The intravenous anesthetic etomidate is optically active and exists in two mirror-image enantiomeric forms. However, although the R(+) isomer is used as a clinical anesthetic, quantitative information on the relative potencies of the R(+) and S(-) isomers is lacking. These data could be used to test the relevance of putative molecular targets.
The anesthetic concentrations for a half-maximal effect (EC50) needed to induce a loss of righting reflex in tadpoles (Rana temporaria) were determined for both etomidate enantiomers. The effects of the isomers on gamma-aminobutyric acid (GABA)-induced currents in stably transfected mouse fibroblast cells was also investigated using the patch-clamp technique. In addition, the effects of the isomers on a lipid chain-melting phase transition were determined.
The EC50 concentrations for general anesthesia for the R(+) and S(-) isomers were 3.4 +/- 0.1 microM and 57 +/- 1 microM, with slopes of n = 1.9 +/- 0.1 and n = 2.9 +/- 0.2, respectively. The R(+) isomer was also much more effective than the S(-) isomer at potentiating GABA-induced currents, although the degree of stereoselectivity varied with anesthetic concentration. R(+) etomidate potentiated the GABA-induced currents by increasing the apparent affinity of GABA for its receptor. Both isomers were equally effective at disrupting lipid bilayers.
These data are consistent with the idea that the GABA(A) receptor plays a central role in the actions of etomidate. Etomidate exerts its effects on the receptor by binding directly to a specific site or sites on the protein and allosterically enhancing the apparent affinity of GABA for its receptor.
MOST general anesthetics in current clinical use are optically active. They can exist as pairs of mirror-image enantiomers that are almost invariably administered as racemic mixtures (a 50:50 mixture of the two enantiomers). Apart from their ability to rotate plane-polarized light in opposite directions, optical isomers have identical physical properties. Nonetheless, for most clinical general anesthetics that have been studied, the mirror-image enantiomers have been found to have different anesthetic potencies in animals. The observed degree of stereoselectivity varies considerably among agents, at about a factor of only 1.5 or less for isoflurane, [1,2]2–4 for barbiturates and ketamine, and up to a factor of 10 or more for anesthetic steroids. Although the increasing degree of stereoselectivity for the more potent anesthetics might be anticipated, what is more surprising is that the extent of stereoselectivity observed for these anesthetics in in vitro preparations is usually comparable with that observed in animals. [5,7–11]These observations support the suggestion that anesthetics may exert their principal effects at a relatively small number of genetically unrelated targets, because, if many different targets were involved, it would be difficult to account for the extent of the observed in vivo stereoselectivities. For the most potent general anesthetics, which might be expected to display the greatest degree of stereoselectivity, the use of optical isomers should provide a particularly powerful approach to identifying those molecular targets that are likely to be most relevant.
The carboxylated imidazole etomidate (Figure 1) is a potent intravenous agent with a high therapeutic index. It is unique among current general anesthetics in being the only optically active agent that is administered in an optically pure form (the R[+] isomer). Although the S(-) isomer is said to be “devoid of hypnotic activity,”there is no published estimate of its anesthetic activity in animals and no comprehensive study of the relative effects of the two anesthetic enantiomers on its most plausible molecular target, the gamma-aminobutyric acidA(GABAA) receptor. [7,15,16]In this article we describe experiments that compare the effects of the two optical isomers of etomidate on animals, on a defined GABAAreceptor, and on lipid bilayers.
Materials and Methods
Anesthetic Potency Measurements in Tadpoles
The general anesthetic potencies of R(+) and S(-) etomidate were determined for 4–6-week-old Rana temporaria tadpoles (Blades Biological Ltd., Cowden, Kent, UK) in the pre-limb-bud stage of development (average length, about 2 cm). Tadpoles were maintained in an aerated aquarium at 20–22 [degree sign] Celsius. During randomized blind anesthetic potency experiments, about 10 tadpoles were placed in each of several beakers containing 300 ml tap water (20 +/- 1 [degree sign] Celsius) and 58 mM propylene glycol with or without etomidate. (Propylene glycol on its own at concentrations up to 174 mM was tested and found to have no observable effect on the tadpoles.) The anesthetic endpoint was defined as the lack of a purposeful and sustained swimming response after a gentle inversion with a smooth glass rod. The number of anesthetized tadpoles was recorded every 10 min for 120 min (equilibrium was complete after 10–20 min), after which the tadpoles were returned to fresh tap water where their recovery was monitored. In most cases, normal swimming activity was restored within 30 min. Tadpoles that did not recover were excluded from the analysis. Tadpole concentration-response data were fitted according to the method of Waud to a logistic equation of the form Equation 1where p is the percentage of the population anesthetized, C is the etomidate concentration, n is the slope, and EC50is the etomidate concentration for a half-maximal effect.
Preparation and Culture of PA3 Cells
Mouse fibroblast L cells stably transfected with bovine alpha sub 1 beta1gamma2Lsubunits of the GABAAreceptor were supplied by P. Whiting (Merck Sharp and Dohme Research Laboratories, Harlow, Essex, UK) at passage number 10. In these PA3 cells, expression of GABAAreceptor mRNAs is under the control of a hormone-sensitive promoter. Little or no expression occurs without the addition of the hormone dexamethasone to the culture medium. PA3 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10%(vol/vol) heat-inactivated fetal bovine serum, L-glutamine (2 mM), penicillin (100 units/ml), and streptomycin (100 micro gram/ml), all obtained from Gibco Life Technologies, Paisley, UK. Cells were maintained in an atmosphere of 95% air and 5% carbon dioxide at 37 [degree sign] Celsius. The PA3 cells were passaged every week using Cell Dissociation Solution (Sigma, Poole, Dorset, UK) and a passage ratio of 1:10. As the passaged cells approached confluence (usually after 4 days), they were exposed to culture medium containing the antibiotic geneticin (2 mg/ml) to select for cells with geneticin resistance. For the electrophysiologic experiments, PA3 cells were grown on glass coverslips (6 x 18 mm) in culture medium supplemented with 1 micro Meter dexamethasone and incubated for 4 days before recording. PA3 cells with passage numbers ranging from 12–34 were used for these experiments, with no noticeable differences in the levels of expression.
Recording Technique for PA3 Cells
Inward currents, a result of the efflux of chloride ions, evoked by the application of GABA to PA3 cells were recorded using the standard whole-cell patch-clamp technique with an Axopatch 200 amplifier (Axon Instruments, Foster City, CA). Recording pipettes were fabricated from thin-walled filamented borosilicate glass capillary tubes (GC150TF; Clark Electromedical Instruments, Reading, Berkshire, UK) using a two-stage pull (Narishige PB-7 micropipette puller, Tokyo, Japan) and lightly fire-polished to give typical electrode resistances of 5 M Omega. Pipettes filled with intracellular recording solution containing 130 mM CsCl, 1 mM MgCl2, 10 mM HEPES, and 11 mM EGTA (titrated to a pH of 7.2 with CsOH) easily formed “giga-ohm” seals with the cells. Once the whole-cell configuration was established, the cells were voltage clamped at -40 mV. During recording, the PA3 cells were bathed in an extracellular solution containing 124 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 5 mM HEPES, and 11 mM D-glucose (titrated to a pH of 7.4 with NaOH). Membrane currents were filtered at 50 Hz (-3 dB; eight-pole Bessel filter, FE-301-SF; Fylde Electronic Laboratories, Preston, UK). Membrane voltages were compensated for series resistance by > 80% for all recordings. All drugs were applied rapidly via a double-barreled glass capillary tube as described previously. Experiments were done at room temperature (20–23 [degree sign] Celsius).
Apart from the determination of GABA concentration-response curves, most experiments were done with a low, non-desensitizing control concentration of GABA (typically 1 micro Meter, which induces approximately 1% of the maximal current). For experiments with anesthetic, pairs of stable GABA-activated control currents were recorded before and after application of the anesthetic and averaged to give a control response. R(+) and S(-) etomidate were obtained in pure form (a gift from Janssen Research Foundation, Beerse, Belgium) and the R(+) enantiomer was also obtained as Hypnomidate (2 mg/ml R[+] etomidate in 35% propylene glycol and water) from Janssen-Cilag Ltd. (High Wycombe, Buckinghamshire, UK). Stock etomidate solutions were prepared with the same proportion of propylene glycol as that used in the commercial formulation. In the electro-physiologic experiments, propylene glycol concentrations varied from 0–174 mM. However, propylene glycol alone was found to have only a small effect on the GABA response (< 10% inhibition at 174 mM propylene glycol, the highest concentration used). Nonetheless, propylene glycol was always present at the same concentrations in the control and test solutions. At concentrations (> 10 micro Meter) at which R(+) etomidate was found to have a significant GABA mimetic effect, the anesthetic was preapplied before the application of GABA to establish an accurate baseline. (The current activated by 100 micro Meter R[+] etomidate was inhibited by 94 +/- 5% in the presence of 10 micro Meter bicuculline methiodide; n = 6.) However at lower doses, where there was no measurable mimetic effect of R(+) etomidate, the anesthetic and GABA were applied simultaneously. For experiments with the optical isomers, the order of application of the isomers was varied from cell to cell. Unless otherwise stated, all chemicals were obtained from Sigma Chemical Company.
The GABA concentration-response data were fitted (unweighted least-squares) to a Hill equation of the form Equation 2where E is the peak GABA-induced current expressed as a percentage of the maximal current, A is the GABA concentration, nHis the Hill coefficient, and EC50is the GABA concentration for a half-maximal effect.
Anesthetic potentiation (%) was defined as 100 x (I - I0)/I0where I0is the peak of the control GABA-induced current and I is the peak of the GABA-induced current in the presence of anesthetic. Concentration-response data were fitted (unweighted least-squares) to a Hill equation of the form Equation 3where y is the percentage potentiation, ymaxis the maximal percentage potentiation, C is the etomidate concentration, nHis the Hill coefficient, and EC50is the etomidate concentration for a half-maximal effect.
Values throughout the article are given as means +/- SEMs. Statistical significance was assessed using the Student's t test. 
Lipid Phase-Transition Measurements
The effects of the etomidate enantiomers on the main chain-melting phase transition of dipalmitoyl-L-alpha-phosphatidylcholine (DPL; Sigma) was determined using a method similar to that described previously. An aqueous suspension of DPL vesicles was prepared in buffered saline by adding 20 micro liter of a stock suspension to 1.9 ml of extracellular solution (with or without etomidate) in a 4-ml cuvette. The stock suspension of DPL (30 mg/ml) was prepared by vigorously mixing dried lipid with buffer at a temperature (50 [degree sign] Celsius) well above the chain-melting phase transition (41–42 [degree sign] Celsius). To facilitate the solubilization of etomidate, a final concentration of 94 mM ethanol was used. An identical concentration of ethanol was present in the control solutions. The cuvette was placed in the heated stage of a Beckman DU 650 spectrophotometer (Beckman Instruments Ltd., Fullerton, CA) and the absorbance at 450 nm was measured as the temperature was increased at the rate of about 2 [degree sign] Celsius/min. The temperature was monitored using a thermocouple placed in the cuvette, and the data output from the spectrophotometer and a digital thermometer were fed into a computer. The transition temperature was defined as the midpoint in the abrupt change in absorbance.
We investigated the relative potencies of the etomidate enantiomers in producing a loss-of-righting reflex in tadpoles. We found that concentrations of R(+) etomidate in excess of about 1 micro Meter rapidly ([approximately] 15 min) induced loss-of-righting reflex and that the concentration-response curve could be fitted by a logistic equation (see Materials and Methods, Equation 1) with an EC50of 3.4 +/- 0.1 micro Meter R(+) etomidate and a slope n = 1.9 +/- 0.1 (Figure 2). Considerably higher concentrations of the S(-) isomer were required to have a similar effect, and even at 120 micro Meter, the highest concentration tested, < 80% of the tadpoles were anesthetized. These data were also fitted to a logistic equation and yielded an EC50of 57 +/- 1 micro Meter S(-) etomidate and a slope of n = 2.9 +/- 0.2.
In the concentration range of 0.3–300 micro Meter, GABA induced whole-cell currents in the PA3 cells (at a membrane potential of -40 mV) of 10 pA to 11 nA. These currents reversed at a membrane potential of 4.1 +/- 0.4 mV, which is reasonably close to the theoretical reversal potential for chloride ions (for our experimental solutions) of -1 mV. The responses to low concentrations of GABA were markedly potentiated by clinically relevant concentrations of R(+) etomidate (< 10 micro Meter), and this potentiation increased up to a maximum value of > 3,000%(see Figure 3). At low concentrations of anesthetic, the GABA-induced current was potentiated, but its shape remained unchanged (Figure 3). As the etomidate concentration increased, however, not only did the potentiation increase greatly but the current response showed substantial desensitization. Although we did not investigate this effect in any detail, it was apparent that the degree of desensitization induced by etomidate was approximately the same as that observed with control currents (i.e., with GABA alone) of comparable size. In other words, the degree of desensitization appeared to simply reflect the extent to which the current had been activated. The peaks of the potentiated currents could be fitted by a Hill equation (see Materials and Methods, Equation 3) with an EC50of 40 +/- 9 micro Meter R(+) etomidate, a maximal potentiation of 3,200 +/- 300%, and a Hill coefficient of nH= 1.06 +/- 0.13.
Next we investigated the effects of low concentrations of R(+) etomidate on the control GABA concentration-response curve. At low concentrations of GABA, the control currents did not desensitize with time; however, as the GABA concentration was increased, the rate of desensitization also increased (see insets to Figure 4). The peak GABA-induced control currents were well fitted by a Hill equation (see Materials and Methods, Equation 2) with EC50= 17.9 +/- 1.2 micro Meter and nH= 1.5 +/- 0.1. The effect of R(+) etomidate on the control GABA concentration-response curve was to shift the curve leftward in a parallel manner, with increasing concentrations of etomidate resulting in decreasing values of the GABA EC50(Table 1). Thus, although the responses to low concentrations of GABA were potentiated by R(+) etomidate, responses to higher concentrations of GABA were less affected. At the highest concentration of GABA used (300 micro Meter), 10 micro Meter etomidate did not significantly affect either the peak of the response (P > 0.1) nor the time course of desensitization. Figure 4shows representative traces.
We investigated the effects of the two optical isomers of etomidate over a range of concentrations likely to be clinically relevant (< 10 micro Meter). The observed potentiations (of a 1 micro Meter GABA control response) by the R(+) and S(-) isomers are shown in Figure 5(A), where it can be seen that the R(+) isomer was considerably more effective than the S(-) isomer in potentiating the GABA-induced current (although it is unclear from our data if this is due to differences in efficacy or to binding). In experiments designed to determine the precise degree of stereoselectivity, the isomers were applied to individual cells, the ratios of the potentiations measured, and the data from many cells were then averaged (this is considerably more precise than averaging observed potentiations for the individual isomers acting on different cells). The resulting ratios are shown in Figure 5(B) together with some representative current traces illustrating the time courses of the responses. The degree of stereoselectivity increased significantly from [approximately] 5 at the lowest etomidate concentrations to [approximately] 10 with 10 micro Meter etomidate. We observed a comparable degree of stereo-selectivity in the extent to which high concentrations of the etomidate isomers could activate a current on their own in the absence of GABA (see the inset to Figure 5(A)).
To facilitate a molecular interpretation of our observed stereoselectivity on tadpoles and GABAAreceptors, we investigated the effects of the two etomidate enantiomers on lipid bilayers. Probably the most sensitive measure of the ability of an anesthetic to partition into and disrupt a lipid bilayer is the extent to which the anesthetic can decrease the main chain-melting phase transition temperature of lipid bilayers. We found that the phase transition temperature of dipalmitoyl phosphatidylcholine was decreased linearly (data not shown) with increasing concentrations of etomidate. However, not only were high concentrations required to observe substantial shifts, but the reduction in the phase transition temperature for the two isomers was the same. This is illustrated in Figure 6, where the effects of 1 mM of the R(+) and S(-) isomers appear to be essentially identical (P > 0.3).
In any in vitro study of the effects of general anesthetics on a putative target, it is imperative to consider the effects observed in the context of the free aqueous concentrations that are present during general anesthesia in vivo. Although it is relatively straightforward to make such estimates for inhalational agents, it is more problematic for intravenous anesthetics that can be rapidly metabolized. Nonetheless, good estimates are available for quasi-steady-state EC50concentrations of thiopental, pentobarbital, and propofol, which are necessary to achieve defined anesthetic endpoints. In the case of R(+) etomidate, the best estimates that are available from the literature are the plasma concentrations that are present in patients on awakening after a bolus intravenous injection. [24,25]Taking an average value of 0.25 micro gram/ml plasma, together with a plasma/buffer partition coefficient of 4.3 gives a value for the free aqueous concentration of R(+) etomidate of [approxiamately] 0.25 micro Meter. Our value of 3.4 +/- 0.1 micro Meter R(+) etomidate for the loss-of-righting reflex in tadpoles is considerably higher, but it provides another estimate. Although a direct comparison between mammalian and tadpole EC50values for general anesthesia is fraught with difficulties, it does appear that for many general anesthetics (e.g., thiopental, [12,27]propofol, [23,28]and most inhalational agents [22,29]), the concentrations required to induce the loss-of-righting reflex in tadpoles is surprisingly close to the concentrations required to prevent movement in response to a painful stimulus in mammals. It would seem, therefore, that we can consider free aqueous concentrations of R(+) etomidate of a few micromolar or less to be clinically relevant.
In the micromolar range, R(+) etomidate has been shown to allosterically modulate ligand binding to GABAAreceptors, to cause marked prolongation of GABAergic IPSPs (about a fourfold increase in decay time at 10 micro Meter) and miniature IPSCs (about a threefold increase in charge transfer at 8.2 micro Meter), inhibition of synaptic transmission in hippocampal neurons (IC50[nearly =] 6 micro Meter), and substantial potentiation of currents induced by low concentrations of GABA in cultured hippocampal neurons and oocytes expressing GABAAreceptors, [31,32]although the degree of the potentiations observed are highly variable and clearly depend on the GABA receptor subunit composition. [31–33]The potentiations we observed at etomidate concentrations < 10 micro Meter fall within the range found by other workers. Similarly, the lack of any significant mimetic effect (i.e., a direct activation of a GABAergic current by etomidate) in this concentration range is consistent with previous reports, although this action of etomidate is also subunit-dependent. [31,32]Some investigators have argued that this direct effect is responsible for the anesthetic action of etomidate. However, we believe this is unlikely given the rather small effects that are observed at relevant concentrations. For example, even the relatively high concentration of 10 micro Meter R(+) etomidate depolarizes frog primary afferent fibers by less than 100 micro V and activates a GABAergic current of only about 10% of the maximum GABA-induced current, even for the most etomidate-sensitive GABA receptor. For the GABAAreceptor we studied, the direct effect is considerably smaller (see inset to Figure 5(A)). This direct effect is to be compared to the potentiating effects of 10 micro Meter R(+) etomidate on GABA-activated currents: 1,500% potentiation for the most sensitive GABA receptor and the 600% potentiation we have observed (Figure 5(A)). Although it is obviously difficult to extrapolate from these in vitro effects and draw unambiguous conclusions regarding the actions of etomidate in whole animals, or humans, it would seem safe to conclude that, at the free aqueous concentrations of etomidate that are present during etomidate general anesthesia, GABAAreceptors are likely to be substantially potentiated (a 50% potentiation of an inhibitory current can be thought to be equivalent to a 50% inhibition of an excitatory current ). The potent effects of etomidate on GABAAreceptors can be contrasted with the relative insensitivities of glycine receptors, nicotinic acetylcholine receptors, [36,37]5-HT3receptors, voltage-gated calcium [39–42]and potassium channels, and gap junctions. 
The leftward shifts in GABA concentration-response curves in the presence of etomidate (Figure 4) and the corresponding decrease in the GABA EC50concentrations (Table 1) have been observed for many general anesthetics, including etomidate, and are consistent with R(+) etomidate acting by allosterically enhancing the binding of GABA for its receptor. [7,44]Our observation that etomidate enhances the rate of desensitization of GABA-induced currents (see Figure 3) has also been seen by others [30,32]; it can be accounted for by this allosteric action and appears to reflect the extent to which the GABA current has been activated. Similar behavior has been reported for barbiturates potentiating GABAAreceptors and volatile anesthetics potentiating glycine and 5-HT3receptors. In general, it would appear that when anesthetics potentiate the activities of members of this superfamily of neurotransmitter-gated receptors, their effects on desensitization rates are mainly a consequence of an apparent increase in agonist binding.
The stereoselectivity of etomidate as a general anesthetic has been established unequivocally, although the S(-) enantiomer has simply been stated to be “devoid of hypnotic activity,”and there has been no previous estimate available of its EC50concentration for general anesthesia. The concentration-response data in Figure 2, showing the loss-of-righting reflex in tadpoles, gives some quantitative information on the relative potencies of the two enantiomers. These data show that the S(-) isomer is approximately 15 times less potent as a general anesthetic than the R(+) isomer. Studies with the etomidate enantiomers in vitro have been few and far between and have generally included the application of only a single concentration of the S(-) isomer. However, the data that do exist show that 10 micro Meter S(-) etomidate has virtually no effect on synaptic transmission or the time course of GABAergic IPSPs in hippocampal neurons, or the enhancement of GABA or diazepam binding to cortical membranes. In one of the few studies that used more than a single concentration, the S(-) isomer was approximately 20 times less effective at depolarizing frog primary afferent terminals, this effect being attributed to the mimetic action of etomidate. Our own data show, similarly, that the R(+) isomer is about 10 times more effective than the S(-) isomer at activating a direct current (see Fiagure 5(A)). A comparable degree of stereoselectivity has also been reported for the etomidate enantiomers allosterically modulating the binding of various ligands to GABAAreceptors. 
The data illustrated in Figure 5show the relative effects of the two etomidate enantiomers in potentiating currents induced by low concentrations of GABA. Although both enantiomers can potentiate GABA-induced currents, the R(+) isomer is significantly more effective than the S(-) isomer. In addition, the degree of stereoselectivity changes significantly with anesthetic concentration. The fact that the degree of stereoselectivity increases with concentration is, a priori, rather surprising because we might anticipate that a greater degree of stereoselectivity would be observed at the lowest concentrations because the tightest (and presumably the most stereoselective) binding sites would be recruited first. However, we have observed similar behavior for the potentiation of GABA responses by the optical isomers of isoflurane. This increase in stereoselectivity is most simply explained by the anesthetic binding site(s) on the receptor being modified by the anesthetic-induced allosteric conformational change.
The fact that the presence of general anesthetics often results in the tighter binding of neurotransmitters. (e.g., see Figure 4) means, as a general corollary, that in these cases the binding of neurotransmitters in turn must enhance the binding of the anesthetic. Such a conclusion follows naturally from simple equilibrium schemes of the following general type:Equation 4where the binding of GABA (G) to its receptor (R) can occur in the presence or absence of anesthetic (A) to form an active complex (RG or RAG). If, in the presence of anesthetic, the dissociation constant for GABA is reduced (i.e., GABA binds tighter and K sup *G< KGthen it follows that the dissociation constant for the anesthetic is also reduced because Equation 5.
Here the binding of the anesthetic favors the binding of the neurotransmitter and vice versa. (We have observed essentially analogous behavior with the firefly luciferase enzyme, where increasing concentrations of one of its substrates, adenosine triphosphate, enhanced the affinity of the anesthetics for the enzyme.) Thus, interestingly, at the high concentrations of neurotransmitters that are thought to be released across synaptic clefts, one would predict that anesthetics would bind tighter to the postsynaptic receptors (and perhaps exhibit a higher degree of stereoselectivity) than they do when low concentrations of neurotransmitter are applied during bath application in in vitro experiments.
Thus, although the degree of stereoselectivity that we have observed in vitro is comparable qualitatively and quantitatively with that found in animals and is certainly consistent with the view that etomidate exerts its principal effects on GABAAreceptors, [7,15,16]possible differences between the responses of receptors at intact synapses and those of receptors expressed in transfected cells need to be investigated. Further, until the stereo-selectivity of etomidate acting on other putative targets is determined, a definitive statement on the molecular mechanisms underlying etomidate action is premature. Nonetheless, at the molecular level, the high degree of stereoselectivity that we and others have found for etomidate acting on GABAAreceptors, together with the complete lack of stereoselectivity for etomidate dissolving in and disrupting lipid bilayers (see Figure 6), means there can be little doubt that etomidate is acting by binding to a specific site or sites on the GABA sub A receptor rather than by exerting its effects by some nonspecific mechanism.
The authors thank the Medical Research Council for its support and Dr. P. Whiting for the gift of the stably transfected PA3 cells.