Vascular endothelium plays an important role in the regulation of vascular tone. Volatile anesthetics have been shown to attenuate endothelium-mediated relaxation in conductance arteries, such as aorta. However, significant differences in volatile anesthetic pharmacology between these large vessels and the small vessels that regulate systemic vascular resistance and blood flow have been documented, yet little is known about volatile anesthetic action on endothelial function in resistance arteries. Furthermore, endothelium-dependent relaxation mediated by factors other than endothelium-derived relaxing factor (EDRF) has recently been recognized, and there is no information available regarding volatile anesthetic action on non-EDRF-mediated endothelium-dependent relaxation.
Employing isometric tension recording and microelectrode methods, the authors first characterized the endothelium-dependent relaxing and hyperpolarizing actions of acetylcholine (ACh) in rabbit small mesenteric arteries, and tested the sensitivities of these actions to EDRF pathway inhibitors and K+ channel blockers. They then examined the effects of the volatile anesthetics isoflurane, enflurane, and sevoflurane on ACh-induced endothelium-dependent relaxation that was sensitive to EDRF inhibitors and that which was resistant to the EDRF inhibitors but sensitive to blockers of ACh-induced hyperpolarization. The effects of the volatile anesthetics on endothelium-independent sodium nitroprusside (SNP)-induced relaxation were also studied.
Acetylcholine concentration-dependently caused both endothelium-dependent relaxation and hyperpolarization of vascular smooth muscle. The relaxation elicited by low concentrations of ACh (< or = 0.1 microM) was almost completely abolished by the EDRF inhibitors NG-nitro-L-arginine (LNNA), oxyhemoglobin (HbO2), and methylene blue (MB). The relaxation elicited by higher concentrations of ACh (> or = 0.3 microM) was only attenuated by the EDRF inhibitors. The remaining relaxation, as well as the ACh-induced hyperpolarization that was also resistant to EDRF inhibitors, were both specifically blocked by tetraethylammonium (TEA > or = 10 mM). Sodium nitroprusside, a NO donor, produced dose-dependent relaxation, but not hyperpolarization, in the endothelium-denuded (E[-]) strips, and the relaxation was inhibited by MB and HbO2, but not TEA (> or = 10 mM). One MAC isoflurane, enflurane, and sevoflurane inhibited both ACh relaxation that was sensitive to the EDRF inhibitors and the ACh relaxation resistant to the EDRF inhibitors and sensitive to TEA, but not SNP relaxation (in the E[-] strips). An additional finding was that the anesthetics all significantly inhibited norepinephrine (NE) contractions in the presence and absence of the endothelium or after exposure to the EDRF inhibitors.
The results confirm that ACh has a hyperpolarizing action in rabbit small mesenteric resistance arteries that is independent of EDRF inhibitors but blocked by the K+ channel blocker TEA. The ACh relaxation in these resistance arteries thus appears to consist of distinct EDRF-mediated and hyperpolarization-mediated components. Isoflurane, enflurane, and sevoflurane inhibited both components of the ACh-induced relaxation in these small arteries, indicating a more global depression of endothelial function or ACh signaling in endothelial cells, rather than a specific effect on the EDRF pathway. All these anesthetics exerted vasodilating action in the presence of NE, the primary neurotransmitter of the sympathetic nervous system, which plays a major role in maintaining vasomotor tone in vivo. This strongly indicates that the vasodilating action of these anesthetics probably dominates over their inhibitory action on the EDRF pathway and, presumably, contributes to their known hypotensive effects in vivo. Finally, the vasodilating action of these anesthetics is, at least in part, independent from endothelium.
Key words: Anesthetics, volatile: enflurane; isoflurane; sevoflurane. Endothelium: endothelium-derived hyperpolarizing factor; endothelium-derived relaxing factor; nitric oxide. Muscle, smooth: vascular.
VASCULAR endothelium appears to play a crucial role in the regulation of vascular smooth muscle tone by releasing vasoactive substances, such as endothelium-derived relaxing factor (EDRF), prostacyclin, or endothelin, in response to various stimuli. It has been suggested that volatile anesthetics, including halothane, enflurane, isoflurane, and sevoflurane, may significantly affect such endothelial function. [1–4] Muldoon et al. first reported that halothane attenuated endothelium-mediated vasorelaxation induced by acetylcholine (ACh) or bradykinin in rabbit aorta and canine femoral and carotid arteries, while endothelium-independent relaxation induced by nitroglycerin (NTG) was not affected.  These findings have subsequently been confirmed by Uggeri et al., who studied enflurane and isoflurane, in addition to halothane, and reported that all three anesthetics attenuated both receptor- and nonreceptor-mediated endothelium-dependent relaxations in rat aorta without affecting endothelium-independent relaxation induced by sodium nitroprusside (SNP).  Similarly, Toda et al. recently reported that halothane and isoflurane inhibited ACh-induced, but not NTG- or SNP-induced, relaxation in rat aorta,  and the lack of effect of halothane on SNP-induced relaxation was more recently confirmed by Blaise et al. in rabbit aorta.  Although these findings have led to the belief that halothane and other volatile anesthetics may inhibit the synthesis, release, or both of EDRF or inactivate EDRF, Hart et al. have recently demonstrated that halothane inhibited the relaxation and cyclic GMP production induced by nitric oxide (NO)—believed to be identical to EDRF—in endothelium-denuded strips from rat aorta.  Thus, halothane-induced attenuation of endothelium-mediated relaxation may also involve inhibition of guanylate cyclase in vascular smooth muscle cells.
Although various volatile anesthetics have been reported to attenuate endothelium-dependent relaxation in large arteries, such as rabbit or rat aorta, as described above, little information is available regarding the effects of the volatile anesthetics on endothelium-mediated vasorelaxation in small resistance arteries. It is well established that these small arteries and arterioles play an important role in the regulation of local blood flow and systemic arterial pressure. [7,8] Understanding direct anesthetic actions on endothelial function in these vessels may, therefore, prove important in understanding the cardiovascular actions of these agents. Furthermore, previous studies, both in vivo and in vitro, have documented that there exist significant differences in volatile anesthetic pharmacology between conductance and resistance arteries, [9–14] indicating the need to further investigate volatile anesthetic action in small resistance arteries.
Endothelium-dependent relaxation that is resistant to EDRF inhibitors has recently been reported, [15–18] and such relaxation may be mediated by an endothelium-derived hyperpolarizing factor (EDHF)[19–24] that is distinct from EDRF. [15,19,23,24] However, although the relative importance of EDRF and EDHF in the endothelium-dependent relaxation in various types of vessels is under active investigation,  there is, as yet, no information regarding a difference in volatile anesthetic actions on endothelium-dependent relaxation that is sensitive or resistant to EDRF inhibitors.
In this report, we describe the actions of the volatile anesthetics isoflurane, enflurane, and sevoflurane on endothelium-dependent relaxation in small rabbit mesenteric resistance vessels. Employing isometric tension recording and microelectrode methods, we first characterized the vasorelaxing and membrane-hyperpolarizing actions of ACh and SNP, and the sensitivity of the ACh and SNP responses to EDRF inhibitors. After this, we examined the effects of isoflurane, enflurane, and sevoflurane on the endothelium-dependent ACh-induced relaxations that were sensitive and resistant to the EDRF inhibitors, and the effects of these anesthetics on the endothelium-independent relaxation produced by SNP.
Materials and Methods
After receiving institutional approval for this animal study, male albino rabbits (2–2.5 kg) were given sodium pentobarbital (40 mg/kg intravenously) and exsanguinated. The mesentery in the jejunal region was immediately placed in a dissecting chamber filled with preoxygenated Krebs-bicarbonate solution, and the mesenteric artery was rapidly excised. The distal portion of the third order branch (0.2–0.3 mm in diameter; the main truncus was not counted) were used for the current experiments. Under a binocular microscope, the fat and adhering connective tissue were carefully removed, and thin circumferential strips (0.4–0.6 mm long and 0.1–0.12 mm wide) were prepared for tension-recording. The length and diameter of the preparations for the electrophysiologic experiment were about 10 and 0.2 mm (without dissection), respectively. In some experiments, the endothelium was removed by gently rubbing the intimal surface with the round surface of a small pin, as previously described.  Each vascular preparation was obtained from a different animal in both the tension-recording and electrophysiologic experiments.
Recording of Mechanical Activity and Membrane Potential Measurements
Mechanical responses were measured by attaching the strips to a strain gauge (UL-2 type, Shinko Co., Tokyo) in a small chamber (0.9 ml capacity) filled with Krebs solution, as previously reported. [17,26,27] The chamber was placed on a microscope stage and the strip was stretched to 1.1 times its resting (without tension) length to obtain maximal contractile response to high Potassium sup +. The protocol for the experiments to determine the optimal stretch was as follows. After a 30-min equilibration time, high Potassium sup +(40 mM) was applied to the strip (without stretch) for 2 min with an interval of 8 min to allow the contractile response to high Potassium sup + to become constant. The length was then cumulatively increased in a stepwise fashion, and the contractile response was allowed to stabilize. Solution changes were accomplished by rapid perfusion from one end while aspirating simultaneously from the other end. All tension experiments were performed at 35 degrees Celsius.
Electrophysiologic experiments were performed separately from tension experiments. The membrane potential of the smooth muscle cell was measured using a glass capillary microelectrode filled with 3 M KCI (tip resistance, [nearly equal] 40–80 M Omega), as previously reported. [15,20,26,27] The tissue segment was pinned onto a rubber plate in a chamber (1.5 ml volume) with a stretch of 30% and superfused with Krebs solution (35 degrees Celsius) at a flow rate of 3 ml/min. The electrode was inserted into the muscle cell from the outer surface of the vessel. Although, ideally, the vessel should be stretched to the same degree as that in tension experiments, it was impossible to impale the microelectrode without a stretch of this degree ([nearly equal] 30%). With the 30% stretch, the contractile response to high Potassium sup + was [nearly equal] 90% of that obtained with a stretch of 10%.
Application of agonists was initiated after a 30-min equilibration after mounting of the vascular tissues in the chambers. The experiments themselves were started after the vascular responses to norepinephrine (NE), ACh, or SNP all became constant ([nearly equal] 2–3 h and [nearly equal] 1 h after mounting the strips in endothelium-intact and endothelium-denuded strips, respectively). Constant vascular responses to NE, ACh, or SNP were then observed for [nearly equal] 3–6 h. In tension-recording experiments, only one experiment was performed in each strip; in electrophysiologic experiments, several experiments were performed when the effects of inhibitors on the ACh-induced hyperpolarization were reversible.
Solutions and Drugs
The composition of the Krebs solution was as follows (mM): NaCl 111.9, KCl 3.7, MgCl21.2, CaCl22.6, NaHCO325.5, KH3PO 1.2, glucose 11.4. Solutions with various concentrations (1.2–77 mM) of Potassium sup + were prepared by replacing NaCl with KCl, isoosmotically. The solution was bubbled with 95% O2/5% CO2(pH 7.3–7.4, 35 degrees Celsius).
Norepinephrine HCl, methemoglobin (bovine), methylene blue (MB), NG-nitro L-arginine (LNNA), glibenclamide, and apamin were obtained from Sigma Chemical Co. (St. Louis, MO). Tetraethylammonium chloride, barium chloride, 4-aminopyridine, and SNP were obtained from Nacalai Tesque (Kyoto, Japan). Isoflurane and enflurane were obtained from Dainabot Co. (Tokyo, Japan), and sevoflurane was obtained from Kodama Pharmaceutical Co. (Osaka, Japan).
Oxyhemoglobin (HbO2) was prepared by reducing commercial bovine Hb containing 75% methemoglobin. After adding 10 times sodium dithionite to Hb (1 mM), dialysis by 200 times distilled water was carried out at 4 degrees Celsius. The identification and final concentration of oxyhemoglobin were determined spectrophotometrically.
Characterization of ACh and SNP Responses
The effects of endothelial removal, various concentrations of extracellular potassium (1.2–80 mM), EDRF pathway inhibitors (LNNA, MB, and HbO2), and various Potassium sup + channel blockers (Glibenclamide, BaCl2, 4-aminopyridine [4AP], apamin, and tetraethylammonium [TEA]) on the ACh-induced hyperpolarization and relaxation were studied. The ACh-induced hyperpolarizations were studied in the presence and absence of NE (10 micro Meter). In tension-recording experiments, the strips were precontracted with NE, the primary neurotransmitter of sympathetic vasoconstrictor system, which plays a significant role in maintaining vasomotor tone in vivo. After the contractile response to NE reached a steady state, various concentrations (0.03–10 micro Meter) of ACh were applied to the strip in a random sequence either in the presence or absence of inhibitors. In experiments using the EDRF pathway inhibitors, the strips were pretreated with the inhibitor (s) for 20 min before application of ACh or NE, and the inhibitors were then applied throughout the remainder of the experiment. Because preliminary experiments showed that indomethacin (3 micro Meter), a cyclooxygenase inhibitor, did not significantly affect the ACh relaxation in this artery (n = 5, not shown), subsequent experiments were performed in the absence of indomethacin. In addition, because preliminary experiments indicated that ACh exerted its maximal hyperpolarizing effects within 1 min, ACh was applied for 1 min in the electrophysiologic experiments described here.
The concentrations of the EDRF inhibitors used in these experiments were sufficient to exert their maximal effects or were the maximum possible concentrations that could be used without producing significant toxicity. The maximal concentrations of LNNA and HbO2used, 100 micro Meter and 10 micro Meter, respectively, produced no more inhibition of ACh relaxation compared with 30 micro Meter LNNA or 3 micro Meter HbO2(analyzed by a two-factor [concentration, treatment] ANOVA, not shown). In contrast, 10 micro Meter MB produced significant additional inhibition compared with 3 micro Meter MB (P < 0.05 by a two-factor ANOVA), but 30 micro Meter MB was apparently toxic to the vascular tissues, as evidenced by deterioration of contractile function that occurred [nearly equal] 60–90 min after application of this concentration (30 micro Meter) of MB.
The Potassium sup + channel blockers tested in this study were glibenclamide (3 and 10 micro Meter), BaCl2(30 and 100 micro Meter), 4AP (30 and 100 micro Meter), apamin (0.1 and 0.3 micro Meter), and TEA (0.1–30 mM). Glibenclamide and apamine are known to selectively inhibit the ATP-sensitive Potassium sup + channel and the low conductance Calcium2+-activated Potassium sup + channel, respectively, while Barium2+, 4AP, and TEA have been reported to block various types of Potassium sup + channels, including voltage-dependent Potassium sup + channels, Calcium2+-activated Potassium sup + channels, or ATP-sensitive Potassium sup + channels. .
The effect of SNP on resting membrane potential was studied in endothelium denuded (E[-])D strips. The relaxing effects of SNP, as well as the effects of MB (10 micro Meter), HbO2(10 micro Meter), and TEA (10 and 30 mM) on the SNP (1–10 micro Meter)-induced relaxation were studied in E(-) strips preconstricted with NE (10 micro Meter).
The effects of 1 MAC isoflurane (2% in rabbit), enflurane (2.8%), and sevoflurane (3.7%) on ACh-induced relaxation in endothelium intact (E[+]) strips and on SNP-induced relaxation in E(-) strips were studied. [29,30] The effects of the anesthetics on ACh and SNP relaxations were studied in strips preconstricted with NE and the anesthetics were applied for 5 min before, and then continuously during, application of NE and either NE + ACh or NE + SNP. As noted below, all three anesthetics significantly inhibited the contractions induced by 10 micro Meter NE. To compensate for this effect, the NE concentration was increased to produce approximately the same level of tension (i.e., contraction) as 10 micro Meter NE produced in control strips (i.e., without anesthetic) before measuring relaxation produced by ACh or SNP. This technique of compensating for vasodilating effects has previously been employed for both volatile anesthetics and basic polypeptides when examining the effects of these drugs on ACh relaxation. [6,31].
Volatile Anesthetics Delivery, and Analysis
The volatile anesthetics were delivered via calibrated agent-specific vaporizers in line with the Oxygen2/CO2equilibration gas mixture aerating the Krebs solution. The anesthetic concentration in the resulting gas mixture was monitored with a calibrated infrared multi-anesthetic gas analyzer (Capnomac; Datex, Tewksburry, MA). Preliminary experiments indicated that 10 min was sufficient for the Krebs solution to be saturated with the anesthetics at each partial pressure. Therefore, the Krebs solution was equilibrated with the anesthetics for 15 min before the study. The experimental chamber was covered with a thin glass plate to prevent the equilibration gas from escaping into the atmosphere. The concentrations of isoflurane, enflurane, and sevoflurane in the Krebs solution were also determined by gas chromatography yielding values of 0.37 plus/minus 0.01 mM, 0.71 plus/minus 0.05 mM, and 0.42 plus/minus 0.05 mM, respectively (mean plus/minus SD, n = 3–5). These values were 86, 88, and 84% of the theoretical values predicted by the partition coefficient of isoflurane and enflurane in Krebs solution  or by the partition coefficient of sevoflurane in water (the partition coefficient of sevoflurane in Krebs solution is not available). Thus, these concentrations were equivalent to 0.86, 0.88, and 0.84 MAC in rabbit, respectively, although the concentration of sevoflurane may be slightly underestimated, because the partition coefficients of other volatile anesthetics in Krebs solution is about 90% of that in water. .
Calculation and Statistical Analysis
All results were expressed as mean plus/minus SD. In the tension-recording experiments, n denotes the number of animals. In the electrophysiologic experiments, n denotes the number of impalements (cells); the number of preparations (animals) is noted separately. The ACh relaxation in the presence of NE and the ACh hyperpolarization were assessed at the points at which its relaxing or hyperpolarizing effects reached a maximum. The statistical assessment of the data was made by an ANOVA (one or two factor), Scheffe's F test, and Student's t test, as appropriate. A level of P < 0.05 was considered significant.
Acetylcholine and Sodium Nitroprusside-Induced Changes in Membrane Potential in Vascular Smooth Muscle
In E(+) strips, ACh hyperpolarized the membrane potential (MP) of smooth muscle cells in both the absence and the presence of NE (3 and 10 micro Meter)(Figure 1). Norepinephrine (3 and 10 micro Meter) itself significantly depolarized the MP, and ACh (3 micro Meter) reversed the NE-induced depolarization in E(+) strips (Figure 1). Acetylcholine (up to 10 micro Meter) had no significant effect on MP in E(-) strips in the absence (n = 12, five preparations) or in the presence (n = 15, five preparations) of NE (not shown).
As shown in Figure 2(A), 0.1 - 10 micro Meter ACh hyperpolarized the resting MP of the smooth muscle cells, in both 3.6 and 5.9 mM extracellular Potassium sup +([K sup +]o), in a concentration-dependent manner. The hyperpolarizations in the presence of 3.6 mM [K sup +]owere consistently larger (P < 0.05) than those in the presence of 5.9 mM [K sup +]o(Figure 2(B)).
(Figure 2(C)) shows that the hyperpolarization induced by 3 micro Meter ACh varies as a function of extracellular Potassium sup + concentration. At low [K sup +]ovalues, MP deviated significantly from the equilibrium potential for Potassium sup +(EK) predicted by the Nernst's equation, and the ACh-induced hyperpolarizations were the largest. Acetylcholine-induced hyperpolarizations almost disappeared in the presence of [K sup +]oabove 10 mM where MP was essentially equal to EK.
To obtain more distinct hyperpolarizing responses to ACh, subsequent experiments using EDRF and Potassium sup + channel inhibitors were performed in the presence of 3.6 mM [K sup +]o. Under these conditions, LNNA (100 micro Meter) and HbO2(10 micro Meter) did not significantly affect the resting MP, and ACh (3 micro Meter) still hyperpolarized (P < 0.05) the smooth muscle MP in the presence of these inhibitors (Figure 3(A)). The hyperpolarizations in the presence of LNNA and HbO2were 9.6 plus/minus 3.6 mV (n = 8, four preparations) and 10.3 plus/minus 3.2 mV (n = 5, three preparations), respectively, both of which were not significantly different from the control hyperpolarization measured in the absence of the inhibitors (10.5 plus/minus 2.2 mV [n = 9, four preparations]). Methylene blue (10 micro Meter) and a cocktail application of LNNA (100 micro Meter). HbO2(10 micro Meter), and MB (10 micro Meter) significantly depolarized the resting MP (Figure 3(A)). However, the ACh-induced hyperpolarizations were at least as large as the control response after the inhibitors, and the absolute level of MP achieved after ACh was not different than control with any of the inhibitors.
The Potassium sup + channel blockers glibenclamide (Glib, 10 micro Meter), BaCl2(100 micro Meter), 4-aminopyridine (4AP, 100 micro Meter), or apamin (0.3 micro Meter) had little if any effect on the hyperpolarizing action of ACh (Figure 3(B)). The hyperpolarization in the presence of apamin (9.2 plus/minus 3.7 mV, n = 5, three preparations), which was the only Potassium sup + blocker that did not affect MP, was not significantly different from the control hyperpolarization (10.5 plus/minus 2.2 mV, n = 9, four preparations). Once again, the magnitude of the ACh-induced hyperpolarizations in the presence of all these Potassium sup + channel blockers were at least as large as the control responses, and the absolute level of MP achieved after ACh was not different from control.
The Potassium sup + channel blocker TEA (3–30 mM) also caused some depolarization in MP, but TEA substantially attenuated the ACh-induced hyperpolarization such that the ACh response was completely abolished at TEA concentrations greater or equal to 10 mM (Figure 4). Among all the Potassium sup + channel and EDRF blockers tested, TEA (greater or equal to 10 mM) appeared to be the only agent capable of blocking the hyperpolarizing action of ACh.
In contrast to the effects of ACh on MP, SNP (up to 10 micro Meter) had no apparent effect on the resting MP in the E(-) strips (Figure 4(C)). There was no significant difference in the MP measured before (-73.4 plus/minus 2.1 mV, four preparations) and after 10 micro Meter SNP (72.9 plus/minus 3 mV, n = 12, four preparations).
Acetylcholine- and Sodium Nitroprusside-Induced Relaxation of Norepinephrine Contracted Vascular Smooth Muscle
In E(+) strips, 0.03 to 10 micro Meter ACh relaxed the strips precontracted with NE (10 micro Meter) in a concentration-dependent manner (Figure 5(A and D)). In contrast, even very high concentrations of ACh (up to 10 micro Meter) failed to relax E(-) strips precontracted with NE (10 micro Meter)(not shown). The application of the EDRF pathway inhibitors LNNA (100 micro Meter), HbO2(10 micro Meter), MB (10 micro Meter), or a cocktail application of LNNA (100 micro Meter), HbO2(10 micro Meter), and MB (10 micro Meter) significantly enhanced the contractile response to NE in the E(+) strips: LNNA, 2.4 plus/minus 1.8 times control (n = 31); HbO2, 1.7 plus/minus 0.5 times control (n = 8); MB, 2.5 plus/minus 1.5 times control (n = 29); and LNNA + HbO2+ MB, 3.3 plus/minus 2.1 times control (n = 28, Figure 5(B and C)). These EDRF inhibitors almost abolished the relaxations elicited by the low concentrations (less or equal to 0.1 micro Meter) of ACh, but failed to completely abolish the relaxations elicited by high concentrations (greater or equal to 0.3 micro Meter) of ACh (Figure 5(B-D)). Even a cocktail application of LNNA, HbO2, and MB failed to completely abolish the relaxations elicited by high concentrations (greater or equal to 0.3 micro Meter) of ACh (Figure 5(C and D)).
The residual ACh relaxation in the presence of the EDRF pathway inhibitors was completely inhibited by the additional application of 10 and 30 mM TEA (Figure 5(C and D)). However, the Potassium sup + channel blockers Glib (3 and 10 micro Meter), apamin (0.1 and 0.3 micro Meter), 4AP (30 and 100 micro Meter), and BaCl2(30 and 100 micro Meter) had little or no effect on the residual ACh relaxations that were resistant to the EDRF inhibitors (Figure 6).
As shown in Figure 7, SNP dose-dependently (1–10 micro Meter) relaxed the E(-) strips precontracted with NE (10 micro Meter)(Figure 7). However, in contrast to the actions on ACh relaxation in the E(+) strips, TEA (10 and 30 mM) had no significant effect on SNP-induced relaxation (Figure 7). The SNP-induced relaxation was substantially attenuated by the cocktail application of the EDRF pathway inhibitors HbO2(10 micro Meter) and MB (10 micro Meter). These EDRF inhibitors almost completely abolished the relaxations elicited by 1 and 3 micro Meter SNP (Figure 7(A-C)).
Effects of Volatile Anesthetics on Acetylcholine- and Sodium Nitroprusside-Induced Relaxations
As shown above, the relaxations produced by the low concentrations (0.03–0.3 micro Meter) of ACh were abolished or strongly inhibited by the EDRF inhibitors (Figure 5). lsoflurane, enflurane, and sevoflurane all caused significant and reversible inhibition of these endothelium-dependent ACh-induced relaxations that were sensitive to the EDRF pathway inhibitors, as shown in Figure 8, and no significant differences were observed in their inhibitory effects on ACh relaxation among these anesthetics.
Because all three anesthetics significantly inhibited the NE(10 micro Meter) contractions (Figure 8(A-D)), the concentrations of NE were increased in the strips pretreated with the anesthetics to obtain a NE precontraction with the same amplitude as that in the control. The concentrations of NE required to obtain precontractions with the same amplitude as the control were 36.7 plus/minus 12.9 micro Meter (n = 7), 25.6 plus/minus 5.3 micro Meter (n = 9), and 50.6 plus/minus 14.1 micro Meter (n = 6) in the presence of isoflurane, enflurane, and sevoflurane, respectively. Using these concentrations of NE, there were no significant differences in the amplitude of the NE precontractions before and after the anesthetic in all groups. There were no significant differences among the anesthetics for their inhibitory effects on precontraction induced by 10 micro Meter NE (Figure 8(D)).
The effects of volatile anesthetics on ACh relaxation resistant to the EDRF inhibitors were studied using the relaxation induced by 3 micro Meter ACh in the presence of the (EDRF) inhibitors. This relaxation was sensitive to the Potassium sup + channel blocker TEA (Figure 5), which selectively blocked the endothelial-dependent hyperpolarizing action of ACh (Figure 4) that was resistant to the EDRF inhibitors (Figure 3). All three anesthetics significantly and reversibly inhibited this 3 micro Meter ACh-induced relaxation in the presence of the EDRF inhibitors (Figure 9(D)), and no significant differences were observed in the magnitudes of their inhibitory effects on ACh relaxation in the presence of the EDRF inhibitors. All three anesthetics again produced significant inhibition of the precontraction induced by 10 micro Meter NE (P < 0.05), and the concentrations of NE in the strips pretreated with the anesthetics were increased to obtain a NE-induced precontraction of the same amplitude as in the control (Figure 9). There were no significant differences among the three anesthetics in the amount of inhibition of the 10 micro Meter NE contraction (Figure 9(E)), and the concentrations of NE required to obtain a precontraction of the same amplitude as the control were 14.3 plus/minus 4.6 (n = 6), 14.8 plus/minus 3.5 (n = 4), and 16.5 plus/minus 4 micro Meter (n = 4), in the presence of isoflurane, enflurane, and sevoflurane, respectively.
The effects of volatile anesthetics on the endothelium-independent relaxation induced by SNP (3 micro Meter) in the E(-) strips are shown in Figure 10. As described above, this SNP-induced relaxation was almost completely abolished by the EDRF inhibitors MB and HbO2(Figure 7). In contrast to their effects on ACh-induced relaxation, however, neither isoflurane, enflurane, nor sevoflurane had any significantly inhibitory effect on the SNP (3 micro Meter [nearly equal] IC50values)-induced relaxation (Figure 10(A-D)). As in the E(+) strips, the three anesthetics significantly inhibited the NE-induced precontraction in the E(-) strips (Figure 10(A-C and E)). The concentrations of NE in the strips pretreated with the anesthetics were again increased to obtain almost the same NE precontraction amplitude as the control before application of SNP. The concentrations of NE required to obtain precontractions of the same amplitude as the control were 18.33 plus/minus 4.1 (n = 6), 14.3 plus/minus 5.3 (n = 4), and 20 plus/minus 0 micro Meter (n = 4), in the presence of isoflurane, enflurane, and sevoflurane, respectively. There were no significant differences in the amplitude of the NE contraction before and after the anesthetics in all three groups, and there were no significant differences in the inhibitory effects of the anesthetics on the contraction induced by 10 micro Meter NE (Figure 10(E)).
One additional finding of interest was the transient constriction induced by enflurane in both the E(+) and E(-) strips (Figure 8(B), Figure 9(B), and Figure 10(B)). Similar transient constrictions were not observed with either isoflurane and sevoflurane (not shown). The amplitudes of the enflurane-induced contraction in the E(+) strips, in the E(+) strips pretreated with the EDRF inhibitors, and in the E(-) strips were 0.81 plus/minus 0.20 (n = 9), 0.72 plus/minus 0.12 (n = 4), and 0.73 plus/minus 0.08 (n = 6) times the amplitude of a 40 mM Potassium sup +-induced phasic contraction in each condition, respectively. Furthermore, enflurane generated a similar transient contraction in Calcium2+-free, 2 mM EGTA solution in the E(-) strips with an amplitude equivalent to 0.75 plus/minus 0.12 (n = 3) times that of the 40 mM Potassium sup +-induced phasic contraction (in the presence of 2.6 mM Calcium2+).
Characteristics of the Acetycholine-Induced Relaxation and Hyperpolarization
The current study demonstrated that ACh causes endothelium-dependent relaxation and hyperpolarization of the smooth muscle in this small splanchnic resistance artery, as observed previously in other arteries. [15–26] In addition, our results strongly indicate that the relaxation induced by low concentrations (less or equal to 0.1 micro Meter) of ACh is totally dependent on the action of EDRF, but the relaxation induced by higher concentrations (greater or equal to 0.3 micro Meter) of ACh is, in part, mediated by a hyperpolarizing effect of ACh that is independent of the action of EDRF. In support of this, only the relaxation induced by the low concentrations of ACh was abolished by the inhibitors of the EDRF (NO) pathway, including LNNA, HbO2, and MB. The concentrations of ACh that produced relaxation resistant to the EDRF inhibitors, i.e., more than 0.3 micro Meter, corresponded to those that produced hyperpolarization that was not affected by the EDRF inhibitors. Moreover, both the ACh relaxation resistant to the EDRF inhibitors and the ACh hyperpolarization were strongly inhibited or abolished by TEA (10–30 mM). Finally, the effects of SNP strongly indicate that the ACh hyperpolarization is independent from the action of EDRF. The vasorelaxation produced by SNP, which results from NO (i.e., EDRF) generation in smooth muscle independent of the endothelium, was strongly inhibited by the EDRF inhibitors; however, SNP had no effect on membrane potential and SNP-induced relaxation was not affected by TEA (10–30 mM). The proposed mechanisms of the hyperpolarization-mediated vasorelaxation are inhibition of the voltage-dependent Calcium2+ channels, inhibition of IP3synthesis, or both. [33–35].
Whether the hyperpolarization is mediated by some humoral substance(s), i.e., EDHF, or by electrotonic spread of endothelial hyperpolarization to smooth muscle cell membrane still appears to be controversial. [23,36–38] Various EDRF releasers, including ACh, have been shown to hyperpolarize “endothelial” cell membrane, which is thought to result from increased intracellular Calcium2+([Calcium sup 2+]i) and subsequent activation of Calcium2+-activated Potassium sup + channels in endothelial cells. [37,39–41] Because EDRF is synthesized from L-arginine by a constitutive Calcium2+-dependent NO-synthase and released by a Calcium2+-dependent mechanism, [40,41] the endothelial hyperpolarization has been proposed to serve as an important signal for synthesis and release of EDRF by providing the necessary driving force for transmembrane Calcium sup 2+ influx, and sustained increases in [Calcium2+]i in endothelial cells. This endothelial hyperpolarization could then be transmitted to smooth muscle cells electrotonically via some low-resistance electrical pathway, such as gap junctions, which are known to exist between endothelial and smooth muscle cells. [37,38,42,43] According to this model, the hyperpolarization of the smooth muscle membrane would thus be generated electronically as part of the signal transduction pathway involved in the synthesis and release of EDRF from endothelial cells; and TEA, a Potassium sup + channel blocker, may block the EDRF inhibitor-resistant component of the ACh relaxation by inhibiting endothelial Calcium2+-activated Potassium sup + channels, thereby preventing endothelial and smooth muscle hyperpolarization.
Alternatively, several recent studies have demonstrated that ACh can generate smooth muscle hyperpolarization in situations in which electrotonic transmission cannot occur. [21,36,44] It has thus been suggested that ACh hyperpolarization may be mediated by some endothelial-derived hyperpolarizing factor(s)(EDHF). In further support of this model, Beny has demonstrated that injection of electrical pulses into endothelial cells does not evoke any electrical response in the smooth muscle cells of rabbit aorta, indicating that electrotonic spread of hyperpolarization through gap junctions does not occur.  Assuming that ACh hyperpolarizations are thus mediated by some endothelial humoral substance, the lack of effect of LNNA, HbO2, or MB on the ACh-induced hyperpolarization, as well as the lack of effect of SNP on membrane potential, indicates that EDHF is distinct from EDRF (NO). This leads to the proposal that the ACh relaxation in these small mesenteric arteries consists of both EDRF- and EDHF-mediated components. According to this model, TEA may block Potassium sup + channels in endothelial cells that are important in the generation of the EDHF signal, or TEA may block Potassium sup + channels in vascular smooth muscles that are involved in the EDHF hyperpolarizing response.
Methylene blue and the cocktail application of LNNA, HbO2, and MB significantly depolarized the membrane potential. However, it is difficult to conclude from this that EDRF itself has a hyperpolarizing effect in this artery, because LNNA, which is known to be the most specific inhibitor of the EDRF pathway,  did not affect the membrane potential. Additionally, SNP, which presumably results in generation of NO in vascular smooth muscle, had no effect on membrane potential. The mechanism of the depolarizing effects of MB is currently unknown, but could reflect a toxic effect because higher concentrations of MB caused rapid deterioration of the preparation.
Ionic Mechanisms of the Acetylcholine-Induced Hyperpolarization
Our results strongly indicate that the ACh hyperpolarization is mediated through an increase in Potassium sup + permeability, as previously described. [20,24,33] In low [Potassium sup +]O, the membrane potential deviated most from EKand the ACh hyperpolarizations were the largest. Conversely, in the presence of high [Potassium sup +]O, where the membrane potential was essentially the same as EK, ACh-induced hyperpolarizations were absent. In addition, the hyperpolarizations were almost completely inhibited in the presence of TEA (greater or equal to 10 mM), a nonspecific Potassium sup + channel blocker. [28,47,48] The lack of effect of glibenclamide (10 micro Meter), a specific blocker of ATP-sensitive Potassium sup + channels, and Barium2+(100 micro Meter) indicate that the ATP-sensitive or Barium2+-sensitive Potassium sup + channels are not involved in the generation of the hyperpolarization in the small rabbit mesenteric artery. Glibenclamide (10 micro Meter) and Barium2+(50 micro Meter) have been shown to inhibit the ACh-induced hyperpolarization in rabbit cerebral arteries [33,49]; however, our results are similar to those recently reported in the guinea-pig coronary artery  and indicate there may be regional or species differences in the mechanisms behind the generation of ACh hyperpolarization. In addition, both 4AP-sensitive Potassium sup + channels, e.g., the low conductance (10 pS), Calcium2+-dependent, ATP-sensitive Potassium sup + channel in portal vein, [50,51] and apamin-sensitive channels (0.1 and 0.3 micro Meter), i.e., the low conductance Calcium2+-activated Potassium sup + channels, [28,52,53] appear to play little, if any, role in the ACh-induced hyperpolarization or relaxations, because 4AP and apamin had little, if any, effect on the responses to ACh. Finally, although we did not examine the effects of charybdotoxin, a specific inhibitor of high-conductance Calcium2+-activated Potassium sup + channels (Big Potassium sup +[BK] channels),  the lack of significant effect of low concentrations (0.1–1 mM) of TEA on the hyperpolarization indicates that the BK channels do not play a significant role in the generation of the hyperpolarization. The BK channels are highly sensitive to TEA, with the reported EC50values for block by external TEA being 100–300 micro Meter, and BK currents are almost completely abolished by low concentrations of TEA. [28,54–57] In addition, we have previously reported that CTX (100 nM) did not affect the ACh-induced endothelium-dependent hyperpolarization in rabbit coronary artery,  and it is also known that CTX (100 nM) does not affect ACh (3 micro Meter)- and bradykinin (0.1 micro Meter)-induced endothelium-dependent hyperpolarization in rat mesenteric and canine coronary arteries, respectively (Drs. M. Nakashima and P. M. Vanhoutte, personal communication, 1994). Because nitric oxide has recently been shown to directly activate the BK channels in rabbit aortic smooth muscle,  the lack of involvement of BK channels is consistent with the proposal that the hyperpolarization is independent from the action of NO.
Although the observed depolarizing action of all the Potassium sup + channel blockers with the exception of apamin implies that Potassium sup + channels that are sensitive to these blockers may contribute to maintenance of the resting membrane potential, our observations indicate that the ACh hyperpolarization is mediated through the activation of TEA-sensitive Potassium sup + channels that are distinct from those inhibited by glibenclamide (10 micro Meter), Barium sup 2+(less or equal to 100 micro Meter), 4AP (less or equal to 100 micro Meter), or apamin (less or equal to 0.3 micro Meter). However, a variety of Potassium sup + channels are blocked by TEA, and the TEA sensitivity does not allow for a specific identification of the Potassium sup + channels involved. [28,47,48] Further investigation will be necessary to clarify the specific channels, as well as the tissue type (i.e., smooth muscle versus endothelium), that are responsible for this effect.
Effects of Volatile Anesthetics on the Acetylcholine Relaxation
Isoflurane, enflurane, and sevoflurane all caused a reversible inhibition of the relaxation induced by low concentrations of ACh (0.03–0.3 micro Meter), which was highly sensitive to the EDRF (NO) inhibitors. This indicates that these volatile anesthetics inhibit the EDRF-mediated relaxation in this small splanchnic resistance artery, as observed in other large conductance arteries. [1–5] Furthermore, all three anesthetics had no effect on the SNP (3 micro Meter [nearly equal] IC50value)-induced relaxation, which is presumably mediated through endothelium-independent NO generation from SNP in smooth muscle. Consistent with this, SNP-induced relaxation was inhibited by MB and HbO sub 2. These latter results indicate that the inhibition of the ACh-induced relaxation in these small mesenteric arteries was caused by inhibition of EDRF production, release from endothelium, or both, but not by an effect on vascular smooth muscle responsiveness to NO. This is consistent with previous observations regarding isoflurane and enflurane in rat aorta. [1,2,4,5] However, our data also may not be inconsistent with the mechanism proposed by Yoshida et al. for sevoflurane-induced inhibition of endothelium-dependent relaxation, which involved formation of oxygen free radicals that inactivated EDRF.  The potential involvement of guanylate cyclase inhibition in vascular smooth muscle, which has been proposed to be involved in the inhibitory effects of halothane,  remains controversial. However, the lack of effect of the anesthetics we studied on SNP relaxation indicate that the 1 MAC concentrations we used in our experiments did not interfere with guanylate cyclase activation in vascular smooth muscle.
In this small resistance artery, isoflurane, enflurane, and sevoflurane also inhibited the ACh relaxation that was resistant to the EDRF inhibitors. This relaxation, as discussed above, appears to be mediated by ACh-induced hyperpolarization. If the hyperpolarization were mediated by some humoral substance(s), i.e., EDHF, these findings indicate that these anesthetics inhibit the EDHF-mediated relaxation in addition to the EDRF-mediated relaxation. This may result from inhibition of a critical step in the ACh signal transduction pathway common to release of both EDRF and EDHF. Alternatively, if the hyperpolarization was caused by electrotonic transmission from the endothelial cell to the smooth muscle cell, and was not caused by EDHF release, these findings may indicate that anesthetic inhibition of EDRF production, release, or both may be related to inhibition of endothelial hyperpolarization (see above). Further electrophysiologic investigations regarding volatile anesthetic effects on the ACh-induced hyperpolarization of vascular smooth muscle cell membrane, as well as of vascular endothelial cell membrane, will be necessary to definitively solve these issues.
One possible explanation for the current findings is that volatile anesthetics may globally inhibit the two pharmacologically distinct ACh-induced vascular responses by interfering with muscarinic receptors. However, Muldoon et al. showed that halothane inhibits both ACh- and bradykinin-induced relaxations  and, more recently, Uggeri et al. showed that halothane, isoflurane, and enflurane inhibited not only receptor-mediated (methacholine-induced) relaxation but also nonreceptor-mediated (A23187-induced) endothelium-dependent relaxation.  Similarly, Yoshida has also shown that sevoflurane inhibits both bradykinin-induced and A23187-induced relaxation, as well as ACh-induced relaxation.  These results thus strongly indicate that the site of action is distal to the endothelial cell receptors. The observed volatile anesthetic effects on the ACh-mediated vascular responses are thus probably not caused by interference with muscarinic receptor activation.
Vasodilating Action of Volatile Anesthetics
Various EDRF inhibitors greatly enhanced the NE contraction, indicating that NE stimulates the endothelial cell to release EDRF, as has been proposed,  and that the EDRF system is an important determinant of the contractile response to NE. It is, therefore, conceivable that volatile anesthetics could have an enhancing effect on the contractile response to NE under certain conditions, like the other EDRF inhibitors. However, the anesthetics inhibited the NE contraction in both the endothelium-intact and endothelium-denuded strips, indicating that the direct vasodilating effect predominated over any vasoconstricting effect because of EDRF inhibition. Because the sympathetic nervous system plays a major role in maintaining splanchnic vasomotor tone, and NE is the primary sympathetic neurotransmitter, the observed inhibitory effects of these anesthetics on the contractile response to NE indicate that the vasodilating action of these anesthetics in these vessels is probably dominant over their inhibitory action on the EDRF and EDHF pathway in vivo.
Isoflurane, enflurane, and sevoflurane inhibited the NE contractions in the endothelium-intact strips, in the EDRF inhibitor-treated strips, and in the endothelium-denuded strips. In each condition, there were no significant differences in the vasodilating action of isoflurane, enflurane, and sevoflurane when compared at equi-MAC concentrations in rabbit. We did not compare the vasodilating action of each anesthetic between endothelium-intact and endothelium-denuded strips because all the inhibitors of the EDRF pathway significantly enhanced (1.7–2.4 times) the NE contraction in this artery, as previously observed. [17,26,27] This strongly indicates the possibility that the contractile response to NE would also be enhanced by the mechanical endothelial denudation. As previously discussed, [60,61] comparisons between the E(+) and E(-) strips are problematic when the contractile responses to NE are different and the vasodilating action of ACh is estimated according to the percentage decrement from the control response. However, we can conclude from our data that all these anesthetics have direct vasodilating actions on these isolated small splanchnic arteries, and that the vasodilating actions of all these anesthetics have a component that is independent of endothelium.
Finally, enflurane produced transient contractions in various conditions, i.e., in the E(+) strips, in the EDRF inhibitor-treated strips and in the E(-) strips both in the presence and absence of extracellular Calcium2+, indicating that enflurane causes a contraction by releasing Calcium2+ from intracellular Calcium2+ stores, independent of the endothelium.
In conclusion, ACh caused both endothelium-dependent vasorelaxation and hyperpolarization of the smooth muscle cell membrane in the rabbit small mesenteric artery. The relaxation induced by low concentrations of ACh is mediated by EDRF, while the relaxation induced by high concentrations of ACh probably consists of EDRF-mediated and hyperpolarization (or EDHF)-mediated components. Acetylcholine causes hyperpolarization by increasing Potassium sup + permeability via activation of TEA-sensitive Potassium sup + channels independently from the action of EDRF. Isoflurane, enflurane, and sevoflurane all produced reversible inhibition of EDRF-mediated relaxation, probably by inhibiting the production, release, or both of EDRF, or by inactivating EDRF, but not by interfering with guanylate cyclase within smooth muscle cell. The inhibitory effects of these anesthetics on the ACh relaxation resistant to the EDRF inhibitors indicates that these anesthetics may also inhibit ACh-induced hyperpolarization (or EDHF)mediated relaxation. Finally, isoflurane, enflurane, and sevoflurane all have vasodilating action in this isolated small resistance artery, as evidenced by their effects on the NE-induced contractions, and their vasodilating action is, at least in part, endothelium independent.
The authors thank Dr. Hirosi Kuriyama, Professor, Department of Pharmacology, Faculty of Medicine, Kyushu University, and Professor Alex Evers, Head, Department of Anesthesiology, Washington University, St. Louis, Missouri, for their encouragement and critical comments, and Dr. Chris Lingle, Department of Anesthesiology, Washington University, St. Louis, Missouri, for his comments.