Mechanisms of Vasodilation of Cerebral Vessels Induced by the Potassium Channel Opener Nicorandil in Canine In Vivo Experiments

Transcription

1 1644 Mechanisms of Vasodilation of Cerebral Vessels Induced by the Potassium Channel Opener Nicorandil in Canine In Vivo Experiments Tadahiko Ishiyama, MD; Shuji Dohi, MD; Hiroki Iida, MD; Shigeru Akamatsu, MD; Shuichirou Ohta, MD; Hiroyuki Shimonaka, MD Background and Purpose Nicorandil, a potent antianginal agent characterized as a potassium channel opener, could produce cerebrovascular dilation in in vitro studies. Our aim was to investigate the pharmacologic response to the topical application of nicorandil on the vasomotor tone of pial vessels in vivo. To elucidate its mechanism, we also studied the inhibitory action of methylene blue and glibenclamide against nicorandil-induced vasodilation. Methods In 14 dogs prepared with a parietal cranial window, we administered five different concentrations of nicorandil solution (1CT 7,10"*, lfr 5,10" 4, and 10" 3 mol/l) under the window and measured pial arteriolar and venular diameters. pretreating pial vessels with either 10~ 3 mol/l methylene blue or 10" 5 mol/l glibenclamide, we examined inhibitory action after the application of 10" 3 mol/l nicorandil. In additional experiments with 9 dogs, we evaluated the effects of nitroglycerin and cromakalim on pial vessels in Nicorandil belongs chemically to the nitrate group 1 and was originally introduced as a coronary vasodilator. 2-3 Another nitrate agent, nitroglycerin, causes endothelium-independent vasodilation in rat basilar arteries 4 and produces a doserelated dilation of basilar and pial arterioles in in vivo experiments. 5 Nicorandil may also possess a cerebrovascular dilating action, but its pharmacologic effect is different from classic nitrates. 6 Although nitroglycerin mainly dilates the capacitance vessels, 7 nicorandil has been reported to dilate coronary arterioles. 8 Resistive vessel dilators are more beneficial than conductive vessel dilators as cerebrovascular-dilating drugs. However, it has not been confirmed whether nicorandil dilates cerebral vessels. The vascular effect of nicorandil is characterized by opening of the potassium channels of the smooth muscle cell membranes. 9 A recent in vitro observation in smooth muscles of rabbit cerebral arteries suggests that K + channels located in cerebral arterial smooth muscles serve as a negative feedback pathway to control the Received August 5, 1993; final revision received May 16, 1994; accepted May 20, From the Department of Anesthesiology and Critical Care Medicine, Gifu University School of Medicine, Gifu City, Gifu, Japan. Reprint requests to Shuji Dohi, MD, Department of Anesthesiology and Critical Care Medicine, Gifu University School of Medicine, 40 Tsukasa-machi, Gifu City, Gifu 500, Japan American Heart Association, Inc. the absence or presence of 10 5 mol/l methylene blue and 10 5 mol/l glibenclamide, respectively. Results Nicorandil produced significant, concentration-dependent dilation of pial vessels (P<.05). Methylene blue blocked nicorandil-induced dilation, whereas glibenclamide only attenuated such action of nicorandil. Nitroglycerin and cromakalim also produced a concentration-dependent increase in pial arteriolar and venular diameters (P<.05), and those effects were blocked in the presence of methylene blue or glibenclamide, respectively. Conclusions Our in vivo study demonstrates that topical application of nicorandil dilates both pial arterioles and venules in a concentration-dependent manner and suggests that the mechanisms of such action are most likely due to both cyclic GMP-mediated vascular smooth muscle dilation and the regulation of K + flux. (Stroke. 1994^5: ) Key Words microcirculation nitroglycerin dogs potassium channels vasodilation See Editorial Comment, page 1650 degree of membrane depolarization and vasoconstriction. 10 In a few in vitro experiments, nicorandil has been suggested to produce dilation on rat basilar arteries 11 and dog basilar and cerebral arteries 12 ; however, the mechanism for its putative cerebral-vasodilating action has not been studied in in vivo experiments. The endothelium-derived relaxing factor (EDRF) is involved generally in the regulation of systemic vasomotor tone and more specifically in the regulation of cerebral vascular activity. To our knowledge, the effects of topical application of nicorandil on pial vessels have not been reported in in vivo experiments. We investigated its effects on vasomotor tone of normal pial arterioles and venules in anesthetized dogs using the cranial window technique. To elucidate its mechanisms, we also evaluated the inhibitory effect of methylene blue and glibenclamide against nicorandil-induced vasodilation in the current in vivo model. Materials and Methods We studied 23 dogs weighing between 6 and 10 kg. Experimental protocols were approved by our institutional animal care committee. All dogs were initially anesthetized with a bolus infusion of pentobarbital sodium (20 mg/kg) and thereafter maintained with a continuous intravenous infusion (2 mg/kg per hour). tracheal intubation, each dog was mechanically ventilated with a positive-pressure respirator and received vecuronium intravenously for muscle paralysis. Tidal

2 Ishiyama et al Nicorandil and Pial Microcirculation 1645 TABLE 1. Physiological Measurements In Rrst Experimental Stage During Topical Application of NIcorandll Concentration, mol/l 10" 7 io-« MAP, 111±1 115±2 117±2 113±2 HR, bpm 123±8 124±9 131 ±6 134±7 BT, C ± ±0.1 ph ± ± ±0.01 PacOj, 40.3± ± ± ±0.7 Pao,, ± Na, meq/l 146.2± ± ± ± ' ±8 36.8± ± ± ±1.1 MAP Indicates mean arterial blood pressure; HR, heart rate; and BT, body temperature. Values are mean+sem. K, meq/l 3.06± ± ± ± ±0.15 volume and respiratory rate were adjusted to maintain PacOj between 35 and 45, and supplemental oxygen was given to maintain Pao? between 90 and 120. The left femoral vein was cannulated to administer fluid and drugs. The left femoral artery was also cannulated to measure mean arterial pressure (MAP) continuously and to provide blood samples for analyzing arterial blood gas tensions, ph, and serum electrolytes. Rectal temperature was maintained between 36.5 and 37J5 C by a water-circulating warming blanket. A closed CTanial window was used for visualization of the pial microcirculation. The scalp was retracted and temporal muscle was removed. A 2-cm-diameter hole was made in the parietal cortex. the coagulation of dural vessels with a bipolar electrocoagulator, the dura was cut and retracted over the bone. A stainless steel ring with a glass coverslip was placed over the hole and secured with bone wax and dental acrylic. The space under the window was filled with artificial cerebrospinal fluid (acsf), and four polyethylene catheters were inserted into the ring. One catheter was attached to a reservoir bottle containing acsf to maintain a constant intracranial pressure of 7 cm H 2 O. Two other catheters were used for infusion of acsf or study drug solutions and the last one was for continuous monitoring of intracranial pressure. The volume below the window was between 0.5 and 1 ml. The composition of acsf was Na +, 151 meq/l; K +, 4 meq/l; Ca 2+, 3 meq/l; Cl~, 110 meq/l; and glucose, 100 mg/dl; ph was adjusted to 7.48, and the solution was bubbled with 5% CO 2 and air at 37.0 C. Nicorandil was freshly dissolved in acsf; five different concentrations (10~ 7, 10" 6, 10" 5, 10" 4, and 10" 3 mol/l) were prepared and used for the current study. The diameters of three or four pial arterioles and venules were measured sequentially with each different nicorandil concentration by a videomicrometer (Olympus Flovel model VM-20) attached to a microscope (model OMK-1, Olympus). All in vivo experiments were carried out in the following manner. 30 minutes of stabilization, pial arteriolar and venular diameters, MAP, heart rate, rectal temperature, arterial blood gas tensions, ph, and serum electrolytes were measured before and after topical application of the five different nicorandil concentrations into the cranial window in sequential stages in eight dogs. To establish the baseline vessel size, the window was continuously flushed with acsf at the rate of 0.5 to 1 ml/min for 60 minutes after each measurement. 60 minutes from the last administration of nicorandil solution, the pial vascular diameter returned to control values. The effects of methylene blue and glibenclamide on pial vessels were evaluated against the action of topical 10" 3 mol/l nicorandil solution in six dogs with the aforementioned experimental design. Methylene blue was dissolved in acsf to make a 10-' mol/l solution. the completion of baseline measurements following topical use of 10" 3 mol/l methylene blue, 10-' mol/l nicorandil solution was applied and the measurements were repeated. Glibenclamide was dissolved in 100% dimethyl sulfoxide (DMSO) and then diluted with acsf to make a 10" 3 mol/l solution. The DMSO concentration in the 10 3 mol/l glibenclamide solution was 0.1%. A similar experiment was repeated with either 0.1% DMSO or 10" 3 mol/l glibenclamide alone followed by 10~ 5 mol/l nicorandil solution. In additional experiments with nine dogs, we also examined the role of guanylate cyclase and K + channels in mediating the pial vascular responses to nicorandil. control measurements of pial vascular diameter and laboratory data, three different concentrations (10" 7, 10" 3, and 10" 3 mol/l) of nitroglycerin, a classic nitrate, and cromakalim, an ATP-sensitive K + channel opener, were administered under the window, and pial arteriolar and venular responses were evaluated. To elucidate the efficacy and specificity of blockade with methylene blue or glibenclamide, we evaluated the effects of preapplication of 10" 5 mol/l methylene blue or 10" 3 mol/l glibenclamide after the topical application of nitroglycerin or cromakalim, respectively. Cromakalim was dissolved in DMSO and diluted with acsf. Pial vessels were divided into two groups on the basis of initial diameter greater than 200 /xm or less than 200 ^.m. All physiological variables; concentration-dependent effects of nicorandil, nitroglycerin, or cromakalim; and inhibitory effects of methylene blue or glibenclamide for nicorandil-, nitroglycerin-, or cromakalim-induced pial vascular dilation were examined via one-way ANOVA and Schefte's F test for post hoc comparison. Comparison between absolute values for pial arteriolar or venular diameter was made using a paired r test. A value of P<.05 was considered statistically significant. Values are represented as mean±sem. Results MAP, heart rate, body temperature, arterial blood gas tensions, ph, and serum electrolytes did not change significantly in the first experiment with five different nicorandil concentrations (Table 1). Nicorandil produced significant pial arteriolar and venular dilations (Table 2). Concentration-dependent dilations in large and small arterioles and large venules between two stepwise concentration differences (P<.05, Fig 1) were also detected. In small pial venules, a concentration-related increase in diameter was detected at the lower concentration but not at higher concentrations (P<.05). The percent increase in diameter was more prominent in small arterioles at concentrations of i<r 6 (p<.05), 10-5 (p<.ooi), icr 4 (p<.oooi), and 10" 3 mol/l (P<.01, Fig 1A). There was no significant percent change in diameter between large and small venules (Fig IB). Methylene blue, DMSO, and glibenclamide had no effect on pial vascular diameter (Table 3). Methylene blue inhibited 10" 5 mol/l nicorandil-mediated pial arteriolar (large, P<.005; small, P<.001) and venular (large, P<.0005; small, P<.05) dilations (Fig 2). There was no significant difference in percent change in diameter with topical application of 10" 5 mol/l nicorandil

3 1646 Stroke Vol 25, No 8 August 1994 TABLE 2. Effects of Nicorandil on Pial Arterioles and Venules Concentration, mol/l ± ± ± ± ± ± ± ± ±9.12* ± Values are mean±sem. *P<.05, tp<-0005, + P<.0001 compared with corresponding control. 10" ± ± ±7.21t ± ± ± " ± ± ± ±8.05* ± ± ±7.98 io ± ± 30.46* ± ±9.28t ±32.93* ± ± " ±34.70t ± 10.02* ± ± ±9.8* alone or DMSO plus 10 5 mol/l nicorandil. Glibenclamide reduced the dilation of small pial arterioles (/><.05) and venules (P<.05) induced by 10" 5 mol/l nicorandil (Fig 2). MAP, heart rate, body temperature, arterial blood gas tensions, ph, and serum electrolytes did not change before and after topical methylene blue, glibenclamide, or DMSO (Table 4). Topical application of nitroglycerin caused a concentration-dependent increase in diameter in small pial arterioles and large and small venules (P<.05; Table 5, Fig 3). Methylene blue inhibited pial arteriolar (small, P<.005) and venular (large, P<.005; small, /><.0001) dilation mediated with lo" 3 mol/l nitroglycerin, but A O 9 change of large alcriolcs 40" 0 % change of small ancriolcs 35" 30 25" 20" " nicorandil concenlration(-logm) 20" % change of large venules % change of small vcnulcs nicorandil concenlralion(-logm) FIG 1. Line graphs show concentration-related effect of nicorandil on pial arteriolar (A) or venular (B) diameter in 14 dogs. Data are expressed as percent change in diameter. Nicorandil produced a concentration-dependent dilation of large and small pial arterioles and large venules within the range of 10~ 7 to 10" 3 mol/l (large arterioles, 10~ 7 mol/l, 10~ 6 < 10" 5 < 10" 3 mol/l; small arterioles, 10~ 7 mol/l, 10-6 <10-5 <10-'<10-3 mol/l; large venules, 10" 7 mol/l, 10-6 <10-5 <10~ 3 mol/l, P<.05). Percent increase in diameter was more prominent in small than large arterioles. Values are mean±sem. *P<.05, tp<.01, + P<.001, P< glibenclamide did not affect the nitroglycerin-induced dilation (Fig 4). A concentration-dependent vasodilation in pial arterioles and venules (P<.05) was observed after topical application of cromakalim (Table 6, Fig 5). The vasodilation induced by cromakalim was attenuated in the presence of glibenclamide (large arterioles, F<.005; small arterioles, / > <.0005; large venules, i > <.005) but not in the presence of methylene blue (Fig 6). Discussion The results of the present in vivo study demonstrate that the topical application of nicorandil dissolved in acsf dilates pial arterioles and venules in a concentration-dependent manner and that vasodilative responses TABLE 3. Effects of Methylene Blue, Dimethyl Sulfoxide, or Glibenclamide on Pial Arterioles and Venules Methylene Blue DMSO Glibenclamide ± ± ± ± ± ± ± ± ± ± ± ±7.61 DMSO indicates dimethyl sulfoxide. Topically applied methylene blue, DMSO, and glibenclamide had no effect on pial vascular diameter. Values are mean±sem.

4 o Large Arterioles Small Arterioles Small Venules FIG 2. Bar graphs show effects of methylene blue or glibenclamide on nicorandil-induced dilation of pial arterioles (A) or venules (B). Methylene blue inhibited nicorandil-induced pial arteriolar and venular dilation. Glibenclamide attenuated the dilation of small pial arterioles and large and small venules induced by nicorandil. Values are mean±sem. Black bars indicate 10" 5 mol/l nicorandil; hatched bars, 10~ 5 mol/l methylene blue plus 10~ 5 mol/l nicorandil; and shaded bars, 10~ 5 mol/l glibenclamide plus 10~ 5 mol/l nicorandil. *P<.05, tp<-005, $P<.001, P<.0005 compared with corresponding 10~ 5 mol/l nicorandil. to nicorandil are more prominent in small arterioles. Since nicorandil-induced dilation of pial arterioles and venules was inhibited by methylene blue and attenuated by glibenclamide, mechanisms of such action are most likely caused by both cyclic GMP-mediated vascular smooth muscle dilation and the regulation of K + flux. We also observed that topical application of nitroglycerin, a classic nitrate, and cromakalim, an ATP-sensitive K + channel opener, caused a concentration-dependent dilation in canine pial arterioles and venules and that each dilation was blocked with topically applied methylene blue or glibenclamide. In addition, the presence of methylene blue and glibenclamide did not affect pial vasodilation induced by cromakalim and nitroglycerin, respectively. These data obtained from the present in vivo study allow us to conclude that the mechanisms of action of nicorandil on pial vessels are specifically mediated by cyclic GMP and a K + channel opener. Vasodilator effects of nicorandil in the coronary artery have been well established, but little is known about the effect of nicorandil on the cerebral vascular circulation. Nitrates, when topically applied, induce endothelium-independent vasodilation in basilar 4-5 and cerebral 13 arteries. Pinacidil and cromakalim as K + channel openers have been reported to elicit vasodilation in in vitro experiments Since the vascular effects of nicorandil are based on its action as a nitrate Ishiyama et al Nicorandil and Pial Microcirculation 1647 TABLE 5. Effects of Nitroglycerin on Pial Arterioles and Venules Concentration, mol/l 10" 10 10" ± ± ± t ± ± ± 12.84* t t ± ± S * ±7.61* t ± ± : ±13.84* ±11.204: Values are mean±sem. Nitroglycerin induced dilations in small pial arterioles and large and small venules. *P<.05, tp<01, $P<.005 compared with corresponding control. and as a K + channel opener, we suspected that nicorandil may also possess cerebrovascular dilating action. The present in vivo study confirms the fact that nicorandil dilates large and small pial arterioles and venules, with the greatest effect on small pial arterioles. Several studies have noted the correlation between blood flow and vessel diameter in the brain stem and peripheral microcirculation. Theoretically, nicorandil may have a clinical value, improving cerebral blood flow through its vasodilator effect. Alteration in arterial diameter may modulate changes in blood flow. The present results demonstrate that a concentration-dependent increase in small pial arteriolar diameter was observed with a higher concentration of nicorandil, but no increase was detected in small pial venules within the concentration range studied. Because of its nitrate moiety in chemical structure, nicorandil has been suggested to cause a sustained dilation of both venous capacitive and arterial resistance vessels. 19 In the present acute in vivo experiments in dogs anesthetized with pentobarbital, which has been known to decrease cerebral blood flow associated with a decrease in cerebral oxygen consumption, it is possible that the basal anesthetic state might affect the cere- TABLE 4. Physiological Measurements During Topical Application of Methylene Blue, Glibenclamide, Dimethyl Sulfoxide, and Subsequent Nicorandil MAP, HR, bpm BT, C PH Paco 2, Pao 2, Na, meq/l K, meq/l NC MB+NC GC+NC DMSO+NC 117± ± ±6 127± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± MAP indicates mean arterial blood pressure; HR, heart rate; BT, body temperature; NC, topical application of 10" 5 mol/l nicorandil solution; MB+NC, topical application of 10~ 5 mol/l methylene blue and subsequent 10~ 5 mol/l nicorandil; GC+NC, topical application of 10~ 5 mol/l glibenclamide and subsequent 10~ 5 mol/l nicorandil; and DMSO+NC; topical application of 10~ 5 mol/l dimethyl sulfoxide and subsequent 10~ 5 mol/l nicorandil. Values are mean±sem.

5 1648 Stroke Vol 25, No 8 August 1994 TABLE 6. Effects of Cromakalim on Pial Arterioles and Venules Concentration, mol/l concentration (-log M) concentration (-log M) FIG 3. Line graphs show concentration-related effect of nitroglycerin on pial arteriolar (A) or venular (B) diameter. Nitroglycerin induced a concentration-dependent increase in small pial arteriolar and large and small venular diameters (P<.05). Values are mean±sem. brovascular tone in both arterioles and venules. Although pentobarbital has been shown not to affect cerebrovascular tone in isolated dog experiments, 20 we cannot exclude the possibility that nicorandil may work equivalently on pial arterioles and venules directly. Activation of guanylate cyclase and the consequent increase in the amount of cyclic GMP are considered to be the mechanism of action of nitrates. Nicorandil also relaxes vascular smooth muscles by stimulating the synthesis of cyclic GMP. 21 Since methylene blue blocks guanylate cyclase, we applied methylene blue to determine whether guanylate cyclase interferes with the nicorandil-induced pial vasodilation and found that methylene blue did not affect pial vascular diameter. Other investigators have reported that methylene blue per se did not change baseline arteriolar diameter and that methylene blue did not affect topically applied nitrovasodilators; therefore, guanylate cyclase could not be affected by topical methylene blue These data also indicate that methylene blue cannot penetrate into smooth muscle in small arterioles but rather generates a hydroxyl radical that inhibits EDRF. Contrary to these reports, the present result demonstrated that topically applied nitroglycerin-induced pial arteriolar or venular o ' ~'"" " " o* FIG 4. Bar graphs show effects of methylene blue or glibenclamide on nitroglycerin-induced dilation of arterioles (A) or venules (B). Methylene blue inhibited small pial arteriolar and large and small venular dilation. Values are mean±sem. *P<.005, tp< Black bars indicate 10" 3 mol/l nitroglycerin; hatched bars, 10" 5 mol/l methylene blue plus 10" 3 mol/l nitroglycerin; and shaded bars, 10~ 5 mol/l glibenclamide plus 10~ 3 mol/l nitroglycerin. 10" t ± ± t ±16.12* ±21.34t : ± ± ± ± : ± ± : ± 12.74* ± : : ± ±8.70t ±7.61t Values are mean±sem. Cromakalim induced dilations in large and small pial arterioles and venules. *P<.05, tp<-005, P<.001, P<.0005 compared with corresponding control. dilation was attenuated in the presence of methylene blue. Nicorandil has been established to dilate blood vessels without activating EDRF 24 and to release a second messenger, distinct from EDRF, that then activated guanylate cyclase and subsequently elevated cyclic GMP. 25 Furthermore, topically applied methylene blue was reported to inhibit production of cyclic GMP through blockade of guanylate cyclase. 26 We interpret these results to mean that methylene blue blocked the vasodilative response in pial vessels elicited by nicorandil and that the mechanism of the vasoactive effects of nicorandil may in part depend on cyclic GMPmediated second messenger systems. ATP-sensitive K + channels are involved in a dilative action of resistance arteries in response to vasodilative agents. 27 Using dispersed cerebral arterial smooth muscle cells, Bonnet et al 28 demonstrated that outward K + current resulted in cerebrovascular dilation. In in vitro 45" p l^rg< arterioles 40- a 35 ' m sma 1 anerioles 1 30" ea c 20" 5-0- # / T/ '' / i k concentration (-logm) ;e venules concentration (-logm) FIG 5. Line graphs show concentration-related effect of cromakalim on pial arteriolar (A) or venular (B) diameter. Cromakalim induced a concentration-dependent increase in large and small pial arteriolar and venular diameters (P<.05). Values are mean±sem.

6 Ishiyama et al Nicorandil and Pial Microcirculation " I 15 " I 30- ID- S' 0 FIG 6. Bar graphs show effects of methylene blue or glibenclamide on cromakalim-induced dilation of pial arterioles (A) or venules (B). Glibenclamide inhibited large and small pial arteriolarand large venular dilation. Values are mean±sem. *P<.005, tp< Black bars indicate 10~ 3 mol/l cromakalim; hatched bars, 10~ 5 mol/l methylene blue plus 10~ 3 mol/l cromakalim; and shaded bars, 10~ 5 mol/l glibenclamide plus 10~ 3 mol/l cromakalim. experiments, cromakalim, an ATP-sensitive K + channel opener, 29 has been reported to cause vasodilation in basilar and middle cerebral arteries. 14 We used glibenclamide, an ATP-sensitive K + channel blocker, 2730 to examine the mechanism of small arteriolar dilation caused by nicorandil. Since glibenclamide could not dissolve in acsf, DMSO was used as the vehicle for glibenclamide. As reported by other investigators, 3134 our results detected that topically applied glibenclamide and DMSO revealed no vascular effect on pial arterioles or venules. ATP-sensitive K + channels are reported to be present only in very low concentrations in small cerebral arterial segments. 27 Indeed, in rat pial arterioles, cromakalim and pinacidil do not seem to exhibit vasodilative effects Although it was reported that nicorandil-induced vasodilation of isolated cerebral artery was not affected with the presence of glibenclamide, 12 our present in vivo experiments showed that the nicorandil-induced vasodilations were attenuated by glibenclamide in small but not in large arterioles and venules. Recent in vivo studies suggest that ATP-sensitive K + channels are present in cerebral arterioles In our additional experiments, we confirmed that topical application of glibenclamide could effectively block the cromakalim-induced pial arteriolar and venular dilation. The results support the contention that the inhibitory effects of glibenclamide are specific to ATPsensitive K + channels in our experimental model. We interpret these data to suggest that nicorandil-induced small pial arteriolar or venular dilation is at least in part induced by an ATP-sensitive K + channel activator. Although there might be some different response to nicorandil between topical application and intravenous administration, the present in vivo experiments may suggest some clinical applications. Intravenous infusion of nicorandil (10 /xg/kg per minute) has been reported to reverse experimental vasospasm, 38 and prior treatment with nicorandil attenuated the contractile responses to prostaglandin F^39 and endothelin. 24 These results suggest that nicorandil dilates spastic cerebral vessels. Cerebral vasospasm after experimental subarachnoid hemorrhage accompanies marked membrane depolarization in cerebral arterial smooth muscle cells. 40 Membrane hyperpolarization elicited by a K + channel opener may antagonize cerebral vasospasm after subarachnoid hemorrhage. Furthermore, ATPsensitive K + channels in the brain are known to be associated with the release of y-aminobutyric acid (GABA). 41 Inhibition of GABA-ergic inhibitory control after transient brain ischemia leads to delayed neuronal death, and ATP-sensitive K + channels are proposed to play a significant role in postischemic delayed neuronal death. 42 Since downregulation of ATP-sensitive K + channels induced by brain ischemia exerts adverse effects on the brain, 42 nicorandil may be beneficial for cerebral vasospasm after subarachnoid hemorrhage and may protect the brain from ischemia. In conclusion, topical application of nicorandil dilates both pial arteriolar and venular vessels in a concentration-dependent manner. Nicorandil is a cerebrovascular dilator. The mechanisms of vasodilative response caused by nicorandil are involved in cyclic GMP-mediated second messenger systems, and regulations of K + flux also play an important role in the action of nicorandil on small pial arteriolar and venular dilation. Acknowledgment This work was supported by Grant-in-Aid for Scientific Research No (Ministry of Education, Science and Culture, Japan). References 1. Sakai K. Nicorandil. Animal pharmacology. Am J Cardiol. 1989;63: 2J-10J. 2. Aizawa T, Ogasawara K, Kato K. Effects of nicorandil on coronary circulation in patients with ischemic heart disease: comparison with nitroglycerin. / Cardiovasc Pharmacol. 1987;10:S123-S Gross G, Pieper G, Farber NE, Warltier D, Hardman H. Effects of nicorandil on coronary circulation and myocardial ischemia. Am J Cardiol. 1989;63:11J-17J. 4. Faraci FM. Role of nitric oxide in regulation of basilar artery tone in vivo. Am J Physiol. 1990;259:H1216-H Faraci FM, Mayhan WG, Heistad DD. Responses of rat basilar artery to acetylcholine and platelet products in vivo. Stroke. 1991; 22: Taira N. Similarity and dissimilarity in the mode and mechanism of action between nicorandil and classical nitrates: an overview. J Cardiovasc Pharmacol. 1987;1O:S1-S9. 7. Lagerkranser M. Effects of nitroglycerin on intracranial pressure and cerebral blood flow. Ada Anesthesiol Scand. 1992;36: Taira N. Nicorandil as a hybrid between nitrates and potassium channel activators. Am J Cardiol. 1989;63:18J-24J. 9. Yanagisawa T, Taira N. Effect of 2-nicotinamidethyl nitrate (SG75) on the membrane potential of left atrial muscle fibers of the dog: increase in potassium conductance. Naunyn Schmiedebergs Arch Pharmacol. 1980;312: Brayden JE, Nelson MT. Regulation of arterial tone by activation of calcium-dependent potassium channels. Science. 1992;256: Zhang H, Stockbridge N, Weir B. Effects of pinacidil, cromakalim, and nicorandil on potassium currents of rat basilar artery smooth muscle. In: Moreland RS, ed. Regulation of Smooth Muscle Contraction. New York, NY: Plenum Publishing Corp; 1991: Zhang H, Stockbridge N, Weir B, Vollrath B, Cook D. Vasodilatation of canine cerebral arteries by nicorandil, pinacidil and lemakalim. Gen Pharmacol. 1992;23: Wei EP, Moskowitz MA, Boccalini P, Kontos HA. Calcitonin gene-related peptide mediates nitroglycerin and sodium nitroprusside-induced vasodilation in feline cerebral arterioles. Circ Res. 1992;70: Nagao T, Sadosima S, Kamouchi M, Fujishima M. Cromakalim dilates rat cerebral arteries in vitro. Stroke. 1991;22: Parsons AA, Ksoll E, Mackert JRL, Schilling L, Wahl M. Comparison of cromakalim-induced relaxation of potassium precontracted rabbit, cat, and rat isolated cerebral arteries. Arch Pharmacol. 1991;343: Masuzawa K, Asano M, Matsuda T, Imaizumi Y, Watanabe M. Possible involvement of ATP-sensitive K + channels in the relaxant

7 1650 Stroke Vol 25, No 8 August 1994 response of dog middle cerebral artery to cromakalim. / Pharmacol Exp Ther. 1990;255: Fujii K, Heistad DD, Faraci FM. Role of the basilar artery in regulation of blood flow to the brain stem in rats. Stroke. 1991;22: Burrows ME, Johnson PC. Diameter, wall tension, and flow in mesenteric arterioles during autoregulation. Am J PhysioL 1981; 241:H829-H Kinoshita M, Sakai K. Pharmacology and therapeutic effects of nicorandil. Cardiovasc Drugs Ther. 1990;4: Hatano Y, Nakamura K, Moriyama S, Mori K, Toda N. The contractile responses of isolated dog cerebral and extracerebral arteries to oxybarbiturates and thiobarbiturates. Anesthesiology. 1989;71: Sumitomo K, Domae M, Yamanaka K, Nakao K, Hashimoto T, Kitamura K, Kuriyama H. Actions of nicorandil on vascular smooth muscle. J Cardiovasc Pharmacol 1987;10.S66-S Busija DW, Leffler CW, Wagerle LC. Mono-L-argininecontaining compounds dilate piglet pial arterioles via an endothelium-derived relaxing factor-like substance. Circ Res. 1990;67: Marshall JJ, Wei EP, Kontos HA. Independent blockade of vasodilation from acetylcholine and nitric oxide. Am J PhysioL 1988; 255:H847-H Sakakura T. Vasodilative effect of nicorandil on canine basilar arteries. Mie Med J. 1990;40: Greenberg SG, Cantor E, Ho E, Walega M. Comparison of nicorandil-induced relaxation, elevation of cyclic guanosine monophosphate and stimulation of guanylate cyclase with organic nitrate esters./pharmacol Exp Ther : Watanabe M, Rosenblum WI, Nelson GH. In vivo effect of methylene blue on endothelium-dependent and endotheliumindependent dilations of brain microvessels in mice. Circ Res. 1988; 62: Brayden JE. Hyperpolarization and relaxation of resistance arteries in response to adenosine diphosphate distribution and mechanism of action. Circ Res. 1991;69: Bonnet P, Rusch NJ, Harder DR. Characterization of outward K + current in freshly dispersed cerebral arterial muscle cells. Pflugcn Arch. 1991;418: In this interesting article the authors examined the effects of nicorandil on the pial microcirculation of the dog. They examined responses in both arteries and veins, and they demonstrated that topical application of nicorandil dilates both pial arteries and veins in a dose-dependent manner. The authors also demonstrated that, since nicorandil-induced dilation of both pial arteries and veins was inhibited by methylene blue and attenuated by glibenclamide, the mechanism of action of nicorandil is most likely due to both cyclic GMP-mediated vascular smooth muscle dilation and the regulation of K + flux. The new and interesting aspect of this manuscript is that this is the first time that topical applications of nicorandil have been examined Editorial Comment 29. Standen NB, Quayle JM, Davics NW, Brayden JE, Huang Y, Nelson MT. Hyperpolarizing vasodilators activate ATP-sensitive K* channels in arterial smooth muscle. Science. 1989^245: Brayden JE. Membrane hyperpolarization is a mechanism of endothelium-dependent cerebral vasodilation. Am J PhysioL 1990; 259:H668-H Yang ST, Mayhan WG, Faraci FM, Heistad DD. Endotheliumdependent responses of cerebral blood vessels during chronic hypertension. Hypertension. 1991;17: Busija DW, Leffler CW. Effects of phorbol esters on pial arteriolar diameter and brain production of prostanoids in piglets. Circ Res. 1991;69: Murray MA, Faraci FM, Heistad DD. Signal transduction pathways in constriction of the basilar artery in vivo. Hypertension. 1992;19: McCarron JG, Quayle JM, Halpem W, Nelson MT. Cromakalim and pinacidil dilate small mesenteric arteries but not small cerebral arteries. Am J PhysioL 1991;261:H287-H McPherson GA, Stork AP. The resistance of some rat cerebral arteries to the vasorelaxant effect of cromakalim and K + channel openers. Br J Pharmacol 1992;105: Lee WS, Kwon YK, Yu SS, Rhim BY, Hong KW. Disturbances in autoregulatory responses of rat pial arteries by sulfonylureas. Life ScL 1993;52: Mayhan WG, Faraci FM. Responses of cerebral arterioles in diabetic rats to activation of ATP-sensitive potassium channels. Am J PhysioL 1993;265:H152-H Ken K, Naotake S, Hajime N. Effects of SG-75 (nicorandil) on experimental cerebral vasospasm. Jpn Pharmacol Ther. 1987;15: Yamada K, Ohta T, Shimizu K, Yasuda A. Effect of nicorandil on cerebral vasospasm. Jpn J Clin Pharmacol Ther. 1989;20: Harder DR, Dernbach P, Waters A. Possible cellular mechanism for cerebral vasospasm after experimental subarachnoid hemorrhage in the dog. J Clin Invest. 1987;80: Amoroso S, Schmid AH, Fosset M, Lazdunski M. Glucose, antidiabetic sulfonylureas and neurotransmitter release: role of ATPsensitive K + channels. Science, 199O;247: Mourre C, Smith ML, SiesjS BK, Lazdunski M. Brain ischemia alters the density of binding sites for glibenclamide, a specific blocker of ATP-sensitive K + channels. Brain Res. 1990^26: on pial vessels in in vivo experiments in which both arteries and veins have been examined simultaneously in the same experimental preparation by the same investigators. In addition, the authors demonstrated that the mechanism involved in nicorandil vasodilation via cyclic GMP and/or in part the regulation of K + flux appear similar in both arteries and veins. Little work in the past has been done with pial veins, and this article makes an interesting contribution along those lines. Richard J. Traystman, PhD AnesthesiologylCritical Care Medicine The Johns Hopkins Medical University Baltimore, Md