Effects of the Anesthetic Gases Xenon, Halothane, and Isoflurane on Calcium and Potassium Currents in Human Atrial Cardiomyocytes

Right atrial appendages were obtained as surgical specimens from patients (age range, 48–79 yr; mean, 61.2 yr; n = 20) undergoing heart surgery because of coronary artery insufficiency (n = 19) or aortic valve disease (n = 1). All patients were in sinus rhythm and had no evidence of right atrial dysfunction. The cardiovascular medication that most patients received (β-adrenergic antagonists, diuretics, angiotensin-converting enzyme inhibitors, or nitric oxide donors) was stopped at least 12 h before surgery. The investigations were performed in accordance with the principles outlined in the Declaration of Helsinki and approved by the local ethical board. All patients gave informed consent prior to surgery.

Atrial myocytes were prepared, as described previously in detail, 21by digestion with collagenase (type V; Sigma, St. Louis, MO), followed by gradual recalcification. Only well-striated, bleb-free, rod-shaped myocytes were used for the electrophysiologic studies, done within 4 h after the isolation.

Membrane currents were recorded by means of the whole cell patch clamp technique 22with use of an amplifier (EPC-9; HEKA, Lambrecht, Germany) and a personal computer equipped with Pulse 8.0 software (HEKA) for data acquisition and analysis.

The patch pipettes were pulled from borosilicate glass (Hilgenberg, Malsfeld, Germany) with a vertical puller and heat-polished with a microforge. Pipette resistances ranged from 3.0 to 7.0 MΩ when they were filled with pipette solution. The EPC-9 provides automatic procedures for both fast and slow capacitance subtraction. Cell capacitance was determined before any series of depolarizing or hyperpolarizing pulses. The capacitance did not noticeably change during any experiment. Series resistance compensation with supercharging was not performed.

Cardiac myocytes were allowed to adhere to a glass coverslip, which was transferred into a bath chamber. The bath chamber was filled with bath solution to a height of approximately 2 mm (total volume, approximately 400 μl; surface area, approximately 200 mm²). In some experiments, the bath solution was covered with a thin paraffin oil layer (DAB10; Wasserfuhr, Bonn, Germany) saturated with xenon. The chamber was continuously perfused at a rate of 3 to 4 ml/min. The inflow occurred through high-performance-liquid-chromatography-grade steel capillaries.

After obtaining the whole cell configuration, an equilibrium period of 5 min was allowed, during which current amplitudes and access resistance stabilized. Afterward, rundown of Ca²⁺and K⁺currents was negligible (less than 5%) over the time course of the experiments (less than 15 min). For measurements of L-type Ca²⁺currents (I^Ca,L), a resting membrane potential of −60 mV was chosen to minimize Na⁺currents; K⁺currents were blocked with intracellular Cs⁺. I^Ca,Lwas elicited with depolarizing pulses for 400 ms. Ca²⁺currents are given here as peak current amplitude of the respective depolarizing step. For measurements of K⁺currents, Ca²⁺currents were blocked with 200 μm Cd²⁺and extracellular Na⁺was replaced by N -methyl-d-glucamine. Inward rectifying K⁺currents (I^K,ir) were measured during stepwise hyperpolarizing pulses 50 ms in duration, from a holding potential of −60 mV to test potentials ranging from −80 to −130 mV. Outward K⁺currents were measured during stepwise depolarizing pulses 300 ms in duration, from a holding potential of −90 mV to test potentials ranging from −60 to +60 mV. All experiments were performed at room temperature (21°C).

For measurements of I^Ca,L, the bath solution contained tetraethylammonium chloride, 136 mm; CaCl², 1.8 mm; MgCl², 1.8 mm; glucose, 10 mm; and HEPES, 10 mm; this was pH-adjusted to 7.40 with tetraethylammonium hydroxide. The pipette solution contained CsCl, 140 mm; MgCl², 2 mm; adenosine 5′-triphosphate, 0.3 mm; guanosine 5′-triphosphate, 0.3 mm; ethylene glycol-bis(β-aminoethylether)-N,N,N′,N′-tetraacetic acid, 10 mm; HEPES, 10 mm. This was pH-adjusted to 7.20 with CsOH.

For measurements of K⁺currents, the bath solution contained N -methyl-d-glucamine, 132 mm; CaCl², 1 mm; MgCl², 2 mm; KCl, 5 mm; CdCl², 0.2 mm; glucose, 10 mm; and HEPES, 10 mm. This was pH-adjusted to 7.40 with HCl. The pipette solution contained K-aspartate, 110 mm; KCl, 20 mm; MgCl², 1 mm; Na²-phosphocreatine, 5 mm; adenosine 5′-triphosphate, 5 mm; guanosine 5′-triphosphate, 0.1 mm; ethylene glycol-bis(β-aminoethylether)-N,N,N′,N′-tetraacetic acid, 10 mm; and HEPES, 10 mm. This was pH-adjusted to 7.20 with KOH.

Bath solutions containing xenon or volatile anesthetics were prepared by passing an appropriate gas mixture through the solutions at a flow of 30–200 ml/min in a 50-ml glass flask equipped with a glass frit and a membranous septum. Gas left the flask through a valve. After at least 30 min of gassing, a fraction (approximately 5–8 ml) of the solutions was taken through the septum into gas-tight glass syringes (1010 TLL; Hamilton, Bonaduz, Switzerland) and immediately perfused through the bath chamber. Each solution was presented to the cells for at least 60–120 s until current amplitudes remained stable.

For the preparation of solutions containing halothane or isoflurane, the gas composition used for gassing was set with a calibrated vaporizer (Fluotec3 and Isotec3; Ohmeda, Steeton, UK). The respective gas concentrations were as follows: halothane, 0.75% (vol/vol; corresponding to 1 MAC) or 1.5% (2 MAC); and isoflurane, 1.2% (1 MAC). The O²concentration was 35%; the remainder was N². An anesthetic gas analyzer (AGM II; Heyer, Bad Ems, Germany) was used to continuously monitor the concentrations of halothane (Rüsch, Böblingen, Germany) and of isoflurane (Abbott, Wiesbaden, Germany) in the gas phase.

During gassing with xenon, the gas composition was set with a mass flowmeter–controller (E-7000; Bronkhorst HI-TEC, AK Ruurlo, Netherlands) as a mixture of xenon (80% or 95%) and O². Xenon was provided by Messer-Griesheim (Krefeld, Germany).

Samples of 50 μl were taken from the perfused bath chamber with a gas-tight glass syringe (1710 RN, Hamilton) equipped with an integrated steel needle. Care was taken to obtain the sample from a spot where cells would have been in an experiment. Each sample was injected into gas-tight vials (1 ml; type N 11-1 [Macherey-Nagel, Düren, Germany]). The gas concentrations were determined with head space–gas chromatography/mass spectrometry/selected ion monitoring (GC/MS/SIM; SSQ 7000; Finnegan, Bremen, Germany), as described by Maruyama et al.  23As standards, samples were taken from the glass flask after gassing of the bath solution with various concentrations of halothane, isoflurane, or xenon. The precision of the determination of all concentrations was better than 5% of the mean. Measurements of O²concentrations of solutions were performed with a conventional blood gas analyzer (ABL 505; Radiometer, Copenhagen, Denmark).

Current densities were calculated by dividing current amplitudes by the whole cell capacitance. Values are expressed as mean ± SD. Curve-fitting and statistical comparisons were made with commercially available software (Prism 2, San Diego, CA). Paired Student t  tests (two-tailed) were used to determine the statistical significance of differences between raw data obtained before and after the application of a given anesthetic. Differences with P < 0.05 were considered significant.

When the bath solution was gassed with 80% xenon, the xenon concentration in the bath chamber during perfusion was 60.9 ± 7.8% (n = 6) when the bath solution was not covered with paraffin oil. In the presence of oil, the xenon concentration was 68.6 ± 4.5% (n = 4). Thus, a loss of xenon occurred during perfusion that amounted to approximately 19% in the absence and to 11% in the presence of paraffin oil. Accordingly, the patch clamp experiments were performed either with bath solutions gassed with 80% xenon and covered with oil or with bath solutions gassed with 95% xenon, to make sure that one MAC of xenon was presented to the cells.

In preliminary experiments, we noticed the importance of gas-tight syringes and steel capillaries. With a plastic syringe instead of a gas-tight glass syringe, a reduction in the xenon concentration occurred by 21%, even when steel capillaries and paraffin oil were used. For comparison, we determined the loss of oxygen from the bath solution gassed with 100% O². No loss of O²was detected through plastic syringes during the 5-min period. A loss of O²by approximately 5% occurred through plastic perfusion tubes, which consequently were not used for any experiments. A further loss of 5% through the surface of the bath could be prevented by covering the bath solution with paraffin oil.

The mean concentrations of halothane in the bath solution at room temperature (0.75% and 1.5% in the gas phase) were 0.39 mm and 0.85 mm, respectively, whereas the mean concentration of isoflurane in the solution (1.2% in the gas phase) was 0.51 mm. No measurable loss of halothane and of isoflurane was demonstrated under the standard experimental conditions (gas-tight glass syringes, steel capillaries, no paraffin oil). Changes in the concentration during passage from the syringe to the bath chamber were in the range of precision of the analysis method.

Representative tracings of I^Ca,Lin human atrial myocytes are shown in figures 1A and B. Exposure to xenon (in the range of 1 MAC) left the I^Ca,Lvirtually unchanged with respect to the peak amplitude as well as to the time course of the decay of I^Ca,Lafter the peak (fig. 1A). Current densities of I^Ca,Lat various membrane potentials, before and during exposure to xenon, are shown in figure 2A. The maximal activation was observed between 10 and 20 mV. During depolarizing steps to +20 mV, the peak I^Ca,Lamplitude in the presence of xenon was 97.5 ± 4.8% (range, 92–107%; n = 7; 4 patients), as related to the peak I^Ca,Lamplitude of −2.5 ± 0.7 picoamperes (pA)/picofarad (pF) obtained immediately before xenon exposure. The values after washout of xenon were 99.2 ± 2.3% (n = 7) of those prior to xenon exposure. Even when xenon was applied at a higher concentration (approximately 80%; solution gassed with 95% xenon; bath covered with oil; n = 4), no effect on I^Ca,Lwas noticed (data not shown). For comparison, a solution gassed with 95% N²and 5% O²did not change I^Ca,Lto a noticeable amount either (data not shown).

In contrast to xenon, halothane produced a concentration-dependent depression of I^Ca,L. A representative tracing for the effect of halothane is shown in figure 1B. Halothane (1.5%, corresponding to 2 MAC) depressed peak I^Ca,Lamplitudes at +20 mV to 53.9 ± 8.0% (P < 0.05; n = 10; 4 patients), from −2.3 ± 0.6 pA/pF to −1.2 ± 0.3 pA/pF (fig. 2B). When a halothane concentration of 0.75% (1 MAC) was used, peak I^Ca,Lamplitude was reduced to 68.1 ± 4.1% (P < 0.05; n = 4; 3 patients) of control. After washout, a rapid and complete recovery of peak I^Ca,Lto 95.9 ± 7.0% (n = 14) of control was observed.

Similar to halothane, isoflurane (1.2%, corresponding to 1 MAC) induced a depression of the peak I^Ca,Lamplitude to 78.3 ± 9.2% (P < 0.05; n = 7; 2 patients), from −1.6 ± 0.7 pA/pF to −1.2 ± 0.6 pA/pF. This effect was reversible; washout of isoflurane restored peak I^Ca,Lto 96.8 ± 6.3% (n = 7) of the current values before the application of isoflurane.

Three subtypes of K⁺currents in human atrial cardiomyocytes were discriminated: inward rectifying K⁺currents (I^K,ir), measured as inward currents during hyperpolarization, and the two voltage-gated K⁺outward currents of the transient (I^K,to) and sustained (I^K,sus) type.

A representative tracing of I^K,iris shown in figure 3A. The apparent reversal potential of the currents (approximately −65 mV) is different from the value expected from the NERNST equation, indicating some voltage error, such as a liquid junction potential, not corrected in the current experiments. I^K,irremained constant within 7% over the time of one experiment in all cells studied (n = 19). No significant effect on I^K,irwas demonstrated for any of the three anesthetics tested. In the presence of xenon, I^K,ir(as measured during hyperpolarizing pulses to −130 mV) were reduced to 97.5 ± 2.8% (n = 5; 3 patients), from −4.0 ± 2.2 pA/pF to −3.9 ± 2.2 pA/pF (fig. 3B). The respective values for halothane 0.75% were 102.9 ± 4.4% (n = 4; 2 patients); for halothane 1.5%, they were 98.8 ± 3.8% (n = 4; 4 patients); and for isoflurane 1.2%, they were 96.4 ± 4.3% (n = 6; 3 patients) of control.

Voltage-gated K⁺outward currents exhibited a remarkable variability between different cell populations, with respect to the current amplitude and the relative contribution of I^K,toand I^K,sus. The sustained component was measured at the end of depolarizing pulses 300 ms in duration. It was present in all cells studied at membrane potentials of −10 mV and greater (fig. 4A). The current density at +60 mV (6.1 ± 2.3 pA/pF; n = 28; 8 patients) ranged from 2.8 to 12.0 pA/pF. In any one cell over the time of one experiment, however, I^K,suswere sufficiently stable to test effects of anesthetics. No attempt was made to discriminate further subtypes of the sustained K⁺outward current.

The transient component I^K,tothat produced a peak outward current within the first 10 ms after depolarization (fig. 4A) was quantified as the difference between this peak and the current sustained after 300 ms. The current densities as measured at +60 mV (7.8 ± 4.1 pA/pF; n = 27; 8 patients) ranged from 2.8 to 17.6 pA/pF. Again, I^K,towas stable in individual cells over the duration of one experiment.

Effects of anesthetics on voltage-gated K⁺outward currents are shown in figures 4B and Cand figure 5. In the presence of xenon (figs. 4B and 5), I^K,towas suppressed by 6.1 ± 3.7% (n = 10; 4 patients), from 8.2 ± 6.0 pA/pF to 7.7 ± 5.8 pA/pF. Even though this effect was small, it was consistently observed and fully reversible (I^K,toafter washout of xenon: 100.4 ± 8.8%; n = 10). Therefore, the xenon-induced inhibition of I^K,towas highly significant (P < 0.001). In contrast to I^K,to, I^K,suswas virtually unaffected by xenon. Halothane (0.75%) induced a reversible depression of I^K,toas well as of I^K,sus(figs. 4C and 5). Stronger effects on both currents (fig. 5) were evoked by a higher halothane concentration (1.5%). Isoflurane (1.2%) preferentially depressed the I^K,to, whereas no significant effect was demonstrated on I^K,sus(fig. 5).

In the current study, we analyzed the effects of the three anesthetic gases halothane, isoflurane, and xenon on Ca²⁺and K⁺currents in human myocardial cells prepared from the atria of patients without evidence of atrial dysfunction. Inhibitions of Ca²⁺as well as of K⁺currents were demonstrated for halothane and for isoflurane, whereas xenon depressed only K⁺currents. Anesthesia may be understood, according to one of several currently discussed hypotheses, as action on specific ion channels. 24,25It has been suggested that inhibition of the N -methyl-d-aspartate (NMDA) receptor is the main effect of the anesthetic drug ketamine 24; an inhibition of the same channel has been reported for xenon and perhaps may explain its anesthetic properties. 26For other anesthetics and sedatives such as the barbiturates and benzodiazepines, interactions with γ-aminobutyric acid type A (GABA^A) receptor are considered essential. 24If interference with one particular ion channel is the principal action of any given anesthetic drug, it is not surprising when it affects other ion channels as well. The spectrum of side effects is then dependent, to a large part, on the type of channels involved, the physiologic role of these channels, and the potency of the drug to interfere with these channels. In any case, actions on myocardial channels may be the basis for negative inotropic and proarrhythmic effects on the heart. In this respect, the present study suggests that xenon may be superior to halothane and isoflurane since only xenon left L-type Ca²⁺currents unaltered and evoked the least inhibition of voltage-gated K⁺outward currents. Thus, clinical observations that favor xenon as the anesthetic drug with the least cardiac side effects are supported, at a cellular and molecular level, by the results of the current study.

To the best of our knowledge, this is the first study in which human myocardial cells have been used to determine whether anesthetic gases alter ion currents. Moreover, it is the first one to demonstrate that isoflurane and xenon modify I^K,toin the heart; at the same time, it provides further evidence of an action of halothane similar to that recently been found in a study on rat ventricular myocytes. 27I^K,tois characteristic for cells prepared from mammalian myocardial tissues. 28,29In the human heart, it has been found in the atria and in the ventricles, whereas I^K,susis less prominent in the ventricles. 30I^K,toconsists of two components, 31which were not discriminated in the current study. Moreover, Cd²⁺used to block contaminating Ca²⁺currents is known to shift the steady-state activation and inactivation curves of I^K,to. 32Therefore, our study does not assess I^K,toin its full complexity but demonstrates a significant albeit slight inhibition of its amplitude by xenon. The inhibition was so small that I^K,tovaried between different cells to a much higher degree than the amount by which it was depressed by xenon in any one individual cell. The variability of voltage-gated K⁺currents in the human heart has been consistently high in many studies. 30,33–35Obviously, there are regional differences within the heart, 30but even within cells of the same regional origin, 34,36the variability was comparably high, as in our cells from the auricles. In spite of the variability of I^K,to, the effects of xenon could be resolved because they occurred consistently and were fully reversible.

A markedly stronger inhibition of I^K,towas observed during exposure to halothane and isoflurane. At equipotent anesthetic concentrations, corresponding to 1 MAC of either gas, isoflurane evoked a more pronounced effect on I^K,tothan halothane. This order of relative potencies of the two volatile anesthetics was turned around when I^K,susand I^Ca,Lwere considered. Halothane was the only drug that depressed I^K,susand evoked the largest inhibitions of I^Ca,L.

The inhibitory effects on both K⁺and Ca²⁺currents are likely to be of pathophysiologic relevance. Slowly inactivating K⁺channels that together produce I^K,susare important for the repolarization and the timing and duration of the refractory period. 37An even more important role is attributed to the L-type Ca²⁺channels. Diminished expression of this channel type is not only a hallmark of the ventricular myocardium in several different states of insufficiency 38but also a characteristic finding of the atrial myocardium in patients with persistent atrial fibrillation. 39The reduced contractile force and the electrical instability of persistent atrial fibrillation are accounted for by the reduced number of L-type Ca²⁺channels. 40A comparison of L-type Ca²⁺channels between the atrium and the ventricle reveals that ventricular cells express no other α1 subunit mRNA than α1C, whereas α1C as well as a small portion of α1D is found in atrial cells. 41However, as the α1C subunit mRNA was approximately 40-fold more abundant, there is no reason to expect different effects of anesthetics on Ca²⁺currents in ventricles than in atria. The situation may be more complex in the case of voltage-gated K⁺channels, because an exact assignment of currents to particular gene products is not possible. Given the considerable regional heterogeneity of members of the K v  gene family within the heart, 42especially with respect to differences between atria and ventricles, 43the observations of the effects of the anesthetic gases in atrial cells should not be extrapolated without caution to other cardiac regions.

Because we studied human cells, it is of interest to relate our results to those obtained in animal studies. In our investigation, halothane inhibited I^Ca,Lby 32% and 46% at concentrations corresponding to 1 and 2 MAC. In various studies of I^Ca,Lin guinea pig 10,15,16,44and canine 9,45ventricular myocytes, halothane at approximately 2 MAC inhibited by 40–70%. The strongest inhibition was observed when the experiments were performed at 37°C instead of at room temperature;44a similarly strong inhibition was observed in atrial myocytes at 37°C by 2.6 MAC halothane. 16Thus, the susceptibility to halothane does not seem to vary markedly between the investigated species. The situation may be different with isoflurane, as far as can be deduced from the few published studies. Isoflurane at 1 MAC inhibited I^Ca,Lby only 10% in guinea pig cardiomyocytes 10but by approximately 30% in canine cardiomyocytes. 9,45Incidentally, our own experiments reveal an inhibition just in between those two extremes. Moreover, when I^K,susare considered, the inhibition evoked by halothane (2 MAC) on I^K,susin human atrial cells (21%) appears moderate in relation to the reported 54%16or 57%13in the guinea pig at 2.6 MAC.

A further point with regard to which our study of human cells has produced results contrasting with those previously reported for cells from other mammals is the apparent lack of any inhibition of I^K,irby all three anesthetic gases. Although the same finding has been reported with regard to xenon in guinea pig cardiomyocytes, 15there was significant inhibition by halothane as well as by isoflurane in another study of guinea pig cardiomyocytes. 46However, it is not certain whether this reflects true species differences or whether the differences may be related to other experimental conditions, because another study of guinea pig cardiomyocytes did not show any alteration of I^K,irin the presence of halothane. 16Inwardly rectifying K⁺channels are not voltage-gated, and their major role is in controlling the resting potential. Therefore, our results suggest that the electrophysiologic influence of anesthetic gases on the human myocardium preferentially bears impact on the action potential, particularly its duration and repolarization.

A limitation of our study is the fact that we performed all experiments at room temperature. Although ion currents in general tend to be much larger at 37°C than at 21°C, it would be interesting to know whether the extent of inhibitory effects exerted by anesthetic gases might be temperature-dependent. To the best of our knowledge, no study has compared the effects of volatile anesthetics on cardiac ion currents in dependence of the temperature. In our laboratory, exposure of cells to solutions with defined concentrations of gases, especially xenon, could be performed only at room temperature. On the other hand, application of xenon and of the two volatile anesthetics at defined concentrations was a major aim of our study. We consider this particularly important with regard to xenon, because few or no effects were observed, and the high MAC value precludes the determination of the concentration–effect relation over a large range. The technical problems that arose in preliminary experiments came as a surprise. For example, we would not have expected that a sizeable loss of gases occurs through plastic tubes. Even with all the precautions that we took for the final series of experiments, we could not completely prevent loss of xenon. Therefore, we had to resort to gassing with mixtures with a higher xenon percentage than was finally provided to the cells. In contrast, no such loss was observed for halothane, isoflurane, or oxygen. Therefore, we believe that xenon concentrations should be rigorously measured and controlled in all further in vitro studies.

In conclusion, our study demonstrates inhibition of voltage-gated ion currents in human atrial myocytes by the three anesthetic gases halothane, isoflurane, and xenon. Each gas had its individual profile of effects. The results contribute to our understanding of how the three anesthetics exert their cardiac effects at the cellular and molecular level.

The authors thank Wolfgang Bettray, Ph.D. (Assistant Professor, Department of Chemistry, Rheinisch-Westfällische Technische Hochschule [RWTH] Aachen, Aachen, Germany), and Manfred Erkens, M.D. (Senior Fellow, Institute of Forensic Medicine, RWTH Aachen), for their analysis of gas concentrations with GC/MS/SIM. The authors acknowledge the excellent cooperation of the surgeons and the staff of the Department of Cardiothoracic Surgery, University Hospital RWTH Aachen. Part of the xenon supply was a kind gift of Messer-Griesheim (Krefeld, Germany).