Introduction

Summary Introduction Procedures Results Discussion References

A wide variety of chemically diverse compounds, including barbiturates, steroids, propofol, alcohols, and inhalation agents such as halothane, isoflurane, and enflurane, induce general anesthesia in animals and humans (1). Biochemical and electrophysiological studies have shown that all of these drugs potentiate the inhibitory signals mediated by GABA_A receptors in the brain and that their potencies and efficacies in exerting this action correlate with their abilities to induce anesthesia (2-7). These observations support the notion that GABA_A receptors play an important role in anesthesia, although the precise molecular mechanism of action of general anesthetics at GABA_A receptors remains to be fully elucidated.

In addition to facilitating the action of GABA at GABA_A receptors, general anesthetics elicit a GABA-like direct effect that can be detected by an increase in Cl channel permeability or ³⁶Cl uptake into brain membrane vesicles in the absence of GABA (8-13). Although this effect occurs at pharmacologically relevant concentrations, they are generally higher than those required to potentiate GABA responses.

Molecular cloning studies have identified a variety of GABA_A receptor subunits and demonstrated receptor population heterogeneity (14), suggesting that the actions of anesthetics, as well as those of other GABAergic modulators, may be influenced by receptor subunit composition. Indeed, Lin et al. (15) showed that enflurane potentiated GABA-induced currents to a greater extent at 11 than at 112 receptors, indicating that the degree of potentiation produced by enflurane can be altered by the 2 subunit and that at variance with benzodiazepines (16, 17), the modulatory effect does not require this subunit. Similar results have been obtained with other anesthetics, such as isoflurane (18), pentobarbital (19), and propofol (20). We have previously shown that both propofol and pentobarbital activate currents directly and potentiate GABA-induced currents in Xenopus laevis oocytes expressing 1 homomeric receptors (13). In addition, Cestari et al. (21) reported that pentobarbital directly activates Cl currents at homomeric receptors composed of murine 3, but not of 2, subunits expressed in oocytes; chimeras of the two  isoforms revealed that the difference in responsiveness is attributable to a three-amino acid difference in the amino-terminal domains of the two subunits. The influence of  subunits was also demonstrated by Harris et al. (22), who showed that positive modulation of GABA_A receptors by the injectable anesthetics alphaxalone and pentobarbital, but not by the volatile anesthetics isoflurane and enflurane, strictly depended on the coexpression of 2 or 3 subunits in transfected cell lines.

We have now investigated further the influence of  subunit isoforms (1, 2, or 3), coexpressed with human 1 and 2S subunits in X. laevis oocytes, on the actions of the anesthetics alphaxalone and etomidate. Because we previously showed that alphaxalone does not activate 1 homomeric receptors directly (13), we predicted that this steroid derivative might be representative of a class of anesthetics that act at sites localized on subunits other than the  subunit. On the other hand, preliminary studies in our and other laboratories (23, 24) have shown that both direct and modulatory effects of etomidate are dependent on the type of subunit isoform expressed. In addition, by expressing an array of homomeric and dimeric receptor combinations, we attempted to determine whether specific subunits are required for the actions of these anesthetics.

	Experimental Procedures

Summary Introduction Procedures Results Discussion References

Materials. Adult X. laevis females were obtained from Dipl.Biol.-Dipl.Ing. Horst Kähler (Hamburg, Germany). Alphaxalone and etomidate were kindly provided by Glaxo Group Research (Greenford, UK) and Janssen Pharmaceutica (Beerse, Belgium), respectively; stock solutions (10 mM) were prepared in dimethylsulfoxide and stored at 20° until use. SR 95531 was obtained from Research Biochemicals International (Natick, MA), and GABA, picrotoxin, and other reagents of analytical grade were from Sigma Chemical (St. Louis, MO).

Preparation of cDNAs. The cDNAs encoding the human 1, 1, 2, 3, and 2S GABA_A receptor subunits were subcloned into the pCDM8 expression vector (InVitrogen, San Diego, CA) (25). Plasmids were purified with the Promega Wizard Plus Miniprep DNA Purification System (Madison, WI) and then resuspended in sterile distilled water, divided into portions, and stored at 20° until used for injection.

Microinjection of and electrophysiological recording from X. laevis oocytes. Oocyte isolation and cDNA microinjection were performed essentially as previously described (20). Isolated oocytes were placed in modified Barth's saline [containing 88 mM NaCl, 1 mM KCl, 10 mM HEPES-NaOH, pH 7.5, 0.82 mM MgSO₄, 2.4 mM NaHCO₃, 0.91 mM CaCl₂, and 0.33 mM Ca(NO₃)₂]. Various mixtures of GABA_A receptor subunit cDNAs (1.5 ng of each in a total volume of 30 nl) were injected into the nucleus of oocytes according to the "blind" method. The injected oocytes were cultured at 19° in sterile modified Barth's saline supplemented with 10 µg/ml streptomycin, 10 units/ml penicillin, 50 µg/ml gentamicin, 0.5 mM theophylline, and 2 mM sodium pyruvate. Recordings were obtained 1-4 days after injection from oocytes placed in a 100-µl rectangular chamber. The animal pole of oocytes was impaled with two glass electrodes (0.5-3 M) filled with 3 M filtered KCl, and the voltage was clamped at 70 mV with an Axoclamp 2-B amplifier (Axon Instruments, Burlingame, CA). Currents were continuously recorded on a strip-chart recorder. Resting membrane potential usually varied between 30 and 50 mV. Drugs were perfused for 20 sec unless otherwise noted. Intervals of 5 min were allowed between applications of low concentrations of GABA alone and of 10 min when high concentrations of GABA or anesthetics were applied.

Statistical analysis. Currents were expressed as a percentage of the control response (in nA) obtained with GABA alone. A GABA control response was obtained before and after each drug application to take into account possible shifts in the control currents. Oocytes from at least two frogs were used for each experiment, and the total number of oocytes is given. Data are presented as mean ± standard error and were analyzed by Student's t test or by one- or two-way analysis of variance followed by Scheffé's post hoc test. p < 0.05 was considered statistically significant.

	Results

Summary Introduction Procedures Results Discussion References

Direct effects of alphaxalone and etomidate at GABA_A receptors. As expected, alphaxalone and etomidate induced inward Cl currents in the absence of GABA in oocytes expressing human GABA_A receptors composed of 122S subunits (Fig. 1). Responses to both anesthetics were concentration dependent and reversible; a typical slow current decay was observed, which was most marked at the highest concentrations of these drugs. Cl currents induced by either anesthetic were inhibited by the coapplication of either the GABA competitive antagonist SR 95531 or the Cl channel blocker picrotoxin (Fig. 2); total inhibition was apparent at a 25 µM concentration of either SR 95531 or picrotoxin (data not shown).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Direct activation of human recombinant GABAA receptors by alphaxalone (ALPX) (A) and etomidate (ETO) (B). Tracings were obtained from single oocytes expressing 122S receptors and represent Cl currents induced by the two anesthetics in the absence of GABA compared with the response to 10 mM GABA. Horizontal bar above each response, drug application (20 sec).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Effects of SR 95531 and picrotoxin on GABAA receptor-mediated Cl currents induced by alphaxalone (ALPX) (A) and etomidate (ETO) (B). Tracings were obtained from single oocytes expressing 122S receptors and represent the inhibition of the responses to the two anesthetics by 5 µM SR 95531 (SR) and 5 µM picrotoxin (PICTX). Drugs were perfused for 20 sec.

Role of  subunits in the direct actions of alphaxalone and etomidate. To evaluate the influence of  subunits on the direct activation of the receptor-associated Cl conductance by alphaxalone and etomidate, we injected oocytes with cDNAs encoding 1 and 2S subunits together with those encoding 1, 2, or 3 subunits. Alphaxalone activated Cl currents at all receptors with similar efficacies (Fig. 3A), although the maximal effect tended to be greatest at receptors containing the 1 subunit and smallest at those containing the 3 subunit (Table 1). In addition, the potency of alphaxalone at 1-containing GABA_A receptors was approximately twice that at the other two receptor subtypes (Table 1).

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Direct effects of alphaxalone (A) and etomidate (B) at GABAA receptor assemblies containing different  subunit isoforms. Evoked Cl currents were in response to various concentrations of the two anesthetics measured from oocytes expressing 112S (), 122S (), or 132S () receptors. Values are expressed as mean ± standard error percentage of the control response obtained with 10 mM GABA (from five to seven oocytes).

View this table: [in this window] [in a new window]   TABLE 1 Direct activation of CI currents by alphaxalone and etomidate in X. laevis oocytes expressing recombinant GABAA receptors Maximal direct activation of CI current is expressed as a percentage of the control response obtained with 10 mM GABA; values are mean ± standard error for the indicated number (n) of oocytes.

Direct actions of alphaxalone and etomidate at homomeric and dimeric GABA_A receptors. To investigate whether specific subunits of the GABA_A receptor are required for the direct activation of Cl currents by the two anesthetics, we compared the actions of these compounds at 1 homomeric receptors and at receptors formed from two or three different subunits.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Direct actions of alphaxalone (A) and etomidate (B) at homomeric, dimeric, and trimeric GABAA receptor assemblies. Values represent the Cl currents induced by the two anesthetics in the absence of GABA and are expressed as mean ± standard error percentage of the control response obtained with 10 mM GABA (from 5-10 oocytes).

Role of  subunits in the modulation of GABA-induced currents by alphaxalone and etomidate. Both anesthetics have previously been shown to enhance markedly the action of GABA at GABA_A receptors (9, 10, 26-28). Concentration-response curves for the modulatory effect of alphaxalone (0.1-100 µM) on Cl currents induced by GABA (20% of maximally effective concentration) revealed similar efficacies at 112S, 122S, and 132S receptor subtypes (Fig. 5A, Table 2). However, the potency of alphaxalone at receptors containing the 2 subunit was approximately two to three times that at the other two subunit assemblies.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Modulatory actions of alphaxalone (A) and etomidate (B) at GABAA receptor assemblies containing different  subunit isoforms. Values are expressed as mean ± standard error percentage of the potentiation of the control response to GABA (EC20) by various concentrations of the two anesthetics measured in oocytes expressing 112S (), 122S (), or 132S () receptors (from six to nine oocytes). Actual EC20 concentrations of GABA for the different receptor combinations were determined experimentally for each oocyte: 112S, 5-15 µM (mean, 9.3 ± 0.9 µM); 122S, 2-10 µM (mean, 5.4 ± 0.9 µM); and 132S, 1-10 µM (mean, 3.5 ± 0.6 µM).

View this table: [in this window] [in a new window]   TABLE 2 Modulation by alphaxalone and etomidate of GABA-induced CI currents in X. laevis oocytes expressing recombinant human GABAA receptors Maximal potentiation is expressed as the percentage increase in the current induced by GABA at EC20. Data are mean ± standard error for the indicated number (n) of oocytes.

Modulation of homomeric and dimeric GABA_A receptors by alphaxalone and etomidate. Alphaxalone and etomidate each potentiated Cl currents evoked by GABA (20% of maximally effective concentration) to a similar extent in oocytes expressing 1, 11, 12S, 12S, or 112S receptors (Fig. 6). Thus, no single subunit or subunit combination was required for the modulatory action of these anesthetics.

View larger version (30K): [in this window] [in a new window]   Fig. 6.   Modulatory actions of alphaxalone (A) and etomidate (B) at homomeric, dimeric, and trimeric GABAA receptor assemblies. Values represent the potentiation of Cl currents induced by GABA by the two anesthetics and are expressed as mean ± standard error percentage increase in the control response obtained with GABA (EC20) (from 5-10 oocytes). EC20 concentrations of GABA for the different receptor combinations were determined experimentally for each oocyte: 1, 15-32 µM (mean, 22.3 ± 2 µM); 11, 0.5-6 µM (mean, 2.1 ± 0.4 µM); 12S, 10-30 µM (mean, 16.8 ± 2 µM); 12S, 5-20 µM (mean, 8.6 ± 2 µM); and 112S, 5-15 µM (mean, 9.3 ± 0.9 µM).

	Discussion

Summary Introduction Procedures Results Discussion References

Recent studies have suggested that the  subunit of the GABA_A receptor may be required for the action of certain anesthetics (13, 20-22). Thus, we focused our attention on the role of this subunit in both the GABA-mimetic and modulatory actions of alphaxalone and etomidate at recombinant human GABA_A receptors expressed in X. laevis oocytes. Our results show that the two anesthetics differ markedly in the subunit dependence, especially the  subunit sensitivity, of their multiple actions.

Alphaxalone and etomidate have previously been shown to exert a GABA-mimetic action in the absence of GABA (9, 10, 13, 29). Here, we demonstrate that in addition to 1x2S receptors, alphaxalone activates receptors formed by 1 and 2S subunits. In contrast, even high concentrations of this anesthetic exert only a weak activating effect at 1 homomeric receptors, as previously shown (13). These results suggest that this steroid anesthetic interacts preferentially with 1 and 2S subunits but not with the 1 subunit. In contrast, Puia et al. (27) showed that human 1 homomeric receptors expressed in human embryonic kidney 293 cells were directly activated by the steroid derivatives 3,21-dihydroxy-5-pregnan-20-one and 3-hydroxy-5-pregnan-20-one. Differences in potency or efficacy between these endogenous steroids and alphaxalone may account for this apparent discrepancy. In addition, because absolute values for steroid-induced currents were provided in the previous study, whereas we expressed alphaxalone-induced currents as a percentage of the maximal GABA response in the current study, the results of the two studies are not readily compared.

Coexpression of either 1 or 2S subunits with the 1 subunit restored sensitivity of the resulting receptors to alphaxalone. However, currents elicited by this anesthetic were lower at 11, 12S, and 112S receptor constructs than at those composed of only 1 and 2S subunits. These observations suggest that in addition to generating functional GABA-sensitive Cl channels (30), both 1 and 2S subunits contribute to the formation of a sensitive site for the interaction of alphaxalone and that the presence of the 1 subunit reduces the sensitivity of the receptor to this anesthetic. Furthermore, alphaxalone sensitivity seems to require the presence of at least two different types of subunits, one of which must be either 1 or 2S. The notion that the  subunit may not be important in the direct action of alphaxalone is also supported by the observation that the anesthetic showed similar efficacies at trimeric receptors containing different  subunit isoforms, although its potency at 112S receptors was twice that at receptors containing the 2 or 3 subunit.

Etomidate differed from alphaxalone in that its direct action at GABA_A receptors required the  subunit. Thus, etomidate activated 1 homomeric receptors with an efficacy higher than that of GABA itself, but it had no effect at 12S assemblies. In addition, coexpression of 1, but not of 2S, markedly reduced the sensitivity of the receptor to etomidate. These observations suggest that unlike alphaxalone, etomidate may interact directly with the 1 subunit and that the site of action on this subunit is affected in a negative manner by the presence of the 1 subunit.

In contrast to alphaxalone, the importance of the  subunit in the GABA-mimetic action of etomidate is further supported by the observation that within the range of concentrations tested, the efficacy of this anesthetic depended on which  subunit isoform was expressed together with the 1 and 2S subunits, with the rank order 3 > 2 > 1. It should be noted, however, that the concentration-response curve for etomidate at 11S receptors does not reach clear saturation, and therefore the actual efficacy of this compound may not be accurately determined. A similar influence of the  subunit has been demonstrated for the direct action of pentobarbital (31). Together, these observations indicate that alphaxalone and etomidate directly activate GABA_A receptors by interacting at sites localized on different subunits: the  subunit for etomidate and 1 or 2S subunits for alphaxalone. Together with the results of previous studies (13, 20), the current data indicate that with regard to  subunit specificity, the action of etomidate, but not that of alphaxalone, is similar to that of the general anesthetics propofol and pentobarbital.

Despite the fact that alphaxalone and etomidate seem to act at different subunits of the GABA_A receptor, the direct effects of both drugs were blocked by the GABA competitive antagonist SR 95531. This observation is consistent with previous studies showing that the direct actions of these as well as other anesthetics are inhibited by the GABA competitive antagonist bicuculline (8, 9, 12, 20, 26). Given that general anesthetics are thought to be allosteric modulators of GABA_A receptors (1, 5), it is possible that they may induce a conformational change in the heteromeric receptor complex that encompasses the GABA binding site, an event that can be allosterically inhibited by a competitive antagonist. In contrast, Thompson et al. (31) reported that SR 95531 failed to block Cl currents induced by pentobarbital at GABA_A receptors expressed in oocytes. The reason for this apparent discrepancy is not clear.

As shown previously with both native and recombinant GABA_A receptors (1, 5), both alphaxalone and etomidate markedly potentiate the action of GABA. In contrast to the subunit specificity of the direct actions of alphaxalone and etomidate, potentiation of the GABA effect by these anesthetics did not require any specific subunit. Indeed, as shown previously (13, 27), alphaxalone enhanced GABA-evoked Cl currents at 1 homomeric receptors, at which the same compound failed to exert a direct effect. Similarly, etomidate potentiated the action of GABA at all receptors tested, including 12S receptors, which were insensitive to direct activation by this drug. Thus, it seems that the multiple actions of these anesthetics are mediated by different binding sites: one on the  subunit or on the 12S subunits for the direct effect of etomidate and alphaxalone, respectively, and a second, which is present in all receptors, for potentiation of GABA action. The fact that alphaxalone potentiates the action of GABA at 1 homomeric receptors but fails to activate directly these same ion channels suggests that the interaction of GABA with its recognition site may unmask the allosteric site responsible for the potentiating effect of this anesthetic. A similar scenario may account for the action of etomidate at 12S receptors. Multiple sites of action on GABA_A receptors have also been suggested for propofol and pentobarbital (20, 31). Such multiple sites of action may differ in affinity, as suggested by the fact that higher concentrations of anesthetics are required for direct activation than for modulatory action. The efficacy of the modulatory action of etomidate, but not that of alphaxalone, at trimeric receptors depended, as determined with the range of anesthetic concentrations tested, on the specific  subunit isoforms present, with the rank order 3 > 2  1, which is consistent with the notion that these anesthetics influence GABA_A receptor function by acting at different modulatory sites.

The relatively low efficacies and potencies of etomidate with regard to both direct and modulatory effects at 1-containing receptors are consistent with the subunit specificity of its analog loreclezole, an anticonvulsant compound devoid of anesthetic properties (32) that interacts selectively with a site located on the subunit of the GABA_A receptor (33, 34); the affinity of loreclezole for receptors containing 2 or 3 subunits is ~300-fold that for 1-containing receptors. The physiological and pharmacological consequences of the marked difference between alphaxalone and etomidate with regard to subunit specificity for their actions at GABA_A receptors, especially at 1 homomeric and dimeric receptors, are questionable in view of the fact that such subunit assemblies are unlikely to be expressed as such in neurons (35). However, the expression of single- or double-subunit combinations in X. laevis oocytes represents a useful model for evaluation of the role of subunits in the action of general anesthetics. Because etomidate, propofol, and pentobarbital activate 1 homomeric receptors, which are regarded as an ancestral form of GABA receptor (36), sensitivity to these compounds and their physiological counterparts probably evolved early during phylogenesis, whereas sensitivity to steroids may have arisen with the appearance of the  and 2 subunits.