Interaction of Halothane with Inhibitory G-proteins in the Human Myocardium

Left ventricular myocardium from 13 terminally failing hearts (10 from men and 3 from women; age 51 yr plus/minus 4 yr) was obtained from patients after cardiectomy during cardiac transplantation. All patients gave written informed consent before the operation. Drugs used for general anesthesia were flunitrazepam, or midazolam, fentanyl, and pancuronium bromide with isoflurane. Cardiac surgery was performed on cardiopulmonary bypass during hypothermia. The cardioplegic solution used was a modified Bretschneider solution containing (in millimolar units): NaCl 15, KCl 10, MgCl²4, histidine HCl 180, tryptophan 2, mannitol 30, and potassium dihydrogen oxoglutarate 1.

Left ventricular myocardial tissue was chilled in 30 ml ice-cold homogenization buffer (10 mM tris(hydroxymethyl)aminomethane [Tris]-HCl, 1 mM ethyldiamine tetraacetic acid, and 1 mM dithiothreitol, pH 7.4). Connective tissue was trimmed away, myocardial tissue was minced with scissors and membranes were homogenized with a motor-driven glass-polytetrafluorethylene homogenizer for 1 min. The membrane preparation was then homogenized by hand with a glass-glass homogenizer. The homogenate was spun at 484g (JA 20 rotor, Beckmann, Munich, Germany) for 15 min. The supernatant was filtered through four layers of gauze, diluted with an equal volume of KCl (1 M) and stored on ice for 10 min. This suspension was centrifuged at 100,000g for 30 min. The pellet was resuspended in 50 volumes incubation buffer (adenosine 5'-triphosphate [ATP] 2.5 mM, MgCl²2.5 mM, KHCO³1 mM and Tris-HCl 2 mM, pH 7.4). and homogenized for 1 min with a glass-glass homogenizer. This suspension was recentrifuged at 100,000g for 45 min. The final pellet was resuspended in KHCO³(1 mM) and stored at -80 degrees Celsius. Storage did not alter the results.

The activity of adenylyl cyclase was determined in a reaction mixture containing 50 micro Meter [sup 32 Phosphorus]alpha-ATP (approximately 0.3 micro Ci/100 micro liter), 50 mM triethanolamine HCl, 5 mM MgCl², 100 micro Meter ethyleneglycol-bis-(beta-aminoethyl ether) tetraacetic acid, 1 mM 3-isobutyl-1-methylxanthine, 5 mM creatine phosphate, 0.4 mg/ml creatine kinase, and 0.1 mM cyclic adenosine monophosphate at pH 7.4 in a final volume of 100 micro liter. The mixture was preincubated for 5 min at 37 degrees Celsius. The reaction was started by addition of the membrane suspension (30 micro gram/100 micro liter). The incubation time was 20 min at the same temperature. Reactions were stopped by the addition of 500 micro liter of 120 mM zinc acetate. Next, the zinc acetate was neutralized by 600 micro liter Na²CO³(144 mM). After centrifugation for 5 min at 10,000g, 0.8 ml of the supernatant was applied to neutral alumina columns equilibrated with 0.1 M Tris-HCl, pH 7.5. The effluent was collected and the [sup 32 Phosphorus]cyclic adenosine monophosphate was determined by measuring radioactivity in a liquid scintillation spectrometer. Halothane was applied with carbogen for 5 min to the membrane suspension and incubated in closed vials for 20 min for measurements of adenylyl cyclase activity. Control membranes were treated with carbogen (95% O²+ 5% CO²) only in parallel and handled identically. In all experiments, 5 micro Meter guanylylimidodiphosphate (Gpp(NH)p) was present. In the forskolin experiments, guanine-nucleotides were omitted from the reaction to avoid interferences from activated G-proteins with the catalyst.

[sup 32 Phosphorus]Adenosine diphosphate ([sup 32 Phosphorus]ADP)-ribosylation of Gi alpha by pertussis toxin was performed for 12 h at 4 degrees Celsius in a volume of 50 micro liter containing Tris-HCl (100 mM) (pH 8.0 at 20 degrees Celsius), dithiothreitol (25 mM), ATP (2 mM), guanosine 5'-triphosphate (1 mM),³²Phosphorus-nicotinamide adenine dinucleotide (oxidized form) (NAD sup +) (50 nM; 800 Ci/mmol) and pertussis toxin (10 micro gram/ml) that had been activated by incubation with dithiothreitol (50 mM) for 1 h at 20 degrees Celsius before the labeling reaction. Samples were subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (10% (weight in volume) acrylamide, 16 cm total gel length). Gels were stained with Coomassie Blue and dried before autoradiography was performed. There was a linear relation between the amount of incorporated [sup 32 Phosphorus]ADP-ribose and the protein amounts loaded on the gel.

Pertussis toxin treatment was performed under the same incubation conditions as used for [sup 32 Phosphorus]ADP-ribosylation except that³²Phosphorus-NAD was replaced by 3 mM NAD in the reaction. After two washings, membranes were subjected to [sup 32 Phosphorus]ADP-ribosylation or determination of adenylyl cyclase activity. Control membranes were subjected to the same incubation conditions except that pertussis toxin was omitted from the medium. The ADP-ribosylation was quantified by using autoradiography.

Immunoblotting techniques were performed according to Gierschik et al. [7]The polyclonal antiserum MB 1 was raised in rabbits against the carboxyl-terminal decapeptide of retinal transducin alpha (KENLKDCGLF). After electrophoretic separation, proteins were transferred from the sodium dodecyl sulfate polyacrylamide gel (10%, 16 cm length) (125 mA, 12 h, Transblot, BioRad, Richmond, CA). Under these conditions, one immunoreactive Gi alpha band was detected. The sheets were immersed in 100 ml of 3% gelatin in buffer composed of Tris-HCl 20 mM and NaCl 500 mM, pH 7.5 (TBS) and shaken for 1 h at room temperature. Then they were incubated in the antibody solution (e.g., MB 1) containing 100 micro liter antiserum in 50 ml of 1% gelatin in TBS (24 h, room temperature, shaker). After two washings for 10 min with 100 ml TBS containing 0.05% Tween 20, the paper was incubated with the second antibody solution (200 micro liter peroxidase-labeled goat antirabbit immunoglobulin G [Kierkegaard and Perry Laboratories, Gaithersburg, MD]) for 2 h. After repeated washings with 0.05% Tween in TBS, the sheets were transferred to 33 mg nitro blue tetrazolium and 15 mg of 5-bromo-4-chloro-3-indolyl phosphate in 100 ml Tris-HCl (0.1 M) containing NaCl (100 mM) and MgCl²(5 mM) at pH 8.5. Color development was stopped after 10 min by rinsing with water, and nitrocellulose was dried between two sheets of filter paper. Under the conditions used, this serum was strongly reactive against transducin alpha, Gi alpha¹and Gi alpha². The amounts of Gi alpha were evaluated by densitometric scanning (Video 300 A, BioRad). To test the linearity of the immunodetection system, distinct amounts of protein were analyzed, giving a linear relation between protein amounts loaded on the gel and immunoreactive signals of Gi alpha.

Halothane was administered with a Vapor 19 (Drager CO, Lubeck, Germany) by passing the carbogen (95% O²+ 5% CO²) and directing the output through the membrane suspension. The suspended membranes were equilibrated for 5 min. Halothane was administered at 1% corresponding to 1.25 minimum alveolar concentration. Control membranes were equilibrated with halothane and handled identically.

The concentration of halothane in the membrane suspension was measured by gas chromatography using a head space analysis. Separation was done on a 60 m capillary column (RTX 1701, ID 530 micro meter).

Protein was determined according to Lowry et al. [8]using bovine serum albumin as standard. Sodium dodecyl sulfate polyacrylamide gel electrophoresis was performed as described by Laemmli. [9].

Gi alpha/Go alpha standards were provided by Professor Dr. med. P. G. Gierschik (Universitat Ulm, Germany). The second antibody was the affinity purified goat antirabbit immunoglobulin G (peroxidase labeled) from Kierkegaard and Perry Laboratories. Halothane was obtained from Hoechst AG (Frankfurt/Main, Germany). (plus/minus)-Isoproterenol HCl was obtained from Boehringer (Ingelheim, Germany). Gpp(NH)p, guanosine 5'-triphosphate, ATP, and creatine kinase were obtained from Boehringer. Pertussis toxin was obtained from Sigma (Deisenhofen, Germany) or List Biological Laboratories (Campbell, CA). Forskolin was donated by Dr. Metzger (Hoechst AG, Frankfurt, Germany). All other compounds used were of analytical or best grade commercially available. Only deionized and twice distilled water was used.

The data shown are mean plus/minus SEM. Statistical significance was estimated with Student's t test for unpaired observations and analysis of variance. A P value of less than 0.05 was considered significant.

Halothane has been shown to sensitize the myocardium toward exogenous catecholamines in isolated electrically driven preparations obtained from human hearts. [3]To test the hypothesis of whether an interaction of halothane with the G-protein adenylyl cyclase system is part of this sensitization, the influence of halothane on adenylyl cyclase activity was investigated in human cardiac membrane preparations. Halothane increased basal adenylyl cyclase activity, the efficacy of isoproterenol-stimulated adenylyl cyclase activity (Figure 1(A)), and the efficacy of the poorly hydrolyzable guanosine 5'-triphosphate analog Gpp(NH)p, which is able to stimulate G-protein alpha subunits directly (Figure 1(B)). To test whether the catecholamine-sensitizing effect of halothane is attributable to an effect on adenylyl cyclase regulating G-proteins or to a direct effect on the catalytic subunit of the enzyme, forskolin-stimulated adenylyl cyclase activity was studied alone and in the presence of halothane. As shown in Figure 1(C), halothane did not affect forskolin-stimulated adenylyl cyclase activity.

However, G-proteins have been reported to influence the forskolin-stimulated adenylyl cyclase activity in S49 lymphoma cells. [10]To study whether the effects of forskolin on adenylyl cyclase activity are a true measure of the catalyst activity under the experimental conditions chosen, experiments with manganese ions were carried out. These ions have been reported to uncouple the catalyst from the regulatory influences of G-proteins. [11,12]The effect of halothane on isoproterenol-, Gpp(NH)p-, and forskolin-stimulated adenylyl cyclase activity were studied in the presence of MnCl². As shown in Figure 2, the stimulatory actions of halothane on isoproterenol- and Gpp(NH)p-stimulated adenylyl cyclase activity were abolished. Numbers of experiments and basal values are given in the legend to Figure 2. Taken together, the unchanged effects of forskolin and the augmented effects of Gpp(NH)p in the presence of halothane strongly suggest that halothane interacts with G-proteins but has no effect on the catalytic subunit of adenylyl cyclase.

To study whether halothane affects the function of Gi alpha, adenylyl cyclase activity was determined in control and in pertussis toxin-treated membranes. As shown in Figure 3, isoproterenol- and Gpp(NH)p-stimulated adenylyl cyclase activity was augmented in control membranes, whereas the stimulation of halothane was abolished in pertussis toxin-treated membranes, [sup 32 Phosphorus]ADP-ribosylation of pertussis toxin-treated and control membranes with subsequent autoradiography revealed that about 95% of Gi alpha was covalently modified (not shown). These findings are compatible with a pertussis toxin-sensitive inhibition of Gi alpha-function in human heart by halothane.

Further experiments were carried out to specify the mechanisms of this interaction. It has been reported that halothane acts in a nonspecific way by interfering with the lipid bilayer of the membrane. [13]To study whether halothane inhibits the effect of Gi alpha by a nonspecific solubilization of the Gi alpha-protein from the membrane, a membrane-suspension was treated with carbogen plus halothane or with carbogen alone as control condition. After centrifugation, the content of Gi alpha was determined in the particulate fraction and in the supernatant. There was no difference in the content of Gi alpha between the membranes treated with halothane plus carbogen or with carbogen (Figure 4).

To test the hypothesis of whether halothane interacts with the binding of pertussis toxin to Gi alpha or interferes with the alpha-beta gamma interaction, the incorporation of ADP-ribose into Gi alpha was measured alone and in the presence of halothane. As shown in Figure 5, halothane did not influence the efficacy of ADP-ribosylation of Gi alpha. This held true whether the effect of halothane was tested on pertussis toxin catalyzed [sup 32 Phosphorus]ADP-ribosylation of native Gi alpha or whether the efficiency of pertussis toxin plus NAD treatment in the presence of halothane was tested by subsequent [sup 32 Phosphorus]ADP-ribosylation with pertussis toxin.

Halothane facilitates the ability of catecholamines to induce cardiac arrhythmias in anesthetized patients. [2]The effects of halothane on the positive inotropic effects of isoproterenol have been studied as a model for the catecholamine-sensitizing effect of the anesthetic. [3]The authors reported that halothane increases the positive inotropic effect of the catecholamine isoproterenol. This finding is in contrast to observations of Thurston and Glusman, [14]who reported decreased efficacy and potency of the effect of isoproterenol on twitch tension in left atrial myocardium obtained from rats. Lynch [15]reported that halothane did not affect the positive inotropic effect in depolarized muscles obtained from patients during surgery due to arrhythmia. The differences to the presented data may be attributable to methodologic differences, the different origins of the material studied, [16]or species differences. [17]Because isoproterenol is not a substrate for the uptake into presynaptic stores [18]the catecholamine-sensitizing effect of halothane should be located on the postsynaptic site of the cellular membrane. To characterize the effect of the hydrocarbon anesthetic on adenylyl cyclase activity, basal, isoproterenol- and Gpp(NH)p-stimulated adenylyl cyclase activity was studied. In agreement with observations on rat uterine membranes, [19]halothane was able to increase not only the isoproterenol- but also the Gpp(NH)p-stimulated adenylyl cyclase activity in human cardiac membranes. [6]In rat cardiac membranes halothane increased the basal adenylyl cyclase activity as well but had no effect on the stimulation by isoproterenol. [5]From the current observations, it appears that halothane possesses direct effects on the G-protein adenylyl cyclase complex, independent from beta-adrenoceptors.

Adenylyl cyclase activity is dually regulated by stimulatory and inhibitory G-proteins. [20,21]Therefore, halothane could activate adenylyl cyclase activity either by an increase of the catalytic subunit or by an increase of the stimulating G-protein or an impairment of the function of the inhibitory G-protein. To study the first hypothesis, the effects of the diterpene derivative forskolin on adenylyl cyclase activity alone and in the presence of halothane have been studied. Forskolin is able to stimulate the catalytic subunit of the enzyme. [10]The unchanged effect of forskolin in the presence of halothane points toward an unchanged activity of the catalyst. However, it has been reported that the effects of forskolin can be influenced by the activity of stimulatory G-protein (Gs alpha) in S49 lymphoma cells. [22]Although previous results provide evidence for an unchanged effect of forskolin in human myocardial membranes, one might speculate that other effects on inhibitory and stimulatory pathways could mask direct effects of halothane on the catalyst of adenylyl cyclase. Therefore, further experiments with manganese ions were performed. Manganese has been reported to uncouple the catalyst from the G-proteins. [12]In the presence of manganese, isoproterenol and Gpp(NH)p did not stimulate adenylyl cyclase activity alone or in the presence of halothane. These observations correlate with the observation that halothane did not affect the effect of forskolin on cardiac adenylyl cyclase activity.

Therefore, it seems clear that halothane exerts its stimulating effect on adenylyl cyclase activity by an effect on the stimulatory G-protein or a functional desensitization of inhibitory G-protein alpha subunits or an inhibition of the effects of beta gamma subunits on effectors. Hypothetical effects of halothane on Gs alpha proteins were addressed in a previous study. [6]The stimulatory effect on adenylyl cyclase was similar in S49 cyc lymphoma cells, which genetically lack Gs alpha as compared with S49 wild-type cell membranes. [6]Thus, an effect of halothane on Gs alpha function seems to be very unlikely. In addition, reconstitution of recombinant Gs alpha from transformed Escherichia coli into S49 cyc membranes did not alter the effect of halothane compared with native membranes. [6].

The alpha subunits of heterotrimeric inhibitory G-proteins are subject to the covalent modification by pertussis toxin of a cysteine at the fourth position from the carboxyl terminus. [23]This modification functional inactivates Gi alpha resulting in a decreased inhibition of the basal and receptor modulated adenylyl cyclase activity. [24,25]Treatment of cardiac membranes with pertussis toxin leads to an increased adenylyl cyclase activity. [26]The authors of the latter study suggested that treatment of myocardial membranes with pertussis toxin provides a pool for studying the Gi function. Pretreatment of human cardiac membranes with pertussis toxin increased the basal adenylyl cyclase activity in previous work [26]and in the current study. This provides evidence for a tonic Gi inhibition of adenylyl cyclase activity in human cardiac membranes. After pertussis toxin treatment, basal, isoproterenol- and Gpp(NH)p-stimulated adenylyl cyclase activities were enhanced and the effect of halothane abolished. From these data, it appears that halothane stimulates adenylyl cyclase activity by diminishing the inhibitory effect of Gi alpha. The abolished effect of the inhibitory G-protein represents a mechanism for the catecholamine-sensitizing effect of the hydrocarbon anesthetic halothane. Recently, it has been reported that a hydrophobic peptide inhibited the activation of Gi or Go by M2 muscarinic cholinergic receptor in reconstituted phospholipid vesicles. [27]Besides these peptides, hydrocarbon anesthetics seem to be another class of agents that act as potential Gi alpha-protein antagonists. In rat myocardium as well as in rat brain halothane was reported to disrupt the coupling of muscarinic serotonin 1a and adenosine A1 receptor to the inhibitory G-protein. [28]This disruption may potentially cause an increase in adenylyl cyclase activity. However, in human myocardium halothane reduced binding of quinuclidinylbencylate to muscarinergic cholinergic receptors but did not affect the receptor coupling to Gi alpha or the negative inotropic effect of carbachol. [3]The unchanged coupling of inhibitory receptors in the presence of an altered Gi alpha function may be due to the receptor G-protein stoichiometry. The number of Gi alpha proteins exceeds the number of M-cholinergic receptors in human myocardium about 1,000-fold. This may explain the differences to cat heart or brain where an interruption of receptor G-protein coupling was observed previously. [4,5,28-30].

Halothane has been suggested to act in a nonspecific way by producing interfacial dehydration of phospholipid membranes. [13]To test the hypothesis of whether pretreatment with halothane facilitates the solubilization of the inhibitory G-protein and its loss from the membrane compartment, resulting in loss of its inhibitory activity on adenylyl cyclase activity, the amount of Gi alpha in the soluble and the membrane phase before and after treatment with halothane was investigated. The unchanged "membrane-attachment" after the treatment of myocardial membranes with halothane strongly argues against this possibility.

The abolished effect of halothane on adenylyl cyclase activity in the presence of pertussis toxin may head the assumption that halothane influences the ADP-ribosylation with pertussis toxin. However, in human cardiac membranes, halothane had no effect on the ADP-ribosylation with pertussis toxin. Therefore, it is likely that the anesthetic does not interact with the binding of pertussis toxin at the alpha subunit of the G-proteins. Isolated beta gamma subunits augment the [sup 32 Phosphorus]ADP-ribosylation with pertussis toxin, [31]because the heterotrimeric guanosine diphosphate-bound alpha-beta gamma complex is the best substrate for the ADP-ribosylation with pertussis toxin. [32]The unchanged ADP-ribosylation in the presence of halothane argue that halothane may not affect the interaction between alpha and beta gamma subunits of the inhibitory G-proteins.

We conclude that halothane exerts its catecholamine-sensitizing effect by interference with the inhibitory G-protein followed by stimulation of adenylyl cyclase activity. The mechanism of action of halothane appears to be an effect of halothane on the G-protein alpha subunits or beta gamma subunits activating effectors, whereas an effect of alpha-beta gamma subunit interaction or the membrane attachment of the G-protein alpha subunits seems not play a role. These findings raise conceptual perspectives on pharmacologic intervention to modulate G-protein function in certain pathologic states, such as heart failure [33]and autonomous thyroid adenoma. [34].

The authors thank Elisabeth Ronft and Judith Sabo for their excellent technical assistance. They thank colleagues from the Herzchirurgische Klinik (Director, Professor Dr. B. Reichart) for their support in using human myocardium. They thank Professor M. v. Meyer from the Department of Forensic Medicine for measuring the concentration of halothane in the membrane preparation.