Intravenous Anesthetic Propofol Inhibits Multiple Human Cardiac Potassium Channels

PROPOFOL is an intravenously administered anesthetic that is commonly used clinically to induce and/or maintain anesthesia. This anesthetic has multiple advantages, for example, quick onset and rapid recovery, with minimal side effects. In addition, propofol has been reported to have both proarrhythmic and antiarrhythmic effects.¹  It has been reported that a slowed atrioventricular conduction was frequently observed in children with paroxysmal supraventricular tachycardia undergoing radiofrequency catheter ablation,²  and a completed atrioventricular blockade was found in an elderly patient undergoing knee replacement arthroplasty.³  Sinus arrest,⁴  severe bradycardia, atrioventricular blockade,⁵  polymorphic ventricular tachycardia,^6,7  and some particular arrhythmias similar to Brugada syndrome⁸  have been reported in patients receiving propofol. Among the arrhythmias induced by propofol, bradycardia (<50 beats/min) was the most frequent (4.8% of patients with propofol).⁹  Experimental studies demonstrated that blockade of cardiac L-type calcium current (I_Ca.L),^10–12  sodium current (I_Na),^13,14  and/or pacemaker current^15,16  likely contributes to the proarrhythmia observed in patients receiving propofol.

However, propofol has been reported to suppress ventricular tachycardia^17,18  and improve QTc dispersion^19,20  and has been found to be effective in the conversion of supraventricular tachycardia^21,22  or atrial fibrillation²³  to sinus rhythm. Nonetheless, the ionic mechanism underlying the antiatrial tachycardia or antiatrial fibrillation effects of propofol is not fully understood. The ultrarapid delayed rectifier potassium current I_Kur (encoded by hKv1.5) is present in the atria, but not in the ventricles of the human heart,²⁴  and is therefore believed to be a target for antiatrial fibrillation therapy.²⁵  No information is available in the literature regarding the effect of propofol on human atrial I_Kur and action potential. We hypothesized that propofol would inhibit multiple human cardiac potassium channels including I_Kur. The current study was therefore designed to determine the potential effect of propofol on I_Kur, transient outward potassium current (I_to), and action potential in human atrial myocytes using the whole cell patch clamp technique. We also investigated the effects of propofol on hKv1.5, human ether-à-go-go-related gene (hERG or KCNH2, coding for the α subunit of the rapid delayed rectifier potassium current I_Kr) channels, and hKCNQ1/hKCNE1 (coding for the slow delayed rectifier potassium current I_Ks) channels stably expressed in HEK 293 cells. Our results demonstrated that propofol inhibited multiple human cardiac ion channels including I_to, I_Kur, hERG, and hKCNQ1/hKCNE1 channels, and slightly prolonged human atrial action potential duration.

Atrial myocytes were isolated from right atrial appendage specimens obtained from 15 patients (46 to 71 yr old) undergoing coronary artery bypass. The procedure for obtaining the tissue was approved (reference No. UW-10–174) by the institutional review board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (Hong Kong, China) and a written consent was obtained from each patient. All atrial specimens were grossly normal at the time of cardiac surgery, and all patients were free of supraventricular tachyarrhythmias and symptomatic congestive heart failure. The patients were administered angiotensin-converting enzyme inhibitors, β-blockers, or calcium channel blockers before surgery. After excision, the samples were quickly immersed in an oxygenated calcium-free cardioplegic solution and immediately transported to the laboratory. Atrial myocytes were enzymatically dissociated with the procedure described previously.^25,26  The isolated myocytes were kept in a high potassium medium^26,27  at room temperature for at least 1 h before experimental recording. No randomization or blind methods were used in the present electrophysiological study.

The HEK 293 cell line stably expressing hKv1.5 (KCNA5, coding I_Kur),²⁸  hKCNQ1/hKCNE1 (I_Ks),²⁹  or hERG (I_Kr)³⁰  was maintained in Dulbecco modified Eagle medium (DMEM; Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum and 400 μg/ml G418 (Sigma-Aldrich, St. Louis, MO) for the cell line expressing hKv1.5 and hERG, or 100 μg/ml hygromycin (Invitrogen) for the cell line expressing hKCNQ1/hKCNE1. Cells used for electrophysiology were seeded on a glass cover slip.

Membrane currents were recorded with the whole cell patch clamp technique as described previously.^26,28,31  In brief, a small aliquot of the solution containing the isolated myocytes was placed in an open perfusion chamber (1 ml) mounted on the stage of an inverted microscope (IX50; Olympus, Tokyo, Japan). Myocytes were allowed to adhere to the bottom of the chamber for 10 to 20 min and then superfused at 2 to 3 ml/min with Tyrode solution. For current recording in cell lines, the glass coverslips with HEK 293 cells expressing different ion channels were transferred into the cell chamber and superfused with Tyrode solution.

Borosilicate glass electrodes (1.2-mm outside diameter) were pulled with a Brown-Flaming puller (model P-97; Sutter Instrument Co., Novato, CA) and had tip resistances of 2 to 3 MΩ when filled with pipette solution. Membrane currents were recorded in voltage clamp mode, and action potentials were recorded in current clamp mode using an EPC-9 amplifier and Pulse software (HEKA, Lambrecht, Germany). After a giga-ohm seal was obtained, the cell membrane was ruptured by gentle suction to establish whole cell configuration. The cell membrane capacitance (C_m) was directly measured using the lock-in module of the Pulse software and used for normalizing the current in individual cells. The series resistance (3 to 5 MΩ) was compensated by 50 to 80% to minimize voltage errors. Electrical signals were recorded with a sampling rate of 5 kHz and filtered at 2 kHz.

Voltage-dependent I_to traces were recorded using a protocol with 300-ms voltage steps in 10-mV increments between −40 and +60 mV from a holding potential of −50 mV at 0.2 Hz. The current was measured from the peak to “‘quasi”-steady-state level. I_Kur was determined using a 100-ms prepulse to +40 mV (to inactivate I_to) followed by a 10-ms interval before 200-ms test potentials between −40 to +60 mV from a holding potential of −50 mV, then to −30 mV (pulse interval of 10 s). The current was measured from zero current level to the current at the end of the voltage step. Voltage-dependent hKv1.5 current was elicited by 5-s voltage pulses to potentials between −40 mV and +60 mV from a holding potential of −80 mV (pulse interval, 20 s). The current amplitude was measured at the end of the 5-s depolarizing pulse. Voltage-dependent hERG current was recorded with 3-s voltage steps from −80 mV to potentials between −60 to +60 mV, then to −50 mV every 15 s. The step current was measured as the difference between zero current and the level at the end of voltage step. The tail current was measured at its peak. Voltage-dependent hKCNQ1/hKCNE1 current traces were recorded with 3-s voltage steps from −80 mV to potentials between −40 to +60 mV, then to −50 mV (pulse interval of 15 s). The voltage step-activated current was measured from the beginning of the significant activation of time-dependent current to the level at the end of the depolarization step.

To obtain IC₅₀ values, the fractional blocks obtained at different drug concentrations were fitted with the Hill equation: E = E_max/[1 + (IC₅₀/C)^b], where E is the inhibition of currents in percentage at concentration C, E_max is the maximum inhibition, IC₅₀ is the concentration for 50% inhibition of maximum effect, and b is the Hill coefficient. The activation or inactivation conductance variables of I_to, hKv1.5, hERG, or hKCNQ1/hKCNE1 were determined with normalized currents. Current activation and inactivation were fitted by the Boltzmann distribution: y = 1/{1 + exp[(V_m − V_0.5)/S], where V_m is the membrane potential, V_0.5 is the midpoint, and S is the slope factor. The relation of 1/τ_block against the concentration is described by the linear function: 1/τ_block = k[D] + l, where 1/τ_block is the time constant of development of block, and k and l are the apparent rate constants for association and dissociation of the drug.

Tyrode solution contained 140 mM NaCl, 5.4 mM KCl, 1 mM MgCl₂, 1 mM CaCl₂, 0.33 mM NaH₂PO₄, 5 mM HEPES, and 10 mM glucose; pH was adjusted to 7.4 with NaOH. The pipette solution for current recording contained 20 mM KCl, 110 mM K-aspartate, 1 mM MgCl₂, 10 mM HEPES, 5 mM EGTA, 0.1 mM guanosine-5'-triphosphate, 5 mM Na₂-phosphocreatine, and 5 mM adenosine triphosphate; pH was adjusted to 7.2 with KOH. The pipette solution for action potential recording contained 20 mM KCl, 110 mM K-aspartate, 1 mM MgCl₂, 10 mM HEPES, 0.05 mM EGTA, 0.1 mM GTP, 5 mM Na₂-phosphocreatine, and 5 mM adenosine triphosphate; pH was adjusted to 7.2 with KOH. For I_to and I_Kur recording, BaCl₂ (200 μM) and CdCl₂ (200 μM) were added to the bath solution to block I_K1 and I_Ca.L. Atropine (1.0 μM) was used to minimize possible I_K,ACh contamination during the current recording. Diphenyl phosphine oxide-1 (DPO-1; 2 μM) was added to inhibit I_Kur for determining I_to.²⁷  A stock solution of propofol (Sigma-Aldrich) at 200 mM was made in dimethyl sulfoxide.

A group size of number of 5 or more was determined based on previous experience.^25,26  Nonlinear curve fitting was performed using PulseFit (HEKA) and Sigmaplot (SPSS, Chicago, IL). Paired and/or unpaired Student two-tailed t test was used as appropriate to evaluate the statistical significance of differences between two group means. Two-way ANOVA followed by the Newman–Keuls test was used for multiple comparisons, which is referred to as “ANOVA” in the results, unless otherwise noted. Data are presented as mean ± SEM. Values of P less than 0.05 were considered to indicate statistical significance.

It is well established that both I_to and I_Kur are present in human atrial myocytes.^24,26,27,32  Voltage-dependent membrane currents were elicited in a representative human atrial myocyte with a voltage protocol as shown in the inset in the absence and the presence of propofol (fig. 1A). Propofol at 50 μM inhibited both I_to peak and the sustained I_Kur. To accurately evaluate the propofol effect on I_to, the selective I_Kur/Kv1.5 blocker DPO-1 was used to separate I_to as in a previous study.²⁷ Figure 1B shows the voltage-dependent I_to traces during control, in the presence of 2 μM DPO-1, co-presence of DPO-1 and 30 μM propofol, and washout of propofol. DPO-1 (2 μM) induced a slight reduction of peak current amplitude and almost a full inhibition of I_Kur. Propofol at 30 μM remarkably suppressed I_to, and the effect was partially reversed by washout. Figure 1C displays the time course of I_to recorded with a protocol as shown in the inset in the absence or presence of 30 μM propofol in a typical experiment with DPO-1 treatment. I_to was gradually decreased by 30 μM propofol, and the inhibition reached a steady-state level within 4 min and was partially reversed upon washout. Experiments to determine the blocking properties of I_to by propofol were conducted in the presence of 2 μM DPO-1.

The effect of propofol on current–voltage (I–V) relations of I_to was determined at 3, 10, 30, and 100 μM (fig. 2A). Propofol inhibited I_to in a concentration-dependent manner. Propofol at 10 to 100 μM suppressed I_to at test potentials of 0 mV to +60 mV (n = 6, P < 0.05 or P < 0.01 vs. control, ANOVA). No significant change in voltage dependence was observed at any concentration of propofol (fig. 2B). The concentration–response relation for inhibiting the peak current of I_to by propofol was evaluated at +50 mV and fitted with a Hill equation (fig. 2C). The IC₅₀ for inhibiting peak current of I_to by propofol was 33.5 ± 2.0 μM (Hill coefficient 1.8 ± 0.2, n = 6). Propofol decreased the peak current and increased the inactivation of I_to, which implies open-channel blockade.^33,34  Therefore, the inhibition was also estimated from the integral of the total current charge crossing the membrane at +50 mV, and propofol inhibited the I_to charge with an IC₅₀ of 20.9 ± 5.0 μM (Hill coefficient 2.1 ± 0.4, n = 6; fig. 2C).

To analyze the open-channel blocking property of I_to by propofol, the time to peak and the inactivation time constant were determined in human atrial myocytes in the absence and the presence of propofol. Figure 3A illustrates the expanded I_to traces recorded at +40 mV in the absence and the presence of 30 μM propofol. The time to peak of I_to was clearly reduced by propofol. The mean value of the time to peak of the current at 0 to +60 mV was significantly reduced by 30 μM propofol (fig. 3B, n = 6, P < 0.01 vs. control, ANOVA). I_to traces (+40 mV) were fitted by a mono-exponential function with the time constants shown before and after the application of 30 μM propofol (fig. 3C). The time constant of I_to inactivation was reduced by 30 μM propofol. The mean value of the time constant of I_to inactivation at 0 to +60 mV is shown in figure 3D. The time constant was significantly reduced by 10 or 30 μM propofol at all test potentials (n = 6, P < 0.05 or P < 0.01 vs. control, ANOVA).

The onset of the open-channel blockade was further analyzed with the method described by Slawsky and Castle³⁵  using the current traces recorded at +50 mV with 10, 30, and 100 μM propofol. Acceleration of I_to inactivation by 10, 30, and 100 μM propofol suggested an open-channel blocking effect. The drug-sensitive current expressed as a proportion of the current in the absence of the drug [(I_control − I_propofol)/I_control], where I_control and I_propofol are the current in the absence and the presence of propofol. The drug-induced block was then plotted as a function of time. The blockade developed in a time-dependent manner with an exponential onset as shown by the curve fits in figure 3E. The rate of blocking development was enhanced as the concentration increased; the time constants averaged 26.3 ± 2.9 ms, 19.1 ± 1.1 ms, and 7.6 ± 0.6 ms, at 10, 30, and 100 μM propofol, respectively (n = 6). The 1/τ_block was plotted as a function of propofol concentrations as shown in figure 3F. The straight line is a regression fit of the equation 1/τ_block = k[D] + l, and the apparent rate constants for association (k) and dissociation (l) were (1.1 ± 0.2) × 10⁶ M⁻¹ s⁻¹ and 25.2 ± 5.8 s⁻¹, respectively. The apparent K_d (K_d = l/k) derived from this relation for I_to current block by propofol was 22.9 μM, which is close to the IC₅₀ of 20.9 μM obtained from the concentration–response curve determined by the integral of current charge.

Figure 4A illustrates the current and protocol (1-s conditioning pulses from −100 to +30 mV at a holding potential of −80 mV, followed by a 300-ms test pulse to +50 mV after a 30-ms interval at −50 mV) used for determining the steady-state inactivation (availability, I/I_max) of I_to, while figure 4B shows the tail current recorded by a voltage protocol (8-ms voltage steps from −50 mV to potentials between −40 to +50 mV, then to −40 mV) for determining the steady-state activation (g/g_max) of I_to. The normalized variable of I/I_max or g/g_max was calculated and fitted with a Boltzmann function in individual cells to obtain the half potential (V_0.5) of I_to availability or activation and slope factor. Figure 4C displays the mean values of I/I_max fitted by a Boltzmann distribution in the absence or the presence of 30 μM propofol. The V_0.5 of I_to availability was −22.0 ± 1.3 mV in the control and −24.6 ± 2.1 mV in the presence of propofol (n = 8, P = 0.168, paired Student t test). The fractional blockade was plotted as a function of the voltage of the preceding pulse as described by Caballero et al.,³⁶  demonstrating that the blockade remained unchanged at potentials between −100 mV (48.1 ± 3.0%) and −50 mV (48.3 ± 3.1%, n = 8, P > 0.05, ANOVA). At more positive potentials, it tended to increase as the amount of inactivated channels increased, reaching a maximum at −20 mV (57.7 ± 6.4%), but the difference was not statistically significant (n = 8, P > 0.05 vs. blockade at −100 mV, ANOVA).

Figure 4D displays the mean values of g/g_max fitted by a Boltzmann distribution in the absence or the presence of 30 μM propofol. The V_0.5 of I_to activation was 21.9 ± 1.5 mV in the control and 15.9 ± 1.6 mV in the presence of propofol (n = 7, P = 0.055, paired Student t test). No difference was observed for the slope factor of I_to availability or activation before and after the application of propofol.

Time-dependent recovery of I_to from inactivation was determined with a paired-pulse protocol as shown in figure 4E. The recovery curves were fitted by a mono-exponential function in the absence and the presence of 30 μM propofol in individual cells. The recovery time constant of I_to from inactivation was 55.7 ± 6.1 ms in the control and 50.3 ± 7.8 ms in the presence of propofol (n = 6, P = 0.253, paired Student t test). These results indicate that propofol inhibits human atrial I_to without affecting the voltage-dependent kinetics and recovery time constant of the channel.

The use-dependent and rate-dependent effects of I_to by propofol were determined using 20 repetitive 200-ms depolarizations from a holding potential of −50 mV to +50 mV at 0.5, 1, 2, and 3 Hz. There was no significant difference when decreasing I_to by 30 μM propofol at the rate of 0.5, 1, 2, and 3 Hz (fig. 4F), suggesting that the use-dependent or rate-dependent blocking effect is not involved in the I_to inhibition by propofol in human atrial myocytes.

Figure 5A shows the time course of I_Kur recorded in a representative cell with a voltage protocol: as shown in the inset, a 100-ms prepulse to +40 mV to inactivate I_to, followed by a 200-ms test pulse to +50 mV, then to −30 mV every 15 s.^26,27  The current was slowly reduced by 30 μM propofol and reached a steady state in 10 min, and the inhibitory effect was partially reversed by washout. The voltage-dependent I_Kur (fig. 5B) was recorded in a representative cell with a voltage protocol shown in the inset. Propofol at 30 μM substantially inhibited both I_Kur and tail current, and the effect partially recovered on washout. At test potential of +50 mV, I_Kur was reduced by 57.4 ± 5.2% with 30 μM propofol (n = 7, P = 0.002 vs. control, paired Student t test).

I–V relations of I_Kur are illustrated in figure 5C in the absence and the presence of 3, 10, 30, and 100 μM propofol in a total of six myocytes. Propofol inhibited I_Kur in a concentration-dependent manner. The inhibition fraction of the current was more at potentials positive to +20 mV than that at 0 mV with 10 to 100 μM propofol (fig. 5D, n = 6, P < 0.05 or P < 0.01 vs. 0 mV, ANOVA). Figure 5E shows the concentration–response relation of propofol, which was fitted by a Hill equation. The IC₅₀ (at +50 mV) of propofol for inhibiting I_Kur was 35.3 ± 1.9 μM (Hill coefficient 1.1 ± 0.05, n = 6).

I_Kur/Kv1.5 channel inhibitors usually block the open state of the channel and induce an increased inactivation of the current, that is, blocking increase during depolarization.^33,34,37,38  However, propofol inhibited I_Kur without showing the increased inactivation in human atrial myocytes (fig. 5), suggesting that the open-channel blockade may not be involved in the current inhibition. Nevertheless, I_Kur was recorded in human atrial myocytes using short-duration test pulses with a prepulse, which might cover the open-channel blocking property of propofol. We therefore determined the effect of propofol on hKv1.5 channels expressed in HEK 293 cells using longer-duration pulses.

Figure 6A illustrates hKv1.5 current elicited by 5-s voltage pulses between −40 mV and +60 mV from a holding potential of −80 mV with a pulse interval of 20 s. Propofol at 50 μM significantly inhibited hKv1.5 current, and the effect partially recovered on washout. Figure 6B shows the I–V relations of hKv1.5 current measured at the end of the depolarization step before and after the application of 10, 30, 50, and 100 μM propofol. The current was inhibited by propofol in a concentration-dependent manner (n = 9, P < 0.05 or P < 0.01 vs. control at 0 mV to +60 mV, ANOVA). Concentration–response relation for inhibiting the peak current, the “quasi”-steady-state current measured at the end of the voltage pulse, and the current charge crossing the membrane by propofol at 10 to 300 μM were evaluated at +50 mV and fitted with a Hill equation (fig. 6C). The IC₅₀ values were 132.2 ± 6.9 μM (Hill coefficient 1.4 ± 0.1, n = 7), 58.1 ± 2.7 μM (Hill coefficient: 1.7 ± 0.1, n = 7), and 67.0 ± 2.8 μM (Hill coefficient 1.6 ± 0.1, n = 7), respectively.

Although the IC₅₀ values are similar for inhibiting the “quasi”-steady-state current measured at the end of the voltage pulse and the current charge, the IC₅₀ for inhibiting the peak current is greater, suggesting that propofol enhanced the inactivation of hKv1.5 current elicited by 5-s pulses, and open-channel blockade is involved in the current inhibition. The open-channel blocking properties were analyzed as for I_to (fig. 3).^36,39  The current traces (at +50 mV) were fitted by a mono-exponential function with time constants shown in the absence and the presence of 50 μM propofol (fig. 7A). The mean values of the time constant (fig. 7B) were significantly reduced by 50 μM propofol at potentials of +10 to +60 mV (n = 11, P < 0.05 or P < 0.01 vs. control, paired Student t test). Figure 7C illustrates the current traces at +50 mV in the absence and the presence of 30, 50, and 100 μM propofol. The drug-sensitive hKv1.5 current expressed as a proportion of the current in the absence of the drug [(I_control − I_propofol)/I_control]. 1/τ_block as a function of the propofol concentration for data obtained at 30, 50, and 100 μM is shown in figure 7D. The line is a regression fit of the equation, 1/τ_block = k[D] + l (n = 6). The apparent rate constants for association (k) and dissociation (l) were (1.5 ± 0.1) × 10⁴ M⁻¹ s⁻¹ and 0.88 ± 0.09 s⁻¹, respectively. The apparent K_d (K_d = l/k) derived from this relation for hKv1.5 blocking by propofol was 58.7 μM, which is close to the IC₅₀ of 57.3 μM obtained from the concentration–response curve.

To obtain the steady-state inactivation of hKv1.5 channels, a protocol with 10-s conditioning pulses from −80 to +50 mV and then a 1-s test pulse to +50 mV was used with 20-s pulse interval. Figure 7E shows the mean variables of I/I_max of hKv1.5 current in the absence and the presence of 50 μM propofol. The variables were fitted with a Boltzmann distribution in individual cells to obtain the V_0.5 of hKv1.5 inactivation. The inactivation V_0.5 was −9.5 ± 0.7 mV in control and −8.7 ± 0.9 mV in 50 μM propofol (n = 9, P = 0.433, paired Student t test). Fractional block by propofol showed that current blockade did not change significantly even in the voltage range of channel inactivation.

Recovery of hKv1.5 current from inactivation was determined with a paired-pulse protocol (a 5,000-ms step to +50 mV from a holding potential of −80 mV, followed by a 300-ms step to +50 mV with variable P1–P2 interval between 2 and 9,000 ms, pulse interval of 25 s). Both of the recovery curves in the absence and the presence of 50 μM propofol were well fitted with a bi-exponential function (fig. 7F). Propofol slightly slowed the recovery of hKv1.5 current from inactivation, but the difference was not significant (fast time constant: 285 ± 15 ms in the control, 469 ± 86 ms in the presence of propofol, P = 0.125, paired Student t test; slow time constant: 3,956 ± 399 ms in the control, 4,075 ± 437 ms in the presence of propofol, n = 8, P = 0.873, paired Student t test).

It has been demonstrated that both I_Kr and I_Ks are present in human cardiac myocytes and play an important role in cardiac repolarization in human heart.^24,40  It is interesting to investigate whether the inhibition of human cardiac I_Kr and/or I_Ks contributes to the antiatrial tachycardia/fibrillation of propofol. It has been difficult to record these two currents in cardiac myocytes from the chunk dissociation method with a small atrial specimen from patients undergoing coronary bypass surgery. Therefore, HEK 293 cells stably expressing hERG (coding for I_Kr)³⁰  or hKCNQ1/hKCNE1 (coding for I_Ks)²⁹  were used here to determine the effects of propofol on these two types of currents.

Figure 8A shows the time course of hERG tail current recorded with a voltage step as shown in the inset in the absence and the presence of 30 and 100 μM propofol. The current was inhibited by 30 and 100 μM propofol, and the inhibition was reversed by washout. Similar results were observed in voltage-dependent hERG current. Both hERG tail current and step current were clearly decreased by propofol (fig. 8B). Figure 8, C and D, illustrates the I–V relations of mean values of hERG tail and step current in the absence and the presence of 30, 100, and 300 μM propofol. Propofol at 30 μM inhibited hERG tail and step current from +20 mV to +60 mV (n = 10, P < 0.05 vs. control, ANOVA), while at 100 and 300 μM, propofol significantly decreased the current from 0 mV from +60 mV (P < 0.05 or P < 0.01 vs. control, ANOVA). The concentration–response curves (fig. 8E) of propofol for inhibiting hERG tail current and step current at +40 mV were fitted with a Hill equation. The IC₅₀ of propofol was 84.4 ± 15.5 μM for inhibiting hERG tail current and 73.2 ± 16.6 μM (n = 7) for inhibiting hERG step current.

The steady-state activation (g/g_max) of hERG channels was determined by normalized tail current in the absence and the presence of 30 and 100 μM propofol (fig. 8F). The activation curves were fitted with a Boltzmann function. The V_0.5 of hERG channel activation was 4.1 ± 1.6 mV in control, 0.5 ± 2.1 mV in 30 μM propofol (n = 7, P > 0.05 vs. control), and −10.1 ± 2.2 mV in 100 μM propofol (n = 7, P < 0.05 vs. control, one-way ANOVA followed by the Newman–Keuls test). The activation of hERG channels was significantly shifted to the negative potential by propofol.

Figure 9A shows the time course of human cardiac hKCNQ1/hKCNE1 channels expressed in HEK 293 cells in a typical experiment with the voltage step as shown in the inset. Propofol at 30 and 100 μM significantly inhibited hKCNQ1/hKCNE1 step current, and the inhibition was partially reversed by washout. The voltage-dependent hKCNQ1/hKCNE1 current was also suppressed by propofol (fig. 9B). Figure 9C illustrates the I–V relations of mean values of hKCNQ1/hKCNE1 current density in the absence and the presence of 10, 30, and 100 μM propofol. Propofol significantly inhibited hKCNQ1/hKCNE1 current at test potential of 0 to +60 mV (n = 6, P < 0.05 or P < 0.01 vs. control, ANOVA). The concentration–response curves (fig. 9D) of propofol for decreasing hKCNQ1/hKCNE1 current at +40 mV were fitted with a Hill equation. The IC₅₀ of propofol was 32.4 ± 6.0 μM (n = 6) for inhibiting hKCNQ1/hKCNE1 channels.

The steady-state activation (g/g_max) of hKCNQ1/hKCNE1 channels was determined by normalized tail current in the absence and the presence of 30 μM propofol (fig. 9E). The activation curves were fitted with the Boltzmann function. The V_0.5 of hKCNQ1/hKCNE1 current activation was 16.3 ± 4.5 mV in control and 20.1 ± 4.1 mV in 30 μM propofol (n = 7, P = 0.088 vs. control, paired Student t test). The activation conductance of hKCNQ1/hKCNE1 channels was not significantly affected by propofol.

The effect of propofol on cardiac action potential was determined in human atrial myocytes at 36°C. Figure 10A displays the action potential traces recorded in a representative cell at 2 Hz. Propofol at 30 μM slightly prolonged action potential duration. Action potential duration at 90% repolarization was significantly increased by the application of propofol (fig. 10B, n = 6, P = 0.024 vs. control, paired Student t test). The measured dV/dT of action potential phase 0 was reduced from 78.2 ± 5.7 V/s to 72.7 ± 5.9 V/s (n = 6, P = 0.034 vs. control, paired Student t test), which may be related to the inhibition of I_Na as reported previously.^13,14  However, resting membrane potential was not affected by propofol (−73.0 ± 1.7 mV vs. −72.4 ± 1.1 mV, n = 6, P = 0.528, paired Student t test).

The current study demonstrates that the intravenous anesthetic propofol blocks human atrial native I_to, I_Kur, hERG, and hKCNQ1/hKCNE1 channels stably expressed in HEK 293 cells with concentrations ranging from 3 to 100 μM and it slightly prolongs human atrial action potential duration.

Previous reports demonstrated that propofol inhibits several cardiac ion channels.¹  Propofol at 1 to 100 μM decreases the cardiac depolarization currents including I_Ca.L in cardiac ventricular myocytes from rat¹²  and guinea pig,¹¹  rabbit,¹⁴  and human atrial myocytes,¹⁰ I_Na in rat and rabbit ventricular myocytes,^13,14  and pacemaker current (i.e., hyperpolarization-activated cyclic nucleotide–regulated channels) expressed in HEK 293 cells.^16,41  These effects may count for the ionic mechanisms of sinus arrest, bradycardia, and/or atrioventricular blockade observed in patients with propofol as an intravenous anesthetic.^4,5,9 

For cardiac repolarization currents, the reports for cardiac inward rectifier potassium current (I_K1) are controversial. One study reported that propofol at 3 to 30 μM decreased I_K1 in rabbit ventricular myocytes,¹⁴  whereas other reports demonstrated that propofol (28 to 60 μM) had no effect on I_K1 in ventricular myocytes from guinea pig⁴²  and canine¹⁷  hearts. Our results support the notion that propofol has no significant effect on cardiac I_K1 because propofol at 30 μM only slightly inhibited human cardiac Kir2.1 channels expressed in HEK 293 cells without statistical significance (Supplemental Digital Content 1, https://links.lww.com/ALN/B107, figure 1, n = 8, inhibited by approximately 3.3% at −120 mV, P = 0.124 vs. control, paired Student t test). This may explain why no change is observed in the resting membrane potential of human atrial myocytes with application of propofol.

The transient outward potassium current I_to plays a role in the early rapid repolarization of cardiac action potential in mammalian heart, including that of humans. Propofol (3 to 60 μM) inhibited I_to in ventricular myocytes from rat,⁴³  rabbit,¹⁴  and canine¹⁷  hearts. The current study demonstrated the novel information that propofol decreased I_to by binding to the open channels in human atrial myocytes. However, propofol showed no use- or rate-dependent blockade of I_to, which is similar to the effect of propafenone³⁵  and allitridi³³  on cardiac I_to.

The ultrarapid delayed rectifier potassium current I_Kur is present in human atria but not in the ventricles of the human heart.²⁴  It is believed that I_Kur is a target for developing atrial-selective antiatrial fibrillation drugs.^24,44,45  No information is available in the literature regarding the effect of propofol on I_Kur/Kv1.5. In this study, we provide the novel information that propofol inhibits human atrial I_Kur with an IC₅₀ of 35.3 μM. We found that the blocking fraction of propofol for I_Kur was slightly but significantly higher at potentials positive to +20 mV than that at 0 mV, which suggests a stronger inhibition of the current at positive potential of action potential. Because I_to and I_Kur play a crucial role in the repolarization of human atrial myocytes,^25,32,45,46  blockade of I_to and I_Kur would significantly prolong atrial action potential duration and therefore would exert the antiatrial arrhythmic effect.^25,45  Therefore, the antiatrial tachycardia and antiatrial fibrillation observed in patients with propofol^21–23  is likely related to the inhibition of I_to and I_Kur.

Although propofol did not show the development blockade of I_Kur using short depolarization pulses with a prepulse in human atrial myocytes, the development blockade was observed in HEK 293 cells expressing hKv1.5 channels with a longer duration of depolarization pulses without prepulse. Propofol enhanced the inactivation of hKv1.5 current elicited by 5-s pulses, suggesting that propofol inhibits hKv1.5 by preferentially binding to the open channels. It is interesting that propofol is more sensitive to inhibit native I_Kur than hKv1.5 expressed in HEK 293 cells.

The delayed rectifier potassium currents I_Kr and I_Ks are also important repolarization currents in human heart.^24,40  Blockade of I_Kr may be antiarrhythmic or proarrhythmic.^47–49  The proarrhythmic effect of I_Kr/hERG blockade is mainly related to induction of prolonged QT interval of electrocardiography or Torsades de Pointes.^47,48  An earlier report did not find the inhibition of I_Kr in guinea pig cardiac myocytes,⁵⁰  while in Xenopus oocytes expressing hERG channels, propofol significantly inhibited hERG current at a high concentration of 100 μM.⁵¹  The current study showed that significant inhibition of hERG channels expressed in HEK 293 cells was observed at 30 μM. The IC₅₀ of propofol was 73.2 μM for hERG step current and 84.4 μM for hERG tail current, which is much greater than plasma concentrations in patients.^52,53  Experimental study did not find the prolonged QT interval in guinea pigs.⁵¹  Clinical reports demonstrated that QTc interval was unaffected^54,55  or shortened by propofol.⁵⁶  Interestingly, propofol reversed the QT prolongation induced by sevoflurane²⁰  and reduced the QT dispersion.⁵⁷  These clinical reports suggest that the slight hERG/I_Kr inhibition would not induce a prolonged QT interval.

In Xenopus oocytes expressing the cardiac I_Ks gene IsK, propofol showed less sensitivity for the current inhibition with an estimated IC₅₀ value of 250 μM,⁵⁸  while in guinea pig cardiac myocytes propofol at 300 μM fully inhibited I_Ks.⁵⁰  However, a recent report demonstrated that propofol inhibited I_Ks in guinea pig ventricular myocytes with an IC₅₀ of 23 μM,¹¹  which is close to that observed in the current study. We found that propofol inhibited human cardiac hKCNQ1/hKCNE1 channels expressed in HEK 293 cells with an IC₅₀ of 32.4 μM. Reduction of human cardiac I_Ks may also contribute to the inhibition of supraventricular tachycardia/fibrillation in patients with propofol.^21–23 

Pharmacokinetic studies have shown that an intravenous dose of propofol for anesthesia may reach a peak level at 44 μM in blood plasma and generally 10 to 20 μM for anesthetic maintenance.^52,53  In the current study, we found that significant inhibition of I_to, I_Kur, and hKCNQ1/hKCNE1 current by propofol was observed at 3 to 10 μM, which is in the range of anesthetic maintenance. Therefore, the conversion of supraventricular tachycardia/fibrillation to sinus rhythm in patients with propofol^21–23  may be related to its inhibition of I_to, I_Kur, and I_Ks.

It should be noted that the current study demonstrated that significant inhibition of I_to, I_Kur, and hKCNQ1/hKCNE1 current was observed with 30 μM propofol. The effect was supposed to significantly prolong action potential duration. However, we found that action potential duration was slightly prolonged by propofol in human atrial myocytes. It is well recognized that for the effect of a drug on action potential duration, I_Ca.L blockade has a counter effect with potassium current inhibition. The slight increase in action potential duration with propofol is most likely due to the I_Ca.L blockade observed previously in human atrial myocytes¹⁰  and other species.¹²  This phenomenon is similar to that of Ca²⁺ channel blocker diltiazem on potassium currents and cardiac action potential.³⁶  The conversion of supraventricular tachycardia/fibrillation to sinus rhythm in patients with propofol^21–23  may be related to inhibiting more for potassium currents than for I_Ca.L.

The limitation of the current study was that we were unable to record I_Kr and I_Ks in human atrial myocytes isolated from small atrial specimens with the tissue chunk method. It is unclear whether the channels are digested by enzymes. If so, it may lead to an alteration of the action potential waveform and action potential duration prolongation and induce an underestimation of propofol effect on the action potential. The observed slight prolongation of action potential duration by propofol may be related partially to the loss of these two channels.

In summary, the current study reported for the first time that propofol inhibited human cardiac atrial repolarization potassium currents including human atrial I_to and I_Kur, and slightly prolonged human atrial action potential duration. Propofol also inhibited hERG and hKCNQ1/hKCNE1 channels expressed in HEK 293 cells. Reduction of I_to, I_Kur, and I_Ks likely contributes to the suppression of supraventricular tachycardia and atrial fibrillation observed in patients given propofol.

The work was supported by a grant from the Important National Science and Technology Specific Projects (grant no. 2009ZX09301-014), Beijing, China.

The authors declare no competing interests.