Isobolographic Analysis of the Interactions between Midazolam and Propofol at GABA^AReceptors in Embryonic Mouse Neurons 

Animal treatment protocols were approved by the Animal Care Committee of the University of Toronto. Cultures of embryonic hippocampal neurons were prepared from Swiss white mice, as previously described. [29]Briefly, fetal pups (17 days in utero) were removed from mice killed by cervical dislocation. The hippocampi were dissected from each fetus and placed in an ice-cooled culture dish. Neurons were then dissociated by mechanical trituration using two Pasteur pipettes (tip diameter, 150-200 [micro sign]m) and plated on 35-mm culture dishes at a density of approximately 1 x 10⁶cells/ml. The culture dishes were coated with collagen (from calf skin) or poly-D-lysine (Sigma Chemical Company, St. Louis, MO). For the first 10 days in vitro, cells were maintained in minimal essential media supplemented with 10% fetal bovine serum and 10% horse serum (Life Technologies, Grand Island, NY). The neurons were cultured at 36.5 [degree sign]C in a 5% carbon dioxide-95% air environment. After the background cells had grown to confluence, 0.1 ml of a mixture of 4 mg 5-fluorodeoxyuridine and 10 mg uridine in 20 ml minimal essential media was added to the extracellular solution to reduce the number of dividing cells. Subsequently, the media was supplemented with 10% horse serum and changed every 3 or 4 days. Cells were studied 12 to 16 days after dissociation. For all reported results, data were acquired from the cells from at least three different dissections.

Before recording, neurons were rinsed with an extracellular recording solution containing 140 mM NaCl, 1.3 mM CaCl², 5.4 mM KCl, 25 mM N-2-hydroxy-ethylpiperazine-N'-2-ethanesulphonic acid (HEPES), 33 mM glucose, 1 mM MgCl², and 0.3 mM tetrodotoxin (TTX)(pH 7.4, 325-335 mOsm). Whole-cell patch-pipettes (3-10 M Omega) were fabricated from borosilicate glass capillary tubing and filled with a solution containing 140 mM choline-Cl, 10 mM HEPES, 11 mM EGTA, 1 mM CaCl², 10 mM tetraethylammonium chloride (TEA-Cl), 2 mM MgCl², and 4 mM Mg-ATP (pH 7.4, 290-300 mOsm). Neurons were voltage-clamped at -60 mV, and all experiments were conducted at room temperature. Whole-cell currents were recorded using the Axopatch 1-D amplifier (Axon Instruments, Foster City, CA). Data were digitized, filtered (2 kHz), and acquired on-line using the pClamp 6 program (Axon Instruments). A multibarreled perfusion system was used to achieve a rapid exchange of solutions. Solutions flowed continuously through the barrels, and the barrel position was controlled by a computer-driven DC stepping motor (Vexta; Oriental Motor USA, Torrance, CA). The amplitude of GABA-evoked whole-cell currents gradually, and irreversibly, decreases progressively to a stable amplitude. Presumably, this slow “run-down” of current results from changes in the intracellular contents of the cell after disruption of the membrane patch. To allow for run-down, no responses were used for data analyses until 10 min after breakthrough.

Intralipid, at concentrations equivalent to those used in the experiments, had no effect on the amplitude or time course of GABA-activated current. However, propofol at high concentrations (approximately 1 to 6 [micro sign]M) directly activated the GABA^Areceptor in hippocampal neurons. [11,29]In our experiments, it was not possible to distinguish current resulting from receptor activation by propofol from the GABA-evoked (but propofol-modulated) response. Propofol-activated current could not simply be subtracted from the total current because GABA enhances propofol-induced receptor activation. Therefore, for this analysis, we did not try to differentiate between propofol- and GABA-evoked currents.

An underlying assumption of the isobolographic analysis is that the peak current accurately reflects receptor activation. However, the peak current could be underestimated because of desensitization that occurs during the onset of the response. Experiments were designed to minimize the effect of desensitization on the amplitude of the peak response. To ensure that the rate of activation was considerably faster that the rate of desensitization, low concentrations of GABA (less than the mean effective concentration [EC (⁵⁰)]) were used to elicit control responses. This analysis also assumed that the peak current was measured during equilibrium conditions. However, the rate of equilibration of propofol with the GABA^Areceptor is likely to be slower than that of GABA or midazolam. [29]Nevertheless, the effect of slow equilibration of propofol was minimal because the current gradually increased to a relatively stable amplitude.

Whole-cell currents were analyzed using the CLAMP-FIT program of the pClamp software (Axon Instruments), and the concentration-response analysis was performed using Graph Pad Prism software (Graph Pad, San Diego, CA). Current amplitudes were plotted against the drug concentrations and fit using a standard logistic Equation 1Where I is the current amplitude, I (^max) is the maximum current amplitude, C is the concentration of the agonist, EC⁵⁰is the concentration of agonist that produced 50% of the maximal response, and n is the Hill coefficient. Because the ability of anesthetics to enhance GABA-evoked currents is greatest at low concentrations of agonist, control responses were elicited by low concentrations of GABA (0.3, 1, and 3 [micro sign]M). These concentrations of GABA were selected based on our previous studies [29]of GABA-evoked currents from embryonic hippocampal neurons. In the current study, 0.3, 1, and 3 [micro sign]M GABA produced 3%, 9%, and 23%, respectively, of the peak current evoked by a saturating concentration of GABA (600 [micro sign]M). Thus, the concentrations of GABA used for control responses were much lower than the EC (⁵⁰) value.

Analysis of Drug Interactions. The potency of a drug is measured as a dose (or concentration) that produces a specified effect level. In this study, two methods were used to determine whether midazolam and propofol interacted in an additive, supraadditive (synergistic), or subadditive (antagonistic) manner. [31-34]In the first series of experiments, a fixed-concentration model was used to evaluate drug interactions among a range of drug and agonist concentrations. [3]The fixed-concentration model predicts that the combination of drug 1 and drug 2 will produce an effect equal to the predicted sum of each effect alone (or the theoretical additive response). This analysis assumes that the concentrations of the drugs used for the analysis fall on the linear portion of the dose-response curve. [32]Furthermore, although the drugs may differ in potency, it also assumes that the shapes of the dose-response curves are similar. The concentration-response curves for midazolam- and propofol-induced enhancement of GABA^Areceptor activity may not be parallel, because drugs that modulate the GABA^Areceptor can greatly differ in potency. Thus, the potency ratio of these two drugs could change with the level of effect, and synergism may not be detected unless a wide range of concentrations is tested.

In the second series of experiments, drug interactions were evaluated using isobolographic analysis. The isobolographic analysis evaluates a range of drug concentrations and considers equieffective concentrations of two drugs and the combination of the two drugs that yield a specified level of effect. The isobologram is a Cartesian plot of pairs of doses that, in combination, produce a specified level of effect. The isobologram was constructed by determining the potency of propofol and midazolam alone to enhance GABA-evoked currents. We arbitrarily selected an effect level of 300%, or three times the amplitude of current evoked by a low concentration of GABA alone (i.e., a 200% increase in the response to GABA alone). The concentration of GABA was chosen such that even a response that was three times the control response was not near saturation. The concentrations of midazolam and propofol that increased the GABA-evoked responses by a factor of three were then plotted on the x and y axes of the isobologram, respectively. A diagonal line, called the “isobol of additivity,” was then joined to these two end points. Subsequently, we selected a lower, fixed concentration of midazolam and determined the concentration of propofol that produced (in combination with midazolam) the desired effect level. If the two drugs have an additive interaction, the concentration of propofol would lie on the isobol of additivity. [30,32]However, if the two drugs produce a synergistic interaction, the concentration of propofol would fall below the isobol of additivity because less propofol would be needed to produce the desired effect. Conversely, if the drugs interact in an antagonistic or subadditive manner, the propofol dose would lie above the isobol of additivity. The advantage of this method is that it allows for a rigorous statistical test of synergy. Statistical considerations for isobolographic analysis have been described before. [32-34] 

Propofol was prepared from Diprivan (Zeneca Pharma, Missassauga, Ontario, Canada), where each milliliter of Diprivan contained 10 mg 2,6 di-isopropylphenol, 100 mg soybean oil, 12 mg egg lecithin, and 22.5 mg glycerol. Stocks of propofol (1 and 10 mM) were prepared using filtered water every 1 to 3 days and were stored at 4 [degree sign]C. Midazolam was prepared from Versed (Hoffmann-LaRoche, Missassauga, Ontario, Canada) and dissolved in extracellular solution. Each milliliter of Versed contained 1 mg midazolam, 0.1 mg disodium edetate, 10.45 mg benzyl alcohol, and 8 mg NaCl. In addition, purified midazolam (donated by Hoffmann-LaRoche) was dissolved in dimethylsulfoxide. There was no difference in the potentiation of GABA-evoked currents when midazolam was prepared from Versed versus the purified compound. Flurazepam was also dissolved in dimethyl-sulfoxide. Other chemicals, including EGTA, HEPES, and potassium chloride, were purchased from Fluka Biochemika (Bucks, Switzerland); dimethylsulfoxide was received from Fischer Scientific (Fairlawn, NJ), CNQX was received from Tocris Cookson (St. Louis, MO), TTX was received from Precision Biochemicals, (Vancouver, British Columbia, Canada), and NaCl, CaCl², glucose, and NaOH were received from BDH (Toronto, Ontario, Canada). Unless otherwise specified, all other compounds were purchased from Sigma Chemical Company.

Data are presented as the mean +/− SEM or are normalized to control values, unless otherwise indicated. An analysis of variance was used to compare multiple groups of data. The EC⁵⁰values and Hill slopes of different concentration-response curves were compared using paired Student's t tests. Statistical significance was accepted when P was less than 0.05.

The application of exogenous GABA (0.1-600 [micro sign]M) elicited inward current in all hippocampal neurons that were voltage-clamped at -60 mV. Midazolam (0.5 [micro sign]M) and propofol (10 [micro sign]M) both increased the amplitude of currents evoked by subsaturating concentrations of GABA (3 [micro sign]M), as illustrated in Figure 1A. The peak amplitude of the current activated by 3 [micro sign]M GABA, in the absence or presence of midazolam, propofol, or the combination of midazolam and propofol, was less than the maximal current evoked by a saturating concentration of GABA (600 [micro sign]M). GABA (3 [micro sign]M)-activated currents were enhanced by midazolam and propofol by factors of 1.45 +/− 0.21 and 3.44 +/− 0.41, respectively, whereas the combination of midazolam and propofol enhanced currents by a factor of 4.15 +/− 0.54 (n = 8;Figure 1B). The increase in the peak current, recorded in the presence of propofol and midazolam, was not significantly different from the predicted additive response (P > 0.05; Student's t test).

Similar results were obtained when 0.5 [micro sign]M flurazepam was substituted for midazolam; 0.5 [micro sign]M flurazepam, 10 [micro sign]M propofol, and 10 [micro sign]M propofol and 0.5 [micro sign]M flurazepam increased the peak current by a factor of 1.26 +/− 0.06, 2.49 +/− 0.36, and 3.24 +/− 0.56, respectively (n = 7;Figure 1C). These data suggest that midazolam and propofol or flurazepam and propofol had an additive interaction with respect to their ability to enhance the peak current evoked by 3 [micro sign]M GABA in hippocampal neurons (Figure 1D).

The ability of anesthetics to enhance GABA-evoked responses is influenced by both the concentration of the anesthetic and the concentration of GABA. [20]In the next series of experiments, the concentration of propofol was decreased from 10 to 1 [micro sign]M, whereas the concentration of midazolam was unchanged. Midazolam at doses of 0.5 [micro sign]M, 1 [micro sign]M propofol, and 1 [micro sign]M propofol plus 0.5 midazolam enhanced currents evoked by 3 [micro sign]M GABA by a factor of 1.62 +/− 0.40, 1.89 +/− 0.60, and 2.72 +/− 1.08 (n = 5), respectively. The interaction between midazolam and propofol was again additive at this lower concentration of propofol (P > 0.05; Student's t test). Because the enhancement of GABA-evoked current by anesthetics is greatest at low concentrations of GABA, in the next series of experiments, the concentration of GABA was reduced from 3 [micro sign]M to 0.3 [micro sign]M. Interestingly, a supraadditive interaction was observed between 0.5 [micro sign]M midazolam and 1 [micro sign]M propofol (Figure 2) in the presence of this lower concentration of GABA. GABA-evoked currents were enhanced by midazolam, propofol, and midazolam plus propofol by a factor of 3.54 +/− 0.39, 4.93 +/− 0.80, and 14.73 +/− 3.03, respectively (n = 4; P < 0.05, Student's t test).

The supraadditive interaction observed between midazolam and propofol in the enhancement of current activated by low concentrations of GABA was further investigated using isobolographic analysis. The concentration of GABA used to elicit the control response was increased from 0.3 [micro sign]M to 1 [micro sign]M because the small current activated by the lower concentration of GABA was difficult to resolve from the background noise. GABA at a dose of 1 [micro sign]M activated approximately 9% of the maximum response evoked by a saturating concentration of GABA. The peak amplitude of current activated by 1 [micro sign]M GABA and of various concentrations of propofol (0.1 to 100 [micro sign]M) was measured, and the concentration-response curve for propofol-induced enhancement of GABA-evoked current was constructed (Figure 3A).

The maximal increase of current was produced by 100 [micro sign]M propofol, which enhanced the response by a factor of 12.6 +/− 3.4 (n = 9, Figure 3B). Thus, the current activated by a low concentration of GABA (EC (⁹)) was increased 12.6 times and had an amplitude similar to that evoked by a saturating concentration of GABA. The similar maximum response suggests that the same population of receptors is activated by GABA (EC¹⁰⁰) and GABA (EC⁹) plus propofol. The concentration of propofol that produced 50% of the maximal increase and the calculated Hill coefficient were 7.12 +/− 1.34 [micro sign]M and 1.27 +/− 0.15 (n = 9), respectively (Figure 3B). The concentration of propofol that increased the amplitude of the control response by a factor of three was 2.42 +/− 0.43 [micro sign]M.

Midazolam also produced a concentration-dependent increase in the amplitude of GABA (1 [micro sign]M)-evoked currents (Figure 4A). The concentration-response relation for midazolam-modulated current is illustrated in Figure 4B. The EC⁵⁰and Hill coefficients for potentiation of GABA-evoked current were 0.10 +/− 0.03 [micro sign]M and 1.18 +/− 0.18, respectively. The maximal increase observed with midazolam was considerably less than that observed with propofol, presumably because midazolam does not directly activate the receptor. GABA-evoked currents were maximally increased by a factor of 4.25 +/− 0.55 (n = 9). However, in two cells, the maximal enhancement of the GABA-evoked responses by midazolam failed to reach the desired effect level (2.25 and 2.37, respectively), and data from these two cells were not included in further analysis. The concentration of midazolam that enhanced GABA-evoked responses by a factor of three was 0.18 +/− 0.08 [micro sign]M (n = 7, Figure 4B).

To evaluate the interaction between midazolam and propofol and to complete the isobolographic analysis, a low concentration of midazolam (0.06 [micro sign]M) was selected. Currents were recorded in the presence of 1 [micro sign]M GABA, 0.06 [micro sign]M midazolam, and various concentrations of propofol (0.06 to 6 [micro sign]M;Figure 5A). The concentration of propofol necessary to increase the control response by a factor of three in the presence of 0.06 [micro sign]M midazolam was determined from individual concentration-response curves constructed for current from each cell (Figure 5B).

The concentrations of midazolam and propofol alone that produced the increase in current to the effect level were plotted on the x and y axes of the isobologram, respectively (Figure 6), and a straight line was drawn between these points to form the isobol of additivity. The interpolated concentrations of propofol that caused the threefold increase in current recorded in the presence of 1 [micro sign]M GABA and 0.06 [micro sign]M midazolam were then plotted on the isobologram. In all cells tested (n = 9), the concentration of propofol that produced the effect level was lower than the concentration predicted by the isobol of additivity. Furthermore, the average concentration of propofol (0.63 +/− 0.20 [micro sign]M, n = 9) was significantly lower than the predicted additive concentration (1.63 +/− 1.22 [micro sign]M, n = 9; P < 0.05, Student's t test), showing that propofol and midazolam interact synergistically to enhance GABA (1 [micro sign]M)-evoked currents.

The supraadditive effect of midazolam and propofol was confirmed in a subsequent experiment using the fixed-concentration method. We evaluated the effects of 0.62 [micro sign]M propofol, 0.6 [micro sign]M midazolam, and the combination of 0.62 [micro sign]M propofol and 0.6 [micro sign]M midazolam on currents activated by 1 [micro sign]M GABA. Current amplitudes were increased by factors of 2.0 +/− 0.04, 1.78 +/− 0.1, and 3.34 +/− 0.14 (n = 4). The predicted additive response (2.81 +/− 0.12) was significantly less than the observed response, confirming a synergistic interaction.

In this study, we evaluated the interactions between benzodiazepines and propofol for modulating GABA^Areceptors in embryonic mouse neurons. We observed an additive enhancement of GABA (3 [micro sign]M)-evoked currents by propofol (1 or 10 [micro sign]M) and midazolam (0.5 [micro sign]M) in hippocampal neurons. In contrast, a synergistic interaction between midazolam (0.5 [micro sign]M) and propofol (1 [micro sign]M) was apparent when the concentration of GABA was reduced from 3 [micro sign]M to 0.3 or 1 [micro sign]M. These observations are consistent with previous data that suggest that the potentiation of GABA-evoked currents by anesthetics depends critically on the concentration of agonist. [20]This concentration dependence of synergism may explain previous conflicting results. An additive interaction between diazepam (0.3 [micro sign]M) and propofol (0.3 to 3 [micro sign]M) was reported for enhancement of³H-GABA binding or muscimol-stimulated³⁶Cl^-uptake in brain preparations. [30]In contrast, propofol and flurazepam interacted synergistically to enhance GABA-activated current in Xenopus oocytes expressing human recombinant [small alpha, Greek]¹[small beta, Greek]²[small gamma, Greek]^2Lor [small alpha, Greek]²[small beta, Greek]²[small gamma, Greek]^2Lreceptor constructs. [35] 

The interactions between midazolam and propofol on GABA^Areceptor function are complex, and several possible mechanisms could account for the supraadditive increase in GABA-evoked current. Dissimilar to clinical studies in which pharmacokinetic or pharmacodynamic factors may contribute to the apparent synergism, in our study we assumed that the effective concentration of the drug at the receptor is known and is not influenced by the presence of the other drug. Thus, synergism can be attributed to direct modulation of GABA^Areceptor function or functional changes to other unidentified receptors or intracellular signaling pathways. Here we will only consider possible changes to the binding and gating properties of the GABA (^A) receptor.

Sedative-hypnotic compounds such as midazolam, flurazepam, and diazepam bind to the flumazenil-sensitive benzodiazepine site on the GABA^Areceptor. [13,14,26]These benzodiazepines increase the affinity of the receptor for GABA and, reciprocally, GABA increases the binding affinity for benzodiazepines. [36-39]Electrophysiologic studies indicate that benzodiazepines, such as diazepam, enhance the amplitude of whole-cell currents but do not directly activate the receptor. [36,40,41]Studies of single-channel currents indicate that the enhancement of the macroscopic current results from an increase in the frequency of channel opening. [29,42]Diazepam also increases single-channel conductance, although it is not known whether midazolam produces similar effects. [43]Conversely, propofol has at least three distinct effects on GABA^Areceptor function, including an increase in the frequency of channel opening, [29,42]activation of the receptor in the absence of GABA, and alteration of the kinetics of desensitization. [9-11,29]Propofol binds at a site that is functionally distinct from the benzodiazepine site, because flumazenil does not effect³⁵S-TBPS binding inhibition by propofol. Conversely, propofol does not displace³H-flunitrazepam from cerebral synaptosomes. Neither propofol-evoked currents nor the clinical effects of propofol are influenced by flumazenil. [11,27,34,36,37,42] 

Although the binding domains for propofol and midazolam are distinct, radiolabeling and electrophysiologic studies indicate that they are functionally coupled. Propofol stimulated³H-flunitrazepam binding to rat membranes. [44]Conversely, diazepam and midazolam increased the amplitude of propofol-induced currents. [11,45] 

An allosteric interaction, whereby propofol and midazolam increase the binding affinity for GABA, may account for the supraadditive interaction for current activated by 0.1 to 1 [micro sign]M but not by 3 [micro sign]M GABA observed in our study. Synergism was no longer apparent when receptors were activated by higher concentrations of GABA, possibly because of the saturation of binding sites. However, 3 [micro sign]M GABA is less than the EC⁵⁰value, and the amplitude of current, recorded in the presence or absence of drugs, was less than that activated by 600 [micro sign]M GABA. Therefore, we favor an explanation that attributes the synergistic interaction to direct activation of the receptor by propofol. Propofol-evoked currents are enhanced by midazolam, [45]therefore the supraadditive interaction between midazolam and propofol could result from midazolam-induced potentiation of the propofol-evoked current. This interaction may be occluded at higher concentrations of GABA, because propofol has a lower potency than GABA for activating the receptor.

Synergism between midazolam and propofol was observed only at low concentrations of GABA. The functional significance of this interaction to synaptic transmission is questionable because near-saturating concentrations of GABA are released from the synaptic terminal. [46]At most GABAergic synapses, high concentrations of transmitter rapidly activate current that peaks and then decays within the millisecond time range. However, an additional tonic inhibition can arise from persistent activation of receptors by low ambient concentrations of GABA. [47]Although the functional significance of this persistent inhibition remains to be elucidated, midazolam and propofol could enhance the tonic activation of GABA^Areceptors.

Schultz and MacDonald [48]proposed that “subtle” clinical actions of barbiturates observed at low drug concentrations (such as anticonvulsant and sedative effects) might be caused by an augmentation of GABA-mediated inhibition. In contrast, anesthetic actions observed at higher drug concentrations may result from direct activation of the receptor. Consistent with this suggestion, low, clinically relevant concentrations of midazolam and propofol interact synergistically to enhance GABA-evoked currents and also to induce hypnosis. [49-53]Higher concentrations of propofol and midazolam do not interact synergistically to enhance GABA-evoked current or to induce surgical anesthesia.

Various methods have been used to relate the in vitro effects of anesthetics to their behavioral actions. These include studies of optical isomers; rank-order potency; drug interactions, including synergism and antagonism; and, more recently, the use of genetically engineered animals. Here we evaluated the interactions between propofol and midazolam in the modulation of GABA^Areceptors to account for synergism observed for the induction of hypnosis. It has been postulated that a synergistic interaction implies that the two anesthetics act through distinct mechanisms. [5]Others [54]suggest that an additive interaction is consistent with the drugs acting at independent sites. Here we demonstrate an additive and synergistic interaction between propofol and midazolam for modulating a single target receptor. Thus, drug interactions cannot account for drugs acting at one or more sites. Furthermore, it was postulated that if different clinical actions of the anesthetic have the same underlying mechanisms, the drugs should achieve these components in a similar proportion. [5]However, our results indicate that even when both drugs influence the same receptor, the ratio of potency will not be the same for behavioral effects observed at high and low drug concentrations. The important clinical implication of our results is that the benefits of synergism, observed at one concentration ratio, may not be apparent at another. In this regard, the synergistic interaction between midazolam and propofol for hypnosis does not support the use of these drugs for clinical actions observed with higher drug concentrations, such as general anesthesia.

The authors thank Dr. P. Pennefather and Ms. P. Watson for their comments and review of the manuscript.