Studies using mice with point mutations of GABAA receptor α subunits suggest that the sedative and anxiolytic properties of 1,4-benzodiazepines are mediated, respectively, by GABAA receptors bearing the α1 and α2 subunits. This hypothesis predicts that a compound with high efficacy at GABAA receptors containing the α1 subunit would produce sedation, whereas an agonist acting at α2 subunit-containing receptors (with low or null efficacy at α1-containing receptors) would be anxioselective. Electrophysiological studies using recombinant GABAA receptors expressed in Xenopus oocytes indicate that maximal potentiation of GABA-stimulated currents by the pyrazolo-[1,5-a]-pyrimidine, DOV 51892, at α1β2γ2S constructs of the GABAA receptor was significantly higher (148%) than diazepam. In contrast, DOV 51892 was considerably less efficacious and/or potent than diazepam in enhancing GABA-stimulated currents mediated by constructs containing α2, α3, or α5 subunits. In vivo, DOV 51892 increased punished responding in the Vogel conflict test, an effect blocked by flumazenil, and increased the percentage of time spent in the open arms of the elevated plus-maze. However, DOV 51892 had no consistent effects on motor function or muscle relaxation at doses more than 1 order of magnitude greater than the minimal effective anxiolytic dose. Although the mutant mouse data predict that the high-efficacy potentiation of GABAA1a receptor-mediated currents by DOV 51892 would be sedating, behavioral studies demonstrate that DOV 51892 is anxioselective, indicating that GABA potentiation mediated by α1 subunit-containing GABAA receptors may be neither the sole mechanism nor highly predictive of the sedative properties of benzodiazepine recognition site modulators.
GABAA receptors are heteropentamers consisting of protein subunits from at least eight homologous families assembled into a chloride-permeable ion channel (Barnard et al., 1998; Korpi et al., 2002). GABAA receptors are most commonly formed from α, β, and γ subunits arranged as a γ-β-α-β-α pentamer (Minier and Sigel, 2004). The principal (∼60%) GABAA receptor isoform in adult mammalian brain consists of α1, β2, and γ2 subunits, with GABAA receptors containing α2 or α3 subunits constituting an additional 10 to 20% of the total population (McKernan and Whiting, 1996). Binding of GABA to its receptor in the complex opens the ionophore, increasing chloride conductance, which leads to neuronal inhibition (Hamill et al., 1983). GABAA receptors also contain recognition sites for a number of pharmacologically important compounds that can either directly activate or modulate the function of this family of ligand-gated ion channels (Rudolph et al., 2001; Korpi et al., 2002; Whiting, 2003).
Perhaps the most extensively studied of these recognition sites is the “benzodiazepine receptor” (Choi et al., 1977; Möhler and Okada, 1977; Squires and Braestrup, 1977), which resides at the interface of the α-γ subunits (Minier and Sigel, 2004). Benzodiazepines remain widely used for the treatment of anxiety (Stahl, 2002), and also act as anticonvulsants, hypnotics, myorelaxants, and amnestic agents. Although this range of pharmacological properties results from the allosteric enhancement of the actions of GABA at its receptor, the roles of different GABAA receptor isoforms in mediating these discrete effects of benzodiazepines are not fully resolved (Basile et al., 2004). A better understanding of the GABAA receptor(s) mediating each of the pharmacological properties of benzodiazepines may facilitate the development of novel compounds devoid of undesirable side effects typically associated with the use of 1,4-benzodiazepines in treating anxiety disorders (e.g., ataxia, muscle relaxation, and sedation).
Data obtained from mice with point mutations of GABAA receptor α subunits have contributed to the concept that individual GABAA receptor subtypes mediate specific pharmacological actions of benzodiazepines. Thus, mice with point mutations of the α1 subunit (Rudolph et al., 1999) rendering the receptor insensitive to benzodiazepines (i.e., “knockins”) show a substantial reduction in diazepam-induced sedation (Rudolph et al., 1999; McKernan et al., 2000). Furthermore, diazepam is inactive in the elevated plus-maze (a test predictive of anxiolytic activity) in α2 knockin mice (Low et al., 2000). These results indicate that compounds selective for GABAA receptors containing α2 subunits (or compounds with little or no efficacy at other GABAA receptor isoforms) will exhibit anxiolytic properties, with a reduction in the sedative characteristics presumptively mediated by GABAA receptors containing α1 subunits (Atack, 2003; Basile et al., 2004). Although α2 subunit-selective ligands exhibit an “anxioselective” profile in preclinical studies [e.g., SL-651498 (Griebel et al., 2003) and L-838417 (McKernan et al., 2000)], the results of other pharmacological studies do not support the conclusions reached using these mutant mice. Thus, TP-003, a selective, positive modulator of α3 subunit-containing GABAA receptors produces a robust, anxiolytic-like effect in rodent and nonhuman primate models of anxiety (Dias et al., 2005). Furthermore, 2-pyridinyl[7-(4-pyridinyl)pyrazolo[1,5-a]pyrimidin-3-yl]methanone (ocinaplon) exhibits an anxio-selective profile in animals and produces an anxiolytic effect in patients with generalized anxiety disorder in the absence of benzodiazepine-like side effects such as sedation and dizziness (Lippa et al., 2005). This pyrazolopyrimidine demonstrates the highest potency and efficacy in enhancing GABA-mediated currents in Xenopus oocytes expressing GABAA1 receptors composed of α1, β2, and γ2 subunits, compared with GABA receptors with α2, α3, and α5 subunits (Lippa et al., 2005).
Although ocinaplon is most active at GABAA1 receptors, it is only ≈80% as efficacious as diazepam. One possible explanation for the anxioselective actions of ocinaplon compatible with the knockin mouse findings posits the requirement for high efficacy (equivalent to a benzodiazepine such as diazepam) of a drug at α1 subunit-containing GABAA receptors to produce sedation but a much lower efficacy at α2 subunits to reduce anxiety (Lippa et al., 2005). Thus, the degree of potentiation of GABA-gated Cl– currents mediated by α1 subunit-containing GABAA receptors must be equivalent or greater than those elicited by diazepam to elicit sedation. A corollary of this hypothesis is that an agent that potentiates GABAA1a receptor activation with greater efficacy than diazepam may express more sedative than anxiolytic activity, as exemplified by zolpidem (Graham et al., 1996). We now describe the pharmacological properties of DOV 51892, a structural analog of ocinaplon, which permits a test of the hypothesis that a highly efficacious, positive modulator of GABAA receptors containing α1 subunits will possess primarily sedative actions.
Materials and Methods
Male Sprague-Dawley (behavior) or Wistar (radioligand binding) rats weighing 200 to 300 g were used. The rats were group-housed (five/cage) in standard laboratory cages and kept in a temperature-controlled colony room (21 ± 2°C) with a 12-h light/dark cycle (lights on, 7:00 AM; lights off, 7:00 PM) according to the Association for the Assessment and Accreditation of Laboratory Animal Care guidelines. Commercial food and tap water were available ad libitum. Female, oocyte-positive Xenopus laevis frogs were purchased from either Xenopus One (Ann Arbor, MI) or Nasco (Fort Atkinson, WI).
Radioligand Binding Assays
A modification of previously described techniques was used to determine the affinity of DOV 51892 for the benzodiazepine recognition site (Liu et al., 1996; Skolnick et al.,1997). Rat cerebella were used as the receptor source and were obtained under Association for the Assessment and Accreditation of Laboratory Animal Care guidelines. Cerebellar membranes were repeatedly washed and aliquots (∼0.1 mg of protein) added to tubes containing either DOV 51892 (final concentrations 312.5 nM–20 μM) or diazepam (100 fM–100 μM). All tubes received an aliquot of [3H]flumazenil (Ro 15-1788; 78.6 Ci/mmol; 1 nM final concentration; PerkinElmer Life and Analytical Sciences, Boston, MA) and sufficient Tris citrate buffer to yield a final volume of 0.3 ml. Nonspecific binding was determined in the presence of 10 μM diazepam. The assay was incubated for 2 h at 0–4°C and terminated by vacuum filtration. The IC50 values for radioligand displacement by unlabeled competing agents were determined using nonlinear regression techniques (Prism; GraphPad Software Inc., San Diego, CA).
Recording of GABA-Gated Currents from GABAA Receptor Constructs in Xenopus Oocytes
cRNAs encoding GABAA receptor α1, α2, α3, or α5, β2, and γ2S subunits were injected into oocytes from X. laevis. Forty-eight hours later, measurements of the effects of DOV 51892 and diazepam on GABA-gated Cl– currents from oocytes expressing GABAA receptors were performed using a Warner two-electrode voltage-clamp amplifier (Warner Instruments, Inc., Foster City, CA) (Park-Chung et al., 1999). Microelectrodes of 1 to 3 MΩ when filled with 3 M KCl were used to record from oocytes in a recording chamber continuously perfused with ND-96 buffer solution containing 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl2 · 2H2O, 1 mM MgCl2 · 6H2O, and 5 mM HEPES adjusted to pH 7.4 with 5 mM NaOH. Oocytes were clamped at a holding potential of –70 mV during data acquisition. Drugs were applied by perfusion at a rate of approximately 50 μls–1 for 10 to 20 s followed by a 90-s wash. All experiments were performed at room temperature (22–24°C).
Data acquisition and external perfusion were controlled by computer, and dose-response data from each oocyte were fitted to the Hill equation by nonlinear regression techniques. Based on the GABA concentration-response curve fit, an EC10 for GABA was determined for each subunit combination, and this concentration was used for subsequent modulator concentration-response studies. Peak current measurements were normalized and expressed as a fraction of the peak control current measurements. Control current responses to an EC10 concentration of GABA were redetermined after every two to four modulator applications.
The anxiolytic and sedative actions of DOV 51892 and diazepam (Sequoia Research, Pangbourne, UK) were determined using a battery of behavioral assays. In all cases, animals were orally administered DOV 51892, diazepam, or vehicle (0.5% methylcellulose containing 0.1% Tween 80; 2 ml/kg), and their effects were evaluated after 1 h. The sequence of testing after administration of the test agents was as follows: 1 h after administration of the test agents, the animals were tested in the open field for 3 min and then placed into the elevated plus-maze for 5 min. After completion of the elevated plus-maze, grip strength was tested in three trials lasting less than a minute, followed by 2 min on the rotarod. All of these studies were completed in 1 day.
Tests Predictive of Anxiolysis. The Vogel conflict test was used to determine the anxiolytic activities of the test substances (Vogel et al., 1971). Rats were deprived of water for approximately 16 h before acclimation in the test chamber. Only rats that drink in the apparatus following a 2-day selection process were used in subsequent testing. On the third day, water-deprived rats were placed in the test chamber 1 h after drug treatment and given free access to the drinking tube. After 20 licks, a 2-s, 0.2-mA, two-phase shock was delivered to the metal drinking tube, and it was repeated after every 20 licks. Each session lasted for 3 min from the delivery of the first shock, and the number of punished responses was recorded. To control for any effects of the test compound on thirst, a group of rats were provided with access to water in the test chamber, but they received no electrical stimulus. These nonpunished responses were then recorded. The ability of the benzodiazepine binding site antagonist flumazenil (5 mg/kg i.p.) to reverse the anxiolytic effects of DOV 51892 in the Vogel conflict test was also tested.
A second test, the elevated plus-maze (Pellow et al., 1985) was also used to assess the anxiolytic-like activity of diazepam and DOV 51892. The test was initiated by placing a rat on the central platform of the maze facing an open arm. Testing lasted for 5 min, and the time spent in various sections of the maze, including the central platform, was recorded. The percentage of time spent in the open arms of the maze [time spent in open arm/(total duration of the session – time spent on the neutral platform) × 100] and the total number of arm crosses are presented.
Tests of Sedation, Ataxia, and Myorelaxation. The sedative, ataxic, and myorelaxant effects of DOV 51892 and diazepam were determined using the open field, rotarod, and grip strength meter (Kelley, 1993), although the open field also provides information predictive of anxiolytic activity (Crawley, 1985).
In the open-field test, rats were placed in the corner of a dimly lit (40-lux) open field of black plywood (66 × 56 × 30 cm), and the distance traveled over a 3-min period was measured using the Anymaze tracking system (Stoelting Co., Wood Dale, IL). The presence of ataxia was assessed using a rotarod apparatus (ENV-577; MED Associates, St. Albans, VT). The degree of myorelaxation was determined next using a grip strength meter (Columbus Instruments, Columbus, OH). The forepaws of a rat were placed on the metal mesh attached to the meter, and the body was gently pulled until the rat released the grid. Three measures of grip strength were taken sequentially for each rat, averaged, and corrected for body weight. Finally, rats were placed on a rod rotating at 6 rpm. Those animals that did not fall off the apparatus within 2 min were considered to have normal balance and coordination.
To minimize the potential interference of nonspecific arousal induced by animal handling, the test “battery” was always performed in the following order, from minimally to maximally arousing (open field, elevated plus-maze, grip strength, and rotarod tests) and typically took less than 15 min per subject.
Preparation of DOV 51892. DOV 51892 was synthesized from ocinaplon in two steps, using 3-chloroperoxybenzoic acid to form the intermediate pyridine N-oxide.
A solution of 20.0 g (0.066 mol) of 2-pyridinyl[7-(4-pyridinyl)pyrazolo[1,5-a]pyrimidin-3-yl]methanone in 1 liter of methylene chloride was stirred at room temperature and treated with 19.6 g (75%) of 3-chloroperoxybenzoic acid for 20 h. The resulting precipitate was collected by filtration, washed with CH2Cl2 (30 ml × 3) and dried in vacuo. The crude product was slurried in Na2CO3 solution (13.5 g in 300 ml of water) at room temperature for 2 h and then filtered, washed with water (50 ml × 2), and dried in vacuo at 70°C to yield 11.2 g of yellow powder, which is further purified by flash chromatography on silica gel eluting with CHCl3/MeOH (98:2). Fractions containing the desired product were collected and evaporated to dryness yielding 7.6 g (0.024 mol; 28.8%) of 2-pyridinyl[7-(4-pyridinyl)pyrazolo[1,5-a]pyrimidin-3-yl]methanone N7-oxide with >95% purity. In the second step, the pyridine N-oxide was converted to the 2-chloropyridinyl-analog (DOV 51892) by treatment with phosphorous oxychloride followed by alkalinization with K2CO3 and extraction with chloroform.
The 2-pyridinyl[7-(2-chloropyridin-4-yl)pyrazolo[1,5-a]pyrimidin-3yl]methanone was synthesized by adding 8.0 g (0.025 mol) of 2-pyridinyl[7-(4-pyridinyl)pyrazolo[1,5-a]pyrimidin-3-yl]methanone N7-oxide to 100 ml of phosphorous oxychloride at room temperature. The reaction mixture was heated with stirring in an oil bath to 110–120°C for 8 h and concentrated in vacuo. To the dark brown residue was added crushed ice and solid potassium carbonate. The alkaline solution was extracted with chloroform (50 ml × 3). The combined organic layer was washed with water (20 ml × 2), dried with sodium sulfate, and evaporated in vacuo. The brown residue was purified by column chromatography on silica gel, eluting with chloroform and then chloroform/methano1 (99:1). Thin layer chromatography (chloroform/methanol; 9:1) was used to monitor the purification. Fractions containing the desired product were collected and evaporated to dryness to give 2.2 g (0.0069 mol; 27.2% yield) of 2-pyridinyl-[7-(2-chloropyridin-4-yl)pyrazolo[1,5-a]pyrimidin-3-yl]methanone as a pale yellow powder in 98.7% purity. M/e+ 336, 1H NMR (CDCl3) 7.22 (1H, d), 7.53 (1H, m), 7.92 (2H, m), 8.05 (1H, s), 8.27 (1H, d), 8.68 (1H, d), 8.77 (1H, d), 8.955 (1H, d), 9.44 (1H, s).
Preparation of [D3]-DOV 51892. [D3]-DOV 51892 was synthesized from [D4]-ocinaplon using techniques similar to those described above. [D4]-Ocinaplon was synthesized using the following steps. [D4]-methyl isonicotinate. Methanol (1.5 liters) was added to 5.3 g of [D4]-isonicotinic acid (0.41 mol). Hydrogen chloride gas was bubbled and the mixture was heated under reflux for 48 h. The solvent was evaporated, and water and ether were added to the solid. Sodium bicarbonate was added to pH 9, and then ether extraction and purification by distillation were performed.
[D4]-4-acetylpyridine. Lithium hydride (1.2 g) was added to a solution containing 14 g of [D4]-methyl isonicotinate in 17 g of ethyl acetate, which was then heated to 95°C for 18 h. Water was added, and the solid was filtered. An additional 90 ml of water was added to the solid followed by 35 ml of concentrated HCl. The mixture was then heated under reflux for 2 h. Then, 10 N NaOH was added to pH 5, and potassium carbonate was added to pH 10. The solution was ether extracted, and the product was purified by distillation.
[D4]-pyridine enamine. Dimethylformamide dimethylacetal (8.5 g) was added to [D4]-4-acetylpyridine, and the mixture was heated to 70°C for 24 h. After the solution cooled, ether was added and the solid was filtered.
[D4]-ocinaplon. Two hundred milliliters of acetic acid and 4.2 g of aminopyrazole were added to 4.6 g of [D4]-pyridine enamine. The mixture was heated under reflux for 24 h, and then the solvent was evaporated. The final product was crystallized using isopropanol and was used to make [D3]-DOV-51892.
Determination of DOV 51892 Concentrations. Male Sprague-Dawley rats were dosed orally with 6 mg/kg DOV 51892 suspended in 0.5% methylcellulose/0.1% polysorbate 80 (10 ml/kg). One hour after dosing, the rats were euthanized, and samples of whole blood and brain were collected. Plasma was isolated from the blood and stored at –70°C, as was the brain. Rat brains were homogenized using a hand-held homogenizer in 9 volumes of acetonitrile/water [1:9 (v/v)] containing 1% formic acid. The homogenates were processed using the same experimental procedures as described for the plasma samples.
To make the standard curve, blank plasma and brain homogenates were spiked with DOV 51892 to give concentrations from 2.5 to 2000 ng/ml. Samples were extracted using C8 solid phase extraction. Aliquots (0.10 ml) of standard, quality control, and samples were diluted with equal volumes of the [D3]-DOV 51892 internal standard and transferred to prewetted solid phase extraction cartridges. After washing with deionized water, the analytes were eluted with 0.8 ml of acetonitrile. The solvent was evaporated to dryness under vacuum, and the samples were reconstituted in 50 μl of mobile phase and mixed by vortexing. The reconstituted samples were analyzed by a liquid chromatography/tandem mass spectrometric assay using a Restek UltraCyano column (100 × 2.1 mm; 5-μm particle size). The mass transitions for quantitation were m/z 336 → 78 and 338 → 78 for DOV 51892 and 339 → 78 for the internal standard. The lower limit of quantitation was 2.5 ng of DOV 51892/ml.
Statistics. The significance of the differences in behavioral responses of rats treated with the test substances and vehicle was determined using one-way analysis of variance with Dunnett's post hoc test.
Dose-effect curves for potentiation of the GABA-induced current were fitted individually for each oocyte by nonlinear least-squares regression using the logistic equation, and parameter values were averaged. EC50 values were log-parameterized for fitting, averaging, and determination of 95% confidence limits. Confidence limits were calculated using the t-distribution.
DOV 51892 inhibited [3H]flumazenil binding to cerebellar membranes with an IC50 of 1.5 ± 0.16 μM, whereas the archetypical benzodiazepine diazepam inhibited radioligand binding with much higher affinity (IC50 = 5.8 ± 0.83 nM). Both substances fully inhibited [3H]flumazenil binding (i.e., Imax = 100%; data not shown).
Electrophysiological investigations on GABA-gated currents were performed using Xenopus oocytes expressing recombinant GABAA receptor constructs differing in α subunit composition. Consistent with previous studies (Lippa et al., 2005), diazepam approximately doubled the size of the GABA-gated current (at an EC10 concentration of GABA) mediated by α1β2γ2S constructs (103 ± 2.6%), with an EC50 of 240 nM (Fig. 1A; Table 1). DOV 51892 was ∼3-fold less potent (EC50 = 750 nM), but it was more efficacious than diazepam (DOV 51892 Emax/diazepam Emax = 1.50) (Fig. 1A; Table 1). DOV 51892 exhibited both lower potency and lower efficacy at the other “diazepam-sensitive” GABAA receptor constructs tested (α2β2γ2S, α3β2γ2S, α5β2γ2S), both in an absolute sense and relative to diazepam (Fig. 1, B–D; Table 1). Thus, the rank order potency of DOV 51892 at enhancing GABA-gated chloride currents in different α subunit-containing constructs was α1 > α2 = α3 > α5, compared with α2 = α3 = α5 > α1 for diazepam. Furthermore, DOV 51892 maximally enhanced GABA-gated current amplitude in various α subunit-containing constructs with a rank order of absolute efficacy of α1 > α2 > α3 = α5, compared with α3 > α2 > α1 > α5 for diazepam. Compared with the rank order of absolute efficacy for DOV 51892, the maximal potentiation of GABA-gated chloride currents by DOV 51892 relative to diazepam showed a substantially different profile: α1 > α5 > α2 > α3.
DOV 51892 was subsequently examined in animal models used to predict the anxiolytic activity and “side effects” (sedation, muscle relaxation, and ataxia) produced by benzodiazepine recognition site agonists. DOV 51892 (1–24 mg/kg) dose-dependently increased the number of punished responses in the Vogel conflict assay with a minimal effective dose (MED) of 3 mg/kg (Fig. 2). Moreover, it increased the response rate to a maximum of 3.8-fold above that of vehicle-treated animals (Fig. 2A). The increase in punished responding produced by DOV 51892 (6 mg/kg p.o.; 15.6 ± 3.09 licks) was suppressed by coadministration of a dose of flumazenil (5 mg/kg i.p.; 6.1 ± 2.3 licks; p < 0.01; Tukey's test following analysis of variance) that had no effect when administered alone (5.3 ± 0.6 versus 5.3 ± 1.0 licks; vehicle versus flumazenil, respectively). Diazepam (2.5–20 mg/kg; Fig. 2B) significantly increased punished responding only at the 10-mg/kg dose. However, only 4 of 10 rats tested showed any punished drinking after treatment with 20 mg/kg diazepam (Fig. 2B). Neither DOV 51892 (12 and 24 mg/kg) nor diazepam (10 mg/kg) significantly altered the unpunished rate of responding (Fig. 2C).
DOV 51892 (1–24 mg/kg) also exhibited anxiolytic-like activity in the elevated plus-maze (Fig. 3A), with a MED of 3 mg/kg, and a significant, 5-fold increase above vehicle levels in the percentage of time spent in the open arms. Diazepam also enhanced the percentage of time spent in the open arms, with significant increases observed after administration of the 5- and 20-mg/kg doses (2.8-fold increase above vehicle levels at 20 mg/kg; Fig. 3C). Although there was a trend toward an increase in the total number of arm crossings following administration of 3 mg/kg DOV 51892 or 10 mg/kg diazepam, this did not reach significance. Higher doses (96 mg/kg DOV 51892; 50 mg/kg diazepam) significantly reduced the total number of arm crossings (Fig. 3, B and D).
When tested under 40-lux lighting conditions, DOV 51892 (3 and 6 mg/kg; Fig. 4A) significantly increased the distance traveled in the open field by 154 to 167%. This activity declined at higher doses of DOV 51892, but it never fell below control activity levels, even at the highest dose tested (96 mg/kg). Diazepam significantly increased locomotion in the open field at 10 mg/kg, but it reduced the distance traveled by a maximum of 40% after 50 mg/kg (Fig. 4B).
DOV 51892 did not consistently affect rotarod performance at doses of up to 96 mg/kg (Fig. 5A), whereas 50 mg/kg diazepam significantly reduced the latency to fall from the rotarod by an average of 8 s (Fig. 5B). DOV 51892 (Fig. 6A) did not alter grip strength at the highest dose tested (96 mg/kg). In contrast, diazepam significantly reduced grip strength at a dose of 50 mg/kg (Fig. 6B).
The plasma and brain levels of DOV 51892 at 1 h after oral administration of a dose of 6 mg/kg to male rats were determined to be 4.6 ± 0.5 (n = 6) and 5.6 ± 0.6 μM (n = 6), respectively. Plasma levels of DOV 51892 after an oral dose of 10 mg/kg to male rats averaged 2300 ng/ml from 0.5 to 8 h after dosing.
Results from mice rendered insensitive to diazepam by means of a His→ Arg point mutation of the α subunit indicate that GABAA receptors containing α1 and α2 subunits mediate the anxiolytic and sedative actions of benzodiazepine recognition site ligands, respectively (Low et al., 2000; McKernan et al., 2000). Nonetheless, there is an emerging body of pharmacological evidence, including the anxioselective actions of the α3 subunit-selective positive modulator TP-003 (Dias et al., 2005) and the pyrazolopyrimidine ocinaplon (Lippa et al., 2005) that cannot be adequately explained by the findings from these point-mutated mice. The present observations demonstrate that DOV 51892 is a highly efficacious, positive modulator of currents mediated by α1 subunit-containing constructs and also exhibits an anxioselective profile in vivo. Relative to diazepam, DOV 51892 can be regarded as a “supermodulator” of the α1β2γ2S GABAA receptors.
Radioligand binding assays using a GABAA receptor preparation from rat cerebellum (which contains primarily the α1 subunit GABAA receptor isoform; Saxena and Macdonald, 1996) demonstrated that the potency of DOV 51892 in inhibiting [3H]flumazenil binding was ∼250-fold less than diazepam. The potency of DOV 51892 in this assay is slightly higher than the value reported for the structurally related compound ocinaplon (Lippa et al., 2005). However, in electrophysiological assays using the GABAA1a receptor construct, there is only approximately a 3-fold difference in potency between DOV 51892 and diazepam (Table 1; Fig. 1). This difference may result from the markedly lower potency of diazepam in this assay, together with a small increase in potency of DOV 51892. The EC50 of diazepam for enhancing GABA-gated currents mediated by recombinant GABAA1a receptors in this study is consistent with previously reported values for diazepam assayed under similar or identical conditions (Lippa et al., 2005). One factor contributing to the difference in potency of diazepam in these two assay systems is that the binding of benzodiazepines to native GABAA receptors is enthalpy-driven, and increasing the assay temperature by ≈20°C in the electrophysiology assay would significantly reduce ligand affinity (Maguire et al., 1992). Perhaps more surprising than the high potency of DOV 51892 in modulating GABA-gated Cl– currents mediated by recombinant GABAA1a receptors was its high efficacy (Emax) relative to diazepam. The efficacy of ocinaplon at recombinant GABAA1a receptors was reported to be ≈85% of that of diazepam (Lippa et al., 2005). In contrast, DOV 51892 potentiated the size of the GABA-gated current to 148% of that of diazepam (Table 1; Fig. 1) and may therefore be classified as a supermodulator at GABAA1 receptors. This high efficacy was selective for the GABAA1a receptor isoform, since DOV 51892, like ocinaplon (Lippa et al., 2005), is less potent and/or behaves as a partial modulator relative to diazepam at GABAA2a and A3a receptors.
Subsequently, the effects of DOV 51892 and diazepam were characterized in vivo. Because systemically (i.p.) administered diazepam causes significant ataxia at doses close to or overlapping its anxiolytic actions (Soderpalm et al., 1989; P. Popik, unpublished observations), both diazepam and DOV 51892 were administered orally to better reflect the primary clinical application of these agents as well as to better separate the anxiolytic and sedative actions of the drugs. The anxiolytic-like actions of DOV 51892 were manifested in the suppression of punished licking behavior in the Vogel conflict test and the enhancement of open arm exploration in the elevated plus-maze with a potency equal to or greater than that of diazepam. These anticonflict actions of DOV 51892 were abolished by Ro 15-1788, demonstrating that these behaviors are GABAA receptor-mediated. Based on the high efficacy of DOV 51892 in enhancing GABA-gated Cl– currents mediated by GABAA1a receptors and the results of studies in His→ Arg mutant α1 subunit-containing GABAA receptor knockin mice (McKernan et al., 2000), the molecular pharmacology of DOV 51892 would predict that it have predominantly sedative-like characteristics in vivo. The apparent inability of DOV 51892 to cause both ataxia and myorelaxation at doses manyfold higher than the MED for anxiolytic activity should be viewed as surprising. In contrast, orally administered diazepam causes ataxia and significant myorelaxation at doses only 2 to 10 times higher than the anxiolytic MED.
The in vivo pharmacological profile of DOV 51892 would not be predicted by the hypothesis that α1 subunit-containing GABAA receptors are responsible for mediating the sedative properties of benzodiazepine positive modulators (Rudolph et al., 1999; McKernan et al., 2000), whereas activation of GABAA receptors containing α2 and/or α3 subunits reduces anxiety (Low et al., 2000). Based on the anxiolytic activity of compounds with low efficacy (i.e., partial modulators) at α2 and α3 subunit-containing GABAA receptors, it has been proposed that a reduction in anxiety requires only partial potentiation of these receptors, compared with that achieved with, for example, a benzodiazepine (McKernan et al., 2000), whereas sedation would require maximal potentiation of α1 subunit-containing GABAA receptors (Facklam et al., 1992; Jones et al., 1994; Basile et al., 2004; Lippa et al., 2005).
Because of the relatively low potency (750 nM) of DOV 51892 at recombinant GABAA1a receptors, it could be argued that brain concentrations of DOV 51892 were not sufficient to achieve full potentiation, or at least an enhancement equivalent to 20 mg/kg diazepam. However, a dose of 6 mg/kg DOV 51892 active in both the Vogel conflict test and elevated plus-maze yielded a total brain concentration (≈5 μM) sufficient to saturate GABAA1a receptors. Although the total brain concentrations of DOV 51892 may not accurately reflect the amount of drug in the vicinity of GABAergic synapses, GABAA1a receptors are likely to be fully activated at doses of DOV 51892 that are more than 10-fold higher (96 mg/kg) than the anxiolytic dose. Even after these high doses of DOV 51892, there is no remarkable impairment of motor function.
The anxioselective actions of DOV 51892, a very efficacious GABAA1a receptor modulator, raises the possibility that the data obtained from recombinant receptors in vitro does not reflect the effects of DOV 51892 on native receptors in vivo. This possibility will require additional studies both in vivo and using transiently or stably transfected recombinant GABAA receptors in vitro, to evaluate the pharmacological properties of these compounds. Nonetheless, the hypothesis that GABAA receptors containing α1 subunits mediate the sedative actions of benzodiazepine recognition site ligands should be revisited in response to the pharmacological data from novel structural classes of benzodiazepine recognition site ligands.
A corollary to this premise is the possibility that facilitation of GABAA1a receptors may, in fact, contribute to the anxiolytic effects. Although the present data do not provide direct evidence that GABAA1a receptors are involved in anxiety, some support may be derived from previous reports of anxiolysis produced by agonists at α1 subunit-containing GABAA receptors, such as ocinaplon (Lippa et al., 2005) and CL 218,872 (Lippa et al., 1979), as well as reports (Cox et al., 1995; Harvey et al., 2002) that a selective antagonist at GABAA1a receptors (tert-butyl-β-carboline carboxylate) blocks the anxiolytic but not the motor-impairing actions of benzodiazepines (Shannon et al., 1984).
This work was done with either funding from (to P.P., E.K., M.K., G.N., B.S., S.R., T.T.G., D.H.F.) or as employees of (P.K., Z.C., P.S., A.S.L., A.S.B.) DOV Pharmaceutical, Inc.
P.P. and E.K. contributed equally to this work.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
ABBREVIATIONS: SL-651498, 6-fluoro-9-methyl-2-phenyl-4-(pyrrolidin-1-yl-carbonyl)-2,9-dihydro-1H-pyrido[3,4-b]indol-1-one; L-838417, 3-(2,5-difluorophenyl)-7-(1,1-dimethylethyl)-6-(1-methyl-1H-1,2,4-triazol-5-yl)methoxyl-1,2,4-triazolo[4,3-b]pyridazine; Ro 15-1788, flumazenil; TP-003, 4,2′-difluoro-5′-[8-fluoro-7-(1-hydroxy-1-methylethyl) imidazo[1,2-a]pyridine-3-yl]biphenyl-2-carbonitrile; DOV 51892, 7-(2-chloropyridin-4-yl)pyrazolo-[1,5-a]-pyrimidin-3-yl](pyridin-2-yl)methanone); MED, minimal effective dose; CL-218872, 3-methyl-6-[-3-(trifluoromethyl)phenyl]-1,2,4-triazolo[4,3-b]pyridazine.
- Received May 2, 2006.
- Accepted September 11, 2006.
- The American Society for Pharmacology and Experimental Therapeutics