Cocaine and heroin (‘speedball’) self-administration: the involvement of nucleus accumbens dopamine and μ-opiate, but not δ-opiate receptors

Psychopharmacology (2005) 180: 21–32DOI 10.1007/s00213-004-2135-9

ORIGINAL INVESTIGATION

Jennifer L. Cornish . Jaclyn M. Lontos .Kelly J. Clemens . Iain S. McGregor

Cocaine and heroin (‘speedball’) self-administration:the involvement of nucleus accumbens dopamine and μ-opiate,but not δ-opiate receptors

Received: 7 April 2004 / Accepted: 22 November 2004 / Published online: 29 January 2005# Springer-Verlag 2005

Abstract Rationale: The combined administration ofheroin and cocaine (‘speedball’) is common among intra-venous drug users. Dopamine receptors in the nucleus ac-cumbens play a key role in cocaine self-administration;however, their role in speedball self-administration is un-known, as is the role of opiate receptors in this region.Objectives: The effect of blocking dopamine D1, D2, μ-opiate or δ-opiate receptors in the nucleus accumbens onthe intravenous self-administration of combined heroin andcocaine was examined in rats. Methods: Rats with bilat-eral cannulae implanted into the nucleus accumbens weretrained to self-administer intravenous speedball (ratio ofcocaine/heroin, 17:1) under a progressive ratio (PR) sched-ule. Prior to their self-administration session, rats were thenmicroinjected with the dopamine D1 receptor antagonistSCH 23390 (1 and 6 nmol side−1), the D2 receptor antag-onist raclopride (3 and 10 nmol side−1), the μ-opiate re-ceptor antagonist CTOP (0.1, 0.3 and 1.0 nmol side−1), theδ-opiate receptor antagonist naltrindole (1.0, 3.0 and 10nmol side−1) or a cocktail of SCH 23390 (1 nmol side−1)and CTOP (0.1 nmol side−1) into the nucleus accum-bens. Results: Microinjection of SCH 23390, raclopride orCTOP into the nucleus accumbens produced dose-de-pendent decreases in breakpoints under the PR schedule,while naltrindole was without effect. The highest dose ofSCH 23390 also significantly reduced locomotor activ-ity measured during speedball self-administration. Thecombination of SCH 23390 and CTOP significantly re-duced breakpoints, while not affecting locomotor activity.Conclusions: These results indicate that dopamine and μ-opiate receptors, but not δ-opiate receptors, in the nucleusaccumbens are involved in the reinforcing effects of speed-ball. Combined administration of D1 and μ-opiate receptorantagonists may be more selective at reducing the rein-

forcing effects of speedball self-administration than eitherdrug alone.

Keywords Speedball . Cocaine . Heroin . Self-administration . Dopamine . Opiate . Reward .Reinforcement

Introduction

It is increasingly recognized that drug users typically par-take of many different drugs and that multiple drugs areoften consumed within a single session (National Instituteon Drug Abuse, 1998). One relatively common drug com-bination involves the simultaneous intravenous administra-tion of cocaine and heroin, commonly referred to as a‘speedball’ (Schutz et al. 1994). Health and socio-economicproblems associated with the use of this drug combinationare increasingly recognized and current pharmacotherapiesappear largely ineffective in controlling speedball use (seeLeri et al. 2003 for review).

Previous studies suggest that taking low dose cocaineand heroin in combination may synergistically enhance thereinforcing effects of either drug taken alone (Foltin andFischman 1992; Hemby et al. 1999). Rats and primatesappear more motivated to receive combined heroin/psy-chostimulant infusions relative to cocaine or heroin alone(Ranaldi andWise 2000; Rowlett et al. 1998). The apparentunique effects produced by combined cocaine and heroinmay render current pharmacotherapies, such as methadone-maintenance treatment, relatively ineffective as they targetonly one drug type (Mendelson and Mello 1996). As such,further investigation may help to establish the neurobio-logical substrates mediating the reinforcing properties ofspeedball.

The mesocorticolimbic dopamine system, projectingfrom the ventral tegmental area (VTA) to the nucleus ac-cumbens and prefrontal cortex, plays a key role in thereinforcing effects of drugs (Wise 2002). Psychostimu-lants, such as cocaine and amphetamine, affect dopamineuptake transporters, resulting in increases in extracellular

J. L. Cornish (*) . J. M. Lontos . K. J. Clemens .I. S. McGregorSchool of Psychology A19, University of Sydney,2006 Sydney, Australiae-mail: [email protected].: +61-2-93513544Fax: +61-2-93518023

concentrations of dopamine in terminal regions (Ritz et al.1987; Sieden et al. 1993). Opiates, such as heroin, pri-marily act at μ-opiate receptors on γ-amino-butyric acid(GABA) cells of the VTA, to disinhibit dopamine cellactivity (Johnson and North 1992). This disinhibition re-sults in increases in dopamine release to act at dopamineD1 and D2 receptor families of the nucleus accumbens(Wise et al. 1995). It is generally accepted that this increasein nucleus accumbens dopamine transmission is involvedin the motivation for drug usage and recent studies suggestthat opioid systems, particularly at the μ-opiate receptor,are selectively involved in reward preference (for review,see Cardinal and Everitt, 2004).

The reinforcing effects of cocaine are reduced by dopa-mine receptor antagonists administered systemically ordirectly into the nucleus accumbens (Ikemoto et al. 1997;McGregor and Roberts 1993; Phillips et al. 1994). How-ever, while opiate administration increases dopamine re-lease in the nucleus accumbens (Gerasimov and Dewey1999; Maisonneuve et al. 2001; Wise et al. 1995), a similarrole for accumbens dopamine receptors in heroin self-administration has not been conclusively shown. Heroin/cocaine combinations produce synergistic elevations innucleus accumbens dopamine concentrations (∼1,000%above baseline) compared to cocaine (∼380–400% abovebaseline) and heroin (∼0–70% above baseline) self-admin-istration alone (Gerasimov and Dewey 1999; Hemby et al.1999).

Opiate receptors of the μ subtype are crucial for therewarding effects of opioids, serving as high-affinity bind-ing sites for heroin and morphine (Yu 1996). In additionto the VTA, μ receptors are abundant in the nucleus ac-cumbens and reside on GABAergic neurons (Svingos et al.1997, 1998). Antagonism of the μ-opiate receptor sys-temically or locally in the nucleus accumbens significantlyreduces heroin (Martin et al. 2002; Killian et al. 1978;Negus et al. 1993) or cocaine reward (Corrigall and Coen1991b; Mello et al. 1990; Ward et al. 2003). Furthermore, μreceptor agonists are self-administered directly into thenucleus accumbens (Goeders et al. 1984).

The δ-opiate receptors are also located in limbic areas(Tempel and Zukin 1987) and reside upon acetylcholiner-gic interneurons in striatal regions (Le Moine et al. 1994).Administration of δ-opioid agonists increases dopaminelevels in the nucleus accumbens (Spanagel et al. 1990),suggesting a role for these receptor types in reward pro-cesses. Systemic administration of the selective δ-opioidreceptor antagonist, naltrindole, reduces heroin reward(Negus et al. 1993) yet shows inconsistent effects on co-caine reinforcement (De Vries et al. 1995; Negus et al.1993, 1995; Reid et al. 1995).

These studies indicate that μ- and δ-opiate receptors mayplay a role in cocaine and heroin self-administration.However, the effect of intra-accumbens dopamine and opi-ate receptor antagonists on speedball self-administrationremains to be assessed. The present study thereforeexamined the role of dopamine D1, D2, μ- and δ-opiatereceptors of the nucleus accumbens in the self-administra-tion of speedball. A progressive ratio (PR) schedule of

reinforcement was employed which allows the measure-ment of the amount of work (lever presses) an animal willundertake for drug infusions (Arnold and Roberts 1997).This allowed us to identify changes in the motivation toreceive speedball reward following nucleus accumbenstreatments.

Materials and methods

Animals

A total of 28 male Hooded–Wistar rats (90–150 days, 250–400 g at the start of the experiment) served as subjects. Therats were bred within the School of Psychology at theUniversity of Sydney. The rats were initially housed ingroups of 7–8 in large polypropylene tubs lined withwoodchips. After surgery, the rats were individuallyhoused in smaller tubs to prevent damage to the intracranialcannulae or intravenous catheters. Food and water wasavailable ad libitum. Both pre- and post-surgery rats weremaintained on a 12-h conventional light–dark cycle (lightsoff at 1900 h).

All experimental procedures were conducted in accord-ance with the Australian Code of Practice for the Care andUse of Animals for Scientific Purposes, 6th edition (Na-tional Health and Medical Research Council, 1997). Allefforts were made to minimize the number of animals usedand their suffering. Ethical approval for all experimentswas obtained from the Sydney University Animal EthicsCommittee.

Apparatus

Operant chambers

Eight standard operant chambers [250 (depth)×310(width)×500 mm (height)] were used. The chambers hadaluminium sides and tops while the front and back wallswere made of Plexiglas. The floor was constructed of 16metal rods (diameter 6 mm) spaced 15 mm apart. Theoperant chambers were housed in sound-attenuating boxes[600 (depth)×580 (width)×670 mm (height)] equippedwith a fan to provide ventilation and masking noise.

To detect locomotor activity, each chamber had twopassive infrared (PIR) detectors (Quantum passive infraredmotion sensor, part no. 890-087-2, NESS Security Prod-ucts, Australia) positioned in the centre of each side wall∼30 mm above the floor. The PIR detectors were capable ofdetecting relatively small movements of the rats’ head andbody. The counts from the detectors were recorded by aMacintosh computer running WorkbenchMac software fordata acquisition (McGregor 1996).

Each chamber was equipped with two 50-mm-wideretractable levers (Med Associates, part ENV112-BM) onthe right-hand side wall. The levers were spaced 110 mmapart and situated 60 mm above the floor. Depression ofone of the levers (the active lever) resulted in a 0.05 ml

22

intravenous infusion of speedball over 2.5 s followed bythe illumination of a cue light (situated 50 mm above thelever) indicating a 20 s time-out period. During the time-out period, depression of either lever had no scheduledconsequences. Depression of the other lever (the inactivelever) had no consequences at any time. Responses on eachlever and delivery of drug infusions were recorded oncomputer.

The drug infusion system, located outside of the operantchamber, consisted of an infusion pump (Med-PC, VT,USA), containing a 10 ml syringe and 23-gauge cut-offneedle connected to Tygon tubing (Daigger, IL, USA). Thetubing was connected to a fluid swivel assembly (Instech,PA, USA) and PE50 tubing (Plastics One, VA, USA)threaded through a spring connector (CG313, Plastics One,VA, USA). Approximately 20 mm from the base of thespring connector, the spring was separated and the PE50tubing exited to insert, via a needle-tip connector (23-gaugehypodermic tubing connector, 10 mm long), into the intra-venous catheter, while the spring connector was attached tothe head mount of the animal.

Intracranial microinjection apparatus

Intracranial microinjections were delivered by the exper-imenter using a microinjection infusion pump (KD200, KDScientific, MA, USA). Two 1-μl glass microsyringes(Series no. 80,100; Hamilton, NV, USA) were each at-tached to polyethylene tubing (PE20, Microtube Extru-sions, NSW, Australia) which terminated with two 15-mmmicroinjectors (33-gauge; Small Parts, FL, USA).

Procedure

Surgery

All rats underwent surgery to implant an indwelling cath-eter into the right external jugular vein, to fit a screw-assembly head mount and to implant bilateral intracranialcannulae in the nucleus accumbens. Procedures for im-planting jugular vein catheters were as previously de-scribed (Norwood et al. 2003). Intracranial cannulae wereimplanted immediately after jugular vein catheter surgerywhile fitting the head mount for self-administration pro-cedures. The head mount consisted of a screw assembly(grooved plastic bead, ∼5 mm in diameter) placed at thecaudal part of the skull, glued in position with cranioplasticcement and four small screws (Small Parts, FL, USA)tapped into the skull. At the same time, animals had guidecannulae (14 mm, 26-gauge) bilaterally implanted 1 mmabove the nucleus accumbens using stereotaxic proceduresat the co-ordinates AP=+1.3 mm, ML=±1.5 mm and DV=−6.5 mm from Bregma using the rat brain atlas of Paxinosand Watson (1998). The bilateral guide cannulae werelowered through two drill holes in the skull and secured inplace by the application of cranioplastic cement around thecannulae, screw assembly and screws.

After surgery, rats recovered in a warm recovery cham-ber and were returned to a clean home cage upon recovery.Immediately after surgery and for two subsequent days,rats were treated with the analgesic Flunixin (2.5 mg kg−1,s.c.; Troy Laboratories, VIC, Australia). Catheter patencywas maintained by the daily intravenous flush of 0.2 ml ofantibiotic (100 g ml−1, Cephazolin Sodium, David BullLaboratories, VIC, Australia) in 100 IU ml−1 of hepa-rinized saline (David Bull Laboratories, VIC, Australia).Body weight and general health were monitored daily. Allinstruments used during surgery were sterilized with a glassbead sterilizer (Fine Science Tools, CA, USA).

Intravenous drug self-administration procedure

After 5–7 days recovery from surgery, the rats began daily2-h speedball self-administration sessions conducted dur-ing the 12-h light cycle. Prior to each session, the operantchambers were wiped down with 70% ethanol solution andall infusion system connectors were flushed with saline tominimize contamination of the intravenous catheter. Ratswere placed in the operant chamber, the spring connectorscrewed onto their head mount, their intravenous catheterflushed with 0.1 ml of heparinized saline (10 IU ml−1), andthe connector to the infusion line inserted via a needle-tipinto the catheter and secured with a clip. The number ofdrug infusions, active-lever presses, inactive-lever pressesand locomotor activity counts were recorded during the 2 hself-administration sessions. At the end of each session, theinfusion line was disconnected, the intravenous catheterwas flushed with 0.2 ml of the antibiotic solution, thecatheter was closed with a pin and the spring connector wasunscrewed from the head mount. Food or water was notavailable during the 2-h session.

Fixed-Ratio 1 (FR1) schedule

Rats were initially shaped to self-administer a training doseof combined heroin (36 μg kg−1 infusion−1) and cocaine(600 μg kg−1 infusion−1) on an FR1 continuous reinforce-ment schedule. This relatively high speedball dose waschosen to facilitate self-administration behaviour. If therat failed to self-administer the speedball through lever-pressing, “priming” injections were given when the rat wasin close proximity to the lever, to shape self-administrationbehaviour. Stable self-administration baselines were estab-lished for every rat within 7 days (number of infusionsobtained varied less than ±10% compared to previous day)and rats were then advanced to a progressive ratio (PR)schedule.

Progressive-Ratio (PR) schedule

The PR schedule used was defined by an exponential equa-tion in which the reinforcement number is a natural logarith-mic function of the ratio value: Ratio=((5×(ereinforcer #×0.2)−5).

23

The lever press requirement for the first 20 successive rein-forcers progresses as follows: 1, 2, 4, 6, 9, 12, 15, 20, 25, 32,40, 50, 62, 77, 95, 118, 145, 178, 219 and 268 (Duvauchelleet al. 1998).

Each PR session was completed after 2 h or if noinfusion had been self-administered for more than 30 min.Once drug infusion baselines had begun to stabilize on thePR schedule (taking approximately 3 days), the trainingdose of heroin and cocaine was diluted eightfold (heroin,4.5 μg kg−1 infusion−1; cocaine 75 μg kg−1 infusion−1),with the ratio of heroin and cocaine kept to 17:1. This doseof speedball lies on the ascending limb of the speedballdose–response curve under a PR schedule (Duvauchelleet al. 1998), thus facilitating the detection of both increasesand decreases in lever responding.

Once stable self-administration baselines were estab-lished on the reduced ‘test’ speedball dose (taking approxi-mately 5 days), sham microinjections commenced. Thisprocedure involved inserting a 15-mm microinjection needle(33-gauge) through the cannulae and 1 mm into the nucleusaccumbens 2 min prior to the self-administration session.After two sham procedures with stable self-administrationbaselines, testing commenced with antagonist treatments.

On test days, a 14-mm obturator was first placed into thecannulae and removed, to ensure that the cannulae were notblocked. For every drug treatment, the microinjectors werefirst wiped with 70% ethanol, dried and placed insidean ampoule containing the antagonist drug or vehicle. A15-mm microinjector was then inserted into each cannula,entering at a depth of 1 mm into the nucleus accumbens.The microinjection apparatus was then activated to deliver0.5 μl of the antagonist drug over a period of 1 min, whilethe rat sat within an open plastic container. The microin-jector remained in the cannula for an additional 30 s afterdrug infusion to ensure total drug infusion. The injectorswere then removed from the cannulae and replaced with14-mmmicroinjection obturators (33-gauge) into each can-nula. The treated rat was then placed in the operant cham-ber for 2 h and self-administration behaviour assessed.

At the completion of each 2-h session, rats were dis-connected from the infusion system and returned to theanimal housing facilities. Rats did not receive more thanfour central injections (vehicle + two or three drug doses)to ensure the structural integrity of the nucleus accumbenswas maintained. The experiment was a within subjectsdesign where each animal received all doses of a particularantagonist plus vehicle. To control for any order effect ofdrug administration, doses were administered following aLatin-square design. There were four groups. Group 1received the dopamine D1 receptor antagonist SCH 23390[vehicle (distilled water), 1 and 6 nmol side−1]. Group 2received the μ-opiate antagonist CTOP [vehicle (saline),0.1, 0.3 and 1.0 nmol side−1]. Group 3 received the do-pamine D2 receptor antagonist raclopride [vehicle (saline),3 and 10 nmol side−1) and one combination mixture of lowdoses of SCH 23390 (1 nmol side−1) and the μ-opiate re-ceptor antagonist CTOP (0.1 nmol side−1). The combina-tion treatment was tested in the raclopride treated group.This allowed the control over any possible side effects of

repeated SCH 23390 or CTOP exposure in Groups 1 or 2.Group 4 was treated with the δ-opiate receptor antagonistNaltrindole (NAL) [vehicle (distilled water), 1, 3 and 10nmol side−1].

Drugs

In previous studies, speedball solutions have involved com-binations of heroin and cocaine at a ratio of approximately17:1 (Duvauchelle et al. 1998) and this dose ratio was usedfor this study. The speedball solution was prepared dailywith the following method. Heroin (Australian Govern-ment Analytical Laboratories, NSW) was dissolved in ab-solute ethanol (1 mg/0.093 ml), and made to a solutionof 0.18–0.24 mg ml−1 using saline (depending on bodyweight). An equal volume of cocaine hydrochloride (3–4mg ml−1; Macfarlan Smith, Edinburgh, UK) in saline wasadded to make the high dose solution of speedball usedin self-administration training for the experiments. Thisequates to adding equal amounts of heroin (72 μg kg−1

infusion−1) and cocaine (1,200 μg kg−1 infusion−1) solu-tions to obtain the training doses listed below.

The dose of speedball administered under the FR1 re-inforcement schedule consisted of heroin (36 μg kg−1

infusion−1) and cocaine (600 μg kg−1 infusion−1). The doseof speedball administered under the PR reinforcementschedule was initially the training dose, then a saline-diluted solution (8:1): heroin (4.5 μg kg−1 infusion−1) andcocaine (75 μg kg−1 infusion−1).

Intra-accumbens microinjections were made using thefollowing drugs: the dopamine D1 receptor antagonistSCH 23390 (R(+)-7-chloro-8-hydroxy-3-methyl-1-phenyl-2,3,4,5-tetrahydro-1H-3-benzazepine hydrochloride) (Sigma,USA), diluted in distilled water; the dopamine D2 receptorantagonist Raclopride ((S)-3,5-dichloro-N-[1-ethyl-2-pyr-rolidinyl)methyl]-2-hydroxy-6-methoxy-benzamide-L-tar-trate) (Sigma, USA) diluted in 0.9% isotonic saline; CTOP(D-Pen-Cys-Tyr-D-Trp-Orn-Thr-Pen-Thr-NH2) (Tocris, MO,USA), diluted in 0.9% isotonic saline and Naltrindole(Tocris, MO, USA), diluted in distilled water.

Histology

At the completion of each study, the rats were deeplyanaesthetized with pentobarbitone sodium (100 mg kg−1,i.p.) and transcardially perfused with saline, followed by4% formaldehyde in phosphate-buffered saline. The brainwas removed and stored in 4% formaldehyde for at least 1week before sectioning. Coronal sections (40 μm thick)were obtained from the injection site with the use of acryostat and mounted on gelatin-coated slides. The sec-tions were stained with Cresyl Violet and coverslippedwith Eukitt mounting medium. Cannula placements in thenucleus accumbens were verified relative to the rat brainatlas of Paxinos and Watson (1998). Only animals withbilateral cannulae placement into the nucleus accumbenswere included in data analysis.

24

Data analysis

An initial analysis compared responding during FR and PRreinforcement schedules. For this, the number of infusionsper schedule was compared using a one way repeated-measures analysis of variance (ANOVA) with schedule as afactor. Comparisons between active and inactive leverresponses for each schedule were made using a two-wayrepeated measures ANOVA (lever×schedule). Locomotoractivity was compared across schedules using a one-wayrepeated measures ANOVA with schedule as a factor.

The effects of the different antagonist treatments onspeedball self-administration were analyzed using within-subjects analysis. The effect of intra-accumbens antagonisttreatment on infusion breakpoints was analyzed using aone-way repeated measures ANOVAwith dose as a factor.To compare drug effects on active and inactive lever re-sponding, a two-way repeated measures ANOVA wasemployed (treatment×lever). The effect of treatments onlocomotor activity was analyzed using a one-way repeated

measures ANOVAwith dose as a factor. For all tests, post-hoc comparisons employed the Tukey’s Honestly Signif-icant Difference (HSD) with repeated measures test. AllANOVAs were performed using Statview software (ver-sion 5.0.1, 1998).

Results

Histology

Figure 1 displays the bilateral cannula placement into theregion of the nucleus accumbens for each treatment group.One rat in Group 1 (SCH 23390) had incorrect cannulaeplacement and has been removed from all data analyses.Rats in Group 1 (SCH 23390) and Group 3 (Racloprideplus mixture) had injections localized between 0.70 and 1.2mm anterior of Bregma. Rats in the remaining two groupshad injection placement between 0.7 and 1.0 mm anteriorof Bregma.

Fig. 1 Location of microinjec-tion sites in the nucleus accum-bens of rats in each treatmentgroup. Microinjections were inthe core and shell of the nucleusaccumbens from 0.70 to 1.2 mmanterior to Bregma. Drawingsare adapted from the atlas ofPaxinos and Watson (1998)

25

Schedules of intravenous speedball self-administration

Table 1 shows baseline speedball self-administration dataunder the different reinforcement schedules (FR versus PR)and for the two different speedball doses used under the PRschedule (high versus low dose).

Across the three conditions, there was a significant lever(F1,52=42.087, p<0.0001) effect, indicating that the ratshad learned to discriminate between levers for drug reward.There was also a significant schedule effect for lever re-sponding (F2,104=20.395, p<0.0001). Post-hoc compari-sons reveal that either dose with PR schedule had greater

active lever responding (PR high p<0.05; PR low p<0.05)and significantly reduced speedball infusions compared tothe FR schedule (PR high p<0.05; PR low p<0.05). Ani-mals on low dose PR also showed significantly less activelever responding than when on the high dose of speed-ball (p<0.05). The number of infusion breakpoints for PRschedule was also decreased when the dose of speedballwas reduced (p<0.05). Compared to the FR schedule, ani-mals on the PR schedule showed a significant increasein inactive lever responding (PR high p<0.05; PR lowp<0.05). Inactive lever responding was not different be-tween the two doses on PR schedule. Locomotor activityfollowing the high dose of speedball infusions on FR or PRwas not significantly different, yet was significantly re-duced following administration of low dose speedball onPR (FR p<0.05, PR high p<0.05).

Dopamine D1 receptor antagonist effect in the nucleusaccumbens on speedball self-administration

Figure 2 shows the effect of intra-accumbens micro-injection of the D1 receptor antagonist SCH 23390 onspeedball self-administration and locomotor activity. Therewas a significant treatment effect of intra-accumbens SCH23390 administration on speedball infusion breakpoints(F2,10=14.029, p=0.0013). Compared to vehicle treatment,SCH 23390 significantly decreased infusion breakpoints ata high and low dose level (p<0.05).

Two-way repeated measures ANOVA showed a sig-nificant treatment (F2,20=7.127, p=0.0046) and lever

Table 1 Baseline intravenous self-administration data across rein-forcement schedules and heroin/cocaine (speedball) dose

Schedule and group Infusions Active Inactive Activity

FR1 (high dose) 24 (2) 40 (3) 5 (2) 3,802 (212)PR (high dose) 11 (1)* 215 (38)* 14 (3)* 3,887 (151)PR (low dose) 8 (1)*a 103 (16)*a 13 (2)* 2,970 (162)*a

Data are shown as mean (SEM) with n=27 for each condition.Active and inactive refer to responses on the active and inactivelevers. Active lever responding includes those performed during the20-s time-out period. Activity refers to locomotor activity in the 2-hsession. FR1 data are taken from the last session before transferringto the PR (high dose) schedule and PR (high dose) data are takenfrom the last session before transferring to the PR (low dose). PR(low dose) data are taken from the day prior to the first intracranialvehicle injection for rats in each treatment group*Significant difference between FR1 and PR schedule (p<0.05)aSignificant difference between high and low speedball dose underPR schedule (p<0.05)

Fig. 2 Effect of microinjectionof the dopamine D1 receptorantagonist SCH 23390 (SCH)into the nucleus accumbens onA infusion breakpoint, B activelever responses, C locomotoractivity and D inactive leverresponses during 2-h speedballself-administration sessions.Data are shown as mean±SEM,n=6. *Significant differencefrom vehicle treatment (p<0.05)

26

(F1,10=14.396, p=0.0035) effect illustrating that treatmentwith SCH 23390 reduced responding for speedball admin-istration and that rats selectively pressed on the drug-pairedlever. The lack of a significant treatment×lever effect indi-cates that treatment with SCH 23390 inhibited respondingon both active and inactive levers. Post-hoc comparisonsreveal that both doses of SCH 23390 reduced active andinactive lever responding [active: 1 nmol side−1 (p<0.05), 6nmol side−1 (p<0.05); inactive: 1 nmol side−1 (p<0.05), 6nmol side−1 (p<0.05); Fig. 2B, D].

There was a significant treatment effect of SCH 23390on locomotor activity (F2,10=10.268, p=0.0037). Locomo-tor activity was significantly reduced by pretreatment withboth doses of SCH 23390 (p<0.05, Fig. 2C).

μ-Opiate receptor antagonist effect in the nucleusaccumbens on speedball self-administration

Figure 3 shows the effect of intra-accumbens treatmentwith the μ-opiate receptor antagonist CTOP on speedballself-administration. One way repeated measures ANOVArevealed a significant treatment effect of CTOP adminis-tration on speedball infusions (F3,18=5.813, p=0.0058).Only the high dose of CTOP administered into the nucleusaccumbens significantly reduced infusion breakpoints (1.0nmol side−1 CTOP; p<0.05), while the lower doses of 0.1and 0.3 nmol side−1 CTOP were not effective (Fig. 3A).

Rats pretreated with CTOP showed a significant treat-ment (F3,36=4.848, p=0.0062) and lever (F1,12=7.872,p=0.0159) effect, indicating that CTOP treatment reducedlever responding and that rats showed a preference foractivity at the drug-paired lever. A significant treatment×lever interaction (F3,36=4.382, p=0.0099) shows that CTOP

administration significantly reduced active lever respond-ing. Post-hoc comparisons reveal that only the highest doseof CTOP reduced active lever pressing (p<0.05) but waswithout effect on the inactive levers (Fig. 3B, D).

There was not a significant treatment effect of CTOPadministration on locomotor activity during speedball self-administration (F3,18=1.232, p=0.3271; Fig. 3C).

Dopamine D2 receptor antagonist and D1/μ-opiatereceptor antagonist combination effect in the nucleusaccumbens on speedball self-administration

One animal from Group 3 (rat N2) was a high responder forspeedball infusions and had data levels greater than 2standard deviations of the group means [active responses tovehicle treatment—323 (14 speedball infusions)]. Thisanimal has been excluded from statistical analyses forintra-accumbens treatment yet showed a similar pattern ofdrug effect.

Figure 2 shows the effect of intra-accumbens micro-injection of the D2 receptor antagonist raclopride or thecombined low doses of SCH 23390 and CTOP on speed-ball self-administration and locomotor activity. Intra-ac-cumbens treatment produced a significant treatment effecton speedball infusions (F3,15=9.075, p=0.0011). Comparedto vehicle treatment, raclopride significantly decreased in-fusion breakpoints at a high, but not low, dose (10 nmolside−1; p<0.05). The intra-accumbens administration of amixture of low dose SCH 23390 (1 nmol side−1) and CTOP(0.1 nmol side−1) significantly reduced infusion break-points for speedball self-administration (p<0.05; Fig. 4A).

Two-way repeated measures ANOVA showed a sig-nificant treatment (F3,30=6.100, p=0.0023) and lever

Fig. 3 Effect of microinjectionof the μ-opiate receptor antago-nist CTOP into the nucleusaccumbens on A infusionbreakpoint, B active lever re-sponses, C locomotor activityand D inactive lever responsesduring the 2-h speedball PRself-administration session. Dataare shown as mean±SEM, n=7.*Significant difference from ve-hicle treatment (p<0.05)

27

(F1,10=10.893, p=0.0080) effect illustrating that treatmentwith raclopride or mixture reduced responding for speed-ball administration and that rats selectively pressed on thedrug-paired lever. A significant treatment×lever interactionsuggest that the treatment effect is selective for one of thelevers (F3,30=6.101, p=0.0023). Post-hoc comparisons in-dicate that the highest dose of raclopride (10 nmol side−1,p<0.05) and the mixture of SCH 23390 and CTOP (p<0.05)significantly reduced active lever responding for speedball(Fig. 4b, d). None of the treatments affected inactive lever

responses. Locomotor activity was not significantly affectedby intra-accumbens treatment (F3,15= 2.973, p=0.0653;Fig. 4C).

δ-Opiate receptor antagonist effect in the nucleusaccumbens on speedball self-administration

Figure 5 shows the effects of the administration of the δ-opiate receptor antagonist naltrindole into the nucleus

Fig. 4 Effect of microinjectionof the D2 receptor antagonistraclopride (RAC) or the combi-nation of a low dose of SCH23390 (1 nmol side−1) plusCTOP (0.1 nmol side−1) (SCH+CTOP) on A infusion break-point, B active lever responses,C locomotor activity and Dinactive lever responses duringthe 2-h speedball PR self-administration session. Data areshown as mean±SEM, n=6.*Significant difference fromvehicle treatment (p<0.05)

Fig. 5 Effect of microinjectionof the δ-opiate receptor antago-nist naltrindole (NAL) into thenucleus accumbens on A infu-sion breakpoint, B active leverresponses, C locomotor activityand D inactive lever responsesduring the 2-h speedball PRself-administration session. Dataare shown as mean±SEM, n=7.*Significant difference from ve-hicle treatment (p<0.05)

28

accumbens on speedball self-administration. There was nota significant treatment effect of naltrindole administrationon speedball infusions (F3,18=0.456, p=0.7166; Fig. 5A).Two-way repeated measures ANOVA showed a significantlever effect (F1,12=18.931, p=0.0009) indicating specificresponding at the active lever compared to the inactivelever. There was not a significant treatment effect on leverresponding following naltrindole administration (Fig. 5B,D). There was not a significant effect of intra-accumbensnaltrindole treatment on locomotor activity during speed-ball self-administration (F3,18=1.294, p=0.3068; Fig. 5C).

Discussion

In the present study, as in previous studies (Duvauchelleet al. 1998; Hemby et al. 1999), the combination of heroinand cocaine (speedball) was reliably self-administered byrats. As would be expected, FR1 schedules lead to a greaternumber of self-administered infusions than PR scheduleswhile PR schedules produced more active lever responsesdue to the increased response requirement for drug infu-sions. Inactive lever responses moderately increased fromFR1 to PR schedules, indicating that increased respondingat the active lever under PR schedules may also transferacross to the inactive lever due to an overall increase in leverpressing.

While animals on the high dose PR schedule receivedapproximately half of the number of speedball infusionsthan when on the FR1 schedule, locomotor activity didnot significantly change across these two conditions. Thissuggests that fulfilling the response requirement for PRschedules may also significantly enhance locomotor activ-ity, or that the measured amount of locomotor activity forthe 2-h period was maximal at a smaller number of in-fusions. Diluting the speedball dose under the PR schedulesignificantly reduced infusions, responses and locomotoractivity, consistent with a reduced level of reinforcementand locomotor activation with a lower concentration ofspeedball.

The present study indicates that the reinforcing effects ofintravenous speedball self-administration are reduced bythe intra-accumbens administration of dopamine receptorand μ-opiate receptor antagonists, but not by a δ-opiatereceptor antagonist. In addition, the intra-accumbens ad-ministration of combined low doses of a D1 and a μ-opiatereceptor antagonist into the nucleus accumbens was alsoeffective at reducing speedball self-administration.

Rats microinjected with the D1 receptor antagonist SCH23390 in to the nucleus accumbens showed a dose-dependent reduction in breakpoints under the PR schedule,with reduced responding on both the active and inactivelevers. Locomotor activity was also dose-dependently re-duced by SCH 23390. Intra-accumbens treatment with thedopamine D2 receptor antagonist raclopride dose-depen-dently reduced breakpoints and active lever presses butwithout affecting locomotor activity. Overall, these resultssuggest that both D1 and D2 dopamine receptors in thenucleus accumbens may mediate the self-administration of

speedball, although with an inhibitory effect of SCH23390 on locomotor activity and inactive lever presses, ageneral performance decrement with this treatment cannotbe ruled out.

As animal subjects are required to complete an increas-ing number of lever-presses under a PR reinforcementschedule to obtain a drug, the drug must provide them withsufficient motivation to fulfil the response requirement. Inthe present study, dopamine receptor antagonists may havedecreased this motivation, thereby resulting in the observeddecreases in drug self-administration (Arnold and Roberts1997; Cardinal and Everitt 2004). These findings aresimilar to those of previous studies conducted at the sys-temic level, which demonstrate that selective dopamineD1 and D2 receptor antagonists attenuated cocaine self-administration in rats (Awasaki et al. 1997; Corrigall andCoen 1991a; Richardson et al. 1993). Additionally, thesefindings are comparable to those of studies conducted at thelevel of the nucleus accumbens, which demonstrate thatdopamine transmission is directly reinforcing in this region(Carlezon et al. 1995; Ikemoto et al. 1997). Consequently,intra-accumbens and intravenous cocaine self-administra-tion can be disrupted by the local microinjection of either adopamine D1 or D2 receptor antagonist into the nucleusaccumbens (Ikemoto et al. 1997; McGregor and Roberts1993; Phillips et al. 1994).

Although systemic administration of opiates modest-ly increases dopamine levels in the nucleus accumbens(DiChiara and Imperato 1988; Gerasimov and Dewey1999; Maisonneuve et al. 2001; Wise et al. 1995), the in-volvement of nucleus accumbens dopamine transmissionin opiate reward is not well defined. Following dopamin-ergic lesions in the nucleus accumbens, rats continued toself-administer intravenousmorphine or heroin (Dworkin etal. 1988; Pettit et al. 1984). The systemic administration ofthe D2 receptor antagonist sulpiride inhibited intracranialself-administration of morphine into the VTA but not intothe nucleus accumbens (David et al. 2002). Conversely, thereinforcing efficacy of heroin is reduced following inhibi-tion of dopaminergic cells in the VTA which is accom-panied by a significant reduction in nucleus accumbensextracellular dopamine levels (Xi and Stein 1999). Further-more, cellular adaptations of the dopamine system in thenucleus accumbens have been demonstrated with chronicheroin self-administration (Self et al. 1995). While thepresent study does not provide data for heroin self-ad-ministration alone, clearly the self-administration of com-bined heroin and cocaine depends upon both dopamine D1and D2 receptors in the nucleus accumbens.

The finding that SCH 23390 also produced a significantdecrease in speedball-induced locomotor activity is worthnoting. Other studies have reported that microinjections ofSCH 23390 into the nucleus accumbens decreases loco-motor activity (McGregor and Roberts 1993). Hypoactivitycould result in parallel dose-dependent decreases in theself-administration of a drug or a non-drug reinforcer, suchas food, reflecting a general disruption in operant behav-iour (Mello and Negus 1996). However, others have dem-onstrated that low doses of SCH 23390 selectively reduce

29

cocaine self-administration without altering responding forfood (Caine and Koob 1994) and similar doses of SCH23390 infused into the nucleus accumbens increase leverpressing for cocaine on a continuous reinforcement sched-ule (Caine et al. 1995; Cornish et al. 1999).

The administration of the μ-opiate receptor antagonistCTOP into the nucleus accumbens also dose-dependentlyreduced infusion breakpoints for speedball self-adminis-tration and decreased active lever responding. This effectwas not associated with decreased locomotor activity. Onceagain, this suggests that locomotor activity and infusionbreakpoints are, to a certain extent, independently mediatedas the reduction of one behaviour did not decrease theother. This suggests that μ-opiate receptors in the nucleusaccumbens mediate the reinforcing effects of heroin/cocaine combinations. This agrees with previous findingsthat systemically administered μ-opiate receptor antago-nists modulate cocaine (Corrigall and Coen 1991a,b; Melloet al. 1990; however, see also Killian et al. 1978; Rowlettet al. 1998), heroin (Killian et al. 1978; Koob et al. 1984;Rowlett et al. 1998) and speedball (Hemby et al. 1996;Rowlett et al. 1998) self-administration in rat and rhesusmonkey. Additionally, these findings agree with demon-strations that heroin, morphine and opioid peptides aredirectly reinforcing when injected into the nucleus accum-bens (Goeders et al. 1984; Olds 1982; Van Der Kooy et al.1982) and that intra-accumbens infusions of μ-opiatereceptor antagonists decreases intra-accumbens or intrave-nous heroin, morphine, opioid peptide and cocaine self-administration (Goeders et al. 1984; Olds 1982; Ward et al.2003). Recent data also suggests that accumbens μ-opiatereceptors may be selectively involved in regulating he-donic aspects of reinforcement processes (for review, seeCardinal and Everitt, 2004). Collectively, these results sug-gest that intra-accumbens μ-opiate receptors mediate re-ward processes initiated by heroin, cocaine and heroin/cocaine combinations.

In contrast to the other antagonists, the δ-opiate receptorantagonist naltrindole was without effect on speedball self-administration at the doses tested. While it is possible thatthe doses used may have been too low, it is worth notingthat these doses were effective in reducing lever respondingfor ethanol reward following administration into thenucleus accumbens (Hyytia and Kiianmaa 2001). Usinghigher doses may have brought into question the receptorselectivity of any drug effect. The present study indicatesthat effects of the highest dose of naltrindole were quitevariable.

Previous studies have shown that systemic adminis-tration of naltrindole reduced heroin self-administration(Negus et al. 1993), yet the involvement of δ-opiate re-ceptors in cocaine self-administration are less clear (DeVries et al. 1995; Negus et al. 1995). While there arelimited reports on the role of δ-opiate receptors in thenucleus accumbens in heroin or cocaine self-administra-tion, it is known that intracranial self-stimulation can beenhanced by μ-opiate and δ-opiate receptor agonists in-jected in the nucleus accumbens. However, the effective-ness of the δ-opiate agonist is regionally specific to the

shell area of this region (Johnson et al. 1995). In addition,intracisternally administered naltrindole inhibits amphet-amine-induced dopamine release in the striatum but not thenucleus accumbens (Schad et al. 1996). Together with thepresent findings, these studies suggest that δ-opiate re-ceptors in the nucleus accumbens are not essential me-diators of drug reinforcement.

The current findings concur with those of studies exam-ining the effects of systemically administered dopamineand opioid receptor antagonists on speedball self-admin-istration. Mello and Negus (1996, 1999) reported that thecombination of a dopamine and opioid receptor antagonistsignificantly reduced speedball self-administration in pri-mates, whereas independent administration of the sameantagonist doses did not alter speedball intake. Here, wereport that speedball self-administration was reduced by amixture of a D1 and μ-opiate receptor antagonist injectedinto the nucleus accumbens. Our data also shows that athigher doses of D1 or μ-opiate receptor antagonists, theself-administration of speedball was reduced by inhibitingone receptor type. However, in the case of SCH 23390,higher doses also greatly reduced general locomotoractivity and inactive lever presses. While the CTOP/SCH23390 mixture reduced speedball self-administration to asimilar extent as the low dose of the D1 antagonist alone,the addition of the μ-opiate receptor antagonist removedthe non-selective effects of the D1 antagonist on locomotoractivity and inactive lever pressing. Intra-accumbens ad-ministration of the μ-opiate receptor antagonist CTOPtended to increase locomotor activity (see Fig. 3) and ithas been demonstrated that activation of μ-opiate receptorsin the nucleus accumbens leads to hypoactivity followed byhyperactivity (Meyer et al. 1994; Pert and Sivit, 1977).Combined with the complex neuronal localization of μ-opiate receptors in the nucleus accumbens (Svingos et al.1997), the locomotor effects of μ-opiate receptor modu-lation suggests that antagonism of these receptors at apresynaptic or interneuronal level to alter γ-amino-butyricacid (GABA) release could override the post-synaptic lo-comotor effects of dopamine D1 receptor antagonism. Fur-ther, comparable neuronal organization exists in the VTA,where the local administration of CTOP results in activa-tion of efferent projections to enhance locomotor activity(Badiani et al. 1995). Therefore, combined CTOP and SCH23390 administration into the nucleus accumbens maywork in opposition to each other to maintain the locomotoreffect of speedball self-administration while reducing thereinforcing efficacy of speedball. With respect to pharma-cotherapies for speedball addiction, it is important to con-sider the benefits of combined drug treatment, so that highdose side effects of individual antagonists can be avoided.It would clearly be of interest to examine whether lowerdoses of SCH 23390, with no intrinsic effect, would be-come effective in combination with low doses of CTOP.

In conclusion, the self-administration of heroin andcocaine combinations is sensitive to both reinforcementschedule and infusion concentration. Under a PR scheduleof reinforcement, the inhibition of D1, D2 dopamine or μ-opiate but not δ-opiate receptors in the nucleus accumbens

30

results in a significant reduction in speedball self-admin-istration. The combined administration of a D1 and μ-opi-ate receptor antagonist was more selective at reducingspeedball self-administration than either drug dose alone.These data suggest that dopamine and μ-opiate mecha-nisms in the nucleus accumbens are important in mediatingcombined heroin and cocaine use and are consistent withthe idea that combined medications aimed at dopamine andopiate systems may prove beneficial as pharmacotherapiesfor speedball addiction.

Acknowledgements This work was supported by a University ofSydney Sesqui Fellowship Grant (J.L.C.).

References

Arnold JM, Roberts DC (1997) A critique of fixed and progressiveratio schedules used to examine the neural substrates of drugreinforcement. Pharmacol Biochem Behav 57:441–447

Awasaki M, Nishida TL, Sasaki Y, Sato S (1997) Dopamine D1antagonist SCH-23390 attenuates self-administration of bothcocaine and fentanyl in rats. Environ Toxicol Pharmacol 3:115–122

Badiani A, Leone P, Stewart J (1995) Intra-VTA injections of themu-opioid antagonist CTOP enhance locomotor activity. BrainRes 690(1):112–116

Caine SB, Koob GF (1994) Effects of dopamine D1 and D2 antag-onists on cocaine self-administration under different schedulesof reinforcement in the rat. J Pharmacol Exp Ther 270:209–218

Caine SB, Heinrichs SC, Coffin VL, Koob GF (1995) Effects of thedopamine D1 antagonist SCH 23390 microinjected into theaccumbens, amygdala or striatum on cocaine self-administra-tion in the rat. Brain Res 692:47–56

Cardinal RN, Everitt BJ (2004) Neural and psychological mecha-nisms underlying appetitive learning: links to drug addiction.Curr Opinion Neurobiol 14:156–162

Carlezon WA, Devine DP, Wise RA (1995) Habit-forming actions ofnomifensine in nucleus accumbens. Psychopharmacology 122:194–197

Cornish JL,Duffy P,Kalivas PW (1999)A role for nucleus accumbensglutamate transmission in the relapse to cocaine-seeking behav-ior. Neuroscience 93:1359–1367

Corrigall WA, Coen KM (1991a) Cocaine self-administration isincreased by both D1 and D2 dopamine antagonists. PharmacolBiochem Behav 39:799–802

Corrigall WA, Coen KM (1991b) Opiate antagonists reduce cocainebut not nicotine self-administration. Psychopharmacology104:167–170

David V, Durkin TP, Cazala P (2002) Differential effects of thedopamine D2/D3 receptor antagonist sulpiride on self-admin-istration of morphine into the ventral tegmental area or thenucleus accumbens. Psychopharmacology 160:307–317

De Vries TJ, Babovic-Vuksanovic D, Elmer G (1995) Lack ofinvolvement of delta-opioid receptors in mediating the reward-ing effects of cocaine. Psychopharmacology 120:442–448

DiChiara G, Imperato A (1988) Opposite effects of mu and kappaopiate agonists on dopamine release in the nucleus accumbensand the dorsal caudate of freely moving rats. J Pharmacol ExpTher 244:1067–1080

Duvauchelle CL, Sapoznik T, Kornetsky C (1998) The synergisticeffects of combining cocaine and heroin (“speedball”) using aprogressive-ratio schedule of drug reinforcement. PharmacolBiochem Behav 61:297–302

Dworkin SI, Guerin GF, Co C, Goeders NE, Smith JE (1988) Lackof an effect of 6-hydroxydopamine lesions of the nucleusaccumbens on intravenous morphine self-administration. Phar-macol Biochem Behav 30:1051–1057

Foltin RW, Fischman MW (1992) The cardiovascular and subjectiveeffects of intravenous cocaine and morphine combination inhumans. J Pharmacol Exp Ther 261:623–632

Gerasimov MR, Dewey SL (1999) Gamma-vinyl γ-aminobutyricacid attenuates the synergistic elevations of nucleus accumbensdopamine produced by a cocaine/heroin (speedball) challenge.Eur J Pharmacol 380:1–4

Goeders NE, Lane LD, Smith JE (1984) Self-administration ofmethionine enkephalin into the nucleus accumbens. PharmacolBiochem Behav 20:451–455

Hemby SE, Smith JE, Dworkin SI (1996) The effects of eticloprideand naltrexone on responding maintained by food, cocaine,heroin and cocaine/heroin combinations in rats. J PharmacolExp Ther 277:1247–1258

Hemby SE, Co C, Dworkin SI, Smith JE (1999) Synergisticelevations in nucleus accumbens extracellular dopamine con-centrations during self-administration of cocaine/heroin combi-nations (speedball) in rats. J Pharmacol Exp Ther 288:274–280

Hyytia P, Kiianmaa K (2001) Suppression of ethanol responding bycentrally administered CTOP and Naltrindole in AA and Wistarrats. Alcohol Clin Exp Res 25:25–33

Ikemoto S, Glazier BS, Murphy JM, McBride WJ (1997) Role ofdopamine D1 and D2 receptors in the nucleus accumbens inmediating reward. J Neurosci 17:8580–8587

Johnson SW, North RA (1992) Opioids excite dopamine neurons byhyperpolarization of local interneurons. J Neurosci 12:483–488

Johnson PI, Goodman JB, Condon R, Stellar JR (1995) Rewardshifts and motor responses following microinjections of opiate-specific agonists into either the core or shell of the nucleusaccumbens. Psychopharmacology 120:195–202

Killian AK, Bonese K, Schuster CR (1978) The effects of naloxoneon behaviour maintained by cocaine and heroin injections inrhesus monkeys. Drug Alcohol Depend 3:243–251

Koob GF, Pettit HO, Ettenberg A, Bloom FE (1984) Effects ofopiate antagonists and their quaternary derivatives on heroinself-administration in the rat. J Pharmacol Exp Ther 229:481–486

Le Moine C, Kieffer B, Gaveriaux-Ruff C, Befort K, Bloch B(1994) Delta-opioid receptor gene expression in the mouseforebrain: localization in cholinergic neurons of the striatum.Neuroscience 62:635–640

Leri F, Bruneau J, Stewart J (2003) Understanding polydrug use:review of heroin and cocaine use. Addiction 98:7–22

Maisonneuve IM, Warner LM, Glick SD (2001) Biphasic dose-related effects of morphine on dopamine release. Drug AlcoholDepend 65:55–63

Martin TJ, Kim SA, Lyupina Y, Smith JE (2002) Differentialinvolvement of mu-opioid receptors in the rostral versus caudalnucleus accumbens in the reinforcing effects of heroin in rats:evidence from focal injections of β-funaltrexamine. Psycho-pharmacology 161:152–159

McGregor A, Roberts DC (1993) Dopaminergic antagonism withinthe nucleus accumbens or the amygdala produces differentialeffects on intravenous cocaine self-administration under fixedand progressive ratio schedules of reinforcement. Brain Res624:245–252

McGregor IS (1996) Using strawberry tree workbenchmac andworkbench PC software for data acquisition and control in theanimal learning laboratory. J Behav Res Meth Instrum Comput28:38–48

Mello NK, Negus SS (1996) Preclinical evaluation of pharma-cotherapies for treatment of cocaine and opioid abuse usingdrug self-administration procedures. Neuropsychopharmacolo-gy 14:375–424

Mello NK, Negus SS (1999) Effects of flupenthixol andquadazocine on self-administration of speedball combinationsof cocaine and heroin by rhesus monkeys. Neuropsychophar-macology 21:575–588

Mello NK, Mendelson JH, Bree MP, Lukas SE (1990) Buprenor-phine and naltrexone effects on cocaine self-administration byrhesus monkeys. J Pharmacol Exp Ther 254:926–939

31

Mendelson JH, Mello NK (1996) Management of cocaine abuse anddependence. N Engl J Med 334:965–972

Meyer ME, McLaurin BI, Allen M, Meyer ME (1994) Biphasiceffects of intra-accumbens μ-opioid peptide agonist DAMGOon locomotor activities. Pharmacol Biochem Behav 47:827–831

National Institute on Drug Abuse (1998) Epidemiologic trends indrug abuse. NIH Publication No 98-4300, Washington DC,USA, p 79

Negus S, Henriksen S, Mattox A, Pasternak G, Portoghese P (1993)Effect of opioid agonists selective for mu, delta and kappaopioid receptors on the reinforcing effects of heroin in rats. JPharmacol Exp Ther 265:1245–1252

Negus SS, Mello NK, Portoghese PS, Lukas SE, Mendelson JH(1995) Role of delta opioid receptors in the reinforcing anddiscriminative stimulus effects of cocaine in rhesus monkeys. JPharmacol Exp Ther 273:1245–1256

Norwood CS, Cornish JL, Mallet PE, McGregor IS (2003) Pre-exposure to the cannabinoid receptor agonist CP 55940 en-hances morphine behavioral sensitization and alters morphineself-administration in Lewis rats. Eur J Pharmacol 465:105–114

Olds M (1982) Reinforcing effects of morphine in the nucleusaccumbens. Brain Res 237:429–440

Paxinos G, Watson C (1998) The rat brain in stereotaxic co-ordinates, 4th edn. Academic Press, CA, USA

Pert A, Sivit C (1977) Neuroanatomical focus for morphine andenkephalin-induced hypermotility. Nature 265:645–647

Pettit HO, Ettenberg A, Bloom FE, Koob GF (1984) Destruction ofdopamine in the nucleus accumbens selectively attenuatescocaine but not heroin self-administration in rats. Psychophar-macology 84:167–173

Phillips GD, Howes SR, Whitelaw RB, Wilkinson LS, Robbins TW,Everitt BJ (1994). Isolation rearing impairs the reinforcingefficacy of intravenous cocaine or intra-accumbens D-amphet-amine: impaired response to intra-accumbens D1 and D2/D3dopamine receptor antagonists. Psychopharmacology 115:419–429

Ranaldi R, Wise RA (2000) Intravenous self-administration ofmethamphetamine–heroin (speedball) combinations under aprogressive-ratio schedule of reinforcement in rats. NeuroRe-port 11:2621–2623

Reid LD, Glick SD, Menkens KA, French ED, Bilsky EJ, Porreca F(1995) Cocaine self-administration and naltrindole, a delta-selective opioid antagonist. NeuroReport 6:1409–1412

Richardson NR, Piercey MF, Svensson K, Collins RJ, Myers JE,Roberts DCS (1993) Antagonism of cocaine self-administrationby the preferential dopamine autoreceptor antagonist, (+)-AJ76. Brain Res 619:15–21

Ritz MC, Lamb RJ, Goldberg SR, Kuhar MJ (1987) Cocainereceptors on dopamine transporters are related to self-admin-istration of cocaine. Science 237:1219–1223

Rowlett JK, Wilcox KM, Woolverton WL (1998) Self-administra-tion of cocaine–heroin combinations by rhesus monkeys:antagonism by naltrexone. J Pharmacol Exp Ther 286:61–69

Schad CA, Justice JB Jr, Holtzman SG (1996) Differential effects ofdelta- and mu-opioid receptor antagonists on the amphetamine-induced increase in extracellular dopamine in striatum andnucleus accumbens. J Neurochem 67:2292–2299

Schutz CG, Vlahov D, Anthony JC, Graham NMH (1994) Com-parison of self-reported injection frequencies for past 30 daysand 6 months among intravenous drug users. J Clin Epidemiol47:191–195

Self DW, McClenahan AW, Beitner-Johnson D, Terwilliger RZ,Nestler EJ (1995) Biochemical adaptations in the mesolimbicdopamine system in response to heroin self-administration.Synapse 21:312–318

Sieden LS, Sabol KE, Ricuarte GA (1993) Amphetamine: effectson catecholamine system and behaviour. Annu Rev PharmacolToxicol 33:634–677

Spanagel R, Herz A, Shippenberg TS (1990) The effects of opioidpeptides on dopamine release in the nucleus accumbens: anvivo microdialysis study. J Neurochem 55:1734–1740

Svingos AL, Moriwaki A, Wang JB, Uhl GR, Pickel VM (1997) μOpioid receptors are localised to extrasynaptic plasma mem-branes of GABAergic neurons and their targets in the ratnucleus accumbens. J Neurosci 17:2585–2594

Svingos AL, Clarke CL, Pickel VM (1998) Cellular sites foractivation of μ-opioid receptors in the rat nucleus accumbensshell: relationship with Met5-enkephalin. J Neurosci 18:1923–1933

Tempel A, Zukin SR (1987) Neuroanatomical patterns of the μ, δ,and κ opioid receptors of rat brain as determined by quanti-tative in vitro autoradiography. Proc Natl Acad Sci 84:4308–4312

Van Der Kooy D, Mucha RF, O’Shaughessy M, Bucenieks P (1982)Reinforcing effects of brain microinjection of morphine re-vealed by conditioned place preference. Brain Res 243:107–117

Ward S, Martin T, Roberts D (2003) Beta-funaltrexamine affectscocaine self-administration in rats responding on a progressiveratio schedule of reinforcement. Pharmacol Biochem Behav75:301–307

Wise RA (2002) Brain reward circuitry: insights from unsensedincentives. Neuron 36:229–240

Wise RA, Leone P, Rivest R, Leeb K (1995) Elevations of nucleusaccumbens dopamine and DOPAC levels during intravenousheroin self-administration. Synapse 21:140–148

Xi Z-X, Stein EA (1999) Baclofen inhibits heroin self-administra-tion behaviour and mesolimbic dopamine release. J PharmacolExp Ther 290:1369–1374

Yu L (1996) The mu opioid receptor: from molecular cloning tofunctional studies. Addict Biol 1:19–30

32