Blockade of Myocardial ATP-sensitive Potassium Channels by Ketamine 

Our investigation conformed with the Guiding Principles in the Care and Use of Animals as approved by the Council of the American Physiology Society.

Cell Isolation. Single rat ventricular myocytes were prepared by enzymatic digestion, as described previously. [19] Male Sprague-Dawley rats (weight, 250–300 g) were stunned with a blow on the neck and killed by spinal cord dislocation. After opening the chest cavity, hearts were quickly excised and immersed in Krebs-Henseleit (KH) buffer solution (pH 7.35). Excised hearts were perfused in a retrograde manner via the aorta in a Langendorff apparatus with KH solution saturated with a 95% oxygen and 5% carbon dioxide gas mixture for 5 min to clear visible blood. Isolated hearts were perfused with Ca²+-free KH solution until they stopped beating, and then they were perfused with Ca sup 2+-free KH solution containing 0.075% collagenase (CLS 2; Worthington Biomedical Co., Freehold, NJ) for about 30 min. After enzymatic digestion, ventricular muscle was removed mainly from the lower left ventricular wall near the ventricular septum. The muscle was cut into small pieces and then mechanically dissociated into single cells by gentle agitation in Ca²+-free KH solution containing 1% bovine serum albumin. The cells were allowed to settle at the bottom of the container by gravity and the supernatant of the cell suspension was removed by replacing it with Ca²+-free KH solution containing 1% bovine serum albumin. All cells used for the experiments were rod-shaped and had clear striations. Krebs-Henseleit solution contained 118 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO sub 4, 1.2 mM KH²PO⁴, 10 mM HEPES, 25 mM NaHCO³, 10 mM pyruvate, 11 mM dextrose, and 1 mM CaCl².

Current Recording and Data Analysis. Gigaseals were formed with Sylgard-coated pipettes (borosilicate, Kimax-51, Kimble Glass Inc., Owens, IL) with 4 or 5 M Ohms resistance. Single-channel currents were recorded using the inside-out and cell-attached configurations of the patch-clamp method described by Hamil et al. [20] Channel currents were recorded with an Axopatch 1D patch-clamp amplifier (Axon Instruments, Foster, CA) and stored on video tapes via pulse code modulator (PCM-501ES; Sony, Tokyo, Japan) for subsequent computer analysis. Electric signals were converted into digits and stored on a hard disk of a personal computer (486 DX4; Hyundai, Seoul, Korea) at a sampling rate of 330 kHz with an analog-to-digital converter (Digidata 1200; Axon Instruments). pClamp software (Version 5.7.2; Axon Instruments) was used for data acquisition and analysis. The 50% threshold method was used to detect events. The open probability (P^o) was calculated using an equation derived from Spruce et al. [21]:Equation 1where t^jis the time spent at current levels corresponding to j = 0,1,2,…,N channels in the open state; T^dis the duration of the recording; and N is the number of the channels active in the patch. Recordings of 30–60 s were analyzed to determine P sub o. The channel activity was expressed as N [center dot] P^o. Changes of channel activity in the presence of drugs were calculated as the relative ratio of the channel activity between the values obtained before and after drug treatment. All experiments were done at 22 +/- 2 [degree sign] Celsius. The standard bath and pipette solutions contained 140 mM KCl, 2 mM MgCl², 5 mM EGTA, and 10 mM HEPES (pH 7.2).

The rat hearts were rapidly excised and transferred to a dissection bath filled with Tyrode's solution oxygenated with a 97% oxygen and 3% carbon dioxide mixture. Papillary muscles were carefully dissected free from the right ventricular wall, mounted horizontally in a narrow channel of a tissue chamber, and continuously superfused with oxygenated Tyrode's solution at 36 [degree sign] Celsius or 37 [degree sign] Celsius. The size of dissected papillary muscles ranged from 0.5–1 mm in width and was 2 or 3 mm in length. The mural end of the muscle was fixed by an insect pin to the bottom of the chamber coated with Sylgard. The portion of the muscle adjacent to the insect pin was pressed against the floor by stimulating electrodes, which were used to elicit action potentials and contractions. The action potentials were elicited by stimulating the cardiac cells with square pulses (1 Hz, lasting 1 ms, 20–30% above threshold voltage) by a stimulator via a stimulus isolation unit (WPI, Sarasota, FL). Action potentials were recorded with a 3-M KCl-filled microelectrode (10–20 M Ohms) connected to an amplifier (KS-700; WPI) and were displayed on an oscilloscope (dual-beam storage 5113; Tektronix, Beaverton, OR). Tracings on the oscilloscope screen were photographed using 35-mm film and also recorded on a chart recorder (RS 3400; Gould, Cleveland, OH). Rat papillary muscle was superfused with Tyrode's solution at a constant rate (5 ml/min). Tyrode's solution contained 137 mM NaCl, 5.4 mM KCl, 1.05 mM MgCl², 0.45 mM NaH²PO⁴, 11.9 mM NaHCO sub 3, 1.8 mM CaCl², and 5 mM dextrose.

The following compounds were used: ketamine hydrochloride (Yuhan, Seoul, Korea), 2.4-dinitrophenol (Sigma Chemical Co., St. Louis, MO), pinacidil monohydrate (Leo Pharmaceutical, Copenhagen, Denmark), glibenclamide (RBI, Natick, MA), and ATP (Sigma Chemical Co.).

All data were presented as means +/- SD. Repeated-measures analysis of variance, followed by Scheffe's multiple-range test, was applied to identify significant differences among the effects observed with different concentrations of ketamine. Paired statistical comparisons were made using paired t tests. In all comparisons, P < 0.05 was considered significant.

To evaluate the inhibitory effect of ketamine on K^ATPchannels, the effect of ketamine on K^ATPchannel activity in single rat ventricular myocytes was examined using the inside-out and cell-attached configurations of the patch-clamp techniques. Activation of the K^ATPchannels observed in inside-out patches was inhibited by 1 mM ATP or 10 micro Meter glibenclamide and showed a unitary conductance of 65–70 pS. These channel properties were consistent with K^ATPchannel currents recorded previously. [22]

When inside-out patches were formed in an ATP-free bath solution, maximal K^ATPchannel activities were observed. Ketamine inhibited K^ATPchannel activity at a concentration as low as 1 micro Meter and produced a concentration-dependent inhibition of K^ATPchannel activity. K^ATPchannel activity was almost completely inhibited by 10 sup -3 M ketamine. Washout of ketamine resulted in more than 80% recovery of the channel activity, and glibenclamide completely blocked the channel activity (Figure 1(A)). On formation of inside-out patches in the ATP-free bath solution, K^ATPchannel activity gradually decreased with time. This phenomenon is known as “run-down.”[4] Therefore, data obtained from such experiments, as shown in Figure 1(A), may not represent an accurate ketamine-K^ATPchannel activity relation. To minimize the time-dependent decrease of the channel activity (i.e., the run-down phenomenon) and to obtain an accurate ketamine-K^ATPchannel activity relation, we determined the effect of a single concentration of ketamine from each inside-out patch within 3 min of patch excision. Under these conditions, the average percent recovery of K^ATPchannel activities after ketamine washout was 94 +/- 5% of N [center dot] P^obefore ketamine treatment (n = 96). The plot of relative channel activities as a function of ketamine concentration was fitted to the Hill equation using the least-squares method [23](Figure 1(B)):Equation 2where y is the relative N [center dot] P^o, [D] is the concentration of ketamine, Kⁱis the ketamine concentration at which half-maximal inhibition of the channels occurs, and H is the Hill coefficient. The Kⁱand Hill coefficient for ketamine were 62.9 micro Meter and 0.54, respectively.

To observe the effect of ketamine on K^ATPchannel activity in intact ventricular myocytes, we next examined the effect of ketamine on pinacidil- or 2,4-dinitrophenol (DNP)-induced openings of K^ATPchannels in cell-attached patches (Figure 2and Figure 3).

No channel openings were observed in the absence of pinacidil or DNP in cell-attached patches. Pinacidil (5 x 10 sup -4 M) increased K sub ATP channel activity, which was inhibited by ketamine in a concentration-dependent manner (Figure 2). Removal of ketamine from the bath solution resulted in more than 80% recovery of K^ATPchannel activity. An inhibitor of mitochondrial ATP synthesis, DNP (5 x 10 sup -5 M), opened K^ATPchannels in cell-attached patches. This opening of K sub ATP channels was also inhibited by ketamine in a concentration-dependent manner (Figure 3). These results indicate that ketamine inhibits K^ATPchannel activity in intact myocardial cells.

(Figure 4) shows the current-voltage (I-V) relation obtained in four inside-out membrane patches before and after adding 10 sup -4 M ketamine. The I-V curves before and after ketamine treatment at negative membrane potential displayed a linear relation with conductances of 69.6 +/- 0.7 pS and 70.1 +/- 0.6 pS, respectively. The positive membrane potentials of 40 mV resulted in I-V curves with inward rectification. Application of ketamine to the intracellular side of membrane patch did not affect the conductance of single-channel currents of K^ATPchannels.

To determine whether ketamine blocks the K^ATPchannel, we studied its effects on the action potential characteristics in rat papillary muscle. Action potentials were recorded from isolated rat papillary muscle with 3 M KCl-filled conventional microelectrodes. From the same papillary muscle preparations stimulated at 1 Hz, the resting membrane potential was -79.1 +/- 2.4 mV (n = 6), and the action potential duration at 100% repolarization was 142.4 +/- 17.5 ms. The maximum effect of ketamine on action potentials was observed about 5 min after exposure to the drug, and the ketamine effect was readily reversible. The action potential duration at 90% repolarization (APD⁹⁰) was increased by 10 sup -5 M, 3 x 10 sup -5 M, and 10 sup -4 M of ketamine from 65 +/- 11 ms to 66 +/- 11 ms, 71 +/- 12 ms, and 89 +/- 14 ms, respectively. Glibenclamide (10 sup -5 M) had no effect on normal action potentials. The effects of ketamine on action potential duration were also examined using pinacidil. Papillary muscles were superfused with a solution containing 3 x 10 sup -4 M pinacidil, followed by 10 sup -5 M, 3 x 10 sup -5 M, and 10 sup -4 M ketamine or 10 sup -5 M glibenclamide. Pinacidil shortened the APD⁹⁰from 55 +/- 17 ms to 28 +/- 4 ms (Figure 5(A)). The pinacidil-induced APD⁹⁰was increased by 10 sup -5 M, 3 x 10 sup -5 M, and 10 sup -4 M ketamine from 28 +/- 4 ms to 36 +/- 5 ms, 40 +/- 5 ms, and 55 +/- 9 ms, respectively (Figure 5(B)). The pinacidil-induced APD shortening was completely reversed by glibenclamide. Resting membrane potential was slightly depolarized at 10 sup -4 M ketamine from 79.5 +/- 3.1 mV to 76.8 +/- 2.7 mV (P < 0.01).

Our study examined the effects of ketamine on K^ATPchannel activity in the rat myocardium. The principal findings are that ketamine inhibits K^ATPchannel activity and increases action potential duration.

The K^ATPchannel is only activated at low ATP concentrations, with half-maximal inhibition of channel opening at 20–100 micro Meter, [1,24] whereas the normal intracellular ATP concentration is much higher (5–10 mM). However, the measured [ATP]ⁱat the time of action potential shortening (or in metabolic blockade of ischemia) is in the millimolar range, which is close to normal level. [25] The role of the K^ATPchannel in response to myocardial ischemia, therefore, has been questioned. In contrast, several recent studies have shown that because of the high density of cardiac K^ATPchannels, only a small increase in the open state probability (less than 1%) was sufficient to shorten action potential duration during ischemia. [26,27] Production of ATP is expected to decrease during hypoxic or ischemic conditions, and the levels of other metabolites (e.g., ADP, lactate, and protons) that affect channel activities may increase. The dose-response curve resulting from the cellular metabolite that affects the channel activities shifts to the right, causing the amount of ATP required to hold all of the channels closed to increase. From these considerations, K^ATPchannels may play a role in shortening action potential duration, even if [ATP]ⁱdoes not fall substantially below the normal intracellular level. In our study, we used DNP, an inhibitor of oxidative phosphorylation in mitochondria, to induce the K^ATPchannel current. The lag periods after DNP exposure, but before the induction of K^ATPchannel current, ranged from 3 to 20 min in ventricular myocytes. This latency was considered to reflect the time for the depletion of endogenous energy sources before the intracellular ATP levels decreased. However, the opening of K^ATPchannel by DNP may be different from the activation of K^ATPchannel by ischemia or hypoxia.

The K^ATPchannel underlies the increase in potassium permeability and the consequent reduction in action potential duration during hypoxia or ischemia. The effects of drugs responsible for opening potassium channels indicate that the basic electromechanical phenomena during ischemia (e.g., action potential shortening, loss of contractility, and cellular inexcitability) can be replicated by the activation of K^ATPchannels. [28] The K^ATPchannel antagonist glibenclamide either attenuates or abolishes hypoxia-induced action potential shortening. [10] In our findings, pinacidil-induced reduction of action potential duration was abolished by ketamine in a dose-dependent manner, and the pinacidil-induced shortening of APD was completely reversed by glibenclamide. These results indicate that the target for ketamine-induced action potential prolongation is the same K^ATPchannel; that is, a target for glibenclamide.

During ischemia or hypoxia, K^ATPchannels open to induce several protective responses in the heart. With reperfusion or reoxygenation, oxygen free radicals are generated, which could trigger a chain of damaging chemical reactions, resulting in “reperfusion injury.” The free radical-induced injury can be attenuated by potassium channel openers acting on K^ATPchannels. [29] In addition, the K^ATPchannels may also be involved in “ischemic preconditioning,” a cardioprotective phenomenon wherein brief ischemia increases the tolerance of myocytes to subsequent prolonged ischemic insults. [30] During ischemia, opening of K^ATPchannels benefits the heart, possibly by reducing Ca²+ influx through voltage-operated calcium channels, thus slowing ATP depletion and decreasing calcium-induced toxicity. Gross and Auchampach [31] reported that glibenclamide, a K^ATPchannel blocker, abolished the effects of preconditioning. In our study, ketamine inhibited K^ATPchannel activities in inside-out and cell-attached patches. These results suggest that inhibition of the K^ATPchannel by ketamine may aggravate ischemia-reperfusion injury and attenuate ischemic preconditioning.

Extracellular potassium concentration increases rapidly during myocardial ischemia because of increased potassium efflux. Weiss et al. [32] reported that potassium efflux during metabolic inhibition was most likely to arise from the increased membrane potassium conductance possibly due to the activation of K^ATPchannels. Increased extracellular potassium promotes depolarization of tissue in the vicinity of the ischemic region. The flow of this injury current (i.e., the current flowing between ischemic and normal cells) has been implicated as a potential cause for the initiation of premature ventricular beats. [33] Although K^ATPchannel activation is thought to be highly protective in the ischemic myocardium, activation of these channels may, during certain circumstances, prove to be arrhythmogenic. Previous studies have shown that K^ATPchannel antagonists can prevent ventricular arrhythmias in vitro and in vivo, whereas potassium channel agonists reduce the time for the onset of fibrillation. [34,35] Like the action of other K^ATPchannel antagonists, ketamine may protect against arrhythmia formation during myocardial ischemia.

The effects of ketamine on the other potassium channels have been reported. Kulkarni et al. [36] reported that ketamine reduced the maximal conductance of Kv2.1 in the mammalian brain. They suggested that the inhibition of potassium currents leads to increased neuronal excitability. Endou et al. [37] reported that ketamine inhibited the Ca sup 2+-insensitive transient outward potassium current, resulting in prolonged action potential duration in rat hearts. In the present study, ketamine increased action potential duration in rat papillary muscle. Ketamine increased action potential duration more remarkably in the presence of pinacidil, which markedly shortens the action potential duration in rat papillary muscle. These results indicate that ketamine inhibits transient outward potassium channels and K^ATPchannels in rat ventricular myocytes.

In humans, the peak plasma concentration of ketamine is 3–60 micro Meter during anesthesia with an intravenous dose of 2 mg/kg. [38] Although 45–50% of ketamine is bound to plasma proteins, principally alpha 1-acid glycoprotein, these concentrations may be sufficient to inhibit the K^ATPchannel activity and potentially reverse anti-ischemic effects mediated by this channel. However, it is difficult to extrapolate the present results to the human heart because of the in vitro model and possible species differences.

Ketamine inhibits the opening of K^ATPchannels. Therefore, ketamine may inhibit the cardioprotective effects of K^ATPchannels, promote ischemic injury, and produce irreversible damage. Further studies are needed to characterize the clinical significance of these findings in the human heart.

The authors thank Ki-Chan Ha and Jung-Ar Park for technical assistance, and Amy Terhune and Dr. Kwang-Jin Kim for critical reviews.