Additive Effects of Sevoflurane and Propofol on γ-Aminobutyric Acid Receptor Function

γ-Aminobutyric acid type A receptor complementary DNAs (cDNAs; gift from Neil Harrison, Ph.D., Professor of Pharmacology and Director, C.V. Starr Laboratory of Molecular Neuropharmacology, Department of Anesthesiology, Weill Cornell Medical College, New York, New York) were expressed via  the vector pCIS2 in human embryonic kidney (HEK) 293 cells (American Type Culture Collection, Manassas, VA) as previously described.17HEK-293 cells were cultured on poly-d-lysine–treated coverslips in a solution containing Eagle minimum essential medium supplemented with 5% fetal bovine serum (Hyclone, Logan, UT), l-glutamine (0.292 μg/ml), penicillin G sodium (100 U/ml), and streptomycin sulfate (100 μg/ml). For the transient expression of GABA^Areceptors, cells were transfected using the CaPO⁴precipitation technique.18,19The GABA^Areceptor cDNAs and adeno-associated virus–green fluorescent protein cDNA (gift from H. Trent Spencer, Ph.D., Assistant Professor, Department of Pediatrics, Division of Hematology/Oncology and Bone Marrow Transplantation, Emory University School of Medicine, Atlanta, Georgia) were precipitated for 30 min at room temperature in a 160-μl solution containing 125 mm CaCl², 140 mm NaCl, 750 μm Na²HPO⁴, and 2.5 μg of each cDNA. After 30 min, the mixture was added to cells grown on coverslips. The cDNA was in contact with the HEK cells for 24 h in an atmosphere containing 3% CO²(37°C) before being removed and replaced with fresh culture medium in an atmosphere of 5% CO²(37°C).

Coverslips of transfected cells were transferred 48–72 h after cDNA removal to a recording chamber and perfused continuously with extracellular solution (145 mm NaCl, 3 mm KCl, 1.5 mm CaCl², 1 mm MgCl², 6 mm D-glucose, 10 mm HEPES/NaOH adjusted to pH 7.4). Whole cell patch clamp recordings from fluorescing HEK-293 cells (voltage clamped at −60 mV) were made using a Multiclamp 700A amplifier (Axon Instruments, Foster City, CA) as described previously.20The resistance of the patch pipette was 4–6 MΩ when filled with intracellular solution (145 mm N -methyl-d-glucamine hydrochloride, 5 mm dipotassium ATP, 1.1 mm EGTA, 2 mm MgCl², 5 mm HEPES/KOH, 0.1 mm CaCl²adjusted to pH 7.2). In addition to the continuous bath perfusion with extracellular medium, solutions including GABA, general anesthetics, or both were applied rapidly to the cell using a motor-driven solution exchange device (Rapid Solution Changer RSC-160; Molecular Kinetics, Indianapolis, IN). Solutions were exchanged within approximately 50 ms. Laminar flow out of the rapid solution changer head was achieved by driving all solutions at identical flow rates (1.00 ml/min) via  a multichannel infusion pump (KD Scientific, Holliston, MA). The solution changer was driven by protocols in the acquisition program pCLAMP 9 (Axon Instruments).

For GABA concentration–response measurements, cells were superfused with extracellular saline before application of one of eight GABA concentrations for 2 s, followed by a return to saline for at least 8 s before any subsequent GABA application. Below 100 μm GABA, the responses did not desensitize; at and above 100 μm, the amplitude of the responses declined by 10–15% in the continued presence of the agonist. Responses were low-pass-filtered (100 Hz; −3 dB, four-pole Bessel) and digitized with a 1322A interface (Axon Instruments) using pCLAMP 9 and stored for off-line analysis. Because intracellular and extracellular solutions contained equal chloride concentrations (145 mm), the chloride equilibrium potential was around 0 mV. All experiments were performed at room temperature (21°–24°C).

Stock solutions of GABA and propofol were diluted in extracellular solutions shortly before use. Sevoflurane solutions were prepared by injection of liquid anesthetic with a gas-tight syringe (Hamilton, Reno, NV) into intravenous drip bags containing defined volumes of extracellular solutions (100 ml) and used for up to 4 h. Clinically relevant concentrations of general anesthetics were used throughout the study; the aqueous concentration for 1 minimum alveolar concentration (MAC) sevoflurane was taken to be 330 μm,5and the anesthetic EC⁵⁰(AC⁵⁰) concentration for propofol was taken to be 2 μm.21Losses of general anesthetics in this perfusion system have been measured using gas chromatography and typically represent only 5–10% of the initial total drug concentration.22Sevoflurane was obtained from Abbott Laboratories (North Chicago, IL), and propofol (2,6 di-isopropylphenol) was obtained from Sigma (St. Louis, MO).

For each GABA exposure, the peak current amplitudes were measured and the GABA concentration–response data for each cell with and without general anesthetic were extracted from the raw data using our own software package. The analysis software was written to calculate nonlinear dose response curve parameters using Visual Basic macros within Microsoft Excel (Microsoft Corp., Redmond, WA) to facilitate efficient data organization. Dose–response parameters were optimized using GRG2, a version of the Generalized Reduction Gradient algorithm included in Microsoft Excel.23Using iterative processing and extensively automated file handling, we were able to process multiple data streams simultaneously.

The current peaks were fitted to a Hill equation of the form I = I^max*[GABA]^nH/([GABA]^nH+ EC^50nH), where I is the peak of each current, I^maxis the maximum whole-cell current amplitude, [GABA] is the GABA concentration, EC⁵⁰is the GABA concentration eliciting a half-maximal current, and n^His the Hill coefficient.

Concentration–response relations were recorded in the absence and presence of general anesthetic in the same cell. This enabled us to determine the control and anesthetic-modulated GABA EC⁵⁰s, defined as EG⁵⁰and EG⁵⁰′ respectively. The fractional effect on EC⁵⁰was defined as (1 − EG⁵⁰′/EG⁵⁰).

Concentration–effect relations were calculated for propofol and sevoflurane. The fractional effects of anesthetic alone on GABA EC⁵⁰were fitted to a Hill equation of the form E =[anesthetic]^slope/([anesthetic]^slope+ C^50slope), where E is the fractional effect of the anesthetic on GABA EC⁵⁰. For sinistral concentration–response shifts (0 < E ≤ 1), [anesthetic] is the concentration of propofol or sevoflurane, C⁵⁰is the anesthetic concentration eliciting a half maximal effect, and slope is the Hill coefficient for the anesthetic concentration–effect relation. Statistical significance was assessed using a one-way analysis of variance with Dunnett test for multiple comparisons. Data are presented as mean ± SEM.

The response surface for the modulation of GABA^Areceptor function by propofol and sevoflurane was determined using the method of Minto et al.  24Briefly, anesthetic concentrations were normalized to the C⁵⁰s of each drug:

We used U^Aand U^Bto define a new variable θ, the drug ratio of A and B.

In the absence of sevoflurane (propofol alone), U^A= 0 and hence θ= 1. Conversely, in the absence of propofol (sevoflurane alone), U^B= 0 and hence θ= 0. When equal quantities of the drugs are present, θ= 0.5.

Substituting these terms into the Hill equation, we obtain the following function:

U⁵⁰(θ) defines the potency of the drug combination relative to the potency of either drug alone, and γ(θ) is the sigmoidicity of the response surface. When U⁵⁰(θ) > 1, the interaction is synergistic. When U⁵⁰(θ) < 1, the interaction is subadditive or antagonistic. When U⁵⁰(θ) = 1, the interaction is additive.

It has been demonstrated24that most isoboles have a simple inward or outward curvature and that they can be well approximated by a simple second order polynomial of the form

and

These functions were fitted to the concentration–effect data using the Gauss–Newton nonlinear least-squares method (Statistics Toolbox function “nlinfit”; MATLAB, Natick, MA). The results were verified using the NONMEM program (Globomax, Hanover, MD).

After transfection with adeno-associated virus–green fluorescent protein and GABA^Areceptor α¹, β², and γ^2sand cDNAs, HEK-293, more than 90% of the cells were found to fluoresce, indicating that successful transfection conditions had occurred. Fluorescing cells were whole cell voltage clamped at −60 mV and superfused with extracellular saline. Application of GABA at eight different concentrations (0.3–1,000 μm) to the cells under these conditions elicited inward chloride currents in a concentration-dependent manner. Addition of 2 μm propofol to the extracellular medium resulted in an increase in the amplitudes of currents activated by 1–30 μm GABA, a small decrease in peak currents activated by 100–1,000 μm GABA, and an increase in the baseline noise of the recording. A representative whole cell voltage clamp recording showing the effect of 2 μm propofol on GABA^Areceptor function is shown in figure 1. These effects were fully reversible. Propofol, 2 μm, reduced the GABA EC⁵⁰by a factor of 0.59 ± 0.05, a fractional effect of 0.41 ± 0.05 (table 1). These experiments were repeated using 0.2–10 μm propofol. Propofol was found to reduce the GABA EC⁵⁰in a concentration-dependent manner. The C⁵⁰for the effect of propofol on the GABA EC⁵⁰was C^50,P= 2.1 ± 0.1 μm, and the slope was 1.02 ± 0.05.

Next, we repeated these experiments using sevoflurane. Addition of 330 μm sevoflurane to the extracellular saline also resulted in an increase in the amplitudes of currents activated by 1–30 μm GABA and an increase in the baseline noise of the recording (see representative trace in fig. 2). However, in contrast to the experiments with propofol, no reduction in the amplitudes of currents activated by 100–1,000 μm GABA was observed. These effects were also fully reversible. Sevoflurane, 330 μm, reduced the GABA EC⁵⁰by a factor of 0.46 ± 0.06, a fractional effect of 0.54 ± 0.06 (table 2). These experiments were repeated using 33–1,650 μm sevoflurane. Sevoflurane also reduced the GABA EC⁵⁰in a concentration-dependent manner. The C⁵⁰for the effect of sevoflurane on the GABA EC⁵⁰was C^50,S= 306 ± 15 μm, and the slope was 1.00 ± 0.04.

Finally, we measured the effect of combinations of propofol and sevoflurane on GABA receptor activation. The fractional effects of combinations of propofol and sevoflurane on GABA EC⁵⁰are shown in table 3. Figure 3shows a representative recording of the combined effect of 1 μm propofol and 165 μm sevoflurane on currents activated by 0.3–1,000 μm GABA. As was observed with sevoflurane and propofol applied alone, the drug combination resulted in an increase in the amplitudes of currents activated by 1–30 μm GABA, a small decrease in peak currents activated by 100–1,000 μm GABA, and an increase in the baseline noise of the recording. These effects were fully reversible. A combination of 1 μm propofol with 165 μm sevoflurane resulted in a fractional effect of 0.52 ± 0.05 (table 3). The data shown in table 3was used to construct a response surface model for the modulation of GABA^Areceptor function by propofol and sevoflurane (fig. 4).

We found that when propofol and sevoflurane were applied alone and together at sufficient concentrations, they increased the amplitudes of submaximal GABA responses. Furthermore, they significantly increased the apparent affinity of GABA for human α¹β²γ^2sGABA^Areceptors heterologously expressed in HEK-293 cells, in 29 of the 33 combinations of anesthetics tested (P < 0.05).

Using the response surface modeling method,24we used these results to define the response surface for the actions of propofol and sevoflurane on the fractional effect on GABA EC⁵⁰. Using the methods described, we fitted the functions U⁵⁰(θ) and γ(θ) for 0 < θ < 1 (fig. 5). First, we noticed that the interaction surface for drug combinations retained the sigmoidicity of the two anesthetics applied alone. Second, we noticed that both β^2Uand β^2γwere both close to zero. A NONMEM verification revealed that β^2U= 0.25 ± 0.49 and b^2g= 0.16 ± 0.47 and were not significantly different from zero. Because neither of the fitted functions significantly deviated from unity, we concluded that neither synergism nor antagonism was occurring and that combinations of propofol and sevoflurane resulted in a purely additive effect on receptor function.

The primary aim of this study and its companion article25was to determine whether propofol and sevoflurane had additive or synergistic effects in humans and on human receptors. In humans, synergy with anesthetic drugs is common. Opioids have been shown to reduce general anesthetic requirement (MAC) but are without any intrinsic anesthetic potency when administered alone. Therefore, the interactions between opioids and general anesthetics can be considered to be highly synergistic. A similar argument can be made for the synergistic actions of midazolam and thiopental.1Propofol and sevoflurane, however, are both general anesthetics in their own right, and we might therefore expect there to be less synergism between them. In the companion study to this report, Harris et al.  25have shown that in humans, there is indeed no synergism between propofol and sevoflurane for immobility and loss of consciousness. Instead, the two anesthetics were shown to be additive. In this study, we have also shown that these two general anesthetics are additive in their actions on GABA^Areceptors, the most common fast inhibitory neurotransmitter receptor in both the brain and the spinal cord, sites thought to be critical for loss of consciousness and immobility, respectively.

The two drugs used in this study were selected for three important reasons. First, although both sevoflurane and propofol can be used alone for both induction and maintenance of anesthesia, these two anesthetics are commonly used together in the clinical setting. Second, these two anesthetics are chemically very different; sevoflurane is a small fluorinated ether, whereas propofol is a large phenol, and it is therefore unlikely that both compounds are capable of making the same array of bonds within a common anesthetic binding site. Finally, both propofol and sevoflurane exist as a single optical isomer. Conducting an additivity study with a pair of anesthetics, one of which exists as a racemic mixture of two or more stereoisomers, would have required an initial investigation into the synergistic, additive, or antagonistic activity of each of the isomers with respect to the other, before any consideration of the second drug could be undertaken. By using propofol and sevoflurane, we removed this requirement from our study.

We hypothesized that it would be unlikely for propofol and sevoflurane to compete for the same binding within the GABA^Areceptor because of the large differences in their molecular structure. This hypothesis is strongly supported by the results of site directed mutagenesis experiments that are consistent with the hypothesis that inhaled general anesthetics mediate their effects within an anesthetic binding cavity located with in the GABA^Aα subunit,11–13,26whereas propofol interacts directly with a genetically related cavity located within the β subunit.15,27,28 

Although there are many data for the separate site hypothesis for propofol and sevoflurane, the same cannot be said of volatile anesthetics that have been shown to interact with the same cavity in GABA^Areceptors.12However, it is unlikely that different anesthetics interact with this cavity in an identical manner. Rather than having simply antagonistic or additive actions, it is tempting to speculate whether a combination of chloroform and halothane, for example, could produce a better anesthetic combination for modulating GABA^Areceptor function than an ether alone.

In addition to having different binding sites, these two drugs have been reported to modulate receptor function via  overlapping but distinctly different mechanisms. Bai et al.  3concluded that propofol slowed the desensitization and deactivation of GABA^Areceptors. This finding was supported by O’Shea et al. ,29who proposed that propofol acted by modulating receptor gating, not ligand binding. Sevoflurane, however, has been shown to increase the apparent affinity of GABA for its receptor30and to block receptor function at high anesthetic concentrations, possibly via  the same mechanism as observed with isoflurane.31 

A recent study32compared the effects propofol and another halogenated inhaled anesthetic, halothane, on GABA receptor single-channel currents and concluded that propofol acted by shortening slow closed times, whereas halothane prolonged the slow open times. It is important to note, however, that the effect of both drugs was to ultimately increase the open probability of the receptor’s ion channel.

Under equilibrium conditions, such as those presented here, the anesthetic-mediated increase in amplitudes of currents activated by submaximal GABA concentrations is thought to be underpinned by the stabilization of the open state of the receptor.28In synapses, where nonequilibrium conditions exist and the shape of the postsynaptic current is dictated by the intrinsic stochastic properties of the synaptic receptors, general anesthetics have been shown to prolong the decay of the postsynaptic current also by stabilizing the open state of the receptor.3 

Currently, the overwhelming weight of experimental evidence suggests that propofol27,28and sevoflurane5,6,14alter the function of the nervous system by enhancing GABA^Areceptor function. Notwithstanding this evidence, there are many novel proteins currently under investigation that may also play an important role in generating, or contributing to, the anesthetized state.33–38In a recent study,39propofol and sevoflurane were both shown to depress the activity of ventral horn interneurons in cultured spinal cord slices. These two anesthetics yielded different effects on the patterns of action potential firing, which led the authors to suggest that in the spinal cord, propofol only acted via  modulation of GABA^Areceptors. The actions of sevoflurane were more complex and most likely result from interactions with glycine receptors, GABA^Areceptors, and a third, unspecified component.

In this study, we examined the combined effect of clinically relevant concentrations of sevoflurane and propofol on the function of GABA^Areceptors containing only the α¹, β², and γ^2sreceptors. Although this receptor combination accounts for approximately 40% of the GABA^Areceptors in the central nervous system, there are several other important tissue-specific combinations that have different kinetics and different general anesthetic sensitivities. It would be interesting if future studies investigated the effects of these two drugs on other subunit combinations (e.g. , α², β³, and γ²) using the techniques described here and also electrophysiologic methods with higher temporal resolutions, e.g. , single channel recording.

In conclusion, the data presented in this study and the results in the companion article25show that propofol and sevoflurane modulate GABA^Areceptor function and generate the anesthetized state (immobility and loss of consciousness) in an additive manner. Neither of the two studies detected a significant degree of synergism or antagonism between propofol or sevoflurane in the three assays. Therefore, it seems that the most likely explanation for the results described here is that propofol and sevoflurane enhance GABA^Areceptor function, in both the brain and the spinal cord. The two drugs do this by interacting at distinctly different binding sites within the same protein. However, the binding of anesthetic to either of the modulatory sites converges on a single effect: the enhancement of GABA^Areceptor gating, resulting in an increase in the open probability of the integral ion channel. The resulting alteration in coincidence detection and synchrony in neuronal networks containing these receptors40,41is likely to be fundamental in the generation of the anesthetized state.

The authors thank Steven Shafer, M.D. (Professor of Anesthesia, Stanford University, Stanford, California), for the NONMEM verification; Peter Sebel, M.B., B.S., Ph.D., M.B.A. (Professor of Anesthesiology, Emory University School of Medicine and Adjunct Professor of Psychology, Emory College of Emory University, Atlanta, Georgia), Jay Johansen, M.D., Ph.D. (Associate Professor, Department of Anesthesiology, Emory University School of Medicine, Grady Memorial Hospital, Atlanta, Georgia), Meagan Ward, A.B. (Neuroscience Graduate Student, Department of Physiology and Graduate School of Arts and Sciences, Emory University, Atlanta, Georgia), Adam Hall, Ph.D. (Assistant Professor, Department of Biologic Sciences, Smith College, Northampton, Massachusetts), and Nick Franks, Ph.D. (Professor of Biophysics and Anesthetics, Biophysics Section, The Blackett Laboratory, Imperial College London, United Kingdom), for helpful discussions.