SINCE the often-cited articles by Sowton and Sherrington, 1Brooks and Eccles, 2Bremer and Bonnet, 3and Larrabee and Pasternak, 4spanning nearly 50 yr of research into the cellular mechanisms of anesthesia, it has been almost axiomatic that depression of synaptic transmission by general anesthetics was not secondary to effects on the action potential. In this issue of the Journal, Wu et al. 5apply cutting-edge electrophysiologic methods and a unique preparation from the mammalian brain to revisit this issue, and their results challenge this long-held and common conception—at least at a specialized synapse, the calyx of Held, and possibly at other excitatory glutamatergic synapses as well.
In the central nervous system (CNS), our understanding of the presynaptic effects of anesthetic drugs lags far behind our detailed knowledge of postsynaptic drug-receptor interactions. This mirrors the general delay in understanding the presynaptic machinery compared to the postsynaptic side of a prototypical CNS synapse. 6The principal reason for this discrepancy lies in the technical difficulty of studying such small structures as axon terminals using electrophysiologic methods. A technical breakthrough came in 1994, when Forsythe 7(followed within months by Borst et al. 8) described a preparation in which both presynaptic and postsynaptic elements of a mammalian CNS synapse were accessible to direct electrophysiologic investigation. Since then, this preparation has contributed significantly to our understanding of presynaptic physiology in general, and to the release of glutamate, the most common excitatory neurotransmitter, in particular. Wu et al. have now used this preparation to address an important question in the field of anesthetic mechanisms.
The calyx of Held is a sign-inverting switch in the brainstem auditory pathway. Located in the medial nucleus of the trapezoid body, it plays a role in determining the spatial location of a sound source based on interaural intensity differences. This unique structure is specialized to provide reliable and rapid excitatory transmission from glutamate-releasing cells that originate in the anterior ventral cochlear nucleus onto the glycinergic neurons of the medial nucleus of the trapezoid body (note the switch from excitatory to inhibitory transmitter in the pathway). By CNS standards, it is a giant presynaptic terminal of 10–15 μm in diameter, and this makes it accessible to electrophysiologic recording using glass micropipettes. Each principal neuron in the medial nucleus of the trapezoid body receives input from only one calyx-type axon terminal, which, in a 9-day-old rat, comprises about 600 active release zones. Simultaneous recordings from the calyx presynaptic terminal and the postsynaptic neuron permit an unprecedented level of access to both partners of this excitatory CNS synapse.
The notion that the excitatory presynaptic terminal is a likely target for the depressant action of volatile anesthetics on neurotransmission has been suggested previously, based on electrophysiologic studies in the hippocampal slice preparation 9–11and on biochemical studies of isolated cortical nerve terminals (synaptosomes). 12,13However, the link between action potential invasion of the presynaptic terminal and response of the postsynaptic neuron involves multiple processes, including sodium channels, calcium channels, intracellular calcium stores, and a variety of proteins involved in vesicle docking and membrane fusion. 6,14These and other presynaptic mechanisms have all been considered possible targets for volatile anesthetic action. Several observations, including the finding that some sodium channels may be more sensitive to depression than was previously appreciated, led to the suggestion that effects on voltage-gated sodium channels may play a role in reducing transmitter release, 15,16but the relative importance of this and other targets remains unresolved. Wu et al. have now obtained direct measurements of the relative effects of isoflurane on two central events leading to the release of neurotransmitter: the presynaptic action potential and the fusion of transmitter-filled vesicles with the presynaptic membrane. The effect of isoflurane on the link between these two events—the increase in intracellular Ca2+—was not measured directly but instead was extrapolated by simulating action potentials of varying amplitude in the presynaptic terminal and by measuring the resulting Ca2+currents. Simultaneously, the postsynaptic responses to the released transmitter were also measured.
Wu et al. start by showing that anesthetic effects at the calyx of Held are qualitatively and quantitatively similar to those obtained in the more standard preparations (e.g. , the hippocampal slice) having synapses that are considered more representative of those in the CNS. The concentration of isoflurane that reduces by 50% the postsynaptic response at the calyx (0.49 mm) is comparable to the concentrations of other volatile anesthetics that impair excitatory synapses in the hippocampus, 9as is the observation that paired-pulse depression is reduced. 10,11Having demonstrated that volatile anesthetic modulation of glutamate release at the calyx of Held is not unlike that of other CNS synapses, the authors provide new details of anesthetic interactions with the presynaptic terminal. Biologic membranes with electrochemical gradients (all excitable membranes) can “store” some electrical charge (capacitance). The fusion of transmitter-containing vesicles with the membrane of the presynaptic terminal leads to the incorporation of tiny amounts of new membrane into the terminal that can be measured as minute increases in the membrane capacitance of the presynaptic terminal. Because Wu et al. recorded directly from the presynaptic terminal, they were able to demonstrate that isoflurane reduced the capacitance change in response to presynaptic stimulation, indicating that isoflurane reduced the number of glutamate-containing vesicles that fused with the presynaptic membrane to release transmitter. Postsynaptic recordings demonstrated that the depressant effect of isoflurane on the capacitance increase quantitatively matched its depression of the postsynaptic response (43%vs . 50% at 0.7 mm isoflurane), supporting a presynaptic locus of action for its depressant effects on glutamatergic transmission.
Having thus demonstrated that isoflurane depresses transmitter release, Wu et al. addressed possible causes by studying the effects of isoflurane on two processes intimately linked to transmitter release: action potential and subsequent calcium entry. When an action potential traveling from the soma of the neuron invades the presynaptic terminal, Na+entry through voltage-activated channels leads to the initial depolarization of the terminal. Once depolarization reaches a certain threshold, various classes of voltage-gated Ca2+channels open and Ca2+enters, initiating the transmitter release process. The authors found that isoflurane depressed the action potential invading the presynaptic terminal only modestly (5.5% at 0.7 mm). Because of a nonlinear relationship, however, this modest effect on Na+entry translates into a substantial reduction of Ca2+entering the terminal (approximately 12%). Transmitter release is in turn nonlinearly related to Ca2+influx (the cooperativity ranges from 3–4 at various synapses). Therefore, in a short amplification cascade, a mere 5.5% depression of action potential amplitude translated into an approximately 50% reduction in the amount of transmitter released. Not all of the reduction in transmitter release caused by isoflurane could be accounted for by this effect, but a large fraction could—approximately 70%. The remaining 30% depression remains unresolved but might relate to direct effects on voltage-gated Ca2+channels, or on the biochemical machinery that uses Ca2+to allow transmitter-containing vesicles to fuse with the plasma membrane.
Wu et al. are not the first investigators to study the interaction of volatile anesthetics with axonal action potential propagation in the mammalian CNS. The general consensus has been that the effect of various anesthetics on presynaptic Na+channels was insignificant in myelinated axons. 17–19The discrepancy between previous findings and the observations of Wu et al. must be reconciled. One possibility is that Na+channels expressed in the axon differ from those expressed in the terminal in their susceptibility to anesthetic block. Another possibility is that the susceptibility of the calyx demonstrated here is a developmental peculiarity: the shape of the action potential changes dramatically within days as the animals reach the age at which hearing begins (10–12 days in rats 20). Wu et al. conducted their experiments using tissue from animals younger than 10 days old. However, similarities between anesthetic effects on overall function at these synapses and more mature synapses suggest that their results may be generally applicable. Last, it is certainly possible that extracellular recording techniques used in previous studies were not sensitive enough to consistently resolve such small changes in the action potential amplitude.
Wu et al. present strong evidence that isoflurane depresses glutamatergic synaptic transmission at relevant concentrations by reducing the amplitude of the action potential in the nerve terminal. Although it resolves some issues, this work, like all discovery, also leads to new questions. For example, do other volatile anesthetics act similarly? Does this result apply to all excitatory transmitters in the CNS? Is inhibitory transmission similarly depressed at the presynaptic level? The amplitude of evoked inhibitory responses is indeed decreased by isoflurane, but this has been attributed to direct anesthetic effects on postsynaptic inhibitory receptors. 21There is also evidence that halothane can augment transmitter release. 22,23What is different about these synapses, or about the action of this anesthetic? Finally, does depression of glutamatergic neurotransmission contribute to any endpoint of the multifaceted anesthetic state? In this context, it is interesting to recall that hypernatremia increases the minimum alveolar concentration of volatile anesthetics required to suppress movement in response to a noxious stimulus (MAC). 24If glutamate release is so exquisitely sensitive to changes in the amplitude of the action potential, then elevation of [Na+] from 130 to 180 mm could, simply by increasing the driving force, more than account for the 75% increase in MAC—does this underlie the effect of sodium concentration on MAC?
Approximately one century ago, the calyx first described by Hans Held played a significant role in the debate between the supporters of the reticular hypothesis of the organization of the CNS and the proponents of the “individual neuron.” Held’s observations led Ramon y Cajal and others to conclude that the CNS is made up of individual neurons. In the past decade, work on this synapse has helped to clarify numerous issues relating to the mechanisms of transmitter release, especially the contribution of various voltage-gated channels to this process. It is now helping us to understand the molecular mechanisms by which volatile anesthetics produce their long-studied, but heretofore poorly understood, effects.