The inverse agonist of CB1 receptor SR141716 blockscompulsive eating of palatable food

Riccardo Dore1, Marta Valenza1,2, Xiaofan Wang1, Kenner C. Rice3, Valentina Sabino1 &Pietro Cottone1

Laboratory of Addictive Disorders, Departments of Pharmacology and Psychiatry, Boston University School of Medicine, Boston, MA, USA1, Department ofBiomedical Sciences and Human Oncology, University of Bari Aldo Moro, Bari, Italy2 and Chemical Biology Research Branch, National Institute on Drug Abuse andNational Institute on Alcohol and Alcoholism, Rockville, MD, USA3

ABSTRACT

Dieting and the increased availability of highly palatable food are considered major contributing factors to the largeincidence of eating disorders and obesity. This study was aimed at investigating the role of the cannabinoid (CB) systemin a novel animal model of compulsive eating, based on a rapid palatable diet cycling protocol. Male Wistar rats werefed either continuously a regular chow diet (Chow/Chow, control group) or intermittently a regular chow diet for 2 daysand a palatable, high-sucrose diet for 1 day (Chow/Palatable). Chow/Palatable rats showed spontaneous and progressivelyincreasing hypophagia and body weight loss when fed the regular chow diet, and excessive food intake and body weightgain when fed the palatable diet. Diet-cycled rats dramatically escalated the intake of the palatable diet during the firsthour of renewed access (7.5-fold compared to controls), and after withdrawal, they showed compulsive eating andheightened risk-taking behavior. The inverse agonist of the CB1 receptor, SR141716 reduced the excessive intake ofpalatable food with higher potency and the body weight with greater efficacy in Chow/Palatable rats, compared tocontrols. Moreover, SR141716 reduced compulsive eating and risk-taking behavior in Chow/Palatable rats. Finally,consistent with the behavioral and pharmacological observations, withdrawal from the palatable diet decreased thegene expression of the enzyme fatty acid amide hydrolase in the ventromedial hypothalamus while increasing that ofCB1 receptors in the dorsal striatum in Chow/Palatable rats, compared to controls. These findings will help understandthe role of the CB system in compulsive eating.

Keywords Compulsive eating, food intake, palatability, rimonabant, risk-taking behavior, SR141716.

Correspondence to: Pietro Cottone, Laboratory of Addictive Disorders, Departments of Pharmacology and Psychiatry, Boston University School ofMedicine, 72 E Concord St, R-618, Boston, MA 02118, USA. E-mail: cottone@bu.edu

INTRODUCTION

The subjective evaluation of foods is an important deter-minant of their acceptability and an important evolu-tionary mechanism promoting the consumption ofenergy-dense foods in scarce environments (Birch 1999).However, given the increased availability of calorie-dense, highly palatable foods in Western societies, thismotivational mechanism is now largely contributing tothe obesity epidemic and eating disorders (Hill et al.2003; Mathes et al. 2009). Interestingly, as a compensa-tory cultural mechanism, many individuals attempt tocontrol body weight by limiting themselves to perceived‘safe’ foods, which are generally less palatable than the‘forbidden’ ones (high-sugar or high-fat), from which

they abstain. Dieting in turn increases craving for palat-able food, promotes overeating and results in repeated,discrete alternations in food palatability across time(Polivy & Herman 1985). This maladaptive pattern offood intake has raised the question whether obesity andeating disorders, like drug addiction, can be conceptual-ized as chronic relapsing conditions with alternatingperiods of abstinence (e.g. dieting) and relapse (compul-sive overeating of palatable foods; APA 2000; Corwin &Grigson 2009; Cottone et al. 2012).

In pre-clinical research, a variety of animal modelshave been designed to mimic the aberrant feeding patternobserved in humans as well as to study the underlyingneurobiological mechanisms (Cottone et al. 2008b,2009a; Cifani et al. 2009; Corwin, Avena & Boggiano

ORIGINAL ARTICLE

bs_bs_bannerAddiction Biologydoi:10.1111/adb.12056

© 2013 Society for the Study of Addiction Addiction Biology

2011; Parylak et al. 2012; Blasio et al. 2013). A success-ful procedure to induce excessive food intake in rodents isto provide limited access to highly palatable foods withoutthe use of food restriction or food deprivation (Corwin &Buda-Levin 2004). Highly rewarding foods, rich in fatsand sugars, are offered ad libitum to experimental subjectsintermittently within limited periods of time [brief:1–2 hours (Corwin 2004; Cottone et al. 2012) orextended: 2 days (Cottone et al. 2009b)], followed by avariable period of exposure (23 hours to 5 days) to aregular chow diet, which is typically undereaten byanimals. Palatable diet withdrawal is critically importantto induce overconsumption upon renewing its access,according to a classical deprivation effect (McBride & Li1998; George et al. 2007; Maier et al. 2010). Ultimately,cycles of palatable food access/withdrawal are essentialto induce escalation and maintenance of excessiveintake.

The first aim of this study was to characterize thebehavioral effects of a novel experimental procedure ofintermittent, but extended access to palatable food, whichconsisted of 2-day access to a regular chow diet followedby 1 day of access to a highly palatable, sugary diet. Giventhe short cumulative length of regular chow and palat-able accesses (3 days), this procedure would have theadvantage of consisting of faster experimental cyclingcompared to other existing intermittent, extended accessprotocols (Cottone et al. 2008a, 2009a). Since the can-nabinoid (CB) system modulates feeding (Matias, Bisogno& Di Marzo 2006; Mechoulam et al. 2006; Kunos 2007;DiPatrizio & Piomelli 2012) and is engaged by palatablediets (DiPatrizio & Simansky 2008; Timofeeva et al.2009; Bello et al. 2012), the second aim of our study wasto investigate whether the consummatory and motiva-tional outcomes of intermittent, extended access tohighly palatable food were CB type-1 (CB1) receptordependent. For this purpose, in this animal model, wetested the effects of the CB1 receptor inverse agonistSR141716 on (1) escalated excessive intake of palatablediet and (2) time spent and the amount of food eaten(modeling risk-taking behavior and compulsive eating,respectively) in an aversive, open compartment, wherethe palatable diet was offered, using a light/dark conflicttest. Finally, the third aim of this study was to determinethe effects of withdrawal from the intermittent, extendedaccess to palatable diet on the gene expression of compo-nents of the endocannabinoid machinery, including theCB1 receptor, and the enzymes involved in the synthesisand degradation of the two main endocannabinoids,anandamide (AEA) and 2-arachidonoylglycerol [N-acylphosphatidylethanolamine phospholipase D (NAPE-PLD), diacylglycerol lipase a (DAGLa), fatty acid amidehydrolase (FAAH) and monoacylglycerol lipase (MAGL)]in brain areas involved in food intake and motivation.

MATERIALS AND METHODS

Subjects

Male Wistar rats, weighing 180–230 g and 41–47 daysold at arrival (Charles River, Wilmington, MA, USA),were single housed in wire-topped, plastic cages(27 ¥ 48 ¥ 20 cm) on a 12-hour reverse light cycle (lightsoff at 11:00 am), in an Association for Assessment andAccreditation of Laboratory Animal Care International-approved humidity- (60%) and temperature-controlled(22°C) vivarium. Rats had access to corn-based chow(Harlan Teklad LM-485 Diet 7012 [65% (kcal) carbohy-drate, 13% fat, 21% protein, metabolizable energy310 cal/100 g; Harlan, Indianapolis, IN, USA]) and waterad libitum at all times unless otherwise specified. Proce-dures adhered to the National Institutes of Health Guide forthe Care and Use of Laboratory Animals (NIH publicationnumber 85-23, revised 1996) and the Principles of Labo-ratory Animal Care and were approved by Boston Univer-sity Medical Campus Institutional Animal Care and UseCommittee.

Drugs

The CB1 receptor inverse agonist SR141716 [rimona-bant or 5-(4-Chlorophenyl)-1-(2,4-dichloro-phenyl)-4-methyl-N-(piperidin-1-yl)-1H-pyrazole-3-carboxamide]HCl was synthesized as reported previously (Kumar et al.2008). SR141716 HCl was solubilized in an 18:1:1mixture of saline : ethanol : cremophor and was admin-istered intraperitoneally (0, 0.3, 1, 3 mg/kg, 1 ml/kg,-30 minutes pre-treatment).

Development of a shortened, ad libitum palatablediet alternation

Figure 1 shows a schematic representation of the inter-mittent, extended access to a palatable diet procedure,which was modified from a previously described experi-mental protocol (Cottone et al. 2008a, 2009a,b; Iemoloet al. 2012). Briefly, after 1 week of acclimation, rats(n = 20) were divided in two groups matched for foodintake, body weight and feed efficiency from a previous3–4 days baseline period. One group was then provided achow diet (‘Chow’) ad libitum (Chow/Chow), whereas asecond group was provided chow ad libitum for 2 daysfollowed by 1 day of ad libitum access to a highly palat-able, chocolate-flavored, high-sucrose diet (‘Palatable’;Chow/Palatable). The chow diet was the above describedHarlan Teklad LM-485 Diet 7012. The palatable dietwas a nutritionally complete, chocolate-flavored, high-sucrose (50% kcal), AIN-76A-based diet that is com-parable in macronutrient proportions and energy densityto the chow diet [chocolate-flavored Formula 5TUL:66.7% (kcal) carbohydrate, 12.7% fat, 20.6% protein,

2 Riccardo Dore et al.

© 2013 Society for the Study of Addiction Addiction Biology

metabolizable energy 344 cal/100 g; TestDiet, Rich-mond, IN, USA; formulated as 45 mg precision foodpellets to increase its preferredness]. For brevity, the first 2days (chow only) and the following 1 day (chow or palat-able according to experimental group) are referred in allexperiments as C and P phases. Diets were never concur-rently available. Food intake and body weight were meas-ured daily. Average food intake was calculated as the kcalintake in a certain phase divided by the number of days ofthat phase (2 for C phase, 1 for P phase). Average bodyweight change was calculated as the difference betweenthe body weight at the end and at the beginning of aphase divided by the number of days of that phase (2 forC phase, 1 for P phase). Average and cumulative feedefficiency was calculated as mg body weight gained in acertain time interval (phase or cycle) divided by the kcalfood intake in the same time interval. Cumulative foodintake and cumulative body weight gain was calculatedas the kcal food intake or body weight gained at the end ofeach cycle since the beginning of the study. To evaluatethe escalation of palatable diet intake, food intake wasmeasured 1 hour after diet switch at the beginning ofeach P phase. For the time course studies, food intake wasmeasured at the beginning of either the C or the P phase(at the time of the diet switch), and then 1 hour, 3 hours,6 hours and 24 hours following the switch of the diet. Aspreviously published (Cottone et al. 2008a,b, 2009b), the5TUL Chocolate Diet (sugary palatable diet) was uni-formly preferred compared to the Harlan LM-485 chowdiet.

Effects of SR141716 on intake of palatable food andbody weight

SR141716 (0, 0.3, 1, 3 mg/kg, i.p., -30 minutes) wasadministered at the onset of P phase in a within-subjectLatin square design (n = 11/group). Rats were then pro-vided with a pre-weighed amount of food and intake wasrecorded 1 hour, 3 hours, 6 hours and 24 hours later.Body weights were recorded right before and 24 hoursfollowing drug administration.

Effects of SR141716 on risk-taking behavior andcompulsive eating

The task was performed as previously described(Teegarden & Bale 2007; Cottone et al. 2012). Rats(n = 44) were tested for 10 minutes in a light/dark rec-tangular box (50 ¥ 100 ¥ 35 cm) in which the aversivelight compartment (50 ¥ 70 ¥ 35 cm) was illuminatedby a 60 lux light. The dark side (50 ¥ 30 ¥ 35 cm) had anopaque cover and ~0 lux of light. A shallow, metal cupcontaining a pre-weighed amount of the same foodreceived during the ad libitum diet alternation (HarlanTeklad LM-485 Diet 7012 for Chow/Chow rats or 45-mgchocolate pellets for Chow/Palatable rats) was positionedin the center of the light compartment. The two compart-ments were connected by an open doorway, whichallowed the subjects to move freely between the two.White noise was present. On the test day (beginning of Pphase), following the 2 days (C phase) of withdrawal frompalatable food, rats were pre-treated with SR141716 (0and 1 mg/kg, i.p.) 30 minutes before being placed intothe light compartment, facing both the food cup and thedoorway. The time spent in the open compartment andthe amount of food eaten during the test were measured.The two dependent variables were then used to opera-tionalize the constructs of ‘risk-taking behavior’ and‘compulsive-like eating’, respectively. Because of rats’innate fear for bright, aversive environments, the timespent exploring the light compartment of the light/darkbox under normal, control conditions is minimal. Anincreased time spent in this compartment, as comparedto control conditions, resulting from the presence of thehighly palatable diet, was operationalized as ‘risk-takingbehavior’ (Teegarden & Bale 2007; Cottone et al. 2012).Moreover, under normal, control conditions, eatingbehavior is typically suppressed when a rat faces adversecircumstances; a significant increase in food intake inspite of the adverse conditions, as compared to controlconditions, was operationalized as a construct of‘compulsive-like eating’ (Heyne et al. 2009; Johnson &Kenny 2010; Cottone et al. 2012). Water was not

Figure 1 Schematic representation of theintermittent, extended access to a palatablediet procedure used in this study. Briefly,subjects were divided in two groups: a firstgroup received a regular chow diet ad libitumdaily (Chow/Chow), and a second groupreceived 2 days of chow diet followed by aday of a highly palatable, chocolate flavored,high-sugary, diet (Chow/Palatable). Forbrevity, the first 2 days (chow only) and lastday (chow or palatable per diet condition) ofeach cycle are referred to as C and P phases

CB system and compulsive eating 3

© 2013 Society for the Study of Addiction Addiction Biology

available during the 10-minute test. At the end of thetest, rats were removed from the light/dark box appara-tus, returned to their home cages, and food was provided(chow or palatable diet, according to the experimentalgroup).

Quantitative real-time polymerase chainreaction (RT-PCR)

The quantitative RT-PCR was performed as shownpreviously (Sabino et al. 2011; Cottone et al. 2012).Chow/Chow and Chow/Palatable rats (n = 7–12) wereanesthetized and decapitated at the end of the C phase,before the putative renewed access to the highly palatablediet (same time point of pharmacological studies). Brainswere quickly removed and coronally sliced in a brainmatrix. Punches containing dorsal striatum (DS),nucleus accumbens (NAcc), lateral hypothalamus (LH)and ventromedial hypothalamus (VMH) were collectedon an ice-cold stage. Total RNA was prepared from eachbrain punch using RNeasy mini kit (Qiagen, Valencia,CA, USA) following the standard protocol for RNA extrac-tion from animal tissues. Total RNA was prepared fromeach punch using the RNeasy lipid mini kit (Qiagen) asrecommended for animal tissue. Total RNA was quanti-fied by Nanodrop 1000 (Thermo Scientific, Wilmington,DE, USA), and then reverse transcribed with QuantiTectReverse Transcription Kit (Qiagen), which includes aDNA removal step. For quantitative RT-PCR, Roche LightCycler 480 Master-plus SYBR Green mix (Roche AppliedScience, Indianapolis, IN, USA) was used. Reactions werecarried out in a 10 ml volume using a 96-well plate Real-plex2 machine (Eppendorf, Hauppauge, NY, USA). Theprimers (0.5 mM final concentration) for CB1 receptor,NAPE-PLD, DAGLa, FAAH, MAGL and the housekeeping

gene cyclophilin A (CypA) were synthesized by Sig-maAldrich (St. Louis, MO, USA) with a standard desaltingpurification method. Fragments were amplified using athree-step RT-PCR protocol that included an initial 94°Cfor 10 minutes step to activate Taq polymerase, followedby 40 cycles of amplification. The sequences of theprimers and the conditions are summarized in Table 1.Standard curves were constructed using purified andsequenced fragments. The results were analyzed by thesecond derivative methods were expressed in arbitraryunits normalized by the expression levels of the referencegene, CypA. The standard curve and the samples wereanalyzed in duplicate. Gene-specific amplification wasdetermined by melting curve analysis as one peak at theexpected melting temperature and by agarose gel electro-phoresis.

Statistical analysis

Average daily food intake, average daily body weightchange and average daily feed efficiency were analyzedusing three-way mixed analyses of variance (ANOVAs)with Diet Schedule as a between-subjects factor, and Cycleand Phase as within-subject factors. First hour of palat-able food consumption, cumulative food intake, cumula-tive body weight change and cumulative feed efficiencywere analyzed using two-way ANOVAs with Diet Scheduleas a between-subjects factor and Phase as a within-subjectfactor. Time course of food intake in C and P phase wasanalyzed using two-way ANOVAs with Diet Schedule as abetween-subjects factor and Time as a within-subjectfactor. To analyze the time course of ingestion followingSR141716 treatment, three-way ANOVAs on incremen-tal food intake were performed, with Diet Schedule as abetween-subjects factor and Treatment and Time as

Table 1 Primers and conditions used in qPCR for the detection of mRNA levels of cannabinoid receptor type 1, N-acyl-phosphatidylethanolamine phospholipase D, diacylglycerol lipase a, fatty acid amide hydrolase, monoacylglycerol lipase and thehousekeeping gene cyclophilin A in punches obtained from the brains of Chow/Chow and Chow/Palatable rats.

Target geneRGDsymbol Primers

Temperatures(°C)

Times(seconds)

Cyclophilin A (CypA) Ppia Primer A: TAT CTG CAC TGC CAA GAC TGA GTG 95–58–72 20–15–10Primer B: CTT CTT GCT GGT CTT GCC ATT CC

Cannabinoid receptor type 1 (CB1) Cnr1 Primer A: TGG GCA CCT TCA CGG TTC TG 95–60–72 15–15–20Primer B: GGA AGG CCT GCA TCG GAG ACT

Fatty acid amide hydrolase (FAAH) Faah Primer A: CAC TGT GAC CGC CGA GGA CGA TG 95–60–72 15–15–20Primer B: CAG CTG TTC CAC CTC CCG CAT GAA

Monoacylglycerol lipase (MAGL) Mgll Primer A: ACC TGG TCA ATG CGG ATG GAC 95–58.7–72 15–10–10Primer B: GTC ATA ACG GCC ACA GTG TTC CC

Diacylglycerol lipase a (DAGLa) Dagla Primer A: GCT CTT TGC CCT CGC TGC CTA TGG 95–60–72 15–15–20Primer B: GGG CGT GCA GGG CAG AGA CAA CAC

N-acyl-phosphatidylethanolaminephospholipase D (NAPE-PLD)

Napepld Primer A: TGC TAG GGC CTT GGA ATC GGT TCT T 95–58.1–72 15–10–15Primer B: TCA GTT TCA CCG GCG GCT CTA AGT A

4 Riccardo Dore et al.

© 2013 Society for the Study of Addiction Addiction Biology

within-subject factors. The effects of SR141716 on thetime spent in the open compartment and the amount offood eaten were analyzed using two-way ANOVAs, withDiet Schedule and Treatment as between-subjects factors.The effects of diet alternation on gene expression wereanalyzed using Student’s t-tests. Pairwise effects wereinterpreted using Fisher’s least significant difference tests.The software/graphic packages used were Systat 12.0,SigmaPlot 11.0 (Systat Software Inc., Chicago, IL, USA),InStat 3.0 (GraphPad, San Diego, CA, USA) and Statistica7.0 (Statsoft, Inc., Tulsa, OK, USA).

RESULTS

Effects of diet alternation on food intake, body weightand feed efficiency

Figure 2a shows that alternating access to the chocolate-flavored, sugary, palatable diet progressively altered dailyfood intake of male rats in a diet-specific manner[Cycle ¥ Diet Phase ¥ Diet Schedule: F(7,126) = 30.72,P < 0.001]. Chow/Palatable rats were already overeatingcompared to control rats during the very first access tothe palatable diet [t(18) = -3.22, P < 0.01; Fig. 1a, P

(c)

(e)

(d)

(a) (b)

(f)

Chow/ChowChow/Palatable

Chow/ChowChow/Palatable

Chow/ChowChow/Palatable

Chow/ChowChow/Palatable

Chow/ChowChow/Palatable

Chow/ChowChow/Palatable

140

120

100

80

0

Ave

rage

dai

ly fo

od in

take

(kc

al)

Ave

rage

dai

ly

body

wei

ght c

hang

e (g

)A

vera

ge d

aily

fe

ed e

ffici

ency

(m

g/kc

al)

Cum

ulat

ive

feed

effi

cien

cy (

mg/

kcal

)C

umul

ativ

ebo

dy w

eigh

t cha

nge

(g)

Cum

ulat

ive

food

inta

ke (

kcal

)

C P1

C P C P C P C P C P C P C P2 3 4 5 6 7 8

Diet phase (cycle)

C P1

C P C P C P C P C P C P C P2 3 4 5 6 7 8

Diet phase (cycle)

Diet phase (cycle)

3000

2000

1000

01 2 3 4 5 6

Cycle

1 2 3 4 5 6

Cycle

1 2 3 4 5 6

Cycle

20

15

10

5

0

−5

200

150

100

50

0

150

100

50

0

−50

120

100

80

60

0

7 8

7 8

7 8C P1

C P C P C P C P C P C P C P2 3 4 5 6 7 8

Figure 2 Effects of intermittent, extended access to a palatable diet on (a) average daily food intake, (b) cumulative food intake, (c) averagedaily body weight change, (d) cumulative body weight change, (e) average daily feed efficiency, and (f) cumulative feed efficiency in maleWistar rats (n = 20). Values of C phase represent the average of 2 days access to chow food. Data represent mean (�SEM). Symbolsdenote significant difference from Chow/Chow group **P < 0.01, ***P < 0.001

CB system and compulsive eating 5

© 2013 Society for the Study of Addiction Addiction Biology

phase of cycle 1], and upon returning to the regularchow diet, cycled rats showed a significant hypophagiaduring the following day compared to Chow/Chowrats [M � SEM: 94.1 � 1.4 kcal versus 72.9 � 3.0 kcal;t(18) = 6.43 P < 0.001; Chow/Chow and Chow/Palatablerats, respectively, not shown], which lasted up to the com-pletion of the 2 days of the C phase [t(18) = 4.62,P < 0.001; Fig. 1a, C phase of cycle 2]. Therefore, Chow/Palatable rats spontaneously cycled their food intake, asfunction of the diet which was provided, and the magni-tude of both chow diet hypophagia and palatable dietovereating progressively increased with further dietswitches. The magnitude of the chow diet hypophagia inChow/Palatable rats (M � SEM: -238.1 kcal, -15.3%versus Chow/Chow rats, difference of intakes during the2-day C phases across eight cycles) was similar to themagnitude of the palatable diet overeating (M � SEM:+207.2 kcal, +26.9% versus Chow/Chow rats, differenceof intakes during the 1-day P phases across eight cycles),which resulted in a non-statistically significant differencein cumulative energy intake between the two groups,across the eight cycles [Diet Schedule: F(1,18) = 0.10,NS; Cycle ¥ Diet Schedule: F(7,126) = 0.37, NS; Fig. 2b].

Chow/Palatable rats spontaneously cycled not onlytheir food intake but also their body weight and feed effi-ciency [average daily body weight change: Cycle ¥ DietPhase ¥ Diet Schedule: F(7,126) = 22.54, P < 0.001;average daily feed efficiency: Cycle ¥ Diet Phase ¥ DietSchedule: F(7,126) = 18.58, P < 0.001; Fig. 1c,e]. Begin-ning from cycle 3, Chow/Palatable rats disproportionallygained more body weight than Chow/Chow rats during theP phases, resulting in a significantly increased local feedefficiency. The magnitude of these effects progressivelyamplified, cycle by cycle. Similarly, beginning from cycle 5,Chow/Palatable rats disproportionally gained less bodyweight than Chow/Chow rats during C phases, resulting ina decreased feed efficiency. These effects progressivelyintensified to the point that Chow/Palatable rats startedlosing absolute body weight during the C phase of cycle 8.

Although cumulative body weight gain did not differbetween groups [Diet Schedule: F(1,18) = 1.89, NS;Cycle ¥ Diet Schedule: F(7,126) = 1.55, NS; Fig. 2d],Chow/Palatable rats showed an increased feed efficiency,compared to Chow/Chow rats [Diet Schedule: F(1,18) =2.36, NS; Cycle ¥ Diet Schedule: F(7,126) = 2.19, P <0.05; Fig. 2f]. However, post hoc comparisons did notreveal any statistically significant effect at any of the timepoints analyzed.

Effects of diet alternation on escalation of palatablefood consumption

Figure 3a shows the effects of diet alternation on foodintake during the first hour of access to the palatable diet.

Chow/Palatable rats progressively and dramatically esca-lated the intake of the palatable diet during the firsthour of access, as a function of the number of cycles[Cycle ¥ Diet Schedule: F(7,126) = 30.72, P < 0.001].The intake of diet-cycled rats became significantly higherthan the intake of chow in control rats starting from thesecond palatable diet access; by the seventh access, Chow/Palatable rats were able to consume approximately5.5-fold the intake of Chow/Chow rats [M � SEM;

(a)

(b)

(c)

Chow/ChowChow/Palatable

Chow/Chow P phaseChow/Palatable P phase

Chow/Chow C phaseChow/Palatable C phase

1-ho

ur fo

od in

take

(kc

al)

50

40

30

20

10

01 2 3 4 5 6 7 8

Cycle

Cum

ulat

ive

food

inta

ke (

kcal

)C

umul

ativ

efo

od in

take

(kc

al)

150

120

90

60

30

0

150

120

90

60

30

0

1 3 6 24Time (hour)

1 3 6 24Time (hour)

Figure 3 Effects of intermittent, extended access to a palatable dieton (a) food intake during the first hour of the P phase (whenChow/Palatable rats are fed the sugary, palatable diet), (b) time coursefood intake during P phase, and (c) time course food intake duringC phase in male Wistar rats (n = 20). Food intake was measured 1, 3,6 and 24 hours. Data represent mean (�SEM). Symbols denotesignificant difference from Chow/Chow group ***P < 0.001

6 Riccardo Dore et al.

© 2013 Society for the Study of Addiction Addiction Biology

43.4 � 1.8 kcal. versus 7.9 � 1.3 kcal., respectively;t(18) = -15.52, P < 0.001].

Effects of diet alternation on time course of food intakeduring P and C phases

As shown in Fig. 3b, diet-cycled rat overeating during Pphase was greatest toward the first hours following theswitch to the palatable diet. Indeed, the degree to whichChow/Palatable rats overate compared to control Chow/Chow rats peaked at the first hour of access, anddecreased during the remaining 23 hours [Diet Schedule:F(1,18) = 66.6, P < 0.001; Time ¥ Diet Schedule:F(3,54) = 23.02, P < 0.001]. Their cumulative intakewas 745.2 � 60.1% that of controls after 1 hour,349.0 � 19.5% at 3 hours, 233.0 � 9.2% at 6 hoursand 129.6 � 6.2% at 24 hours (M � SEM).

Analogously, the degree to which Chow/Palatablerats underate compared to control Chow/Chow ratsbottomed at the first hour of access, and increasedduring the remaining 23 hours [cumulative intake was18.2 � 7.4% that of controls after 1 hour, 36.4 � 8.7%at 3 hours, 64.0 � 5.9% at 6 hours and 74.3 � 3.8%at 24 hours (M � SEM), Fig. 3c]. However, while asignificant effect of the Diet Schedule was detected[F(1,18) = 33.6, P < 0.001], the Time ¥ Diet Scheduleinteraction did not reach statistical significance[F(3,54) = 1.46, NS].

Effects of SR141716 on food intake and body weight

Pre-treatment with the CB1 receptor inverse agonistSR141716 significantly decreased food intake [Dose:F(3,60) = 16.91, P < 0.001], but, as revealed by the sig-nificant Diet Schedule ¥ Dose interaction, the anorecticeffect was dissimilar between Chow/Chow and Chow/Palatable rats [F(3,60) = 3.35, P < 0.05]. As shown inFig. 4, although SR141716 maintained the same efficacyin both groups at the highest dose injected (3 mg/kg), thedrug treatment more potently reduced the intake of thehighly palatable diet in Chow/Palatable rats, than itreduced the consumption of the chow in control Chow/Chow rats. Indeed, pairwise comparisons showed thatSR141716, at the dose of 1 mg/kg in Chow/Palatable rats,consistently decreased the intake of the palatable dietduring the entire 24 hours observation period, comparedto the vehicle condition, whereas the same dose onlyreduced intake in Chow/Chow rats during the first hour(Fig. 4b,f). SR141716 treatment also reduced bodyweight gain [Dose: F(3,60) = 25.79, P < 0.001], butagain dissimilarly between Chow/Chow and Chow/Palatable rats [Diet Schedule ¥ Dose: F(3,60) = 3.91,P < 0.05; Fig. 4c,f]. Indeed, although drug treatmentmaintained the same potency in decreasing body weight

gain in both the groups (lowest effective dose: 3 mg/kg),SR141716 reduced body weight more efficaciously inChow/Palatable rats compared to control Chow/Chow rats.Indeed, the body weight loss induced by the 3 mg/kg dosewas ~2.5-fold bigger in Chow/Palatable rats than incontrol Chow/Chow rats [M � SEM: 4.6 � 1.6 g versus11.1 � 1.3 g; t(20) = 3.14 P < 0.01; as different from therespective vehicle condition in Chow/Chow and Chow/Palatable rats, respectively].

Effects of SR141716 on risk-taking behavior andcompulsive eating

As shown in Fig. 5a, the SR141716 effect on the timespent in the light, aversive compartment was selective forChow/Palatable rats as drug treatment did not influencethis measure in control Chow/Chow rats [DietSchedule ¥ Drug Treatment: F(1,40) = 5.99, P < 0.02].As revealed by post hoc analysis, vehicle-treated Chow/Palatable rats spent more time in the aversive light com-partment, where the palatable diet was positioned,compared to vehicle-treated Chow/Chow rats (P < 0.02).However, when Chow/Palatable rats were pre-treated with1 mg/kg of SR141716 (the lowest effective dose in thefood intake experiment), the time spent there did not sig-nificantly differ from vehicle-treated Chow/Chow rats, andtended to differ from vehicle-treated Chow/Palatable rats(P = 0.07).

Moreover, during the 10 minute test, Chow/Palatablerats compulsively consumed ~70-fold the food consumedby control Chow/Chow rats, in spite of the aversive envi-ronmental conditions. This effect was potently reducedby SR141716 pre-treatment [-49%, P < 0.001; DietSchedule ¥ Drug Treatment: F(1,40) = 9.37, P < 0.01;Fig. 5b]. SR141716 did not affect food consumption inChow/Chow rats.

Quantitative RT-PCR

Chow/Palatable rats showed an increased gene expressionof CB1 receptors in the DS, compared to Chow/Chow rats[t(11) = 2.2, P < 0.05; Fig. 6a]. No differential effect inthe mRNA expression of CB1 between groups wasobserved in the NAcc, LH and VMH [ts(12–13) < 1.36,NS].

In addition, a decreased gene expression of theenzyme FAAH was observed in the VMH of Chow/Palatable rats as compared to Chow/Chow rats [t(9) = 3.9,P < 0.001; Fig. 6b]. Chow/Chow and Chow/Palatable ratsdid not differ in the mRNA expression of FAAH in theNAcc, DS and LH [ts(10–13) < 0.63, NS].

Chow/Chow and Chow/Palatable rats did not differ inthe expression of MAGL, DAGLa and NAPE-PLD in any ofthe brain areas analyzed [ts(9–14) < 1.7, NS; Fig. 6c–e].

CB system and compulsive eating 7

© 2013 Society for the Study of Addiction Addiction Biology

DISCUSSION

Consummatory and motivational outcomes ofintermittent, extended access to a palatable diet

In the present study, we developed a novel animal modelof compulsive eating resulting from an intermittent,extended access to a sugary, highly palatable diet, which,compared to other existing models (Cottone et al.2009a,b), is characterized by a considerably fastercycling protocol. Rats undergoing palatable diet alterna-tion developed maladaptive feeding adaptations whichwere a function of the diet offered and were characterizedby progressively increasing hypophagia of the less palat-able, regular chow diet, and progressively increasingovereating of the sugary diet. Although palatable diet-cycled rats showed large fluctuations of food intake,cumulative food intake of Chow/Palatable rats did notsignificantly differ from Chow/Chow group since the

magnitude of the hypophagia was comparable to themagnitude of the overeating. Palatable diet alternationalso resulted in spontaneous fluctuations in body weightand feed efficiency. Interestingly, starting from theseventh palatable food withdrawal, cycled rats beganundereating the regular chow diet at the point of sponta-neously losing absolute body weight. Similarly to whatwas observed with food intake, cumulative body weightgain and feed efficiency were not affected by the intermit-tent, extended access to the palatable diet. Although pal-atable diet alternation did not influence body weight, wecannot exclude whether the two groups could still differin fat mass. Future body composition analysis will beneeded to specifically address this point. The alternatingpatterns of food intake, body weight gain and feed effi-ciency are analogous to the ones observed with the pre-viously published protocol of intermittent, extendedaccess to palatable food (Cottone et al. 2008a, 2009b),

Chow/Chow

Chow/Palatable

(a) (b) (c)

(d) (e) (f)

Foo

d in

take

(kc

al)

Foo

d in

take

(kc

al)

Foo

d in

take

(kc

al)

Foo

d in

take

(kc

al)100

75

50

25

0

100

75

50

25

0

VehicleSR141716A 0.3 mg/kgSR141716A 1.0 mg/kgSR141716A 3.0 mg/kg

VehicleSR141716A 0.3 mg/kgSR141716A 1.0 mg/kgSR141716A 3.0 mg/kg

1 3 6

Time (hour)

1 3 6

Time (hour)

150

120

90

60

30

0

24 hours

150

120

90

60

30

0

24 hours

24 hours

Bod

y w

eigh

t cha

nge

(g)

20

15

10

5

0

−5

24 hours

Bod

y w

eigh

t cha

nge

(g)

20

15

10

5

0

−5

Figure 4 Effects of pre-treatment (-30 minutes) with the inverse agonist of the type-1 cannabinoid (CB1) receptor SR141716 (0, 0.3, 1 and3 mg/kg, intraperitoneal) on 1, 3, 6 and 24 hours food intake and body weight change of Chow/Chow (a, b and c) and Chow/Palatable (d, e andf) male Wistar rats (n = 22). Drug treatment was performed at the beginning of the P phase. Data represent mean (�SEM). Symbols denotesignificant difference from vehicle-treated rats *P < 0.05, **P < 0.01, ***P < 0.001, linear trend dose effect $P < 0.05, $$P < 0.01, $$$P < 0.001,main dose effect #P < 0.05, ##P < 0.01, ###P < 0.001

8 Riccardo Dore et al.

© 2013 Society for the Study of Addiction Addiction Biology

with the difference in that the same number of cycles inthis study were completed in less than half of the time.

Importantly, intermittent, extended access to thehighly palatable diet induced a dramatic escalation offood consumption during the first hour of access, as afunction of the number of cycles. The escalation of pal-atable food intake during the first hour of access was anexperience-dependent phenomenon as indicated by thestrong Cycle ¥ Diet Schedule interaction and by the excel-lent fit of intake to the sigmoidal associative learningfunction (min: 5.9 � 1.8, max: 44.2 � 3.0, ET50:2.6 � 0.3, Hillslope: 2.2 � 0.6, r2 = 0.98, P < 0.0001,M � SEM). During the first hour of renewed access to the

palatable diet, cycled rats were able to consume up to~7.5-fold the intake of control rats; the amount of foodeaten by Chow/Palatable rats in 1 hour was equivalent to~42% of the daily caloric intake of Chow/Chow rats. Esca-lation of a behavioral response similar to the one shownhere has been also observed with intermittent, extendedaccess to drugs of abuse like cocaine (Ahmed & Koob1998), heroin (Ahmed, Walker & Koob 2000) and nico-tine (George et al. 2007), as well as with short, intermit-tent access to palatable diets (Cottone et al. 2008b, 2012;Parylak et al. 2012). Given the very similar caloricdensity of the chow and the palatable diet, the excessivecaloric intake reflected indeed an active binge-like eatingof food. The observed excessive food intake in such a briefperiod of time was, therefore, consistent with a hedonic,rather than a homeostatic mechanism.

Withdrawal from intermittent, extended access to thepalatable diet increased the motivation to compulsivelyeat the sugary diet while facing aversive conditions.Indeed, Chow/Palatable rats spent significantly more timein an aversive compartment and ate ~70 times more thancontrol Chow/Chow rats when they were offered the pal-atable diet in an open, aversive compartment of a light/dark conflict box at the end of the 2 days of withdrawal ofthe C phase. The increased motivation for the palatablediet observed in the light/dark conflict test in Chow/Palatable rats compared to Chow/Chow rats was depend-ent on the diet history and not on the diet offered in thelight compartment. Indeed, in performing a light/darkconflict test in Chow/Chow rats where we provided eitherthe regular chow diet or the palatable in the aversive com-partment, we found that neither of the two groups ate thefood and they spent a similar amount of time in the lightcompartment (P = 1.00 and P = 0.47, food intake andtime in the light compartment, respectively, not shown).Overall these results suggest that withdrawal from inter-mittent, extended access to a highly palatable diet causesthe emergence of compulsive eating and risk-takingbehavior, similarly to what has been observed in alcoholand drug addiction and in certain forms of eating disor-ders and obesity (Vanderschuren & Everitt 2004;Teegarden & Bale 2007; Heyne et al. 2009; Hopf et al.2010; Johnson & Kenny 2010; Cottone et al. 2012).

Effects of SR141716 on excessive intake ofhighly palatable food, compulsive eating andrisk-taking behavior

In this study, we have found that the CB1 receptor inverseagonist, SR141716, reduced the excessive intake of thehighly palatable diet in diet-cycled rats more potentlythan it reduced the intake of chow in control rats. Duringthe first hour of access to the highly palatable diet, wherethe largest difference between Chow/Chow and Chow/

Figure 5 Effects of pre-treatment (-30 minutes) with theinverse agonist of the type-1 cannabinoid (CB1) receptor SR141716(0, 1 mg/kg, intraperitoneal) on risk-taking behavior and compulsiveeating in male Wistar rats (n = 44). (a) Time spent in the open,aversive compartment; (b) food intake. Data represent mean(�SEM). Symbols denote significant difference from Chow/Palatablevehicle-treated group *P < 0.05, ***P < 0.001. § denotes a trend(P = 0.07)

CB system and compulsive eating 9

© 2013 Society for the Study of Addiction Addiction Biology

Palatable rats’ intake is observed, the 3 mg/kg dose ofSR141716 reduced intake of the palatable diet by 43%.In addition, at the highest dose injected, the CB1 inverseagonist reduced with greater efficacy body weight in diet-cycled rats compared to control chow-fed rats. Theseresults are consistent with the hypothesis that CB1 recep-tor antagonism decreases the hedonic value of food and,therefore, preferentially decreases the intake of highlypalatable diets (Arnone et al. 1997; Simiand et al. 1998;Jarrett, Scantlebury & Parker 2007; Mathes, Ferrara &Rowland 2008). On the other hand, our results mayappear in contrast with recent evidence showing that theCB1 receptor antagonist SR147778 is less effective inreducing binge-like eating of a sweet fat diet compared toa control diet (Parylak et al. 2012). However, manydifferences between the two studies may account fordifferent effects observed. Indeed, here, we used an inter-mittent, extended access (24 hours) to a sugary diet andmale rats were ad libitum fed for the entire diet alternationprocedure, while in the Parylak’s study, binge-like eatingwas induced by a highly limited (10 minutes) access to asweet fat diet and female rats were exposed daily to a2-hour food deprivation. In addition, contrarily to whatobserved here, binge-like eating rats showed increasedbody weight gain, feed efficiency, as well as adiposity(Parylak et al. 2012).

SR141716 injected at the lowest effective dose (1 mg/kg) in the dose-response study, reduced both risk-takingbehavior and compulsive-like eating in the light/dark

conflict test, without affecting behavior in control Chow/Chow rats. These observations suggest that CB1 receptorsplays a role in the loss of control and in the compulsive-ness associated with excessive intake of palatable food.More generally, our findings support the hypothesis thatCB1 receptors mediate the rewarding properties of addic-tive substances and that its blockade represents a prom-ising pharmacological tool to treat compulsive palatablefood/drug seeking (Serrano & Parsons 2011). Indeed,pharmacological blockade of CB1 receptor has been dem-onstrated to reduce compulsive seeking behavior formany drugs of abuse (De Vries et al. 2001; Fattore et al.2003; Cippitelli et al. 2005; Cohen et al. 2005).

Quantitative RT-PCR analysis, performed during with-drawal from intermittent, extended access to a palatablediet (same time point of the dose-response and risk-taking behavior/compulsive eating experiments) revealeddecreased mRNA levels of the AEA-degrading enzymeFAAH in the VMH and an increased gene expression ofCB1 receptors in the DS in cycled rats. Both results point atan increased cannabinergic tone in the brain of Chow/Palatable rats, which is consistent with the preferentialeffects of SR141716 in reducing excessive intake of pal-atable food, compulsive eating and risk-taking behaviorin food-cycled rats. Indeed, the decreased expression ofthe enzyme FAAH in the VMH observed here suggests areduced degradation mechanism and therefore increasedlevels of the endocannabinoid AEA. The VMH is a keynucleus in the regulation of appetite (Suzuki et al. 2010),

(a) (b)

(d) (e)

(c)150

125

100

75

0

150

125

0

150

125

100

75

0

150

125

100

75

0

Chow/ChowChow/Palatable

mR

NA

CB

1/C

yclo

phili

n A

mR

NA

FA

AH

/Cyc

loph

ilin

A

mR

NA

DA

GLa

lpha

/Cyc

loph

ilin

A

mR

NA

NA

PE

-PLD

/Cyc

loph

ilin

A

mR

NA

MA

GL/

Cyc

loph

ilin

A

DSNAcc LH

VMH DS

NAcc LHVM

H

DSNAcc LH

VMH DS

NAcc LHVM

H

DSNAcc LH

VMH

100

75

150

125

0

100

75

Figure 6 (a) Cannabinoid receptor type 1 (CB1), (b) fatty acid amide hydrolase (FAAH), (c) monoacylglycerol lipase (MAGL), (d)diacylglycerol lipase a (DAGLa) and (e) N-acyl-phosphatidylethanolamine phospholipase D (NAPE-PLD) mRNA expression in the dorsalstriatum (DS), nucleus accumbens (NAcc), lateral hypothalamus (LH) and ventromedial hypothalamus (VMH) of Chow/Chow and Chow/Palatable (n = 7–12) rats. Data represent mean (�SEM) expressed as percent of Chow/Chow group. Symbols denote significant differencefrom Chow/Chow group *P < 0.05, **P < 0.01

10 Riccardo Dore et al.

© 2013 Society for the Study of Addiction Addiction Biology

and the CB1 receptor mRNA within this area has beenshown to be highly expressed compared to other hypotha-lamic nuclei (Mailleux & Vanderhaeghen 1992). Veryinterestingly, microinfusion of AEA in the VMH has beenshown to induce overeating in pre-satiated rats (Jamshidi& Taylor 2001). This effect was prevented by pre-treatment with SR141716, suggesting a primary role forAEA in the VMH in promoting appetite through the acti-vation of CB1 receptors (Jamshidi & Taylor 2001). Thus,we can hypothesize that withdrawal from intermittent,extended access to palatable food induced an increasedAEA tone in the VMH, which may have contributed tothe excessive intake observed upon renewing access tothe palatable diet. Consistent with this hypothesis,SR141716 decreased palatable food intake with higherpotency in cycled rats compared to control chow-fed rats.

In addition, we observed an increased gene expres-sion of CB1 receptors in the DS of rats withdrawn fromthe palatable diet. The DS is a brain area heavilyinvolved in reward processing and decision making, andit integrates sensorimotor, cognitive and motivational/emotional information (Balleine, Delgado & Hikosaka2007). CB1 receptors, which are heavily expressed inthis area (Pettit et al. 1998), play an important role inmodulation of reward, and their stimulation has beendemonstrated to increase dopamine release (Szabo,Muller & Koch 1999). Therefore, it can be hypothesizedthat the increased expression of CB1 receptors in the DSof palatable food withdrawn rats would determine adopamine overflow, which may in turn be responsiblefor the heightened motivation for the sugary diet even inadverse conditions.

In summary, our behavioral, pharmacological andmolecular observations give important insights into theneurobiology of compulsive eating and will be importantto develop novel pharmacological treatments.

Acknowledgements

We thank Stephen St Cyr and Patrick De Souza for tech-nical assistance and Rahul Rao for editorial assistance.This publication was made possible by grant numbersDA023680, DA030425, MH091945, MH093650 andAA016731 from the National Institute on Drug Abuse(NIDA), the National Institute of Mental Health (NIMH)and the National Institute on Alcohol Abuse and Alco-holism (NIAAA), by the Peter Paul Career DevelopmentProfessorship (P.C.) and by Boston University’s Under-graduate Research Opportunities Program (UROP). Thisresearch was also supported by the NIH IntramuralResearch Programs of the National Institute on DrugAbuse and the National Institute of Alcohol Abuse andAlcoholism, NIH, DHHS. Its contents are solely theresponsibility of the authors and do not necessarily

represent the official views of the National Institutes ofHealth.

Disclosure/Conflict of Interest

The authors declare no conflict of interest.

Authors’ Contribution

RD, VS and PC were responsible for the study concept anddesign; RD and MV performed behavioral experiments;XW performed quantitative RT-PCR; KCR provided criti-cal reagents; RD, MV and PC performed data analysis; RD,VS and PC drafted the manuscript. All authors providedcritical revision of the manuscript for important intellec-tual content and approved final version for publication.

References

Ahmed SH, Koob GF (1998) Transition from moderate to exces-sive drug intake: change in hedonic set point. Science282:298–300.

Ahmed SH, Walker JR, Koob GF (2000) Persistent increase in themotivation to take heroin in rats with a history of drug esca-lation. Neuropsychopharmacology 22:413–421.

APA (2000) Diagnostic and Statistical Manual of MentalDisorders, 4th edn-Text Revision. Washington, DC: AmericanPsychiatric Association.

Arnone M, Maruani J, Chaperon F, Thiebot MH, Poncelet M,Soubrie P, Le Fur G (1997) Selective inhibition of sucrose andethanol intake by SR 141716, an antagonist of centralcannabinoid (CB1) receptors. Psychopharmacology (Berl)132:104–106.

Balleine BW, Delgado MR, Hikosaka O (2007) The role of thedorsal striatum in reward and decision-making. J Neurosci27:8161–8165.

Bello NT, Coughlin JW, Redgrave GW, Ladenheim EE, Moran TH,Guarda AS (2012) Dietary conditions and highly palatablefood access alter rat cannabinoid receptor expression andbinding density. Physiol Behav 105:720–726.

Birch LL (1999) Development of food preferences. Annu RevNutr 19:41–62.

Blasio A, Steardo L, Sabino V, Cottone P (2013) Opioid system inthe medial prefrontal cortex mediates binge-like eating. AddictBiol. doi: 10.1111/adb.12033. [Epub ahead of print].

Cifani C, Polidori C, Melotto S, Ciccocioppo R, Massi M (2009) Apreclinical model of binge eating elicited by yo-yo dieting andstressful exposure to food: effect of sibutramine, fluoxetine,topiramate, and midazolam. Psychopharmacology (Berl)204:113–125.

Cippitelli A, Bilbao A, Hansson AC, del Arco I, Sommer W, HeiligM, Massi M, Bermudez-Silva FJ, Navarro M, Ciccocioppo R, deFonseca FR (2005) Cannabinoid CB1 receptor antagonismreduces conditioned reinstatement of ethanol-seeking behav-ior in rats. Eur J Neurosci 21:2243–2251.

Cohen C, Perrault G, Griebel G, Soubrie P (2005) Nicotine-associated cues maintain nicotine-seeking behavior in ratsseveral weeks after nicotine withdrawal: reversal by thecannabinoid (CB1) receptor antagonist, rimonabant(SR141716). Neuropsychopharmacology 30:145–155.

Corwin RL (2004) Binge-type eating induced by limited access inrats does not require energy restriction on the previous day.Appetite 42:139–142.

CB system and compulsive eating 11

© 2013 Society for the Study of Addiction Addiction Biology

Corwin RL, Avena NM, Boggiano MM (2011) Feeding andreward: perspectives from three rat models of binge eating.Physiol Behav 104:87–97.

Corwin RL, Buda-Levin A (2004) Behavioral models of binge-type eating. Physiol Behav 82:123–130.

Corwin RL, Grigson PS (2009) Symposium overview—foodaddiction: fact or fiction? J Nutr 139:617–619.

Cottone P, Sabino V, Roberto M, Bajo M, Pockros L, Frihauf JB,Fekete EM, Steardo L, Rice KC, Grigoriadis DE, Conti B, KoobGF, Zorrilla EP (2009a) CRF system recruitment mediatesdark side of compulsive eating. Proc Natl Acad Sci U S A106:20016–20020.

Cottone P, Sabino V, Steardo L, Zorrilla EP (2008a) Intermittentaccess to preferred food reduces the reinforcing efficacy ofchow in rats. Am J Physiol Regul Integr Comp Physiol295:R1066–R1076.

Cottone P, Sabino V, Steardo L, Zorrilla EP (2008b) Opioid-dependent anticipatory negative contrast and binge-likeeating in rats with limited access to highly preferred food.Neuropsychopharmacology 33:524–535.

Cottone P, Sabino V, Steardo L, Zorrilla EP (2009b) Consumma-tory, anxiety-related and metabolic adaptations in female ratswith alternating access to preferred food. Psychoneuroendo-crinology 34:38–49.

Cottone P, Wang X, Park JW, Valenza M, Blasio A, Kwak J, IyerMR, Steardo L, Rice KC, Hayashi T, Sabino V (2012) Antago-nism of sigma-1 receptors blocks compulsive-like eating. Neu-ropsychopharmacology 37:2593–2604.

De Vries TJ, Shaham Y, Homberg JR, Crombag H, Schuurman K,Dieben J, Vanderschuren LJ, Schoffelmeer AN (2001) A can-nabinoid mechanism in relapse to cocaine seeking. Nat Med7:1151–1154.

DiPatrizio NV, Piomelli D (2012) The thrifty lipids: endocannabi-noids and the neural control of energy conservation. TrendsNeurosci 35:403–411.

DiPatrizio NV, Simansky KJ (2008) Activating parabra-chial cannabinoid CB1 receptors selectively stimulatesfeeding of palatable foods in rats. J Neurosci 28:9702–9709.

Fattore L, Spano MS, Cossu G, Deiana S, Fratta W (2003) Can-nabinoid mechanism in reinstatement of heroin-seeking aftera long period of abstinence in rats. Eur J Neurosci 17:1723–1726.

George O, Ghozland S, Azar MR, Cottone P, Zorrilla EP, ParsonsLH, O’Dell LE, Richardson HN, Koob GF (2007) CRF-CRF1system activation mediates withdrawal-induced increases innicotine self-administration in nicotine-dependent rats. ProcNatl Acad Sci U S A 104:17198–17203.

Heyne A, Kiesselbach C, Sahun I, McDonald J, Gaiffi M, DierssenM, Wolffgramm J (2009) An animal model of compulsivefood-taking behaviour. Addict Biol 14:373–383.

Hill JO, Wyatt HR, Reed GW, Peters JC (2003) Obesity and theenvironment: where do we go from here? Science 299:853–855.

Hopf FW, Chang SJ, Sparta DR, Bowers MS, Bonci A(2010) Motivation for alcohol becomes resistant to quinineadulteration after 3 to 4 months of intermittent alcoholself-administration. Alcohol Clin Exp Res 34:1565–1573.

Iemolo A, Valenza M, Tozier L, Knapp CM, Kornetsky C, SteardoL, Sabino V, Cottone P (2012) Withdrawal from chronic, inter-mittent access to a highly palatable food induces depressive-like behavior in compulsive eating rats. Behav Pharmacol23:593–602.

Jamshidi N, Taylor DA (2001) Anandamide administration intothe ventromedial hypothalamus stimulates appetite in rats. BrJ Pharmacol 134:1151–1154.

Jarrett MM, Scantlebury J, Parker LA (2007) Effect of delta9-tetrahydrocannabinol on quinine palatability and AM251 onsucrose and quinine palatability using the taste reactivity test.Physiol Behav 90:425–430.

Johnson PM, Kenny PJ (2010) Dopamine D2 receptors inaddiction-like reward dysfunction and compulsive eating inobese rats. Nat Neurosci 13:635–641.

Kumar AA, Singh SG, Venkate AC, Lalit W (2008) An improvedprocess for the preparation of rimonabant. Patent publicationWO200862480 A2.

Kunos G (2007) Understanding metabolic homeostasis andimbalance: what is the role of the endocannabinoidsystem? Am J Med 120 (9 Suppl 1):S18–S24; discussionS24.

Maier EY, Ahrens AM, Ma ST, Schallert T, Duvauchelle CL(2010) Cocaine deprivation effect: cue abstinence over week-ends boosts anticipatory 50-kHz ultrasonic vocalizations inrats. Behav Brain Res 214:75–79.

Mailleux P, Vanderhaeghen JJ (1992) Distribution of neuronalcannabinoid receptor in the adult rat brain: a comparativereceptor binding radioautography and in situ hybridizationhistochemistry. Neuroscience 48:655–668.

Mathes CM, Ferrara M, Rowland NE (2008) Cannabinoid-1receptor antagonists reduce caloric intake by decreasingpalatable diet selection in a novel dessert protocol in femalerats. Am J Physiol Regul Integr Comp Physiol 295:R67–R75.

Mathes WF, Brownley KA, Mo X, Bulik CM (2009) The biology ofbinge eating. Appetite 52:545–553.

Matias I, Bisogno T, Di Marzo V (2006) Endogenous cannabi-noids in the brain and peripheral tissues: regulation of theirlevels and control of food intake. Int J Obes (Lond) 30 (Suppl1):S7–S12.

McBride WJ, Li TK (1998) Animal models of alcoholism: neuro-biology of high alcohol-drinking behavior in rodents. Crit RevNeurobiol 12:339–369.

Mechoulam R, Berry EM, Avraham Y, Di Marzo V, Fride E(2006) Endocannabinoids, feeding and suckling—from ourperspective. Int J Obes (Lond) 30 (Suppl 1):S24–S28.

Parylak SL, Cottone P, Sabino V, Rice KC, Zorrilla EP (2012)Effects of CB1 and CRF1 receptor antagonists on binge-like eating in rats with limited access to a sweet fat diet:lack of withdrawal-like responses. Physiol Behav 107:231–242.

Pettit DA, Harrison MP, Olson JM, Spencer RF, Cabral GA (1998)Immunohistochemical localization of the neural cannabinoidreceptor in rat brain. J Neurosci Res 51:391–402.

Polivy J, Herman CP (1985) Dieting and binging. A causal analy-sis. Am Psychol 40:193–201.

Sabino V, Cottone P, Blasio A, Iyer MR, Steardo L, RiceKC, Conti B, Koob GF, Zorrilla EP (2011) Activationof sigma-receptors induces binge-like drinking in Sardi-nian alcohol-preferring rats. Neuropsychopharmacology 36:1207–1218.

Serrano A, Parsons LH (2011) Endocannabinoid influence indrug reinforcement, dependence and addiction-related behav-iors. Pharmacol Ther 132:215–241.

Simiand J, Keane M, Keane PE, Soubrie P (1998) SR 141716, aCB1 cannabinoid receptor antagonist, selectively reducessweet food intake in marmoset. Behav Pharmacol 9:179–181.

12 Riccardo Dore et al.

© 2013 Society for the Study of Addiction Addiction Biology

Suzuki K, Simpson KA, Minnion JS, Shillito JC, Bloom SR (2010)The role of gut hormones and the hypothalamus in appetiteregulation. Endocr J 57:359–372.

Szabo B, Muller T, Koch H (1999) Effects of cannabinoids ondopamine release in the corpus striatum and the nucleusaccumbens in vitro. J Neurochem 73:1084–1089.

Teegarden SL, Bale TL (2007) Decreases in dietary preferenceproduce increased emotionality and risk for dietary relapse.Biol Psychiatry 61:1021–1029.

Timofeeva E, Baraboi ED, Poulin AM, Richard D (2009) Palatablehigh-energy diet decreases the expression of cannabinoid type1 receptor messenger RNA in specific brain regions in the rat.J Neuroendocrinol 21:982–992.

Vanderschuren LJ, Everitt BJ (2004) Drug seeking becomes com-pulsive after prolonged cocaine self-administration. Science305:1017–1019.

CB system and compulsive eating 13

© 2013 Society for the Study of Addiction Addiction Biology