Although previous studies have proposed that anesthetics may influence signal transduction systems, their effects on the beta-adrenoceptor-mediated system have not been fully characterized in vascular smooth muscle. The aim of this study was to determine how halothane and isoflurane affect beta-adrenoceptor-mediated vasodilation in rat aorta and what mechanisms were involved.
Isometric tension and the intracellular calcium ion concentration ([Ca2+]i) were measured concomitantly in rat aortic strips from which the endothelium was removed. Strips precontracted with norepinephrine were dilated with the beta-adrenoceptor agonist, isoproterenol; the adenylyl cyclase activator, forskolin; or the membrane-permeable dibutyryl cyclic adenosine monophosphate (cAMP) with or without halothane or isoflurane. The effects of the anesthetics on each vasodilator were compared with the control responses. Beta-adrenoceptor binding characteristics and affinity for agonists were determined with [125I]-iodocyanopindolol with and without halothane or isoflurane. Furthermore, concentrations of cAMP induced by either isoproterenol or forskolin were measured with or without the anesthetics using an enzyme immunoassay procedure.
Halothane and isoflurane attenuated vasodilation and [Ca2+]i decreases induced by isoproterenol, whereas both anesthetics only slightly affected vasodilation and [Ca2+]i decreases induced by forskolin and dibutyryl cAMP. Halothane and isoflurane had no effect on beta-adrenoceptor binding characteristics and affinity for agonists. Three percent halothane or 4% isoflurane significantly reduced cAMP levels induced by isoproterenol but not by forskolin.
Halothane and isoflurane, at clinically relevant concentrations, can interfere with beta-adrenoceptor-mediated responses in the rat aorta at the steps after the agonist-receptor binding but before the adenylyl cyclase activation.
SEVERAL steps in the signal transduction pathway have been reported for relaxation of vascular smooth muscles by [small beta, Greek]-adrenoceptor agonists. Once the agonists interacts with the [small beta, Greek]-adrenoceptor, stimulatory G proteins activate adenylyl cyclase, which increases the intracellular concentration of cyclic adenosine monophosphate (cAMP). The increased concentration of cAMP activates a specific protein kinase, protein kinase A, and inhibits calcium ion (Ca2+) influx mainly by acting on inositol 1,4,5-triphosphate or Ca2+-activatedpotassium channel, resulting in a decrease in cytosolic Ca2+concentrations ([Ca2+]i) and vasorelaxation. In addition to this mechanism, [small beta, Greek] agonists induce vasorelaxation through an endothelium-dependent pathway. 
Volatile anesthetics are likely to influence several hydrophobic sites within the cell membrane, particularly membrane-binding proteins such as receptors, G proteins, catalysts, and ion channels. Recent evidence favors specific sites of action at the level of protein-protein [5,6]or lipid-protein interactions [7,8]at clinically relevant concentrations. However, the effects of volatile anesthetics on agonists-receptor bindings, G proteins, catalysts, and ion channels vary. For example, halothane has been shown to inhibit stimulatory G proteins without affecting inhibitory stimulatory G proteins in mice hearts. Conversely, halothane inhibited inhibitory G proteins without any effect on stimulatory G proteins in rat hearts and failing human hearts. Furthermore, halothane antagonized events in the signaling cascade sequence after activation of adenylyl cyclase in rat hearts and adipocytes. 
Although volatile anesthetics are suggested to have diverse effects on the signal transduction system of the vascular contraction and relaxation, little information is available about how they affect vasodilation mediated by the [small beta, Greek]-adrenergic receptor-G protein-effector pathway. In the current study, we investigated the effects of halothane and isoflurane, at clinically relevant concentrations, on the [small beta, Greek]-adrenoceptor-mediated responses in the vascular smooth muscle of the rat aorta, and we characterized the underlying mechanism of action. We chose rat aorta because it has plenty of [small alpha, Greek]- and [small beta, Greek]-adrenoceptors and allows us to measure [Ca2+]iand muscle tension concomitantly.
Materials and Methods
This study was approved by the Committee on Animal Research of the Sapporo Medical University. All experiments were conducted on male Sprague-Dawley rats (average weight, 280 g) that were 7 or 8 weeks old.
Preparation of Rat Aortic Strips
Rats were anesthetized with isoflurane, and the descending thoracic aorta was isolated and placed into a normal physiologic saline solution composed of 136.9 mM NaHCO3, 5.5 mM glucose, and 0.01 mM EDTA and bubbled with 95% oxygen plus 5% carbon dioxide at 37 [degree sign]C, pH 7.4. Fat and connective tissue were removed carefully, and vessels were cut into spiral strips that were about 2 mm wide and 7 mm long. The endothelium was removed by gently rubbing the intimal surface with a moistened cotton swab. The absence of the endothelium was verified by lack of relaxation with acetylcholine (10-6M).
Simultaneous Measurement of Muscle Tension and [Ca2+]i
The [Ca2+]iwas measured as described previously. Briefly, muscle strips were loaded with 10-5M acetoxymethyl ester of fura-2 (fura-2/AM) solution for 3-5 h at room temperature. A noncytotoxic detergent, 0.05% Cremophor EL, was added to solubilize the fura-2/AM in the solution. After the fura-2 loading, muscle strips were washed with physiologic saline solution and bubbled with 95% oxygen plus 5% carbon dioxide at 37 [degree sign]C, pH 7.4. The [Ca2+]iwas measured by a fura-2-Ca2+method using a fluorimeter (CAF-100; Japan Spectroscopic, Tokyo, Japan). The muscle strip was held horizontally in a temperature-controlled organ bath with a 7-ml volume. The solution was aerated with 95% oxygen plus 5% carbon dioxide. One end of the muscle strip was connected to a strain gauge transducer (TB-612T; Nihon Kohden, Tokyo, Japan) to monitor the mechanical activity. The muscle strips were stretched at 5-min intervals in increments of 0.1 g to achieve optimal resting tension. Optimal resting tension was defined as the minimum amount of stretch required to achieve the largest contractile response to 40 mM KCl and determined in our pilot experiments to be 0.7 g for the muscle strips used in these studies. The optimal resting tension was applied and allowed to equilibrate before the experiments were started. The muscle strip was illuminated alternately at 50 Hz at excitation wavelengths of 340 +/− 10 and 380 +/− 10 nm, and the amount of 500 +/− 20 nm fluorescence by 340 nm excitation (F340) and that induced by 380 nm excitation (F380) was measured. The ratio of F340to F380(R340:380) was used as an indicator of [Ca2+]i, and the quantitative comparison of [Ca2+]iwas made by taking resting and norepinephrine-stimulated [Ca2+]ias 0% and 100%, respectively.
After the equilibrating period, 0%(control), 1%, 2%, and 3% halothane or 2% and 4% isoflurane, delivered by a vaporizer, was introduced into the oxygen-carbon dioxide mixture (n = 6 each). From our pilot experiments, 10-7M norepinephrine elicited approximately 80% of the maximal norepinephrine-induced contraction both with and without halothane or isoflurane, although the maximum responses were decreased with the anesthetics. Ten minutes after bubbling of one concentration of either anesthetic the muscle strip was contracted with 10-7M norepinephrine, and the changes in muscle tension and R340:380 were regarded as the reference point (100%). Five minutes after norepinephrine administration, one of the following drugs was cumulatively administered to the organ bath to induce vasodilation: isoproterenol (10-8.5 M to 10-7M), a [small beta, Greek]-adrenoceptor agonist; forskolin (10-7.5 M to 10-6M), an adenylyl cyclase activator; or dibutyryl cAMP (DBcAMP; 10-4.5 M to 10-3M), a membrane-permeable cAMP.
The concentration of halothane or isoflurane in the gas mixture was monitored continuously by a precalibrated anesthetic agent monitor (model 303; Atom, Tokyo, Japan). The anesthetic concentrations in the bath fluid were measured by gas chromatography with flame ionization detection (GC-17A; Shimadzu, Kyoto, Japan). Gas chromatographic measurements of halothane and isoflurane in the bath fluid were 0.30 +/− 0.01 mM, 0.56 +/− 0.03 mM, and 0.90 +/− 0.03 mM for 1%, 2%, and 3% halothane, and 0.48 +/− 0.01 mM and 0.90 +/− 0.08 mM for 2% and 4% isoflurane (n = 6 each).
Radioligand-binding Receptor Studies
The rats were anesthetized with isoflurane, and the descending thoracic aorta was removed and cleaned of fat and connective tissue. The plasma membrane preparation, a modification of the method of Shaul et al., was performed at 4 [degree sign]C. Tissue was pooled from three rats for each assay. After harvesting, the aortas were placed in ice-cold 0.25 M sucrose buffer (10 times volume) with 10 mM Tris-HCl (pH 7.4), and the adventitia was dissected off. The sucrose buffer was changed, and the tissue was minced with scissors and homogenized twice with a Polytron homogenizer (Kinematica, Lucerne, Switzerland). The homogenate was centrifuged at 4 [degree sign]C at 1,500 g for 10 min, and the resulting supernatant was filtered through two layers of gauze and then centrifuged at 4 [degree sign]C at 100,000g for 30 min. The pellet formed was resuspended in 3 ml of 5 mM HEPES buffer (sodium salt) with 1.0 mM MgSO4, pH 7.4, yielding a protein concentration of 50-100 [micro sign]g/ml measured by the method described by Lowry et al. 
Binding assays were done in 5-ml glass vials that were tightly sealed with a cap to maintain the constant anesthetic concentration. Anesthetic was added to the solution in a liquid form, with a Hamilton microliter syringe. Our pilot experiments showed that the addition of 0.5 [micro sign]l halothane or 0.8 [micro sign]l isoflurane to 0.5 ml of HEPES buffer at 37 [degree sign]C produced approximately 3% halothane or 4% isoflurane and was maintained for at least 120 min. Actual anesthetic concentrations measured by gas chromatography (model GC-17A, Shimadzu) were 0.87 +/− 0.09 mM for halothane and 0.89 +/− 0.07 mM for isoflurane (n = 6 each). These are equivalent to 2.94 +/− 0.31% halothane and 3.96 +/− 0.32% isoflurane.
[small beta, Greek]-adrenoceptor binding assays were performed in duplicate using [(125) I]-iodocyanopindolol (2,000 Ci/mmol). The total binding incubation volume was 500 [micro sign]l HEPES buffer with 50 [micro sign]l [(125) I]-iodocyanopindolol dilution at concentrations ranging from 10 to 100 pM and 100 [micro sign]l membrane preparation, in the presence or absence of 10-5M l-alprenolol to distinguish nonspecific from specific binding. Incubations were maintained at 37 [degree sign]C with constant shaking for 90 min. The reaction was terminated by rapid dilution with 4 ml of 20 mM KH2PO4buffer, pH 7.4, with 1 mM MgSO4at 37 [degree sign]C, and separation of bound and free ligand by filtration through Whatman GF/C filters (Whatman International Ltd., Maidstone, UK) under vacuum, followed by three 4-ml rinses of the filter with the same buffer. Filters were counted for 1 min in a gamma counter (CompuGamma; Wallac, Turku, Finland) with a 70% counting efficiency. The effects of halothane and isoflurane were also studied under identical conditions. Each of the [(125) I]-iodocyanopindolol dilutions as standards was counted in each run under similar conditions. Specific binding was defined as total binding minus nonspecific binding. The density of receptors (Bmax) and the dissociation constant for the ligand (Kd) were determined using linear regression and the Scatchard transformation (n = 6 each).
Measurement of Cyclic Adenosine Monophosphate Concentration
Aortic strips were incubated with physiologic saline solution containing 10-7M norepinephrine for 5 min, followed by the administration of 10-7M isoproterenol or 10-6.5 M forskolin, and no addition for control, with and without the exposure of 3% halothane or 4% isoflurane (n = 6 each). The preparations were frozen quickly in liquid nitrogen and stored at -80 [degree sign]C until use. Aortic strips were homogenized in 10 volumes of Hank's balanced salt solution (without calcium and magnesium) containing 5 mM EDTA with a homogenizer (PowerHomogenizer S-203; Ikeda Scientific, Tokyo, Japan). The homogenates were centrifuged at 100g for 10 min at 4 [degree sign]C. Cyclic AMP was extracted from the supernatants by the solid-phase extraction method using Amprep SAX columns (Amersham, Buckinghamshire, UK). The concentrations of cAMP were measured by the Biotrak cAMP enzyme immunoassay system (Amersham). The resulting pellets were solubilized and used to measure protein concentrations by the method described by Lowry et al. 
Data were expressed as mean +/− SD. In smooth muscle tissue, the ratio of fluorescence to autofluorescence is much smaller than that in a single isolated cell. Because of the large amount of background fluorescence, the absolute value of [Ca2+]imay not be calculated accurately from the F340:F380ratio. To measure [Ca2+]iand muscle tension, norepinephrine-induced changes in [Ca2+]iand muscle tension 5 min after the start of contraction were used as references (100%). Changes in [Ca2+]iand muscle tension produced by isoproterenol, forskolin, or DBcAMP were expressed as a percentage value of those caused by norepinephrine-induced contraction. The effects of halothane and isoflurane on the concentration-response curves to isoproterenol, forskolin, or DBcAMP were assessed by calculating the inhibitory concentration causing 50% relaxation of the contraction to norepinephrine (IC (50)). The IC50value for each muscle strip was obtained from nonlinear regression according to the Hill Equation usingPrism software (Graphpad, San Diego, CA) and was presented as log IC50. The mean and SD were calculated from those values. Data were compared using one-way analysis of variance and post hoc analysis with Scheffe's F test. The significant value was considered < 5%.
The following drugs and chemicals were used: fura2/AM (Dojindo Laboratories, Kumamoto, Japan), Cremophor El, norepinephrine, isoproterenol, forskolin, l-alprenolol, acetylcholine chloride (Sigma Chemical Co., St. Louis, MO), and DBcAMP (Daiichi, Tokyo, Japan). Forskolin was dissolved in absolute ethanol. The final concentration of ethanol in the bath fluid was <0.1%, which did not affect either muscle contraction or fura2-Ca2+signals.
Effects of Halothane and Isoflurane on Isoproterenol-, Forskolin-, and DBcAMP-induced Relaxation
Norepinephrine elicited muscle contraction and the increment of [Ca2+]i. Halothane and isoflurane caused a 10-20% reduction of the contractile response to 10-7M norepinephrine (control, 1%, 2%, 3% halothane; and 2% and 4% isoflurane; 0.29 +/− 0.05 g, 0.26 +/− 0.04 g, 0.24 +/− 0.06 g, 0.24 +/− 0.07 g, 0.24 +/− 0.06 g, 0.23 +/− 0.06 g, respectively). Isoproterenol dose dependently reduced muscle tension and [Ca2+]iin the muscle strip precontracted with norepinephrine. Figure 1shows original recordings of fluorescence and force development induced by norepinephrine, followed by dose-dependent reduction with isoproterenol. These responses were blunted in the presence of 3% halothane.
(Figure 2and Figure 3) show concentration-response curves of halothane and isoflurane against norepinephrine-induced muscle contraction and [Ca2+]i. The effects of halothane and isoflurane on IC50for isoproterenol, forskolin, or DB-cAMP are summarized in Table 1. Halothane significantly attenuated both muscle relaxation and decreases of [Ca2+]iinduced by isoproterenol in a concentration-dependent manner. Isoflurane also attenuated responses, although the degrees of attenuation were similar between 2% and 4% isoflurane. From the IC50values, 3% halothane was more potent than 4% isoflurane in antagonizing the isoproterenol-induced relaxation, although 3% halothane and 4% isoflurane are approximately equianesthetic concentrations. In contrast, halothane and isoflurane significantly but slightly augmented the decreases in muscle tension and the changes of [Ca2+]iinduced by forskolin and DBcAMP. These results indicate that halothane and isoflurane can predominantly interfere with both muscle relaxation and decreases of [Ca2+]iat a step before adenylyl cyclase activation.
Radioligand Binding Studies
[small beta, Greek]-adrenoceptor binding experiments using [(125) I]-iodocyanopindolol in the presence and absence of l-alprenolol revealed saturable and high-affinity binding. Scatchard transformation of the data yielded linear plots from which Bmaxand Kdcould be determined. Treatment of aortic membranes with 3% halothane or 4% isoflurane had no effect on [small beta, Greek]-adrenoceptor density and antagonist affinity for iodocyanopindolol (Table 2). These results indicate that halothane and isoflurane do not interfere with binding characteristics of the [small beta, Greek]-adrenoceptors.
Effects of Halothane and Isoflurane on Isoproterenol or Forskolin-induced Cyclic Adenosine Monophosphate Production
The contents of cAMP in the vascular smooth muscle were directly measured to delineate whether the reduced [small beta, Greek]-adrenoceptor-mediated responses by halothane and isoflurane could be related to reduced cAMP production. Figure 4shows the effects of 3% halothane and 4% isoflurane on isoproterenol or forskolin-induced cAMP contents after isoproterenol or forskolin treatment. Isoproterenol increased cAMP contents in the smooth muscle, whereas preincubation of the aortic strips with halothane or isoflurane significantly inhibited these increments. In contrast, both halothane and isoflurane did not influence forskolin-induced increases in cAMP contents.
This study was designed to determine whether the volatile anesthetics halothane and isoflurane inhibit [small beta, Greek]-adrenoceptor-mediated responses, and which signal transduction steps are important for the inhibition if they do. We found that the volatile anesthetics used at clinically relevant concentrations can interfere with responses mediated by the [small beta, Greek]-adrenoceptor agonist, isoproterenol. Halothane and isoflurane impaired isoproterenol-induced vasodilation and a decline of [Ca2+]i, whereas they only slightly affected forskolin- and DBcAMP-induced vasodilation and a decrease of [Ca2+]i. This finding is consistent with the additional result that both anesthetics inhibited isoproterenol-induced, but not forskolin-induced, increases in cAMP contents. Furthermore, both agents did not interfere with the ligand-binding binding property to the [small beta, Greek]-adrenoceptor. These results indicate that halothane and isoflurane predominantly interfere with [small beta, Greek]-adrenoceptor-mediated responses at the step between the [small beta, Greek]-adrenoceptor and adenylyl cyclase; the logical site would therefore include G proteins.
There have been conflicting reports detailing the effects of anesthetics on ligand-receptor binding. Marty et al. have reported that halothane decreases [small beta, Greek]-adrenoceptor density by 10% without changing its affinity in lymphocyte membrane. In contrast, halothane had no effect on [small beta, Greek]-adrenoceptor binding to brain [23,24]and cardiac tissue. [25,26]The latter reports are consistent with our result that halothane and isoflurane had no effects on V (max) and Kdin the ligand-receptor binding study. In addition, even if a 10% reduction in the number of available receptors was elicited by volatile anesthetics, it would result in little attenuation of the [small beta, Greek]-adrenoceptor-mediated mediated responses because many spare receptors usually exist in the membranes. Therefore, the [small beta, Greek]-agonist-receptor coupling should not be a major site of inhibition by the volatile anesthetics.
The G protein is a family of membrane-bound proteins involved in the signal transduction between a receptor and an intracellular effector. Several studies have indicated that volatile anesthetics could interfere with receptor-G protein interactions in the muscarinic, [small alpha, Greek] 2-adrenergic, [small beta, Greek]-adrenergic, 5-HT1A, and adenosine adrenoceptor receptors. Volatile anesthetics inhibited receptors from being converted into the high-affinity state by G proteins. [26,27]Furthermore, Rooney et al. found that halothane, at clinically relevant concentrations, inhibited phospholipase C activity stimulated by [small beta, Greek] agonists, but not by G-protein activators, in turkey erythrocyte membranes. They suggested that halothane would modify the interaction between the activated receptor and the G protein. These previous reports support our findings that halothane and isoflurane would modulate the receptor-G-protein interactions in [small beta, Greek]-adrenoceptor-mediated responses.
Halothane and isoflurane had similar effects in the steps of the [small beta, Greek]-adrenergic signaling pathway. It is, however, possible that halothane could cause slightly greater suppression of isoproterenol-induced vasodilation than isoflurane despite the similar inhibition of cAMP production and [Ca2+]iby both anesthetics. This difference, therefore, might be attributed to the [Ca2+]i-tensionrelation rather than the cAMP-[Ca2+]ione. The difference could not be explained by the diverse effect of both anesthetics on the intracellular Ca2+storage sites, [17,31]because decreases in [Ca2+]iwere similar between the anesthetics during vasodilation.
G proteins coupled with [small beta, Greek]-adrenoceptors can act on sodium channels, calcium channels, and adenylyl cyclase in the heart or adenosine triphosphate-sensitive and Ca2+-activated potassium channels in smooth muscle. [33,34]Recent evidence suggests that G protein [small beta, Greek]-[small gamma, Greek] subunits play a major part in signal transmission, including calcium and potassium channels, which are depressed more potently by halothane than by isoflurane. Weigt et al. found that halothane and isoflurane have different effects on a G-protein-dependent and cAMP-dependent pathway activated by isoproterenol in ventricular myocytes. Thus a G-protein-dependent pathway that directly regulates ion channels could be responsible for the difference between the anesthetics observed during isoproterenol-induced vasodilation. It is possible that high concentration of isoflurane can augment a G-protein-dependent pathway, because no significant difference in muscle relaxation was observed between 2% and 4% isoflurane. The exact site(s) of the inhibition of [small beta, Greek]-adrenoceptor-mediated vasodilation by halothane and isoflurane will need to be determined by future investigations using direct G-protein activators.
An additional unexpected finding of the current study was that halothane and isoflurane might enhance vasodilation induced by forskolin or DBcAMP, both of which bypass the G-protein cascade. It is possible that halothane and isoflurane could also influence the sites after cAMP elevation. This stimulatory effect may be related to the potential anesthetic dilator properties. Volatile anesthetics are known to depress cardiac function and smooth muscle contractility, [14,39]mainly because they antagonize the L-type Ca2+channels and decrease the Ca2+sensitivity in skinned fiber preparations. Therefore, halothane and isoflurane might enhance forskolin-and DBcAMP-mediated vasodilation, probably because of the direct modulation of the Ca2+influx, the sensitivity of the contractile elements by both anesthetics, or both. However, further investigations are needed because this enhancement was small.
Halothane has diverse effects on [small beta, Greek]-adrenoceptor-G protein-adenylyl cyclase interactions. [9-13]These different effects of halothane on the [small beta, Greek]-receptor-G protein-adenylyl cyclase system may result not only from the differences of the distribution or the amount of G proteins depending on each species or tissues but also from the molecular and functional diversity of G-protein-stimulated adenylyl cyclases. 
In conclusion, we found that halothane and isoflurane, at clinically relevant concentrations, interfere with [small beta, Greek]-adrenoceptor-mediated responses in rat aortic smooth muscles. Halothane and isoflurane inhibit the [small beta, Greek]-adrenoceptor-mediated vasodilation but only slightly affect cAMP-mediated vasodilation. Furthermore, both anesthetics reduced the isoproterenol-induced, but not forskolin-induced, increases in cAMP. Halothane and isoflurane interfere with [small beta, Greek]-adrenoceptor-mediated responses at the levels occurring after agonist-receptor binding but before adenylyl cyclase activation.
The authors thank Professor Akiyoshi Namiki, M.D., Ph.D., Chairman, Department of Anesthesiology, Sapporo Medical University School of Medicine, for the direction of this research and review of the manuscript, and Paul A. Murray, Ph.D., Director, Center for Anesthesiology Research, The Cleveland Clinic Foundation, for reviewing the manuscript.