Psychopharmacology (2008) 197:591–600 DOI 10.1007/s00213-008-1077-z
ORIGINAL INVESTIGATION
Discriminative stimulus effects of tiagabine and related GABAergic drugs in rats L. M. McDonald & W. F. Sheppard & S. M. Staveley & B. Sohal & F. D. Tattersall & P. H. Hutson
Received: 19 December 2006 / Accepted: 10 January 2008 / Published online: 9 February 2008 # Springer-Verlag 2008
Abstract Rationale Tiagabine is an anticonvulsant drug which may also have sleep-enhancing properties. It acts by inhibiting reuptake at the gamma-aminobutyric acid (GABA) transporter (GAT-1). Objectives The aim of the study was to determine whether tiagabine acted as a discriminative stimulus and, if so, whether other GABAergic compounds would generalise to it. Materials and methods Rats were trained to discriminate tiagabine (30 mg/kg p.o.) from vehicle, and generalisation to drugs that modulate GABA was assessed. Results Gaboxadol (5–20 mg/kg p.o.), a selective extrasynaptic GABAA agonist, generalised to tiagabine, although the extent of the generalisation was inconclusive. Indiplon (1 mg/kg p.o.), a benzodiazepine-like hypnotic, also partially generalised to tiagabine, although zolpidem and S-zopiclone did not. Baclofen, a GABAB receptor agonist, and gabapentin, which increases synaptic GABA, did not generalise to L. M. McDonald (*) : W. F. Sheppard : S. M. Staveley : B. Sohal : F. D. Tattersall : P. H. Hutson Merck, Sharp and Dohme, Terlings Park, CM21 2QR, Harlow, Essex, UK e-mail:
[email protected] W. F. Sheppard : P. H. Hutson Merck and Co. West Point, Pennsylvania, USA S. M. Staveley : F. D. Tattersall Discovery Biology Pfizer Ltd., Ramsgate Road, Sandwich CT13 9NJ, UK B. Sohal Novartis Horsham Research Centre, Wimblehurst Road, Horsham RH12 5AB, UK
tiagabine. (+)-Bicuculline (3 mg/kg i.p.), a GABAA receptor antagonist, blocked the tiagabine cue, but the less brainpenetrant salt form, bicuculline methochloride, had no effect. Conclusions These data suggest that tiagabine generates a discriminative stimulus in rats, and provides a central GABA-mediated cue, but is distinct from the other GABAergic compounds tested. Keywords Tiagabine . GABA . Drug discrimination . Rat . Gaboxadol . Indiplon . Baclofen . Gabapentin . Bicuculline . Bicuculline methochloride
Introduction Tiagabine (R(−)-N-[4,4-bis(3-methylthien-2-yl)but-3-enyl] nipecotic acid) is a brain-penetrant derivate of nipecotic acid and potently binds to the GAT-1 gamma-aminobutyric acid (GABA) transporter in neurons and glial cells (Brodie 1995). It is used clinically as add-on treatment in epilepsy (Leach and Brodie 1998). Tiagabine has also been shown to enhance sleep in healthy elderly people (Mathias et al. 2001) and to improve the efficiency of sleep in rats, with an EEG profile resembling that of gaboxadol (Lancel et al. 1998). It was therefore of interest to determine whether the discriminative stimulus effects of tiagabine could be accounted for by actions at GABA receptors or by increasing synaptic GABA. In addition to its GABA reuptake inhibition, tiagabine has a low affinity for the histamine H1 receptor and for the benzodiazepine receptor and negligible affinity for other receptors (Brodie 1995). Tiagabine dose-dependently increased extracellular GABA overflow by at least 200% and prolonged GABA-induced hyperpolarisation in hippocampal slices (Braestrup et al. 1990). It showed an anticonvul-
592
sant profile in many preclinical models, including pentylenetetrazol-induced seizures and amygdala kindling (Nielsen et al. 1991). Consistent with this profile, tiagabine is used clinically as add-on therapy in epilepsy, where seizures are not fully controlled by traditional anticonvulsants, such as carbemazepine (Leach and Brodie 1998). Sedation has been reported as a side effect of tiagabine (Schachter 1999), consistent with its impairment of rotarod performance and locomotor activity in rats (Nielsen et al. 1991), and indicated possible sedative or hypnotic potential. Tiagabine was tested for its sedative potential in rats and humans, and showed a profile that resembled that of gaboxadol (Lancel et al. 1998; Mathias et al. 2001). Drug discrimination in rats is a well-established method of comparing the in vivo stimulus properties of compounds. It has distinguished more than 40 different classes of compounds and shows considerable pharmacological specificity (Overton 1991). It has been extensively used to study GABAergic compounds, including the benzodiazepines, barbiturates, muscimol and gaboxadol, and has revealed some important distinctions (see Lelas et al. 2000 for a review). In rats trained to discriminate the direct GABAA agonist, muscimol, from vehicle, diazepam, midazolam, baclofen, vigabatrin and tiagabine produced no more than partial generalisation (Grech and Balster 1997; Jones and Balster 1998). In rats trained to discriminate gaboxadol from vehicle, zolpidem, zopiclone and tiagabine did not generalise (McDonald et al. 2007). To our knowledge, tiagabine has not previously been shown to generate a cue in rat-drugdiscrimination studies. Therefore, it was of interest to determine whether or not tiagabine acted as a discriminative stimulus and, if it did, to test several drugs with actions at different GABA receptors or that increased synaptic GABA, to see if any of these generalised to tiagabine. Several GABAergic compounds were selected for testing against the tiagabine discriminative stimulus. Gaboxadol is a selective extrasynaptic GABA agonist, which is thought to act primarily through GABAAα4βδ receptors (Brown et al. 2002; Wafford et al. 1996). Zolpidem, S-zopiclone and indiplon are all allosteric modulators at the benzodiazepine site of GABAAαβγ-containing receptors, and are moderately selective for the GABAAα1 receptor subunit (Crestani et al. 2000; Davies et al. 2000; Foster et al. 2004). Baclofen is a direct GABAB receptor agonist. Gabapentin increases synaptic GABA concentration and interacts with the α2δ calcium channel (Gee et al. 1996; Löscher et al. 1991). These compounds were tested for generalisation to tiagabine. (+)-Bicuculline, a GABAA receptor antagonist, was used to block the tiagabine cue, as was bicuculline methochloride which, in contrast to bicuculline, does not induce convulsions and is therefore thought to be a less brain-penetrant salt form of bicuculline.
Psychopharmacology (2008) 197:591–600
Materials and methods Male Sprague–Dawley rats weighing about 200–250 g (10 weeks old) at the start of training were housed in groups of three, under a 12-h light–dark cycle, with lights on at 7 A.M. and were food-restricted to maintain a body weight of no less than 80% of their free-feeding weight. After arrival, rats had been allowed to settle for 1 week and were then food-restricted for 1 week before the start of training. All procedures in these studies were conducted in accordance with the UK Animals (Scientific Procedures) Act 1986 and associated UK Home Office guidelines. Standard operant test chambers were used (Med Associates, St. Albans, Vermont, USA). Each was equipped with two retractable levers and a feeder for the delivery of food pellets. During initial training, a single lever was extended, and this was alternated between left and right each day. Rats were trained to press a single lever to obtain food rewards in a 15-min session. For each reward, the lever had to be pressed ten times (fixed-ratio 10 (FR10) schedule), and animals were trained until they collected at least 20 of 30 pellets per session, on at least two consecutive training days. Rats were then trained to discriminate tiagabine 30 mg/ kg p.o. from vehicle. (Training had begun with 10 mg/kg p.o. tiagabine, which was raised to 20 and then 30 mg/kg, after which the rats started to learn the discrimination). Each day, rats were dosed with tiagabine or vehicle in a randomised fashion, 30 min before being placed in the operant boxes, with both levers extended, for a 15-min session. For each rat, one lever was randomly designated as the drug lever, and the other was designated as the vehicle lever. To receive a food reward, each rat had to make ten consecutive responses on the appropriate lever, according to whether it was given the training drug or vehicle. If a rat made an incorrect response, the response requirement on the correct lever was reset. Training sessions took place once daily, generally 5 days per week. Rats were considered fully trained to the drug cue when they had achieved at least 80% responding to the appropriate lever over five consecutive training sessions. In addition, individual rats had to complete ten consecutive responses on the appropriate lever and no more than five responses on the inappropriate lever before the delivery of the first reward, in at least eight of ten consecutive training sessions.
Generalisation testing One or two generalisation test sessions (GTs) were run per week. At least two training sessions were carried out between each GT, to return performance to criterion. For all
Psychopharmacology (2008) 197:591–600
generalisation tests, rats had to show correct vehicle and drug lever responding in training sessions immediately before the test day, otherwise they were omitted from that test day. All available rats that met the criterion were used. The GTs had identical parameters to training sessions, except that no food rewards were given. Doses for generalisation tests were chosen based on published literature or on pilot experiments to determine the maximum dose of a compound that did not cause a cessation of responding. The test of tiagabine, and the GTs of gaboxadol, zolpidem and indiplon were run within-subjects. The tests of tiagabine and gaboxadol were run simultaneously on separate groups of rats, followed by simultaneous running of the tests of indiplon and zolpidem. Each rat received at least three test doses of that drug, plus vehicle and a positive control (tiagabine) over the course of several generalisation test days, using a Latin square design to determine the order of dosing for each rat. All doses for each drug were completed before moving to the next drug for that group of rats. Due to a requirement to complete the studies much earlier than originally planned, the GTs for S-zopiclone, gabapentin, baclofen, (+)-bicuculline and bicuculline methochloride were run as between-subjects designs. All rats from the within-subjects tests participated in the between-subjects tests. There were insufficient numbers of rats to allow each rat to participate only once in each GT, so some rats participated in a drug test group on test day 1 and a vehicle or positive control test group on test day 2, or vice versa, across a maximum of two test days for each drug, following the method of Bristow et al. (1998).
Drug preparation Tiagabine HCl, gaboxadol HCl, gabapentin (base) and (±)baclofen (base) were dissolved in distilled water. Bicuculline methochloride was dissolved in saline (0.9% NaCl). Zolpidem, S-zopiclone and (+)-bicuculline were dissolved in 300 μl propylene glycol (1.3% of total solution volume), 200 μl glacial acetic acid (0.9%) and 1.5 ml 1 M NaOH (6.8%), plus saline or distilled water to give the correct solution volume. This method reliably gave a clear solution at pH 5. Indiplon was suspended in 0.5% w/v methylcellulose. All drugs were given 30 min before behavioural testing, except for gabapentin, which was given 60 min before testing. All drugs were dosed p.o. by oral gavage into the stomach at 3 ml/kg, except (+)-bicuculline and bicuculline methochloride, which were dosed i.p. at 1 ml/kg. All vehicles were pH-matched. All doses refer to the base form of each compound. Tiagabine HCl was supplied by
593
R&S Pharmachem (China), gaboxadol HCl and indiplon were the kind gift of Lundbeck (Copenhagen), zolpidem was supplied by Tocris Cookson (UK), zopiclone was supplied in its racemic form by Tocris Cookson and separated into enantiomers in-house. Baclofen, gabapentin, (+)-bicuculline and bicuculline methochloride were supplied by Sigma (UK).
Data analysis The main dependent variable was the percentage of druglever responses compared to total lever presses up to the point at which the first reward would have been delivered (GTs were conducted in extinction). Response rate data for that set of lever presses was also analysed. For all statistical analyses, α=0.05. Positive control data were excluded from all statistical analyses. For experiments that were conducted within-subjects using a balanced Latin square design (testing of tiagabine against itself and GTs for gaboxadol, zolpidem and indiplon), a one-way repeated measures analysis of variance (ANOVA) was run with dose as the factor. Maunchly’s test of sphericity was carried out on all repeated measures ANOVAs and, where sphericity was violated, a Greenhouse–Geisser correction to the degrees of freedom was used. Where main effects were significant, differences between groups were analysed with t tests, followed by a Bonferroni correction for comparisons with the control group. For tests that were conducted betweensubjects (GTs for S-zopiclone, gabapentin, baclofen, (+)bicuculline and bicuculline methochloride), a one-way ANOVA was used, with dose as the only factor. Significant main effects were followed-up with Dunnett’s test, to determine significant differences between each drug group and the vehicle group. Rats were deemed to have generalised fully to the training compound if they made 80% or more responses on the drug-associated lever. Graphs show means and SEs. Rats with two or more missing data points were excluded from the analysis. In the within-subjects studies, if a rat had been excluded at a particular dose, that dose was repeated in that rat if possible when the Latin square design was concluded. Rats that made fewer than ten lever presses in a session were excluded from drug lever responding data for that test day, but were included in response rate data. In two cases, one missing data point was replaced by the mean of the other rats at that dose, to prevent the statistical package from deleting all data for those rats. All calculations were made using SPSS v14.0 (SPSS, Chicago), except that some Dunnett’s tests were calculated by hand. In the current study, partial generalisation was defined as a statistically significant (p
