Inhibitory Effects of Diazepam and Midazolam on Ca²⁺and K⁺Channels in Canine Tracheal Smooth Muscle Cells 

The Sapporo Medical University Ethical Committee on Animal Research approved the study. Adult mongrel dogs weighing 9-12 kg were anesthetized with 10 mg/kg intramuscular ketamine and killed by exsanguination. The tracheas were excised quickly and placed in modified Krebs' solution equilibrated with 95% oxygen and carbon dioxide at 4 [degree sign]C (composed of 118 mM NaCl, 4.7 mM KCl, 21 mM NaHCO³, 1.2 mM MgSO (⁴), 1.2 mM KH²PO⁴, 10 mM glucose, and 2.5 mM CaCl²; pH [tilde operator] 7.4). Cells were dispersed according to previously described methods. [13,14]Briefly, tracheal smooth muscle was minced and incubated for 10 min in Ca^2+-freemodified Tyrode's solution at room temperature (22-24 [degree sign]C). The modified Tyrode's solution contained 135 mM NaCl, 5.4 mM KCl, 1 mM MgCl², 5 mM glucose, 5 mM N-[2-hydroxyethyl]piperazine-N'-[2-ethanesulfonic acid](HEPES), and 0.1%(wt/vol) bovine serum albumin, with the pH adjusted to 7.4 with 0.5 M tris-[hydroxymethy]aminomethane (Tris). The tissue was digested for 20 min at 37 [degree sign]C in Ca^2+-freemodified Tyrode's solution that contained 0.08%(wt/vol) collagenase, 0.05% trypsin inhibitor, and 0.03% protease. Cells were dispersed by trituration, filtered through nylon mesh, and centrifuged. The pellet was resuspended in a modified Krafbruhe solution [15]and stored at 4 [degree sign]C for as long as 5 h before being used. The modified Kraftbruhe solution contained 85 mM KCl, 30 mM K²HPO⁴, 5 mM MgSO⁴, 5 mM Na²ATP, 5 mM pyruvic acid, 5 mM creatine, 20 mM taurine, 5 mM [small beta, Greek]-hydroxybutyrate, and 0.1%(wt/vol) fatty acid-free bovine serum albumin, with the pH adjusted to 7.25 with Tris.

All experiments were performed at room temperature (22-24 [degree sign]C). Micropipettes were pulled from soda lime hematocrit tubing (GC-1.5; Narishige, Tokyo, Japan) using a two-stage puller (model PP-83, Narishige) and were heat polished. These had resistances of 3 to 5 m Omega when filled with solution. An aliquot (approximately 0.5 ml) of the cell suspension was placed in a perfusion chamber on the stage of an inverted microscope (IX-70; Olympus, Tokyo, Japan). At X600 magnification, a three-dimensional oil-driven micromanipulator (ONM-1; Narishige) was used to position the patch pipettes against the membrane of the tracheal smooth muscle cells. After obtaining a high-resistance seal (5-50 G Omega) with slight suction (5-20 cm water), the patch membrane was disrupted by strong negative pressure, which allowed the voltage of the entire cell membrane to be controlled [16]and permitted the pipette solution to diffuse into the cytoplasm. Membrane currents were monitored using a CEZ-2400 patch-clamp amplifier (Nihon Kohden, Tokyo, Japan), and the amplifier output was low-pass filtered at 2,000 Hz. Leak currents, estimated by appropriate scaling of currents during 20-mV hyperpolarizing pulses, were subtracted from each of these records. Membrane capacitance and series resistance were compensated for by using the internal circuitry of the patch-clamp amplifier. All data were digitized (10,000 samples per s), stored on a hard disk, and analyzed using a 8100/100AV Power Macintosh computer (Apple, Cupertino, CA) using the Pulse+PulseFit 8.02 and Igor Pro 2.04 analysis software programs (Heka, Wiesenstrasse, Lambrecht, Germany).

To measure inward Ca²⁺currents (I^Ca) through VDCCs, recording solutions were chosen to inhibit K⁺currents and enhance Ca (²⁺) currents. The pipette solution contained 130 mM CsCl, 4 mM MgCl², 10 mM EGTA, 5 mM Na²ATP, and 10 mM HEPES, with the pH adjusted to 7.2 with Tris. The bath solution contained 130 mM tetraethylammonium chloride, 1 mM MgCl², 10 mM CaCl², 10 mM glucose, and 10 mM HEPES, with the pH adjusted to 7.4 with Tris. Whole-cell I^Cas were elicited at 5-s intervals by 150-ms depolarizing pulses (-50 to +40 mV in 10-mV increments) from a holding potential of -70 mV. Inactivation curves were determined using a double-pulse protocol that consisted of a 3-s prepulse to a potential of -70 to +20 mV, followed by a 150-ms depolarization to +20 mV. The peak change in the current during the test pulse was expressed as a fraction of that obtained with the -70-mV prepulse, and this quantity was fit to a Boltzmann expression [17,18]using least-squares analysis to estimate the potential of half-maximal inactivation (V^1/2) and the slope factor (k).

To measure outward K⁺currents (I^K), recording solutions were chosen to enhance the K⁺currents. The pipette solution contained 70 mM KCl, 60 mM K^+-glutamate, 5 mM K²ATP, 1 mM MgCl², 2.5 mM EGTA, 1.8 mM CaCl², and 10 mM HEPES, with the pH adjusted to 7.2 with Tris; the computer-calculated [Ca²⁺]ⁱwas [tilde operator] 10^-6M. A variant of this solution contained 10 mM EGTA and no CaCl², giving a [Ca²⁺]ⁱof <or= to 10^-9M. The bath solution contained 135 mM NaCl, mM KCl 5.2, 1.8 mM CaCl², 1 mM MgCl², 10 mM HEPES, and 10 mM glucose, with the pH adjusted to 7.4 with Tris. Whole-cell I^Ks were elicited at 5-s intervals by 150-ms depolarizing pulses (-40 to +60 mV) from a holding potential of -70 mV.

Voltage-pulse protocols were performed in control solutions for more than 5 min to obtain a stable baseline. Cells were exposed to a single concentration of one of the benzodiazepines (diazepam, 10^-8to 10^-3M; or midazolam, 10^-8to 10^-3M) tested by changing the inflow perfusate of the chamber to one of a similar composition but with benzodiazepine. The perfusion chamber consisted of a glass coverslip bottom, with needles placed for rapid solution changes. The chamber volume was approximately 1 ml, and complete solution changes in the chamber could be obtained within 1 min using a peristaltic pump (CTP-3; Iuchi, Tokyo, Japan) attached to the input and output ports. After a 6-min exposure, the perfusate was switched again to the control solution. The G Omega seal was maintained for a period sufficient to evaluate the reversibility of the effects of benzodiazepine in 206 of 238 experiments (86%). In another experiment, the effects of the benzodiazepine antagonists flumazenil (10^-5M, a specific central type [19]) and PK11195 (10^-5M, a specific peripheral type [20,21]) on these channels were tested alone and with these benzodiazepine agonists.

To identify the characteristics of the I^Caseen in this study, the effects of the L-type VDCC antagonist nifedipine (10^-6M) and the agonist Bay K 8644 (10^-6M) on I^Cawere evaluated. The effects of charybdotoxin (40 nM), a specific Ca^2+-activatedK⁺(K^Ca) channel blocker, [22]and 4-aminopyridine (1 mM), a specific Ca^{2+-independent}delayed rectifier K⁺(K^DR) channel blocker, [22]on I^Kwere also evaluated to identify the characteristics of the I^Ks seen in this study.

The following drugs and chemicals were used: trypsin inhibitor (from soybean), bovine serum albumin, Na²ATP, pyruvic acid, creatine, taurine, [small beta, Greek]-hydroxybutyrate, EGTA, TEACl, nifedipine, Bay K 8644, dimethyl sulfoxide, charybdotoxin, 4-aminopyridine (Sigma Chemical Co., St. Louis, MO), type-I collagenase (Gibco Laboratories, Grand Island, NY), protease (Calbiochem, La Jolla, CA), and PK11195 (Research Biochemicals, Natick, MA). Diazepam, midazolam, and flumazenil were donated by Yamanouchi Pharmaceutical Company (Tokyo, Japan). Nifedipine and Bay K 8644 were dissolved in ethanol, and diazepam was dissolved in dimethyl sulfoxide (0.01% final concentrations for both).

Data are expressed as the mean +/− SD. The IC⁵⁰s of the effects of the benzodiazepines on I^Caand I^Kwere obtained using a Boltzmann expression. [17,18]Changes in peak whole-cell currents (I^Caor I^K) or in the inactivation parameters V^1/2and k with exposure to each drug were compared at each applied potential using the paired, two-tailed t test. The percentage of control peak whole-cell currents (I^Caor I^K) and the values of V^1/2and k after treatment were compared for the benzodiazepines using one-factor analysis of variance and a Kruskal-Wallis test. In all comparisons, P < 0.05 was considered significant.

The I^Caseen in enzymatically dispersed canine tracheal smooth muscle cells during step depolarizations from -70 mV peaked at approximately 10 ms and was inactivated with a time constant of approximately 50 to 90 ms (Figure 1A: control). During baseline conditions, threshold activation of I^Caoccurred at approximately -20 mV, and maximum peak current amplitude was obtained at approximately +20 mV. In 138 cells, the maximum peak I^Cawas -318 +/− 26 pA (range, -201 to -612 pA). The inactivation parameters obtained in 28 cells during control conditions were V (^1/2)=-20.4 +/− 2.9 mV and k = 7.2 +/− 1.3 mV. As previously reported in porcine tracheal smooth muscle cells, [13,14]the addition of 10^-6M nifedipine, a blocker of slowly inactivating (L-type) Ca²⁺channels, virtually eliminated the I^Caof canine tracheal smooth muscle cells by approximately 93%, and 10^-6M Bay K 8644, an agonist of L-type Ca²⁺channels, enhanced the magnitude of I^Ca(by approximately 2.4 times) but did not alter the time course of the currents (n = 3 in each case, data not shown). Inward Ca²⁺currents with a similar time course were observed in the inactivation experiments.

As shown in a representative trace for depolarization from -70 to +20 mV (Figure 1A), midazolam (10^-4M) inhibited the magnitude of I (^Ca) but did not obviously alter the time course of the current. Peak I^Caobtained with repeated steps to +20 mV increased in a few minutes after obtaining the whole-cell configuration at time 0 to a stable plateau, decreased rapidly by approximately 50% during exposure to 10^-4M midazolam and recovered completely with washout (Figure 1B). Similar results were obtained with diazepam. Figure 2shows the relation between peak I^Caagainst applied potential before and after exposure to 10^-6M diazepam and 10^-5M midazolam. Each of these benzodiazepines significantly inhibited I^Caat step potentials ranging from of -10 to +40 mV and decreased peak I^Caat +20 mV by approximately 50%(n = 7). The actual percentage inhibitions of peak I^Caachieved by these agents at these concentrations (50.3 +/− 8.4% and 45.8 +/− 8.8%, respectively) were not significantly different. There was no apparent shift in the voltage dependence of I^Cawith either of the benzodiazepines.

We determined the dose dependence of the inhibition of peak I^Caby each of these benzodiazepines. Figure 3shows the relation between the percentage of control peak I^Caat +20 mV and the molar concentration of the agents in the bath solution. Each of the two benzodiazepines significantly inhibited peak I^Cain a dose-dependent manner. Midazolam (IC⁵⁰= approximately 1.2 x 10^-5M) required a 10-fold greater concentration to achieve the same inhibitory effect as that of diazepam (IC (⁵⁰)= approximately 10^-6M).

(Figure 4and Table 1) summarize the effects of the benzodiazepines diazepam and midazolam at equieffective inhibitory concentrations (10^-6M and 10^-5M, respectively) on the inactivation curves of I^Ca. Each of these agents shifted the inactivation curve to a more negative potential. The induced changes in V^1/2brought about by these agents were statistically significant in each case, and there was no significant difference in V^1/2between these agents. The slope factor k was not changed by exposure to either of the benzodiazepines.

The effects of the benzodiazepine antagonists flumazenil and PK11195 were also tested on the control I^Caand on the inhibitory effect on I^Caof the benzodiazepine agonists diazepam and midazolam. Flumazenil (10^-5M) and PK11195 (10^-5M) had no significant effect on the control I^Ca(n = 3 in each case, data not shown). Figure 5shows the time course of the peak I^Caobtained in a representative cell with repeated steps to +20 mV during exposure to 10^-5M flumazenil and 10^-6M diazepam. Despite pretreatment with a high concentration of flumazenil, 10^-6M diazepam still induced an approximate 50% inhibition. Similar results were obtained with 10^-5M PK11195 and 10^-6M diazepam, with 10^-5M flumazenil and 10^-5M midazolam, and with 10^-5M PK11195 and 10^-5M midazolam (n = 3 in each case, data not shown).

(Figure 6A) shows a macroscopic outward K⁺current (I^K) obtained from a freshly dispersed canine tracheal smooth muscle cell dialyzed with a pipette solution containing a [Ca²⁺]ⁱof [tilde operator] 10^-6M to enhance I^Kthrough K^Cachannels. The I^Kwas activated progressively by 150-ms depolarizing pulses from a holding potential of -70 mV to consecutively more positive membrane potentials. Stepwise depolarization from a holding potential of -70 mV to more than -30 mV elicited an outward I^Kwith a mean peak amplitude of 1,840 +/− 201 pA at +60 mV (n = 92). The addition of 40 nM charybdotoxin, a specific K^Cachannel blocker, significantly decreased peak I^Kwithout any change in the time course of the current (Figure 6B). Figure 6C summarizes the current-voltage (I-V) relation plotted as percentages of maximum I^Kbefore and after exposure to charybdotoxin. Application of charybdotoxin reduced the peak current at +60 mV by 65 +/− 14%(n = 4).

The effects of benzodiazepines on this charybdotoxin-sensitive I (^K) was examined in 84 cells. Midazolam (10^-4M) caused an approximately 15% reduction in I^Kwithout any apparent effect on the time course of the current (Figure 7A), an effect that was reversible when the drug was washed out (data not shown). Figure 7B shows the effects of 10 (^-4) M midazolam on the current-voltage (I-V) relation for K⁺channel activation. This benzodiazepine significantly suppressed the I^Kamplitude over the entire voltage range studied without shifting the voltage dependency of the I-V relation. Figure 7C shows the relation between peak I (^K) at +60 mV, expressed as a percentage control, and the bath concentrations of the benzodiazepines. Both the benzodiazepines diazepam and midazolam significantly and dose dependently inhibited I^Kbut required high concentrations of more than 10^-5M and more than 10^-4M, respectively, to show significant effects. Flumazenil (10^-5M) and PK11195 (10^-5M) had no effect on the control I^Kor on the inhibitory effect on the I^Kof these benzodiazepines (n = 3 in each case, data not shown).

In a separate series of experiments, we used a pipette solution in which [Ca²⁺]ⁱwas strongly buffered with 10 ma EGTA to minimize the outward I^Kthrough K^Cachannels, and we examined the effects of benzodiazepines on Ca^{2+-independent}I^K. Figure 8A shows a representative trace of I^Kunder these conditions. The I^Ks were activated progressively by 150-ms depolarizing pulses from a holding potential of -70 mV to consecutively more positive potentials. The mean peak amplitude in 92 cells was 518 +/− 110 pA at +60 mV. The application of 40 nM charybdotoxin had no effect on the I^K(n = 3, data not shown). Then we examined the effect of 1 mM 4-aminopyridine on this Ca^{2+-independent}I (^K). As shown in Figure 8, B and C, 4-aminopyridine decreased the peak I (^K) amplitude without changing the time course of the current at +60 mV by 72 +/− 14%.

The effects of benzodiazepines on this 4-aminopyridine-sensitive I (^K) were examined in 84 cells. Diazepam (10^-4M) reversibly suppressed the I^Kwithout changing the time course of the current (Figure 9A). This agent significantly suppressed the I^Kamplitude for the entire voltage range studied without shifting the voltage dependency of the I-V relation (Figure 9B). Figure 9C shows the relation between the percentage control of peak I^Kat +60 mV and the bath concentrations of the benzodiazepines. Diazepam and midazolam significantly and dose dependently inhibited I^Kbut also needed high concentrations (>10^-5M and >10 (^-4) M, respectively) to depress I^K. Flumazenil (10^-5M) and PK11195 (10^-5M) had no effect on the control I^Kor on the inhibitory effects on the I^Kof these benzodiazepines (n = 3 in each case, data not shown).

As previously reported in porcine tracheal smooth muscle cells, [13,14]we measured depolarization-induced inward Ca²⁺currents (I (^Ca)) in freshly dispersed canine tracheal smooth muscle cells under ionic conditions designed to inhibit K⁺and Na⁺currents and to enhance Ca²⁺currents. Based on their time and voltage dependencies, their sensitivity to a nifedipine blockade, and their enhancement by the Ca²⁺channel agonist Bay K 8644 (data not shown), these currents are presumed to reflect the activity of long-lasting (L-type) VDCCs. [13,14,23,24] 

Both of the benzodiazepines tested inhibited I^Cathrough VDCCs of canine tracheal smooth muscle cells without an apparent change in the kinetics of activation or inactivation (Figure 1A). The onset of inhibition was rapid and the effect was reversible (Figure 1B) and dose related (Figure 3). None of the benzodiazepines altered the voltage dependence of I^Ca(Figure 2), suggesting that the drug has no effect on membrane surface charge or on the voltage sensor of the channel. These data indicate a cellular effect of benzodiazepines that can account for the airway smooth muscle relaxant effects of these agents. [1,2,7]The concentrations of midazolam required to inhibit VDCCs shown in Figure 3are not similar to those required to directly relax preconstricted airway smooth muscle preparations. [1,2,25]This discrepancy may result from differences in species and concentrations of contractile agonists and from differences in the experimental techniques we used.

To evaluate the inhibitory actions of these benzodiazepines on VDCCs of tracheal smooth muscle cells further, we studied the effects of these agents on steady-state, voltage-dependent inactivation of I^Ca. During prolonged depolarization (prepulse), a fraction of the VDCCs enters an unavailable or “inactivated” state. The degree of steady state inactivation depends on the prepulse potential (Figure 4). [14,26]The mean potential of half-maximal inactivation (V^1/2= approximately 20.4 mV) and the mean slope factor (k = 7.2 mV) that we obtained under control conditions are similar to values reported from porcine [14,26]and bovine [24]tracheal smooth muscle cells. Both of the benzodiazepines significantly shifted the inactivation curves to more negative potentials without changing the sigmoid shapes of the curves (Figure 4, Table 1). A qualitatively similar shift induced by a dihydropyridine-sensitive Ca²⁺antagonist, such as nifedipine, in canine colonic [27]and porcine tracheal [26]smooth muscle cells has been interpreted as evidence of a drug-induced stabilization of the inactivated state. [28]Because a phenylalkylamine Ca (²⁺) antagonist, such as verapamil, has no effect on the inactivation curve of I^Ca, [26]we speculate that benzodiazepines have a dihydropyridine-sensitive Ca²⁺antagonist-like inhibitory effect on VDCCs in canine tracheal smooth muscle cells, reflecting the stabilization of the VDCCs in their activated and inactivated states because of the high affinity of the agents for these states. [28]The reported resting membrane potential typically is approximately -50 to -60 mV for airway smooth muscle. [29]Based on the voltage-dependent blockade and a shift in inactivation of I^Ca, these agents would have a more inhibitory effect on I^Caat a lower membrane potential.

In this study, using 2.5 mM EGTA and 1.8 mM CaCl²([Ca²⁺]ⁱof [tilde operator] 10^-6M) in the pipette solution, we recorded whole-cell I^Kvalues in freshly dispersed canine tracheal smooth muscle cells. Because I^Khad a large conductance, and peak I^Kat +60 mV was significantly suppressed by the K^Cachannel blocker [23]charybdotoxin by approximately 65%(Figure 6), this current may include a considerable amount of K⁺current through K^Cachannels. [30-32]Similar K^Cachannel blockers (charybdotoxin) or TEA-sensitive L^Ks have been observed in vascular smooth muscle cells. [33-35]In another study using a pipette solution containing no Ca²⁺and 10 mM EGTA to buffer the intracellular free Ca²⁺(Ca²⁺]ⁱof <or= to 10 (^-9) M), we also recorded a whole-cell outward I^Kthat was charybdotoxin insensitive but 4-aminopyridine sensitive (Figure 8). Under these conditions, I^Kdisplays the properties of I^Kthrough delayed rectifier K⁺(K^DR) channels. [31-33]Similar outward K^DRcurrents also have been observed in vascular smooth muscle. [35,36] 

Based on pharmacologic and electrophysiologic properties, these two types of K⁺channels have been identified in airway smooth muscle cell membranes. [32]The most common K⁺channel in airway smooth muscle is the K^Cachannel. [31,32]The open-state probability of this channel is low at the resting membrane potential but increases in proportion to membrane depolarization and elevation of [Ca²⁺]ⁱ. The resulting enhanced K⁺efflux will induce membrane repolarization or hyperpolarization, reduce the open-state probability of VDCCs, and in turn cause bronchodilation. In addition of this K⁺channel type, voltage- but not Ca^2+-dependentK^DRchannels also were identified in airway smooth muscle cell membranes. [32]The K^DRchannels may play a central role in setting the level of the restine membrane potential. [32] 

In this study, the whole-cell outward I^Ks through K^Caand K^DRchannels were inhibited significantly by the two benzodiazepines tested (Figure 7and Figure 9), which might lead to membrane depolarization and muscle contraction. The results of this study indicate that diazepam and midazolam did not enhance the K⁺current, which would be expected to promote bronchodilation, but instead they reduced the activity of K⁺channels in canine tracheal smooth muscle cell membranes. This suggests that mechanisms other than K⁺channel opening are likely to mediate benzodiazepine-induced bronchodilation. Because these agents also needed high concentrations to inhibit I^Ks compared with those that exhibited significant inhibitory effects on I^Cas in canine tracheal smooth muscle cells (Figure 3, Figure 7, and Figure 9), the suppression of I^Cas through VDCCs may be one of the most important mechanisms by which benzodiazepines relax airway smooth muscle.

Airway smooth muscle tone also could be regulated by some neuropeptides. [37][small gamma, Greek]-Aminobutyric acid (GABA) has an inhibitory effect on postganglionic cholinergic neurotransmission in ferret airways. [38]The benzodiazepine receptor is a positive modulatory subunit of the GABA receptor and enhances the chloride channel currents by increasing its opening frequency. [39]Therefore, in addition to directly inhibiting VDCCs, benzodiazepines might inhibit airway smooth muscle contraction by stimulating some benzodiazepine receptor, which leads to GABA receptor activation. In the current study, however, 10^-5M flumazenil and 10^-5M PK11195 (specific central [19]and specific peripheral [20,21]benzodiazepine antagonists, respectively) had no effect on the control I^Caand I^Kor on the changes in I^Caand I^Kinduced by the benzodiazepine agonists (Figure 5). Therefore, the diazepam and midazolam benzodiazepines probably relax the airway smooth muscle by binding cell membranes relating to VDCCs, rather than by activating benzodiazepine receptors. In support of our findings, studies have shown that flumazenil and PK11195 have no effect on benzodiazepine-induced relaxation of airway smooth muscle. [1,2,7]The concentration of flumazenil (10^-5M) used in this study is greater than the estimated levels of plasma concentrations used clinically. [20,40,41] 

The benzodiazepines tested showed dose-dependent inhibition of I (^Ca) and I^K(Figure 3, Figure 7, and Figure 9). Diazepam is more potent than midazolam in terms of I^Caand I^K. Our data should be extrapolated to the clinical situation cautiously because of possible species differences, in vivo and in vitro differences, and the fact that our patch-clamp experiments were performed at low, nonphysiologic (ambient) temperature and using intracellular (pipette) and extracellular (organ bath) electrolytes. Nonetheless, the plasma concentrations of the benzodiazepines used clinically are approximately 3 x 10^-7to 10^-5M. [42-44]In the current study, the bath concentrations of the benzodiazepines diazepam and midazolam used to induce 50% inhibition of I^Cawere approximately 10 (^-6) M and 10^-5M, respectively, which seems relevant to clinical concentrations. We must note, however, that these agents are highly bound to plasma protein (>90% bound), [42]and the estimated plasma concentrations of free agents seem to be approximately 10^-9to 10^-6M. Therefore, the similarity between the clinical concentrations [42-44]and the concentrations needed to inhibit I^Caby approximately 50% in this study simply may be a coincidence. Other mechanisms, such that benzodiazepines acts on the medulla to cause airway dilatation, [45]also should be considered.

In conclusion, the benzodiazepines diazepam and midazolam decreased the inward Ca²⁺current of canine tracheal smooth muscle cells, indicating the inhibition of VDCCs. This response could contribute to the ability of these agents to relax airway smooth muscle in vitro. A shift in the inactivation curve by these agents to more negative potentials can be interpreted as evidence of drug-induced stabilization of the inactivated state. Because the benzodiazepines reduced the K⁺channel activity at high concentrations, K⁺channel opening could not be responsible for the mechanisms of benzodiazepine-induced bronchodilation. The lack of antagonized effects of their antagonists flumazenil and PK11195 is related to the non-GABA-mediated electrophysiologic effects of benzodiazepines on airway smooth muscle contractility.

[Section] Ducros L, Plaisance P, Joseph T, Bard M, Salmeron S, Payen D, Lecarpentier Y: Determinants of ketamine-induce bronchodilation in guinea pig tracheal smooth muscle (abstract). Am J Respir Crit Care Med 1996; 153:A741