IN this issue of ANESTHESIOLOGY, Joo et al. 1examine whether reduced synaptic excitation contributes to anesthetic effects of barbiturates. Shortly after the chemical basis of neurotransmission was established by pioneers of neuroscience, the synapse became a focus of investigation for researchers striving to explain the neural basis of anesthesia. The early discoveries that different transmitters could excite or inhibit activity in adjacent neurons, and that these separate influences are integrated to determine the firing rate of an individual cell, led to the intuitive and plausible hypothesis that anesthesia could be brought about by an altered balance of these two processes, either an increase in inhibition or a decrease in excitation, with different drugs acting via different mechanisms.
Although an apparently simple and straightforward principle, testing this hypothesis has not proved to be an easy task. How does one establish the functional role of changes in a widespread neurotransmitter system in complex nervous systems? One approach is to look for common actions of different agents and to test whether potencies for their actions are correlated with behavioral potencies. In this way, the γ-aminobutyric acid type A receptor has been identified as a likely candidate for producing anesthesia, 2and the well-known action of barbiturates to enhance the activity of this receptor fits nicely with its proposed behavioral relevance. However, many agents have been found to alter a number of physiologic processes that may be functionally important. When barbiturates were found to also block excitation mediated by the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) class of glutamate receptors, 3–5and selective AMPA receptor antagonists were found to reduce anesthetic requirements, 6the following question naturally arose: Does the action of barbiturates on AMPA receptors contribute to the anesthetic properties of these agents?
The article by Joo et al. addresses this question. Using a null mutant animal deficient in a glutamate receptor subunit (GluR2) that normally confers sensitivity to barbiturates, these investigators tested whether pentobarbital is still able to produce hypnosis, despite a reduction in its effect on AMPA receptors in these animals. The finding that it is not only still able, but even more potent in inhibiting the loss of righting, corneal, and pineal reflexes led the authors to conclude that AMPA receptor antagonism in these animals did not contribute appreciably to the hypnotic action of barbiturates. A similar conclusion was also reached recently by a separate group, based on the similarities of the action of sedative and convulsant barbiturate isomers on AMPA receptors. 7
The rationale behind these experiments is noteworthy because it shows both the promise and potential pitfalls of using transgenic animals to relate molecules to behavior. A great advantage of the approach is that now, for the first time, investigators are able to realistically target specific components of a system under investigation and look for changes in whole-animal behavior, similar to the approach that biochemists have used for years but now no longer limited to test-tube experiments. Rather than relying on a correlation of different effects of different drugs or rank orders of potency, Joo et al. were able to take advantage of a change in a single subunit of interest to test for changes in barbiturate effects. Like the use of specific binding site antagonists or stereoisomers to alter components of interest, the ability to differentially target even closely related receptors is an important advance. Of course, questions regarding compensatory effects in global knockouts arise, but more sophisticated techniques such as conditional transgenics are being developed, in which time- and location-specific genetic alterations may be manipulated in ways that reduce the likelihood that confounding changes have developed.
However, interpreting the changes observed in genetically altered animals, even without considering compensatory alterations, may not be entirely straightforward. The problem is not unlike the use of ablation studies to define the function of different brain areas: The investigator is able to say what the system does when something has been altered or lost, but this may or may not show the role that it played in the native state. Joo et al. were able to conclude that barbiturate antagonism of AMPA receptors did not play a role in hypnosis in this transgenic animal model. This result is important because it clearly indicates that other drug actions, such as γ-aminobutyric acid type A receptor enhancement, do play important roles. However, this does not necessarily answer the question of whether AMPA antagonism contributes to hypnosis in wild-type animals. If partial AMPA receptor block is functionally significant under normal conditions, what would be the expected outcome of fully reducing the function of these receptors via subunit knockout rather than with a drug? According to the hypothesis of balanced inhibition and excitation, the requirement for enhanced inhibition would be reduced, leading to a decreased drug requirement in the knockout animals. This is precisely what was observed, thus suggesting that AMPA receptor inhibition may, in fact, play a role in hypnosis. To test this possibility, it will be useful to assess the relative reductions in anesthetic requirement for other drugs that do not inhibit AMPA receptors in this animal model, for it would be expected that a relatively greater reduction would be observed for these other agents. Joo et al. noted that in preliminary experiments, the potency of halothane, which does not reduce AMPA receptor responses, was also reduced. However, a quantitative comparison of the reduced requirement may be necessary to distinguish between a causal role of AMPA receptors in hypnosis and the alternative suggestions of a global reduction in baseline levels of excitation or of altered excitation of inhibitory interneurons.
Even with such experiments, it may prove difficult to define the relative importance of multiple actions of a single agent to the production of anesthesia or even to more specifically defined end points such as hypnosis, analgesia, and immobility. Although it has been a useful concept, the idea that the balance of inhibition and excitation in the nervous system yields a particular behavioral state is simply too limiting. We know that the brain shows a complexity in anatomy and richness in function that is not shared by other organs. Inhibition does not simply reduce the firing rate of target cells, but establishes precise temporal firing relationships in widely distributed ensembles of neurons. 8Excitation does not simply get integrated in a target cell to determine whether it reaches threshold, but interacts with multiple, spatially segregated time- and voltage-dependent processes to alter enzymatic activities and gene expression patterns as well as local dendritic processes. 9,10Although approaching anesthesia with this level of complexity seems a daunting task, steady headway is being made in fields pertinent to anesthetic mechanisms, such as defining the cellular and molecular basis of learning and memory. Improvements in our understanding of the neural basis of anesthesia must incorporate these advances. The study by Joo et al. tackles an important and difficult question and provides a glimpse of the future in which powerful new genetic tools may be targeted to regions or systems in the brain that subserve processes related to specific end points of anesthesia.