Isoflurane Modulation of Neuronal Nicotinic Acetylcholine Receptors Expressed in Human Embryonic Kidney Cells

HEK cells stably expressing the human α4β2 subunit combination were obtained from SIBIA Neuroscience, Inc. (La Jolla, CA) (now Merck Research Laboratories, San Diego). Cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 2 mm L-glutamine, 100 U penicillin, 100 mg streptomycin (Invitrogen, Carlsbad, CA), 6% iron-supplemented calf serum (Sigma- Aldrich, St. Louis, MO), and 100 mg/ml G418 (Mediatech, Herndon, VA) and were kept at 34.7°C in air + carbon dioxide (93% + 7% by volume). For patch clamp experiments, cells were plated on glass coverslips coated with poly-L-lysine and cultured for 1 to 3 days.

The whole-cell currents were recorded using the whole-cell patch clamp technique at room temperature (22°C). Pipette electrodes were made from 0.8 mm (inner diameter) borosilicate glass capillary tubes and fire-polished. The electrodes had a resistance of 2–3 MΩ when filled with standard pipette solution. The membrane potential was clamped at −50 mV and a 5–10-min period were allowed after rupture of the membrane to equilibrate the cell interior with pipette solution. Currents through the electrode were recorded by an Axopatch 200 amplifier (Axon Instruments, Union City, CA) after filtering at 2 kHz.

Single-channel currents were recorded using the outside-out patch clamp technique at room temperature (20°–25°C). Pipette electrodes were coated with SigmaCote (Sigma-Aldrich) to minimize background noise. The electrodes had a resistance of 10 to 15 MΩ when filled with pipette solution. The membrane was clamped at −70 mV.

The external solution for both whole-cell and outside-out single-channel patch clamp experiments contained (in mM): NaCl 150, KCl 5, CaCl²2.5, MgCl²1, glucose 10, HEPES acid 5.5, and Na⁺HEPES 4.5. The pH was adjusted to 7.3 with HCl, and osmolarity was adjusted to 320 mOsm by D-glucose. The internal solution contained (in mM): K-gluconate 140, MgCl²2, CaCl²1, EGTA 11, HEPES acid 10, Mg²⁺adenosine triphosphate 2, and Na⁺GTP 0.2. pH was adjusted to 7.3 with KOH, and osmolarity was adjusted to 300 mOsm by D-glucose.

Acetylcholine (Sigma-Aldrich) was first dissolved in distilled water to make stock solutions. Isoflurane, sevoflurane, and halothane were obtained from Anaquest (Madison, WI) (a division of BOC Inc), Abbott Laboratories (North Chicago, IL), and Ayerst Laboratories (New York, NY), respectively. Saturated isoflurane, sevoflurane, and halothane solutions were made by stirring the respective anesthetic in the external solution over 8 h in a sealed glass container with little air space and their concentrations had been estimated to be 15.3, 11.8,6and 18 mm,20respectively. Anesthetic test solutions were prepared immediately before experiments by diluting the saturated solution and were kept in air-free, closed glass bottles to prevent evaporation of the anesthetics.

The rapid-exchange application system was used for applications of test solutions. With this system solution was exchanged with a rise time of 20 ms as measured by a change in junction potential using the patch electrode. In a few experiments, a faster solution exchange using a piezo-driven switch was used, which provided solution exchange within 2 ms. In the current study, the term “coapplication” is referred to as the simultaneous application of anesthetic and acetylcholine through a three-barrel square glass tube, whereas the term “preperfusion” is referred to as the application of anesthetic before coapplication of anesthetic and acetylcholine.

EC⁵⁰values and their slope factors (Hill coefficients) were calculated from the equation:

where I is the amplitude of acetylcholine-induced current, I^maxthe maximum current, C the drug concentration, and n the Hill coefficient. The nonlinear regression analysis was carried out using the least squares fitting method (Sigmaplot, Version 8.0, SPSS Inc., Chicago, IL) by a microcomputer.

Two conductance levels were observed, and the largest conductance level was manually chosen for the analysis of amplitude because it represented the major single-channel opening events. Opening and closing of the channels were detected using the 50% threshold criterion.21Amplitude histograms were fitted by a sum of Gaussian functions using the least-squares method. The interburst intervals were determined by the method of Colquhoun and Sakmann.22 

Whole-cell data are expressed as the mean ± SD and n represents the number of experiments. For the single-channel data, n represents the number of events that were used for estimating the values including the means of amplitude, open time, closed time, and burst duration.

At a holding potential of −50 mV, inward currents generated by 3 s applications of 30 μm acetylcholine showed some decay, indicating desensitization of the receptor. When acetylcholine was applied at 2-min intervals, the peak inward currents were maintained at a stable level.

First, to compare the potency of three inhalational anesthetics on the α4β2 nAChRs expressed in HEK cells, concentration-response relationships for suppression of acetylcholine-induced currents by coapplication with the anesthetics were examined. Isoflurane, sevoflurane, and halothane inhibited acetylcholine-induced currents in a concentration-dependent manner (fig. 1). Hump tail currents could be seen upon termination of acetylcholine pulse at high concentrations of isoflurane (300 μm, 1 mm) and to a lesser extent with sevoflurane (fig. 1A). The 50% inhibitory concentrations (IC⁵⁰s) for halothane, isoflurane, and sevoflurane were 39.8 μm (0.21 minimum alveolar concentration [MAC]), 67.1 μm (0.24 MAC), and 183.3 μm (0.61 MAC), respectively. Thus, the IC⁵⁰values were less than 1 MAC for each of the anesthetics. Hill coefficients were estimated to be 0.8, 0.7, and 0.8 for halothane, isoflurane, and sevoflurane, respectively. In terms of IC⁵⁰MAC values, halothane and isoflurane were almost equipotent and sevoflurane was less potent.

We have previously reported that halothane block of α4β2-type nAChRs increases with increasing acetylcholine concentration in the cortical neurons.8To compare halothane with two other anesthetics, acetylcholine concentration-response relationships in the presence of anesthetics were examined. Figure 2shows the effects of three anesthetics on the acetylcholine concentration-response relationships. Isoflurane at 100 μm inhibited the peak current at acetylcholine concentrations ranging from 10 μm to 1 mm. The peak current was suppressed to 33% of the control with coapplications of 100 μm isoflurane and 10 μm acetylcholine (fig. 2A). With increasing concentration of acetylcholine from 10 μm to 1 mm, the isoflurane inhibition became less and less (fig. 2, A and C). Figure 2Bshows that the apparent affinity of the receptor for acetylcholine was also decreased by all anesthetics: halothane at 92 μm (0.4 MAC), sevoflurane at 132 μm (0.4 MAC), and isoflurane at 100 μm (0.36 MAC) increased the acetylcholine EC⁵⁰from the control of 11.8 μm to 53.3 μm, 26.0 μm, and 55.9 μm, respectively. This competitive nature of isoflurane and sevoflurane inhibition was seen only with the coapplication of anesthetic but not with the preapplication and coapplication of anesthetic. Halothane, however, behaved differently from isoflurane and sevoflurane in that it greatly decreased the maximum acetylcholine current (fig. 2B).

We have previously demonstrated that halothane significantly slows the decay phase of acetylcholine current of the α4β2-type receptor in cortical neurons.8To see whether anesthetics behave similarly, their effects on the decay phase of acetylcholine-activated currents were examined. When isoflurane was coapplied with ACh, the peak amplitude of current was reduced and the decay phase was altered (fig. 3A). The decay phase of the control current could be well fitted by a single exponential function, whereas the decay phase in the presence of anesthetics was fitted by the sum of two exponential functions (fig. 3B). The fast time constant was on the order of 300 ms and the slow time constants were around 1.2 s, which was not significantly different from the control value of 1.4 ± 0.29 s (n = 10) (fig. 3B). The fast component probably represented the time course of washing-in of anesthetics.

We have previously shown that halothane blocks the α4β2-type nACh receptors of cortical neurons in both resting and activated states.8To see whether isoflurane blocks the α4β2 receptor in both states, experiments were performed by comparing the effects of coapplication of isoflurane and acetylcholine to those with preperfusion of isoflurane.

Records in figure 4Ashow that 3-s coapplication of 30 μm acetylcholine and 100 μm isoflurane suppressed the peak amplitude of currents and accelerated the decay phase (record b). A second coapplication 2 min later produced a current (record c) nearly identical to that of the first coapplication. This indicated that isoflurane blocked the open channel quickly to a steady state. The current recovered completely within 2 min, as indicated by the washout experiment. Washing the cell for 2 min with isoflurane-free solutions restored the acetylcholine current (record d). The time courses of the block and unblock of the peak current are illustrated by white circles in figure 4B.

To evaluate the isoflurane block of the resting receptor, preapplication experiments were performed as shown in figure 4C. Bath perfusion of 100 μm isoflurane commencing 2 min before coapplication of isoflurane and acetylcholine reduced the current without showing a prominent peak (record b). Continuous bath perfusion of isoflurane for 2 min longer resulted in slightly more suppression of current (record c), suggesting that the resting block was essentially established within the first 2 min. Washing with isoflurane-free solution for 2 min restored the acetylcholine current completely (record d). The time course of suppression of the resting receptor during bath perfusion of isoflurane is shown by black circles in figure 4B. Thus, isoflurane blocked both the activated receptor and the resting receptor. Similar results were obtained in four other cells.

Current records of figure 5illustrate the comparison between the resting and activated receptor block. Record B shows the activated receptor block by coapplication of 100 μm isoflurane and 30 μm acetylcholine. Record C represents the resting receptor block produced by preapplication of isoflurane. Record D shows current recovery by washing with normal solution after preapplication of isoflurane. Recovery from block on termination of isoflurane application occurred slowly when the receptor was activated by acetylcholine.

Records A, B, and C of figure 5are superimposed in E. When the block was measured as a reduction of the peak current, isoflurane blocked the activated receptor to 48% ± 7.0% of the control and the resting receptor to 20% ± 4.5% of the control (n = 5). However, when the block was measured at the end of pulse, currents following the activated receptor block and the resting receptor block attained the same steady-state level. These results suggested that isoflurane had the same affinity for the resting receptor and for the activated receptor. Figure 5Fshows the activated receptor recovery by superimposing records A and D. Current recovered slowly with a time constant of 161 ± 16.5 ms (n = 5) when the receptor was activated by acetylcholine after preapplication of isoflurane and attained the same steady-state level as the control current generated by acetylcholine alone. The slow rise of the current suggested that the escape of isoflurane molecule from the activated receptor was slow. Thus, it was conclude that isoflurane blocked the α4β2 nACh receptor in both resting and activated states.

When anesthetics were coapplied with acetylcholine using the perfusion method having a solution exchange time on the order of hundreds of milliseconds, the effects of anesthetics on the peak current and on its kinetics would be distorted. To achieve the steady-state anesthetic effect, the cell was first exposed to the anesthetic for 2 min and then activated by coapplication of acetylcholine and anesthetic. Figure 6shows the effects of 100 μm isoflurane on the peak current and its decay kinetics as activated by 3, 30, and 300 μm acetylcholine. Isoflurane reduced the currents activated by 3, 30, and 300 μm acetylcholine to 26.3% ± 9.0%, 27.4% ± 5.0%, and 27.5% ± 11.0% of the control currents, respectively (n = 4). Thus, isoflurane inhibition of the human α4β2 receptors was independent of acetylcholine concentration.

In addition, the effect of the preapplication of isoflurane on the decay kinetics of acetylcholine-activated current was found to be different from that obtained with the coapplication only. The control currents decayed single exponentially, becoming faster with increasing acetylcholine concentration and the time constants were 1.93 ± 0.23 s, 1.43 ± 0.26 s, and 1.10 ± 0.12 s (n = 4), respectively, for the currents activated by 3, 30, and 300 μm acetylcholine. In the presence of 100 μm isoflurane, the decay of acetylcholine current remained single exponential with a time constant of 2.5 s ± 0.30 s (n = 12) irrespective of acetylcholine concentration. These time constants were significantly larger than the control values (P < 0.05).

Coapplication of acetylcholine and isoflurane could not faithfully reveal the onset and offset kinetics of blocking action because the time courses of activation and deactivation of the receptor might overlap those of the blocking and unblocking action of isoflurane. To overcome this difficulty, a fast solution exchange using a piezo-driven switch was utilitized. This provided a complete solution exchange within 2 ms. In this protocol, the receptor was first activated by acetylcholine and the acetylcholine-containing solution could be swiftly switched to a solution containing acetylcholine and various concentrations of isoflurane. Figure 7Adepicts the concentration-dependent isoflurane block of the current activated by 100 μm acetylcholine. Figure 7Bshows the dose-response relationship for isoflurane to inhibit the activated receptors. The fit to the data gave an IC⁵⁰of 62.3 μm and a Hill coefficient of 1.0. The onset of block of the activated receptor was very rapid with a time constant of 2 ms for 84 μm isoflurane and 1 ms for higher concentrations. The weak concentration dependence of the onset of block is most likely attributable to the limitation of solution exchange. The recovery time constant was on the order of 20 ms regardless of isoflurane concentrations. Because this time constant is much larger than the solution exchange time (∼2 ms) and because the recovery time constant is concentration independent, this constant is interpreted as representing the unblocking time constant.

Acetylcholine-induced whole-cell currents in α4β2 AChRs are inwardly rectified.23–25Isoflurane did not change the control current-voltage relationship. Little or no currents were induced at positive potentials, and the linear relationship of the inward current was not altered by isoflurane, suggesting that the suppression by isoflurane as seen in negative membrane potentials was independent of membrane potential (data not shown).

To further elucidate the mechanism of action of isoflurane, acetylcholine-induced singlechannel currents were studied with or without isoflurane. When 30 nm acetylcholine was applied to the outside-out membrane patch clamped at −70 mV, single-channel currents occurred as either brief isolated openings or longer openings interrupted by short closures or gaps. Sample records of currents without and with 84 μm isoflurane are shown in figure 8, A and B, respectively. Two conductance state currents, −1.8 and −2.9 pA, were observed. Coapplication of 100 nm dihydro-β-erythroidine completely blocked the currents, indicating that the currents were generated by the activation of α4β2 nAChRs (data not shown). Only the main large conductance currents were analyzed in the current study as the small conductance currents were observed much less frequently (<10% of the total events). Figure 8, C and Dshow current amplitude histograms. Amplitudes of single-channel currents induced by 30 nm acetylcholine were −2.93 ± 0.46 pA (n = 5) and −2.88 ± 0.31 pA (n = 5) in the absence and presence of 84 μm isoflurane, respectively. Thus, the amplitude of main conductance state current was not altered by 84 μm isoflurane (fig. 8, C and D) or by 280 μm isoflurane (data not shown).

Figure 9shows the open time distribution for currents induced by 30 nm acetylcholine in the absence and presence of 84 μm isoflurane. The time axis is drawn on a logarithmic scale so that the effective bin width increases exponentially from left to right. This displays a multi-exponential distribution as a series of skewed bell-shaped curves whose peaks overlie the time constants of several exponential components. The open time distributions of main conductance currents in the presence of acetylcholine and acetylcholine plus isoflurane indicate multi-exponential components. The open time distribution in 30 nm acetylcholine had time constants of 7.4 ms (68.6% of the total observations) and 1.6 ms (31.4%) (fig. 9A). The mean open time was estimated to be 6.72 ± 3.99 ms (n = 5). When 84 μm isoflurane was coapplied with 30 nm acetylcholine, the time constants were 7.8 ms (48.5%) and 1.3 ms (51.5%) and the mean open time was 4.54 ± 2.23 ms (n = 5) (fig. 9B). Thus, isoflurane shortened the mean open time (paired Student t  test, P < 0.01).

Distributions of the closed time of the main conductance state currents in the presence of 30 nm acetylcholine and of 30 nm acetylcholine plus 84 μm isoflurane also show multi-exponential components (fig. 10). There were at least three components, suggesting that the channel could be in at least three closed states. The time constants in 30 nm acetylcholine were 63.1 ms (51.1% of the total observations), 5.3 ms (33.5%), and 0.65 ms (15.4%) (fig. 10A). The mean closed time was estimated to be 42.2 ms (n = 5). The time constants of the closed time in the presence of 30 nm acetylcholine and 84 μm isoflurane were 231 ms (49.4%), 8.45 ms (37.1%), and 0.52 ms (13.5%) (fig. 10B); the mean closed time was 113.6 ms (n = 5). Thus, isoflurane increased the time constant associated with the slow component of the closed time distributions, resulting in a significant increase in the mean closed time to 113.6 ± 26.5 ms from the control of 42.2 ± 8.20 ms (n = 5) (paired Student t  test, P < 0.01).

Figure 11shows the distributions of burst durations of acetylcholine-induced currents without and with isoflurane. The distributions were fitted by three exponential functions. In 30 nm acetylcholine, the time constants of burst durations were estimated to be 13.7 ms (74.5%), 1.7 ms (10.7%), and 0.7 ms (14.7%) (fig. 11A). The mean burst duration was estimated to be 13.5 ± 4.03 ms (n = 5). When 84 μm isoflurane and 30 nm acetylcholine were coapplied, the time constants were 9.0 ms (68.9%), 1.1 ms (20.3%), and 0.8 ms (10.8%) (fig. 11B). Thus, two of the three time constants were slightly decreased by isoflurane. The mean burst duration was estimated to be 8.4 ± 1.8 ms (n = 5), which was significantly less than the control (paired Student t  test, P < 0.01).

Figure 12shows the time courses of changes in the probability of openings in the presence of 30 nm acetylcholine and in 30 nm acetylcholine plus 84 μm isoflurane. The mean open probability was decreased from 0.078 ± 0.027 (n = 5) in the control to 0.022 ± 0.007 (n = 5) in the presence of isoflurane, representing a 72.0 ± 4.90% decrease. Thus, it was concluded that the open probability was decreased by isoflurane.

The current study demonstrated that isoflurane, sevoflurane, and halothane at clinical relevant concentrations reversibly blocked the whole-cell currents of the human α4β2 nAChRs expressed in HEK cells. The IC⁵⁰values were less than 1 minimum alveolar concentration, and the potencies were halothane > isoflurane > sevoflurane. At the single-channel level, isoflurane shortened the open time without causing flickerings or reducing the single-channel current amplitude. It also prolonged the mean closed time. As a result of these two effects, the single-channel opening probability was greatly reduced by isoflurane.

The current study has confirmed the results of previous studies that neuronal nAChRs are more sensitive than muscle-type nAChRs to anesthetic block.5,6In addition, responses to isoflurane are different between muscle and neuronal nAChRs. In muscle-type nAChRs, isoflurane inhibition was acetylcholine concentration dependent, being larger at high concentrations of acetylcholine.26In the current study using the human α4β2 nAChR expressed in HEK cells or in the previous study using rat α4β2 nAChRs expressed in Xenopus  oocytes,6isoflurane block at equilibrium was independent of acetylcholine concentration. This difference is also corroborated with the results of single-channel studies.

The current study also demonstrated that the method of application affected the blocking potency of an anesthetic and its nature of interaction with acetylcholine at the receptor. When isoflurane was preapplied for 2 min and then coapplied with acetylcholine, the blocking potency of isoflurane increased, as reflected in a decrease in IC⁵⁰to 36 μm from 68 μm as obtained from the coapplication experiment. The IC⁵⁰of 36 μm is almost identical to the 34 μm obtained by Violet et al.  6using the preapplication and coapplication method in rat α4β2 receptors. The IC⁵⁰of 67 μm obtained with the coapplication only is similar to 85 μm reported by Flood et al. ,5who used the coapplication of isoflurane in chicken α4β2 receptors. Thus, our two methods of application of isoflurane in human α4β2 receptors yielded results almost identical to those obtained with either rat or chicken α4β2 receptors using the corresponding methods of anesthetic application. Volatile anesthetic action on the α4β2 nAChR appears to be species independent.

The isoflurane block was more potent when the preapplications and coapplications were used than when only the coapplication method was used. This is as expected because the former method provided the steady-state concentration of isoflurane whereas the latter provided a decreased concentration at the time when the peak current was measured. A similar effect was observed with thiopental inhibition of neuronal nicotinic receptors.27 

In addition to difference in potencies, these two methods of drug application yielded different results with respect to the acetylcholine concentration dependence of anesthetic block. Using the coapplication of anesthetic with acetylcholine, anesthetic block was dependent on acetylcholine concentration, being lesser with increasing acetylcholine concentration, as seen with isoflurane by Flood et al.  5Using the preapplication and coapplication method, we found that isoflurane block was acetylcholine concentration independent, as observed by Violet et al.  6for halothane.

Anesthetics also affected the kinetics of the acetylcholine-activated current. When an anesthetic was preapplied, the decay of the acetylcholine-activated current became slower irrespective of acetylcholine concentrations (fig. 6). Using the coapplication protocol, anesthetics caused biphasic decay in the acetylcholine-activated currents in contrast to a single exponential decay in the control. The fast time constants were on the order of 200–300 ms and the time constants for the slow component were approximately 1.5 s, which is not significantly different from the control value of figure 3B. The fast time constant probably reflects the solution exchange time when the coapplication method is used.

The coapplication method affected isoflurane block of nAChRs as well. The onset of isoflurane block of the resting and activated receptors and the recovery from anesthetic block were on the order of 150∼200 ms, which is rate-limited by an unstirred layer in the vicinity of the cell membrane, as seen previously.8However, when isoflurane was preperfused using a rapid application system with the solution exchange time less than 100 μs, the time constants of isoflurane binding to and unbinding from the muscle type nAChRs were estimated to be around 500 μs.26Our conclusion, based on the slow coapplication, that isoflurane blocks the resting receptors may actually reflect an ultra-fast block of the activated receptors.

Several investigators have reported that inhalational anesthetics alter the kinetic properties of muscle nAChRs.26,28–31Anesthetics were reported to cause receptor channels to flicker rapidly between open and closed states and to decrease channel open time with little or no effect on the current amplitude.30In the current study, however, isoflurane did not induce flickerings during channel openings in the human α4β2 nAChR channels. Isoflurane exerted a greater effect in decreasing the frequency of channel openings rather than reducing the open time with flickerings, as observed in muscle nAChRs.26Once the isoflurane molecule binds to the α4β2 nAChR, it may require time to dissociate from the binding site. This interpretation is supported by the experiment with a fast solution exchange showing the unblocking rate constant of 50 s⁻¹. The blocking rate constant estimated from a K^dof 40 μm and an unblocking rate constant of 50 s⁻¹was on the order of 1∼5 × 10⁶M⁻¹·s⁻¹. Thus, the blocking rate constant of isoflurane in the α4β2 nAChRs is on the same order of magnitude as that in the muscle nAChRs. The unblocking rate constant for the α4β2 nAChRs is at least 20 times smaller that that for the muscle nAChRs. These observations would account for the lack of flickering in the single-channel activity and for an enhanced blocking potency. The difference in the single-channel blocking action might explain the difference in the whole-cell results between the muscle nAChR and the neuronal nAChR. The site-directed mutagenesis study in combination with the single-channel analysis of muscle-type nAChR supports the notion that the M²domain that forms the pore could modulate isoflurane blocking action.31In addition, inhalational anesthetics could bind to an extracellular, water-accessible pocket formed by amino acids from δM¹-δM³α helices.32Anesthetic binding to this site may contribute to its inhibitory action on the resting closed receptors.

Anesthetic modulation of neuronal nAChRs is deemed to have great impact on the function of the brain. Acetylcholine is one of the neurotransmitters released from the brain stem, hypothalamus, basal forebrain, and cerebral cortex that participate in cognition, memory, alertness, learning, and antinociception. It has been reported that halogenated inhalational anesthetics have hyperalgesic effects at low clinical concentrations.33,34In addition, inhalational anesthetics may trigger substantial changes in brain function via  a cascade of multisynaptic events by inhibiting nAChRs.12–14Therefore, the high sensitivity of neuronal nAChRs to inhalational anesthetics could contribute to various stages of anesthesia such as amnesia,35inattentiveness, and delirium, although nAChRs are not directly involved in the hypnotic component of anesthesia.

We thank Julia Irizarry, B.A., Program Assistant II (Department of Molecular Pharmacology and Biological Chemistry, Northwestern University Medical School, Chicago, Illinois) for secretarial assistance, and Barbara Kowalska, M.D., Research Technologist II (Department of Molecular Pharmacology and Biological Chemistry, Northwestern University Medical School, Chicago, Illinois) for technical assistance.