Oxygen Tension Modulates Inhibition of L-type Calcium Currents by Isoflurane in Human Atrial Cardiomyocytes

Right atrial appendages were obtained as surgical specimens from patients undergoing heart surgery. The patients (n = 30, 23 men, 7 women) had a mean age of 67 yr (range, 45–85 yr), and all underwent a coronary artery bypass operation (n = 23) or aortic valve replacement (n = 7). All patients were in sinus rhythm and had no evidence of right heart failure. Their obligatory medication included β blockers and nitrates. The examinations were performed in accordance with the principles outlined in the declaration of Helsinki and approved by the local institutional review board (Ethik Kommission, Rheinisch-Westfälische Technische Hochschule Aachen, Germany). All patients gave written informed consent before surgery.

Human atrial cardiomyocytes were prepared according to Hatem et al. ,21as modified and described in detail by Skasa et al.  22Briefly, the myocardial specimens were carefully cut into chunks and washed in Ca²⁺-free buffer twice for approximately 5 min. Afterward, the tissue was incubated in Ca²⁺-free buffer that contained protease XXIV (Sigma, Deisenhofen, Germany) and collagenase V (Sigma). Incubation was finished as soon as microscopic examination revealed intact single cardiomyocytes. After centrifugation, cardiomyocytes were resuspended in a buffer containing the following: 120 mm NaCl, 5.4 mm KCl, 5 mm MgSO⁴, 5 mm sodium pyruvate, 20 mm glucose, and 10 mm HEPES, pH adjusted to 7.4 with NaOH. Because visible cell damage to these cells would occur if a physiologic Ca²⁺concentration was immediately restituted, the CaCl²concentration was gradually increased to 1.8 mm immediately before measurements of I^Ca,L. Only well-striated, bleb-free, rod-shaped myocytes were used for the studies, performed within 4 h after isolation.

Cardiomyocytes were allowed to adhere to a glass coverslip that was transferred into a perfusion chamber. The cells were continuously superfused as described previously.5I^Ca,Lwere recorded with the whole cell patch clamp technique using an amplifier (EPC-9; HEKA, Lambrecht, Germany) and a personal computer equipped with Pulse 8.5 software (HEKA) for data acquisition and analysis. The patch pipettes, pulled from borosilicate glass and fire polished, had a tip resistance of 3–5 MΩ when filled with pipette solution. A holding potential of −60 mV was chosen to minimize sodium currents. I^Ca,Lwere evoked by a series of depolarization pulses to potentials ranging from −50 to +60 mV. Ca²⁺currents are given here as peak current amplitude of the respective depolarizing step. To save time, pulses only to +10 mV were used to evaluate the effects of low Po²and anesthetics in some experiments.

The bath solution contained 136 mm tetraethylammonium chloride, 1.8 mm CaCl², 1.8 mm MgCl², 10 mm glucose, and 10 mm HEPES, pH adjusted to 7.4 with tetraethylammonium hydroxide. The pipette solution contained 140 mm CsCl, 2 mm MgCl², 0.3 mm adenosine 5′-triphosphate, 0.3 mm guanosine 5′-triphosphate, 10 mm EGTA, and 10 mm HEPES, pH-adjusted to 7.2 with CsOH.

Bath solutions were bubbled with either 100% nitrogen or with air in a gastight glass flask equipped with a frit and a membranous septum. Gas could leave the flask through a valve. The Po²values were measured with a calibrated oxygen-sensitive probe (Unisense, Aarhus, Denmark; tip diameter, 500 μm). Gassing with nitrogen for approximately 30 min resulted in a decrease of the Po²to approximately 1 mmHg. Bath solutions containing isoflurane were bubbled with an appropriate gas mixture (isoflurane plus air or nitrogen). An anesthetic gas analyzer (Vamos; Dräger, Lübeck, Germany) was used to continuously monitor the concentration of isoflurane in the gas phase. A fraction (10–20 ml) of gassed solutions was taken through the septum of the reservoir into gastight syringes (1010 TLL; Hamilton, Bonaduz, Switzerland) and immediately used for experiments. The bath chamber (volume approximately 300 μl) was continuously perfused through high-performance liquid chromatography–grade steel capillaries at a rate of 2–5 ml/min. The surface of the bath was covered with a thin layer of paraffin oil (DAB10; Wasserfuhr, Bonn, Germany) saturated with nitrogen to minimize diffusional exchange of oxygen and isoflurane. The isoflurane concentration in the reservoir as well as in the bath chamber was verified by headspace gas chromatography-mass spectrometry–selected ion monitoring (GC/MS/SIM, SSQ 7000; Finnegan, Bremen, Germany) as described previously.5 

To examine the effects of a lowered Po²on I^Ca,L, the myocytes were exposed to a hypoxic solution in the bath chamber by rapid solution exchange, followed by continuous perfusion of the chamber with the same hypoxic solution. The Po²was continuously recorded with an oxygen-sensitive probe (Unisense) placed into the chamber close to the examined cell. A final Po²of 12 ± 1 mmHg was achieved in the bath chamber. This condition could be held stable throughout the measurements of I^Ca,Lin one myocyte (approximately 5–7 min).

The thiol-specific oxidizing agent DTNB (Sigma) was dissolved in ethanol (100%) to yield a stock solution with 2 mm that was stored at 4°C. Bath solutions contained 200 μm DTNB; in all experiments using DTNB, ethanol was added to control solutions resulting in the same final concentration (1% vol/vol) as the DTNB solutions. The thiol-specific reducing agent DTT (Sigma) was dissolved in water.

Most experiments were carried out at room temperature (21°C). Some experiments were performed at 36°C. In this case, the temperature of the solution in the double-barreled stainless steel tubes and the bath chamber was kept constant with a feedback-controlled Peltier device (Strothmann, Aachen, Germany). The temperature was continuously monitored with a thermocouple probe mounted in the chamber wall.

Current amplitudes mark the peak current during one depolarizing pulse. Current densities were calculated by dividing current amplitudes by the whole cell capacitance amounting to 85 ± 7 pF in 59 cells. Values are expressed as mean ± SEM. Statistical comparisons were performed with SPSS version 12.0 software (SPSS Inc., Chicago, IL). Differences in I^Ca,Lwere tested for statistical significance using the Wilcoxon nonparametric test for paired data. Unpaired data were tested using the nonparametric Mann–Whitney U test. A probability of error of P < 0.05 was considered significant.

To study effects of the oxygen tension on I^Ca,Lin cardiomyocytes, we continuously measured the Po²in the immediate neighborhood of a patch clamped cell. When the superfusate was changed from a solution with a Po²of 150 mmHg to a solution gassed with nitrogen, a decrease in Po²resulted that fully developed over approximately 1 min. Then a stable plateau of 12 ± 1.0 mmHg (range, 9–15 mmHg) was reached. Lower Po²values could not be obtained. When the superfusate was switched back to the original solution with the high Po², Po²values were restored within 20 s. I^Ca,Lwas measured in intervals of approximately 20 s as peak current values after depolarizing steps from a holding potential of −60 mV to +10 mV. Unexpectedly, I^Ca,Ldid not exhibit any noticeable changes when the Po²was changed from the control conditions (150 mmHg) to 12 mmHg (fig. 1A) and back to 150 mmHg (not shown). In the mean, I^Ca,Lwas 2.00 ± 0.30 pA/pF (n = 11) in control conditions, 2.03 ± 0.31 during low Po², and 1.97 ± 0.31 after restoration of the control Po²(fig. 1B). I^Ca,Lat other membrane potentials ranging from −30 to +30 mV were not significantly altered by changes in the Po², either (fig. 1C). Moreover, no modifications of I^Ca,Lwere observed during the transition phases of the Po². Likewise, the kinetics of the currents during each depolarizing step were virtually identical in each Po²condition (fig. 1A).

Although the range over which the Po²was varied in our experiments was well within the range for which significant changes of I^Ca,Lhave previously been reported,9,10we reasoned that the Po²might not be sufficiently low to affect I^Ca,Lin human cardiomyocytes. Because oxygen contamination, probably due to diffusion, led to a failure of our attempts to reduce the Po²further by oxygen-free superfusion of the cells, we tried to remove oxygen enzymatically23,24by adding 100–300 mU/ml Oxyrase (Oxyrase Inc., Mansfield, OH) to the superfusate. Indeed, very low Po²values could be obtained with Oxyrase (even below 1 mmHg). I^Ca,Lwas completely suppressed by these solutions. However, this suppression was irreversible. Moreover, the effects of Oxyrase could not be attributed to Po²reductions because exposure to Oxyrase evoked the same irreversible inhibition of I^Ca,Lin the presence of a Po²of 150 mmHg (data not shown).

The experiments reported so far were performed at room temperature. We also tested whether a potential sensitivity of I^Ca,Lto decreased oxygen tensions might reveal in physiologic temperatures (36°C). The problem with experiments at this temperature is that a significant run-down of I^Ca,Loccurs in the majority of experiments, in contrast to those at room temperature when I^Ca,Lwas usually stable over 10–15 min. Typically, run-down at 36°C reduces I^Ca,Lby as much as 50% over 8 min, making the unequivocal detection of small inhibitions on top of the run-down impossible. However, in none of the experiments in which a low Po²was applied to cells at 36°C (n = 7) was any decrease of I^Ca,Lobserved that was not explainable by run-down. Moreover, the restoration of a Po²of 150 mm did not result in a deceleration of time-dependent run-down of I^Ca,L. In two experiments, run-down was slower than usual (fig. 1D); in these cells, I^Ca,Lwas identical in a Po²of 150 and of 14 or 12 mmHg. Therefore, no evidence was obtained that a low Po²suppressed I^Ca,Lat 36°C.

Next, we studied the interaction between oxygen and isoflurane with respect to their effects on L-type calcium channels. Isoflurane was added to superfusates at a final concentration of 1.2% (vol/vol), resulting in 0.38 mm and corresponding to a minimum alveolar concentration of 1 at 21°C. Previous experiments at a Po²of 150 mmHg had shown that isoflurane evokes a rapid, stable, and quickly reversible inhibition of I^Ca,L.5,17,18In the current experiments, cells were kept at 150 mmHg Po²(control conditions) and then exposed either to isoflurane at a low Po²(9–15 mmHg) or to isoflurane at a control Po²(150 mmHg). Afterward, the original control conditions were restored. The results (fig. 2) show a weaker inhibition of I^Ca,Lby isoflurane at the lower than the control Po². Isoflurane reduced I^Ca,Lby 17 ± 2.0% at the control Po²but only by 5.8 ± 2.9% at the low Po². When the decreases in I^Ca,Linduced by isoflurane at a Po²of 150 mmHg (n = 8) were compared with those at 9–15 mmHg (n = 7), a slightly significant difference was revealed (P = 0.037, Mann–Whitney U test). To demonstrate more conclusively that isoflurane-mediated inhibitions were dependent on the oxygen tension, we designed a protocol in which the effect of two Po²values on isoflurane-induced current inhibitions was studied in each cell, allowing the application of a paired test for statistical significance. Isoflurane was applied at a Po²of 9–15 mmHg to cells kept previously at 150 mmHg Po². Then, with isoflurane still present in the superfusate, the Po²was increased to 150 mmHg. Finally, isoflurane was removed while the Po²was kept at 150 mmHg (fig. 3). In these experiments, there was a significant stronger inhibition of I^Ca,Lby isoflurane at 150 than at 12 mmHg Po²(P = 0.018, Wilcoxon rank test, n = 7). The data furthermore show that hardly any inhibition of I^Ca,Lwas evoked by isoflurane at the low Po². Washout of isoflurane consistently led to a full recovery of I^Ca,L.

The results so far are consistent with the view that not so much isoflurane but rather the high Po²(because 150 mmHg represents an oxygen tension considerably exceeding the physiologic values) is the decisive factor that accounts for the reduction of I^Ca,Lwhen isoflurane is applied at a Po²of 150 mmHg. Because the Po²might affect membrane proteins such as calcium channels through oxidation and reduction of sulfhydryl groups, we tested the thiol-specific oxidant DTNB in comparison with the thiol-specific reducing agent DTT. The reducing agent DTT (2 mm) did not affect I^Ca,Lat basic conditions (fig. 4A) or in the presence of isoflurane (fig. 4B). Specifically, DTT did not change I^Ca,Lwhen the agent was added to a standard bath (3.41 ± 1.38 vs.  3.46 ± 1.36 pA/pF, n = 6). Furthermore, when a control value of I^Ca,Lwas determined in a standard bath and then the known inhibition by isoflurane was provoked, the further addition of DTT (in the continuous presence of isoflurane) left I^Ca,Lat the same level as in the presence of isoflurane alone (2.80 ± 0,88 vs.  2.86 ± 0.94 pA/pF, n = 5). In contrast to DTT, the oxidizing agent DTNB (200 μm) had profound effects on I^Ca,L(figs. 4C and D). Addition of the agent to a standard bath evoked an immediate reduction of I^Ca,Lby 11 ± 2% (n = 10). However, this effect was irreversible; after extended washout, I^Ca,Lremained on the diminished level for several minutes, demonstrating that inhibitions of I^Ca,Lby DTNB cannot be attributed to run-down of I^Ca,L(fig. 4C). Therefore, I^Ca,Lin human cardiomyocytes are more sensitive to oxidation and high oxygen tensions than to reduction and low oxygen tensions.

In this study, we have examined how oxygen tension affects I^Ca,Lin human atrial cardiomyocytes at basal conditions as well as in the presence of the anesthetic gas isoflurane, previously shown to inhibit these currents at high Po²values.5As main findings, a Po²between 9 and 150 mmHg did not affect I^Ca,Lin the absence of isoflurane, in discrepancy to previous findings on guinea pig cardiomyocytes. However, the Po²exerted significant effects on the inhibition of I^Ca,Lby isoflurane. Only at high Po²levels were inhibitory effects of the anesthetic gas observed. Inhibition of I^Ca,Lwas also evoked by the oxidizing agent DTNB, whereas the reducing agent DTT did not modify I^Ca,L.

Our results indicate that I^Ca,Lin the human heart are inhibited by oxidizing agents and by high Po²levels (in the presence of isoflurane) rather than by reducing agents and low Po². What could the reasons be for the apparent discrepancy of the current results, when compared with those obtained in guinea pig cardiomyocytes? First, one could assume that the temperature is relevant. The author of the study in guinea pig cells10chose a temperature of 37°C, and such experimental conditions are more difficult to establish than room temperature. From our experience, experiments at physiologic temperatures have a high failure rate due to rapid run-down of I^Ca,L.25The situation could not be improved by use of the perforated patch technique because this requires more time without producing more stable conditions in our hands. Despite the problems with run-down, we considered it essential to test the hypothesis that Po²effects on I^Ca,Lwould become apparent at physiologic temperatures. Not any indication was found for an inhibition of I^Ca,Lby decreased Po²values, but run-down of I^Ca,Lprevented the definite exclusion of a Po²-induced inhibition in most experiments. However, we have two experiments in which only moderate run-down of I^Ca,Loccurred. Even in these experiments, we observed no reduction of I^Ca,Lby a low Po², indicating that human cardiomyocytes are not sensitive to Po²in the way of guinea pig cardiomyocytes.10 

The second reason for the discrepancy could be the existence of species-dependent (human vs.  guinea pig) and cell-dependent (atria vs.  ventricles) differences in the oxygen sensitivity of cardiomyocytes. In this context, it is noteworthy that no strict correlation between the extent of Po²reductions and the effects on I^Ca,Lwas reported in guinea pig cardiomyocytes.10On the other hand, such a correlation was clearly established in human embryonic kidney 293 cells expressing α^1Csubunits.9Moreover, no saturation of the Po²-induced I^Ca,Lsuppression was evident in the latter cell model, even at the lowest attainable Po²values. Therefore, the oxygen-sensing capability of a cell may apply to different ranges of Po²variations. Unfortunately, we could not achieve hypoxic conditions with Po²values considerably lower than 12 mmHg. Our attempts with Oxyrase did not allow us to attribute the observed marked alterations of I^Ca,Lto hypoxia but rather to some oxygen-independent effects of the enzyme. However, it should be stressed that our experiments tested a range of Po²values that produced dramatic changes of I^Ca,Lin both guinea pig cardiomyocytes and human embryonic kidney 293 cells expressing human α^1Csubunits, again pointing to distinctly different Po²sensitivity in human cardiomyocytes than in these other models.

Finally, the results that have been reported from studies on different isoforms of Ca²⁺channel α^1Csubunits deserve consideration. Only one of three isoforms, or splice variants, of the proteins conferred oxygen sensitivity to I^Ca,Lin human embryonic kidney 293 cells.19Although all three isoforms are likely to exist in the human heart,20a quantitative or functional analysis has not yet been performed. Such an analysis should also consider regional differences, in particular between atria and ventricles. Currently, a quantitative determination of the expression pattern and of I^Ca,Lcould not be performed simultaneously and in identical cells. If there were a preponderance of oxygen-insensitive over oxygen-sensitive Ca²⁺channel proteins in our cells, the experimental results would be explained.

The molecular basis of how the two gases oxygen and isoflurane modulate ion currents has not been fully elucidated. An interaction of isoflurane with channel proteins is likely and may be a predominant factor in many conditions.18Similarly, oxygen may interact with L-type Ca²⁺channels directly, as proposed on the basis of the experiment of heterologously expressed channels26and of results obtained with thiol-specific reducing and oxidizing agents.27It has been concluded that the open probability of L-type Ca²⁺and other channels is critically dependent on the oxidative state of various sulfhydryl groups which are differently amenable to redox-modulating compounds.26,27However, the literature on effects of thiol-reducing and oxidizing agents is full of inconsistencies, as summarized in a recent review.28For example, in guinea pigs where I^Ca,Lis sensitive to decreased Po²values, the reducing agent DTT mimicked the effects of low Po², whereas the oxidizing agent DTNB partly reversed the inhibition by low oxygen tensions.10Mostly in line with these results, DTT attenuated and DTNB stimulated basal I^Ca,Lin ferret cardiomyocytes.11In definite contrast, the sulfhydryl oxidant 2,2′-dithiopyridine was inhibitory on heterologously expressed α^1Csubunits of smooth muscle from rabbits, whereas DTT had only negligible effects on basal I^Ca,L.27The oxidizing agent p-chloromercuribenzene sulfonic acid led to an irreversible inhibition of calcium currents,26as did DTNB in our hands. To explain the discrepancies between various studies, again general species differences may be considered as well as differences in the oxygen sensitivity of various channel splice variants. Moreover, even in guinea pigs, the effects of oxidizing agents are not unequivocal because DTNB alone evoked some effects inconsistent with a pure antagonism of low Po²,10and another oxidizing agent, p-hydroxy-mercuric-phenylsulfonic acid, reduced rather than enhanced I^Ca,L, and these effects were at best partly reversed by DTT.29To our best knowledge, the current study is the first one that analyzes effects of DTNB and DTT in human cardiomyocytes; as in the case of the oxygen sensitivity, a distinct divergence to the situation in guinea pigs is evident. However, our results are fully in line with those of Fearon et al. ,26who heterologously expressed one human α^1Csubunit and demonstrated lack of effects of a reducing agent and irreversible inhibition by an oxidizing agent.

From a clinical point of view, it is an interesting question how the cardiac actions of isoflurane are modulated by clinically relevant Po²values. In the past, there has been much debate regarding which Po²around a cell represents physiologic conditions and which hypoxia.30–34A Po²of 150 mmHg is likely to exceed the physiologic range,33although such a Po²is standard in most experiments in isolated cells. Therefore, it could be argued that isoflurane inhibitions of I^Ca,Lat a Po²of 150 mmHg represent a situation where the anesthetic gas acts in the presence of another gas, i.e. , oxygen, in unphysiologic and potentially toxic concentrations. In this line, isoflurane may be viewed as an agent unmasking inhibitory effects of oxygen on I^Ca,L, whereas isoflurane-dependent inhibitions of I^Ca,Lare dramatically reduced when the Po²is decreased to probably more physiologic values. It remains to be clarified how sulfhydryl groups in L-type channel proteins govern the interaction of isoflurane and oxygen.

In conclusion, the current experiments reveal a striking functional difference of I^Ca,Lin human atrial cardiomyocytes, in comparison with cells from guinea pig ventricles, because I^Ca,Lwas not modulated by Po²in human cells in the way demonstrated for guinea pig cells. Our results might be explained by a predominant expression of oxygen-insensitive α^1Csubunits. Regional differences of the expression pattern of α^1Csubunits may be present in the human heart, but species differences seem to be more relevant. Furthermore, the study demonstrates a clear dependence of the isoflurane-induced inhibition of I^Ca,Lon Po². An inhibition occurs mostly at standard experimental Po²conditions, but these are likely to represent hyperoxia not normally present in the microenvironment of living cells.33,34At lower oxygen tensions, isoflurane becomes almost inert on I^Ca,Lin human cardiomyocytes from atrial appendages.

The authors thank Manfred Möller, Ph.D. (Assistant Professor, Department of Hygiene and Environment Medicine, University Hospital, Rheinisch-Westfälische Technische Hochschule Aachen, Aachen, Germany), for analysis of gas concentrations with gas chromatography. The authors acknowledge the excellent cooperation of the staff of the Department of Thoracic and Cardiovascular Surgery, University Hospital, Rheinisch-Westfälische Technische Hochschule Aachen, Aachen, Germany.