Abstract
Inhibitory GABAA receptor modulators are widely used therapeutic agents for a variety of central nervous system disorders. Ltk– cells stably expressing human recombinant GABAA subunits (α1β1–3γ2s) were seeded into 96-well plates, loaded with chlorocoumarin-2-dimyristoyl phosphatidylethanolamine and bis(1,3-diethyl-2-thiobarbiturate)trimethineoxonol, and rapid fluorescence resonance energy transfer technique (FRET) measurements were made of GABA-evoked depolarizations in low-Cl– buffer using a voltage/ion probe reader. The influence of different βsubunits on the ability of agents to modulate and directly activate the ion channel was examined. GABA evoked concentration-dependent decreases in FRET, increasing fluorescence emission ratio (460/580 nm) at α1β1γ2, α1β2γ2, and α1β3γ2 receptors with similar maximal amplitude (P > 0.05, n = 17) and EC50 values of 2.4 ± 0.2, 2.5 ± 0.2, and 1.3 ± 0.1 μM, respectively. Piperidine-4-sulfonic acid and 4,5,6,7-tetrahydroisoxazolo[5,4-c]pyridin-3-ol were less potent, with EC50 values of 8.7 ± 0.9, 9.2 ± 0.5, and 11.7 ± 1.2, and 43.7 ± 6.4, 24.8 ± 1.6, and 26.1 ± 2.4 μM, respectively. Potency and maximal efficacy of propofol, methyl 6,7-dimethoxy-4-ethyl-β-carboline-3-carboxylate, pentobarbital, and steroids, 5α-pregnan-3α-ol-20-one and 5β-pregnan-3α-ol-20-one, were unaffected by the β isoform present in the receptor complex. However, several compounds displayed β2/3 subunit selectivity, notably loreclezole, R(–)-etomidate, and a group of anti-inflammatory agents including mefenamic acid, flufenamic acid, meclofenamic acid, tolfenamic acid, niflumic acid, and diflunisal. The anti-inflammatories exhibited varying levels of efficacy at β2/3 subunits, with micromolar potency, while having antagonist or weak inverse agonist profiles at α1β1γ2. Diflunisal was the most efficacious compound, eliciting greater potentiation than loreclezole (90 ± 14% and 109 ± 14% at β3 and β2, respectively, compared with 62 ± 6% and 56 ± 3%), whereas niflumic acid exhibited the lowest efficacy. An additional agent, olsalazine, weakly potentiated responses at all three receptors without any selectivity. This study identifies and characterizes a variety of allosteric modulators for which βsubunits are an important determinant of efficacy and potency.
GABA is the predominant inhibitory neurotransmitter in mammalian central nervous system, and ligand-gated GABAA receptor ion channels are the target for a wide range of therapeutic agents including benzodiazepines, barbiturates, steroids, convulsants, and anesthetics (reviewed in Mehta and Ticku, 1999). GABAA receptors are pentameric assemblies of multiple subunits (α1-6, β1-3, γ1-3, δ, ϵ, π, and θ; Whiting et al., 1999) with an integral chloride channel and a variety of allosteric binding sites through which rapid inhibitory synaptic neurotransmission can be modulated. Subunit assembly and stoichiometry studies have revealed a variety of receptor subunit combinations that are differentially expressed in mammalian brain (Fritschy and Mohler, 1995; McKernan and Whiting, 1996), with the most abundant receptor subtypes generally having a 2α:2β:1γ stoichiometry (Farrar et al., 1999). Multiple subunits provide the potential for enormous structural diversity, although evidence suggests differential assembly signals within α and βsubunits operate to restrict the number of possible combinations (Bollan et al., 2002). The largest GABAA receptor population in rat brain has an α1β2γ2 composition, whereas α2β3γ2 and α3βγ2/γ3 together constitute the next most prevalent subtypes (McKernan and Whiting, 1996).
The precise array of subunits present in the receptor complex governs the pharmacological and functional properties of receptor subtypes. The GABA agonist binding site is thought to occur at the interface of α and βsubunits (Ebert et al., 1997), whereas affinity and efficacy at the well-characterized benzodiazepine site are influenced by residues in both α and γ (Wingrove et al., 1997) but not β (Hadingham et al., 1993) subunits. The βsubunit variant has been shown to influence loreclezole and etomidate sensitivity, with a single amino acid at the carboxyl-terminal end of the putative channel-lining domain TM2, being a key determinant of selectivity (Wingrove et al., 1994; Belleli et al., 1997). More recently, it has also been suggested that the βsubunit plays a dominant role in determining the ion selectivity of heteromeric GABAA receptors (Jensen et al., 2002).
Several groups have described regional localization of different βsubunits in rat brain (Wisden et al., 1992; Miralles et al., 1999; Pirker et al., 2000). All βsubunits are widely distributed throughout the brain, although β2 is the most abundant, especially in cerebellum and cortex, and immuno-reactivity is also highly concentrated in thalamic nuclei with the exception of the reticular nucleus. In hippocampus, β3 is the most abundant of the βsubunits, but some interneurons express high levels of β2, and there is some β1 expression on both dendrites and interneurons. There are high levels of β3 expression, with little β2, in corpus striatum and in the granule cells of the olfactory bulb, whereas in the cerebellum, β2 and β3 subunits predominate over β1.
Given the wide diversity of GABAA receptor subtypes, often in distinct neuronal circuits, and the variety of allosteric sites on the receptor through which ion channel activity can be subtly modulated, specificity is clearly key to the development of novel compounds for this therapeutic target. Progress has recently been made in dissecting functional roles of different GABAA receptors using knockout and point mutation strategies (Rudolph et al., 2001). Evidence to date supports the notion that subunit-selective modulators may elicit desirable clinical efficacy with reduced side effects, one such example being at the benzodiazepine site where it is possible to separate anxiolysis and sedation, each being mediated via different GABAA α-subunits (McKernan et al., 2000). Identification of such compounds that can discriminate between GABAA receptor subunits requires a functional assay with sufficiently high sensitivity to detect subtle modulation of ion channel activity.
In the present study, we have utilized a rapid ratiometric, voltage-sensitive fluorescence resonance energy transfer technique (FRET) (Gonzalez et al., 1999) to measure GABA-evoked changes in membrane potential. Using this approach, we have characterized the influence that different βsubunits exert on the pharmacological modulation of GABAA receptor ion channels.
Materials and Methods
Cell Culture. Combinations of human GABAA receptor subunits (α1βxγ2s) were stably expressed in mouse Ltk– fibroblast cells by transfection of the appropriate subunit cDNAs in vector pMSGneo using standard calcium phosphate transfection techniques. Cells, passaged weekly, were maintained at 37°C in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen, Carlsbad, CA), supplemented with 10% (v/v) FetalClone II serum (Hyclone) and 1 mg/ml geneticin. For experiments, cells were removed from stock plates with 0.05% trypsin/0.53 mM EDTA solution (Invitrogen) and resuspended in DMEM supplemented with 10% FetalClone II serum but without geneticin. Cells were seeded into black-sided Porvair 96-well microtiter plates at densities of 3 to 8 × 104 cells/ml in a volume of 200 μl/well and grown in the presence of 10% serum for 5 to 8 days in an incubator at 37°C. Receptor expression, which is under the control of a dexamethasone-sensitive promoter, was induced 24 h prior to experiment in the confluent cell monolayers using serum-containing DMEM supplemented with 1 μM dexamethasone (Sigma Chemical, Poole, Dorset, UK).
Fluorescence Measurement of Membrane Potential. Experiments were performed, as previously described (Adkins et al., 2001), in low-Cl– buffer (160 mM sodium-d-gluconate, 4.5 mM potassium-d-gluconate, 2 mM CaCl2, 1 mM MgCl2, 10 mM d-glucose, and 10 mM HEPES, pH 7.4). Cells were washed twice, leaving 35 μl of residual volume, and 65 μl of dyes were added to give final concentrations of 4 μM chlorocoumarin-2-dimyristoyl phosphatidylethanolamine (CC2-DMPE; FRET donor) and 1 μM bis(1,3-diethyl-2-thiobarbiturate)trimethineoxonol [DiSBAC2(3); FRET acceptor]. Cells were dye-loaded for 30 min at room temperature in darkness, washed again, and dye solutions were then added to give a final concentration of 1 μM DiSBAC2(3) with 0.5 mM tartrazine to quench extracellular fluorescence. Plate preparation was automated using a CCS Packard Platetrak.
Plates were then placed in a voltage/ion probe reader (Aurora Biosciences Corp., San Diego, CA), which performs automated additions using a Hamilton 2200 pipetter and records fluorescence emissions at 460 and 580 nm simultaneously from eight wells (Gonzalez et al., 1999). A 400DF15 filter was used in the excitation pathway and 460DF45 and 580DF60 filters in the respective emission pathways. Rapid ratiometric FRET measurements were made of GABA-evoked depolarizations in low-Cl– buffer, and the ability of compounds to modulate an EC50 response to GABA was examined. Experiments were performed using a two-addition protocol whereby basal fluorescence was read for 8 s before addition of test compounds, and a half-maximal concentration of GABA was added 22 s later, with fluorescence emissions recorded at 1 Hz. For assessment of the direct effect of compounds on channel opening, the initial addition was buffer only, and test compound alone was added as a second addition.
Data Analysis. Data were analyzed as previously described (Adkins et al., 2001). For each time point at each fluorescence emission wavelength, background fluorescence was subtracted (recorded from wells without cells in the same plate), and the ratio of fluorescence at 460 to 580 nm was calculated. GABA-evoked depolarizations were then expressed as a fractional change in this ratio, with the mean plateau ratio taken between 10 and 15 s after GABA addition normalized to the mean ratio over the first 6 s of recording before agonist addition. Algorithms written as Excel 97 (Microsoft, Redmond, WA) macros were used for automated calculations of fluorescence ratio, and GABA responses and concentration response curves were fitted to a four-parameter logistic equation using Prism (Prism version 2.0b; GraphPad Software Inc., San Diego, CA). Effects of compounds were expressed as percent change in the EC50 GABA response in the presence of compound versus the EC50 response alone, determined in control wells on the same plate.
Agonist stimulation and modulatory responses were assessed by performing one-way analysis of variance using Prism version 2.0b (GraphPad Software Inc.) followed by post hoc testing using Tukey's multiple comparison test to determine significance. Mean EC50, IC50, and efficacy values are presented as the arithmetic mean and S.E.M. from a number of independent and separate determinations (n), typically with at least four replicate test wells per concentration point on each plate tested.
Materials. CC2-DMPE was obtained from Aurora Biosciences Corp., DiSBAC2(3) was from Molecular Probes (Eugene, OR), and tartrazine was obtained from Sigma. GABA, 4,5,6,7-tetrahydroisoxazolo[5,4-c]pyridin-3-ol (THIP), piperidine-4-sulfonic acid (P4S), gluconate salts, and all other GABAA receptor modulators were also from Sigma. 2,6-diisopropylphenol (propofol) was from Aldrich, and etomidate was obtained from Janssen Pharmaceuticals (Antwerp, Belgium) as Hypnomidate, 2 mg/ml injectable solution. S(–)-Etomidate was synthesized in-house.
Results
GABA Site Ligands. Under low-chloride conditions, GABA evoked concentration-dependent depolarizations in cells loaded with CC2-DMPE and DisBac2(3) and expressing GABAA receptors of composition α1β1γ2s, α1β2γ2s, or α1β3γ2s. These depolarizations were rapidly transduced into decreased FRET and, therefore, an increase in the ratio of fluorescence emission at 460 nm (f460) to that at 580 nm (f580). The fluorescence emission ratio rose to a concentration-dependent plateau within 5 s and was sustained for more than 15 s (Fig. 1). Agonist responses at each receptor subtype were of similar maximal amplitude (P > 0.05; one-way analysis of variance followed by post hoc Tukey test), although GABA exhibited slightly higher potency at β3-containing receptors (P < 0.001) (Table 1) compared with β2 and β1.
Concentration-dependent changes in fluorescence emission ratio with time in response to GABA at α1β2γ2s receptors (data taken from a single experiment). Black bar, presence of GABA.
Potency values for GABA-stimulated depolarizations in low-chloride media in cells loaded with CC2-DMPE and DisBac2(3) and stably expressing a1bxg2s GABAA receptors
Results are mean ± S.E.M. from 17 separate determinations.
Other GABA site ligands also evoked concentration-dependent decreases in FRET. P4S and THIP exhibited lower potency than GABA and, in addition, P4S exhibited reduced efficacy compared with the maximum GABA response (P < 0.01; n = 4) (Table 2). THIP, however, did not exhibit significantly different efficacy compared with maximum GABA.
GABA site ligands, P4S and THIP, evoked concentration-dependent decreases in FRET
Results are mean ± S.E.M. from four separate determinations.
Influence of βSubunit on Modulation of the GABA Response. The ability of compounds to modulate the response to GABA was examined using an equieffective concentration of GABA (EC50), predetermined for each receptor subtype. As previously reported (Hadingham et al., 1993), benzodiazepine site ligands did not discriminate between receptor isoforms containing different βsubunits. At the maximal concentration tested (10 μM), the inverse agonist methyl 6,7-dimethoxy-4-ethyl-β-carboline-3-carboxylate negatively modulated GABA responses at α1β1γ2, α1β2γ2, and α1β3γ2 receptors, with maximal efficacies of –48 ± 4% (n = 5), –42 ± 6% (n = 5), and –42 ± 6% (n = 3), respectively (Fig. 2). In contrast, the anticonvulsant loreclezole exhibited subtype selectivity, with quantitatively greater maximal efficacy and potency at β2- and β3-versus β1-containing receptors (Fig. 2). There was little potentiation of α1β1γ2 receptors up to 10 μM (15 ± 4%, n = 7), whereas α1β2γ2 and α1β3γ2 receptor responses were potentiated by 56 ± 3% (n = 7) and 62 ± 6% (n = 4), respectively, with EC50 values of 1.9 and 0.6 μM. Neither compound elicited any direct activation of the receptor in the absence of GABA.
Cells were pretreated with different concentrations of the benzodiazepine site inverse agonist methyl 6,7-dimethoxy-4-ethyl-β-carboline-3-carboxylate (•) or the anticonvulsant loreclezole (○) before addition of GABA at a half-maximal concentration (EC50), previously established for each receptor subtype. Results (n = 3–7 determinations) are shown as mean percent modulation of the control EC50 response ± S.E.M. Loreclezole selectively potentiated β2- and β3-containing receptors. Neither compound elicited any direct effect on the ion channel alone (data not shown).
Modulation and Activation of α1βxγ2s GABAA Receptors by Anesthetics, Steroids, and Anti-Inflammatory Agents. The i.v. anesthetic propofol (Fig. 3A) and the non-volatile anesthetic barbiturate pentobarbital (Fig. 3B) potentiated the EC50 GABA response at all receptor subtypes. The maximal efficacies and potencies for propofol were not significantly different for β1-, β2-, and β3-containing receptors, being 123 ± 21, 132 ± 29, and 128 ± 24% and 3.8 ± 1.2, 2.4 ± 0.6, and 1.6 ± 0.1 μM, respectively (P > 0.05; n = 5). Pentobarbital was approximately an order of magnitude less potent than propofol, with EC50 values of 22.6 ± 2.3, 22.0 ± 1.9, and 13.4 ± 2.5 μM, and maximal efficacies of 133 ± 29, 168 ± 26, and 140 ± 41%, respectively (n = 3). Both propofol and pentobarbital directly activated the ion channel in the absence of GABA at concentrations approximately 10-fold higher than those eliciting potentiation of the GABA response (EC50 values of 27.2 ± 2.0, 25.3 ± 1.7, and 20.3 ± 3.0 μM, n = 6–9; and 119 ± 7.5, 117 ± 15.9, and 75 ± 3.0 μM, n = 3–6, respectively). For all compounds where a significant direct effect of the compound on the ion channel in the absence of GABA was observed, the modulation effects shown represent the total effect of the compound, given the absence of tools to block this direct component. Thus, for pentobarbital, at concentrations above 100 μM, a significant portion of the modulatory effect is likely to be due to a direct effect on the channel itself.
The potencies and efficacies for potentiation of the GABA EC50 response and direct channel gating by the i.v. anesthetic propofol (A) and the barbiturate pentobarbital (B) were not markedly affected by the βisoform present in the receptor complex. Propofol (n = 5 determinations) was an order of magnitude more potent than pentobarbital (n = 3 determinations) in potentiating GABA. Both compounds elicited direct channel activation at all subtypes at concentrations approximately 10-fold higher than those that potentiated GABA (n = 3 determinations each). The modulation graph represents the total effect of the compounds.
The anesthetic agent etomidate similarly both potentiated the EC50 GABA response and, at higher concentrations, caused direct activation (Fig. 4A). Like loreclezole, etomidate also displayed functional selectivity for β2/β3 subunits over β1, having both increased potency (EC50 values 0.5 ± 0.1 and 0.3 ± 0.1 μM, respectively) and higher maximal efficacy (152 ± 20 and 158 ± 23% potentiation, respectively) than at α1β1γ2 receptors (EC50 2.9 ± 0.8 μM; efficacy 107 ± 19%; n = 4). Etomidate evoked a higher degree of potentiation of the GABA response than loreclezole at all three subunit combinations examined. Unlike loreclezole, etomidate elicited direct effects on GABAA receptors, and this direct activation of the ion channel at higher concentrations also displayed βsubunit selectivity (n = 6). Etomidate exhibited enantiomeric selectivity, with the S(+)-enantiomer displaying reduced potency and no βsubunit selectivity (Fig. 4B), in contrast to the R(–)-enantiomer (Fig. 4A). The S(+)-enantiomer could not be tested at high enough concentrations to determine whether there was any direct activation.
The anesthetic agent R(–)-etomidate exhibited higher potency and efficacy for potentiation of GABA responses at receptors containing a β2 or a β3 subunit (mean ± S.E.M. from four independent experiments) and also elicited direct ion channel activation at higher concentrations (n = 6), potency and efficacy being reduced at β1-containing receptors (A); etomidate also exhibited stereoselectivity with the S(+)-enantiomer displaying lower potency and no subunit selectivity (B).
The steroid pharmacology of these GABAA receptor isoforms was also investigated. The progesterone metabolite 5α-pregnan-3α-ol-20-one, together with 5α-pregnan-3β-ol-20-one and 5β-pregnan-3α-ol-20-one, were all tested for their ability to modulate ion channel opening (Fig. 5). 5α-Pregnan-3α-ol-20-one robustly potentiated GABA responses and did not discriminate between different βsubunits, eliciting quantitatively similar maximal efficacies at all receptors with little difference in potency (Fig. 5a). In contrast, 5α-pregnan-3β-ol-20-one had no significant effect on the GABA response except at the top concentration (200 μM) where there was a modest inhibition (β1, –18 ± 4%; β2, –38 ± 7%; β3, –35 ± 8%; n = 9) (Fig. 5b). 5β-Pregnan-3α-ol-20-one elicited potentiation of all three receptor subtypes with similar potency and maximal efficacy (Fig. 5c).
Modulation of GABA EC50 responses at α1βxγ2s GABAA receptors in response to steroids. a, response to 5α-pregnan-3α-ol-20-one (n = 3 determinations. b, 5α-pregnan-3β-ol-20-one (n = 3 determinations). c, 5β-pregnan-3α-ol-20-one (n = 4 determinations).
A group of nonsteroidal anti-inflammatory agents were also examined for modulatory and direct effects on GABAA receptors (chemical structures shown in Fig. 6). The aminoarylcarboxylic acid derivatives mefenamic acid, flufenamic acid, meclofenamic acid, tolfenamic acid, and niflumic acid and the salicylic acid derivatives diflunisal and olsalazine all showed activity at GABAA receptors. Except for olsalazine, the anti-inflammatories all selectively potentiated α1β2γ2s and α1β3γ2s receptors, exhibiting varying levels of efficacy at β2/3 subunits, with micromolar potency, while having antagonist or weak inverse agonist profiles at α1β1γ2. The ability of compounds to modulate the EC50 response to GABA is shown in Table 3. Diflunisal was the most efficacious of these compounds, eliciting greater potentiation than loreclezole (90 ± 14 and 109 ± 14% at β3 and β2, respectively, compared with 62 ± 6 and 56 ± 3%), although it was an order of magnitude less potent. Niflumic acid exhibited the lowest efficacy. The aminoarylcarboxylic acid derivatives were more potent modulators than the salicylic acid derivatives. An additional agent, olsalazine, weakly potentiated responses at all three receptors without any selectivity and elicited some direct activation at the top concentration tested (200 μM). There was no direct receptor activation observed up to 200 μM with niflumic acid or flufenamic acid, a very modest direct effect at the top concentration with mefenamic acid and diflunisal, whereas a more robust direct effect was observed for meclofenamic (Fig. 7) and tolfenamic (Fig. 8) acids.
Structures of loreclezole, etomidate, and anti-inflammatory agents evaluated as GABAA receptor modulators.
The ability of nonsteroidal anti-inflammatory agents to modulate GABA responses at different receptor subtypes (etomidate and loreclezole are shown for comparison)
An equieffective concentration of GABA (EC50) was predetermined for each receptor subtype prior to modulation experiments. Results are the mean ± S.E.M. from 5 to 10 separate determinations and represent the total effect of the compound (modulatory and direct).
The nonsteroidal anti-inflammatory agent meclofenamic acid selectively potentiated α1β2γ2s and α1β3γ2s receptors with similar potency and similar maximal efficacy, while displaying inverse agonism at β1 (n = 7 determinations). There was some direct activation with this compound at concentrations approximately 10-fold higher than those at which potentiation was observed (n = 4 determinations).
A related anti-inflammatory agent, tolfenamic acid, also selectively potentiated GABA responses at α1β2γ2s and α1β3γ2s receptors in a concentration-dependent manner (n = 9 determinations) and elicited direct activation at the top concentration tested (200 μM) (n = 4 determinations).
Discussion
In this study, we have performed the most comprehensive study yet undertaken of the effect of βsubunit expression on the activation and modulation of GABAA receptor subtypes. We demonstrate the importance of βsubunits in influencing the action of a variety of agents, highlighted by the marked differences in efficacy and potency of loreclezole, etomidate, and multiple nonsteroidal anti-inflammatory agents, the GABAA receptor selectivity of some of which are reported here for the first time. An earlier study examining GABA-evoked currents recorded in oocytes expressing rat cortex mRNA (Woodward et al., 1994) was the first to suggest that some anti-inflammatory agents had the potential to modulate GABAA receptors, although the precise receptor subunit composition was unclear, and there was no information about receptor selectivity. Many of the compounds displaying subunit selectivity are used clinically, and it will be interesting to explore further whether differences in their efficacy profiles underlie some of their distinct physiological effects.
Evidence suggests that nonsteroidal anti-inflammatories can effectively cross the blood-brain barrier (Bannwarth et al., 1989). This, together with the relatively high therapeutic plasma concentrations that are achieved in the region of 80 μM and 20 to 70 μM for mefenamic and niflumic acids, respectively (Halliwell et al., 1999; Sinkkonen et al., 2003), suggests that appreciable brain levels might be achieved. Indeed, observations of antiepileptogenic effects (Wallenstein, 1991) and adverse events associated with anti-inflammatory overdose are consistent with activity at central GABAA receptors. Given that low micromolar concentrations are sufficient to potentiate β2/3-containing GABAA receptors, this suggests that modulatory effects could occur at clinically relevant concentrations. In addition, the anti-inflammatory-sensitive α1β2γ2 receptor subtype is the largest GABAA receptor population in mammalian brain (McKernan and Whiting, 1996). As well as activity at GABAA receptor ion channels, fenamate nonsteroidal anti-inflammatories have also been reported to have activity at a variety of other ion channels (e.g., Greenwood and Large, 1995; Malykhina et al., 2002).
Application of FRET techniques to the study of GABAA receptors has been previously described (Adkins et al., 2001). Clearly, in terms of agonist time course, these FRET responses have a slower onset than those determined by more direct electrophysiological measures of channel opening, with peak currents attained within 5 s versus millisecond responses using patch-clamp techniques. Although the GABA response is likely to occur rapidly, changes in FRET are downstream of this and, thus, are a slower readout of channel function. The current technique is influenced by the speed of voltage-dependent redistribution DiSBAC2(3) in the membrane, which occurs subsequent to the flow of chloride ions and membrane depolarization. However, potency values obtained using voltage/ion probe reader fluorescence measurements for GABA-evoked depolarizations and allosteric modulation at a variety of GABAA receptors agree well with values reported from other assays including electrophysiology (e.g., Wafford et al., 1994; Smith et al., 2001). In addition, this technique offers the potential for much increased throughput, allowing scope to more widely explore the pharmacology of different recombinant receptors.
Recent studies of murine β3 homomeric GABAA ion channels indicate these GABA-insensitive receptors can still respond to pentobarbital, propofol, etomidate, and pregnanolone (Wooltorton et al., 1997), although the properties of murine β1 homomers are somewhat different. This appears to be due to a single amino acid, asparagine-265, at the extracellular end of M2 (Cestari et al., 2000), the same residue that regulates potentiation by etomidate (Belleli et al., 1997) but not pentobarbital in heteromeric receptors. This amino acid has also been implicated in the action of loreclezole (Wingrove et al., 1994) and mediates potentiation by high concentrations of β-carbolines (Stevenson et al., 1995). Although the presence of a β2 or β3 subunit is clearly a prerequisite for loreclezole efficacy, different α or γ subunits reportedly influence the level of potentiation (Wafford et al., 1994). In a detailed study of one particular anti-inflammatory agent at different recombinant receptors (Halliwell et al., 1999), the same residue was also implicated in βsubunit-selective modulation of GABAA receptors by mefenamic acid. Additionally, Sinkkonen et al. (2003) reported that potentiation of α1β2γ2 receptors by niflumic acid depended on the presence of a γ2 subunit, also found to influence mefenamic acid modulation (Halliwell et al., 1999). Antagonism of the α6β2γ2 receptor subtype by niflumic acid has also been reported (Sinkkonen et al., 2003), and substitution of an α4 subunit reduced mefenamic potentiation of α1β2γ2 receptors by 50% (Whittemore et al., 1996).
Pentobarbital, propofol, and etomidate all elicited quantitatively larger potentiations of the GABA response than loreclezole, anti-inflammatories, and steroids. Although the nature of the βsubunit is clearly critical for both potentiation and GABA-mimetic activity of etomidate, with lower potency and efficacy evident for β1-containing receptors, the actions of propofol, pentobarbital, and steroids were found in our study to be similar irrespective of the βisoform present in the receptor complex. However, evidence for direct involvement of βsubunit residues in the action of propofol comes from a study showing that methionine 286 in transmembrane domain 3 of the βsubunit controlled a binding cavity for propofol (Krasowski et al., 2001) and was implicated in potentiation of GABA responses but not in direct activation. This may imply the two phenomena are mediated via separate sites, a view supported by Siegwart et al. (2002), who observed differential structural requirements for modulatory and direct effects of a variety of general anesthetics. There is also evidence that a glycine residue at the entrance to TM1 of the β2 subunit is involved in channel gating and anesthetic modulation (Carlson et al., 2000). In the current study, pentobarbital, etomidate, and propofol all had pronounced effects on direct activation at approximately 10-fold higher concentrations than those which potentiated the GABA response.
With regard to pentobarbital, the affinity of potentiation electrophysiologically in oocytes was little affected by altering α or βsubunits, although the maximal degree of potentiation was influenced by the nature of the former but not the latter (Thompson et al., 1996), in agreement with the present results. With regard to direct activation by pentobarbital, the α subunit influenced both the affinity and efficacy of the direct effect, whereas the nature of the βsubunit exerted a lesser influence on these parameters, only showing differences when coexpressed with α1 but not with α6. This is in contrast to the current observations and may reflect cell-type-specific differences. Although the βsubunit appears to have little impact on the action of steroids, it clearly plays a role in mediating their effects since modulation is preserved even at homomeric βsubunit-containing receptors (Wooltorton et al., 1997).
Many compounds examined in the present study showed evidence of declining responses at high concentrations (e.g., Figs. 3, a and b, 4a, 7, and 8), and this was evident both for modulatory and direct effects. Similar observations have been made in electrophysiological studies of the action of pentobarbital, etomidate, and mefenamic acid (Thompson et al., 1996; Belleli et al., 1997; Halliwell et al., 1999). Previous studies have identified several different effects with respect to the action of anesthetic agents: a potentiation of the GABA response, direct activation of the ion channel, and, at higher concentrations, a block of the GABA chloride channel. It is likely, at higher concentrations of test agents, that multiple processes may be occurring simultaneously that may limit the amount of potentiation that can be observed.
Etomidate showed stereoselectivity in potentiating GABA responses, with the S(–)-enantiomer exhibiting lower potency and no βsubunit selectivity, unlike the R(+)-enantiomer. Stereoselectivity has also been reported for steroids with multiple chiral centers (Covey et al., 2000). The latter study indicates that 5α-reduced steroids show a much higher degree of enantiospecificity in their actions as modulators and as anesthetics than 5β-reduced steroids. The limited effect here of 5α-pregnan-3β-ol-20-one agrees with previous published data describing the stereoselectivity of the 3-hydroxy group, which must be in the α configuration for potent receptor modulation (El-Etr et al., 1998; Hawkinson et al., 1998).
An additional compound which is a derivative of salicylic acid, salicylidene salicylhydrazide, has recently been reported by us to also display βsubunit selectivity (Thompson et al., 2004), being a selective inhibitor of β1-containing receptors. This further supports the finding that βsubunits are an important determinant of efficacy and potency for a variety of allosteric modulators and direct activators of human GABAA receptor ion channels. In the current study, we have used a novel fluorescence technique that provides rapid and sensitive measurements of GABAA receptor function to examine the influence of differing βsubunit expression on the potency and efficacy of a variety of ligands. Several compounds discriminated β2/β3 from β1-containing receptors including the anticonvulsant loreclezole, the anesthetic etomidate, and a relatively diverse group of anti-inflammatory agents.
Acknowledgments
We would like to extend our thanks to Matthew Leveridge and Julie Kerby, who provided cell culture support for this work.
Footnotes
-
doi:10.1124/jpet.104.070342.
-
ABBREVIATIONS: FRET, fluorescence resonance energy transfer; DMEM, Dulbecco's modified Eagle's medium; CC2-DMPE, chlorocoumarin-2-dimyristoyl phosphatidylethanolamine; DiSBAC2(3), bis(1,3-diethyl-2-thiobarbiturate)trimethineoxonol; THIP, 4,5,6,7-tetrahydroisoxazolo[5,4-c]pyridin-3-ol; P4S, piperidine-4-sulfonic acid.
- Received April 20, 2004.
- Accepted June 10, 2004.
- The American Society for Pharmacology and Experimental Therapeutics