Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

With a minimum alveolar concentration (MAC) of 104% in humans, nitrous oxide (N₂O) would require hyperbaric conditions to act as an anesthetic gas (Gonsowski and Eger, 1994). Therefore, in clinical practice, N₂O is usually combined with other anesthetics, such as isoflurane (ISO). N₂O has been reported to affect various ligand-gated ion channels, e.g., glutamate receptors (Macdonald and Ramsey, 1995; Jevtovic-Todorovic et al., 1998) and nicotinic acetylcholine receptors (Wachtel, 1995). Currently used intravenous and volatile anesthetics affect the -aminobutyric acid_A receptor (GABA_AR) (Tanelian et al., 1993). The activation of this ligand-gated chloride channel is of pivotal importance for synaptic inhibition (Möhler et al., 1996). A receptor assembly, consisting of 2₁, 2₂, and 1₂ subunits, predominates in the mammalian brain (Chang et al., 1996; Möhler et al., 1996). Several animal studies demonstrate the involvement of the GABA_AR system in the effects of N₂O on, e.g., visually evoked potentials (Dzoljic et al., 1996), analgesia (Emmanouil and Quock, 1989), and anxiolysis (Emmanouil et al., 1994). A direct enhancing effect of N₂O on a GABA-evoked response has been shown in acutely dissociated hippocampal neurons (Dzoljic and van Duijn, 1998).

Most studies report a potentiating effect of volatile anesthetics on GABA_AR channels, i.e., an increase in GABA-evoked chloride flux (Moody et al., 1988; Harrison et al., 1993; Zimmerman et al., 1994; Jenkins et al., 1999). Some studies also found an additional blocking effect of ISO on the GABA-evoked response (Edwards and Lees, 1997; Adelsberger et al., 1998; Neumahr et al., 2000). However, a block of GABAergic transmission is not easily reconciled with the apparent decrease of neuronal excitability during anesthesia. Various studies suggest that volatile anesthetics, such as ISO, halothane, and enflurane, lower the excitability of central neurons by prolonging the decay of GABA-mediated inhibitory postsynaptic currents (IPSCs) (Jones and Harrison, 1993). Other studies report that volatile anesthetics prolong the decay and reduce the amplitude of GABA_A IPSCs (Banks and Pearce, 1999). It has been suggested that a prolonged flickering of GABA_AR channels, which is caused by the anesthetic blocking the channel pore, could be responsible for the increased duration of IPSCs (Jones and Harrison, 1993). Evidence in favor of this assumption comes from observations made at the nicotinic acetylcholine receptor. At this receptor, the single channel burst duration increases on application of channel-blocking compounds (Beam, 1976; Neher and Steinbach, 1978).

There is evidence from studies on GABA receptors in insects that ISO apparently shares the binding site for picrotoxin (Edwards and Lees, 1997), a GABA antagonist that binds to the channel lumen of GABA_AR (Gurley et al., 1995). An open-channel block by ISO was also suggested by studies performed at a GABAergic crayfish muscle synapse (Adelsberger et al., 1998). In addition to its effects at the GABA_AR, ISO is an open-channel blocker at the nicotinic acetylcholine receptor (Scheller et al., 1997). At this site, ISO elicits a transient increase in the agonist-evoked current, when the agent is rapidly withdrawn from the receptor. Such rebound currents are established features of open-channel blockers (Dilger and Liu, 1992). They are considered to signal the unbinding of an open-channel blocker from its binding site in the channel pore (Scheller et al., 1997; Adelsberger et al., 1998; Neumahr et al., 2000).

In the present study, the effects of coadministered ISO and N₂O were investigated on GABA channel activity to determine whether the additive effect observed clinically could be explained by interactions with the GABA_AR.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Preparation. Human embryonic kidney cells (HEK293; Deutsche Sammlung von Mikroorganismen und Zellkulturen, Braunschweig, Germany) were maintained in minimum essential medium, supplemented with 10% fetal calf serum, 4 mM L-glutamine, 100 U/ml of penicillin, and 100 U/ml of streptomycin, in an atmosphere of 5% CO₂, 95% air, and 100% relative humidity at 37°C.

Electrophysiology. For the experiments, performed at 20-23°C, the medium was replaced by extracellular solution containing 162 mM NaCl, 5.3 mM KCl, 0.67 mM Na₂HPO₄, 0.22 mM KH₂PO₄, 2 mM CaCl₂, 15 mM HEPES, 5.6 mM glucose, pH 7.4 adjusted with NaOH. The patch-clamp technique was used to measure GABA-evoked chloride currents under whole-cell voltage-clamp (30 mV) conditions. Borosilicate glass pipettes (GC150TF-10; Clark Electromedical Instruments, Pangbourne Reading, UK) were pulled, using a two-step horizontal puller (Zeitz Instruments, Augsburg, Germany), and heat polished. The resulting tips had a series resistance of 4 to 9 M. Pipettes were filled with intracellular solution containing 140 mM KCl, 11 mM EGTA, 10 mM HEPES, 10 mM glucose, 2 mM MgCl₂, 1 mM CaCl₂, pH 7.3 adjusted with KOH. GABA-induced currents were recorded with an Axopatch 200B patch-clamp amplifier, low-pass filtered at a cutoff frequency of 5 kHz, and then digitized at 10 kHz with a digidata 1200 A/D converter, performed with pClamp 6.0 software (all from Axon Instruments, Foster City, CA).

Agonist and Drug Application. To match the rapid kinetics of ligand-activated ion channels, a piezo-driven system for fast exchange of solutions was used (Franke et al., 1987). GABA was applied alone or combined with the drug under investigation to the whole cell patches. The drugs were administered to the cell via a "liquid filament", i.e., a tiny jet of solution, discharged from a borosilicate glass tube (inner diameter 0.15 mm) inside the recording chamber, which was perfused by extracellular solution (Fig. 1A). This technique allows for a complete exchange of solutions in the vicinity of the cell, held in the whole cell mode, within 1 ms as measured by activation of voltage-operated calcium channels in separate experiments (data not shown). The liquid filament consisted of extracellular solution containing indicated concentrations of GABA alone (controls) or in combination with the respective agent (test solution). The test solution was applied to the whole cell patch in pulses of 1.5 s. An interval of 10 s between the pulses allowed full recovery of the GABA_AR channels from a desensitized state. Each current trace was averaged from at least three stable responses.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   A, piezo-driven liquid filament switch was used to achieve exchange of solutions in the vicinity of the whole cell patch within <1 ms. Test solution and background solution form a laminar flow system. The device, upon activation, shifts the tube upward by 20 µm to immerse the cell in the test solution containing the agonist, inducing GABA-gated currents (upward arrow). The GABA application is terminated after 1.5 s by the downward shift of the liquid filament upon deactivation of the piezo crystal (dashed downward arrow). An interval of 10 s between the pulses of 1.5 s allowed full recovery of the receptors from desensitized states. Each current trace was averaged from at least three recordings. B, recombinant 122 GABAAR channels were transiently expressed in HEK293 cells. Application of GABA (103 M) to the whole cell patches elicited the maximum GABA response (top). In this experiment, GABA induced a peak current of 2350 pA. The current decay in the presence of the agonist is due to desensitization. At the end of the application of GABA, the current decayed to a baseline level. GABA (5 × 106 M, same experiment, below) induced a current with an amplitude of 534 pA. This nonsaturating concentration of GABA did not desensitize the 122 GABAAR. GABA (5 × 106 M) elicited 26 ± 3% (n = 28) of the maximum GABA response.

1

MAC and Solubility of N₂O. The MAC value of N₂O in humans is 1.04 atm (Hornbein et al., 1982). Based on the solubility coefficient for 37°C (Wilhelm et al., 1977), the MAC equivalent for dissolved N₂O was calculated to 20.6 mM. The solubility of N₂O in the extracellular solution, prepared for this study, was measured using a technique for a volumetric evaluation of the solubility of gases in fluids (Krauss and Gestrich, 1977). This technique was modified by Dr. Karl-Heinz Meister (Linde AG, Höllriegelskreuth, Germany). The measurements were performed at 20°C and 1 atm. N₂O was applied to the extracellular solution under the same conditions as in the patch-clamp experiments. Saturation (>95%) was achieved within 2.5 ± 0.3 min. A calculated solubility coefficient of 0.654 ± 0.010 [published value for N₂O in H₂O (20°C, 1 atm) is 0.6788 (Wilhelm et al., 1977)], resulted in a concentration of N₂O of 29.2 ± 0.4 mM in the extracellular solution at 20°C and 1 atm.

Statistical Analysis. Peak current and time to peak (10-90%) were measured using automated detection algorithms (AxoGraph software for MacOS). Data are presented as means ± S.E.M. with the number of experiments indicated. Statistical analysis was performed using Student's paired t test (p < 0.05 was considered as significant).

Sources of Anesthetics and Chemicals. GABA was obtained from Sigma Chemical Co. (St. Louis, MO), N₂O from Linde AG, and isoflurane (Forene) from Deutsche Abbott GmbH (Wiesbaden, Germany).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Agonist Application Using the Piezo-Driven Liquid Filament. At recombinant ₁₂₂ GABA_AR channels, transiently expressed in HEK293 cells, GABA (5 ×10⁶ M) elicited 26 ± 3% (n = 28) of the maximum GABA response, evoked by a saturating concentration of GABA (10³ M). GABA was applied by means of a piezo-driven liquid filament switch to cells recorded in the whole cell patch mode (Fig. 1A). The responses to GABA (5 × 10⁶ M), which was used in the following experiments as a standard test, did not desensitize (Fig. 1B) and reversed at 0 mV, corresponding to the equilibrium potential for chloride ions under the chosen experimental conditions.

Biphasic Effect of ISO on GABA-Induced Currents. Increasing the concentration of ISO in steps from 0.15 to 1.2 mM resulted in a bell-shaped concentration-response curve for GABA-induced currents. The maximum increase in current (1.51 ± 0.14-fold) was seen at 0.45 mM ISO (Fig. 2A). This finding suggests a dual effect of ISO on GABA-induced currents, i.e., an enhancing effect that predominates at lower ISO concentrations (0.8 mM), and a prevailing blocking effect at higher ISO concentrations (>0.8 mM).

View larger version (27K): [in this window] [in a new window]   Fig. 2.   A, within each experiment, GABA (5 × 106 M, control) was applied to the whole cell patches, with or without N2O (29.2 mM), and with or without increasing concentrations of ISO. Solutions were prepared as described under Materials and Methods. The amplitude of the respective control current (GABA alone) was set for 1.0. N2O alone increased the currents through 122 GABAAR channels 1.54 ± 0.10-fold (n = 47) versus control (5 × 106 M GABA). GABA, combined with increasing concentrations of ISO, revealed a bell-shaped concentration-response curve. , 0.15, 0.45, 0.75, 1.0, and 1.2 mM ISO resulted in 1.31 ± 0.10-, 1.51 ± 0.14-, 1.49 ± 0.19-, 1.27 ± 0.17-, and 0.83 ± 0.12-fold current amplitudes. , in combination with N2O, 0.15, 0.45, 0.75, 1.0, and 1.2 mM ISO resulted in 1.57 ± 0.14-, 1.64 ± 0.16-, 1.63 ± 0.22-, 1.25 ± 0.16-, and 0.82 ± 0.10-fold current amplitudes (means ± S.E.M., current amplitudes were normalized to the control; *p < 0.05, **p < 0.01 versus control, n = 7-15 for each concentration of ISO). B, raw current traces are shown: N2O and 0.6 mM ISO increased the GABA-induced current from 466 pA (control) to 686 pA and 679 pA, respectively. N2O and 0.6 mM ISO combined were not worth mentioning as an additive in increasing the GABA-induced current (719 pA). A rebound current, induced by the withdrawal of 0.6 mM ISO, was reversibly increased by N2O. C, amplitudes of these rebound currents increased with increasing concentrations of ISO. N2O further increased the rebound currents (means ± S.E.M., *p < 0.05 ISO + N2O versus ISO, n = 5-13 for each concentration of ISO).

N₂O Increased GABA-Induced Currents. The extracellular solution containing GABA (5 × 10⁶ M) was saturated with N₂O and applied immediately. The concentration of N₂O was evaluated as described under Materials and Methods. N₂O (29.2 mM) significantly increased the response to GABA 1.54 ± 0.10-fold (n = 47, Fig. 2A). After washout, the effect of N₂O was reversible (data not shown; for details, see Hapfelmeier et al., 2000).

Coapplication of ISO and N₂O. ISO and N₂O were coapplied and the effect on GABA-induced currents was studied. On the one hand, the enhancing effect of N₂O on the GABA-evoked response was not significantly affected by the addition of ISO (Fig. 2A). On the other hand, there was no significant difference between the potentiating effect of ISO alone and ISO coapplied with N₂O (Fig. 2A).

ISO Additionally Induced an Open-Channel Block. The rapid withdrawal of ISO from the whole cell patch induced a transient increase in the current response (Fig. 2B). This rebound current increased with increasing ISO concentrations (Fig. 2C). The rebound currents suggest that ISO evokes a dose-dependent open-channel block at the ₁₂₂ GABA_AR. Since the blocking effect of ISO is more prominent at high concentrations, the phenomena of rebound currents was also studied applying 1.5 and 15 mM ISO combined with a range of GABA concentrations (10⁹-10³ M). GABA was applied alone (Fig. 3A) and combined with 15 mM ISO (Fig. 3, B and C) or 1.5 mM ISO (Fig. 3C). An increasing number of blocked GABA_AR, prior to the unbinding of ISO, is probably the reason for the clear dose-response relationship between the amplitude of the rebound current and the concentration of ISO and GABA, respectively (Fig. 3, B and C). It is a widely held belief (Scheller et al., 1997; Adelsberger et al., 1998; Neumahr et al., 2000) that these rebound currents reflect the transition from a channel-blocked to a channel reopened state. Compatible with this view, similar rebound currents were observed, when the GABA_AR channel blocker picrotoxin or the volatile anesthetic sevoflurane was applied. In contrast, the rapid withdrawal of the competitive antagonist bicuculline did not induce any rebound currents (Fig. 4).

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Increasing concentrations of GABA (109-103 M) were applied alone (A) or combined with 15 mM ISO (B) to the same whole cell patch. A saturated concentration of ISO (15 mM) exhibited the maximum blocking effect of ISO on the GABAAR. The withdrawal of ISO from the whole cell patch induced rebound currents (arrow), suggesting the open-channel block by ISO. The rebound currents, reflecting the transition from open-channel-blocked to reopened channels, increased with increasing GABA concentrations. C, averaged amplitudes of the rebound currents (means ± S.E.M., n = 4). Rebound currents, appearing at the end of the application pulse (GABA combined with 1.5 or 15 mM ISO), increased with increasing concentrations of GABA.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   GABA (103 M) was coapplied with the noncompetitive GABAAR antagonist picrotoxin, with the volatile anesthetic sevoflurane, or with the competitive antagonist bicuculline. A, rebound current (arrow) was induced by the rapid discontinuation of picrotoxin or of sevoflurane. B, rapid withdrawal of bicuculline did not induce any rebound currents.

N₂O Increased the Rebound Currents. The rebound currents induced by the withdrawal of ISO were significantly increased by N₂O (29.2 mM) (Fig. 2, B and C). N₂O (29.2 mM), which increased the response to GABA, neither enhanced the potentiating effect of ISO, nor did it induce any rebound current (Fig. 2B). These findings suggest that, in contrast to ISO, N₂O exhibited no channel block at the GABA_AR.

Modeling of the Effects of N₂O and ISO. Our experimental data (Fig. 5) can be well described by a modified kinetic model (Fig. 6) for the GABA_AR (Jones and Westbrook, 1995; Haas and Macdonald, 1999). In respect of the theories of kinetic modeling (Colquhoun and Hawkes, 1981; Colquhoun and Sakmann, 1985), the rate constants were chosen based on our experimental data and kinetic studies on the GABA_AR used in this study. Macroscopic current modeling was performed using BIOQ-Biochemical Equations software (Parnas & Parnas Neurobiology Lab, Hebrew University, Jerusalem, Israel). The simulated currents, which fit well to the experimental data, are shown in Fig. 7.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Withdrawal of GABA alone (A) or combined with N2O (B) did not induce any rebound currents (see insets). The rebound current induced by the withdrawal of ISO (0.6 mM) (C) suggested the open-channel block by ISO. N2O might enhance the action of GABA independently from interactions with ISO and, thus, provide more open channels to be blocked by ISO, which resulted in increased rebound currents (D). Averaged current amplitudes and times to peak are expressed as means ± S.E.M.

View larger version (31K): [in this window] [in a new window]   Fig. 6.   A, modified model for the GABAA receptor (Haas and Macdonald, 1999) was used for computer-based simulations without any respect of desensitized states of the receptor, since, in our experiments, GABA (5 × 106 M) did not induce desensitization. This model provides closed (C), opened (O), and blocked (B ISO) states. For the GABAAR used in this study, a Hill coefficient of 2.2 ± 0.4 was calculated previously (Jahn et al., 1997). Thus, three binding steps for GABA were suggested. B, N2O induced a leftward shift in the dose-response curve of GABA associated with decreased times to peak (Hapfelmeier et al., 2000) (see also Fig. 5B). This suggests an increased binding rate (1.45-fold) of GABA. C, ISO exhibited two different effects on this GABAAR: (i) ISO (0.6 mM) also induced a leftward shift in the dose-response curve of GABA (data not shown), however, associated with increased times to peak (see also Fig. 5C), suggesting a decreased unbinding rate (0.5-fold) of GABA; (ii) ISO induced an open-channel block. The rates of blocking and unblocking of ISO were derived from experiments applying ISO concentrations ranging from 0.15 to 15 mM (data not shown). D, in this possible model, the actions of N2O and ISO were combined independently.

View larger version (23K): [in this window] [in a new window]   Fig. 7.   Simulated current traces derived from the respective model in Fig. 6. The traces represent the opened state (O) of the channels. Simulated current amplitudes and times to peak (A-D) correspond well to the data in Fig. 5. C, this model reflecting the actions of ISO provides a block probability (B ISO state) of 0.30 for the channels along with a relatively small rebound current of 0.047. D, in this model, N2O increased the block probability to 0.35 and the rebound current to 0.059.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, we investigated the effects of ISO and N₂O alone or in combination on GABA-induced currents recorded from HEK293 cells expressing a rat recombinant _122L GABA_AR. We used a piezo-driven application system for fast exchange of solutions (Fig. 1).

Effects of N₂O and ISO on GABA-Induced Currents. The present experiments show that both N₂O and ISO at clinically relevant concentrations enhanced GABA-induced currents in cells carrying the most ubiquitous GABA_AR assembly (₁₂₂) present in the mammalian central nervous system. However, N₂O and ISO are different in affecting the activation kinetics of the currents. According to the presented model, it is possible that N₂O and ISO independently interact with the GABA_AR. Furthermore, in contrast to N₂O, ISO additionally induces a blocking effect that becomes more prominent at higher concentrations.

ISO Induces an Open-Channel Block at the GABA_AR. Conspicuously, the rapid withdrawal of ISO induced a marked dose-dependent transient increase in the response to GABA. The most parsimonious explanation for this observation is that ISO evokes a dose-dependent open-channel block at the ₁₂₂ GABA_AR (Adelsberger et al., 1998; Neumahr et al., 2000). Compatible with our view, rebound currents were observed, when the GABA_AR channel blocker picrotoxin (see Introduction) was applied (Fig. 4). Both ISO and picrotoxin bind to a site inside the channel lumen (Edwards and Lees, 1997). Similar rebound currents were seen applying channel blockers at the nicotinic acetylcholine receptor (Dilger and Liu, 1992; Scheller et al., 1997).

Effects of Coadministered N₂O and ISO. When coadministered, the enhancing effects of ISO and N₂O on the fast GABA-induced response were not additive. However, the rebound currents elicited by the rapid withdrawal of ISO were clearly enhanced by N₂O. This suggests that the ISO-induced open-channel block at the ₁₂₂ GABA_AR was enhanced by N₂O. An additional enhancement of GABA_AR activation by N₂O (our data; Dzoljic and van Duijn, 1998; Yamakura and Harris, 2000), independent from its interaction with ISO, might even provide more open channels to be blocked by ISO. Based on a modified model of the GABA_AR (Haas and Macdonald, 1999), the suggested mechanism is depicted in Fig. 6.

Impact of Open-Channel Block for Synaptic Transmission. It is a widely held assumption that an open-channel block delays the transition from an "open" to a "closed" state of a channel. This delay induces a channel flickering with prolonged single channel burst durations (Neher and Steinbach, 1978), which may result in prolonged GABA-IPSCs (Jones and Harrison, 1993) or in prolonged cholinergic neuromuscular transmission (Legendre et al., 2000). A prolongation of GABA_AR-mediated IPSCs by volatile anesthetics was observed at the ISO concentration of 0.6 mM (Banks and Pearce, 1999). The prolongation of these IPSCs will result in a net enhancement of GABAergic synaptic transmission, despite the decrease in IPSC amplitude observed in this study (Banks and Pearce, 1999). In our experiments, ISO (0.6 mM) induced a significant rebound current (Fig. 2C), indicating the open-channel block of GABA_AR channels. The present data suggest that N₂O could enhance the prolonging effect of ISO on GABA_AR-mediated IPSCs by an increase in open-channel block. Thus, the effect of coadministered N₂O and ISO on GABA-IPSCs and miniature IPSCs recorded from, e.g., hippocampal slices, could be a major point of further investigations. Taken together, the findings may provide an explanation how N₂O enhances ISO actions under clinical conditions.