Halothane and isoflurane previously were reported to attenuate endothelium-derived relaxing factor/nitric oxide-mediated vasodilation and cyclic guanosine monophosphate (cGMP) formation in isolated rat aortic rings. Carbon monoxide has many chemical and physiologic similarities to nitric oxide. This study was designed to investigate the effects of halothane and isoflurane on carbon monoxide-induced relaxations and cGMP formation in the isolated rat aorta.
Isometric tension was recorded continuously from endothelium denuded rat aortic rings suspended in Krebs-filled organ baths. Rings precontracted with submaximal concentrations of norepinephrine were exposed to cumulative concentrations of carbon monoxide (26-176 microM). This procedure was repeated three times, with anesthetics delivered 10 min before the second procedure. Carbon monoxide responses of rings contracted with the same concentration of norepinephrine (10(-6) M and 2 x 10(-6) M) used in the anesthetic-exposed preparations also were examined. The concentrations of cGMP were determined in denuded rings using radioimmunoassay. The rings were treated with carbon monoxide (176 microM, 30 s) alone, or carbon monoxide after a 10-min incubation with halothane (0.34 mM or 0.72 mM). To determine whether the sequence of anesthetic delivery influenced results, vascular rings pretreated with halothane were compared with nonpretreated rings.
Carbon monoxide (26-176 microM) caused a dose-dependent reduction of norepinephrine-induced tension, with a maximal relaxation of 1.51 +/- 0.07 g (85 +/- 7% of norepinephrine-induced contraction). Halothane (0.34 mM and 0.72 mM) significantly attenuated the carbon monoxide-induced relaxations, but only the highest concentration of isoflurane (0.53 mM) significantly attenuated the carbon monoxide-induced relaxations. Carbon monoxide (176 microM) significantly increased cGMP content (+88.1 +/- 7.1%) and preincubation of the aortic rings with halothane (0.34 mM and 0.72 mM) inhibited this increase (-70.7 +/- 6.8% and -108.1 +/- 10.6%, respectively). When aortic rings and carbon monoxide were added simultaneously to Krebs solution equilibrated with halothane (0.72 mM), no inhibition of cGMP formation occurred.
Carbon monoxide-induced endothelium-independent relaxations of rat aortic rings were decreased by clinically relevant concentrations of halothane and isoflurane. The carbon monoxide-induced elevations of cGMP were attenuated by halothane only when the anesthetic was incubated with aortic rings before carbon monoxide treatment. The possible clinical significance of the actions of the anesthetics on this endogenous vasodilator is yet to be determined.
ENDOTHELIUM-DERIVED relaxing factor/nitric oxide is released from vascular endothelium in vivo in response to physical, chemical, and hormonal stimuli. Nitric oxide activates soluble guanylate cyclase (sGC) by combining with the heme moiety. Activated sGC converts 3',5'-guanosine triphosphate to 3',5'-cyclic guanosine monophosphate (cGMP). [1,2]Carbon monoxide is a biologically active molecule that has been reported to relax various types of vascular smooth muscle. [3–6]Its mechanism of action is thought to be similar to nitric oxide, although less potent. Like nitric oxide, carbon monoxide binds to the ferrous atom of the heme moiety, and forms a coordinate bond with sGC to activate the enzyme. Activated sGC increases cGMP, decreases cytosolic Calcium2+, and produces vasorelaxation. [4,5].
The interaction of volatile anesthetics with endogenous regulators of vascular function is an important aspect of their clinical use. Using in vitro models, many investigators have found that halothane and isoflurane attenuate endothelium-derived relaxing factor/nitric oxide-mediated vasodilation. [7–14]The mechanism responsible for these general anesthetic actions are, however, unknown. A direct inhibition of nitric oxide synthase activity has been proposed but is difficult to reconcile with experiments demonstrating the ability of halothane and isoflurane to inhibit vasorelaxation by authentic nitric oxide added to endothelium-denuded vascular rings. Halothane has been shown to decrease nitric oxide stimulation of cGMP production in vascular rings, and in a partially purified preparation of sGC from rat liver. These observations suggest that general anesthetics can inhibit nitric oxide stimulation of sGC and that this may be responsible for their inhibition of vasorelaxation. However, Blaise et al. have proposed that halothane may inactivate nitric oxide by an interaction with a halothane-derived free radical. .
Halogenated general anesthetics could inhibit nitric oxide stimulation of sGC by combining directly with nitric oxide, by preventing nitric oxide binding to the heme prosthetic group, or by an allosteric effect on the enzyme. Because carbon monoxide is more stable than nitric oxide, yet uses the same binding site as nitric oxide, it therefore becomes an ideal gas to study the interaction of anesthetics with the heme binding site in sGC. In the current study, we examined the effects of halothane and isoflurane on carbon monoxide-mediated vasorelaxation and cGMP formation in endothelium-denuded rat aortic rings.
Materials and Methods
Preparation of Isolated Rat Aortic Rings
All animal protocols were performed with the approval of the institutional review board. Male Sprague-Dawley rats (weighing 300–450 g) were anesthetized with halothane and thoracic aortas were carefully excised and placed in Krebs solution of the following composition (mM): 118.2 NaCl, 4.7 KCl, 2.5 CaCl2, 1.2 MgSO4, 1.2 KH2PQ, 25.0 NaHCO3, and 5.6 glucose (pH 7.4). The aortas were cleaned of adhering fat and connective tissue and cut into rings 3–4 mm long. The endothelium was gently removed by rotating the rings around the tip of a small forceps.
The rings were suspended in 25-ml water-jacketed organ chambers filled with Krebs solution (37 degrees C) between two stainless steel wire hangers inserted into the lumen of the aorta. The lower hanger was attached to a hook and the upper hanger to a force transducer (Grass FT 03; Quincy, MA). Isometric tension was continuously recorded on Gould Brush recorders (model RS 3400; Valley Views, OH). Four rings were equilibrated for 60 min in Krebs solution continuously aerated with 95% O (2)/5% CO2(PO2= 506+/-4 mmHg, PCO2= 23+/-0.6 mmHg). The aortic rings were set at optimal tension of approximately 1.5g as determined by previously described length-tension experiments. The Krebs solution was changed at 15-min intervals during equilibration.
After equilibration, cumulative concentrations of norepinephrine (1 nM - 10 micro Meter) were added to the baths to produce concentration-response curves. From these concentration-response curves, the norepinephrine concentration that caused 60% of the maximum norepinephrine tension for each ring (EC60) was determined. The absence of the endothelium was verified by lack of relaxation to acetylcholine (1 micro Meter). Vasoactive substances were removed from the organ baths by repeated washing with Krebs solution until a stable baseline was reestablished. At least 30 min elapsed between successive exposures to norepinephrine, during which time the rings were washed at 5-min intervals.
The protocol was aimed at comparing the effects of equivalent minimum alveolar concentrations (MAC) of halothane and isoflurane on aortic ring relaxations induced by carbon monoxide. An EC60concentration of norepinephrine was added to the endothelium-denuded rings to produce an active contraction. When a stable plateau was reached, a range of concentrations of carbon monoxide (26–176 micro Meter) was used to cause relaxation. Relaxations with carbon monoxide were expressed as a depression (grams) of the norepinephrine-induced tension. When maximal relaxation was reached, the organ baths were repeatedly washed with Krebs solution to remove vasoactive substances until a stable baseline was reestablished. This procedure was repeated three times. Anesthetics were administered 10 min before and throughout the second procedure.
Because halothane and isoflurane caused diminution of norepinephrine (5 x 10-9M) contraction, the norepinephrine concentration was increased during halothane and isoflurane administration to produce tension equal to that obtained before administration of the anesthetics. To determine whether the higher concentrations of norepinephrine influenced the effect of anesthetics on carbon monoxide relaxation, carbon monoxide responses of rings contracted with the same concentrations of norepinephrine (10-8M and 2 x 10-8M) used in the anesthetic-exposed preparations also were examined.
Halothane isoflurane were delivered by calibrated vaporizers (Foregoer DR 1, Smithtown, NY) to the oxygen/carbon dioxide mixture aerating the Krebs. The concentrations in the resulting gas mixtures were measured by an infrared analyzer (Datax, Model 254; Helsinki, Finland), which was calibrated using standard gas mixture. Bath concentrations of anesthetics in Krebs were confirmed by gas chromatography as previously described. A third norepinephrine/carbon monoxide contraction-relaxation procedure was done without anesthetics to assess the recovery from the anesthetic effect. Simultaneous time control rings were run for all experiments in an identical manner but without exposure to anesthetics. Five rings from five rats were used in the protocol.
Carbon Monoxide Preparation
A stock saturated solution of carbon monoxide was prepared by using a gas-tight syringe to inject 10 ml carbon monoxide gas into a sealed flask of 100 ml water (0 degree C), from which an aliquot of carbon monoxide solution was then withdrawn anaerobically and added to the organ bath to appropriate final concentrations. The concentration of carbon monoxide in the saturated solution (1.7 mM) and organ baths (in 37 degrees Celsius) was calculated based on its physical constants. Solutions of carbon monoxide were prepared daily.
Cyclic Guanosine Monophosphate Analysis
Two different approaches were used in the cGMP analysis. The basic difference between these two approaches was the time sequences at which the aortic rings were exposed to halothane relative to carbon monoxide treatment. In the first series, the aorta was exposed to halothane before carbon monoxide treatment, and in the second series, the aorta was not preexposed to halothane, but was added to the bath simultaneously with carbon monoxide.
In the first series, four aortic rings without endothelium were placed in separate organ baths containing 95% O2/5% CO2aerated Krebs solution at 37 degrees C for 60 min without tension. Cyclic nucleotide phosphodiesterase inhibitor 3-isobutyl-I-methylxanthine (1 mM) was added to all of the baths 10 min before and throughout the procedure. One bath was the control and nothing further was added. Carbon monoxide (176 micro Meter) alone was added to the second bath for 30 s. The last two baths were bubbled with 1 or 2 MAC halothane (0.34 or 0.72 mM) for 10 min and then the carbon monoxide (176 micro Meter) added for 30 s.
In the second series, the baths were bubbled with 2 MAC halothane (0.72 mM) for 10 min and then the 3-isobutyl-I-methylxanthine-pretreated aortic rings were added at the same time as the carbon monoxide (176 micro Meter). In this series, a ring with neither halothane or carbon monoxide treatment served as control and a ring with only carbon monoxide treatment was used to determine carbon monoxide-stimulated cGMP formation.
All tissues were frozen in liquid nitrogen for 20 s after treatment with carbon monoxide and were analyzed for cGMP content using an Amersham radioimmunoassay kit (Arlington Heights, IL). Results were expressed as pmol/mg protein.
The following chemicals were used: acetylcholine chloride, norepinephrine HCl, 3-isobutyl-I-methylxanthine (Sigma Chemical, St. Louis, MO); carbon monoxide (Aldrich Chemical, Milwaukee, WI), isoflurane (Anaquest, Madison, WI); and halothane (Halocarbon, N. Augusta, NC).
Relaxations caused by carbon monoxide are expressed as grams from the active tension produced by the EC60concentration of norepinephrine, and are designated as grams relaxation of norepinephrine contraction on the graphs. All data are expressed as means+/-standard error of means. Statistical analysis was performed by repeated-measures analysis of variance between experimental groups followed by Student-Newman-Keuls test when appropriate. A P < 0.05 was considered significant.
Effects of Anesthetics on Carbon Monoxide-induced Relaxations
Carbon monoxide (26–176 micro Meter) caused a dose-dependent reduction of norepinephrine-induced tension with a maximal relaxation of 1.51+/-0.0.7 g (85+/-7% of norepinephrine-induced contraction). Representative tracings of carbon monoxide relaxation and effects of anesthetics on carbon monoxide relaxation are shown in Figure 1and Figure 2. Halothane (0.34 mM and 0.72 mM, which are 1 and 2 MAC equivalents) treatment of the aortic rings for 10 min significantly attenuated the carbon monoxide-induced relaxations. At concentrations similar to 1 MAC halothane, isoflurane (1 MAC, 0.29 mM) reduced, but did not significantly inhibit carbon monoxide-induced relaxations of the rat aorta. However, the higher concentration of isoflurane (2 MAC, 0.53 mM) significantly attenuated carbon monoxide relaxations (Figure 3).
Both anesthetics decreased contractile responses to norepinephrine. Therefore, the concentrations of norepinephrine used to contract the aortic rings during halothane (10-8M for 1 MAC and 2 x 10-8M for 2 MAC) and isoflurane (2 x 10-8M for 1 MAC and 4 x 10-8M for 2 MAC) were increased to achieve contractions approximately equal to those of the preanesthetic group (contracted with 5 x 10-9M norepinephrine). There were no significant differences in norepinephrine-induced tensions among preanesthetic, anesthetic, and postanesthetic groups, nor with time control group. The carbon monoxide concentration-responses of aortic rings precontracted with the higher norepinephrine concentrations (10-8M and 2 x 10-8M) were compared in the absence and presence of halothane (1 and 2 MAC). Halothane also significantly inhibited relaxations induced by carbon monoxide (119 and 176 micro Meter) when these higher concentrations of norepinephrine were used for precontraction (Figure 4).
Effect of Halothane on Carbon Monoxide-mediated Cyclic Guanosine Monophosphate Production
(Table 1) shows the effects of 0.34 mM and 0.72 mM halothane on carbon monoxide-stimulated cGMP content following carbon monoxide treatment of isolated endothelium-denuded rat aortic rings. Stimulation with carbon monoxide (176 micro Meter) increased cGMP content (+ 88.1 +/-7.1%) and preincubation of the aortic rings with halothane (0.34 mM and 0.72 mM) significantly inhibited these increases (- 70.7 +/-6.8% and - 108.1+/-10.6%, respectively). Halothane does not change basal cGMP content. .
However, when the experimental protocol was changed so that the aortic rings and carbon monoxide were added simultaneously to halothane (0.72 mM) equilibrated Krebs solution, carbon monoxide-mediated increases in cGMP formation were not inhibited by halothane (Figure 5).
In the current study, carbon monoxide caused dose-dependent vascular relaxation of endothelium-denuded rat aortic rings. Carbon monoxide was less potent in inducing vascular relaxations than nitric oxide. These results are consistent with the reports of Furchgott et al. and Steinborn et al. Major new findings of this study are that halothane at low and high concentration and isoflurane at a high concentration inhibited carbon monoxide-induced relaxations in endothelium-denuded rat aortic rings. These anesthetics are more effective inhibitors of carbon monoxide-induced relaxations than previously reported nitric oxide-induced relaxations. In addition, halothane attenuated the carbon monoxide-mediated cGMP formation, and tissue pretreatment with this anesthetic agent was required for this effect.
Higher concentrations of norepinephrine were used to contract anesthetic exposed aortic rings to achieve contractions equal to those of the preanesthetic group. This was necessary because it is known that the amount of relaxation caused by a relaxant is inversely related to the amount of preexisting tension. To determine whether these higher concentrations of norepinephrine influenced the results, we compared the carbon monoxide concentration-response in aortic rings precontracted with higher concentrations of norepinephrine in the absence and in the presence of halothane (1 and 2 MAC), and found that halothane still significantly inhibited relaxation induced by carbon monoxide. In addition, the aortic rings used in the cGMP protocol were not contracted or exposed to norepinephrine and halothane significantly inhibited carbon monoxide-induced cGMP production. Thus, the effects observed are not dependent on the higher concentrations of norepinephrine used with the anesthetic-treated aortic rings.
Carbon monoxide can be generated endogenously from at least two sources. One of these is the oxidation of organic molecules and another predominant source is from heme by the action of the enzyme heme oxygenase, which cleaves heme into biliverdin and carbon monoxide. Carbon monoxide is produced in the nervous system and may function as a neurotransmitter in the brain and in nonadrenergic, noncholinergic nerves. [19,20]Inhibitors of heme oxygenase, such as zinc protoporphyrin IX, decrease cGMP concentrations in olfactory cells and increase blood pressure, whereas metabolic inducers of heme oxygenase (such as heme, hemin, and tin) have been reported to lower the mean arterial blood pressure of normotensive rats and spontaneously hypertensive rats. [21–23]Such findings support a physiologic role for carbon monoxide similar to that of nitric oxide.
Nitric oxide and carbon monoxide exert vascular effects by forming a coordinate bond with the ferrous atom of the heme moiety in sGC to activate this enzyme resulting in elevated cGMP, decreased cytosolic Calcium2+, and a decrease in vascular tension. This proposed mechanism for carbon monoxide is supported by results of Ramos et al., who reported a significant increase in the cGMP level in cultured smooth muscle cells from the rat aorta after treatment with carbon monoxide. Thus, the activation of sGC by carbon monoxide in vitro and in isolated cells is similar to that of nitric oxide. Both are accompanied by increases in intracellular cGMP.
Like nitric oxide, carbon monoxide is a simple diatomic gas that has the ability to bind to the ferrous atom of the heme moiety of the sGC. Although sharing some chemical similarities, nitric oxide has an additional electron that is readily lost to give the nitrosonium ion. For this reason, nitric oxide can undergo both oxidative and reductive reactions and therefore is more reactive than carbon monoxide. While carbon monoxide is stable in the presence of oxygen, nitric oxide reacts rapidly with oxygen to form nitrogen dioxide. Nitric oxide binds to the heme moiety of the hemoglobin molecule with an affinity 100–5,000-fold higher than carbon monoxide, and dissociates from hemoglobin at a rate that is 20-fold slower than that of carbon monoxide. .
We have previously suggested that the mechanism of the halothane and isoflurane attenuations of nitric oxide-induced relaxations may be dependent on competition of these halogenated hydrocarbons with nitric oxide for the ferrous heme site of the sGC. Halogenated hydrocarbons have been reported to have an affinity for ferrous heme proteins, and to be metabolized by a variety of heme proteins including cytochrome P-450. Halogens have high electron affinities, and all of the halogens are electrophilic. As a result, molecules containing halogens will be readily attracted to electron donating sites such as the ferrous hemoprotein. In theory, such halogenated molecules also may directly interact with the nitric oxide radical. A similar mechanism may be occurring with carbon monoxide. If the anesthetics interfere with Iron (2+)-carbon monoxide interaction, carbon monoxide would be less effective in activating sGC. However, because carbon monoxide is much more stable and nonreactive than nitric oxide, it is unlikely that the observed attenuation of carbon monoxide-mediated vasorelaxation results from a direct interaction of the anesthetics with carbon monoxide.
Halogenated anesthetics were more effective in inhibiting carbon monoxide relaxations than nitric oxide relaxations. This result is consistent with the lower affinity of carbon monoxide for hemoprotein. An alternative explanation for the current results could be that the anesthetics might bind to nonheme sites of the sGC molecule and cause an allosteric interference with carbon monoxide-heme interaction. The ability of halogenated general anesthetics and carbon monoxide to interact with heme groups favors the former explanation.
The rate of association and affinity of halothane for hemoprotein is unknown but anesthetics are likely to have a much lower affinity for hemoprotein than nitric oxide or carbon monoxide. Our finding that these agents must be preincubated with vascular tissue to elicit inhibitory activity suggests a much slower binding rate as compared with nitric oxide and carbon monoxide. Preincubation with halothane and isoflurane has been shown to interfere with the nitric oxide-mediated relaxations and cGMP increase in vascular smooth muscle, as well as the nitric oxide activation of partially purified sGC from a rat liver preparation. In the current study, we preincubated halothane with vascular smooth muscle. After a 10-min incubation period with halothane, the addition of carbon monoxide induced less vascular relaxation. Halothane also caused a dose-dependent inhibition of carbon monoxide-mediated cGMP production. Previously, we reported that halothane does not change basal cGMP content. .
Assays measuring cGMP concentrations are typically stopped within 20–30 s. This may be enough time for agents with high affinity for heme groups (nitric oxide, carbon monoxide) to modulate sGC activity but may not be enough time for weaker binding agents to exert their effects. Because of the lower affinity of anesthetics for heme proteins, it may not be possible to observe anesthetic-induced inhibition if these agents are mixed with the tissue simultaneously and assayed only for a short time. This was recently demonstrated by Zuo and Johns who used a partially purified sGC preparation from rat brain. They did not equilibrate the enzyme with the anesthetics before the addition of nitric oxide, and the anesthetics had no effect on nitric oxide-stimulated activity of sGC. In our study, when halothane was delivered to the Krebs solution for 10 min followed by simultaneous addition of vascular tissue and carbon monoxide, we did not find any significant inhibition of cGMP formation.
Alternatively, the preincubation requirement for halothane may involve the generation of a halothane metabolite, which is then active. We have not been able to detect any change of halothane concentrations in the organ bath, however, using gas chromatographic analysis. The reversibility of the halothane effect also rules out a possible covalent interaction of halothane with sGC.
In our previous studies, the magnitude of halothane inhibition of acetylcholine and nitric oxide-induced relaxations was greater than that of isoflurane. Toda et al. and Nakamura et al. reported similar findings. In the current experiments, halothane had inhibitory effects at low and high concentrations. Isoflurane only demonstrated inhibition when a high concentration was used. This is consistent with previous results with nitric oxide and also consistent with the lower affinity of isoflurane for hemoprotein. Akata et al. also reported that 1 MAC isoflurane did not inhibit sodium nitroprusside-induced relaxation in rabbit mesenteric resistance arteries.
In summary, carbon monoxide-induced endothelium-independent relaxations of rat aortic rings were decreased by clinically relevant concentrations of halothane and isoflurane. The carbon monoxide-induced increases of cGMP were attenuated by halothane only when the anesthetic was incubated with aortic rings before carbon monoxide treatment. The possible clinical significance of the actions of the anesthetics on this endogenous vasodilator is yet to be determined.
The authors thank Drs. Geoffrey S. F. Ling, Ruth Wachtel, and Russell A. Van Dyke for their comments; Ms. Ida J. Lirette for her secretarial assistance.