The dual orexin receptor antagonist almorexant, alone and in combination with morphine, cocaine and amphetamine, on conditioned place preference and locomotor sensitization in the rat

Abstract

Dual orexin receptor (OXR) antagonists emerge as a novel therapeutic class to treat insomnia that, based on anti-addictive effects of selective OXR type 1 antagonists in rats, might be associated with less abuse liability than commonly used γ-aminobutyric acid (GABA) receptor modulators. Here, we studied the effects of the sleep-enabling dual OXR antagonist almorexant on conditioned place preference (CPP) and locomotor sensitization in rats. First, we compared almorexant to the GABA metabolite γ-hydroxybutyrate (GHB), which is clinically used as a sleep-inducing drug and which is associated with mild abuse liability. Whereas conditioning with GHB induced significant place preference, conditioning with almorexant did not. Second, we tested the potential of almorexant to interfere with the conditioned rewarding or locomotor sensitizing effects related to psychostimulants or opiates. Almorexant attenuated the expression of CPP to high doses of cocaine (15 mg/kg) and d.l-amphetamine (2 mg/kg), but not to high dose of morphine (10 mg/kg). Conversely, almorexant interfered with the expression of locomotor sensitization to morphine, but not with that to cocaine and d.l-amphetamine. Third, we observed that chronic almorexant (12 d) treatment in morphine, cocaine or amphetamine pre-conditioned and locomotor-sensitized rats had no influence on the maintenance of CPP and locomotor sensitization when tested after almorexant washout. Our findings suggest that almorexant itself does not exert conditioned rewarding effects in the rat and that it may acutely interfere with the expression of CPP or locomotor sensitization in a drug-dependent manner (monoaminergic psychostimulants vs. opiates).

Addiction, almorexant, conditioned place preference, locomotor sensitization, orexin

Introduction

The orexin (OX) neuropeptide system, consisting of orexin-A and -B (OX_A/B) and the two G-protein coupled orexin receptors, type 1 and type 2 (OXR_1/2), plays a major role in the regulation of vigilance and wakefulness (Saper et al.2005; Tsujino & Sakurai, 2009). As such, it coordinates the arousal state of the organism with other physiological requirements, such as the demand for food and the reaction to environmental stress and reward (Sakurai et al.2010; Sinton, 2011).

Blocking OXRs pharmacologically has recently been explored in primary insomnia and the first dual OXR antagonists have been successfully tested in clinical settings (Brisbare-Roch et al.2007; Hoever et al.2010; Winrow et al.2011). A common side-effect of current sleep medications, which are mostly based on modulating γ-aminobutyric acid (GABA)_A or GABA_B receptor transmission, such as benzodiazepine-like drugs and γ-hydroxybutyrate (GHB), is their reinforcing properties, which lead to abuse liability in patients (Griffiths & Johnson, 2005; Licata & Rowlett, 2008). Whether abuse potential is intrinsically linked to sleep promotion by dual OXR antagonism is currently unknown.

Both anatomical and functional evidence suggests an important positive interaction of the endogenous OX system with reward pathways of the brain. OX fibres project to the ventral tegmental area (VTA) and to the nucleus accumbens (Balcita-Pedicino & Sesack, 2007; Baldo et al.2003; Fadel & Deutch, 2002), two crucial brain structures of the mesolimbic dopamine reward system. Both OXRs are expressed on dopaminergic and GABAergic neurons within the VTA (Korotkova et al.2003; Narita et al.2006) and direct VTA infusions of OX_A and OX_B increase dopamine efflux within the nucleus accumbens and are able to induce conditioned place preference (CPP; Narita et al.2007). Moreover, OX neurons become activated upon exposure to reward-related contextual cues (Harris et al.2005) and the chronic administration of drugs of abuse leads to long-term changes in OX and/or OXR expression (Kane et al.2000; Zhang et al.2007).

Decreased OXR signalling reduces several addiction-related effects of drugs of abuse. For instance, OX-deficient mice show reduced CPP to morphine and less severe morphine withdrawal signs (Georgescu et al.2003; Narita et al.2006). Selective pharmacological OXR₁ blockade reduces cue- and stress-induced reinstatement of cocaine seeking (Boutrel et al.2005; Smith et al.2009, 2010), cue-induced reinstatement of alcohol seeking (Jupp et al.2011; Lawrence et al.2006) and nicotine self-administration (Hollander et al.2008; LeSage et al.2010). OXR₁ antagonism also attenuates the expression of amphetamine- and cocaine-induced CPP (Gozzi et al.2011; Hutcheson et al.2011), and reduces the expression or development of locomotor sensitization to amphetamine and cocaine (Borgland et al.2006; Quarta et al.2010).

Despite substantial knowledge of the effects of OXR₁ antagonism with regard to reward pathways, data on the effects of selective OXR₂ blockade (Gozzi et al.2011; Shoblock et al.2011; Smith et al.2009) and dual OXR_1/2 antagonism (LeSage et al.2010; Winrow et al.2010) are still sparse, possibly due to the lack of suitable, freely available tool compounds.

Here, we investigated both CPP and locomotor sensitization as two different outcomes associated with exposure to drugs of abuse. We first evaluated the effects of almorexant per se in comparison to those of the sleep-inducer GHB and then we explored almorexant's potential to acutely interfere with the expression of CPP and locomotor sensitization induced by morphine, cocaine and amphetamine. Last, we evaluated whether chronic almorexant treatment could affect the maintenance of pre-established CPP and locomotor sensitization to those drugs of abuse.

Methods

Animals

Male Sprague–Dawley rats (Harlan, The Netherlands) were purchased at an age of 7–8 wk, group-housed under a regular 12-h light/dark schedule (lights on at 06:00 hours) and kept under standard conditions with ad libitum food and water. Rats weighed between 250 and 300 g at the beginning of the experiments, which were performed during the light phase (10:00–18:00 hours) under 900 lx light conditions. All experimental procedures were approved by the local Veterinary Office and strictly adhered to Swiss federal regulations on animal experimentation.

Drugs and formulations

Almorexant (ACT-078573-hydrochloride; Actelion Pharmaceuticals Ltd, Switzerland) was formulated in 0.25% methyl cellulose at 5 ml/kg and administered by oral gavage. Sodium GHB (sodium oxybate, Xyrem^®; UCB Pharma AG, Switzerland) was formulated in purified water (pH adjusted to 7.5 with malic acid) at 5 ml/kg for oral administration. Morphine hydrochloride (Haenseler AG, Switzerland), cocaine hydrochloride (Haenseler AG) and d.l-amphetamine (Actelion) were formulated in 0.9% NaCl at 5 ml/kg and administered by i.p. injection. All drug concentrations and doses were calculated as the free base.

Behavioural testing

CPP apparatus

The CPP apparatus (PanLab, Spain) consisted of two equal-sized plastic compartments (40 × 33 × 45 cm), which were connected by a smaller rectangular transition corridor (25 × 13 × 45 cm) via removable guillotine doors. Both compartments were used for conditioning, and were made distinguishable using different visual and tactile cues (wall pattern, floor texture), which themselves did not induce an initial preference for either compartment (data not shown). Behaviour was monitored via video-tracking (Viewpoint, France).

CPP protocol

The CPP protocol consisted of a 10-d schedule with a pre-conditioning preference test (pre-test) on day 1, an 8-d conditioning phase (days 2–9) and a post-conditioning preference test (test) on day 10. The pre-test and test were identical in design: rats were placed in the centre of the small transition corridor, the guillotine doors were opened and the rats were given free access to the two conditioning compartments for 15 min. The time the animals spent in each of the three compartments was recorded by video-tracking and is shown in the figures. The small transition compartment was regarded as a neutral zone.

During the conditioning phase (days 2–9), rats were alternately treated, once each day, with a particular drug or the corresponding vehicle (e.g. saline) followed by confinement in the associated conditioning compartments for 45 min. Thus, in each experiment rats received, in total, four pairings of the drug with one compartment and four pairings of the vehicle with the other compartment. The compartment chosen for drug- and vehicle-pairings was alternated between rats within each experimental group. The total distance moved in the compartments during each drug- and vehicle-pairing was recorded by video-tracking.

Expt 1: effect of almorexant on CPP and locomotor activity

Following a pre-test, alternating almorexant and vehicle oral administrations were paired with confinement to different CPP compartments for a total of four administrations each over 8 d, as described above. Rats were confined to the compartments either immediately or 2 h after treatment (for both the 100 and 300 mg/kg doses). After the conditioning phase, place preference was tested under drug-free conditions.

Treatment doses of almorexant were chosen at the high end of the sleep-enabling dose–response curve in rats (Brisbare-Roch et al.2007). Oral treatment (100 and 300 mg/kg) reduces latency to persistent non-rapid eye movement (non-REM) sleep within 30 min of oral administration, suggesting that almorexant brain concentrations within the 45 min-lasting CPP conditioning sessions reached levels sufficient to exert psychotropic effects. Both doses were, however, also tested with a 2 h pretreatment time, when brain concentrations and pharmacodynamic effects achieved during CPP conditioning (between 2 and 2 h and 45 min after administration) are expected to be near maximal. A steady-state plasma concentration of about 2 µm almorexant is required to achieve 90% human OXR₂ occupancy in a transgenic rat model over-expressing human OXR₂ (Winrow et al.2012). Three hours following oral gavage of 100 and 300 mg/kg almorexant in the rat, brain concentrations reach about 2 µm and 6 µm, respectively (Brisbare-Roch et al.2007). Thus, estimated OXR₂ occupancy at both doses is expected to be ⩾90%.

Expt 2: effect of GHB on CPP and locomotor activity

Following the pre-test, rats were conditioned during four context-pairings with oral administrations of GHB (350 or 700 mg/kg; 30 min pretreatment time) according to previous reports (Fattore et al.2000; Watson et al.2010). Thereafter, place preference was tested under drug-free conditions. The shorter pretreatment time-point for GHB as compared to almorexant was chosen based on its different pharmacokinetic/pharmakodynamic properties (Brisbare-Roch et al.2007; Lettieri & Fung, 1979; Watson et al.2010).

Expts 3–6: effects of almorexant on the expression of conditioned reward and locomotor sensitization related to morphine, cocaine and amphetamine

Expts 3a, 4 and 5 were designed to test: (1) the effect of acute almorexant treatment on the expression of CPP and locomotor sensitization to morphine (10 mg/kg), cocaine (15 mg/kg) and amphetamine (2 mg/kg); (2) the effect of chronic almorexant treatment on the maintenance of pre-established CPP and locomotor sensitization to morphine, cocaine and amphetamine (see Fig. 1).

Expts 3 a, 4 and 5 were identical in design and, following a CPP pre-test at day 1, differed only in the type of drug of abuse that was used for CPP conditioning and sensitization development (four drug and four saline pairings) during days 2–9: morphine (10 mg/kg) for expt 3 a, cocaine (15 mg/kg) for expt 4 and d.l-amphetamine (2 mg/kg) for expt 5. For each experiment, at day 10 almorexant (100 mg/kg) or vehicle was acutely administered 2 h before the CPP test to assess almorexant's effect on CPP expression to the different drugs of abuse. Thereafter, following a 2-d break, almorexant (100 mg/kg) or vehicle was chronically administered from day 13 onwards once daily for 12 d until day 25 (vehicle and almorexant treatment groups always remained the same throughout the experiment). Following 2 d of drug washout, rats were then exposed to a CPP re-test at day 27 under drug-free conditions to evaluate the effect of chronic almorexant treatment on the maintenance of CPP. Thereafter, another 2-d re-sensitization session was performed (days 28–29), where rats received one additional pairing (the fifth) with drug and one with saline. This re-sensitization test evaluated the effect of chronic almorexant treatment on the maintenance of locomotor sensitization. Finally, another 2-d re-sensitization session was conducted (days 30–31), which involved an acute 2 h almorexant (100 mg/kg) or vehicle pretreatment before the last pairing (the sixth) with the respective drug. The corresponding saline pairing was conducted without pretreatment. This was to assess the effect of acute almorexant treatment on the expression of locomotor sensitization to the different drugs of abuse.

Expt 3b: effects of almorexant on the expression of locomotor sensitization to morphine in a CPP-independent experimental set-up

In order to investigate whether findings from expt 3 a also translated to a CPP-independent locomotor sensitization paradigm with identical injection and testing time-points, one group of rats was primed with four injections of saline (every second day) in the home cage, the other group was primed with four injections of morphine (10 mg/kg). Twenty days after the last injection, both groups were further subdivided into a vehicle and almorexant pretreatment group (Sal-Veh, Sal-Alm, Mor-Veh, Mor-Alm). Then they were all acutely challenged with an injection of morphine (10 mg/kg) and immediately placed into small open fields (46 × 46 × 40 cm; TSE Systems, Germany) for a time-period of 45 min. Horizontal locomotor activity was detected via infrared beam breaks and analysed by Actimot^® software (TSE Systems).

Expt 6: Effects of almorexant on the expression of conditioned reward related to low dose morphine and cocaine

Following a CPP pre-test, rats were conditioned during four context-pairings with low doses of either morphine (3 mg/kg i.p.) or cocaine (5 mg/kg i.p.) according to the CPP protocol described above. Two hours before the CPP test, half of the morphine- and cocaine-conditioned rats were treated with vehicle and the other half with almorexant (100 mg/kg p.o.).

Statistical analysis

Place preference (time spent in the drug-paired vs. the vehicle-paired compartment) or strength of place preference (time spent in the drug-paired compartment at test or re-test vs. pre-test) was analysed using paired Student's t tests. Effects on the interaction of almorexant with the expression of morphine, cocaine or amphetamine induced CPP was assessed by two-way analyses of variances (ANOVAs) for the factors treatment and compartment. Drug effects on general locomotor activity and on the development of locomotor sensitization (increasing locomotor activity with repeated injections) during the CPP conditioning phase were analysed by one-way or two-way ANOVAs for repeated measures (treatment or treatment and time). Maintenance of locomotor sensitization (indicated by maintaining increased locomotor activity after the last compared to the first drug injection) and direct comparisons between the effects of almorexant and vehicle on locomotor sensitization were analysed using two-way ANOVAs, or one- or two-tailed unpaired Student's t tests. Newman–Keuls post-hoc tests or planned comparisons were applied where appropriate. The null hypothesis was rejected at p < 0.05. Sample sizes are indicated in the figure legends.

Fig. 1

Experimental schedule of the extended conditioned place preference (CPP) paradigm (expts 3a, 4 and 5). Following the drug-free CPP pre-test at day 1, rats underwent an 8-d CPP conditioning and sensitization development phase, during which they received four compartment-pairings with saline (Sal) and four with a particular drug of abuse (D): morphine in expt 3a, cocaine in expt 4 and amphetamine in expt 5. In all other aspects the three experiments were identical in design. At day 10, half of the rats received vehicle and the other half received almorexant (100 mg/kg p.o. −2 h) in order to test the effect of acute orexin receptor (OXR)_1/2 blockade on the expression of morphine, cocaine or amphetamine CPP. Thereafter, following a 2 d break (BR; days 11–12), rats underwent a chronic, 12 d treatment period (days 13–24) with vehicle or almorexant (100 mg/kg p.o. once daily). Vehicle and almorexant treatment groups always remained the same throughout the experiment. After a drug washout period of 2 d (BR; days 25–26), rats were exposed to a CPP re-test at day 27 and a re-sensitization session with the fifth drug/Sal pairing on days 28–29. Here we evaluated the effect of chronic OXR_1/2 blockade on the maintenance of CPP expression and locomotor sensitization, to morphine, cocaine and amphetamine. At days 30–31, rats received another acute treatment with almorexant (100 mg/kg p.o.) or vehicle before the sixth drug pairing to assess the effects of acute OXR_1/2 blockade on the expression of locomotor sensitization to morphine, cocaine and amphetamine. The corresponding sixth Sal pairing was conducted without vehicle or almorexant pre-treatment.

Results

Expt 1: effect of almorexant on CPP and locomotor activity

Conditioning with almorexant did not induce preference for the almorexant-paired compartment at the CPP test at either dose tested (100 mg/kg: t₁₀ = 0.95, p = 0.36, Fig. 2a; 300 mg/kg: t₁₃ = 1.47, p = 0.16; Fig. 2c; 100 mg/kg, 2 h pretreatment: t₁₅ = 1.83, p = 0.09; Fig. 2e; 300 mg/kg, 2 h pretreatment: t₁₂ = 0.85, p = 0.41; Fig. 2g). Locomotor activity during the conditioning phase was slightly reduced by the 100 mg/kg almorexant dose (treatment: F_1,10 = 27.00, p < 0.001; Fig. 2b) and slightly increased by the 300 mg/kg dose (treatment: F_1,13 = 4.70, p < 0.05; Fig. 2d). No significant time or time × treatment effects were revealed, indicating that both vehicle- and almorexant-treated rats did not significantly change their locomotor activity over repeated pairings. Administration of almorexant at 100 and 300 mg/kg with a 2 h pretreatment time significantly reduced locomotor activity compared to vehicle at all pairings (100 mg/kg: treatment: F_1,15 = 228.3, p < 0.001; Fig. 2f; 300 mg/kg: F_1,12 = 128.8, p < 0.001; Fig. 2h). In both cohorts of animals, locomotor activity habituated with repeated pairings (100 mg/kg: time: F_3,45 = 15.7, p < 0.001; Fig. 2f; 300 mg/kg: F_3,36 = 12.5, p < 0.001; Fig. 2h), independent of treatment (100 mg/kg: treatment × time: F_3,45 = 2.4, p = 0.08; 300 mg/kg: F_3,36 = 2.6, p = 0.07).

Fig. 2

Effects of almorexant (Alm) on conditioned place preference and locomotor activity. Following the conditioned place preference pre-test at day 1 under drug-free conditions (a, c, e, g), rats underwent an 8-d conditioning phase, where they received four pairings of Alm with one compartment and four pairings of vehicle (Veh) with another compartment. Alm was administered either immediately before placement into the conditioning compartment [100 mg/kg (a, b); 300 mg/kg (c, d); p.o.] or following a pretreatment (p) time of 2 h [100 mg/kg (e, f); 300 mg/kg (g, h); p.o.]. Total distance moved during conditioning was recorded via video tracking (b, d, f, h). At day 10, preference for the Alm-paired vs. the Veh-paired compartment was assessed under drug-free conditions (a, c, e, g). Data are presented as mean±s.e.m.n = 11–16 rats per group. ^&num1;p < 0.05 [significant treatment effect, analysis of variance (ANOVA)]; * p < 0.05, reduced locomotion at fourth vs. first pairing, indicating habituation (post-hoc test following ANOVA).Trans, Transition corridor.

Expt 2: effect of GHB on CPP and locomotor activity

Conditioning with GHB resulted in a significant preference for the GHB-paired compartment at the CPP test at both the 350 mg/kg (t₁₅ = 2.22, p < 0.05, Fig. 3a) and 700 mg/kg (t₁₅ = 2.69, p < 0.05; Fig. 3c) doses. Analyses of the effects of the 350 mg/kg GHB dose on locomotor activity during conditioning revealed significant effects for treatment (F_1,15 = 4.76, p < 0.05) and time (F_3,45 = 11.9, p < 0.001), but not for the treatment × time interaction (F_3,45 = 2.84, p = 0.06; Fig. 3b). These results indicated that GHB slightly increased locomotion in comparison to vehicle across all pairings and that the locomotor activity of this cohort of rats habituated during the conditioning phase, independent of treatment. Analyses of the effects of the 700 mg/kg GHB dose also revealed significant effects for treatment (F_1,15 = 11.1, p < 0.001), time (F_3,45 = 3.14, p < 0.05), and, in addition, a significant treatment × time interaction (F_3,45 = 11.0, p < 0.001; Fig. 3d). Post-hoc comparisons showed that GHB significantly reduced locomotor activity only during the first two pairings compared to the vehicle control (p < 0.05). Finally, locomotion habituated upon repeated conditioning trials with vehicle treatment, but not with repeated GHB treatment.

Fig. 3

Effects of γ-hydroxybutyrate (GHB) on conditioned place preference (CPP) and locomotor activity. Following the CPP pre-test at day 1 under drug-free conditions (a, c), rats underwent an 8-d conditioning phase, where they received four pairings of GHB with one compartment and four pairings of vehicle (Veh) with another compartment. GHB was administered 30 min before placement in the conditioning compartments [350 mg/kg (a, b); 750 mg/kg (c, d); p.o.] Total distance moved during conditioning was recorded via video-tracking (b, d). At day 10, preference for the GHB-paired vs. the Veh-paired compartment was assessed under drug-free conditions (a, c). Data are presented as mean±s.e.m.n = 16 rats per group. * p < 0.05, time spent in the GHB-paired vs. Veh-paired compartment (t test) or reduced locomotion at fourth vs. first pairing, indicating habituation [post-hoc test following analysis of variance (ANOVA)]; ⁺p < 0.05 (significant treatment effect, ANOVA); ^&num1;p < 0.05 vs. respective vehicle group (post-hoc test following ANOVA). Trans, Transition corridor.

Expts 3a, b: effect of almorexant on the expression of conditioned reward and locomotor sensitization related to high dose of morphine (10 mg/kg)

In expt 3 a, conditioning with morphine resulted in significant preference for the morphine-paired compartment at the CPP test (t₁₅ = 5.43, p < 0.001; Fig. 4a), which was still present at the CPP re-test 2 wk later (t₁₅ = 2.85, p < 0.05; Fig. 4a). An additional comparison of the amount of time spent in the morphine-paired compartment before conditioning (pre-test) and after conditioning at the CPP test (t₁₅ = 3.46, p < 0.01) and re-test (t₁₅ = 1.06, p = 0.31) indicated that the strength of conditioning decreased with time and/or repeated testing. Acute administration of almorexant did not affect the expression of CPP for the morphine-paired compartment at the CPP test (t₁₅ = 3.71, p < 0.01; Fig. 4b). This was also indicated by a non-significant treatment × compartment interaction in the direct comparison between the almorexant- and the vehicle-treated groups (F_2,60 = 2.4, p = 0.1). Furthermore, because of its vigilance-decreasing effects, acute almorexant reduced the distance moved during the CPP test (3922 ± 200 cm vs. 5819 ± 226 cm in the vehicle group; t₃₀ = 6.3, p < 0.001), which might explain the larger amount of time spent in the small, neutral transition corridor.

Fig. 4

Effects of almorexant (Alm) on the expression of conditioned place preference (CPP) and locomotor sensitization related to morphine (Mor). Following the drug-free CPP pre-test at day 1 and conditioning with Mor (10 mg/kg i.p.)/saline (Sal) during the next 8 d, place preference for the Mor-paired compartment was assessed at day 10, 2 h after acute treatment with vehicle (Veh) or Alm (100 mg/kg p.o.; test: a, b). Two days later rats were continued to be treated with Veh or Alm (100 mg/kg, p.o., once per day) for 12 d and re-tested again for maintenance of CPP expression following a drug washout period of 2 d (re-test: a, b). Following the establishment of locomotor sensitization by the first four Mor injections during the conditioning phase (pairings no. 1–4) (c, d), rats were re-tested for the maintenance of sensitization by a fifth Mor injection (pairing no. 5), following washout after chronic treatment with Veh (c) or Alm (d). Thereafter, rats were again re-tested for the expression of locomotor sensitization by a sixth injection of Mor (pairing no. 6), but this time 2 h following acute treatment with Veh (c) or Alm (100 mg/kg p.o.; d). The locomotor activity during the corresponding six Sal injection pairings is shown for comparison. The effect of acute almorexant on Mor-induced sensitization was also assessed in a CPP-independent set-up (e). Rats were primed with four injections of Sal or Mor (10 mg/kg i.p.) and then on the test day all rats were challenged with an acute injection of Mor (10 mg/kg) following pretreatment with either Veh or Alm (100 mg/kg p.o.) 2 h before. Locomotion was assessed in small open fields with slightly different dimensions than the CPP conditioning compartments. Data are presented as mean±s.e.m.n = 16 rats per group for (a–d), n = 10 for (e). * p < 0.05, ^&num1;p < 0.05 vs. the respective first Mor injection (post-hoc test following analysis of variance or t test). Trans, Transition corridor.

After washout following chronic treatment withalmorexant for 12 d, CPP for the morphine-paired compartment was still present at the CPP re-test (t₁₅ = 2.72, p < 0.05; Fig. 4b) and was not different from that of the chronically vehicle-treated group at re-test (treatment × compartment: F_2,60 = 2.4, p = 0.1; Fig. 4a).

Four injections of morphine during CPP conditioning produced robust locomotor sensitization (F_3,15 = 49.3, p < 0.001; Fig. 4c; F_3,15 = 49.0, p < 0.001; Fig. 4d). Sensitization was maintained at the fifth morphine injection after washout following chronic vehicle (t₁₅ = 10.15, p < 0.001 vs. first injection; Fig. 4c) or almorexant treatment (t₁₅ = 8.82, p < 0.001; Fig. 4d) and did not differ between both treatment groups (t₃₀ = 0.46, p = 0.65). Sensitization at the sixth morphine injection remained unaffected by acute vehicle treatment (t₁₅ = 5.53, p < 0.001 vs. first injection; Fig. 4c) but was completely blocked by acute almorexant treatment (t₁₅ = 0.82, p = 0.22 vs. first injection; Fig. 4d). This was also indicated by a direct comparison between both treatment groups (t₃₀ = 5.0, p < 0.001).

The above effects of acute almorexant treatment on locomotor sensitization were replicated in a CPP-independent set-up in expt 3b where ANOVA revealed significant effects of priming (morphine or saline; F_1,28 = 22.4, p < 0.001; Fig. 4e) of treatment (vehicle or almorexant; F_1,28 = 39.3, p < 0.001) and a significant interaction (F_1,28 = 11.2, p < 0.001). As expected, post-hoc analyses revealed that, upon acute morphine challenge, vehicle-treated rats that had previously been primed with four morphine injections showed significantly increased (sensitized) locomotion as compared to vehicle-treated rats having been primed with only saline (p < 0.001). The effect of previous morphine priming on locomotion, however, was completely abolished in almorexant-treated rats (p = 0.76), which moved significantly less than morphine-primed, vehicle-treated rats (p < 0.001).

Taken together, acute almorexant treatment did not interfere with the expression of CPP but with the expression of locomotor sensitization to a high dose of morphine. Chronic almorexant treatment, followed by a 2-d washout, had no effect on the maintenance of CPP or locomotor sensitization to morphine.

Expt 4: effects of almorexant on the expression of conditioned reward and locomotor sensitization related to high dose of cocaine (15 mg/kg)

Conditioning with cocaine resulted in a significant preference for the cocaine-paired compartment at the CPP test (t₁₄ = 3.01, p < 0.01; Fig. 5a) and at the CPP re-test 2 wk later (t₁₄ = 2.36, p < 0.05; Fig. 5a). However, the strength of conditioning declined over time, indicated by a significant difference between the time spent in the cocaine-paired compartment between test and pre-test (t₁₄ = 4.52, p < 0.001) but not between re-test and pre-test (t₁₄ = 1.29, p = 0.22). Acute administration of almorexant attenuated the expression of cocaine CPP (no significant difference between the cocaine-paired and saline-paired compartment at the CPP test: t₁₅ = 1.65, p = 0.12; Fig. 5b). This was further verified by a significant treatment × compartment interaction in the direct comparison between the almorexant- and the vehicle-treated groups (F_2,58 = 5.0, p < 0.01). Acute almorexant also reduced the distance moved during the CPP test (4060 ± 166 cm vs. 5740 ± 209 cm in the vehicle-treated group; t₂₉ = 6.3, p < 0.001).

Fig. 5

Effects of almorexant (Alm) on the expression of conditioned place preference (CPP) and locomotor sensitization related to cocaine (Coc). Following the drug-free CPP pre-test at day 1 and conditioning with Coc (15 mg/kg i.p.)/saline (Sal) during the next 8 d, place preference for the Coc-paired compartment was assessed at day 10, 2 h after acute treatment with vehicle (Veh) or Alm (100 mg/kg, p.o.; test: a, b). Two days later rats were continued to be treated with Veh or Alm(100 mg/kg, p.o., once per day) for 12 d and re-tested again for the maintenance of CPP following a drug washout period of 2 d (retest: a, b). Following the establishment of locomotor sensitization by the first four Coc injections during the conditioning phase (pairings no. 1–4; c, d), rats were re-tested for the maintenance of sensitization by a fifth Coc injection at pairing no. 5, following washout after chronic treatment with Veh (c) or Alm (d). Thereafter, rats were again re-tested for the expression of locomotor sensitization by a sixth injection of Coc at pairing no. 6, but this time 2 h following acute treatment with Veh (c) or Alm (100 mg/kg p.o.; d). The locomotor activity during the corresponding six Sal injection pairings is shown for comparison. Data are presented as mean±s.e.m.n = 15–16 rats per group. * p < 0.05 (t test), ^&num1;p < 0.05 vs. the respective first Coc injection (post-hoc test following analysis of variance or t test). Trans, Transition corridor.

After washout following chronic treatment with almorexant for 12 d, significant CPP for the cocaine-paired compartment was still present at the re-test (t₁₅ = 4.45, p < 0.001; Fig. 5b) and was not different from the cocaine CPP expressed by the chronically vehicle-treated group at re-test (treatment × compartment: F_2,58 = 0.37, p = 0.69; Fig. 5a).

Four injections of cocaine produced robust locomotor sensitization (F_3,14 = 4.06, p < 0.05; Fig. 5c; F_3,15 = 8.62, p < 0.001; Fig. 5d). Sensitization was maintained at the fifth cocaine injection given after 2 d washout following chronic vehicle (t₁₄ = 2.14, p < 0.05 vs. first injection; Fig. 5c) or almorexant (t₁₅ = 2.08, p < 0.05; Fig. 5d) treatment and was quantitatively not different between both treatment groups (t₂₉ = 1.04, p = 0.31). Likewise, sensitization at the sixth cocaine injection was still maintained after acute vehicle (t₁₄ = 2.85, p < 0.01 vs. first injection; Fig. 5c) and almorexant treatment (t₁₅ = 2.82, p < 0.01; Fig. 5d) and did qualitatively not differ between these two treatment groups (t₂₉ = 0.72, p = 0.48).

Taken together, acute almorexant treatment attenuated the expression of CPP but not the expression of locomotor sensitization to a high dose of cocaine. Chronic almorexant treatment, followed by a 2 d washout, had no effect on the maintenance of CPP or locomotor sensitization to cocaine.

Expt 5: effects of almorexant on the expression of conditioned reward and locomotor sensitization related to high dose of amphetamine (2 mg/kg)

Conditioning with amphetamine resulted in a significant preference for the amphetamine-paired compartment at the CPP test (t₁₅ = 6.75, p < 0.001; Fig. 6a) and re-test 2 wk later (t₁₅ = 4.02, p < 0.01; Fig. 6a). However, the strength of conditioning declined over time, indicated by a significant difference between the time spent in the amphetamine-paired compartment between test and pre-test (t₁₅ = 3.52, p < 0.01) but not between re-test and pre-test (t₁₅ = 1.69, p = 0.11). Acute administration of almorexant prevented the expression of amphetamine CPP (no significant difference between the amphetamine-paired and saline-paired compartment: t₁₅ = 0.56, p = 0.59; Fig. 6b). This was further verified by the direct comparison between the almorexant- and the vehicle-treated groups (treatment × compartment: F_2,60 = 4.39, p < 0.05; Fig. 6a, b). Acute almorexant also reduced the distance moved during the CPP test (4017 ± 159 cm vs. 5338 ± 170 cm in the vehicle-treated group; t₃₀ = 5.67, p < 0.001).

Fig. 6

Effects of almorexant (Alm) on the expression of conditioned place preference (CPP) and locomotor sensitization related to d.l-amphetamine (Amph). Following the drug-free CPP pre-test at day 1 and conditioning with Amph (2 mg/kg i.p.)/saline (Sal) during the next 8 d, place preference for the Amph-paired compartment was assessed at day 10, 2 h after acute treatment with vehicle (Veh) or Alm (100 mg/kg, p.o.; test: a, b). Two days later rats were continued to be treated with Veh or Alm (100 mg/kg, p.o., once per day) for 12 d and re-tested again for maintenance of CPP following a drug washout period of 2 d (retest: a, b). Following the establishment of locomotor sensitization by the first four Amph injections during the conditioning phase (pairings no. 1–4; c, d), rats were tested for the maintenance of sensitization by a fifth Amph injection at pairing no. 5, following washout after chronic treatment with Veh (c) or Alm (d). Thereafter, rats were again re-tested for the expression of locomotor sensitization by a sixth injection of Amph (pairing no. 6), but this time 2 h following acute treatment with Veh (c) or Alm (100 mg/kg p.o.; d). The locomotor activity during the corresponding six Sal injection pairings is shown for comparison. Data are presented as mean±s.e.m.n = 16 rats per group. * p < 0.05 (t test), ^&num1;p < 0.05 vs. the respective first Amph injection (post-hoc test following analysis of variance or t test). Trans, Transition corridor.

After washout following chronic treatment with almorexant for 12 d, significant CPP for the amphetamine-paired compartment was still present at the re-test (t₁₅ = 2.28, p < 0.05; Fig. 6b) and was not different from the amphetamine CPP expressed by the chronically vehicle-treated group at re-test (treatment × compartment: F_2,60 = 0.21, p = 0.81; Fig. 6a). Four injections of amphetamine produced robust locomotor sensitization (F_3,15 = 15.2, p < 0.001; Fig. 6c; F_3,15 = 9.30, p < 0.001; Fig. 6d). Sensitized locomotion was still maintained at the fifth amphetamine injection after 2-d washout following chronic vehicle (t₁₅ = 3.19, p < 0.01 vs. first injection; Fig. 6c) or almorexant (t₁₅ = 4.51, p < 0.001; Fig. 6d) treatment and was quantitatively not different between both treatment groups (t₃₀ = 0.21, p = 0.83). Sensitization was also still present at the sixth amphetamine injection following acute vehicle (t₁₅ = 3.67, p < 0.01 vs. first injection; Fig. 6c) or almorexant (t₁₅ = 3.97, p < 0.01 vs. first injection; Fig. 6d) treatment and was quantitatively not different between both treatment groups (t₃₀ = 0.26, p = 0.80).

Taken together, acute almorexant treatment prevented the expression of CPP but not the expression of locomotor sensitization to a high dose of amphetamine. Chronic almorexant treatment, followed by a 2-d washout, had no effect on the maintenance of CPP or locomotor sensitization to amphetamine.

Expt 6: effects of almorexant on the expression of conditioned reward related to low doses of morphine (3 mg/kg) and cocaine (5 mg/kg)

Conditioning with low dose morphine or cocaine still resulted in a significant preference for the drug-paired compartment at the CPP test (morphine: t₁₅ = 5.32, p < 0.001; Fig. 7a; cocaine: t₁₅ = 2.93, p < 0.05; Fig. 7c). Acute administration of almorexant did not affect the expression of morphine CPP (t₁₅ = 3.70, p < 0.01; Fig. 7b) but slightly attenuated that of cocaine CPP (t₁₅ = 1.25, p = 0.23; Fig. 7d). However, direct comparison between the almorexant- and the vehicle-treated groups revealed non-significant treatment × compartment interactions for both morphine (F_2,60 = 2.41, p = 0.1) and cocaine (F_2,60 = 0.1, p = 0.9). This ANOVA suggested that almorexant, in fact, did not significantly attenuate the low dose cocaine CPP expression, contrary to what was indicated by the independent t test analyses. Furthermore, almorexant reduced locomotion during CPP testing in both the morphine group (3371 ± 183 cm vs. vehicle: 4721 ± 145 cm; t₃₀ = 5.8, p < 0.001) and the cocaine group (3196 ± 122 cm vs. vehicle: 4845 ± 110 cm; t₃₀ = 10.1, p < 0.001).

Fig. 7

Effects of almorexant (Alm) on conditioned place preference (CPP) related to low doses of morphine (Mor) and cocaine (Coc). Following the drug-free CPP pre-test at day 1 and conditioning with Mor (3 mg/kg i.p.)/saline (Sal; a, b) or Coc (5 mg/kg i.p.)/Sal (c, d) during the next 8 d, place preference for the Mor- or Coc-paired compartment was assessed at day 10, 2 h after acute treatment with vehicle (Veh; a, c) or Alm (100 mg/kg, p.o.; b, d). Data are presented as mean±s.e.m.n = 16 rats per group. * p < 0.05 (t test). Trans, Transition corridor.

Discussion

Comparison of almorexant with GHB on CPP and locomotion

Almorexant has been developed as the first sleep-enabling drug based on dual OXR blockade (Brisbare-Roch et al.2007). In comparison to the currently used positive GABA receptor modulators, almorexant treatment maintains sleep architecture and the physiological relations between REM and non-REM sleep without inducing excessive sedation or myorelaxation (Brisbare-Roch et al.2007; Steiner et al.2011) and with possibly low abuse liability (Griffiths & Johnson, 2005; Scammell & Winrow, 2011). Indeed, current knowledge of the OX system in relation to reward pathways suggests an anti-addictive potential of compounds inhibiting OXRs (Aston-Jones et al.2009). This hypothesis is now further supported by our observations that almorexant did not induce CPP at any dose or condition tested, as opposed to GHB, which was shown here and by others (Fattore et al.2000; Itzhak & Ali, 2002; Watson et al.2010) to induce CPP in rodents, in line with reported abuse liability in patients (Carter et al.2009; Griffiths & Johnson, 2005).

Both almorexant and GHB at high doses decrease vigilance and cause hypolocomotion (Brisbare-Roch et al.2007; Itzhak & Ali, 2002). Interestingly, GHB is known to induce tolerance to its hypolocomotor effects upon repeated administration (Itzhak & Ali, 2002), whereas almorexant does not (Brisbare-Roch et al.2007). Accordingly, in the present study, the hypolocomotor effects of high oral doses of almorexant (100 and 300 mg/kg) given 2 h before CPP conditioning (Fig. 2f, h) was maintained over four repeated administrations, whereas rats developed tolerance to the hypolocomotor effects of a high oral dose (700 mg/kg) of GHB from the third administration onwards (Fig. 3d). Tolerance as well as sensitization are phenomena positively correlated with abused drugs (O'Brien & Gardner, 2005). Thus, the present findings may serve as an additional indicator for potentially lower abuse liability of almorexant compared to GHB.

Acute almorexant in combination with morphine, cocaine and amphetamine

This study also investigated whether almorexant might exert anti-addictive-like properties similar to OXR₁ antagonists or genetic ablation of OX (Gozzi et al.2011; Harris et al.2005; Hutcheson et al.2011; Narita et al.2006; Sharf et al.2010). We found that acute almorexant treatment attenuated high dose cocaine- (10 mg/kg) and amphetamine- (2 mg/kg) induced, but not morphine- (3 and 10 mg/kg) induced, CPP. This is surprising in light of results from Harris et al. (2005), who demonstrated that OX neurons in the lateral hypothalamus become activated independent of whether rats were exposed to cocaine- or morphine-associated cues. Thus, it must be perceived that, depending on the drug of abuse, different neurocircuits besides OXs become co-activated or that different downstream pathways are recruited, perhaps via engagement of particular OX neuron subpopulations or OXR output regions. In line with this hypothesis, Sharf and co-workers found a functional dissociation between the expression of cocaine and morphine CPP, of which only the latter could be attenuated by the OXR₁ antagonist SB-334867 (Sharf et al.2010). Part of the results were, however, later countered by those of Gozzi et al. (2011), demonstrating that CPP expression for cocaine could, in fact, be inhibited by another selective OXR₁ antagonist, GSK-1059865. Furthermore, Hutcheson et al. (2011) also demonstrated a blockade of CPP expression for amphetamine by the OXR₁ antagonist SB-334867. Other, CPP-independent studies confirmed a crucial role for OXR₁ signalling in motivated behaviours towards cocaine-associated cues/contexts (Borgland et al.2009; Smith et al.2009, 2010). Thus, we may speculate that attenuated CPP expression to high dose cocaine and amphetamine by almorexant may be mediated mainly via a reduction of conditioned motivational responses towards drug-associated contextual cues caused by OXR₁ blockade.

It remains an interesting observation that almorexant only showed a trend but did not significantly attenuate the CPP expression related to a low dose of cocaine (5 mg/kg). Cocaine is often viewed as a ‘dirty drug', affecting not only dopamine, but also serotonin and noradrenaline uptake, and, in addition, sodium channels function at high dose (Hill et al.2009). Thus, one may speculate that a slightly different activation of neurocircuits during conditioning with a low dose as opposed to high dose cocaine may have led to a different extent of OX recruitment upon context re-exposure at the CPP test. So far, we are not aware of any other study in the OX field having tested low doses (⩽5 mg/kg) of cocaine and 15 mg/kg remains the most frequently used dose (Gozzi et al.2011; Sharf et al.2010). Ultimately, the cocaine–OX dose–response relationship remains a field for further exploration.

As a sleep-enabling drug, almorexant does not induce irreversible sedation or myorelaxation (Steiner et al.2011). These findings argue against the objection that some of the observed CPP effects could also have been secondary to the sleep-enabling effects of almorexant. In fact, whereas almorexant-treated rats moved about 30% less than vehicle-treated rats during the CPP test, they were not asleep and were well able to differentiate between the drug-paired and non-paired compartment, as demonstrated during the CPP test to 3 and 10 mg/kg morphine. Moreover, the lack of effect of almorexant on the expression of morphine CPP supports the conclusion that conditioned-reward-related memory recall remains intact under almorexant, which has already been ascertained for spatial (Dietrich & Jenck, 2010) and procedural memory (Steiner et al.2011).

Similar to the CPP data, we also observed drug of abuse-dependent differential effects of acute almorexant treatment on the expression of locomotor sensitization. Almorexant blocked locomotor sensitization to morphine but not to the psychostimulants cocaine and amphetamine. These data are in line with studies from Borgland et al. (2006) and Smith et al. (2009), who demonstrated that OXR₁ signalling is not implicated in the expression of sensitization to cocaine or in cocaine reinforcement, but are contrary to a study from Quarta et al. (2010), who showed that OXR₁ antagonism can reduce amphetamine sensitization (Quarta et al.2010). Perhaps compensatory mechanisms at OXR₂ (Dugovic et al.2009) may explain the partially different outcome with regard to almorexant and amphetamine. In the case of morphine, the involvement of orexins in the expression of sensitization had not yet been investigated and our current findings now suggest that OXR signalling is critically involved.

In view of the different molecular mechanisms of action for cocaine, dopamine and morphine (Vanderschuren & Kalivas, 2000), it is conceivable that even though all three drugs lead to similar reward-related effects, i.e. CPP and locomotor sensitization, they do so by recruiting different neuronal circuits. Indeed, some molecular targets have already been identified that are differentially engaged upon exposure to drug-related cues/contexts or during sensitization with psychostimulants as opposed to opiates (e.g. Barik et al.2010; Veeneman et al.2011). We may now extend these findings to the OX system. Indeed, OX neurons can be directly recruited by morphine via co-expressed µ-opioid receptors (Georgescu et al.2003). In addition, OXs may influence the action of morphine in the VTA, where they control synaptic changes and glutamate signalling (Borgland et al.2006, 2008) and this may possibly differentiate their role in the expression of morphine conditioned reward and locomotor sensitization from their role in psychostimulant conditioned reward and locomotor sensitization.

The differential involvement of OXs in CPP and locomotor sensitization also illustrates that both behaviours are not necessarily correlated (in fact, none of our experiments revealed significant correlations between the magnitude of locomotor sensitization effects and CPP effects; data not shown) and may reflect different aspects of addiction. Whereas CPP relates to drug-induced context associations resulting from positive reinforcement (Bardo & Bevins, 2000), locomotor sensitization in animals may translate to ‘incentive sensitization’ in humans, reflecting the exaggerated motivation or craving experienced upon repeated drug exposure (Robinson & Berridge, 1993). A deeper understanding of the exact processes in which OXR₁, OXR₂ or dual OXR_1/2 signalling is involved will be pivotal for designing specific treatments for addiction based on OXR antagonism. Two recent papers, for instance, have nicely separated the differential roles of OXR₁ from OXR₂ signalling, demonstrating a critical role for OXR₁ in the reinstatement of cue-induced cocaine-seeking (Smith et al.2009), but a critical role for OXR₂ in ethanol self-administration and CPP (Shoblock et al.2011). Whether the dual blockade of both receptors combines the effects achieved by single OXR₁ and OXR₂ antagonism or whether compensatory effects may occur if both pathways are targeted simultaneously (Dugovic et al.2009) awaits further investigation.

Chronic almorexant and the maintenance of morphine, cocaine and amphetamine induced CPP and sensitization

Chronic cocaine, morphine and nicotine exposure is known to change OX and/or OXR mRNA expression in the hypothalamus for up to 6 wk after termination of drug exposure (Kane et al.2000; Zhou et al.2006, 2008). These findings suggest that long-term modifications in OX signalling following chronic exposure to drugs of abuse may contribute to the long-term persistence of addiction-like behaviours. We asked the question whether interference with these long-term OX signalling modifications through chronic OXR_1/2 blockade by almorexant treatment may have a functional implication for the maintenance of addictive-like behaviours, such as CPP or locomotor sensitization. However, the lack of effect of chronic almorexant treatment on CPP and locomotor sensitization, when re-tested after almorexant washout, suggests that such molecular OX/OXR signalling modifications were either not permanently affected by transient chronic blockade or not involved in the expression of those behaviours or not present under the particular drug of abuse-priming conditions used in this study.

Summary

Taken together, we show that almorexant itself does not induce CPP in the rat as opposed to another sleep-promoting drug, GHB and may, thus, present with lower abuse liability. Furthermore, we demonstrate that acute almorexant reduces the expression of conditioned rewarding effects related to psychostimulants as well as the expression of locomotor sensitizing effects related to morphine.

Acknowledgements

We thank Hendrik Dietrich (Actelion Pharmaceuticals Ltd, Switzerland) for conceptual assistance with the set-up of the CPP paradigm. We thank Carla Sciarretta (Actelion Pharmaceuticals Ltd, Switzerland) for careful review of the manuscript and Alex Dorr (Actelion Pharmaceuticals Ltd, Switzerland) for editorial assistance. The authors accept full responsibility for the scientific content of this article.

Statement of Interest

All authors are employees of Actelion Pharmaceuticals Ltd, Switzerland.