Isoflurane Is a Potent Modulator of Extrasynaptic GABA_A Receptors in the Thalamus

General anesthesia is characterized by a variety of behavioral endpoints, including amnesia, analgesia, sedation, blunted autonomic responsiveness, unconsciousness, and immobility (Campagna et al., 2003). In all cases, the dose of general anesthetic required to produce amnesia, sedation, and hypnosis is much lower than that required to produce immobility. For example, the commonly used inhaled anesthetic, isoflurane, prevents voluntary response to spoken commands in human subjects (Dwyer et al., 1992) at concentrations that are only 30 to 40% of the levels required for immobility in response to a painful stimulus.

In the waking state, the thalamus is involved in the processing and relay of sensory information to the cortex, and it is also important in the maintenance of high-frequency synchronous oscillatory activity that is associated with attention (Steriade et al., 1993; von Krosigk et al., 1993; Steriade, 2000; Ribary, 2005). During slow-wave sleep, however, relay neurons in the thalamus are hyperpolarized and generate 1 to 3 Hz (delta wave) activity that entrains neurons in the cortex (Steriade, 2000; Steriade, 2003).

Therefore, the thalamo-corticothalamic loop has long been identified as a key target for anesthetic action to induce sedation and hypnosis (Angel, 1993; Alkire et al., 2000; White and Alkire, 2003; Alkire and Miller, 2005). In vivo, the firing rate of thalamocortical relay cells is strongly suppressed by isoflurane (Detsch et al., 1999), and these in vivo inhibitory effects of isoflurane can be reversed by local application of the GABA_A receptor (GABA_A-R) antagonist bicuculline (Vahle-Hinz et al., 2001), indicating that GABA_A-Rs contribute significantly to the effects of the anesthetic on the firing of thalamic neurons, although other mechanisms involving K⁺ conductances have also been proposed to contribute to the anesthetic effects in the thalamus (Ries and Puil, 1999a).

In addition to the “classic” GABA_A-Rs that mediate synaptic inhibition, an additional population of GABA_A-Rs has recently been demonstrated to occur at extrasynaptic sites in the central nervous system, and a population of such receptors exists on thalamocortical relay neurons (Belelli et al., 2005; Cope et al., 2005; Jia et al., 2005; Bright et al., 2007). These extrasynaptic GABA_A-Rs (consisting mainly of α₄, β₂, and δ subunits) are persistently activated by low concentrations of GABA and have distinct pharmacological properties that differentiate them from the synaptic GABA_A-Rs in relay neurons of the thalamus (which consist mainly of α₁, β₂, and γ₂ subunits) (Jia et al., 2005). GABA-mediated tonic inhibition in thalamic neurons requires expression of the GABA_A-R α₄ subunit, because tonic currents are absent in thalamic relay neurons from GABA_A-R α₄-subunit knockout (Gabra4^–/–) mice (Chandra et al., 2006). The inhibitory function of these extrasynaptic GABA_A-Rs has been shown to be enhanced by the novel hypnotic gaboxadol and by the i.v. anesthetic etomidate (Belelli et al., 2005; Cope et al., 2005; Jia et al., 2005).

In the present study, we investigated the actions of isoflurane on thalamocortical neurons in the mouse ventrobasal (VB) thalamus. We found that isoflurane, at clinically relevant concentrations (25–250 μM) (Franks and Lieb, 1994), evoked sustained currents in VB neurons. These sustained currents were blocked by gabazine, a selective antagonist of GABA_A-Rs. In addition, we examined the actions of isoflurane on GABA_A-Rs expressed in human embryonic kidney (HEK)293 cells, using subunit compositions chosen to resemble synaptic and extrasynaptic GABA_A-Rs found in the thalamus. We found that isoflurane directly activated the“extrasynaptic” α₄β₂δ receptors and showed greater potency at α₄β₂δ receptors than at the “synaptic” α₁β₂γ_2s receptors. The isoflurane-activated current was absent in VB neurons from Gabra4^–/– mice, which do not express extrasynaptic GABA_A-Rs (Chandra et al., 2006). Our data suggest that isoflurane is a potent modulator of extrasynaptic GABA_A-Rs in VB neurons.

Materials and Methods

Electrophysiological Recordings in Brain Slices. Experiments were performed in accordance with institutional and federal guidelines. Mice between 22 and 50 days old (C57BL/6, Gabra4^+/+, and Gabra4^–/–) were anesthetized with halothane and sacrificed. The brains were quickly removed and placed in ice-cold slicing solution, which contained the following: 2.5 mM KCl, 26 mM NaHCO₃, 1.25 mM NaH₂PO₄, 220 mM sucrose, 11 mM glucose, 10 mM MgSO₄, and 0.5 mM CaCl₂, before cutting horizontal slices (300-μm thick) on a microslicer (VT 1000S; Leica, Wetzlar, Germany).

Slices were perfused with carbogenated artificial cerebrospinal fluid, which contained the following: 124 mM NaCl, 2.5 mM KCl, 2 mM MgSO₄, 2 mM CaCl₂, 26 mM NaHCO₃, 1.25 mM NaH₂PO₄, and 10 mM glucose. Whole-cell patch clamp recordings from visually identified thalamic neurons were performed using an Axopatch 200A amplifier (Molecular Devices, Sunnyvale, CA) at room temperature (20–22°C) as described previously (Jia et al., 2005). The intracellular solution for voltage-clamp recordings contained the following: 140 mM CsCl, 4 mM NaCl, 1 mM MgCl₂, 10 mM HEPES, 0.05 mM EGTA, 2 mM ATP-Mg, and 0.4 mM GTP-Mg; pH was adjusted to 7.2 with CsOH. Intracellular solution for current-clamp recordings contained the following: 130 mM K⁺-gluconate, 5 mM NaCl, 2 mM MgCl₂, 10 mM HEPES, 0.5 mM EGTA, 2 mM ATP-K⁺, and 0.3 mM GTP-Na⁺; pH was adjusted to 7.25 with KOH. Spontaneous inhibitory postsynaptic currents (IPSCs) were recorded at –60 mV and isolated by bath application of 2 to 5 mM kynurenic acid. Access resistance was monitored throughout the recording period; cells were included for analysis only if the series resistance was less than 20 MΩ and the change in series resistance was less than 20% over the course of the experiment. Data were analyzed as described previously (Jia et al., 2005); very briefly, off-line analysis was performed using MiniAnalysis 5.5 (Synaptosoft, Decatur, GA), SigmaPlot 6.0 (SPSS Inc., Chicago, IL), and Excel 2000 (Microsoft, Redmond, WA). The holding current shift was measured as the difference in the holding current before and during drug application. IPSCs were detected and analyzed using MiniAnalysis as described previously (Jia et al., 2005). Unless otherwise indicated, averaged data are expressed as mean ± S.E.M. Statistical significance was assessed using Student's t test or one-way ANOVA with Dunnett's test, and p < 0.05 was considered statistically significant.

Recordings from HEK293 Cells Expressing Recombinant GABA_A-Rs. The cDNAs encoding the human α₁, mouse α₄, rat β₂, human γ_2s, and human δ subunits were subcloned into the pcDNA3.1 expression vector and transiently expressed in HEK293 cells (American Type Culture Collection, Rockville, MD), as described in Jia et al. (2005). Ligand-gated currents were recorded at room temperature (voltage-clamped at –60 mV) using an Axopatch 200 amplifier (Molecular Devices). The extracellular solution contained the following: 145 mM NaCl, 3 mM KCl, 1.5 mM CaCl₂, 1 mM MgCl₂, 6 mM d-glucose, and 10 mM HEPES; pH was adjusted to 7.4 with NaOH. The intracellular solution used to fill patch pipettes contained the following: 145 mM N-methyl-d-glucamine hydrochloride, 0.1 mM CaCl₂, 5 mM ATP-K, 1.1 mM EGTA, 2 mM MgCl₂, and 5 mM HEPES; pH was adjusted to 7.2 with KOH. GABA and/or isoflurane were applied rapidly (∼50-ms exchange time) to the cell via a multichannel motor-driven solution exchange device (Rapid Solution Changer RSC-100; Molecular Kinetics, Pullman, WA).

Concentration-response amplitude data were analyzed as described previously (Jia et al., 2005); concentration-response data for each individual cell were fitted (using a sum of least-squares method) to a Hill equation of the form as follows: I = I_max × [agonist]^nH/([agonist]^nH + EC₅₀^nH), where I is the peak current, I_max is the maximal whole-cell current amplitude, [agonist] is the agonist concentration, EC₅₀ is the agonist concentration eliciting a half-maximal current response, and n_H is the Hill coefficient. Peak current in response to an EC₂₀ concentration of GABA alone (I_EC20) was defined as the control response, and isoflurane was then preapplied for 20 s before coapplication with GABA(EC₂₀) to ensure that isoflurane had reached equilibrium with the receptors. Isoflurane-induced potentiation was calculated as the percentage increase in peak current relative to control. Currents elicited directly during preapplication of isoflurane were measured and normalized to I_EC20 to quantify the direct activation of GABA_A receptors by isoflurane in the absence of GABA. Statistical significance was assessed using a one-way ANOVA with a Dunnett's multiple comparison post-test. Data are presented as mean ± S.E.M.

Drugs and Preparation of Volatile Anesthetic Solutions. GABA, gabazine [4-[6-imino-3-(4-methoxyphenyl) pyridazin-1-yl] butanoic acid hydrobromide], and kynurenic acid (4-oxo-1H-quinoline-2-carboxylic acid) were purchased from Sigma-Aldrich (St. Louis, MO). Isoflurane [2-chloro-2-(difluoromethoxy)-1,1,1-trifluoro-ethane] was obtained from Abbott Laboratories (North Chicago, IL). A stock solution of 10 mM GABA was prepared daily. Isoflurane solutions were prepared by injection of liquid anesthetic with a gas-tight syringe (Hamilton, Reno, NV) into i.v. solution bags containing 100 ml of extracellular solution and were used within 2 h (Krasowski and Harrison, 2000). Isoflurane solutions were applied to the brain slice preparations via perfusion through polytetrafluoroethylene tubing; the drug-containing solution was applied to the HEK cells locally using a rapid solution changer. We have previously shown that losses of anesthetic to the air and the tubing using this approach are less than 5% (Krasowski and Harrison, 2000).

Generation and Use of α₄-Subunit Knockout Mice. The α₄-subunit gene knockout (KO) mice were generated as described previously (Chandra et al., 2006). All knockout (Gabra4^–/–) and wild-type (Gabra4^+/+) littermates used were age-matched and were on the same genetic background (129 × 1/S1 × C57BL/6J hybrid; F2–F6 generations). Those conducting the experiments were blind to genotype in all studies.

Results

Low Concentrations of Isoflurane Enhance Tonic Inhibition in VB Neurons. Voltage-clamp recordings were made in VB relay neurons held at –60 mV using a CsCl-based intracellular solution. A low concentration of isoflurane (25 μM) elicited a sustained inward current (Fig. 1A); the mean amplitude of this current was 10 ± 2 pA (n = 14). At higher concentrations, isoflurane elicited larger currents [85 μM: 23 ± 3 pA, n = 14; 250 μM: 51 ± 4 pA, n = 17 (Fig. 1, B and C)].

We next examined whether the currents induced by isoflurane were mediated by GABA_A-Rs. As shown in Fig. 2A, 20 μM gabazine, a specific GABA_A-R antagonist, not only completely blocked inhibitory synaptic currents (IPSCs) but it also blocked the inward current elicited by isoflurane (250 μM), indicating that these isoflurane-activated tonic currents are mediated by GABA_A-Rs. The additional outward current induced by the addition of gabazine (Fig. 2A) indicates the presence of a persistent background current in this neuron, consistent with the reports of tonic inhibition in thalamocortical relay neurons (Belelli et al., 2005; Cope et al., 2005; Jia et al., 2005; Chandra et al., 2006; Bright et al., 2007). Isoflurane is able to directly activate GABA_A-Rs (Yang et al., 1992; see also Fig. 5), as well as potentiate the action of GABA at these receptors (Jones and Harrison, 1993; see also Fig. 5), and so for the sake of simplicity, we will refer to the sustained current recorded from neurons in brain slices as the “tonic current” or “isoflurane-activated current.”

We then performed experiments using a voltage-ramp protocol to compare the reversal potential of isoflurane-activated currents (E_Iso) to the Cl^– equilibrium potential (E_Cl). Currents induced by a slow ramp voltage command (+40 to –40 mV, 10 s) were recorded before and after the perfusion of 250 μM isoflurane, and E_Iso was calculated from the subtracted traces as shown in Fig. 2B; E_Iso was 5 ± 1 mV (n = 6), which was very close to the predicted E_Cl (2.1 mV), calculated using the Nernst equation. These results support our interpretation that isoflurane-induced currents arise via the enhancement of a chloride conductance mediated by GABA_A-Rs.

Isoflurane Prolongs the Decay Time of IPSCs. We also examined the effects of isoflurane on synaptic GABA_A-Rs in the thalamus. In thalamic VB neurons, synaptic GABA_A-Rs consist primarily of α₁, β₂, and γ₂ subunits (Jia et al., 2005). Spontaneous IPSCs were readily observed in VB neurons and were completely blocked by gabazine (Fig. 2A). Isoflurane (25–250 μM) had no effect on either the amplitude or frequency of these IPSCs (Fig. 3). Low concentrations of isoflurane (25 μM) had no effect on the decay time of IPSCs (% change: 4 ± 4%, n = 10), but higher concentrations of isoflurane significantly increased IPSC decay time (85 μM, 28 ± 6%, n = 10; 250 μM, 133 ± 12%, n = 10), and these findings are completely consistent with observations made in hippocampal neurons (Jones and Harrison, 1993; Banks and Pearce, 1999; Nishikawa and MacIver, 2001; Caraiscos et al., 2004; Verbny et al., 2005).

Isoflurane Decreases the Excitability of VB Neurons. From a resting membrane potential of approximately –75 mV, most VB neurons in our recordings displayed “burst” firing in response to depolarizing current injection (Llinás and Jahnsen, 1982). Therefore, to facilitate measurements of firing rate, we depolarized the membrane potential to approximately –60 mV by constant current injection. At this membrane potential, VB neurons were generally silent but displayed sustained action potential (AP) firing in response to injection of depolarizing current. The amplitude of the current step (500-ms duration) was adjusted to induce ∼10 APs (Fig. 4A), corresponding to a firing frequency of ∼20 Hz, and we then compared the numbers of APs evoked by depolarizing current steps before and after isoflurane application. Isoflurane (25 μM) significantly reduced the number of evoked APs, from 9.5 ± 1.0 to 7.5 ± 1.0% (p < 0.01, n = 6). In addition to the effect on spike firing, isoflurane (25 μM) also decreased the membrane input resistance (R_m) to 90 ± 2% of control (p < 0.01, n = 6).

Isoflurane Directly Activates GABA_A α₄β₂δ Receptors Expressed in HEK293 Cells. The GABA_A-Rs activated at inhibitory synapses onto VB neurons are thought to be of the α₁β₂γ₂ subtype (Zhang et al., 1997; Huntsman and Huguenard, 2000), whereas the extrasynaptic receptors are thought to be α₄β₂δ (Belelli et al., 2005; Jia et al., 2005; Chandra et al., 2006). To compare the isoflurane sensitivity of these proposed synaptic and extrasynaptic GABA_A-R subtypes, we tested the effect of isoflurane on α₁β₂γ_2s and α₄β₂δ GABA_A-Rs expressed in HEK293 cells. Concentration-response curves for GABA (data not shown) revealed that α₄β₂δ receptors (EC₅₀ 2.1 ± 0.1 μM, n = 43) are more sensitive to GABA than α₁β₂γ_2s receptors (EC₅₀ 22.7 ± 2.1 μM, n = 20; p < 0.001), but the maximal GABA current was smaller in α₄β₂δ receptors than in α₁β₂γ_2s receptors (223 ± 16 and 1741 ± 193 pA, respectively), all consistent with earlier observations (Jia et al., 2005).

We then examined the modulation of these recombinant GABA_A-R subtypes over a range of concentrations of isoflurane (25–800 μM). Isoflurane was applied for 20 s to a HEK cell held under a voltage clamp before coapplication with GABA (EC₂₀) for another 20 s. Typical recordings from both α₁β₂γ₂ and α₄β₂δ receptors are shown in Fig. 5, A and B (two different HEK cells for each subtype). At all of the concentrations tested, isoflurane enhanced the amplitude of the GABA-evoked response to a greater degree in α₄β₂δ receptors than in α₁β₂γ₂ receptors (Fig. 5C).

In the presence of isoflurane, the peak current observed in response to GABA is actually the sum of two components, I_Direct (i.e., the current induced by isoflurane alone, due to direct receptor activation) and I_GABAEC20+ISO (i.e., the isoflurane-potentiated GABA-evoked current). In a more detailed analysis, we compared these two currents in α₁β₂γ₂ and α₄β₂δ receptors. As shown in Fig. 6A, isoflurane, over the entire range of concentrations tested (25–800 μM), directly activated α₄β₂δ receptors in the absence of GABA, whereas in α₁β₂γ_2s receptors, only higher concentrations of isoflurane (≥200 μM) were able to induce direct currents. In addition, isoflurane (at all of the concentrations tested) activated proportionally larger direct currents (I_Direct/I_GABAEC20) in α₄β₂δ receptors than in α₁β₂γ_2s receptors. At the same time, isoflurane was more potent as a modulator (in terms of potentiation of the response to GABA) at α₄β₂δ receptors than at α₁β₂γ_2s receptors (Fig. 6B). Note that these data illustrate the potentiating effect of isoflurane, calculated by subtracting I_Direct (Fig. 6A) from the total current (Fig. 5C).

Because some of the isoflurane currents were small in amplitude (∼10 pA), we decided to test the possibility that the currents were influenced by the change of solutions and the associated movement artifacts. Therefore, we used the same protocol but omitted the isoflurane. Under these test conditions, when switching from one stream of saline to another, there was no significant enhancement of the GABA response (α₁β₂γ_2s:0 ± 10%, n = 5; α₄β₂δ: –5 ± 5%, n = 7) or direct current response to the solution exchange (α₁β₂γ_2s:2 ± 1%, n = 5; α₄β₂δ: –4 ± 2%, n = 7).

Isoflurane-Evoked Currents Are Absent in VB Neurons from Gabra4^–/–Mice. In a final set of experiments, we used a knockout mouse strain to determine whether extrasynaptic GABA_A-Rs are required for the generation of isoflurane-induced currents in the thalamus. We have previously demonstrated that extrasynaptic GABA_A-Rs are absent in thalamic relay neurons from Gabra4^–/– mice (Chandra et al., 2006). No significant isoflurane-evoked currents were detected in VB neurons from Gabra4^–/– mice (25 μM: 0 ± 1 pA, n = 10; 85 μM: 0 ± 1 pA, n = 12; 250 μM: 1 ± 1 pA, n = 11). In contrast, wild-type (WT) neurons showed measurable isoflurane-evoked currents (25 μM: 8 ± 2 pA, n = 8; 85 μM: 17 ± 3 pA, n = 8; 250 μM: 44 ± 4 pA, n = 9), which were comparable to isoflurane-evoked currents recorded from standard C57BL/6 mice (Fig. 7). This difference between the genotypes was highly significant (p < 0.001 at all three concentrations of isoflurane). In contrast, we found that the modulation of IPSCs in VB neurons by isoflurane was similar in wild-type and α₄-subunit knockout mice. As shown in Fig. 7C, isoflurane had a comparable effect on the decay time of IPSCs recorded in VB neurons from wild-type and Gabra4^–/– mice. IPSC decay times in the absence or presence of isoflurane (at the specified concentration) were (WT versus KO) as follows: control, 15 ± 1 versus 14 ± 1 ms; 25 μM, 16 ± 2 versus 15 ± 1 ms; 85 μM, 19 ± 1 versus 19 ± 2 ms; 250 μM, and 33 ± 3 versus 31 ± 3 ms. These results are consistent with the hypothesis that extrasynaptic (and not synaptic) GABA_A-Rs mediate the isoflurane-evoked tonic current in thalamic neurons.

Discussion

Volatile anesthetics have been shown to modulate GABA_A-R function in neurons and heterologous expression systems (Jones and Harrison, 1993; Krasowski et al., 1998; Banks and Pearce, 1999; Krasowski and Harrison, 2000; Li and Pearce, 2000; Nishikawa and MacIver, 2001; Nishikawa and Harrison, 2003; Hemmings et al., 2005). In the present study, we found that isoflurane (25–250 μM) induced sustained currents in thalamic relay neurons and that this was completely dependent on the presence of extrasynaptic GABA_A-Rs (α₄β₂δ subtype). Recordings from α₄β₂δ receptors expressed in HEK293 cells demonstrated that isoflurane not only potentiated the GABA response but it also directly activated these receptors in the absence of GABA. Therefore, the sustained current evoked by isoflurane in thalamic relay neurons probably results from the sum of the direct activation of extrasynaptic GABA_A-Rs and the enhancement of the action of ambient GABA on these receptors.

General anesthesia in mammals is a complex phenomenon, involving a combination of desirable effects such as amnesia, hypnosis, and immobility (Campagna et al., 2003). Although immobilization is commonly used as a measure of anesthetic potency (Hemmings et al., 2005), 3-fold lower concentrations are required for sedation and hypnosis (Dwyer et al., 1992). GABA_A-Rs may not be involved in the immobilizing action of inhaled anesthetics, which seems to occur at the level of the spinal cord, and is resistant to GABA antagonists (Sonner et al., 2003), but there is strong evidence that GABA_A-Rs are involved in the sedation and hypnosis induced by many general anesthetics (reviewed by Hemmings et al., 2005).

The thalamus relays sensory information to the appropriate modality-specific areas in the sensory cortex and also participates in the transitions between waking and sleep states (Saper et al., 2001; Steriade, 2003). The thalamus is therefore a potential target for the sedative and hypnotic actions of volatile anesthetics, and it has been previously argued that the thalamus is a principal site of drug action for generating the anesthetized state (Angel, 1993). A variety of studies support the idea that volatile anesthetics modulate thalamic function. At the whole-brain level, positron emission tomography imaging demonstrates that the volatile anesthetics isoflurane and halothane both produce a particularly large decrease in glucose utilization in the thalamus (White and Alkire, 2003). Consistent with a decrease in metabolic activity, the firing rate of thalamocortical relay cells in vivo is strongly suppressed by anesthetic concentrations of isoflurane (Detsch et al., 1999).

AK⁺ conductance mechanism has been implicated in the inhibitory actions of isoflurane in the thalamus in vitro (Ries and Puil, 1999a,b), but in other studies, the effect of isoflurane to inhibit spike firing induced by a mechanical stimulus in the thalamus in vivo is blocked by local application of the GABA_A-R antagonist bicuculline (Vahle-Hinz et al., 2001). In the present study, we demonstrate that tonic currents activated by isoflurane in thalamic neurons in vitro are mediated by GABA_A-Rs, because the isoflurane-induced currents were totally blocked by a specific GABA_A-R antagonist, gabazine. In addition, the reversal potential of isoflurane-induced currents (E_Iso) is close to the calculated equilibrium potential for chloride ions (E_Cl) in our experiments and far from K⁺ equilibrium potential (E_K).

Extrasynaptic GABA_A-Rs and tonic inhibition have been observed in the thalamus, hippocampus, dentate gyrus, cortex, and cerebellum (Brickley et al., 1996; Nusser and Mody, 2002; Jia et al., 2005; Keros and Hablitz, 2005). The presence of a low concentration of GABA in the extracellular space leads to persistent activation of the extrasynaptic GABA_A-Rs, and the resulting tonic inhibition regulates the excitability of individual neurons and the activity of neural networks (Mody and Pearce, 2004; Semyanov et al., 2004; Farrant and Nusser, 2005; Jia et al., 2007). In the thalamus, δ subunit containing extrasynaptic GABA_A-Rs appear to be a common target for a variety of sedative and hypnotic agents, including gaboxadol (4,5,6,7-tetrahydroisoxazolo[5,4-c]pyridin-3-ol) (Belelli et al., 2005; Cope et al., 2005; Jia et al., 2005) and tetrahydrodeoxycorticosterone (Cope et al., 2005). Several laboratories also report that extrasynaptic GABA_A-Rs containing the δ subunit are sensitive to low concentrations of ethanol (Mody et al., 2007; Olsen et al., 2007; Santhakumar et al., 2007; Smith and Gong, 2007; but see: Borghese and Harris, 2007; Botta et al., 2007).

It is noteworthy that not all extrasynaptic GABA_A-Rs contain δ subunits; in hippocampal CA1 pyramidal neurons, for example, one population of extrasynaptic GABA_A-Rs contains the α₅ subunit (Caraiscos et al., 2004). Like extrasynaptic receptors containing δ subunits, α₅ subunit containing GABA_A-Rs are sensitive to a spectrum of anesthetic agents. At the same concentration of isoflurane as used in the present study, Caraiscos et al. (2004) observed that 25 μM isoflurane potentiated the tonic current in cultured hippocampal pyramidal neurons obtained from wild-type, but not α₅-subunit knockout, mice. In addition, the i.v. anesthetic etomidate was shown to potentiate tonic currents recorded in cultured hippocampal pyramidal neurons expressing the GABA_A-R α₅ subunit. In this case, the amnestic, and not sedative-hypnotic, effect of the drug is dependent on α₅-subunit expression (Cheng et al., 2006). These studies suggest that extrasynaptic GABA_A-Rs are a common molecular target for central nervous system depressants (Orser, 2006).

In the thalamus, extrasynaptic GABA_A-Rs that contain the δ subunit also seem to require inclusion of the GABA_A-R α₄ subunit because tonic currents are absent in relay neurons from GABA_A-R α₄-subunit knockout (Gabra4^–/–) mice (Chandra et al., 2006). Our experiments on thalamic relay neurons from Gabra4^–/– mice suggest that this population of extrasynaptic GABA_A-Rs exclusively mediates the sustained currents elicited by low concentrations of isoflurane, although the prolongation of synaptic currents by isoflurane (≥85 μM) is independent of α₄-subunit expression, because IPSCs were potentiated to a similar extent by isoflurane in wild-type and α₄-subunit knockout mice. We conclude that the α₄ subunit containing GABA_A-Rs provides the molecular basis for the isoflurane-induced current and that this current is completely independent of synaptic GABA_A-Rs.

We examined the pharmacological properties of heterologously expressed GABA_A-Rs and configured in order to resemble natively expressed extrasynaptic (α₄β₂δ) receptors, which are more sensitive to GABA than α₁β₂γ_2s receptors (Brown et al., 2002; Storustovu and Ebert, 2006). We found that isoflurane is a more potent modulator at α₄β₂δ GABA_A-Rs (which are analogous to native extrasynaptic receptors) than at α₁β₂γ₂ receptors (which are analogous to native synaptic receptors). One interesting finding is that isoflurane, at a concentration as low as 25 μM, directly activates α₄β₂δ receptors in the absence of GABA. In contrast, only high concentrations of isoflurane (≥200 μM) are able to induce significant direct currents in α₁β₂γ_2s receptors, consistent with earlier observations (Krasowski et al., 1998; Krasowski and Harrison, 2000; Raines et al., 2003).

In conclusion, we show that extrasynaptic GABA_A-Rs in the thalamus are potently activated by isoflurane. Because the thalamus plays a critically important role in both sensory processing and sleep regulation, we suggest that the sedative and hypnotic actions of isoflurane at subanesthetic concentrations may be mediated in part through this population of receptors.