ORIGINAL INVESTIGATION

Acamprosate attenuates cocaine- and cue-induced reinstatement of cocaine-seeking behavior in rats

M. Scott Bowers & Billy T. Chen & Jonathan K. Chou &

Megan P. H. Osborne & Justin T. Gass & Ronald E. See &

Antonello Bonci & Patricia H. Janak & M. Foster Olive

Received: 20 February 2007 /Accepted: 16 July 2007 / Published online: 2 September 2007# Springer-Verlag 2007

AbstractRationale Acamprosate (calcium acetylhomotaurinate) is aglutamatergic neuromodulator used for the treatment ofalcoholism, but its potential efficacy in the treatment ofpsychostimulant addiction has not been explored.Objectives The purpose of this study was to assess theeffects of acamprosate on cocaine-stimulated locomotoractivity, cocaine self-administration, and cue- and cocaine-induced reinstatement of cocaine-seeking behavior.Materials and methods All experiments utilized once-dailytreatment for 5 consecutive days. First, the effects of salineor acamprosate (100, 300, or 500 mg/kg intraperitoneally)on body weight were examined. On the last day oftreatment, locomotor activity was assessed before and afterdrug treatment, after which all animals received an acutechallenge of cocaine (10 mg/kg). Next, a separate group of

rats were trained to intravenously (IV) self-administercocaine (0.6 mg/kg per infusion), subjected to extinctionprocedures, and then tested for effects of acamprosate oncue- or cocaine-induced reinstatement. A third group of ratswas trained to self-administer cocaine as described aboveand were treated with saline or acamprosate before daily IVself-administration sessions.Results Repeated administration of 500 mg/kg acamprosatebut not lower doses produced reductions in both bodyweight and spontaneous locomotor activity, and thus thisdose was not tested further. Acamprosate at 300 mg/kg butnot 100 mg/kg attenuated both cocaine- and cue-inducedreinstatement without altering baseline patterns of cocaineself-administration or cocaine-stimulated hyperlocomotion.Conclusions Acamprosate attenuates both drug- and cue-induced reinstatement of cocaine-seeking behavior, suggest-ing that this compound may serve as a potential treatment forpreventing relapse in cocaine-addicted humans.

Keywords Cocaine . Self-administration . Reinstatement .

Relapse . Cues . Drug priming . Acamprosate . Glutamate .

Locomotor activity . Body weight

Introduction

Acamprosate prolongs abstinence and reduces subjectivemeasures of alcohol craving in alcoholic patients (seeHeilig and Egli 2006; Rosenthal 2006; Soyka and Roesner2006 for recent reviews). However, the potential efficacy ofacamprosate in reducing the propensity to relapse to the useof abused drugs other than alcohol has received littleattention. We recently demonstrated that acamprosateprevents both the development and cocaine-induced rein-statement of a cocaine-conditioned place preference in mice

Psychopharmacology (2007) 195:397–406DOI 10.1007/s00213-007-0904-y

M. S. Bowers : B. T. Chen : J. K. Chou :A. Bonci : P. H. JanakErnest Gallo Clinic and Research Center,Department of Neurology,University of California at San Francisco,Emeryville, CA 94608, USA

M. P. H. Osborne : J. T. Gass : R. E. See :M. F. Olive (*)Center for Drug and Alcohol Programs,Department of Psychiatry and Behavioral Sciences,Medical University of South Carolina,67 President Street,P.O. Box 250861, Charleston, SC 29425, USAe-mail: olive@musc.edu

R. E. SeeDepartment of Neurosciences,Medical University of South Carolina,Charleston, SC 29425, USA

A. Bonci : P. H. JanakWheeler Center for the Neurobiology of Addiction,University of California at San Francisco,San Francisco, CA 94143, USA

(Mcgeehan and Olive 2003, 2006), suggesting that acam-prosate may modify the rewarding and relapse-inducingeffects of acute cocaine exposure.

Preclinical data indicate that blockade of either N-methyl-D-aspartate (NMDA) ionotropic or type 5 metabotropic(mGlu5) glutamate receptors or normalization of extracellu-lar glutamate levels in regions such as the nucleusaccumbens can attenuate reinstatement of cocaine-seekingbehavior (Backstrom and Hyytia 2006, 2007; Baker et al.2003; Lee et al. 2005; Maldonado et al. 2007; but seeFamous et al. 2007; See et al. 2001). While the preciseneuropharmacological mechanisms of action for acamprosateare not fully understood (De Witte et al. 2005), electrophys-iological studies have shown that acamprosate modulates thefunction of NMDA receptors (Berton et al. 1998; Madambaet al. 1996; Popp and Lovinger 2000; Rammes et al. 2001;Zeise et al. 1990, 1993) and may also antagonize the mGlu5receptor (De Witte et al. 2005; Harris et al. 2002, 2003).Additional studies utilizing in vivo microdialysis have shownthat acamprosate dampens extracellular levels of glutamatewhen the organism is in a hyperglutamatergic state (Dahchourand De Witte 2000, 2003; Dahchour et al. 1998; Spanagelet al. 2005).

Given that acamprosate appears to exert some of its actionsvia modulating glutamate receptor function or extracellularlevels of glutamate, the purpose of the present study was toexamine the effects of acamprosate on cocaine-inducedlocomotion, cocaine self-administration, as well as cocaine-and cue-induced reinstatement of cocaine-seeking behavior.

Materials and methods

Subjects

All experimental procedures conformed to the 1996 NationalInstitutes of Health Guide for the Care and Use of LaboratoryAnimals, the 2003 Guide for the Care and Use of Mammalsin Neuroscience and Behavioral Research, and with theapproval of an Institutional Animal Care and Use Commit-tee. Male Sprague–Dawley rats (300–325 g, Charles RiverLaboratories, Wilmington, MA) were individually housedupon arrival and were weighed and handled daily for 1 weekto assess health and allow acclimation to handling proce-dures. Animals were maintained on a 12-h light–dark cycle(lights on at 07:00 hours) with ad libitum access to foodunless otherwise stated. All experimentation was conductedduring the light phase of the light–dark cycle.

Drugs

Cocaine hydrochloride was obtained from either Sigma-Aldrich (St. Louis, MO) or the National Institute on Drug

Abuse Drug Supply Program (Research Triangle Park, NC).Acamprosate was obtained from EstechPharma (Seoul, SouthKorea). Ketamine and xylazine were obtained from theUniversity of California at San Francisco or MedicalUniversity of South Carolina hospital pharmacy. Pentobarbitalsodium and chloral hydrate were obtained from Sigma-Aldrich. Methohexital was obtained from King Pharmaceu-ticals (Bristol, TN). Cocaine and acamprosate were dissolvedin sterile saline (0.9% w/v sodium chloride, Hospira, LakeForest, IL) before intravenous (i.v.) or intraperitoneal (i.p.)administration. All i.p. administrations were given in avolume of 1 ml/kg. Repeated dosing of acamprosate wasutilized because it has been reported that plasma levels ofacamprosate do not stabilize until at least 5 days of treatment(Durbin et al. 1996; Wilde and Wagstaff 1997).

Monitoring of locomotor activity and body weight

After acclimation to handling procedures, animals wererandomized into one of four groups and received daily i.p.injections of saline or acamprosate (100, 300, or 500 mg/kgi.p.) for 4 days. Body weights were measured dailyapproximately 1 h before drug or saline administration.On the fifth day of treatment, animals were weighed andplaced into 40×40-cm open-field activity-monitoring cham-bers equipped with photobeams (Med Associates, St. Albans,VT) and interfaced to a personal computer. Horizontallocomotor activity was monitored at 100-ms resolution for60 min (habituation period) before administration of saline oracamprosate (as above). After treatment, animals were placedback in the activity monitoring chambers for 20 min. Next, allanimals received an acute injection of cocaine (10 mg/kg i.p.),and locomotor activity was monitored for an additional120 min.

Surgery and catheter maintenance

Before the initiation of self-administration training, ratswere anesthetized with ketamine (66 mg/kg i.p.), xylazine(1.3 mg/kg i.p.), and Equithesin (0.5 ml/kg, consisting of9.72 mg/ml pentobarbital sodium, 42.5 mg/ml chloralhydrate, and 21.3 mg/ml magnesium sulfate heptahydratedissolved in a 44% v/v propylene glycol and 10% v/vethanol solution, Sigma-Aldrich), and silastic catheterswere inserted unilaterally 2.7–3.0 cm into the jugular vein.Catheters were sutured to the underlying muscle tissue with4–0 Ethicon polyglactin absorbable sutures (Johnson andJohnson, Somerville, NJ) through Dacron impregnatedmesh (PlasticsOne, Roanoke, VA). The other end of thecatheter was externalized through a 3-mm dermal biopsyhole made between the scapulae. Catheters consisted of astainless steel cannula (PlasticsOne) attached to silastictubing (Fisher Scientific, Santa Clara, CA) and Dacron

398 Psychopharmacology (2007) 195:397–406

impregnated mesh via dental acrylic (Lang Dental, Wheeling,IL). Dorsal external incisions were closed with 4–0 Ethiconnonabsorbable silk sutures (Johnson and Johnson) and ventralexternal incisions closed with wound clips. Catheters wereflushed daily with cefazolin (0.2 ml of 0.1 g/ml, Sandoz,Holzkirchen, Germany) in heparinized saline (0.2 ml of100 IU, Elkins-Sinn, Cherry Hill, NJ) to help maintain catheterpatency and to protect against infection. Seven days aftersurgery, rats were placed on food restriction for 2 days andthen maintained at 90% normal body weight. This regimenaids in acquisition and maintenance of operant respondingwhile allowing weight gain. Catheter patency was verified byinfusion of methohexital sodium (to effect up to 5 mg/kg i.v.,King Pharmaceuticals) after self-administration procedures.

Self-administration training

Animals were placed in standard operant self-administrationchambers (Coulborn Instruments, Allentown, PA for basalself-administration studies; Med Associates for reinstate-ment studies) equipped with two stimulus lights above tworetractable levers. Operant self-administration chamberswere housed individually in sound-attenuating cubicles thatalso contained a house light and exhaust fan to mask externalodors and noise. Animals were randomly assigned an activelever that, upon depression, resulted in reinforcer delivery.Assignment of the active lever remained constant throughouttraining and testing for any given animal. Depression of theinactive lever resulted in no programmed consequences.Active lever pressing was acquired during overnight,15-h sessions on a modified fixed ratio-1 (FR1) schedule ofreinforcement where each lever press resulted in delivery of asingle 45-mg food pellet (Testdiet, Richmond, IN), which waspaired with illumination of a stimulus light located above theactive lever and activation of a tone for 5 s. Each reinforcerdelivery was followed by a 20-s timeout period where leverpresses were recorded but did not result in additionalreinforcer delivery. The following day, sessions were reducedto 2 h, and the reinforcer was changed to cocaine (0.6 mg/kgper infusion, 2-s infusion time). Training continued with adlibitum home cage access to food and water until active leverpressing stabilized to less than 15% variation over 3–4 days.

Reinstatement testing

After reaching the above criterion for stability of respond-ing, daily extinction trials began, whereby presses on theactive lever no longer produced any programmed conse-quences. After reaching extinction criteria (<15% of theaverage number of active lever presses made during the last2 days of cocaine self-administration, minimum 14 days),animals were subdivided into three groups balanced forself-administration patterns, each receiving daily injections

of saline or acamprosate (100 or 300 mg/kg i.p.) in thehome cage for 5 days 20 min before additional 2-h extinctionsessions. On the fifth day of treatment, at 20 min after salineor acamprosate administration, half of the animals weregiven a single priming injection of cocaine (10 mg/kg i.p.)and placed immediately into the operant self-administrationchamber for assessment of cocaine-seeking behavior (i.e.,active lever presses) for 2 h. The 10-mg/kg priming dose ofcocaine was chosen based on previous reports that lower orhigher priming doses of cocaine result in lower magnitudesof reinstatement (c.f. Cornish and Kalivas 2000; see alsoSchmidt et al. 2005 for review). During the cocaine-inducedreinstatement session, presses on either lever were recordedbut had no programmed consequences. The remaininganimals was placed into the operant self-administrationchamber for cue-induced reinstatement of cocaine-seekingbehavior for 2 h. For cue-induced reinstatement, leverpresses on the lever that previously delivered cocaineresulted in illumination of the stimulus light and activationof the auditory tone for 5 s, but no cocaine was delivered.After each reinstatement test session, animals underwentdaily 2-h extinction sessions until pretreatment patterns ofextinction responding were re-established, and then 5-daydosing procedures were reinitiated. This sequence wasperformed three times so that all animals received salineand both doses of acamprosate (100 and 300 mg/kg i.p.) ina randomized, counterbalanced within-subjects dose–re-sponse design. This “crossover” design allowed quantifica-tion of reinstatement magnitude during each test session byexamining the behavior of saline-treated animals that weretested in parallel with acamprosate-treated animals.

Self-administration testing

To assess the effects of acamprosate on baseline cocaineself-administration, a separate group of animals was trainedas described above. Once stable baseline responding forcocaine was established, animals were administered salineor acamprosate (100 or 300 mg/kg i.p.) once daily in arandomized between-subjects design 20 min before fivesubsequent daily 2-h self-administration sessions.

Statistical analysis

Data were analyzed using SigmaStat 3.0 software (SystatSoftware, San Jose, CA). Pretreatment body weight valueswere averaged across the last 2 days before the initiation ofdrug treatment. Locomotor activity (in 10-min time bins)and body weight were analyzed by a mixed two-wayanalysis of variance (ANOVA, treatment day×treatmentgroup). Locomotor activity (total distance traveled) wasanalyzed by one-way ANOVA. Cocaine intake (mg/kg)was averaged across the last 2 days of baseline self-

Psychopharmacology (2007) 195:397–406 399

administration for each group of animals and was ana-lyzed by one-way ANOVA. Effects of acamprosate oncocaine self-administration were analyzed using a two-way repeated-measures ANOVA with treatment day×leveror number of lever presses or reinforcers delivered asfactors within each treatment group (saline, acamprosate100 mg/kg or acamprosate 300 mg/kg). Effects of acam-prosate on cocaine- or cue-induced reinstatement wereanalyzed by a two-way repeated-measures ANOVA (exper-imental phase [self-administration, extinction, or treatmentbefore reinstatement]×lever) for each reinstatement modal-ity (cue or cocaine). To test for possible order effects onreinstatement, a separate three-way ANOVA was conducted,with reinstatement test×lever×acamprosate dose as factors.All ANOVAs were followed by Holm–Sidak post-hoc mul-tiple comparison procedures. For self-administration and ex-tinction phases, data values were calculated as the averageof active or inactive lever presses during the last 2 daysof active cocaine self-administration or extinction-trainingprocedures, respectively. All data are stated or depicted asmean±SEM.

Results

Effects of acamprosate on locomotor activity and bodyweight

A total of 32 rats (n=8 per group) were tested for effects ofsaline or acamprosate (100, 300, or 500 mg/kg i.p.) on bodyweight and spontaneous and cocaine-stimulated horizontallocomotor activity. These data are presented in Fig. 1. Oneanimal treated with 500 mg/kg acamprosate died after thesecond day of treatment, and the corresponding data wereexcluded from analysis. For the remaining animals, therewas an overall significant effect of treatment group on bodyweight (F[3, 137]=4.46, p<0.05) and a treatment group×treatment day interaction (F[15, 137]=3.10, p<0.001) butno effect of treatment day. Post-hoc analysis revealed thatanimals treated with the 500-mg/kg dose of acamprosateexhibited reduced body weight on days 3, 4, and 5 oftreatment as compared with saline-treated animals (Fig. 1a).Treatment with the lower doses of acamprosate (100 or300 mg/kg) had no effect on body weight (p>0.05).

The effects of acamprosate on spontaneous and cocaine-stimulated locomotor activity are shown in Fig. 1b,c. Anal-ysis of locomotor activity in 10-min time bins (Fig. 1b)revealed a significant effect of treatment group (F[3, 513]=3.151, p<0.05), time (F[19, 513]=40.89, p<0.001), and atreatment group×time interaction (F[57, 513]=1.37, p<0.05). Animals that had been pretreated with 100 mg/kgacamprosate for 4 days showed normal patterns ofhabituation to the novel locomotor activity testing apparatus

during the 60-min habituation period as compared withsaline-pretreated animals. Animals pretreated with the300-mg/kg dose of acamprosate showed reduced locomotoractivity at one time point (40 min) during the habituationperiod as compared with saline-pretreated animals. Animalspretreated with the 500-mg/kg dose of acamprosate showedreduced locomotor activity at two time points (30 and40 min) as compared with saline-pretreated animals. Ananalysis of the overall total distance traveled during the60-min habituation period revealed a significant effect oftreatment group (F[3, 27]=4.77, p<0.01), with animalspretreated with 500 mg/kg acamprosate showing reducedoverall locomotor activity (Fig. 1c).

After administration of saline or acamprosate (according totheir respective treatment groups), animals pretreated with100 mg/kg showed no differences in locomotor activityduring the 20 min after drug treatment as compared tosaline-pretreated animals. Rats treated with 300 mg/kgacamprosate showed reduced locomotor activity during thefirst 10 min after acamprosate treatment relative to saline-treated animals, and rats treated with 500 mg/kg acamprosateshowed reduced locomotor activity during the 20-min periodafter acamprosate treatment relative to saline-treated animals.An analysis of the overall total distance traveled during the20-min postacamprosate period revealed a significant effect oftreatment group (F[3, 27]=7.41, p<0.001), with animalstreated with 300 or 500 mg/kg acamprosate showing reducedoverall horizontal locomotor activity (Fig. 1c).

All animals demonstrated an increase in locomotoractivity during the first 50 min after acute injection ofcocaine (10 mg/kg) as compared to their respective levelsimmediately before cocaine administration. Animals thathad been pretreated with 100 mg/kg acamprosate exhibitedan enhanced locomotor stimulant effect of cocaine during thefirst 10 min after cocaine administration as compared withsaline-treated animals. Animals treated with the 500-mg/kgdose of acamprosate exhibited attenuated locomotor activityduring the 60–80-min time points after cocaine administra-tion. However, an analysis of the overall total distance trav-eled during the 120-min postcocaine period revealed nosignificant group differences (Fig. 1c).

Because we observed adverse effects of the high(500 mg/kg) dose of acamprosate (i.e., reductions in bodyweight and spontaneous locomotor activity as well as onecase of mortality) and because previous reports have shownthat doses of 400 and 450 mg/kg of acamprosate inhibitgeneral behavioral output and reduce body weight in otherrat strains and experimental settings (Czachowski et al.2001; Escher and Mittleman 2006; Gewiss et al. 1991;Heyser et al. 1998; Hölter et al. 1997; Le Magnen et al.1987), we elected not to test any doses of acamprosatehigher than 300 mg/kg on cocaine self-administration andreinstatement of cocaine-seeking behavior.

400 Psychopharmacology (2007) 195:397–406

Effects of acamprosate on cocaine and cue-inducedreinstatement of cocaine-seeking behavior

A total of 24 animals were implanted with indwellingjugular vein catheters for reinstatement experiments. Ofthese, one animal died because of cocaine overdose duringself-administration training, three animals were removedfrom the study because of failure to obtain stable patterns ofcocaine self-administration, and two animals were removedbecause of catheter patency issues. Data from animals thatunderwent cocaine- or cue-induced reinstatement proce-dures (n=9 per group) are shown in Fig. 2. Cocaine intake

was similar between both groups during the last 2 days of themaintenance phase of self-administration (18.0±1.2 mg/kgfor the cocaine-induced reinstatement group, 20.0±1.6 mg/kgfor the cue-induced reinstatement group). A significantinteraction was found between experimental phase (self-administration, extinction, or treatment before reinstatement)and lever for the cocaine-induced reinstatement group (F[4,28]=4.44, p<0.01) and for the cue-induced reinstatementgroup (F[4, 28]=7.02, p<0.001). Post-hoc multiple com-parisons revealed that in both groups, extinction proceduresyielded significant reductions in presses on the active (i.e.,previously cocaine-paired) lever but not the inactive lever

Fig. 1 Effects of acamprosateon a body weight and b, cspontaneous and cocaine-stimulated locomotor activity.a Pretreatment (PreTx) bodyweight values were calculated asan average across the last 2 daysbefore the commencement ofrepeated treatment with salinei.p. (open circles, n=8) oracamprosate 100 mg/kg i.p.(filled circles, n=8), 300 mg/kgi.p. (filled squares, n=8), or500 mg/kg i.p. (filled triangles,n=7). Asterisk, p<0.05 vssaline-treated animals on thesame treatment day. b Locomo-tor activity testing (plotted in10-min time bins) on day 5 ofsaline or acamprosate treatment.Saline (Sal) or acamprosate(Acamp) was administered at thefirst arrow according to therespective treatment group. Allrats subsequently receivedcocaine (10 mg/kg i.p.) at thesecond arrow. Number sign,p<0.05 for animals treated withacamprosate 300 mg/kg vs sa-line at same time point. Asterisk,p<0.05 for animals treated withacamprosate 500 mg/kg vs sa-line at same time point. Plussign, p<0.05 for all groups vstime point immediately beforecocaine. c Analysis of totaldistance traveled during the60-min habituation period,20-min postacamprosate period,or 120-min postcocaine period.Asterisk, p<0.05 vs saline-treated animals. All data arepresented as mean±SEM

Psychopharmacology (2007) 195:397–406 401

vs the respective values observed during the maintenancephase of self-administration (p<0.05). During cocaine- andcue-induced reinstatement testing, a significant increase inactive but not inactive lever responding was observed inanimals pretreated with saline or the low (100 mg/kg) doseof acamprosate (p<0.05 vs extinction for both groups).However, in animals pretreated with the high (300 mg/kg)dose of acamprosate, a significant increase in active leverpresses was not observed in either group (vs extinction forboth groups). Inactive lever presses were not altered byacamprosate treatment in either the cocaine- or cue-inducedreinstatement group. No ordering effects were observedacross reinstatement tests for the cocaine-induced reinstate-ment group or the cue-induced reinstatement group. Treat-ment with saline or acamprosate during the 4 days beforereinstatement testing in either group did not result in changesin the number of active or inactive lever presses on any ofthe treatment days (data not shown).

Effects of acamprosate on cocaine self-administration

A total of 38 animals were implanted with indwellingjugular vein catheters for examination of the effects ofacamprosate on cocaine self-administration. Of theseanimals, five were removed from the study because offailure to obtain stable patterns cocaine self-administration,and three were removed because of catheter patency issues.Cocaine intake was similar across all groups on the last2 days before initiation of saline or acamprosate treatment(12.2±1.8 mg/kg for the saline group; 14.3±1.9 mg/kg forthe acamprosate 100 mg/kg group; 14.4±1.7 mg/kg for theacamprosate 300 mg/kg group). As seen in Fig. 3, there wasno effect of saline or the 100- or 300-mg/kg doses ofacamprosate (n=10 per treatment group) on active cocaineself-administration on any of the treatment days after ad-ministration. Post-hoc comparisons revealed a slight trendtoward an effect of increased cocaine self-administration inanimals treated with 100 mg/kg acamprosate on day 3 ofadministration; however, this effect failed to meet statisticalsignificance (p=0.08). A significant effect of lever wasobserved within each treatment group (F[2, 90]=31.81,p<0.001 for animals treated with saline, F[2, 90]=25.92,p<0.001 for animals treated with acamprosate 100 mg/kg,and F[2, 90]=25.71, p<0.001 for animals treated withacamprosate 300 mg/kg), reflecting the greater respondingon the cocaine (active) lever as compared to the inactive lever.

Discussion

Our data reveal that acamprosate attenuates cocaine- andcue-induced reinstatement of cocaine-seeking behavior.These observations are consistent with our previous reportthat acamprosate also attenuates cocaine-primed reinstate-ment of a cocaine-conditioned place preference in mice(Mcgeehan and Olive 2006). Taken together, these datasuggest that acamprosate may provide some benefit inpreventing relapse in human cocaine addicts. The ability ofacamprosate to attenuate reinstatement of cocaine-seekingbehavior may be relatively specific to cocaine, as otherinvestigators have demonstrated that acamprosate does notalter heroin or stress-induced reinstatement of heroin-seeking behavior (Spanagel et al. 1998) and does not altersucrose reinforcement (Czachowski et al. 2001).

Repeated dosing of acamprosate was utilized because ithas been previously reported that at least 5 days oftreatment with drug is necessary to obtain stable plasmalevels (Durbin et al. 1996; Wilde and Wagstaff 1997),which may be attributable to poor bioavailability. Weobserved that only the 300-mg/kg dose of acamprosatewas effective in attenuating cocaine- and cue-induced

Fig. 2 Acamprosate dose-dependently attenuated a cocaine- or b cue-induced reinstatement of cocaine-seeking behavior (n=9 per group).Baseline self-administration (SA) represents the number of active orinactive lever presses during the last 2 days of self-administrationbefore extinction. Extinction (Ext) represents the number of active orinactive lever presses during the last 2 days of extinction training. Alldata are presented as mean±SEM. Number sign, p<0.05 vs activelever presses during SA. Asterisk, p<0.05 vs active lever pressesduring Ext

402 Psychopharmacology (2007) 195:397–406

reinstatement of cocaine-seeking behavior. While statisticalanalyses revealed that active lever presses in animals treatedwith this dose were not significantly different from valuesafter extinction training, a nonsignificant trend toward anincrease in active lever presses during reinstatement wasevident in rats treated with this dose (p=0.12 for both groups).

However, doses of acamprosate higher than 300 mg/kgwere not tested in reinstatement experiments because weobserved that repeated administration of a 500-mg/kg doseproduced adverse side effects including reduced body weightand hypolocomotion. In addition, doses of acamprosate of400–450 mg/kg have been previously reported to producereductions in general fluid consumption and body weight(Czachowski et al. 2001; Escher and Mittleman 2006;Gewiss et al. 1991; Heyser et al. 1998; Hölter et al. 1997;Le Magnen et al. 1987).

During open-field locomotor experiments, we noted thatdistance traveled was reduced by 300 mg/kg acamprosate,although no overall effects of pretreatment with this dose wereobserved during the preceding 60-min habituation period.Nonetheless, several lines of evidence suggest that thisreduction in motor activity following acute re-exposure to300 mg/kg acamprosate does not underlie the attenuation ofcocaine- and/or cue-induced reinstatement of cocaine-seekingbehavior. First, acamprosate treatment at any dose tested didnot reduce locomotor activity after an acute challenge withcocaine (10 mg/kg i.p.), which was the same dose utilized forcocaine-primed reinstatement. Second, no effects of acam-prosate were observed on inactive lever presses during eitherreinstatement test. Moreover, no effects of acamprosate wereobserved on either active or inactive lever pressing or duringongoing cocaine self-administration. Finally, we have pre-viously demonstrated, in mice, that 300 mg/kg i.p. acam-prosate is devoid of aversive effects that might cause malaiseor nonspecific effects on behavior (Mcgeehan and Olive2003).

The precise mechanism by which acamprosate attenuatedcocaine- and cue-induced reinstatement of cocaine-seekingbehavior remains unknown, primarily because of the complexpharmacological profile of acamprosate. For example, acam-prosate interferes with polyamine regulation of NMDAreceptor function in a bimodal fashion (Berton et al. 1998;Madamba et al. 1996; Popp and Lovinger 2000; Rammes etal. 2001; Zeise et al. 1993; Zeise et al. 1990) that isdependent upon resting NMDA receptor activity (De Witteet al. 2005; Popp and Lovinger 2000). It is interesting to notethat the NMDA receptor antagonist dizocilpine (MK-801)alters active cocaine self-administration and reinstatementinduced by cues but not cocaine-primed reinstatement ofdrug seeking (Backstrom and Hyytia 2006; Hyytia et al.1999; Lee et al. 2005; Park et al. 2002). Taken together, it isunlikely that the attenuation of both cocaine- and cue-induced reinstatement without altering active cocaine self-administration, as reported here, is simply a result of NMDAreceptor antagonism.

Recent studies have demonstrated that acamprosate inhibitsthe neurotoxic effects of (+-)-1-amino-1,3-cyclopentane-trans-dicarboxylic acid, a group I and II mGluR agonist, in

Fig. 3 Acamprosate had no effect on active cocaine self-administration.After establishment of stable baseline responding, rats were adminis-tered saline i.p. (a, n=10), acamprosate 100 mg/kg i.p. (b, n=10), oracamprosate 300 mg/kg i.p. (c, n=10) 20 min before 2-h cocaine self-administration sessions once daily for 5 consecutive days. Active leverpresses=black bars, number of cocaine infusions=hatched bars,inactive lever presses=white bars. p>0.05 for all variables comparedto pretreatment (PreTx) values. All data are presented as mean±SEM

Psychopharmacology (2007) 195:397–406 403

a manner similar to the mGluR5 antagonist SIB-1893 (Harriset al. 2002, 2003). These data suggest that acamprosate canantagonize metabotropic glutamate receptors as well asionotropic glutamate receptors. Indeed, the mGlu5 antagonist2-methyl-6-(phenylethynyl)-pyridine attenuates bothcocaine- and cue-induced reinstatement of cocaine-seekingbehavior (Backstrom and Hyytia 2006; Iso et al. 2006; Lee etal. 2005). However, mGlu5 antagonists also attenuate baselinecocaine self-administration patterns (Chiamulera et al. 2001;Kenny et al. 2003, 2005; Tessari et al. 2004), an effect thatwas not observed with acamprosate in the present study. Thus,it is uncertain whether acamprosate attenuates cocaine- andcue-induced reinstatement by acting at mGlu5 receptors.

Nonetheless, the importance of glutamatergic signalingin the ventral forebrain during cocaine-induced reinstate-ment of drug-seeking behavior has been elegantly studied.Basal extracellular glutamate levels are reduced during long-term withdrawal from repeated cocaine exposure, butcocaine-stimulated glutamate release is significantly elevatedabove that of acutely treated subjects in the nucleusaccumbens (NAc) core (Bell et al. 2000; Hotsenpiller et al.2001; Kalivas et al. 2003; Keys et al. 1998). This dichotomyin ventral forebrain glutamate transmission appears to be acritical mediator of relapse to cocaine seeking, as restorationof basal extracellular glutamate levels prevents cocaine-primed reinstatement of cocaine seeking by inhibitingadditional cocaine-induced glutamate release (Baker et al.2003; Moran et al. 2005). Moreover, stimulation of NMDAor AMPA glutamate receptors in the NAc facilitatesreinstatement of cocaine-seeking (Cornish et al. 1999) whilelocal blockade of these receptors reduces cocaine-inducedreinstatement of cocaine-seeking behavior (Cornish et al.1999; Cornish and Kalivas 2000; McFarland et al. 2003;Park et al. 2002).

Intriguingly, acamprosate has been shown to normalizeelevated extracellular levels of glutamate in the NAc duringethanol withdrawal (Dahchour and De Witte 2000, 2003;Dahchour et al. 1998). However, as mentioned above,withdrawal from cocaine is associated with reduced, ratherthan elevated, extracellular levels of glutamate in the NAc(Baker et al. 2003; Bell et al. 2000; Hotsenpiller et al. 2001;Kalivas et al. 2003; Moran et al. 2005). Thus, additionalmicrodialysis studies are needed to determine whetheracamprosate can “normalize” extracellular levels of gluta-mate in the NAc when basal levels of this neurotransmitterare suppressed (i.e., during cocaine withdrawal), rather thanelevated (i.e., during alcohol withdrawal).

We also observed that acamprosate elicited no effect onresponding for cocaine in actively self-administeringanimals. This lack of effect was unanticipated because wehave previously observed that acamprosate (at similardoses) attenuates the development of a conditioned placepreference to cocaine in mice (Mcgeehan and Olive 2003).

This apparent discrepancy underscores important differ-ences in the effects of certain therapeutic agents on theconditioned rewarding effects of drugs of abuse, asmeasured in a place preference paradigm, as compared toconsummatory or reinforcing aspects that are measured inan operant self-administration paradigm (Bardo andBevins 2000; Calcagnetti et al. 1995; Tzschentke 1998).Accordingly, similar differential effects of acamprosate onrelapse-like behavior, as compared to baseline drug self-administration, have also been reported in the alcoholliterature (Heyser et al. 1998). It is possible, however, thatacamprosate might alter active cocaine self-administrationunder other unit doses of reinforcement, higher responserequirements, or in a progressive ratio paradigm. Alongthese lines, acamprosate has previously been demonstratedto reduce alcohol self-administration under higher responserequirements as opposed to lower (i.e., FR1) responserequirements (Heyser et al. 1998; Spanagel and Hölter2000). Thus, additional studies are needed to determine ifacamprosate reduces cocaine self-administration underconditions of higher response requirements.

With regards to the ability of acamprosate to attenuateongoing cocaine self-administration, it is interesting to notethat a Phase II pilot clinical trial examining the efficacy ofacamprosate in reducing cocaine craving, withdrawal symp-toms, and daily cocaine use was recently initiated (www.clinicaltrials.gov identifier: NCT00385268). The results of thisclinical study will have important implications in determiningthe predictive validity of current preclinical animal models(i.e., IV self-administration and reinstatement procedures) forpredicting the efficacy of potential therapeutic compounds forthe treatment of cocaine addiction.

In summary, we have shown in a rat model of relapsethat acamprosate attenuated both cue- and drug-inducedreinstatement of cocaine-seeking behavior at doses that donot affect ongoing cocaine self-administration or cocaine-stimulated locomotion. Because of a broad mechanism ofaction, further investigation with acamprosate may uncovermultiple receptor systems that contribute to relapse.Regardless of its many potential neurochemical targets,acamprosate lacks abuse potential and presents a mild sideeffect profile in humans (Grant and Woolverton 1989; LeMagnen et al. 1987; Mcgeehan and Olive 2003; Naassila etal. 1998; Nalpas et al. 1990; Schneider et al. 1998;Spanagel et al. 1996). While further preclinical studies areneeded to determine if acamprosate attenuates cocainerelapse induced by other stimuli such as stress, our datasuggest that acamprosate may be a novel pharmacologicaltherapeutic particularly useful in reducing the incidence ofrelapse in cocaine addiction.

Acknowledgments This work was supported by the Neurobiology ofAddiction Research Center (DA015369) and funds provided by the

404 Psychopharmacology (2007) 195:397–406

State of California for medical research on alcohol and substance abusethrough the University of California at San Francisco. The authors wishto thank Shannon Ghee and Brian Wheeler for technical assistance.

References

Backstrom P, Hyytia P (2006) Ionotropic and metabotropic glutamatereceptor antagonism attenuates cue-induced cocaine seeking.Neuropsychopharmacology 31:778–786

Backstrom P, Hyytia P (2007) Involvement of AMPA/kainate,NMDA, and mGlu5 receptors in the nucleus accumbens core incue-induced reinstatement of cocaine seeking in rats. Psycho-pharmacology 192:571–580

Baker DA, McFarland K, Lake RW, Shen H, Tang XC, Toda S,Kalivas PW (2003) Neuroadaptations in cystine-glutamateexchange underlie cocaine relapse. Nat Neurosci 6:743–749

Bardo MT, Bevins RA (2000) Conditioned place preference: whatdoes it add to our preclinical understanding of drug reward?Psychopharmacology 153:31–43

Bell K, Duffy P, Kalivas PW (2000) Context-specific enhancement ofglutamate transmission by cocaine. Neuropsychopharmacology23:335–344

Berton F, Francesconi WG, Madamba SG, Zieglgänsberger W, SigginsGR (1998) Acamprosate enhances N-methyl-D-aspartate recep-tor-mediated neurotransmission but inhibits presynaptic GABABreceptors in nucleus accumbens neurons. Alcohol Clin Exp Res22:183–191

Calcagnetti DJ, Keck BJ, Quatrella LA, Schechter MD (1995)Blockade of cocaine-induced conditioned place preference:relevance to cocaine abuse therapeutics. Life Sci 56:475–483

Chiamulera C, Epping-Jordan MP, Zocchi A, Marcon C, Cottiny C,Tacconi S, Corsi M, Orzi F, Conquet F (2001) Reinforcing andlocomotor stimulant effects of cocaine are absent in mGluR5 nullmutant mice. Nat Neurosci 4:873–874

Cornish JL, Kalivas PW (2000) Glutamate transmission in the nucleusaccumbens mediates relapse in cocaine addiction. J Neurosci 20:RC89 (1–5)

Cornish JL, Duffy P, Kalivas PW (1999) A role for nucleusaccumbens glutamate transmission in the relapse to cocaine-seeking behavior. Neuroscience 92:1359–1367

Czachowski CL, Legg BH, Samson HH (2001) Effects of acamprosateon ethanol-seeking and self-administration in the rat. AlcoholClin Exp Res 25:344–350

Dahchour A, De Witte P (2000) Ethanol and amino acids in the centralnervous system: assessment of the pharmacological actions ofacamprosate. Prog Neurobiol 60:343–362

Dahchour A, De Witte P (2003) Effects of acamprosate on excitatoryamino acids during multiple ethanol withdrawal periods. AlcoholClin Exp Res 27:465–470

Dahchour A, De Witte P, Bolo N, Nédélec JF, Muzet M, Durbin P,Macher JP (1998) Central effects of acamprosate: part 1.Acamprosate blocks the glutamate increase in the nucleusaccumbens microdialysate in ethanol withdrawn rats. PsychiatryRes 82:107–114

De Witte P, Littleton J, Parot P, Koob G (2005) Neuroprotective andabstinence-promoting effects of acamprosate: elucidating themechanism of action. CNS Drugs 19:517–37

Durbin P, Hulot T, Chabac S (1996) Pharmacodynamics andpharmacokinetics of acamprosate: an overview. In: Soyka M(ed) Acamprosate in relapse prevention of alcoholism. Springer,Berlin, pp 47–64

Escher T, Mittleman G (2006) Schedule-induced alcohol drinking:non-selective effects of acamprosate and naltrexone. Addict Biol11:55–63

Famous KR, Schmidt HD, Pierce RC (2007) When administered intothe nucleus accumbens core or shell, the NMDA receptorantagonist AP-5 reinstates cocaine-seeking behavior in the rat.Neurosci Lett 420:169–173

Gewiss M, Heidbreder C, Opsomer L, Durbin P, De Witte P (1991)Acamprosate and diazepam differentially modulate alcohol-induced behavioural and cortical alterations in rats followingchronic inhalation of ethanol vapour. Alcohol Alcohol 26:129–137

Grant KA, Woolverton WL (1989) Reinforcing and discriminativestimulus effects of Ca-acetyl homotaurine in animals. PharmacolBiochem Behav 32:607–611

Harris BR, Prendergast MA, Gibson DA, Rogers DT, Blanchard JA,Holley RC, Fu MC, Hart SR, Pedigo NW, Littleton JM (2002)Acamprosate inhibits the binding of neurotoxic effects on trans-ACPD, suggesting a novel site of action at metabotropicglutamate receptors. Alcohol Clin Exp Res 26:1779–1793

Harris BR, Gibson DA, Prendergast MA, Blanchard JA, Holley RC,Hart SR, Scotland RL, Foster TC, Pedigo NW, Littleton JM(2003) The neurotoxicity induced by ethanol withdrawal inmature organotypic hippocampal slices might involve cross-talkbetween metabotropic glutamate type 5 receptors and N-methyl-D-aspartate receptors. Alcohol Clin Exp Res 27:1724–1735

Heilig M, Egli M (2006) Pharmacological treatment of alcoholdependence: target symptoms and target mechanisms. PharmacolTher 111:855–876

Heyser CJ, Schulteis G, Durbin P, Koob GF (1998) Chronicacamprosate eliminates the alcohol deprivation effect whilehaving limited effects on baseline responding for ethanol in rats.Neuropsychopharmacology 18:125–133

Hotsenpiller G, Giorgetti M, Wolf ME (2001) Alterations in behaviourand glutamate transmission following presentation of stimulipreviously associated with cocaine exposure. Eur J Neurosci14:1843–1855

Hölter SM, Landgraf R, Zieglgansberger W, Spanagel R (1997) Timecourse of acamprosate action on operant ethanol self-administrationafter ethanol deprivation. Alcohol Clin Exp Res 21:862–868

Hyytia P, Backstrom P, Liljequist S (1999) Site-specific NMDAreceptor antagonists produce differential effects on cocaine self-administration in rats. Eur J Pharmacol 378:9–16

IsoY, Grajkowska E,Wroblewski JT, Davis J, Goeders NE, JohnsonKM,Sanker S, Roth BL, Tueckmantel W, Kozikowski AP (2006)Synthesis and structure-activity relationships of 3-[(2-methyl-1,3-thiazol-4-yl)ethynyl]pyridine analogues as potent, noncompetitivemetabotropic glutamate receptor subtype 5 antagonists; search forcocaine medications. J Med Chem 49:1080–1100

Kalivas PW, McFarland K, Bowers S, Szumlinski K, Xi ZX, Baker D(2003) Glutamate transmission and addiction to cocaine. Ann NY Acad Sci 1003:169–175

Kenny PJ, Paterson NE, Boutrel B, Semenova S, Harrison AA,Gasparini F, Koob GF, Skoubis PD, Markou A (2003) Metabo-tropic glutamate 5 receptor antagonist MPEP decreased nicotineand cocaine self-administration but not nicotine and cocaine-induced facilitation of brain reward function in rats. Ann NYAcad Sci 1003:415–418

Kenny PJ, Boutrel B, Gasparini F, Koob GF, Markou A (2005)Metabotropic glutamate 5 receptor blockade may attenuatecocaine self-administration by decreasing brain reward functionin rats. Psychopharmacology 179:247–254

Keys AS, Mark GP, Emre N, Meschul CK (1998) Reduced glutamateimmunolabeling in the nucleus accumbens following extendedwithdrawal from self-administered cocaine. Synapse 30:393–401

Lee B, Platt DM, Rowlett JK, Adewale AS, Spealman RD (2005)Attenuation of behavioral effects of cocaine by the metabotropicglutamate receptor 5 antagonist 2-methyl-6-(phenylethynyl)-pyridine in squirrel monkeys: comparison with dizocilpine. JPharmacol Exp Ther 312:1232–1240

Psychopharmacology (2007) 195:397–406 405

Le Magnen J, Tran G, Durlach J, Martin C (1987) Dose-dependentsuppression of the high alcohol intake of chronically intoxicatedrats by Ca-acetyl homotaurinate. Alcohol 4:97–102

Madamba SG, Schweitzer P, Zieglgänsberger W, Siggins GR (1996)Acamprosate (calcium acetylhomotaurinate) enhances the N-methyl-D-aspartate component of excitatory neurotransmission in rat hippo-campal CA1 neurons in vitro. Alcohol Clin Exp Res 20:651–658

Maldonado C, Rodriguez-Arias M, Castillo A, Aguilar MA, Minarro J(2007) Effect of memantine and CNQX in the acquisition,expression and reinstatement of cocaine-induced conditionedplace preference. Prog Neuropsychopharmacol Biol Psychiatry31:932–939

McFarland K, Lapish CC, Kalivas PW (2003) Prefrontal glutamaterelease into the core of the nucleus accumbens mediates cocaine-induced reinstatement of drug-seeking behavior. J Neurosci23:3531–3537

Mcgeehan AJ, OliveMF (2003) The anti-relapse compound acamprosateinhibits the development of a conditioned place preference toethanol and cocaine but not morphine. Br J Pharmacol 138:9–12

Mcgeehan AJ, Olive MF (2006) Attenuation of cocaine-inducedreinstatement of cocaine conditioned place preference by acam-prosate. Behav Pharmacol 17:363–367

Moran MM, McFarland K, Melendez RI, Kalivas PW, Seamans JK(2005) Cystine/glutamate exchange regulates metabotropic glu-tamate receptor presynaptic inhibition of excitatory transmissionand vulnerability to cocaine seeking. J Neurosci 25:6389–6393

Naassila M, Legrand E, d'Alche-Birée F, Daoust M (1998) Cyame-mazine decreases ethanol intake in rats and convulsions duringethanol withdrawal syndrome in mice. Psychopharmacology140:421–428

Nalpas B, Dabadie H, Parot P, Paccalin J (1990) Acamprosate: frompharmacology to therapeutics. Encephale 16:175–179

Park W-K, Bari AA, Jey AR, Anderson SM, Spealman RD, Rowlett JK,Pierce RC (2002) Cocaine administered into the medial prefrontalcortex reinstates cocaine-seeking behavior by increasing AMPAreceptor-mediated glutamate transmission in the nucleus accu-mbens. J Neurosci 22:2916–2925

Popp RL, Lovinger DM (2000) Interaction of acamprosate withethanol and spermine on NMDA receptors in primary culturedneurons. Eur J Pharmacol 394:221–231

Rammes G, Mahal B, Putzke J, Parsons C, Spielmanns P, Pestel E,Spanagel R, ZieglgansbergerW, Schadrack J (2001) The anti-cravingcompound acamprosate acts as a weak NMDA-receptor antagonist,but modulates NMDA-receptor subunit expression similar tomemantine and MK-801. Neuropharmacology 40:749–760

Rosenthal RN (2006) Current and future drug therapies for alcoholdependence. J Clin Psychopharmacol 26(Suppl 1):S20–S29

Schmidt HD, Anderson SM, Famous KR, Kumaresan V, Pierce RC(2005) Anatomy and pharmacology of cocaine priming-inducedreinstatement of drug seeking. Eur J Pharmacol 526:65–76

Schneider U, Wohlfahrt K, Schulze-Bonhage A, Haacker T, CasparyA, Zedler M, Emrich HM (1998) Lack of psychotomimetic orimpairing effects on psychomotor performance of acamprosate.Pharmacopsychiatry 31:110–113

See RE, Kruzich PJ, Grimm JW (2001) Dopamine, but not glutamate,receptor blockade in the basolateral amygdala attenuates condi-tioned reward in a rat model of relapse to cocaine-seekingbehavior. Psychopharmacology 154:301–310

Soyka M, Roesner S (2006) New pharmacological approaches for thetreatment of alcoholism. Expert Opin Pharmacother 7:2341–2353

Spanagel R, Hölter SM, (2000) Pharmacological validation of a newanimal model of alcoholism. J Neural Transm 107:669–680

Spanagel R, Zieglgansberger W, Hundt W (1996) Acamprosate andalcohol. III. Effects of alcohol discrimination in the rat. Eur JPharmacol 305:51–56

Spanagel R, Sillaber I, Zieglgansberger W, Corrigall WA, Stewart J,Shaham Y (1998) Acamprosate suppresses the expression ofmorphine-induced sensitization in rats but does not affect heroinself-administration or relapse induced by heroin or stress.Psychopharmacology 139:391–401

Spanagel R, Pendyala G, Abarca C, Zghoul T, Sanchis-Segura C,Magnone MC, Lascorz J, Depner M, Holzberg D, Soyka M,Schreiber S, Matsuda F, Lathrop M, Schumann G, Albrecht U(2005) The clock gene Per2 influences the glutamatergic systemand modulates alcohol consumption. Nat Med 11:35–42

Tessari M, Pilla M, Andreoli M, Hutcheson DM, Heidbreder CA (2004)Antagonism at metabotropic glutamate 5 receptors inhibits nicotine-and cocaine-taking behaviours and prevents nicotine-triggeredrelapse to nicotine-seeking. Eur J Pharmacol 499:121–133

Tzschentke TM (1998) Measuring reward with the conditioned placepreference paradigm: a comprehensive review of drug effects,recent progress and new issues. Prog Neurobiol 56:613–672

Wilde MI, Wagstaff AJ (1997) Acamprosate: a review of itspharmacology and clinical potential in the management ofalcohol dependence after detoxification. Drugs 53:1038–1053

Zeise ML, Kasparow S, Capogna M, Zieglgänsberger W (1990)Calciumdiacetylhomotaurinate (CA-AOTA) decreases the actionof excitatory amino acids in the rat neocortex in vitro. Prog ClinBiol Res 351:237–242

Zeise ML, Kasparov S, Capogna M, Zieglgänsberger W (1993)Acamprosate (calciumacetylhomotaurinate) decreases postsynapticpotentials in the rat neocortex: possible involvement of excitatoryamino acid receptors. Eur J Pharmacol 231:47–52

406 Psychopharmacology (2007) 195:397–406