THE traditional dogma on mechanisms of general anesthesia highlights the ease with which many anesthetic agents dissolve in olive oil and other “membrane-like” substances. The suggestion was made that general anesthetics occupy a critical volume of the bulk phospholipid compartment of the neuronal membrane and in doing so alter membrane fluidity and excitability. Several problems have arisen with these ideas, chief of which might be the observation that bulk membrane fluidity changes, such as may be produced by other manipulations (e.g., small changes in temperature) do not elicit anesthesia. Secondly, the so-called Meyer-Overton “rule,” which states that the product of an anesthetic's potency and lipid solubility should be a constant, fails to predict the lack of anesthetic activity of several fluoroalkanes that are extremely hydrophobic, but nonanesthetic. Last, but not least, lipid theories predict that pairs of optical isomers (which have identical chemical and structural formulae, but differ in the arrangement of chemical groups about a single chiral carbon atom) should possess equal potency as anesthetics. The optical isomers of isoflurane have been reported to have unequal potencies as anesthetics. However, the differences in potency between (+) and (-) isoflurane in vivo are small enough to leave this issue in doubt. A new study by Tomlin et al., published in this issue of Anesthesiology, demonstrates a large ([approximately] 15-fold) difference in anesthetic potency in tadpoles between the optical isomers of the potent intravenous anesthetic, etomidate. Such stereospecificity obviously cannot be accounted for by “classical” lipid theories, a conclusion bolstered by the observation that the etomidate isomers have identical effects on the physical properties of lipid bilayers. This degree of stereospecificity is characteristic of a target site on a protein, and so the data reported by Tomlin et al. represent another significant nail in the coffin for the traditional “nonspecific” dogma of anesthetic mechanism.
Tomlin et al.'s study is significant in a second way, as the study of these optical isomers may help illuminate the nature of the protein targets for etomidate anesthesia. The most likely candidates as targets for many anesthetics are the ligand-gated ion channels, which function as receptors for neurotransmitters such as acetylcholine, L-glutamate, and gamma-aminobutyric acid (GABA). One such receptor has been associated with anesthesia since Sir John Eccles noted a prolongation of spinal presynaptic inhibition during Nembutal anesthesia in 1946. Of course, Eccles was unaware of the existence of GABA at the time! The explosive growth of neuroscience in the second half of the twentieth century has enabled an enormous increase in our level of understanding of general anesthetic effects in the brain and spinal cord. Much evidence now links anesthesia and GABA, most crucially the observation that clinically relevant concentrations of many general anesthetics, including isoflurane, halothane, propofol, pentobarbital, and the steroid alphaxalone, enhance the inhibitory actions of GABA on central neurons via an action on postsynaptic GABAAreceptors (although presynaptic effects also may occur). Etomidate also regulates GABAAreceptors at clinical concentrations, and Tomlin et al. show that (+)etomidate is more potent and more efficacious than the (-)isomer in regulating recombinant GABAAreceptors expressed in Xenopus oocytes. This observation argues strongly against indirect effects of the anesthetic on the function of the receptor protein via perturbation of bulk phospholipid and in favor of the existence of specific anesthetic-binding “pockets” or cavities within proteins such as the GABAAreceptor.
The advent of molecular biological techniques now promises to contribute further to dismantling the edifice of the nonselective lipid hypothesis of anesthesia. Two recent studies using site-directed mutagenesis show that mutation of single critical amino acid residues within the GABAAreceptor can completely abolish the allosteric effects of two distinct anesthetic agents. In the case of etomidate, Belelli et al., showed that mutation of asparagine 289 to methionine in the beta3 subunit abolished the allosteric regulation of the GABAAreceptor, whereas in the case of enflurane, the critical residues were serine 270 and alanine 291 in the alpha subunit. These experiments usher in a series of new and exciting experiments with genetically engineered animals; in fact, this era has already arrived. 
In conclusion, the nonspecific hypothesis of anesthetic mechanism has consistently failed to illuminate the events that follow equilibration of anesthetic with the neuronal membrane or to make useful predictions that can be tested experimentally. However, proponents of this idea were so vociferous (and so numerous) as to create a “dark age” in the science of anesthetic mechanism, from which the field has only recently begun to emerge. Although the nonspecific hypothesis of anesthetic action is not yet completely dead, recent findings with isomeric pairs and mutagenesis now threaten to drive the final stake into its heart. Experiments with optical isomers will not only help us to reject an increasingly enfeebled dogma, but they should also throw new light on the spectrum of genuine anesthetic targets.
Neil L. Harrison, Ph.D.
Associate Professor; Department of Anesthesia and Critical Care; The University of Chicago; Whitman Lab; 915. East 57th Street; Chicago, Illinois 60637