Inhibition mediated by γ-aminobutyric acid type A (GABA A) receptors has long been considered an important target for a variety of general anesthetics. In the hippocampus, two types of phasic GABA A receptor-mediated inhibition coexist: GABA A,fast, which is expressed primarily at peri-somatic sites, and GABAA,slow, which is expressed primarily in the dendrites. Their spatial segregation suggests distinct functions: GABA A,slow may control plasticity of dendritic synapses, whereas GABA A,fast controls action potential initiation at the soma. We examined modulation of GABA A,fast and GABA A,slow inhibition by isoflurane at amnesic concentrations, and compared it with modulation by behaviorally equivalent doses of the GABA A receptor-selective drug etomidate.
Whole cell recordings were obtained from pyramidal cells in organotypic hippocampal cultures prepared from C57BL/6 × 129/SvJ F1 hybrid mice. GABA A receptor-mediated currents were isolated using glutamate receptor antagonists. GABAA,slow currents were evoked by electrical stimulation in the stratum lacunosum-moleculare. Miniature GABA A,fast currents were recorded in the presence of tetrodotoxin.
100 μM isoflurane (approximately EC50,amnesia) slowed fast- and slow-inhibitory postsynaptic current decay by approximately 25%. Higher concentrations, up to 400 μM, produced proportionally greater effects without altering current amplitudes. The effects on GABA A,slow were approximately one-half those produced by equi-amnesic concentrations of etomidate.
Isoflurane enhances both types of phasic GABA A receptor-mediated inhibition to similar degrees at amnesic concentrations. This pattern differs from etomidate, which at low concentrations selectively enhances slow inhibition. These effects of isoflurane are sufficiently large that they may contribute substantially to its suppression of hippocampal learning and memory.
What We Already Know about This Topic
Isoflurane enhances synaptic inhibition by potentiation of GABA (γ-aminobutyric acid)ergic synaptic transmission
The role of this mechanism in amnesia produced by subanesthetic concentrations of isoflurane is unclear
What This Article Tells Us That Is New
Isoflurane at amnestic concentrations prolonged both fast and slow forms of phasic GABAAreceptor-mediated inhibition in mouse hippocampal neurons
Enhancement of GABA-mediated synaptic inhibition by isoflurane does contribute substantially to isoflurane-induced amnesia
IONOTROPIC γ-aminobutyric acid type A (GABAA) receptors are key mediators for many receptor-selective anesthetics, including etomidate and propofol.1They are also an important target for benzodiazepines, which by targeting specific receptor subtypes can produce anxiolysis, sedation, and amnesia at low doses, and unconsciousness at higher doses.2GABAAreceptors have been considered likely mediators of at least some components of the anesthetic state induced by inhaled agents, such as isoflurane.3However, since alternative and/or complementary molecular mediators have been identified and defined using advanced neurobiological techniques for some end-points, such as immobility and hypnosis,4and multiple forms of synaptic and nonsynaptic inhibition mediated by GABAAreceptors have been described,5,6the role of GABAAreceptors and inhibitory synapses in anesthesia is undergoing reassessment.
Some of the best evidence for a role of GABAAreceptors in inhaled anesthetic-induced amnesia comes from studies of mice lacking the α1 subunit of the GABAAreceptor. The α1 subunit is present at many inhibitory synapses in the hippocampus and neocortex and is instrumental in benzodiazepine-induced sedation and amnesia.7Mice lacking this subunit, either globally or only in the hippocampus, are resistant to isoflurane-induced amnesia, as measured by fear conditioning to tone or context (no phenotype with respect to sedation was reported).8These results thus support a role for synaptic α1-containing GABAAreceptors in inhaled anesthetic suppression of learning and memory. However, studies of “knock-in” mice carrying isoflurane-resistant α1 or α2 subunits failed to find any change in the concentration of isoflurane required to impair learning.9,10Moreover, inhibitory postsynaptic currents (IPSCs) mediated by α1- and α2-containing GABAAreceptors at prototypical “fast inhibitory synapses” were reported to be relatively insensitive to isoflurane, compared with a tonic form of inhibition that exists in hippocampal neurons that utilizes receptors containing α5 GABAAreceptor subunits.11
These results have thus called into question the behavioral significance of modulation of synaptic α1- and α2-containing receptors, or even synaptic inhibition itself, to isoflurane-induced amnesia. Inhibitory synapses do, however, display a remarkable diversity in their physiologic, pharmacological, and anatomical characteristics.5Is it possible that other types of inhibitory synapses that utilize different subunits play a more important role in modulating memory than those formed by basket cells and other interneurons that impinge on the somata of pyramidal cells? In recent studies we showed that GABAA,slow, a form of synaptic inhibition that is prominent in the dendrites of pyramidal neurons and is particularly well suited to control synaptic plasticity, is mediated in part by α5 subunits of the GABAAreceptor,12and that amnesic concentrations of etomidate enhance GABAA,slowsignificantly more than GABAA,fastIPSCs.13Here we report that amnesic concentrations of isoflurane also enhance GABAA,slowsubstantially, by approximately one-half of the modulation produced by equally effective (amnesic) concentrations of etomidate. Further, we show that in contrast to etomidate, isoflurane modulates both types of phasic inhibition to similar degrees. These findings suggest that modulation of GABAA,slowcontributes to amnesia produced by both etomidate and isoflurane, and that for isoflurane an additional contribution may come from modulation of GABAA,fast.
Materials and Methods
All experiments conformed to the guidelines laid out in the Guide for the Care and Use of Laboratory Animals, and were conducted with the approval of the University of Wisconsin-Madison (Madison, Wisconsin) Animal Care and Use Committee. Organotypic hippocampal brain slice preparation, electrophysiology, materials and chemicals used, and recording conditions were similar to those described recently. Briefly, organotypic hippocampal slice cultures were prepared from 3–8-day-old C57Bl6/129-SvJ hybrid mice as described by Stoppini et al. ,14maintained in an incubator at 36°C in 5% CO2, and used between 10 and 14 days in culture, at which time hippocampal structures remained easily identifiable. Whole cell recordings were obtained from CA1 pyramidal neurons using borosilicate glass pipettes (KG33; Garner Glass, Claremont, CA) filled with a solution containing (in mM): CsCl 135 (for evoked IPSCs recordings, CsCl was partially replaced by 40 mM K-gluconate), Na-HEPES 10, EGTA 10, MgATP 3, GTP 0.5, lidocaine N-ethyl bromide (QX-314) 5, pH = 7.25. Signals were amplified using a MultiClamp 700A amplifier (Axon Instruments, Foster City, CA) and ClampEx software (Axon Instruments), filtered at 5 kHz, and sampled at 10 kHz using a Digidata 1322A (Axon Instruments). Open tip resistance was 4 MΩ. Whole cell access resistance was less than 15 MΩ before compensation by 50–80%. Slices were continuously superfused with artificial cerebrospinal fluid at 24 ± 1°C (evoked responses) or 34 ± 1°C (spontaneous IPSCs), saturated with carbogen (95% O2+ 5% CO2), and containing the glutamate receptor antagonists 6-cyano-7- nitroquinoxaline-2,3-dione (CNQX, 20 μM) and (2R)- amino-5-phosphonovaleric acid (APV, 40 μM) to block excitatory neurotransmission.
Chemicals and Drugs
All chemicals and drugs except for isoflurane were purchased from Sigma-Aldrich (St. Louis, MO). Isoflurane (Abbott Laboratories, Chicago, IL) was prepared from a saturated stock solution stored in 500 ml Teflon gas sampling bags (Fisher Scientific International Inc., Hampton, NH) and transferred to glass syringes for use. To minimize loss of isoflurane, polytetrafluoroethylene tubing was used to connect the glass syringe reservoirs and the recording chamber. The concentration of isoflurane in the recording chamber was measured using gas chromatography (Gow-Mac series 580 flame ionization detector gas chromatograph, 6′×18” stainless steel column packed with 0.2% carbowax 1500 on carbopak C, 60/80; Gow-MAC Instrument Company Bethlehem, PA). The detector and the column were set to 160°C and 140°C, respectively. A nitrogen flow of 30 ml/min resulted in retention times of 45s.
Determination of EC50,amnesia
To compare the effects of isoflurane and etomidate on IPSCs, the EC50,amnesiaconcentration of isoflurane was taken to be 0.28%,9which corresponds to an aqueous concentration of 114 μM,15and the concentration of etomidate as 0.25 μM.16These concentrations were determined using comparable fear-conditioning paradigms, and both in hybrid C57Bl6x129-SvJ mice, the same hybrid strain that we used for the present experiments.
Data were analyzed offline on a personal computer using Mini Analysis (Synaptosoft, Decatur, GA) and ClampFit (Molecular Devices, Sunnyvale, CA). The threshold for event detection was set at three times the root mean square noise level. Spontaneous IPSCs were analyzed by automated event detection that acquires amplitude, 10–90% rise times, and the time to 63% decay. Evoked IPSCs were analyzed by fitting to a monoexponential function using least squares minimization of errors.
Statistical analysis was performed using Microcal Origin (version 7; OriginLab Corporation, Northampton, MA), GraphPad Prism (version 7; GraphPad Software, San Diego, CA), or Microsoft Excel (version 12.3.1; Microsoft Corporation, Redwood, WA). Data are presented as mean ± SD. Statistical comparisons were made using one-tailed Student t tests when a strong expectation existed based on existing literature (e.g. , slowing of decay by isoflurane and etomidate), or Z-tests when testing for deviation of normalized values (iso/control) from unity. Effects were considered significant at P < 0.05.
Isoflurane Slows the Decay of GABAA,slowIPSCs
As reported previously,13electrical stimulation in stratum lacunosum-moleculare of the organotypic hippocampal slice cultures in the presence of glutamate receptor blockers evoked GABAergic IPSCs with rise and decay times characteristic of GABAA,slow. These evoked responses were reversibly slowed by 200 μM isoflurane (fig. 1A), a concentration that is equivalent to 0.5% isoflurane in vivo , which is the lowest concentration that is reliably amnesic.9Both the time from drug application to peak effect and the time from drug termination to recovery of baseline responses were rapid in the organotypic hippocampal slice cultures, typically less than 10 min (fig. 1B). Analysis of evoked responses during drug wash in and washout revealed that 200 μM isoflurane reversibly slowed IPSC decay (fig. 1B, red symbols) but did not alter IPSC amplitude (fig. 1B, blue symbols).
We explored the concentration dependence of isoflurane's effect on evoked GABAA,slowIPSCs by applying isoflurane at concentrations ranging from 55 μM (which is approximately one-half EC50,amnesia)9to 400 μM (approximately 3.5 times EC50,amnesia, or 1.25 times EC50,movement). As illustrated in figure 2, 55 μM isoflurane had no measurable effect, but higher concentrations caused IPSC decay to be slowed in a concentration-dependent fashion. At a concentration of 100 μM, which is slightly less than the EC50,amnesiaconcentration, isoflurane slowed IPSC decay by a statistically significant extent (70 ± 11 ms control vs. 86 ± 9 ms isoflurane, P = 0.001, n = 4, one-tailed paired Student t test).
Isoflurane Also Slows the Decay of GABAA,fastIPSCs
To measure effects of this same range of isoflurane concentrations on GABAA,fastIPSCs, we studied tetrodotoxin-resistant miniature IPSCs, of which more than 99.9% display the rapid rise and decay characteristic of this class.13,17A representative example of the effect of 200 μM isoflurane is shown in figure 3. Comparison of raw traces in the absence versus presence of isoflurane (fig. 3A vs. 3B) shows little obvious effect. A synopsis of all miniature IPSCs recorded from this cell suggests a slight reduction in the peak amplitude (fig. 3C), and a slowing of the decay (fig. 3D), without any change in their frequency (fig. 3E). A summary of all similar experiments over a range of isoflurane concentrations confirmed a dose-dependent effect on the decay of miniature IPSCs (fig. 4A) but revealed no effect on their amplitude (fig. 4B). For comparison, figure 4also presents the effects of isoflurane on the same parameters of evoked GABAA,slowresponses. The decay rates of both types of IPSCs were slowed dose-dependently between 100 and 400 μM isoflurane (fig. 4A). Since the peak amplitude was largely unaffected in this concentration range (fig. 4B), charge transfer was increased for both types of IPSCs (fig. 4C). At 400 μM, i.e. , four times EC50amnesia, it increased charge transfer more for fast than slow IPSCs (P < 0.05, two-tailed Student t test, GABAA,fastn = 3, GABAA,slown = 4). At other concentrations, there were no significant differences (P > 0.05, two-tailed Student t test, GABAA,fastn = 4 at all concentrations, GABAA,slown = 3, 5, and 4 at 50, 100, and 200 μM respectively).
Temperature-sensitivity of Modulation
The effects of anesthetics on molecular targets are often studied in vitro at less than physiologic temperatures. However, these observations can be extrapolated to the in vivo situation by taking into account the temperature-induced changes of gas solubility in aqueous solutions.15No difference in the effect of enflurane, an isomer of isoflurane, on GABAA,fastsynaptic inhibition was found between room and body temperature.18Less is known about temperature-dependent changes of injectable anesthetic effects. We tested the temperature-dependence of intravenous anesthetics by comparing the effect of etomidate on the decay of GABAA,slowIPSCs at 24°C and 34°C. As expected, IPSC decay became faster with increasing experimental temperature (Q10 = 3.3 ± 0.9). Nevertheless, the concentration-dependent relative slowing of the decay by etomidate remained constant (table 1). Therefore we combined data obtained at 24°C and 34°C in figure 5.
We found that isoflurane, at concentrations that effectively block hippocampal memory formation in vivo , prolonged both types of GABAA–ergic phasic inhibition in the murine hippocampus. The effect profile, however, differed substantially from that of the more selective GABAA–ergic drug etomidate: isoflurane affected GABAA,slowless and GABAA,fastmore than etomidate at comparably amnesic concentrations. We discuss the relevance of these findings within the framework of GABAAreceptor-mediated modulation of hippocampal memory formation.
Two Types of Phasic Inhibition
The existence of (at least) two types of phasic GABAA-ergic inhibitory currents, which were originally discovered in the hippocampus, is now well-established in many brain areas.6In the hippocampus, the slow time course of decay of GABAA,slow(30–70 ms, as opposed to 3–8 ms for GABAA,fastat 36°C) is its most striking characteristic. Its slow time course, in combination with its dendritic localization, places this slow synaptic current into an ideal position to balance the equally slow time course of dendritic excitation mediated by the N -methyl-D-aspartate receptor-mediated component of glutamatergic synaptic input.19Given the critical role of N -methyl-D-aspartate receptors in initiating many forms of synaptic plasticity, including long-term potentiation, GABAA,slow-mediated inhibition is thus well suited to control synaptic plasticity, and by extension, hippocampus-dependent learning and memory.
Differential Effects of Isoflurane and Etomidate
We reported recently that amnestic concentrations of etomidate enhance GABAA,slowphasic inhibition more than GABAA,fast, and concluded that etomidate-induced conscious amnesia may be because of the preferential enhancement of GABAA,slowIPSCs. Previous observations linking etomidate-induced amnesia to tonic GABAA-ergic inhibition, specifically to receptors containing the α5 subunit (discovered by using α5 knock-out mice),22are compatible with this interpretation, as GABAA,slowphasic inhibition is also mediated in part by α5-containing receptors.12,23An assumption underlying this reasoning is that since etomidate is a comparatively selective GABAAreceptor agonist, its effects on higher cognitive function are likely to result from the sum of its effects on various types of GABAAreceptor-mediated inhibition. However, the precise quantitative contributions of each of the three distinct forms of GABAA-ergic inhibition (two phasic, one tonic) to sedation, amnesia, hypnosis, and immobility in response to etomidate are yet to be determined.
By contrast, isoflurane modulates numerous targets in addition to the GABAA-ergic system.3In order to determine whether isoflurane's modulation of any individual component is strong enough to contribute substantially to amnesia, we compared its effects, under identical experimental conditions, to changes induced by etomidate. Our results indicate that enhancement of GABAA,slowby isoflurane, approximately one-half of etomidate's effect at behaviorally equivalent concentrations (fig. 5A), is indeed strong enough to contribute substantially. Since GABAA,slowis mediated largely by GABAAreceptor subunits that contain β3 subunits,24the recent finding that mice carrying a forebrain-specific knockout of the β3 subunit are resistant to the amnestic effect of isoflurane provides additional support for a role of GABAA,slowin isoflurane-induced amnesia.25Surprisingly, that same study reported that this selective knockout did not influence etomidate-induced amnesia. The explanation for this lack of effect for etomidate is unclear, but it may reflect the restriction of the knockout to principal (excitatory) cells,26whereas anesthetic-sensitive GABAA,slowIPSCs are also found in interneurons,27where they provide essential timing information through cross-frequency coupling between inhibitory circuits that oscillate at θ- and γ frequencies.28
The other component of phasic inhibition, GABAA,fast, which was barely affected by an amnesic concentration of etomidate, was enhanced by isoflurane to an even greater extent than GABAA,slow(fig. 5B). To the extent that modulation of GABAA,fastcan influence synaptic plasticity, perhaps by altering somatic spiking and back-propagation of action potentials into the dendrites,29modulation of synaptic GABAAreceptors would then play an even greater role in isoflurane-induced amnesia than indicated by comparing only effects of isoflurane and etomidate on GABAA,slow. Since isoflurane modulates both components of phasic inhibition, whereas etomidate preferentially modulates GABAA,slowIPSCs,13this result supports the concept that different agents may achieve similar endpoints by distinct but partially overlapping mechanisms.30
A number of assumptions underlie the approach that we used for quantifying the contribution of GABAAreceptors to isoflurane-induced amnesia. We believe them to be reasonable, and there are supporting data for each one. Nevertheless, to the extent that they represent simplifications of a complex system, and the underlying physiologic basis of memory as well as the clinical relevance of fear conditioning-based models remain incompletely understood, they should be recognized as limitations to the present study. First, we measured drug effects on only a limited set of inhibitory processes. Even within just the hippocampus there exist numerous classes of inhibitory interneurons that differ in their firing patterns, physiologic characteristics, and anatomical projections.31They communicate among themselves and impinge on pyramidal neurons using a physiologically and pharmacologically diverse set of inhibitory synapses as well as nonsynaptic tonic inhibition, and the different forms of inhibition display differential sensitivity to anesthetic drugs.11Although we have presented evidence that slow dendritic IPSCs are well suited to control synaptic plasticity,6,13the precise means by which any form of inhibition controls synaptic plasticity, and learning and memory, remains unclear. Second, to establish a “fractional contribution” of one specific class of anesthetic targets, there must be a linear summation of effects, or at least not a strongly synergistic or antagonistic contribution from different targets. Other investigators have examined this question, at least in relation to endpoints other than amnesia, and have concluded that synergy does exist for some types of receptor-specific agents, but the deviation from additivity is usually small, and for inhaled anesthetics it is generally absent.32,33Third, drug concentrations required to suppress learning depend on the types of learning in humans34and in animal models.9The present study compared the effects of isoflurane and etomidate on only one type of learning: fear conditioning to context, a paradigm that depends upon the hippocampus and amygdala.35Other paradigms or types of learning that engage this same circuitry differently, or that depend on different brain structures, require different drug levels for suppression. Since additional drug targets come into play at higher concentrations, it is possible, even likely, that the precise contributions of the different targets will vary with learning task even for a relatively specific drug such as etomidate. Nevertheless, with these caveats, the present data show that isoflurane modulates GABAAreceptor-mediated inhibition sufficiently strongly even at low “amnesic” concentrations that this receptor family is expected to contribute substantially to this behavioral effect.
Other Targets of Isoflurane that May Contribute to Amnesia
Our quantitative comparison of effects of isoflurane and etomidate on hippocampal inhibition indicates that modulation of GABAA,slowalone by isoflurane is insufficient to cause amnesia. What other targets might contribute? Among the many possible targets, the slow depolarization mediated by N -methyl-D-aspartate receptors (the glutamatergic excitatory counterpart of GABAA,slow) is an attractive complementary target.36,–,38In this “integrative” view, modest effects on the excitatory component (reduction in glutamate release,39,40and a postsynaptic block of the N -methyl-D-aspartate receptor-mediated current37,41,42) paired with a modest but functionally important enhancement of GABAA,slow-mediated hyperpolarization, may cooperatively disrupt synaptic plasticity. The quantitative approach that we employed here may prove useful in establishing which of the multiple targets influenced by isoflurane actually do play important roles.
The authors thank Chong Lor, B.S., Associate Research Specialist; Pee Yang, Student; and Mark Perkins, B.S., Senior Research Specialist, for technical assistance, and Kirsten Martin for secretarial assistance, all at the Department of Anesthesiology, University of Wisconsin-Madison School of Medicine and Public Health, Madison, Wisconsin.