Differential Effects of Etomidate and Midazolam on Vascular Adenosine Triphosphate–sensitive Potassium Channels: Isometric Tension and Patch Clamp Studies

The study was approved by the Animal Investigation Committee of Tokushima University (Tokushima, Japan) and was conducted according to the animal use guidelines of the American Physiologic Society (Bethesda, Maryland).

Isometric tension experiments were performed on 2.5-mm thoracic aortic rings obtained from male Wistar rats (250–300 g) anesthetized with ether. Each ring was placed in modified Krebs-Ringer's bicarbonate solution (control solution) of the following composition: 118.3 mm NaCl, 4.7 mm KCl, 2.5 mm CaCl², 1.2 mm MgSO⁴, 1.2 mm KH²PO⁴, 25.0 mm NaHCO³, 0.026 mm calcium EDTA, and 11.1 mm glucose. The endothelium of all rings was removed mechanically, because vasorelaxation in response to levcromakalim is augmented in the presence of functional endothelium. The removal was confirmed by the absence of relaxation in response to acetylcholine (10⁻⁵m). Several rings cut from the same artery were studied in parallel, with each ring connected to an isometric force transducer (Micro Easy Magnus UC-2A; Kishimoto Medical Instruments Co., Ltd., Kyoto, Japan) and suspended in an organ chamber filled with 2 ml control solution (37°C, pH 7.4) bubbled with 95% oxygen and 5% carbon dioxide. Arteries were gradually stretched to the optimal point of the length–tension curve, as determined by contraction with phenylephrine (3 × 10⁻⁷m). In most of the arteries studied, optimal tension was achieved at approximately 1.0 g. Preparations were equilibrated for 90 min. During submaximal contractions produced in response to phenylephrine (3 × 10⁻⁷m), relaxation after administration of levcromakalim (10⁻⁸to 10⁻⁵m) and a nonselective vasodilator, papaverine (10⁻⁷to 10⁻⁴m), was observed. Etomidate (10⁻⁶, 10⁻⁵, 10⁻⁴m), midazolam (10⁻⁸, 10⁻⁷, 10⁻⁶m), or glibenclamide (10⁻⁸to 10⁻⁵m) was applied to the preparations 15 min before the addition of phenylephrine. Vasorelaxation was expressed as a percentage of the maximum relaxation in response to papaverine (3 × 10⁻⁴m), which was added at the end of the experiments to produce maximum relaxation (100%) of the arteries. Concentration–response curves were obtained in a cumulative manner. Only one concentration–response curve was made for each ring.

A cell line of A10 vascular smooth muscle cells, derived from the thoracic aorta of fetal rats, was obtained from the American Type Culture Collection (Manassas, VA). The cells were incubated in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (Life Technologies; Rockville, MD), 3.7 mg/ml NaHCO³, and 100 μg/ml gentamicin at 37°C in a humidified atmosphere of 95% air and 5% CO². The medium was changed initially at 48 h and then every 2–3 days. When the cells had formed a confluent monolayer after 7–9 days, they were made quiescent by incubation in serum-free medium for 24 h. They were then harvested by the addition of 0.05% trypsin and 0.1% fetal bovine serum. Passages 5–12 were used for experimental purposes. Cultured A10 cells were stimulated with etomidate (10⁻⁶to 10⁻⁴m) or midazolam (10⁻⁸to 10⁻⁴m).

Details of the recombinant experimental design were similar to those of our previous studies.8–10In brief, K^ATP-deficient COS-7 cells were transiently cotransfected with two K^ATPchannel subunits, sulfonylurea receptor (SUR) and inwardly rectifying K⁺channel (Kir) subunits, which comprise specific tissue-type K^ATPchannels, with Lipofectamine, and with Opti-MEM1 reagents. A truncated form of human Kir6.2 lacking the last 36 amino acids at the C-terminus (Kir6.2ΔC36) was obtained by polymerase chain reaction amplification. Mutagenesis complementary DNA (cDNA) of Kir6.2ΔC36 was performed with a site-directed mutagenesis kit (Invitrogen Corp., Carlsbad, CA). All DNA products were sequenced with a BigDye terminator cycle sequencing kit and an ABI PRISM 377 DNA sequencer (Applied Biosystems, Foster City, CA) to confirm the sequence. For electrophysiologic recordings, COS-7 cells were plated onto dishes with glass coverslips, and Kir and SUR subunits were cotransfected with green fluorescent protein cDNA (pEGFP) as a reporter gene. After transfection, cells were cultured for 48–72 h before being subjected to electrophysiologic recordings.

Cell-attached and inside-out patch configurations were used to record the current through single channels via  a patch clamp amplifier, as described previously.16For cell-attached configurations, the bath solution consisted of 140 mm KCl, 10 mm HEPES, 5.5 mm dextrose, and 1 mm EGTA. The pipette solution contained 140 mm KCl, 10 mm HEPES, and 5.5 mm dextrose. For inside-out configurations, the bath solution (intracellular solution) consisted of 140 mm KCl, 10 mm HEPES, 5.5 mm dextrose, 1 mm MgCl², 1 mm EGTA, 0.5 mm magnesium adenosine diphosphate, and 0.5 mm magnesium adenosine triphosphate. The pipette solution (extracellular medium) was of the same composition as that used in cell-attached experiments. The pH of all solutions was adjusted to 7.3–7.4 with potassium hydroxide. Recordings were made at 36 ± 0.5°C. Patch pipettes were pulled with an electrode puller (PP-830; Narishige, Tokyo, Japan). The resistance of pipettes filled with internal solution and immersed in Tyrode solution was 5–7 MΩ. The sampling frequency of the single-channel data was 5 KHz with a low-pass filter (1 KHz).

Channel currents were recorded with a patch clamp amplifier (CEZ 2200; Nihon Kohden, Tokyo, Japan) and stored on a personal computer (Aptiva; International Business Machine Corp., Armonk, NY) with an analog-to-digital converter (DigiData 1200; Axon Instruments, Foster City, CA). pClamp version 7 software (Axon Instruments) was used for data acquisition and analysis. The open probability (P^o) was determined from current amplitude histograms and was calculated as follows:

where 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 channels active in the patch. Recordings of 2–3 min were analyzed to determine P^o. Channel activity was expressed as NP^o. Changes in channel activity in the presence of drugs were calculated as the relative channel activity, i.e. , the ratio between values obtained before and after drug treatment. When the concentration-dependent effects of etomidate and midazolam were studied, increasing concentrations of these drugs were injected into the cell bath with a glass syringe (total volume injected was approximately 10–20 μl).

Drug concentrations needed to induce half-maximum inhibition of the channels (IC⁵⁰) and the Hill coefficient were calculated as follows:

where y is the relative NP^o, [D] is the concentration of drug, Kⁱis the IC⁵⁰, and H is the Hill coefficient.

The following pharmacologic agents were used: midazolam, dimethyl sulfoxide, glibenclamide, papaverine hydrochloride, phenylephrine, pinacidil (Sigma-Aldrich Japan, Tokyo, Japan), levcromakalim, and etomidate (Tocris, Ellisville, MO). Drugs were dissolved in distilled water such that volumes of less than 15 μl were added to the organ chamber. Stock solutions of levcromakalim, glibenclamide, and pinacidil were prepared in dimethyl sulfoxide. The final concentration of dimethyl sulfoxide in the bath solution never exceeded 0.01%; dimethyl sulfoxide at a twofold higher concentration was shown not to affect K^ATPchannel currents. The concentrations of drugs are expressed as final molar (m) concentrations.

All data are presented as mean ± SD. Differences between data sets were evaluated either by repeated-measures one-way analysis of variance followed by the Scheffé F test or by Student t  test with Welch correction. P < 0.05 was considered statistically significant.

First, we examined the effects of various concentrations of selective K^ATPchannel antagonist glibenclamide on relaxation produced by levcromakalim to show whether this potassium channel opener selectively produced relaxation mediated by K^ATPchannels. Concentration–response curves for levcromakalim in the absence and presence of glibenclamide (10⁻⁸to 10⁻⁵m) are shown in figure 1. During submaximum contractions induced by phenylephrine (3 × 10⁻⁷m), selective K^ATPchannel opener levcromakalim (10⁻⁸to 10⁻⁵m) induced concentration-dependent relaxation. Selective ATP-sensitive potassium channel antagonist glibenclamide eliminated this relaxation in a concentration-dependent manner with an IC⁵⁰value of (7.21 ± 7.72) × 10⁻⁸m and a maximally inhibitory concentration of (1.22 ± 1.62) × 10⁻⁴m. The IC⁵⁰and maximally inhibitory concentration of glibenclamide obtained in the current study were consistent with previously reported values.17Glibenclamide (10⁻⁶m) did not affect relaxation produced by the nonselective vasodilator papaverine (10⁻⁷to 10⁻⁴m) (fig. 2).

The effects of etomidate and midazolam on relaxation produced by levcromakalim and papaverine were then examined to determine whether these agents selectively modified relaxation mediated by K^ATPchannels. Etomidate significantly impaired relaxation produced by levcromakalim in a concentration-dependent manner (fig. 3A). The effect of etomidate is rather modest in comparison to that of glibenclamide. However, midazolam did not significantly impair relaxation produced by levcromakalim (fig. 3B). In contrast, even the highest concentrations of etomidate (10⁻⁴m) and midazolam (10⁻⁶m) did not affect relaxation produced by the nonselective vasodilator papaverine (fig. 2).

To investigate whether etomidate and midazolam affect K^ATPchannel activity in vascular smooth muscle cells, we measured single K^ATPchannel currents by the patch clamp technique. As shown in figure 4A, spontaneous single channel activity was observed infrequently in the cell-attached configuration (NP^o< 0.01, n = 10). However, application of 10⁻⁴m pinacidil, a selective K^ATPchannel opener, to the bath solution significantly activated K⁺-selective channels (NP^o0.447 ± 0.171, P < 0.05 vs.  baseline, n = 12). This channel activity was completely blocked by 3 × 10⁻⁶m glibenclamide, a specific K^ATPchannel blocker (fig. 4A; n = 12). The channel showed a single channel conductance of 29.4 ± 4.9 pS (n = 15), as measured by the current–voltage relation between −80 and +60 mV membrane potential. These channel properties were consistent with those reported previously.18Representative examples of the effects of etomidate on pinacidil-induced K^ATPchannel activity in the cell-attached configuration are shown in figure 4B. Application of 10⁻⁶, 10⁻⁵, and 10⁻⁴m etomidate to the outside of the membrane surface inhibited pinacidil-induced K^ATPchannel currents, with relative channel activity decreasing to 0.96 ± 0.07 (n = 8), 0.67 ± 0.22 (n = 10), and 0.21 ± 0.11 (n = 9), respectively. Representative examples of the effects of midazolam on pinacidil-induced K^ATPchannel activity in the cell-attached configuration are shown in figure 4C. In contrast to etomidate, at 10⁻⁸, 10⁻⁶, and 10⁻⁴m, midazolam did not inhibit pinacidil-induced K^ATPchannel currents in the cell-attached configuration, with relative channel activity of 0.98 ± 0.04 (n = 14), 0.96 ± 0.15 (n = 16), and 0.90 ± 0.14 (n = 12), respectively.

Concentration-dependent effects of etomidate on pinacidil-induced K^ATPchannel activity in the cell-attached and inside-out configurations are shown in figure 4D. Etomidate significantly inhibited K^ATPchannel activity at concentrations of 3 × 10⁻⁶m or greater in both the cell-attached and inside-out configurations. The IC⁵⁰values in the cell-attached and inside-out configurations were 1.68 × 10⁻⁵m and 1.52 × 10⁻⁵m, respectively. The Hill coefficients in the cell-attached and inside-out configurations were 1.23 and 1.44, respectively. Etomidate did not change the single channel conductance in either the cell-attached or inside-out configuration (data not shown). Concentration-dependent effects of midazolam on pinacidil-induced K^ATPchannel activity in cell-attached and inside-out configurations are shown in figure 4E. Even high concentrations of midazolam had no significant inhibitory effect on K^ATPchannel activity in either the cell-attached or inside-out configuration.

To determine the tissue-specific effects of etomidate on K^ATPchannel activity, we used inside-out patch clamp configurations to investigate the effects of etomidate on the activities of various types of recombinant Kir6.0/SUR channels. Sarcolemmal K^ATPchannels, SUR2A/Kir6.2 (cardiac type), SUR2B/Kir6.1 (vascular smooth muscle type), SUR2B/Kir6.2 (nonvascular smooth muscle type), and SUR1/Kir6.2 (pancreatic β-cell type) were heterologously expressed in COS-7 cells.19Our previously reported experiments showed that the single-channel characteristics of all types of expressed K^ATPchannels are similar to those of native K^ATPchannels.8–10As shown in figure 5A, application of 10⁻⁵m etomidate to the intracellular membrane surface inhibited pinacidil-induced SUR1/Kir6.2, SUR2A/Kir6.2, SUR2B/Kir6.2, and SUR2B/Kir6.1 channel currents with equivalent potency; relative channel activity decreased to 0.47 ± 0.11, 0.43 ± 0.07, 0.58 ± 0.14, and 0.56 ± 0.18, respectively (fig. 5B). The inhibitory effects of etomidate on recombinant K^ATPchannel activity were reversible; channel activity recovered after washout (fig. 5A).

A C-terminal truncated pore-forming subunit of Kir6.2 (Kir6.2ΔC36), lacking the last 36 amino acids, is capable of forming a functional channel in the absence of SUR.20This has proved to be a useful tool for discriminating the site of action of various agents on K^ATPchannels. As seen in figure 6A, etomidate at 10⁻⁵m inhibited Kir6.2ΔC36 channel currents, with relative channel activity decreasing to 0.49 ± 0.24 (n = 8). This result indicates that the Kir subunit, rather than SUR, is primarily responsible for the effects of etomidate on wild-type K^ATPchannels. The inhibitory effects of etomidate on Kir6.2ΔC36 channel activity were reversible; channel activity recovered after washout (fig. 6A).

We next used site-directed mutagenesis of Kir6.2ΔC36 channels to examine whether the site at which etomidate mediates K^ATPchannel inhibition was identical to that involved in adenosine triphosphate block. We used a double-mutant form of Kir6.2 (Kir6.2ΔC36-K185Q).21In this mutant, the inhibitory potency of adenosine triphosphate was significantly reduced, whereas 10⁻⁵m etomidate inhibited Kir6.2ΔC36-K185Q currents as effectively as it inhibited Kir6.2ΔC36 currents (fig. 6B).

The most important findings in the current study is that the representative imidazoline-derived anesthetics, etomidate and midazolam, have different effects on vascular K^ATPchannels in both isometric tension and patch clamp experiments. The isometric tension experiment in rat aortic ring preparations showed that etomidate reduces relaxation produced by levcromakalim, a selective K^ATPchannel opener, in a concentration-dependent manner, whereas midazolam does not affect this vasorelaxation. Similarly, electrophysiologic study by the patch clamp method showed that etomidate inhibits native vascular K^ATPchannel activity in a concentration-dependent manner, whereas midazolam did not significantly affect K^ATPchannel activity in either the cell-attached or inside-out configuration. These results suggest that etomidate, but not midazolam, inhibits the activities of vascular K^ATPchannels at clinical concentrations.

Etomidate is a rapid-acting sedative/hypnotic agent with little or no cardiovascular or respiratory effects.22Because of these properties, etomidate is widely used as an anesthetic induction agent in patients with poor cardiovascular reserve. Similarly, midazolam is used as a premedication, sedative, and anesthesia induction agent because it has only minimal effects on cardiovascular dynamics.23It 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.24Structurally related imidazoline compounds may be closely related to K^ATPchannels. Lee et al.  12reported that imidazoline compounds inhibit K^ATPchannel activity in guinea pig ventricular myocytes and in pancreatic β cells. However, Choi et al.  14revealed that, when used as an induction agent for emergency department rapid sequence intubation, a 10% decrease in mean systolic blood pressure was observed within 5 min after intubation in the midazolam group but that no significant change was observed in the etomidate group. They also reported that hypotension developed in 19.5% of patients given midazolam but in only 3.6% of patients given etomidate, suggesting that midazolam, even at a low dose, was more likely than etomidate to cause significant hypotension. From these results, they concluded that etomidate is a better alternative.14In addition, Canessa et al.  15reported that etomidate was the only agent that did not decrease arterial blood pressure comparing midazolam and other intravenous anesthetic agents for elective cardioversion. Systolic arterial blood pressure decreased significantly with midazolam (19%), whereas no significant change was evident in patients receiving etomidate. Based on these reports,14,15we hypothesized that the effects of etomidate and midazolam on K^ATPchannel activity might differ.

Adenosine triphosphate–sensitive potassium channels are present in a wide variety of tissues and are believed to link cellular metabolic status and excitability.1–4,18In vascular smooth muscle cells, K^ATPchannels regulate the membrane potential, which controls calcium entry through voltage-dependent calcium channels, and thereby contractility through changes in intracellular calcium.5,18Physiologic studies have suggested that opening K^ATPchannels in vascular smooth muscle causes vasodilation; this occurs physiologically in response to certain neurotransmitters and hypoxia and pharmacologically during therapy with K^ATPchannel openers.4,5A recent clinical study further suggested that prophylactic administration of K^ATPchannel opener nicorandil proved useful for perioperative prevention of cardiac complications.25Previous studies, however, showed that some intravenous anesthetics, including barbiturates, propofol, and ketamine, reduced K^ATPchannel opener–induced rat aortic vasodilation.6,7In a similar experimental system, the current study showed that etomidate also reduced this vasodilation (fig. 3A). This finding agrees in part with recently reported findings, that etomidate reduced K^ATPchannel opener lemakalim-induced vasorelaxation in the canine pulmonary artery.26These results suggest, therefore, that the inhibitory effect of etomidate on K^ATPchannel opener–induced vasorelaxation is independent of the specific vascular site. In contrast to other intravenous anesthetics, midazolam at the concentrations we used did not affect this vasorelaxation (fig. 3B).

Recent patch clamp studies of vascular K^ATPchannels showed that these channels are targets of a wide variety of vasodilators and constrictors, which act through multiple cellular signaling pathways, such as protein kinase A and protein kinase C.5In the current electrophysiologic study, however, etomidate inhibited native vascular K^ATPchannel activity with similar potency in both cell-attached and inside-out configurations (fig. 4D). These results suggest that the inhibitory effect of etomidate on K^ATPchannel activity may be due to direct binding to these channels rather than modulation of the cell-signaling pathway. However, in agreement with the results of our isometric tension study, midazolam had no effect on K^ATPchannel activity in either the cell-attached or inside-out configuration (fig. 4E). These results indicate, therefore, that the differential effects of etomidate and midazolam on levcromakalim-induced vasorelaxation were based on direct but different action on vascular smooth muscle K^ATPchannel activity.

The K^ATPchannel is a hetero-octamer composed of two subunits: a Kir6.0 family subunit (Kir6.1 or Kir6.2) and the SUR subunit (SUR1, SUR2A, or SUR2B).19SUR acts as a regulatory subunit, whereas Kir subunits form the ATP-sensitive channel pore. Different combinations of Kir and SUR subunits generate tissue-specific K^ATPchannel subtypes.19Several K^ATPchannel activators and inhibitors show various tissue specificities, and different types of K^ATPchannels exhibit different pharmacologic properties, which are mainly due to the SUR subunit.18,19In the current study, however, inhibitory potencies of etomidate on recombinant SUR/Kir6.0 channel activity were not influenced by the type of SUR subunit (fig. 5). This result suggests that the Kir6.0 subunit, rather than the SUR subunit, is primarily responsible for the effects of etomidate on K^ATPchannels. This view is strongly supported by the fact that etomidate inhibited Kir6.2ΔC36 channels with the same potency as for the recombinant SUR/Kir6.0 channels (fig. 6A). Furthermore, our results indicate that the binding site of etomidate is not identical to that of adenosine triphosphate at the amino acid level; mutation of K185Q markedly decreased adenosine triphosphate sensitivity but was without significant effects on etomidate inhibition (fig. 6B). Although the precise binding site of etomidate remains unclear, recent electrophysiologic studies showed that other imidazoline compounds, including phentolamine and cibenzoline, blocked recombinant K^ATPchannels via  the Kir6.2 subunit.11,13It is, therefore, possible that inhibitory action via  Kir6.0 is a common feature of imidazoline compounds, which modulate K^ATPchannel activity.

The recommended intravenous dose for inducing anesthesia is approximately 0.3 mg/kg for etomidate and 0.2–0.4 mg/kg for midazolam. Peak plasma concentrations during induction at these doses vary widely but are reported to be approximately 3 ×10⁻⁶m for etomidate27and 0.5 to 3 ×10⁻⁶m for midazolam.24,28Because approximately 75% of etomidate29and 96–97% of midazolam24binds to plasma proteins, the concentrations of etomidate and midazolam used in the current study are clinically relevant. Therefore, it is likely that the differential effects of etomidate and midazolam observed in the current experiments will be encountered in clinical settings. Recently reported physiologic studies showed that vascular K^ATPchannels are involved in the maintenance of resting blood flow in a number of vascular beds, notably the coronary circulation, as well as in vasodilation in response to metabolic demand.5,18Thus, K^ATPchannel openers might be advantageous in the treatment of ischemic conditions such as angina. Therefore, our results suggest that in clinical situations, etomidate, but not midazolam, inhibits vascular K^ATPchannel activity, impairing the vasodilation mediated by K^ATPchannel openers. However, there is increasing evidence that excessive activation of vascular K^ATPchannels plays a role in the catastrophic vasodilation and vascular hyporeactivity of circulatory shock.30Under these conditions, the inhibitory action of etomidate on vascular K^ATPchannels could be advantageous. Further studies are needed to clarify the clinically observed influence of etomidate and midazolam on vascular K^ATPchannels.

In conclusion, our study showed that in the isolated rat aorta, clinically relevant concentrations of etomidate, but not midazolam, inhibited relaxation induced by a K^ATPchannel opener. Electrophysiologic patch clamp measurements indicated that the different actions of the two anesthetics are based on differential direct action on vascular smooth muscle K^ATPchannel activity; etomidate directly inhibited K^ATPchannel activity via  the Kir6.0 subunit, whereas midazolam at clinically relevant concentrations had no effect on K^ATPchannel activity.

The authors thank Susumu Seino, M.D., Ph.D. (Professor and Chairman, Department of Cellular and Molecular Medicine, Chiba University, Chiba, Japan), for providing complementary DNAs (Kir6.2, Kir6.1, SUR1, SUR2A, and SUR2B) and expression vector pCMV6C.