsic regulation of, and BK-mediated decreases of, vascular resistance in the coronary microcirculation of the rat.

It may be argued that NO, rather than a CYP-derived hyperpolarizing factor mediates BK-induced coronary microvascular vasodilation as NO has been shown to activate K_Ca channels in aortic smooth muscle by reacting with a thiol-containing domain of the channel.¹² Moreover, endothelium-dependent vasodilators have also been shown to promote vasodilation of various vascular beds by activation of K_Ca channels.^13,14 Several factors argue against this possibility. First, although BK-induced coronary vasodilation was associated with release of NO, inhibition of BK-induced NO synthesis with L-NIO failed to attenuate the magnitude of BK-induced coronary vasodilation. Second, the K_Ca channel inhibitor iberiotoxin attenuated BK-mediated vasodilation, whereas GTN-mediated vasodilation was refractory to inhibition by this toxin. These data support the conclusion that the magnitude of BK-induced coronary vasodilation and activation of K_Ca channels was not mediated by BK-induced release of NO from the rat coronary vascular endothelium.

Previously we reported that coronary flow was decreased in STZ-induced diabetic rats and that this impaired coronary flow was prevented by treatment with an angiotensin converting enzyme inhibitor.⁴¹ Because angiotensin converting enzyme inhibitor degrades BK, as well as converting angiotensin-I into angiotensin-II, we were interested to determine whether the coronary vasodilator response to BK was altered by diabetes. Our findings suggest that each of the components of the BK response, ie, BK-induced relaxation, NO-mediated relaxation, and K-channel activation, are not significantly impaired in the coronary microvasculature of diabetic rats under the conditions tested. Because chronic diabetes is associated with impaired endothelium-dependent vasodilation in both animals^42–45 and patients,^46–49 our failure to identify any impairment in the coronary response to BK suggests this is not the case for these vessels as assessed by this agonist. However, the effect of an agonist, such as acetylcholine, which is more dependent on the synthesis of NO may yield different results. The inability to find differences in the magnitude of the NO donor-mediated coronary vasodilation between the euglycemic and diabetic rats also suggests that the intracoronary cyclic GMP system is not impaired by diabetes, a finding that is in agreement with several studies.^42,50 In addition, K-channel activation has been reported to be both attenuated and enhanced in diabetic animals.^1,51 However, the inability to demonstrate any differences in response to BK and pinacidil in diabetic rats compared with age-matched controls strongly suggests that these channels are unaltered by diabetes in rat coronary microvessels. The apparent lack of agreement between studies using isolated strips and rings and those using intact perfused organs may reflect the differences between the characteristics of the large arteries and the microcirculation that regulates vascular resistance. Further studies are required to differentiate these possibilities.

In summary, our data indicate that the magnitude of vasodilator response to BK in the rat coronary circulation is independent of NO formation, but that the maintenance of the response is modulated by NO. A CPY-1A-derived substance and K_Ca channels mediate the magnitude of the response to BK, which is consistent with EDHF formation. Although endothelium-dependent reactivity of many vascular beds may be impaired in diabetes, our results suggest that this is not so in the rat coronary vasculature as assessed by bradykinin, although it may be true for agonists, eg, acetylcholine, that are more dependent on the formation of prostaglandins and NO.