Differential Effects of Etomidate, Propofol, and Midazolam on Calcium and Potassium Channel Currents in Canine Myocardial Cells

These experiments were approved by the Medical College of Wisconsin Animal Care Committee, and all experimental procedures strictly conformed to the standards of American Association for Accreditation of Laboratory Animal Care. Adult mongrel dogs weighing 15 to 25 kg were placed in a Plexiglas box and anesthetized with halothane. After surgical anesthesia was attained, the trachea was intubated and the lungs were ventilated with 1.5% halothane in oxygen. The chest was opened and the heart was excised rapidly and placed in cold Krebs' solution. Thin strips of the ventricular tissue were cut into 5-mm lengths and placed in cold cardioplegia solution composed of 10 mM NaCl, 60 mM KCl, 5 mM MgCl², 20 mM glucose, 100 mM sucrose, 5 mM pyruvate, 20 mM taurine, 5 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), and 0.5 mM disodium salt of ethylenediamine tetraacetic acid (Sodium²-EDTA). The washed ventricular segments were placed in Calcium²+ -free Tyrode's solution composed of 140 mM NaCl, 5.4 mM KCl, 5 mM MgCl², 5.5 mM glucose, 5 mM HEPES, 4 mg/ml collagenase (Type I, 161 units/mg; Worthington, Freehold, NJ), 2 mg/ml bovine albumin (Sigma Chemical Co., St. Louis, MO), 40 micro Meter CaCl², 5 mM pyruvic acid, and 5.5 mM disodium salt of adenosine triphosphate (Sodium²-ATP); the pH was adjusted to 6.2 with NaOH. The solution containing ventricular tissue was incubated at 37 degrees Celsius for 1 to 2 h in a slow shaker. After this incubation, single ventricular cells were washed three times in cold potassium glutamate (K-glutamate) solution composed of 130 mM Potassium-glutamate, 5.7 mM MgCl², 5 mM HEPES, 5.5 mM glucose, 5 mM Sodium²-ATP, and 0.12 mM Sodium²-EDTA; the pH was adjusted to 7.4 with KOH. Dispersed cells were stored in Potassium-glutamate solution at 4 degrees Celsius before use.

A drop of dispersed single canine ventricular cells was placed in a perfusion chamber (22 degrees Celsius) on the stage of an inverted microscope (Olympus IMT-2; Leeds Instruments, Minneapolis, MN) equipped with modulation contrast. At 500x magnification, a hydraulic micromanipulator (Narishige, Tokyo, Japan) was used to position heat-polished borosilicate patch pipettes with tip resistance of 4 to 6 M Ohm on the membrane of ventricular cells. High-resistance seals (3 to 30 G Ohm) were formed, after which the pipette patch was removed by negative pressure to obtain the electrical access to the whole cell as previously described. [17]Whole-cell currents were elicited by 200-ms depolarizing pulses generated by a computerized system (pClamp software; Axon Instruments, Burlingame, CA) every 5 to 10 s. The currents were amplified by a List EPC-7 patch-clamp amplifier (Adams & List Associates, Great Neck, NY), and the amplifier output was low-pass filtered at 500 Hz. All data were digitized (sampling rate = 10,000/s) and stored on a hard disk to permit analysis at a later time. For the I sub Ca, the leak and capacitative currents were subtracted from each record by linearly summating scaled currents obtained during 10-mV hyperpolarizing pulses.

The external solution used to measure I^Cacontained: 10 mM BaCl², 135 mM tetraethylammonium chloride, 1 mM MgCl², 10 mM glucose, and 10 mM HEPES (pH = 7.4). The pipette solution used to measure I^Cacontained: 130 mM CsCl, 5 mM adenosine triphosphate (magnesium salt), 5 mM ethyleneglycol-bis-(beta-aminoethyl ether) N,N,N',N'-tetraacetic acid (EGTA), 10 mM HEPES, 1 mM MgCl², and 10 mM glucose (pH = 7.2).

The external solution used to measure I^Kcontained: 135 mM NaCl, 5.5 mM dextrose, 0.5 mM MgCl², 4 mM KCl, and 2 mM CaCl²(pH = 7.4). The pipette solution used to measure I^Kcontained: 125 mM potassium aspartate, 20 mM KCl, 10 mM EGTA, 5 mM adenosine triphosphate (magnesium salt), 1 mM MgCl², and 5 mM HEPES (pH = 7.2).

Repetitive current-voltage curves were obtained in the control solution to monitor time-dependent changes in I^Caand I^K, followed by exposure to 5, 30, or 60 micro Meter etomidate, midazolam, or propofol. Effects of anesthetics on I^Caand I^Kwere complete within 6 min and were reversible with washout.

The recommended intravenous doses to induce anesthesia are approximately 0.2 to 0.4 mg/kg for midazolam, 0.3 mg/kg for etomidate, and 2.5 mg/kg for propofol. It was reported that plasma protein-bound fraction of midazolam and propofol is 97%, whereas for etomidate it is 76%. Peak plasma concentrations during induction with these doses vary widely but are reported to be approximately 0.5 to 3 micro Meter for midazolam, [1,18-20]3 micro Meter for etomidate, [21]and 50 micro Meter for propofol. [22,23],**** For our study, intravenous anesthetics were obtained commercially in their vehicles (2 mg/ml etomidate, 5 mg/ml midazolam, and 10 mg/ml propofol). Each anesthetic was diluted in a given extracellular solution to make 1 mM stock solution that was divided into aliquots, frozen at -15 degrees Celsius, and thawed for daily use. Measured volumes of the 1-mM concentration of each drug were diluted into a known volume of perfusate to obtain perfusate drug concentrations of 5, 30, and 60 micro Meter. Cells were randomly exposed to one of the anesthetic agents, at one of three concentrations. 4-aminopyridine (4-AP; Sigma Chemical Co.) and barium chloride (Barium²+; Fisher Scientific, Itasca, IL) were dissolved directly in the recording solution. Nifedipine (Sigma Chemical Co.) was dissolved in 70% ethanol to make 1 mM stock solution, yielding a final solvent concentration in the recording solution of 0.07%, which by itself did not affect the current. Cobalt chloride (Cobalt²+; Sigma Chemical Co.) was dissolved directly in the recording solution. Tetrodotoxin (TTX; Sigma Chemical Co.) was dissolved in the recording solution to make 1 mM stock solution.

Vehicle preparations of etomidate (350 mg/mL propylene glycol in H²O), propofol (Intralipid; 100 mg/ml soybean oil, 22.5 mg/ml glycerol, and 12 mg/ml egg lecithin in water), and midazolam (0.01% disodium edetate and 1% benzyl alcohol in water, adjusted to pH of 3 with 0.1 NHCl³) were also prepared and administered in separate groups of cells (n = 2 to 4) at the concentrations that would correspond to the highest drug concentration given in this study (60 micro Meter) over the same time period (6 to 10 min). None of the drug vehicles had any effect (change < 5%) on either I^Ca, IK^to, or IK¹.

All currents are expressed as means +/- SEM and were analyzed by two-way analysis of variance. If the F-test showed significance, Fisher's test for least significant differences was performed, with the level of significance designated at P less or equal to 0.05.

Calcium current (I^Ca) was generated by depolarizing pulses (10 mV; 200 ms) from a holding potential of -40 mV to a command potential as high as +40 mV. All experiments were performed with Barium sup 2+ (10 mM) as the charge carrier in place of external Calcium²+ (2 mM); this substitution increased the peak of I^Caand decreased the rate of inactivation. In 64 cells (each group containing 5 to 9 cells), the threshold activation occurred at approximately -30 mV and the maximal activation was reached between -10 and 0 mV. This I^Caresembled high-threshold, long-lasting (L-type) Calcium²+ channel current previously described by researchers in our laboratory [24]and by others. [25-27] Figure 1(upper panel) illustrates the effects of 60 micro Meter etomidate (A), propofol (B), and midazolam (C) on the peak I^Caelicited by the test pulse from a holding potential of -40 mV to 0 mV in three canine ventricular cells. Peak I^Ca(Figure 1(A-C), lower panel) was plotted as a function of membrane potential to analyze the effects of etomidate (A), propofol (B), and midazolam (C) on the current-voltage relations for I^Caactivation. Figure 1(D) summarizes decreases of the peak I^Caamplitude by all three anesthetics. Etomidate at concentrations of 5 micro Meter, 30 micro Meter, and 60 micro Meter decreased the peak I^Caamplitude by 3 +/- 3% (n = 7, NS), 10 +/- 6% (n = 7, NS) and 16 +/- 4% (n = 6; P less or equal to 0.05), respectively. Propofol at concentrations of 5 micro Meter, 30 micro Meter, and 60 micro Meter also decreased the peak I^Caamplitude by 10 +/- 3% (n = 6, NS), 21 +/- 4% (n = 9; P less or equal to 0.05), and 33 +/- 6% (n = 9; P less or equal to 0.05), respectively. In a dose-dependent manner, 5 micro Meter, 30 micro Meter, and 60 micro Meter midazolam decreased the peak I^Caamplitude by 7 +/- 7% (n = 7, NS), 30 +/- 4% (n = 5; P less or equal to 0.05), and 47 +/- 5% (n = 8; P less or equal to 0.05), respectively.

At equimolar concentrations, midazolam was most potent among the intravenous anesthetics in inhibiting I^Ca. The effects of these agents on I^Cawere voltage independent (i.e., there was no difference in percentage decrease among different voltages).

Further analysis of the current traces produced by the depolarization step from a holding potential of -40 mV to 0 mV revealed that all three anesthetics not only decreased the peak I^Cabut also hastened the inactivation of I^Caand thus further contributed to a decrease of total Calcium²+ influx. The inactivation phase of I^Catrace was best fitted with a single exponential curve with an average control inactivation constant (tau) of 0.185 +/- 0.013 s. Although all three anesthetics decreased the inactivation constant at 60 micro Meter, etomidate given at 5 and 30 micro Meter did not significantly alter the inactivation constant (Table 1). Although propofol significantly decreased tau at two higher concentrations, it produced a significantly smaller (P less or equal to 0.05) decrease compared with the same concentration of midazolam. The effects of intravenous anesthetics on I sub Ca were readily reversible with washout.

During the 200-ms depolarizing pulses from -90 mV to consecutively more positive voltages as high as +60 mV, cells showed transient outward Potassium sup + current (IK^to) and an inward-rectifier Potassium sup + current (IK¹). To measure IK^toand IK¹, all experiments were done in the presence of 20 micro Meter tetrodotoxin (TTX) and 2 mM cobalt chloride (Cobalt²+) in the external solution to block Sodium sup + and Calcium²+ currents, respectively. Addition of nifedipine (1 micro Meter) in the external solution produced no significant changes in I^Kamplitude, indicating that this current is not influenced by the Calcium²+ influx via the L-type Calcium²+ channel. Voltage-dependent IK¹was activated approximately at voltages negative to -20 mV and was completely abolished by 1 mM Barium²+ in the external solution, as described previously. [28-30]Voltage-dependent IK^towas identified as IK^to1subtype because it was blocked by 2 mM 4-aminopyridine (4-AP) in the external solution. [31-35] 

(Figure 2(A)) shows actual recordings of the peak IK¹and IK^toelicited by depolarizing pulses from a holding potential of -90 to -50 mV, and from -90 to +60 mV, respectively, in control solution and during exposure to 60 micro Meter etomidate, propofol, and midazolam in three ventricular cells. Peak IK^toand IK¹were plotted as a function of membrane potential to analyze the mean anesthetic effect on the current amplitude (Figure 2(B)). At equimolar concentrations, etomidate, propofol, and midazolam caused reversible decreases in the peak IK^toamplitude (at +60 mV) by 8 +/- 3% (n = 8; P less or equal to 0.05), 9 +/- 2% (n = 6; P less or equal to 0.05), and 23 +/- 3% (n = 15; P less or equal to 0.05), respectively. At the same concentration, etomidate and midazolam produced reversible decreases in the peak IK¹amplitude (at -50 mV) by 20 +/- 6% (n = 8; P less or equal to 0.05), and 14 +/- 5% (n = 15; P less or equal to 0.05). Propofol had no effect on I^K1(n = 6). The effects of intravenous anesthetics on IK^toand IK¹were not voltage dependent. The decrease of IK^toand IK¹amplitude by intravenous anesthetics was completely reversed with washout.

At equimolar concentrations, all three anesthetics significantly decreased IK^to; however, midazolam was more potent than etomidate and propofol (Figure 2(C)). At the same concentration, etomidate and midazolam produced similar and statistically significant decreases in IK¹, whereas propofol had no effect (Figure 2(D)).

Similar to I^Ca, IK^tocurrent that was obtained by a depolarization step from a holding potential of -90 mV to +60 mV was also fitted with a single exponential curve to study its inactivation kinetics. The control inactivation constant (tau) was 0.034 +/- 0.0014 s. Neither etomidate nor propofol changed the inactivation constant but midazolam decreased the inactivation constant by nearly 40% (Table 1).

Intravenous anesthetics cause various degrees of cardiovascular depression in vivo [1-5]and in vitro. [13,15,16],** The differences in cardiovascular depression could result from their differential effects on systemic vascular resistance, [6,7]venous capacitance, [8,9]the autonomic nervous system, and the heart. [10,11] 

Etomidate produces only minimal effects on cardiovascular dynamics [2,11]and therefore is widely recommended for patients with compromised cardiac function and hypotension. On the other hand, propofol produces cardiovascular depression to a larger extent than does etomidate, [4,11],** and thus propofol should be used cautiously in patients with hypovolemia. It was reported that induction of anesthesia with midazolam, even in patients with limited coronary flow, was accompanied by no change in cardiac output or central venous pressure and only a modest reduction in peripheral vascular resistance. [7]At equimolar concentrations as great as 50 micro Meter in isolated guinea pig heart preparation, propofol has been shown to be more potent than etomidate and less potent than midazolam in depressing myocardial contractility. [15]However, the authors suggested that at equivalent induction concentrations (concentration that produces the same depth of anesthesia), propofol would be the most potent, midazolam less potent, and etomidate the least potent negative inotrope. [15]Despite the extensive literature documenting differences in negative inotropic effects among intravenous anesthetics, very little is known about the mechanisms underlying these differences.

Our purpose in this study was to identify and compare the actions of etomidate, propofol, and midazolam on macroscopic Calcium²+ and Potassium sup + currents in isolated canine ventricular myocytes to identify possible mechanisms for the observed differences in negative inotropic effects of these anesthetics.

Inward Calcium²+ current (I^Ca), transient outward Potassium sup + current (IK^to), and inward-rectifier Potassium sup + current (I^K1) contribute to the electrical activity in human, [36]rabbit [37]and canine ventricular myocytes [38]and are the major determinants of the action potential duration in these cells. [34,39]The efflux of Potassium sup + through Potassium sup + channels is functioning as an important modulator of the action potential in canine ventricular myocytes, [35,40,41]offsetting Calcium²+ current and thus preventing early slow-response action potentials [42]while maintaining a high resting membrane potential in latent pacemaker cells. [43] 

Our results show that etomidate, propofol, and midazolam reduce the amplitude of I^Ca, IK^to, and IK¹in these cells. Furthermore, these three channel types were not equally sensitive to block by these anesthetics, and I^Cawas blocked more effectively than IK^toor IK¹. Midazolam was the most potent in decreasing peak I^Ca, whereas etomidate was the least potent. A recent study using the whole-cell and single-channel recordings in guinea pig ventricular myocytes found that 4.4 and 27.4 micro Meter etomidate decreased the whole-cell L-type I^Ca. [27]Similar concentrations of etomidate did not produce attenuation of I^Cain this study, suggesting the possibility of differences between species (guinea pig vs. dog). Another possibility is that the large decrease in I^Caobserved by Takahashi and Terrar [44]is partially due to a significant I^Carundown in guinea pig myocytes as reported by the same authors. The relative potency of intravenous anesthetics in depressing I^Cacorrelates well with in vitro and in vivo depression of contractility, supporting the hypothesis that the block of I^Cais one of the most important factors in determining the negative inotropic effect of intravenous anesthetics. Effects of all three anesthetics on I^Cawere dose dependent and readily reversible.

In addition to decreasing the peak current, intravenous anesthetics could decrease a total I^Caby increasing the rate of inactivation of open Calcium²+ channels. Indeed, analysis (single exponential fit) of the inactivation portion of the I^Carevealed that midazolam dramatically increased the rate of I^Cainactivation. Propofol was less potent, whereas only 60 micro Meter etomidate significantly increased the rate of I^Cainactivation. Absence of the effect of 5 micro Meter etomidate on I^Cainactivation rate that we found correlates with the findings of Takahashi and Terrar, [27]who reported no effect of a similar dose of etomidate on the single Calcium²+ channel kinetics. At 27.4 micro Meter, etomidate reduced the mean open time and increased the mean closed time, favoring the closed state without any effect on conductance. [27]In our study, 30 micro Meter etomidate did not alter the I^Cainactivation rate, which could be consistent with the relatively greater effect seen in their study. The same group reported that a high concentration of propofol (100 micro Meter) also decreased the mean open time and increased the mean closed time without changing I^Cachannel conductance. [44]Decreased peak current and increased rate of inactivation by propofol seen in this study are consistent with their observations at the single-channel level. The increased rate of I^Cainactivation suggests that midazolam, and to a lesser degree propofol and etomidate, bind to the open Calcium²+ channel and through an interaction with the gating mechanism increase the rate of transition of the channel from an open to inactivated state. Further studies investigating frequency and voltage dependence, rate of recovery from inactivation, and the single-channel recordings are needed to determine the exact nature of this interaction.

Midazolam was the most potent and propofol the least potent in depressing IK^toin ventricular cells. The decrease in amplitude of IK^toby these anesthetics suggests that these agents may influence the phase 1 repolarization and early plateau phase of the action potential. [31]These alterations of plateau voltage may result in changes in other voltage-dependent currents during the plateau phase, including augmented I^Ca, and therefore cytosolic Calcium²+ concentration, contractility, and duration of the action potential in the ventricular myocytes. [31,38]However, the magnitude of I^Kdecrease may not be able to overcome a large decrease in I^Ca, as suggested by shortened action potential duration in the presence of propofol. [45]It is of great interest that midazolam increased the rate of IK^toinactivation, whereas propofol and etomidate did not change channel inactivation. The mechanism(s) underlying similar effects of midazolam on I^Caand IK^toinactivation are not clear.

Etomidate exerted the greatest depressant effect on IK¹, whereas propofol had no effect. Baum [46]has reported the lack of propofol's effect on IK¹in guinea pig ventricular myocytes. Blockade of IK¹can affect the duration of the action potential and the rate of repolarization because this current is activated during the repolarization of the cardiac action potential. [47]This current plays a critical role in determining the amplitude and shape of the subthreshold response, [29,48]thus influencing the excitability of ventricular myocytes. Therefore, a depressant effect of intravenous anesthetics on the outward I^Kwould alter resting membrane potential and affect action potential duration and cellular excitability.

Despite the inhibition of both I^Caand I^K, the greater decrease of I^Camay be responsible for the negative inotropic effect observed with these agents. Although midazolam produced the largest decrease of I^Ca, the final effect on myocardial contractility would be dramatically attenuated due to a decrease of I sub K and its significantly smaller induction dose. Because the induction concentration of midazolam is less than 5 micro Meter, midazolam would probably be free of effects on I^Cawhen anesthesia is induced.

Etomidate, propofol, and midazolam produced a dose-dependent decrease of Calcium²+ influx and a type-dependent decrease of Potassium sup + efflux. These anesthetics, particularly midazolam, also increased the rate of I^Caand I^K(only midazolam) inactivation, thus further decreasing the total current. Our data suggest that cardiac depression caused by intravenous anesthetics is due, at least in part, to a decreased amplitude of I^Ca, as directly measured by the whole-cell voltage-clamp method. The relative magnitude of a decrease in I^Cacorrelates well with clinically and experimentally observed negative inotropic effects of these anesthetics. These membrane alterations probably interact with other actions of these intravenous anesthetics, including the changes in Potassium sup + currents leading to an overall effect on myocardial contractility and excitability.

*Rusy BF, Thomas-King PY, King GP, Komai H: Effects of propofol on the contractile state of isolated rabbit papillary muscles under various stimulation conditions (Abstract). Anesthesiology 1990; 73:A560.

**Williams JP, McArthur DJ, Walker EW, Rietsema K, Teunissen E, Bonnenkamp H, Goderie P, Stanley HT: A comparison of the hemodynamics of Diprivan (propofol), thiopental, and etomidate for induction of anesthesia in patients with coronary artery disease. Semin Anesth 1988; 7:112-5.

***Wegrzynowicz ES, Matsuda J, Volk K, Shibata E, Wachtel R: Propofol decreases voltage activated calcium currents in rabbit heart myocytes (Abstract). Anesthesiology 1990; 73:A592.

****White PF: Propofol: Pharmacokinetics and pharmacodynamics. Semin Anesth 1988; 7:4-20.