S33138 [N-[4-[2-[(3aS,9bR)-8-Cyano-1,3a,4,9b-tetrahydro[1]-benzopyrano[3,4-c]pyrrol-2(3H)-yl)-ethyl]phenylacetamide], a Preferential Dopamine D₃ versus D₂ Receptor Antagonist and Potential Antipsychotic Agent. II. A Neurochemical, Electrophysiological and Behavioral Characterization in Vivo

In the accompanying article (Millan et al., 2007), the benzopyranopyrrolidine derivative, S33138, was shown to behave as a potent antagonist at human (h) dopamine D₃ receptors and, at higher concentrations, as an antagonist at hD_2L (long isoform) and hD_2S (short isoform) receptors. In addition, S33138 expressed modest antagonist properties at hα_2C-adrenoceptors (ARs), serotonin (5-HT)_2A receptors, and h5-HT₇ receptors and weak antagonist actions at hD₁ receptors. Thus, S33138 possesses a cellular profile, distinguishing it from previously documented antipsychotics. The purpose of the present studies was 3-fold: first, employing a range of electrophysiological, neurochemical, and behavioral procedures to determine whether S33138 behaves as an antagonist at presynaptic and postsynaptic populations of cerebral D₃ and D₂ receptors; second, employing procedures in which the actions of selective D₃ versus D₂ antagonists have previously been documented (see citations below) to evaluate whether S33138 preferentially antagonizes D₃ versus D₂ receptors in vivo; and third, to establish whether S33138 antagonizes central nervous system populations of α_2C-ARs, 5-HT_2A receptors, 5-HT₇ receptors, and D₁ receptors.

Activity of the cellular marker of neuronal activity, c-fos, is controlled by both D₃ and D₂ receptors. Selective D₃ versus D₂ receptor antagonists preferentially induce its gene expression in the Isles of Calleja (IOC) and nucleus accumbens (Kovacs et al., 2001; Southam et al., 2007), structures enriched in D₃ receptors compared to the striatum wherein D₂ receptors predominate (Landwehrmeyer et al., 1993; Joyce, 2001). Interest in the influence of S33138 upon c-fos is reinforced by studies indicating that the atypical antipsychotic, clozapine, but not haloperidol, also preferentially enhances c-fos expression in the IOC and nucleus accumbens (Vahid-Ansari and Robertson, 1996; Merchant et al., 1996; Guo et al., 1998; Carta and Gerfen, 1999; Kovacs et al., 2001). A further parallel between selective D₃ receptor antagonists and clozapine is that, upon chronic administration, both preferentially decrease the number of spontaneously active dopaminergic neurones in the ventral tegmental area (VTA) versus the substantia nigra, pars compacta (SNPC) (Ashby and Wang, 1996; Ashby et al., 2000). Likewise, this approach suggests that blockade of D₃ receptors principally modulates the activity of D₃ receptor-rich limbic versus striatal structures.

D₃ autoreceptors are colocalized with D_2S receptors on dopaminergic perikarya and terminals (Stanwood et al., 2000; Usiello et al., 2000; Joyce, 2001; Joyce and Millan, 2005) where they play complementary roles in controlling the synthesis, release, and clearance of dopamine (DA), as well as the electrical activity of dopaminergic neurones (Millan et al., 2000d; Usiello et al., 2000; Roberts et al., 2006; Sokoloff et al., 2006). Blockade of tonically active, mesolimbic D_2S autoreceptors by haloperidol enhances DA release in the nucleus accumbens (Millan et al., 2000d), whereas their activation by the partial agonist aripiprazole may, in moderating limbic DA release, contribute to its antipsychotic properties (Davies et al., 2004). Inasmuch as acute administration of selective D₃ receptor antagonists abrogates the phasic suppression of DA release and VTA firing by the preferential D₃ versus D₂ receptor agonist, PD128,907, we examined the influence of S33138 upon its actions.

Discriminative stimulus (DS) properties of dopaminergic agonists reflect recruitment of D₃ and/or D₂ autoreceptors (Cory-Slechta et al., 1996; Bristow et al., 1998; Millan et al., 2000b). By contrast, yawning involves postsynaptic D₂ and/or D₃ receptors on oxytocinergic neurones in the paraventricular nucleus of the hypothalamus (Argiolas and Melis, 1998; Chen et al., 1999; Millan et al., 2000a; Collins et al., 2005). With regard to the hypothermic actions of dopaminergic agonists integrated in the IOC, perifornical hypothalamus, and other structures, the contribution of postsynaptic D₃ versus D₂ receptors remains controversial (Barik and Beaurepaire, 1998; Boulay et al., 1999; Millan et al., 2000a; Perachon et al., 2000; Chaperon et al., 2003). It is interesting that there is compelling evidence for a contrasting influence of postsynaptic D₃ versus D₂ sites upon motor function. Thus, blockade of mesolimbic and striatal D₂ sites disrupts motor behavior, whereas the selective inactivation of D₃ receptors enhances motor behavior (Boulay et al., 1999; Millan et al., 2000a; Joyce, 2001; Sokoloff et al., 2006). This opposing influence of D₃ versus D₂ receptors can be revealed in primates rendered parkinsonian by the neurotoxin 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP). In such akinetic subjects, improvement of motor performance by antiparkinson agents is potentiated and abrogated by selective antagonists at D₃ and D₂ receptors, respectively (Silverdale et al., 2004; Hill et al., 2006). In light of the differential implication of D₃ and D₂ sites in these behavioral procedures, they were also used to examine the actions of S33138.

Dopamine D₁ receptors in limbic and striatal regions control motor drive and coordination, as reflected in the induction of rotation by agonists in rats sustaining a unilateral lesion of the SNPC (Gulwadi et al., 2001; Gerfen et al., 2002). Thus, we examined the influence of S33138 upon rotation elicited by SKF81297. Despite the significance of α_2C-ARs to cognition and mood (Svensson, 2003), in vivo models of drug actions at α_2C-AR sites remain to be established. Nonetheless, α_2C-AR autoreceptors regulate the activity of adrenergic perikarya in the locus coeruleus (LC) (Arima, 1998; Millan et al., 2000d; Owesson et al., 2003), so the influence of S33138 upon its firing rate was determined in anesthetized rats. Finally, activation of 5-HT_2A receptors by the agonist 1-[2,5-dimethoxy-4-iodophenyl]-2-aminopropane (DOI) elicits head twitches (HTW) in rats (Schreiber et al., 1995; Willins and Meltzer, 1997), whereas stimulation of 5-HT₇ receptors in guinea pigs evokes hypothermia (Hagan et al., 2000; Hedlund and Sutcliffe, 2004). In relation, potential antagonist actions of S33138 at these sites were evaluated using these procedures.

Materials and Methods

Animals. Unless otherwise specified, in vivo studies employed male Wistar rats weighing 220 to 250 g (Iffa-Credo, L'Arbresle, France). Male CD1 mice weighing 22 to 25 g (Charles River, Saint-Aubin-les-Elbeuf, France), male C57/Bl6 mice weighing 20 to 25 g (for c-fos studies), and male Hartle guinea pigs weighing 200 to 250 g (Charles River) were also used. Subjects were maintained in sawdust-lined cages with unrestricted access to food and water. The laboratory temperature was held at 21 ± 1°C, and humidity was controlled at 60 ± 5%. There was a 12/12-h light/dark cycle, with lights on from 7:30 AM to 7:30 PM. Before experimentation, all animals were adapted for at least 1 week to laboratory conditions. All animals use procedures conformed to international European ethical standards (86/609-EEC) and the French National Committee (décret 87/848) for the care and use of laboratory animals. In primate studies, adult marmosets (Callithrix jacchus) were used. They were kept in controlled housing conditions, with the temperature held at 26 ± 1°C, 50% relative humidity, and a 12-h light/dark cycle (lights on from 7:30 AM to 7:30 PM). They had free access to food pellets, fresh fruit supplements, and water. All primate experiments were carried under Home Office License PPL 40/01487 in accordance with United Kingdom legal requirements.

Drug Testing. Unless otherwise specified, S33138 was examined by the subcutaneous route, and full dose-response curves were generated. In the protocols of D₃ and D₂ receptor-mediated activity, we have characterized the actions of selective D₃ (and D₂) receptor antagonists in previous studies (see Discussion). The equilibrated D₂/D₃ receptor antagonist, haloperidol, was employed herein as an internal reference ligand to “validate” this series of experiments; data are summarized under Results. For protocols of activity at D₁ receptors, α_2C-ARs, 5-HT_2A receptors, and 5-HT₇ receptors, the selective antagonists SCH23390, idazoxan, MDL100,907, and SB269,970, respectively, were similarly used as “internal” validators in parallel with S33138 (see Results).

Influence of S33138 upon Cerebral Levels of mRNA Encoding c-fos. In mice, the influence of a single i.p. injection of S33138 upon levels of mRNA encoding c-fos was evaluated in cerebral structures differentially expressing D₃ versus D₂ receptors (IOC, nucleus accumbens, and striatum) employing in situ hybridization as described in detail previously (Svenningsson et al., 1997). In brief, mRNA levels encoding c-fos were quantified in coronal, cryostat brain sections using a ³⁵S-labeled cRNA probe and standard hybridization and washing conditions. After washing, sections were dried and apposed to betaMAX hyperfilm for 4 weeks. Autoradiograms were quantified with a Microcomputer Imaging Device system using NIH Image 1.62 software. Slides were subsequently dipped in liquid photographic emulsion and counterstained in cresyl violet for cellular analyses of neurons labeled for c-fos.

Influence of Chronic S33138 upon Spontaneously Active Dopaminergic Neurones in the VTA versus SNPC of Anesthetized Rats. The procedure used was essentially as described previously (Ashby et al., 2000). Male Sprague-Dawley rats (Taconic Farms, NY) (150–175 g at the beginning of the treatment) received one p.o. administration of vehicle or S33138 (0.16 to 10.0 mg/kg) or the “positive control” haloperidol (0.5 and 2.0 mg/kg) once a day for 21 consecutive days. All experiments were conducted 2 h after the last administration on day 21. Animals were anesthetized with chloral hydrate (400 mg/kg i.p.) and mounted in a stereotaxic instrument. Single barrel microelectrodes (glass borosilicate; World Precision Instruments, Sarasota, FL) were used for recording the electrical activity of single dopaminergic neurones in the A10, VTA (AP: 3.0–3.5 mm; L: 0.5–1.0 mm, and DV: 6.0–8.5 mm to the cortical surface), and the A9, SNPC (AP: 3.0–3.5 mm to lambda, L: 1.8–2.5 mm, and DV: 6.0–8.5 mm). The number of spontaneously active dopaminergic neurons in each region was determined across 10 stereotaxic descents.

Influence of S33138 upon the Actions of the D₃/D₂ Receptor Agonist, Ropinirole, in Parkinsonian Primates. The procedure employed was essentially that described previously (Silverdale et al., 2004; Hill et al., 2006). In brief, six adult marmosets were rendered parkinsonian by subcutaneous injection of 2 mg/kg MPTP for 5 consecutive days. This produced a parkinsonian state that was stable after 18 weeks. Marmosets were then treated orally with 12.5 mg/kg l-(3,4-dihydroxyphenyl)-alanine and 3.1 mg/kg benserazide (as Madopar, 62.5 mg dispersible, dissolved in apple juice; Roche Diagnostics, Indianapolis, IN) twice daily for 4 weeks. Thereafter, the studies with S33138 were performed. On the day of experiment, ropinirole plus vehicle, or ropinirole plus S33138, was administered orally in volumes of 5 ml/kg. In a separate experiment, S33138 or vehicle was administered in the absence of ropinirole. Immediately after treatment, the animals were transferred to an observation cage (60 × 55 × 75 cm). Locomotor activity was monitored for 2 consecutive hours using computer-based infrared activity monitors, which provide a quantitative assessment of the total amount of movement in each 5-min time period throughout the experiment.

Influence of S33138 upon the Hypothermia Induced by the Preferential D₃ versus D₂ Receptor Agonist, 7-OH-DPAT. Core temperature (CT) was determined in lightly restrained rats by use of a rectal thermistoprobe as described previously (Millan et al., 2000a). Basal CT was measured; S33138, haloperidol, or vehicle was administered s.c.; and 30 min later, 7-OH-DPAT (0.16 mg/kg s.c.) or vehicle was injected. Thirty minutes later, CT was reevaluated, and the difference (Δ) to pretreatment values was calculated.

Influence of S33138 upon the Yawns Elicited by 7-OH-DPAT. As described previously (Millan et al., 2000a), rats (120–140 g) were individually placed in Plexiglas observation cages (11 × 26 × 30.5 cm), behind which mirrors were positioned to facilitate monitoring of behavior. Thirty minutes after injection of S33138, haloperidol, or vehicle, 7-OH-DPAT (0.04 mg/kg s.c.) was administered, and the number of yawns was measured over a 30-min observation period.

Influence of S33138 upon DS properties of the Preferential D₃ versus D₂ Receptor Agonist, PD128,907. Employing a procedure detailed in previous work (Millan et al., 2000b), rats were trained to discriminate PD128,907 (0.16 mg/kg i.p.) from saline using a two-lever, fixed-ratio 10, food-reinforced procedure. Each 15-min daily session started 15 min after injection. The discrimination criterion was 10 consecutive sessions of correct responses (defined as a maximum of 13 responses on both reinforced and nonreinforced levels before the first reinforcement). Thereafter, tests were performed on Wednesdays and Fridays, and training sessions were performed on the other days. Only rats responding appropriately on the two most recent training days were examined. In test sessions, responses on the selected lever (on which 10 responses were emitted first) were reinforced for the rest of the session. S33138 or haloperidol was injected s.c. 30 min before PD128,907. Data recorded during the test sessions were lever selection and response rate (the total number of presses on both levers).

Blockade by S33138 of the Inhibitory Influence of PD128,907 upon the Electrical Activity of Dopaminergic Cell Bodies. The procedure used has been detailed previously (Millan et al., 2000c, 2004a). After anesthesia with chloral hydrate (400 mg/kg i.p.) and femoral vein cannulation, rats were placed in a stereotaxic apparatus, and a tungsten microelectrode was lowered by an electronic microdrive (Unimechanique, Epinay sur Seine, France) into the VTA (AP: –5.5 from bregma, L: 0.7, H: –7/–8.5 from dura) for recording of extracellular unit activity. Dopaminergic neurones were characterized by their distinctive waveforms and discharge patterns (Millan et al., 2000c). One cell was recorded in each animal. Following baseline recording ((5 min), PD128,907 (0.01 mg/kg i.v.) or vehicle was injected; 1 min later, S33138 (0.001–0.25 mg/kg i.v., one dose per rat) or vehicle was injected, and firing rate was again recorded. In a separate experiment, the influence of cumulative (every 2–3 min) administration of S33138 (0.125–4.0 mg/kg i.v.) versus vehicle was examined. At the end of the treatment, the prototypical dopaminergic agonist, apomorphine (0.031 mg/kg i.v.), was injected, and firing rate was again recorded. Haloperidol was evaluated at a single dose of 0.031 mg/kg i.v. Drug effects were quantified over a 60-s bin at the time of peak action. Spike2 version 5.11 software (CED, Cambridge, England) was used for data acquisition and off-line analysis. Drug effects are expressed as percentage change from baseline (predrug) firing rate (defined as 0%).

Blockade by S33138 of the Inhibitory Influence of PD128,907 upon Dialysate Levels of Dopamine in Freely Moving Rats. Extracellular levels of DA in single dialysate samples of the frontal cortex (FCX), nucleus accumbens, and striatum were determined as described previously (Millan et al., 2000c, 2004a). In brief, guide cannulae were implanted under pentobarbital anesthesia (60 mg/kg i.p.) 1 week before experimentation, and on the test day, a cuprophane CMA/11 probe (4 mm in length for the FCX and striatum, 2 mm for the nucleus accumbens, and, in each case, 0.24 mm outer diameter) was lowered into position. Three basal samples of 20 min each were taken. Vehicle, S33138, or haloperidol was administered s.c. followed 20 min later by PD128,907 (0.16 mg/kg s.c.) or vehicle, and samples was taken for another 3 h. DA levels were quantified by high-performance liquid chromatography followed by amperometric detection. The assay limit of sensitivity was 0.1 to 0.2 pg/sample. Influence of S33138 upon Rotation Elicited by the D₁ Receptor Agonist, SKF81297, in Rats with a Unilateral Lesion of the SNPC. As described previously (Millan et al., 2000a, 2004b), rotation was measured in rats (350–400 g at the time of the study) that had received a unilateral 6-hydroxydopamine (8 μg/4 μl) lesion of the SNPC. Rotation was measured by use of an Rotacount 8 (Columbus Instruments, Columbus, OH) apparatus. Sessions were performed weekly employing an “ABACADA” design where “A” corresponded to injection of the selective D₁ agonist, SKF81297 (0.04 mg/kg s.c.) (control sessions) and “B, C, and D” corresponded to test sessions with s.c. administration of S33138 or the selective D₁ receptor antagonist SCH23390 followed by SKF81297. In SKF81297-control sessions, vehicle was given 30 min before SKF81297, and rotation was measured for 1 h. In test sessions, S33138 or SCH23390 was given 30 min before SKF81297, and likewise, rotation was measured for 1 h. Thus, animals were their “own” controls.

Influence of S33138 upon the Electrical Activity of Adrenergic Compared to Serotonergic Cell Bodies. The basic procedure used was as described above for the VTA and documented previously (Millan et al., 2000c, 2004a). Thus, a tungsten microelectrode was lowered by an electronic microdrive into the LC (AP: –1.0 from 0; L: 1.0–1.2; H: –5.5/–6.0 from the sinus surface) of anesthetized rats. Adrenergic neurones were characterized by their distinctive waveforms and discharge patterns (Millan et al., 2000c, 2004a). One cell was recorded in each animal. Following baseline recording (≥5 min) and vehicle injection (0.5 ml/kg i.v.), S33138 was administered in cumulative doses every 2 to 3 min. Subsequent to S33138 administration, a further injection was made of the α₂-AR autoreceptor agonist clonidine (0.01 mg/kg i.v.), followed by the α₂-AR antagonist idazoxan (0.063 mg/kg i.v.). In a separate group of subjects, the influence of idazoxan alone was also tested. In a further study, the influence of S33138 upon the activity of serotonergic neurones in the dorsal raphe nucleus (DRN) was examined. Coordinates were as follows: AP: –7.8 from bregma; L: 0; H: –5/–6.5 from the sinus surface. Cumulative administration of S33138 was followed by injection of the 5-HT_1A receptor agonist 8-OH-DPAT (0.005 mg/kg i.v.), followed by the selective 5-HT_1A antagonist WAY100,635 (0.1 mg/kg i.v.). In a separate experiment, WAY100,635 was injected followed by 8-OH-DPAT. Effects of S33138 were quantified over 60-s bins at the time of peak action. Drug effects were expressed versus baseline (predrug) firing rate (defined as 0%).

Influence of S33138 upon the Induction of HTW by the 5-HT_2A Receptor Agonist, DOI, in Rats. Using a procedure described previously (Schreiber et al., 1995), DOI (2.5 mg/kg i.p.) was injected into rats placed in Plexiglas observation cages (33.5 × 23.5 × 19 cm). Five minutes after DOI, the number of HTW made in 5 min was counted. Vehicle, S33138, or the selective 5-HT_2A receptor antagonist MDL100,907 was given s.c. 30 min before DOI.

Influence of S33138 upon the Induction of Hypothermia by the 5-HT₇ Receptor Agonist, 5-Carboxytryptamine (5-CT), in Guinea Pigs. Essentially as described elsewhere (Hagan et al., 2000), CT was determined in lightly restrained male guinea pigs with a digital thermistoprobe placed into the ear for 30 s. Basal CT was measured; S33138, the selective 5-HT₇ receptor antagonist SB269,970, or vehicle was administered i.p.; and 30 min later, 5-CT (3 mg/kg i.p.) or vehicle was injected. Ninety minutes later, CT was reevaluated, and the difference (Δ) to pretreatment values were calculated.

Drug Salts and Sources. Haloperidol and S33138 were dissolved in sterile water plus a few drops of lactic acid, and the pH was adjusted close to neutrality (pH ≥5.5) using NaOH. All other drugs were dissolved in sterile water. Doses are in terms of the base. Drug structures, sources, and salts were as follows: MPTP HCl was purchased from Sigma (Dorset, UK). Clonidine HCl, DOI HCl, 7-OH-DPAT HCl, 8-OH-DPAT HBr, PD128,907 HCl, and SCH23390 HCl were obtained from Research Biochemicals International (Natick, MA). Apomorphine, idazoxan, haloperidol, SKF81297, and benserazide HCl were acquired from Sigma Chimie (St Quentin-Fallavier, France). S33138 HCl, 5-carboxytryptamine, maleate, and MDL100,907 were synthesized by G. Lavielle (Servier). Ropinirole, SB269,970, and WAY100,635 were synthesized by J.-L. Péglion (Servier).

Results

Influence of S33138 upon c-fos Gene Expression in Cerebral Structures Differentially Expressing D₃ Compared to D₂ Receptors (Figs.1and2). S33138 dose-dependently (0.04–2.5 mg/kg i.p.) elevated levels of mRNA encoding c-fos in the IOC of mice, a limbic region containing D₃ but not D₂ receptors. S33138 also significantly elevated c-fos gene expression in the nucleus accumbens, a region enriched in D₃ receptors. In contrast, only a “high” dose of S33138 significantly elevated the levels of mRNA encoding c-fos in the striatum, a D₂ receptor-rich structure possessing a few D₃ receptors. Haloperidol showed pronounced effects in all structures, with the greatest increase in the striatum. Values expressed relative to vehicle (100%) are as follows: IOC, vehicle = 100.0 ± 19.0; haloperidol, 0.5 mg/kg, 177.8 ± 49.2 and haloperidol, 2.0 mg/kg, 175.9 ± 19.0, F(2,22) = 3.3, P < 0.05; nucleus accumbens, vehicle = 100.0 ± 10.0; haloperidol, 0.5 mg/kg, 189.8 ± 36.5 and haloperidol 2.0 mg/kg, 189.7 ± 25.1, F(2,19) = 5.9, P < 0.05; and striatum, vehicle = 100.0 ± 7.0; haloperidol, 0.5 mg/kg, 193.9 ± 21.5 and haloperidol, 2.0 mg/kg, 340.1 ± 53.3, F(2,18) = 15.5, P < 0.001.

Influence of Chronic Administration of S33138 on Spontaneously Active Dopaminergic Neurones in Anesthetized Rats (Fig. 3). After chronic (3 week) daily p.o. administration of vehicle, 1.21 ± 0.13 active neurons were identified in the VTA. Chronic administration of S33138 (0.16–10.0 mg/kg p.o.) potently and dose-dependently reduced the number of active neurons. In contrast, S33138 decreased spontaneously active neurons in the SNPC only at the highest dose of 10.0 mg/kg. Moreover, the maximal reduction in number of spontaneously active neurons (–58%) was less than that for the VTA (–83%). Haloperidol (0.5 mg/kg p.o.) produced a significant (P < 0.01) reduction in the number of active neurons in the VTA (–66 + 6.6%) and in the SNPC (–62 ± 6.2%) compared to vehicle (0%).

Influence of S33138 upon Actions of the D₃/D₂ Receptor Agonist, Ropinirole, in Parkinsonian Primates (Fig. 4). In marmosets exposed to MPTP, ropinirole (0.63 mg/kg p.o.) increased motor activity compared to vehicle: 3091 ± 556 versus 235 ± 132 counts, respectively (P < 0.01 in a two-tailed t test). Over a low-dose range, S33138 (0.0025–0.16 mg/kg p.o.) dose-dependently and significantly enhanced this facilitatory influence of ropinirole. In contrast, a higher dose of S33138 (2.5 mg/kg) reduced the action of ropinirole. S33138 (0.16 mg/kg) did not significantly modify motor behavior when given alone (Fig. 4; data not shown).

Antagonism by S33138 of the Hypothermia Elicited by the Preferential D₃ versus D₂ Receptor Agonist, 7-OH-DPAT (Fig. 5A). 7-OH-DPAT (0.16 mg/kg s.c.) elicited a pronounced reduction in CT in rats. This hypothermia was attenuated by S33138 (0.04–2.5 mg/kg s.c.), which did not affect CT alone. Likewise, haloperidol blocked 7-OH-DPAT-induced hypothermia: vehicle + 7-OH-DPAT = –1.44 ± 0.15°C; haloperidol (0.02 mg/kg s.c.) + 7-OH-DPAT = –0.65 ± 0.12°C; haloperidol (0.04) + 7-OH-DPAT = 0.07 ± 0.15°C; and haloperidol (0.16) + 7-OH-DPAT = 0.57 ± 0.10°C, n = 6–8 per dose, F(3,30) = 31.5, P < 0.01. Doses of 0.02, 0.04, and 0.16 were significantly different from vehicle in Dunnett's test. Haloperidol did not affect CT alone (data not shown).

Antagonism by S33138 of Yawns Elicited by 7-OH-DPAT (Fig. 5B). 7-OH-DPAT (0.04 mg/kg s.c.) elicited yawning in rats. This response was blocked by S33138 (0.63–10.0 mg/kg s.c.), which alone did not elicit yawns (data not shown). Haloperidol also blocked yawns elicited by 7-OH-DPAT as follows: vehicle + 7-OH-DPAT = 11.2 ± 1.1; haloperidol (0.01 mg/kg, s.c.) + 7-OH-DPAT = 9.4 ± 0.9; haloperidol (0.04) + 7-OH-DPAT = 6.1 ± 1.0; and haloperidol (0.16) + 7-OH-DPAT = 1.4 ± 0.7, n = 7–8 per dose, F(3,30) = 18.6, P < 0.01. Doses of 0.04 and 0.16 were significantly different from vehicle (Dunnett's test). Haloperidol did not elicit yawns alone (data not shown).

Antagonism by S33138 of the Discriminative Stimulus Properties of the Preferential D₃ versus D₂ Receptor Agonist, PD128,907 (Fig. 5C). PD128,907 (0.16 mg/kg i.p.) elicited a stable DS in rats. S33138 (0.0025–0.63 mg/kg s.c.) blocked its DS properties with an ED₅₀ (95% confidence limits) of 0.02 (0.01–0.04) mg/kg s.c. S33138 did not significantly decrease response rates: for the dose of 0.63 mg/kg, +23 ± 18% versus previous control session (0%). Haloperidol (0.0025–0.04 mg/kg s.c.) also attenuated the PD128,907 cue, exerting a maximal 50% inhibition at 0.04 mg/kg (n = 7). At this dose, there was a pronounced decrease (–75 ± 32%, P < 0.01, paired t test) in the response rate. At a higher dose of haloperidol (0.08 mg/kg), rats were unable to select a lever because of motor perturbation.

Antagonism by S33138 of the Inhibitory Influence of PD128,907 upon the Electrical Activity of Dopaminergic Cell Bodies (Fig. 6). The firing rate (baseline, 4.0 ± 0.3 Hz) of dopaminergic perikarya in the VTA was markedly reduced by PD128,907 (0.01 mg/kg i.v.). S33138 (one dose per subject, 0.001–0.25 mg/kg i.v.) blocked this action of PD128,907 with a minimally effective dose of 0.004 mg/kg i.v. and without itself affecting firing rate. Likewise, haloperidol (0.031 mg/kg i.v.) reversed the action of PD128,907 and yielded an “overshoot” in firing rate relative to baseline values (0%): vehicle/vehicle, + 0.1 ± 2.7%; PD128,907/vehicle, –99.9 ± 0.1%; vehicle/haloperidol, + 46.1 ± 25.6% and PD128,907/haloperidol + 58.7 ± 36.5%, F(3,18) = 10.3, P < 0.01. In a separate experiment, cumulatively administered over a higher range of doses (0.125–4.0 mg/kg i.v.), S33138 slightly increased firing rate with a maximal effect of +29% (1.0 mg/kg). Apomorphine (0.031 mg/kg i.v.) failed to inhibit dopaminergic neurons (+1.7 ± 5.5%) following administration of S33138 compared to vehicle (–98.2 ± 0.4%): P < 0.01 (two-tailed t test). After haloperidol, apomorphine was likewise ineffective (data not shown).

Antagonism by S33138 of the Suppressive Influence of PD128,907 upon DA Levels in the FCX, Nucleus Accumbens and Striatum (Fig. 7). PD128,907 (0.16 mg/kg s.c.) decreased dialysis levels of DA in the FCX of freely moving rats, an action dose-dependently prevented by S33138 (0.16–2.5 mg/kg s.c.), which did not influence DA alone. Haloperidol (0.63 mg/kg s.c.) also blocked the action of PD128,907 and itself increased levels of DA. Relative to basal values (100%), “area under the curve” analysis was as follows: vehicle/vehicle, 98.1 ± 1.8%; vehicle/PD128,907 (0.16), 82.2 ± 3.1%; haloperidol/vehicle, 127.6 ± 2.9% and haloperidol/PD128,907, 132.2 ± 5.9%; n = 5–9 per group. Influence of PD128,907, F(1,16) = 10.0, P < 0.01; influence of haloperidol, F(1,15) = 25.6, P < 0.01 and interaction, F(1,12) = 19.0, P < 0.01. S33138 likewise abolished the influence of PD128,907 (0.16 mg/kg s.c.) on levels of DA in the nucleus accumbens and the striatum (Fig. 7, C and D). Haloperidol (0.63 mg/kg s.c.) also blocked the actions of PD128,907 in nucleus accumbens and striatum and alone increased levels of DA. Nucleus accumbens: vehicle/vehicle, 93.7 ± 1.9%; vehicle/PD128,907, 75.0 ± 3.2%; haloperidol/vehicle, 153.1 ± 3.2% and haloperidol/PD128,907, 154.8 ± 8.2%; n = 5–8 per group. Influence of PD128,907, F(1,13) = 6.9, P < 0.05, influence of haloperidol, F(1,12) = 55.3, P < 0.01 and interaction, F(1,11) = 18.7, P < 0.01. Striatum: vehicle/vehicle, 96.5 ± 1.7%; vehicle/PD128,907, 67.2 ± 2.7%; haloperidol/vehicle, 179.0 ± 5.6% and haloperidol/PD128,907, 205.7 ± 4.0%; n = 5–8 per group. Influence of PD128,907, F(1,14) = 20.1, P < 0.01; influence of haloperidol, F(1,13) = 25.6, P < 0.01 and interaction, F(1,11) = 305.0, P < 0.01.

Lack of Inhibition by S33138 of Rotation Elicited bythe D₁ Receptor Agonist, SKF81297, in Rats with a Unilateral Lesion of the SNPC. SKF81297 (0.04 mg/kg s.c.) elicited marked contralateral rotation: 595.7 ± 74.3 versus vehicle, 10.1 ± 3.0 contralateral turns (P < 0.01 in an unpaired Student's t test). This action of SKF81297 was not significantly affected (matched paired test) by S33138 (2.5–40.0 mg/kg s.c.). Vehicle + SKF81297, 609 ± 81; S33138 (0.16 mg/kg s.c.) + SKF81297, 602 ± 108 and S33138 (10.0) + SKF81297, 760 ± 57; n = 5–8 per dose. In contrast, the potent (rat D₁ receptors, pK_i 8.9) D₁ receptor antagonist, SCH23390, blocked induction of rotation by SKF81297. Vehicle + SKF81297, 580 ± 92; SCH23390 (0.0025) + SKF81297, 484 ± 64; SCH23390 (0.01) + SKF81297, 311 ± 66, SCH23390 (0.02) + SKF81297, 243 ± 77 and SCH23390 (0.04) + SKF81297, 56 ± 34; n = 5–6 per dose. F(4,28) = 5.2, P < 0.01. Doses of 0.02 and 0.04, significantly different from vehicle (matched pairs tests).

Influence of S33138 upon the Electrical Activity of Adrenergic and Serotonergic Cell Bodies in Anesthetized Rats (Fig. 8A). S33138 dose-dependently (0.125–4.0 mg/kg i.v.) increased the activity of LC-localized adrenergic neurons (baseline firing rate, 1.1 ± 0.2 Hz). Relative to basal values (0%), S33138 exerted a maximal effect of +177 ± 7% at 4.0 mg/kg. Following S33138, the α₂-AR agonist, clonidine (0.01 mg/kg i.v.), still inhibited firing (–100.0 ± 0.0%). This effect was reversed by the α₂-AR (rat α₂-adrenoceptors, pK_i, 8.5) antagonist, idazoxan (0.063 mg/kg i.v.), which yielded an “overshoot” of +81 ± 16.3%. Alone, idazoxan (0.063 mg/kg i.v.; n = 5) significantly (+142.0 ± 13.9%, P < 0.01) increased the firing rate relative to vehicle (n = 5), (–2.2 ± 1.3%). In contrast to the LC, S33138 (0.125–4.0 mg/kg i.v.) did not affect the electrical activity of DRN-localized serotonergic neurons (baseline firing rate, 1.2 ± 0.2 Hz). Following S33138, the 5-HT_1A receptor agonist, 8-OH-DPAT (0.005 mg/kg i.v.), still abolished the activity of serotonergic perikarya (–97.6 ± 2.4%). This effect was reversed by the 5-HT_1A receptor antagonist, WAY100,635 (0.1 mg/kg, i.v.: +29.0 ± 12.8% (P < 0.01).

Antagonism by S33138 of the Induction of Head Twitches in Rats by the 5-HT_2A Receptor Agonist, DOI (Fig. 8B). DOI elicited HTW in rats. S33138, which did not itself elicit HTW (data not shown), dose-dependently (0.63–10.0 mg/kg s.c.) abrogated their induction. The potent (rat 5-HT_2A receptors, pK_i, 9.1) 5-HT_2A receptor antagonist, MDL100,907, also abolished induction of HTW: vehicle + DOI, 6.1 ± 0.7; MDL100,907 (0.0025 mg/kg s.c.) + DOI, 5.2 ± 0.5; MDL100,907 (0.005) + DOI, 3.0 ± 1.1; MDL100,907 (0.01) + DOI, 0.8 ± 0.5 and MDL100,907 (0.04) + DOI, 0.0 ± 0.0; n = 5–8 per value, F(4,31) = 11.0, P < 0.01. Doses of 0.01 and 0.04, significantly different (P < 0.05) from vehicle in Dunnett's test. MDL100,907 itself did not elicit HTW (data not shown).

Antagonism by S33138 of the Induction of Hypothermia by the “5-HT₇” Receptor Agonist, 5-Carboxytryptamine, in Guinea Pigs (Fig. 8C). 5-Carboxytryptamine (3.0 mg/kg i.p.) elicited hypothermia in guinea pigs, an effect dose-dependently attenuated by S33138 (2.5–20.0 mg/kg i.p.). S33138 did not modify CT alone (data not shown). By analogy, the potent (rat 5-HT₇ receptors, pK_i, 8.8) 5-HT₇ receptor antagonist, SB269,970, also inhibited induction of hypothermia: Vehicle + 5-CT, –1.81 ± 0.30°C; SB269,970 (0.63 mg/kg i.p.) + 5-CT, –1.56 ± 0.22°C, SB269,970 (2.5) + 5-CT, –0.50 ± 0.36°C and SB269,970 (10.0) + 5-CT, –0.22 ± 0.23°C; n = 5–8 per value, F(3,25) = 4.4, P < 0.01. Doses of 2.5 and 10.0, significantly different (P < 0.05) from vehicle in Dunnett's test. SB269,970 did not affect CT alone (data not shown).

Discussion

Selective Induction of c-fos by S33138 in Limbic Structures. The preferential induction of c-fos by S33138 in D₃ receptor-rich limbic versus striatal regions (Landwehrmeyer et al., 1993; Joyce, 2001; Sokoloff et al., 2006) correlates well with its higher affinity for D₃ over D₂ sites. Furthermore, this finding parallels similar actions of selective antagonists at D₃ versus D₂ receptors (Kovacs et al., 2001; Southam et al., 2007; P. Svenningsson and M. J. Millan, unpublished observations). Interestingly, functional “MRI” imaging studies also revealed greater activation of neurones in limbic versus striatal structures by selective blockade of D₃ receptors (Schwarz et al., 2004). Preferential limbic versus striatal induction of c-fos has been related to a high margin between doses controlling positive symptoms versus those eliciting extrapyramidal symptoms, and a similar pattern of effects to S33138 is seen with atypical antipsychotics (Robertson et al., 2004; Guo et al., 1998). However, it is unclear whether D₃ receptors are involved in their effects. Thus, preferential D₃ versus D₂ receptor agonists blunted limbic induction of c-fos by clozapine in rats (Vahid-Ansari and Robertson, 1996; Guo et al., 1998). However, mice genetically deprived of D₃ receptors revealed no attenuation in the influence of clozapine, which may engage cells different from D₃ antagonists (Merchant et al., 1996; Carta and Gerfen 1999; Kovacs et al., 2001).

Inhibition of Spontaneously Active VTA Neurones by Long-Term Administration of S33138. By analogy to clozapine and olanzapine (Stockton and Rasmussen, 1996; Grace et al., 1997; Ashby and Wang, 1996), and in contrast to haloperidol, chronic administration of S33138 selectively decreased the number of spontaneously active VTA versus SNPC dopaminergic neurones. Supporting a role of D₃ receptor blockade, the action of S33138 was expressed at low doses similar to those inducing c-fos in limbic structures. Furthermore, a selective reduction in spontaneously active VTA versus SNPC neurones was likewise seen employing highly selective D₃ antagonists (Ashby et al., 2000; MacDonald et al., 2003). Chronic administration of haloperidol may decrease the number of spontaneously active VTA and SNPC neurones by “depolarization” blockade because manipulations that hyperpolarize neurones, such as apomorphine administration, reverse its effects (Grace et al., 1997). However, the influence of chronic SB-277011, a selective D₃ receptor antagonist, upon VTA neurones was not blocked by apomorphine (Ashby et al., 2000). Accordingly, depolarization blockade is probably not involved in long-term effects of S33138. Rather, the reduction in spontaneously active VTA neurones produced by S33138 and selective D₃ receptor antagonists may reflect feedback actions initiated in structures like the nucleus accumbens (Grace et al., 1997; Ashby et al., 2000).

Dose-Dependent Facilitation and Attenuation by S33138 of the Antiparkinson Properties of Ropinirole. MPTP selectively lesions nigrostriatal dopaminergic projections, provoking a parkinsonian-like syndrome alleviated by D₂/D₃ receptor agonists (Pearce et al., 1998; Millan et al., 2004b; Silverdale et al., 2004). Blockade of D₂ receptors by a high dose of S33138 probably underlies its abrogation of the motor-relief afforded by ropinirole. By contrast, mirroring the facilitatory influence of selective D₃ receptor antagonists (Silverdale et al., 2004; Hill et al., 2006), low doses of S33138 potentiated the actions of ropinirole. These findings agree well with the notion that D₂ receptors mediate the beneficial actions of antiparkinson agents, whereas D₃ receptor blockade enhances their efficacy and reduces dyskinesias (Bezard et al., 2003; Millan et al., 2004b; Silverdale et al., 2004; Hill et al., 2006; Sokoloff et al., 2006).

Antagonist Properties of S33138 in Behavioral Models. Although D₂ receptors mediate hypothermia in mice (Boulay et al., 1999; Perachon et al., 2000), a predominant role of D₃ sites has been implicated in rats (Millan et al., 2000a; Chaperon et al., 2003). Furthermore, 7-OH-DPAT elicits hypothermia upon microinjection into the IOC, a postsynaptic action unaffected by elimination of dopaminergic neurones (Barik and Beaurepaire, 1998). Nonetheless, a role of both D₃ and (at higher doses) D₂ sites in the antagonism of 7-OH-DPAT-induced hypothermia by S33138 is possible. Collins et al. (2005) asserted that stimulation of D₃ and D₂ receptors, respectively, mediates and opposes induction of yawns by PD128,907 in rats. However, a high dose (56.0 mg/kg s.c.) of the D₃ antagonist, SB-277011, was needed for full blockade, whereas the D₂ antagonist, L741,626, which enhanced yawns, may have been “underdosed” at 1.0 mg/kg (Bristow et al., 1998; Millan et al., 2000a, 2004b; Pak et al., 2006). Moreover, dopaminergic agonists elicit yawns by postsynaptic actions in the paraventricular nucleus, a structure lacking D₃ receptors (Argiolas and Melis 1998; Chen et al., 1999; Joyce 2001); yawns evoked by 7-OH-DPAT are unaffected by selective D₃ antagonists yet abolished by L741,626 (Millan et al., 2000a, 2004b). Accordingly, blockade by S33138 probably reflects its antagonism of D₂ receptors. DS properties of low doses of dopaminergic agonists principally reflect recruitment of presynaptic D₃ and/or D₂ receptors (Cory-Slechta et al., 1996; Bristow et al., 1998; Millan et al., 2000b). Inasmuch as the PD128,907 cue is resistant to selective D₃ receptor antagonists, potent blockade by S33138 was unexpected. Nonetheless, potent antagonism of S33138 at presynaptic D₃/D₂ sites is underpinned by electrophysiological studies (see below), and S33138 is a potent antagonist at D₃/D₂ heterodimers, which may form in dopaminergic neurones (Millan et al., 2007).

Antagonist Properties of S33138 at D₃/D₂ Autoreceptors. Attenuation of the inhibitory influence of PD128,907 upon VTA firing by acute S33138 at doses as low as 0.004 mg/kg i.v. mimics findings with selective D₃ antagonists; this action probably reflects blockade of D₃ autoreceptors, activation of which phasically inhibits DA release and neuronal firing (Millan et al., 2000a,c,d, 2004a; Reavill et al., 2000; Drescher et al., 2005; Roberts et al., 2006). Conversely, full blockade of PD128,907 by higher doses of S33138 presumably reflects additional occupation of D₂ receptors (Usiello et al., 2000). A comparable degree of “resolution” cannot be achieved in dialysis studies, but blockade by S33138 of the PD128,907-induced reduction of DA release in FCX, nucleus accumbens, and striatum (Millan et al., 2000c, 2004a; Reavill et al., 2000; Roberts et al., 2006) corroborates its antagonist properties at D₃ (and D₂) autoreceptors. Mimicking selective D₃ antagonists, and in contrast to D₃/D₂ and preferential D₂ antagonists, S33138 neither activated the VTA nor elevated extracellular levels of DA (Millan et al., 2000c,d; Roberts et al., 2006). There are several possible explanations. First, D₃ autoreceptors are not tonically active. Furthermore, even when occluded by low doses of S33138, DA has access to inhibitory D₂ autoreceptors (Millan et al., 2000c,d; Sokoloff et al., 2006). Second, extinction of “spontaneous” coupling at constitutively active D₂ autoreceptors may account for activation of dopaminergic projections by haloperidol, an inverse agonist (Nilsson et al., 1996; Millan et al., 2000d); the possibility that S33138 behaves as a neutral antagonist is under exploration. Third, at high doses, antagonism by S33138 of excitatory, VTA-localized 5-HT_2A receptors may dampen disinhibition of dopaminergic cell bodies via D₂ receptor blockade (Millan et al., 2000d; Minabe et al., 2001). Regardless of the underlying reasons, this observation is intriguing because increased mesolimbic DA release provoked by haloperidol may interfere with concurrent antagonism of postsynaptic D₂/D₃ receptors and underlie patient resistance.

Lack of Antagonist Properties of S33138 at D₁ Receptors. In rats sustaining a unilateral lesion of the SNPC, D₁ receptor agonists elicit contralateral rotation, principally via supersensitive striatal D₁ receptors, although D₁ sites in the substantia nigra pars reticulata may also be involved (Asin and Montana, 1988; Gulwadi et al., 2001; Gerfen et al., 2002). This response is abolished by D₁ antagonists such as SCH23390. Although S33138 possesses mild affinity for D₁ receptors (Millan et al., 2007), it is >100-fold less potent than SCH23390 and did not influence SKF81297-induced rotation. Accordingly, D₁ antagonism is unlikely to be an important component of its pharmacological properties.

Influence of S33138 upon Noradrenergic Neurones: Blockade of α_2C-ARs. Although S33138 excited noradrenergic perikarya (Millan et al., 2000d; Invernizzi and Garattini, 2004), D₃/D₂ receptors are unlikely to be involved because active doses were substantially higher than those blocking the influence of PD128,907 on dopaminergic neurones. In addition, selective D₃ receptor antagonists do not excite the LC (Millan et al., 2000c). Rather, antagonist properties of S33138 at α_2C-AR autoreceptors may be implicated inasmuch they inhibit adrenergic cell firing (Arima, 1998; Millan et al., 2000d; Owesson et al., 2003). Nonetheless, the response of the LC to agonists is mainly transduced via α_2A-ARs (Mateo and Meana, 1999; Pudovkina et al., 2001). Accordingly, the lack of affinity of S33138 for α_2A-ARs accounts for its inability, in distinction to idazoxan, to prevent the suppressive influence of clonidine upon LC firing. These electrophysiological data provide a “surrogate” measure of the ability of S33138 to block postsynaptic populations of α_2C-ARs inhibitory to motor function, mood, and cognition (Svensson, 2003; Marcus et al., 2005).

Antagonist Properties of S33138 at 5-HT_2A and 5-HT₇ Receptors. Because D₃ receptor antagonists do not affect DOI-induced HTW, the modest antagonist properties of S33138 at 5-HT_2A receptors (Millan et al., 2007) account for its inhibition of DOI-evoked HTW, which are abolished by selective 5-HT_2A receptor antagonists, such as MDL100,907 (Schreiber et al., 1995). 5-HT_2A sites mediating HTW are localized in the FCX (Willins and Meltzer, 1997), a structure implicated in the mood and cognitive deficits of schizophrenia, and 5-HT_2A receptor blockade contributes to the atypical profile of certain antipsychotics, including their low extrapyramidal symptom potential (Meltzer et al., 2003). Activation of hypothalamic 5-HT₇ sites elicits hypothermia, an effect blocked by the selective antagonist, SB269,970 (Hagan et al., 2000; Guscott et al., 2003). In relation, prevention of 5-CT-induced hypothermia by S33138 supports its abrogation of h5-HT₇ receptor-coupled cAMP formation (Millan et al., 2007). Whereas blockade of 5-HT₇ receptors is unlikely to control psychosis per se (Pouzet et al., 2002; Meltzer et al., 2003), it may exert a beneficial effect on comorbid symptoms, such as depressed mood and poor sleep (Hedlund and Sutcliffe, 2004).

Conclusions. In line with cellular studies (Millan et al., 2007), these data demonstrate modest antagonist properties of S33138 at central α_2C-ARs, 5-HT_2A receptors and 5-HT₇ receptors. Most importantly, in line with antagonist properties at human D₃ and D_2L/D_2S receptors, they demonstrate that S33138 blocks cerebral populations of presynaptic and postsynaptic D₃ and D₂ receptors. Furthermore, mimicking in vitro studies, the present in vivo observations support preferential actions of S33138 at D₃ versus D₂ receptors in the central nervous system. Thus, doses of S33138-inducing c-fos in the D₃ receptor-rich IOC were 16-fold lower than those elevating c-fos in striatum, and a similar ratio was seen for doses of S33138 that reduced motor actions of ropinirole in Parkinsonian primates versus those that facilitated its actions. Although comparisons are less easy to make, the more potent inhibition by S33138 of the spontaneous firing of neurons in the VTA versus SNPC is also consistent with more marked D₃ receptor antagonism. A limitation of these analyses is, of course, the contrasting routes, species, and treatment-durations employed, all inherent to the various procedures. Furthermore, whereas PET studies of S33138 in man show dose-dependent (10–70 mg p.o.) occupation of [¹¹C]raclopride-labeled D₂/D₃ sites in basal ganglia (unpublished observations), radiolabeled antagonists differentiating D₃ and D₂ receptors are unavailable. Studies in mice genetically lacking D₃ or D₂ receptors would be useful in further characterizing the roles of D₃ versus D₂ receptors in the actions of S33138. However, certain procedures cannot be performed in this species; “knock-out” mice have their own limitations; and information on novel drugs from nongenetically modified animals is most relevant to their potential actions in man. Finally, it would be interesting to evaluate the actions of other antipsychotics at cerebral D₃ versus D₂ receptors in the procedures employed herein.

Thus, further study of the roles of D₃ and D₂ receptors in the actions of S31338 would be instructive, but the present in vivo and in vitro studies collectively suggest that it acts as a preferential and not selective antagonist of D₃ versus D₂ receptors. These observations provide an instructive framework for interpreting its potential antipsychotic properties in experimental studies and in ongoing (Phase II) clinical trials in man.