Current studies in model organisms like the fruit fly and the nematode are identifying genetic loci that influence the way intact organisms respond to general anesthetics. The presumption that anesthetic action in these lowly invertebrates is similar to that of animals with a backbone is supported by several previously published observations, encouraging us to believe that genes identified as influencing anesthesia in Drosophila will be of general interest for studies of anesthetic action. However, when the endpoint used to isolate mutants is the disruption of a complex fly behavior, one must be concerned that the anesthesia phenotype is a very indirect consequence of an effect of the affected gene on motor strength and/or coordination of the organism. What is needed is a secondary test of anesthesia that does not rely on coordinated motor performance. Moreover, to enhance the possibility of studying the relationship between gene action and anesthetic sensitivity, it would be best if this test involved a circuit of limited complexity. In the past year, we have sharpened our understanding of how general anesthetics alter performance in two electrophysiological assay systems. One useful assay is the electroretinogram (ERG), the light-evoked current recorded extracellularly from the surface of the eye. Modest concentrations of anesthetic agents yielded no significant effects on the maintained (photoreceptor) potential or the light-on transient but did affect the light-off response. Here, we discovered that the effect of general anesthetics on the response elicited with a traditional ~3 second light pulse is distinctly different than that elicited with a much shorter (0.2 sec) pulse. Specifically, the off-transient of a 3 sec ERG from wild-type flies is scarcely affected by modest concentrations of halothane, isoflurane, desflurane but all these agents potently depress the off-transient of a 0.2 sec ERG. The identical pattern is seen when Shaker potassium channels are specifically inhibited by genetic or pharmacological means. Could the similar effects on the off-transient reflect a common mode of action by all these agents? Studies by others have provided little support for the idea that Shaker channels are direct anesthetic targets. But, because these studies involved heterologous expression and may have omitted a critical component, we devised an in vivo test. In the most critical case, we assayed anesthetic sensitivity of flies in which the Shaker gene dosage was raised; if Shaker channels were inactivated by anesthetics, these hypermorphs should have been resistant. Since we found no change in sensitivity, we favor the hypothesis that anesthetics work on other components of the visual circuit. The specific anesthetic effect on the off-transient raises the question: what generates this feature and how does the process differ between long and short light pulses? Since the ERG reports on extracellular current flow, which only indirectly reflects the changing transmembrane potentials in cells underlying the electrodes, the answer must only be speculative. However, the well-known functional anatomy of the fly's visual sysem has provided us with a rational basis for a model. The key fact for our model is that the neurotransmitter released by photoreceptors (PRs) in response to light is inhibitory. There are two major neuronal sub-types that receive such input from the PRs: the large monopolar cells (LMCs) and amacrine cells. We propose that the off-transient is generated by excitatory currents produced by one or both of these cell types when they are relieved of PR inhibition at the end of a light pulse. Amacrines are presumptive excitatory neurons that, in addition to feedback synapses onto PRs make synaptic contacts with three other cell types in the outermost layer of the optic lobe: glia, LMCs and T1 cells. Excitation of these cells in response to amacrine disinhibition could also contribute to and/or enhance the off-transient. How does this model accommodate our findings? Of all the cell types mentioned, Shaker expression is limited to PRs; this suggests that the channel is needed to insure prompt repolarization of PRs at lights-off so that relief from inhibition is synchronous. On the other hand, it is parsimonious to propose that anesthetics act downstream of Shaker, for example by depressing the excitatory response to disinhibition. Regardless of these particulars, the more potent effects on short pulse off-transients implies that during longer illumination either the performance of the circuit is strengthened or the circuit itself is changed by activation of additional / alternative neuronal elements. For the second electrophysiological assay we use a system, the junction between a defined larval motor neuron and its target muscle, in which there is no doubt about the functional anatomy. Moreover, intracellular recording from the muscle under voltage clamp conditions permits a precise analysis of elementary events. At behaviorally-relevant concentrations of isoflurane, there is a robust and reversible decrease in the amplitude of excitatory junctional currents (EJCs) evoked by nerve shock. The amplitude of spontaneous miniature EJCs is not reduced, implying that glutamate receptor function is unaffected by isoflurane and that the effect is therefore due to reduction in neurotransmitter release. EJC amplitude is normal when the synapse is depolarized electrotonically, indicating that isoflurane affects neither the vesicle fusion machinery, nor the maximum intraterminal calcium current. Thus, isoflurane reduces neurotransmitter release by reducing presynaptic excitability. Reduced presynaptic excitability is also manifested as a significant reduction in action potential conduction velocity in the segmental nerve over the same time course. Either of two plausible mechansisms could account for anesthetic reduction of neurotransmitter release: 1) a uniform reduction of the probability of release (p) over the entire dendritic tree; or 2) a localized reduction in the number of active release sites (n). Two lines of evidence argue in favor of the former mechanism. First, the effect of isoflurane is reduced by elevated extracellular calcium, conditions in which synaptic glutamate concentrations are high enough to saturate glutamate receptors. Alteration in n should be insensitive to external calcium concentration. Second, isoflurane increases the paired-pulse ratio, a hallmark of reduced release probability, p. We conclude that, rather than increasing failure of spike proagation along the dendritic tree, such as at branch points, isoflurane acts to reduce calcium influx by reducing the excitability of the entire arbor. This might happen in any of several ways, e.g., by reducing inward current, increasing outward current, creating a shunt, or through a combination of these effects. In order to identify isoflurane target(s) in the presynaptic axon and terminal, we are assessing candidate genes. Our first success has come with paralytic, which encodes the principal sodium channel in Drosophila. Previous work has shown that mutations in para cause behavioral hypersensitivity to volatile anesthetics; we find that these mutations also enhance the effect of isoflurane on conduction velocity. This is the first demonstration of a specific cellular correlate of behavioral hypersensitivity to volatile general anesthetics in a mutant.
Showing the most recent 10 out of 17 publications
