Isoflurane-induced Dilation of Porcine Coronary Arterioles Is Mediated by ATP-sensitive Potassium Channels 

Forty-four porcine hearts were obtained from a slaughterhouse, immediately placed in iced Kreb's buffer, and transported to the laboratory. The composition of the Kreb's buffer was 99.5 mM NaCl, 4.7 mM KCl, 1.2 mM KH (²) PO⁴, 1.2 mM MgSO⁴, 25 mM NaHCO³, 4.9 mM Na pyruvate, 5.4 mM Na fumarate, 1.1 mM CaCl², and 11.5 mM glucose.

A subepicardial microvessel with an internal diameter of 173 +/- 46 (SD)[micro sign]m and a length of 1 - 2 mm was dissected carefully from the distribution of left anterior descending coronary artery. The vessel was placed in a lucite chamber containing cold Kreb's buffer and cannulated at either end with a glass micropipette, which was secured with fine monofilament sutures. One cannula was connected to a pressure reservoir system that maintained intraluminal pressure (measured through a side arm) at 40 mmHg. The other cannula was closed, thus preventing flow through the vessel.

(Figure 1) is a diagram of the microvessel preparation. The chamber containing the coronary arteriole was seated on a stage of an inverted microscope coupled to a video camera. The lumen diameter was determined using an optical density video-detection system. [10]The chamber was supplied with Kreb's buffer via either of two circuits: Circuit 1 supplied isoflurane-free buffer and circuit 2 supplied buffer containing isoflurane. In both circuits, the buffer passed through an oxygenator, which was provided with 95% oxygen and 5% carbon dioxide (thus producing values for P^O2, carbon dioxide pressure [P^CO(2)], and pH in the buffer of 542 +/- 54 mmHg, 40 +/- 4 mmHg, and 7.41 +/- 0.05, respectively). However, only in circuit 2 was the gas supplied to the oxygenator passed through a calibrated vaporizer providing isoflurane. The buffer temperature in the chamber was maintained at 38 [degree sign]C. The circuits providing buffer to the chamber were unidirectional; thus buffer leaving the chamber was discarded. This avoided the possibility that metabolites released from the vessel would return to affect vascular tone.

Initially the cannulated vessel underwent a stabilization period of at least 20 min when it was exposed to Kreb's buffer. Then, in most studies, acetylcholine (10^-8to 10^-6M) was added to the buffer to constrict the vessel to 60 - 70% of its resting diameter. Once an adequate concentration of acetylcholine was determined, it was held constant throughout each study. Acetylcholine was used to provide vasomotor tone, because in contrast to its endothelium-dependent vasorelaxing effect in coronary vessels from most species, acetylcholine produces sustained vasoconstriction in coronary vessels from swine. [20]In series 5, the vasodilating effects of isoflurane were assessed in the same vessels preconstricted sequentially with acetylcholine and U46619.

Series 1 (n = 14): Abrupt versus Gradual Administration of Isoflurane. In nine vessels, buffer (and acetylcholine) was initially supplied through circuit 1. After adequate time for stabilization of the preconstricted vessel (at least 30 min), the baseline diameter was determined and a sample of buffer was obtained for gas analysis. The vessel was exposed to isoflurane according to two protocols: abrupt (ISO-A) and gradual (ISO-G). In ISO-A, the vessel was switched abruptly from isoflurane-free buffer to buffer containing 2% isoflurane; that is, from circuit 1 to circuit 2. The buffer in circuit 2 was recirculated for at least 30 min through the oxygenator to ensure complete equilibration before it was supplied to the vessel. After vessel diameter was stable for 5 min during ISO-A, the value was recorded and samples of buffer were obtained for analysis of gas composition and isoflurane concentration. Then the vessel was returned to circuit 1, which contained buffer with the preconstricting dose of acetylcholine but free of isoflurane. At least 20 min was allowed for recovery. During this period, the buffer in circuit 2 was recirculated three times through the oxygenator with the vaporizer setting at 0% to remove isoflurane. The vessel was switched to circuit 2 and a second set of baseline measurements was obtained. Then the ISO-G protocol was performed, whereby the vaporizer setting was increased stepwise (in 0.5% increments) to 2% during a period of 40 min. Each setting was maintained for 10 min before the vascular diameter was determined and samples of buffer were obtained for analysis. The total elapsed time for the ISO-A and ISO-G protocols was 2 to 2.5 h.

Two approaches were used to address the potential time-dependent effects on vessel reactivity. First, the order of the ISO-A and ISO-G protocols was randomized. Second, an additional set of five vessels, prepared as described before, was subjected to duplicate exposures to isoflurane using the ISO-A protocol.

Series 2 (n = 13): K^ATPChannels and Isoflurane-induced Coronary Vasodilation. Studies were conducted to determine the effect of the K^ATPchannel inhibitor glibenclamide on isoflurane-induced coronary vasodilation. After obtaining an initial baseline measurement for vascular diameter, isoflurane in a concentration of 1.5%(approximating the 1 minimum alveolar concentration [MAC]) was added to the bathing solution using the abrupt protocol described before, and the measurements were repeated (control response). The vessel was then returned to isoflurane-free buffer. After at least 20 min for recovery from the effects of isoflurane, glibenclamide (10 (^-5) M) was added to the buffer and a value for vascular diameter was obtained. While glibenclamide was continued, a second administration of isoflurane was performed. Then the glibenclamide stopped, and isoflurane was administered for a third time. The coronary vasodilating effects of cromakalim (10^-6M), a K^ATPchannel opener, [12]and SNP (10^-5M), a K^ATPchannel-independent vasodilator, [12]were assessed under each condition (before, during, and after glibenclamide) to verify the efficacy of glibenclamide. Measurements of isoflurane concentration in the buffer were obtained before and during, but not after, glibenclamide. The order of isoflurane, cromakalim, and SNP was randomized. Each condition in series 2 required approximately 1 h (about 20 min per dilator); thus the total duration of series 2 was approximately 3 h.

Series 3 (n = 6): Adenosine Receptors and Isoflurane-induced Coronary Vasodilation. The role the coronary adenosine receptors in isoflurane-induced coronary vasodilation was evaluated using 8-PT to block the adenosine receptors. The studies followed the protocol described for series 2, except no post-blockade responses were assessed. The effects of isoflurane (1.5%) were evaluated only under control conditions and during 8-PT (10^-5M). Responses to adenosine (10^-5M) and SNP (10^-5M) were used to evaluate the efficacy of the treatment with 8-PT.

Series 4 (n = 5): Validation Studies Demonstrating the Functional Integrity of Vascular Smooth Muscle and Endothelium. Studies were performed to verify that both the vascular endothelium and the vascular smooth muscle remained functionally intact in the isolated coronary arterioles. Vascular endothelial reactivity was evaluated using the endothelium-dependent vasodilator adenosine diphosphate (5 x 10^-6M, 1.5 x 10^-5M, and 5 x 10^-5M), whereas vascular smooth muscle reactivity was evaluated using the endothelium-independent vasodilator SNP (10^-5M, 3 x 10^-5M, and 10^-4M).

Series 5 (n = 6): Isoflurane-induced Coronary Vasodilation: Acetylcholine versus U46619 Preconstriction. Studies were conducted to determine if isoflurane-induced coronary vasodilation is different if U46619 is substituted for acetylcholine as the preconstrictor. Responses to 0.75% and 1.5% isoflurane were assessed in the same vessel preconstricted sequentially with acetylcholine and U46619. The protocol in these studies followed that described for series 2. The order of the preconstrictors was randomized. Care was taken to ensure that the initial preconstrictor was washed from the vessel before the second preconstrictor was applied.

The concentration of isoflurane in the buffer solution was determined in series 1–3 using a modification of the gas chromatographic technique of Yamamura et al. [21]The data are presented in both units of milligrams per 100 ml and millimolar to facilitate comparisons with previous studies. [17–19] 

Statistical testing was done using an analysis of variance for repeated measurements in conjunction with the Student-Newman-Keuls test, or with the Student's t test for paired samples, as appropriate. [22]All values are mean +/- SD. Probability values <0.05 were considered significant throughout the study.

(Figure 2) shows the effects of isoflurane on vascular diameter of coronary arterioles. In the ISO-G protocol, vaporizer settings >0.5%(i.e., 1%, 1.5%, and 2%) caused increases in vascular diameter. The increases in vascular diameter at 2% isoflurane were more than those at 1% isoflurane, whereas those at 1.5% isoflurane did not differ significantly from those at 1% or 2% isoflurane. The increases in vascular diameter caused by 2% isoflurane using the ISO-A protocol were nearly four times those using the ISO-G protocol. This difference in response was evident regardless of the sequence of the protocols.

The concentration for isoflurane in the buffer solution varied directly with the percentage provided by the vaporizer (Table 1). The anesthetic concentrations in the buffer at 2% were similar during the gradual and abrupt administrations (Table 1).

Duplicate abrupt administrations of 2% isoflurane caused increases in vascular diameter that did not differ (83 +/- 26% and 92 +/- 19%, respectively).

(Table 2) presents the actually measured changes in vascular diameter during isoflurane, SNP, and cromakalim before, during, and after glibenclamide. It shows (1) that glibenclamide itself did not affect vascular diameter;(2) that the vasodilators, including isoflurane, caused responses that were reversible (i.e., vascular diameter returned to baseline); and (3) that glibenclamide blunted considerably (but reversibly) the increases in vascular diameter by isoflurane and cromakalim, but it had no effect on increases caused by SNP. The buffer concentrations for isoflurane were 14.7 +/- 0.8 mg/100 ml (0.80 +/- 0.04 mM) and 14.9 +/- 0.8 mg/100 ml (0.81 +/- 0.04 mM) before and during glibenclamide, respectively.

(Figure 3) shows that 8-PT blunted the increases in vascular diameter by adenosine but it had no effect on those by isoflurane and SNP. 8-PT had no effect on baseline vascular diameter (123 +/- 14 vs. 119 +/- 12 [micro sign]m). Figure 4shows that both adenosine diphosphate and SNP caused concentration-dependent increases in vascular diameter. Figure 5shows that isoflurane caused equivalent concentration-dependent increases in vascular diameter during preconstriction with acetylcholine and U46619. The findings were not affected by the order in which the agents were assessed. The preisoflurane vascular diameters were similar during acetylcholine and U46619, 150 +/- 23 [micro sign]m and 152 +/- 23 [micro sign]m, respectively.

The study yielded several main findings. (1) Isoflurane caused concentration-dependent dilation of porcine coronary arterioles in vitro. (2) The coronary dilating effects of isoflurane were less during a gradual administration than during an abrupt administration. (3) Glibenclamide inhibited isoflurane-induced dilation of coronary arterioles, whereas 8-PT had no effect. (4) Isoflurane caused an equivalent dilation of coronary arterioles that were preconstricted with acetylcholine or U46619.

We used an in vitro microvessel technique developed by Halpern et al. [10]This technique is well established and has been used to study vessels obtained from various vascular beds and species. [17–20,23–26]It had several advantages for our study. (1) The vasomotor responses of coronary arterioles (i.e., resistance vessels) could be studied selectively. (2) The coronary vasomotor responses were independent of the associated changes in myocardial oxygen demand and extravascular compressive forces that are inherent to in vivo models. (3) Because the vessels were studied in a nonperfused state, their responses were assessed in the absence of shear-induced release of nitric oxide. The main disadvantage of the microvessel technique was that, like any in vitro technique, it had artificial and nonphysiologic features, including the use of a drug to establish “background” vascular tone and delivery of drugs via the adventitia. The increases in vascular diameter by adenosine diphosphate and SNP attested to the functional integrity of the vascular endothelium and underlying smooth muscle of the coronary arteriole preparations.

The fidelity of the isoflurane delivery system was demonstrated by buffer concentrations that remained proportional to the percentage supplied by the vaporizer and that did not vary when evaluated twice under different conditions in the same preparation (e.g., during the abrupt and gradual protocols). Our buffer concentrations for isoflurane were in close agreement with those reported by previous investigators using a similar experimental approach. [17,18] 

The coronary microvessels that we studied had diameters in the range of medium coronary arterioles. [27]In vivo, with coronary vasomotor tone intact, approximately 20 - 25% of total coronary vascular resistance is proximal to vessels of this size. [27]Further, studies in vivo have shown that the effects of vasodilators (e.g., adenosine and dipyridamole) may differ in these vessels compared with coronary arterioles that are smaller or larger. [28]Thus it would not be appropriate to extrapolate our findings to other segments of the coronary microcirculation.

We observed that the increases in arteriolar diameter during gradual administrations of isoflurane were less than those during abrupt administrations, which suggested that the vessels adapted over time to the relaxing effects of isoflurane. This ability of the coronary arterioles likely explains the blunted increases in coronary blood flow during gradual administrations of isoflurane [11]and the recovery of coronary vascular tone during prolonged administrations of isoflurane [29]observed in vivo.

The specific mechanism responsible for the ability of the coronary arterioles to adapt to isoflurane is unknown, but it may involve downregulation of specific receptors or of the second messenger system (e.g., K^ATPchannels). In theory, one approach to assess the role of the K^ATPchannels in this phenomenon would be to repeat the abrupt versus gradual comparison in the presence of glibenclamide. However, because glibenclamide nearly abolished the vasodilating effect of the abrupt administrations of isoflurane (Table 2), it would be difficult, if not impossible, to detect a blunting of this effect during the gradual administrations. Further studies using sophisticated molecular approaches to study signal-transduction pathways would appear necessary to clarify the mechanism(s) underlying the ability of the coronary arterioles to adapt to isoflurane.

The validity of our findings in series 1 depends on the duplicate responses in the same vessel being independent of one another; that is, it was necessary to rule out that the blunted response during the gradual administration was due to tachyphylaxis after the abrupt administration. Two lines of evidence suggest that this was so. First, the gradual administrations had less of a vasodilating action even when they were placed before the abrupt administrations. Second, when two abrupt administrations were performed in succession, the vasodilating effects were similar.

Our findings in series 1 - 3 conflicted with those of Park et al. [17–19]that showed that isoflurane, in a similar range of concentrations, caused constriction of isolated coronary arterioles obtained from normal rabbit or rat hearts, or from the normal region of porcine hearts with a chronic coronary occlusion. To determine whether this difference in results occurred because we used acetylcholine whereas Park et al. used U46619 for preconstriction, we did additional studies that compared the vasodilating effects of isoflurane in the same microvessels preconstricted sequentially with acetylcholine and U46619. The similar findings in the presence of acetylcholine and U46619 precluded a different preconstrictor as an explanation for the apparent discrepancy between our current findings and those of Park et al.

Why the coronary microvessels in the studies by Park et al. did not dilate in response to isoflurane remains uncertain. One possible explanation pertains to the protocol they used to administer isoflurane. Each coronary microvessel was exposed for 10 min to each of five increasing concentrations of isoflurane (range, 0.3% to 3%) without interruption for recovery. As shown in the current study, such a gradual and prolonged exposure to isoflurane would tend to blunt its vasodilating effect.

Abrupt administrations of isoflurane were used in the glibenclamide, 8-PT, and acetylcholine versus U46613 studies (series 2, 3, and 5) to minimize the possibility that time-dependent factors (i.e., vascular adaptation and vessel deterioration) would reduce the responsiveness of the coronary arterioles. This was confirmed by the completely restored vasodilating responses to both isoflurane and cromakalim after termination of glibenclamide and by the equivalent responses to the multiple administrations of SNP (Table 2). Another advantage of the abrupt administrations in series 2 and 3 was that they maximized the vasodilating effects of isoflurane, which favored uncovering effects of the pharmacologic antagonists.

Glibenclamide attenuated the increases in coronary arteriolar diameter during isoflurane, implying a significant role for the K^ATPchannels in the response. These findings are consistent with recent open-chest studies conducted in dogs [12]and swine [13]that indicated that glibenclamide blunted the increases in coronary blood flow during intracoronary administrations of isoflurane and of halothane and enflurane. Recent evidence suggests that K^ATPchannels are present in vascular smooth muscle cells and in vascular endothelial cells. [26]Opening of K (^ATP) channels in the vascular smooth muscle cells shifts the membrane potential closer to the K⁺reversal potential. Hyperpolarization then inhibits calcium entry through voltage-dependent calcium channels, leading to reduced vasomotor tone. Opening of the endothelial K^ATPchannels can reduce vascular smooth muscle tone by either production and release of second messengers, such as nitric oxide, or by direct transduction of electrical potentials. Our current findings do not distinguish between the contribution of the K^ATPchannels in vascular smooth muscle and in the endothelial cells to isoflurane-induced coronary vascular relaxation.

Kersten et al. [30]reported that the selective adenosine type 1 (A¹) receptor antagonist (8-cyclopentyl-1,3,dipropyl-xanthine), like glibenclamide in their previous studies, [31,32]inhibited isoflurane-induced cardioprotection in stunned myocardium of dogs. These findings suggested that the myocardial protective effect of isoflurane may involve sequential activation of the A¹receptor and the K^ATPchannels in cardiac cells. We observed no effect of the nonselective adenosine receptor antagonist 8-PT on isoflurane-induced coronary vasodilation, which suggests that isoflurane may act via a different pathway in the coronary vasculature than it does in the cardiomyocytes. The mechanism underlying isoflurane-induced activation of the K^ATPchannels in the coronary arterioles remains to be clarified.

The efficacy of our doses for glibenclamide and 8-PT was evident from their ability of obtund the increases in vascular diameter caused by cromakalim and adenosine, respectively, although they had no effect on the increases in vascular diameter caused by SNP.

In conclusion, our results show (1) that isoflurane causes dilation of porcine coronary arterioles preconstricted with acetylcholine or U46619, which is mediated by the K^ATPchannels but not by the adenosine receptors, and (2) that these vessels exhibit time-dependent adaptation to the vasorelaxing effects of isoflurane.

The authors thank Derrick L. Harris, B.S., and Arun Kumar, B.A., for expert technical assistance.