at SciVerse ScienceDirect

Neuropharmacology 70 (2013) 43e50

Contents lists available

Neuropharmacology

journal homepage: www.elsevier .com/locate/neuropharm

Differential contribution of mesoaccumbens and mesohabenulardopamine to intracranial self-stimulation

Vincent Duchesne, Sandra M. Boye*

Department of Psychiatry, Université de Montréal, 90 Vincent-d’Indy, F-429-3, Outremont, Québec, Canada H2V 2S9

a r t i c l e i n f o

Article history:Received 11 August 2012Received in revised form16 December 2012Accepted 9 January 2013

Keywords:D-AmphetamineDopamineNucleus accumbensShellCoreLateral habenulaIntracranial self-stimulationReward

* Corresponding author. Tel.: þ1 514 343 6111x172E-mail address: sandra.boye@umontreal.ca (S.M. B

0028-3908/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.neuropharm.2013.01.005

a b s t r a c t

The contribution of mesoaccumbens dopamine transmission to intracranial self-stimulation is well-established. However, although the nucleus accumbens comprises two main subregions, the shell andthe core, little is known of the contribution of each to this behaviour. Our first aim was to study theeffects of D-amphetamine infusions into the shell and core in order to understand their relativeimportance to reward and operant responding. Our second aim was to examine the contribution ofa lesser studied group of dopamine neurons, those within the mesohabenular pathway, to intracranialself-stimulation. Male SpragueeDawley rats were implanted with bilateral cannulae in the nucleusaccumbens shell, core or in the lateral habenula and a monopolar stimulation electrode in the posteriormesencephalon, a brain site that is sensitive to changes in dopamine transmission. Using curve-shiftscaling, we measured the reward- and performance-enhancing effects of intra-accumbens (1e20 mg)and intra-habenular (10e40 mg) infusions of D-amphetamine or vehicle. Within the nucleus accum-bens, the use of multiple doses and long test sessions allowed us to observe an interaction between drugeffect and infusion site. We show, for the first time, differences in the minimal doses necessary toenhance rewarding effectiveness and operant responding, in the magnitude of these enhancements aswell as in their duration. Conversely, regardless of dose, intra-habenular D-amphetamine did not alterrewarding effectiveness or operant rate, highlighting the differential contribution of mesoaccumbens andmesohabenular dopamine pathways to intracranial self-stimulation.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The contribution of mesolimbic dopamine (DA) transmission togoal-directed behaviour has been the focus of many years of study.This is especially true of the mesoaccumbens pathway. This path-way is important for mediating goal-directed approach or escapebehaviour in a way that is sensitive to adaptive modifications viaconditioned environmental cues (see Ikemoto and Panksepp (1999)for review). Dopamine in the mesoaccumbens pathway is alsoimportant for the acute rewarding properties of drugs of abuse aswell as of electrical brain stimulation, although few other DA ter-minal sites have received as much research attention. With respectto brain stimulation reward, infusions of DA agonists into the nu-cleus accumbens enhance the rewarding effectiveness of thestimulation (Carlezon and Wise, 1996a; Carr et al., 2009; Colle andWise, 1988; Hayes et al., 2009; Ranaldi and Beninger, 1994a, 1994b)while infusions of DA antagonists attenuate it (Nakajima and

80.oye).

All rights reserved.

Patterson, 1997; Stellar et al., 1983; Stellar and Corbett, 1989).Drug infusions into the more dorsal caudate putamen or prefrontalcortex are relatively ineffective (Colle and Wise, 1988; Ranaldi andBeninger, 1994a, 1994b; Stellar and Corbett, 1989). Other sites, likethe amygdala, have provided inconsistent findings (Stellar andCorbett, 1989; Waraczynski et al., 2010).

On account of neuroanatomy, neurochemistry and function, thenucleus accumbens can be divided into two broad compartments,a ventromedial shell and a dorsolateral core. On the basis of theiranatomical connections, the shell and the core have traditionallybeen designated as preferentially limbic and motoric in function,respectively (Heimer et al., 1991; Zahm and Brog, 1992). For exam-ple, direct infusion of psychostimulants into the nucleus accumbensshell is rewarding (Carlezon et al., 1995; Carlezon and Wise, 1996b;Pontieri et al., 1995; Shin et al., 2008) while lesions of the coreattenuate the locomotor activating properties of such drugs (Boyeet al., 2001; Sellings et al., 2006, 2008; Sellings and Clarke, 2003).Intracranial self-stimulation (ICSS) relies on the rewarding effec-tiveness of the stimulation as well as on the animal’s motoric ca-pacity to emit the operant response. Although early ICSS studiesoften employed the rate of responding as an index of rewarding

V. Duchesne, S.M. Boye / Neuropharmacology 70 (2013) 43e5044

effectiveness, the shortcomings of such measures are now wellestablished (Hodos and Valenstein, 1962) and currently-availablescaling techniques allow dissociation between reward and perfor-mance (Edmonds and Gallistel,1974;Miliaressis et al., 1986). Rate ofresponding remains, however, a valid indicator of the overall ca-pacity of the animal to perform the operant task, which is partic-ularly useful inpharmacological experiments. Despite this,weknowlittle of the role played by the shell and the core of the nucleusaccumbens in ICSS. The first aim of this studywas to infuse differentdoses of D-amphetamine into each subregion in order to charac-terize the relative contribution of increased DA transmissionwithinthese to rewarding effectiveness and operant responding.

The second aim of this study was to examine the contribution ofanother group of DA neurons, those comprising the mesohabenularpathway, to ICSS reward and performance. Mesohabenular DAneurons originate from the rostral portion of the ventral tegmentalarea, ascend via the fasciculus retroflexus and terminate in thelateral habenula (LHb) (Gruber et al., 2007; Herkenham and Nauta,1977; Phillipson and Griffith, 1980). Within the ventral tegmentalarea, their cell bodies are located in paranigral, interfascicular, andparabrachial pigmented nuclei (Gruber et al., 2007; Phillipson andGriffith, 1980), sites that also give rise to DA afferents of the nucleusaccumbens, particularly those that innervate the shell (Ikemoto,2007). Very little is known about the contribution of meso-habenular DA transmission to goal-directed behaviour in generaland to reward function in particular.

The notion that mesohabenular DA may be reward-relevant issupported by at least three observations. First, the LHb plays animportant role in reward error prediction (Bromberg-Martin andHikosaka, 2011) and DA signalling at this site may be pertinent.Second, mesohabenular and mesoaccumbens DA neurons sharenuclei of origin (Gruber et al., 2007; Ikemoto, 2007; Phillipson andGriffith, 1980) suggesting common function. Third, studies of meta-bolic activity, as measured with (14C)-2-deoxyglucose auto-radiography, have consistently shown that pro-dopaminergic stimulisuch as rewarding electrical brain stimulation or systemic admin-istration of DA agonists reduce 2-deoxyglucose uptake in the LHbwhereas anti-dopaminergic treatments such as reward-attenuatingdoses of DA antagonists reliably and significantly increase it(Gallistel et al., 1985; Gomita and Gallistel, 1982; McCulloch et al.,1980; Wechsler et al., 1979). These changes are consistent with theknown effects of DA agonists and antagonists on the firing activity ofmidbrain DA neurons (Bunney et al., 1973). However, since the 2-deoxyglucose technique does not allow identification of the sub-stratemediating the observed changes inmetabolic activity, changesin 2-deoxyglucose uptake may reflect modified afferent input fromdiverse brain sites. Here, we targeted DA terminals specifically byinfusing D-amphetamine directly into the LHb.

2. Materials and methods

2.1. Subjects

Subjects were male SpragueeDawley rats (Charles River, St Constant, Quebec),housed individually in a temperature (21 �C) and humidity (50%) controlled animalcolony with a 12 h light/dark cycle (lights on at 6:30 a.m.). Rats had unrestrictedaccess to food and water and were allowed to habituate to the animal colony for atleast five days prior to surgery. All procedures followed Canadian Council on AnimalCare guidelines and were approved by the Institutional Animal Care Committee.

2.2. Surgery

Rats (300e400 g) were anesthetized with a mixture of oxygen (0.6 L/min) andisoflurane (4%) and mounted onto a stereotaxic apparatus. During surgery, the level ofanesthesia was reduced to 2e3%. Using aseptic techniques, bilateral holes were drilledover the nucleus accumbens or habenula, and a single hole over the posterior mesen-cephalon. Bilateral guide cannulae (26 ga, Plastics One, Roanoke, VA) were implantedinto the LHb (flat skull coordinates: AP: �3.4, ML: �0.6, DV: �5.0 mm), nucleus

accumbens shell (AP: þ1.7, ML: �0.8, DV: �7.7 mm) or nucleus accumbens core(AP: þ1.7, ML �1.5, DV: �6.9 mm), and a stimulation electrode was implanted in theposteriormesencephalon (AP:�7.8, ML: 0, DV:�7.0 to�7.2mm). All coordinateswereobtained fromthePaxinosandWatsonatlasof the ratbrain (PaxinosandWatson,1997).All dorso-ventral (DV) coordinates are in reference to the surface of the skull. Stimu-lationelectrodesweremade fromstainless steelwire (0.27mmdia.) andwere insulatedwith Epoxylite except for the rounded tip. A bare stainless steel wire, connected at oneend to a male amphenol pin and wrapped around four to five miniature screws thatwere threaded into the cranium, servedas the anode.Acrylic dental cementwasused tochronically secure the electrode-cannula assembly to the skull. Prior to the end ofsurgery, all rats received an injection of the non-steroidal anti-inflammatory analgesicKetoprofen (5 mg/kg, sc) and received a second injection the following day.

2.3. Intracranial self-stimulation

Operant conditioning chambers (28 cm wide � 29.4 cm deep � 68.6 cm high)were constructed from PVC (back and side walls) and Plexiglas (front wall). Eachchamber was equipped with a lever located on the left wall, 3.4 cm above the metalgrid floor. Operant conditioning chambers were encased in sound-attenuating boxes(48.6 cm wide � 50.7 cm deep � 95.4 cm high) made from melamine with a Plex-iglas window allowing constant viewing of the rat. Depressions of the lever trig-gered a constant-current generator (PHM-152/2, Med Associates Inc, St Albans, VT)to deliver a single 400-ms train of rectangular cathodal pulses of 0.1 ms in duration,delivered on a FI-1s schedule. Current intensity was monitored on an oscilloscope byreading the voltage drop across a 1 kU resistor in series with the electrode.

Following one week of post-surgical recovery, rats were trained to emit theoperant response by the method of successive approximations. Ten to 15 min oftraining were generally sufficient to establish consistent responding. Rats wereimmediately allowed to self-administer the stimulation for 1 h, at parameters set tosupport vigorous responding. On the following day, rats were allowed to self-administer the same stimulation parameters, but only during 45 s trials that wereseparated by 30 s inter-trial intervals and preceded by five trains of non-contingentstimulation delivered at 1 Hz. All parameters of the non-contingent stimulationwere identical to those available during the 45 s trial. Beginning on the third day, thepulse frequency was systematically reduced by approximately 0.1 log10 units acrosstrials, starting with a frequency that supported maximal responding and endingwith one sufficiently low to induce extinction. The plot of the rate of responding asa function of pulse frequency comprised a single responseefrequency curve; eachcurve took 15 min to complete and 6e8 were determined daily. Reward thresholdwas defined as the pulse frequency required for responding at a half-maximal rateand was derived from a regression line fit to the rising portion of individualresponseefrequency curves. In order to standardize the rewarding effectiveness ofthe stimulation across all rats prior to the start of drug testing, current intensitieswere adjusted to yield reward thresholds of approximately 50 Hz.

2.4. Drug tests

Rats were trained daily until reward thresholds varied by less than 0.1 log10 unitswithin daily sessions and across aminimum of three days. Each drug test beganwithfour threshold determinations (four responseefrequency curves), the first of whichwas excluded from data analysis. Immediately after the end of the fourth curve, ratswere removed from the operant chambers and brought to an adjacent room for druginfusion. The dust cap and obturator were first removed from the cannula assemblyand replaced with an injector (31 ga) connected to two Hamilton microsyringes viapolyethylene tubing. The injectors protruded beyond the tip of the guide cannula by1 mm. Infusions were made with a dual syringe infusion pump (Pump 11 Plus,Harvard Apparatus, Holliston, MA) over 1 min, followed by an additional minuteprior to retraction of the injector in order to allow drug diffusion away from theinjector tip. Obturators and dust caps were then reinstalled and the rats werebrought back to their respective operant conditioning chamber for six additionalthreshold determinations which required approximately 90 min of additionaltesting. At the end of the test session, rats were returned to their home cages.

Rats within each of the three groups (shell, core, LHb) received a total of threedoses of D-amphetamine plus vehicle (sterile 0.9% saline) in a semi-counterbalanced sequence. In order to restrict drug diffusion to the target site, thefirst three doses in each group were delivered in a volume of 0.25 ml. In each of thethree sites, the first three doses were delivered in a fully counterbalanced sequence.For the fourth and last infusion, we doubled the volume by increasing the infusiontime of the highest dose to 2 min. Doses were as follows: shell and core: 0, 1 and10 mg/0.25 ml plus 20 mg/0.5 ml; LHb: 0, 10 and 20 mg/0.25 ml plus 40 mg/0.5 ml. Pilotinfusions of 1 mg/0.25 ml into the LHb of a separate group of rats did not produce anyeffect so we shifted up the range of doses administered to this site in an attempt toobtain an effect. All drug infusions were separated by 3e4 days.

2.5. Histology

At the end of the study, rats were anaesthetized with urethane (1.2 g/kg, ip) anda lesion was created at the tip of the stimulation electrode by passing direct anodalcurrent (100 mA for 15 s). Rats were then perfused with 0.9% saline followed by a 10%

V. Duchesne, S.M. Boye / Neuropharmacology 70 (2013) 43e50 45

formalin solution containing 3% potassium ferrocyanide, 3% potassium ferricyanideand 0.5% trichloroacetic acid. The Prussian blue technique, based on the reaction ofthe perfusion solution with lesion-induced ferric ion deposits in the tissue sur-rounding the electrode tip, allows unambiguous identification of the stimulationsite. Brains were then harvested and kept in 30% sucrose for at least three days.Frozen brains were sliced into 40 mm slices with a cryostat and stained for Nisslsubstance with thionin. Stimulation and infusion sites were located with a lightmicroscope.

2.6. Drugs

D-Amphetamine sulphate (Sigma Aldrich Co Ltd, Dorset, UK) was dissolved insterile 0.9% saline, divided into 20 ml aliquots and frozen (�80�) until used. Doses ofD-amphetamine refer to the weight of the salt. Ketoprofen (Merial Canada Inc, BaieD’Urfé, QC) was injected subcutaneously in a volume of 0.5 ml/kg.

2.7. Statistical analysis

For each rat, the two reward threshold and asymptotic (maximal) response ratevalues obtained from the second and third curves post-infusion were averaged andsubjected to a one-way ANOVA for repeated measures (GraphPad Prism v5,GraphPad Software Inc, San Diego, CA). Significant main effects were followed bypost hoc analyses with Dunnett’s multiple comparison tests. The relation betweenthe site of infusion within the nucleus accumbens and the magnitude of the rewardenhancement produced by 20 mg/0.5 ml D-amphetamine was derived by first aver-aging reward thresholds (% baseline) obtained from the second and third curvespost-infusion and then subtracting the vehicle from the drug response for each rat.Difference scores for shell and core infusions were then evaluated with Pearson’scorrelation coefficient.

3. Results

3.1. Histology

In all, 51 rats underwent stereotaxic surgery. Of these, 43 sur-vived and 29 completed the entire study. Histological analysis

Fig. 1. Distribution of cannula and electrode tips. Injection sites within the nucleus accumbe(c) were located under light microscopic examination of thionin stained tissue. Numbers in

revealed that of those rats with threshold frequencies in the properrange, 21 had correct positioning of both cannulae (shell: n ¼ 6;core: n ¼ 8; habenula: n ¼ 7). Within the nucleus accumbens,infusion sites in the shell and core ranged from 1.0 to 1.7 mm and1.2 to 2.2 mm anterior to bregma, respectively (Fig. 1a). Within theLHb, infusion siteswere locatedwithin its medial half, 3.14e3.6mmposterior to bregma (Fig. 1b). Stimulation sites in the posteriormesencephalon were located close to the midline, in and aroundthe border of the ventral dorsal raphe nucleus and the decussationof the superior cerebellar peduncle, 7.3e8.3 mm posterior tobregma (Fig. 1c).

3.2. Infusions of D-amphetamine into the nucleus accumbens shell

Visual inspection of time course data obtained with shell in-fusions reveals that all three doses caused thresholds to decrease.This effect of D-amphetaminewas maximal at 1 mg/0.25 ml and wasevident immediately after infusion of this dose and 30 min afterinfusion of 10 mg/0.25 ml and 20 mg/0.5 ml (Fig. 2a). Thresholdsremained low for approximately 45 min in comparison to vehicle;vehicle infusions caused thresholds to rise above baseline for mostof the test session. D-amphetamine also increased asymptoticresponse rates when infused into the shell, an effect that was dose-dependent and lasted approximately 60min (Fig. 2b). Results of theone-way ANOVA on mean threshold values revealed a main effectof dose (F(3,15) ¼ 4.02, p < 0.05) and post hoc analysis indicatedthat the decrease in threshold following 1 mg/0.25 ml was sig-nificantly different from vehicle (16%) (Dunnett’s test, p < 0.05)(Fig. 2c). The one-way ANOVA on response rates similarly revealeda significant effect of dose (F(3,15) ¼ 8.36, p < 0.005), and post hocanalysis indicated a significant increase (36%) relative to vehicle

ns (a) and lateral habenula (b) as well as electrode tips in the posterior mesencephalondicate distance (mm) anterior or posterior to bregma (Paxinos and Watson, 1997).

Fig. 2. Reward thresholds and asymptotic response rates following D-amphetamine infusion into the nucleus accumbens shell. Reward thresholds (a) and response rates (b) areexpressed as a percentage of baseline and plotted as a function of time post-infusion. The boxed digit indicates the number of animals that received intra-shell infusions. Symbolsrepresent means � sem. Note the non-zero origin of the ordinate. Infusion of D-amphetamine into the shell reduced reward thresholds and increased response rates with dif-ferential magnitude and time course. The horizontal line above each abscissa indicates the two time points used to obtain mean threshold and response rate values plotted in (c)and (d), respectively. A significant reduction in reward threshold was observed with the lowest dose (1 mg/0.25 ml) whereas only the highest dose (20 mg/0.5 ml) increased responserates (*p < 0.05). Bars show mean � sem thresholds and response rates at each dose.

V. Duchesne, S.M. Boye / Neuropharmacology 70 (2013) 43e5046

following 20 mg/0.50 ml (p< 0.05) (Fig. 2d). The correlation betweeninfusion site and magnitude of drug effect on thresholds was notsignificant (r(6) ¼ �0.15, p > 0.05).

3.3. Infusions of D-amphetamine into the nucleus accumbens core

Infusion of 1 and 10 mg/0.25 ml into the nucleus accumbens coredid not lower reward thresholds. The 20 mg/0.5 ml dose produceda maximal reduction in threshold frequency of approximately 25%,an effect that was evident 30e60 min post-infusion (Fig. 3a). D-amphetamine also produced an increase in asymptotic responserates (Fig. 3b). The maximal increase was approximately 15% abovevehicle, an effect that was present for most of the test session fol-lowing 10 mg/0.25 ml and 20 mg/0.5 ml; the lowest dose of D-amphetamine (1 mg/0.25 ml) similarly elevated response rates butwith a shorter time course. The one-way ANOVA on mean rewardthresholds revealed an effect of dose (F(3, 21) ¼ 3.34, p < 0.05) andpost hoc analysis indicated that only the decrease in thresholdfollowing 20 mg/0.5 ml (23%) was significant (Dunnett’s, p < 0.05)(Fig. 3c). The one-way ANOVA on asymptotic response rates alsorevealed a significant effect of dose (F(3, 21) ¼ 4.96, p < 0.01) andpost hoc analysis indicated significance at 10 mg/0.25 ml and 20 mg/0.5 ml (Dunnett’s, p < 0.05) (Fig. 3d). The correlation betweeninfusion site and magnitude of drug effect on thresholds was notsignificant (r(8) ¼ �0.22, p > 0.05).

3.4. Infusions of D-amphetamine into the LHb

The time course of changes in reward thresholds and asymptoticresponse rates following infusion of D-amphetamine into the LHbare shown in Fig. 4. After the first threshold determination post-infusion (15 min), values tended to rise above vehicle followingtreatment with the two lowest doses (10 and 20 mg/0.25 ml),although doubling the volume of the infusion to 40 mg/0.5 ml did notincrease this effect further (Fig. 4a). At all doses, reward thresholdsremained stable throughout the test session (90 min). We did notobserve any systematic change in asymptotic response rates at any

dose (Fig. 4b). Statistical analysis carried out on mean rewardthresholds (30e45 min) with a one-way ANOVA for repeatedmeasures indicated a significant effect of treatment (F(3, 18) ¼ 3.31,p < 0.05) (Fig. 4c). Post-hoc Dunnett’s multiple comparison test didnot reveal any significant differences with respect to vehicle; thetreatment effect was due to a significant difference between the 20and 40 ug doses, neither of which had an effect on their own.Analysis of variance on mean asymptotic response rates did notreveal a main effect (F(3, 18) ¼ 0.09, p > 0.05) (Fig. 4d).

4. Discussion

Studies using ICSS have generally demonstrated enhancedrewarding effectiveness following psychostimulant drug infusioninto the nucleus accumbens (Carr et al., 2009; Colle andWise,1988;Hayes et al., 2009; Ranaldi and Beninger, 1994a), although littleattention has been paid to the relative contribution of its sub-regions. On the other hand, although LHb involvement in reward(Gomita and Gallistel, 1982) and reward error prediction(Matsumoto and Hikosaka, 2007; Ullsperger and von Cramon,2003) has been demonstrated, little is known about the func-tional contribution of its dopaminergic innervation. We combinedintracranial infusions of the DA agonist D-amphetaminewith ICSS,an animal model of appetitive behaviour, in order to examine thecontribution of these three DA terminal sites to reward functionand operant responding. We describe for the first time differencesbetween the nucleus accumbens shell and core in the minimaldoses required to enhance rewarding effectiveness and operantresponding, in the magnitude of these enhancements as well as intheir duration. Thus, reward thresholds were reduced by the lowestdose of D-amphetamine infused into the shell, but only the highestdose reliably enhanced responding. These findings suggesta greater sensitivity of reward processes than motoric capacity tochanges in DA transmission in the shell. The opposite relation wasobserved in the core. Here, asymptotic respondingwas increased byall doses, but only the highest dose reliably reduced rewardthresholds. These findings suggest a greater sensitivity of motoric

Fig. 3. Reward thresholds and asymptotic response rates following D-amphetamine infusion into the nucleus accumbens core. In the core, only the highest dose (20 mg/0.5 ml)reduced reward thresholds whereas 10 mg/0.25 ml and 20 mg/0.5 ml both increased asymptotic response rates (*p < 0.05). All details are the same as in Fig. 2.

V. Duchesne, S.M. Boye / Neuropharmacology 70 (2013) 43e50 47

capacity than reward processes to changes in DA transmission inthe core. In contrast, we show that regardless of dose, D-amphet-amine did not enhance rewarding effectiveness or operant rate ofresponding when infused into the LHb. In fact, we observeda nonsignificant attenuation of reward with the two lowest doses.The small increase in rewarding effectiveness seenwith 40 mg/0.5 mlwas probably non-specific and due to the uncommonly large doseof D-amphetamine. Our findings highlight the differential con-tribution of mesoaccumbens and mesohabenular DA to reward andperformance.

The lack of effect with LHb infusions was somewhat surprisinggiven that a number of studies have shown that habenular meta-bolic activity is highly sensitive to changes in DA transmission(McCulloch et al., 1980; Wechsler et al., 1979). More particularly forour purposes, brain stimulation reward or DA agonists such asamphetamine reduce metabolic demand whereas DA antagonists

Fig. 4. Reward thresholds and asymptotic response rates following D-amphetamine infusionamphetamine at any dose. All details are the same as in Fig. 2.

in doses that attenuate rewarding effectiveness enhance LHbmetabolic activity (Gallistel et al., 1985; Gomita and Gallistel, 1982).These initial findings suggested that the LHb may constitute a keysite for the reward-altering effects of drugs acting on DA neurons.Although these findings may reflect changes in mesohabenular DAtransmission, the non-specific nature of the 2-deoxyglucose tech-nique also allows the possibility that the observed effects were dueto changes in non-dopaminergic afferents. For instance, the LHbreceives dense GABAergic innervation from basal ganglia nucleisuch as the entopeduncular nucleus, nucleus accumbens and ven-tral pallidum (Herkenham and Nauta, 1977). These nuclei mediatekey behavioural effects of mesoaccumbens DA and may do so viatheir LHb projections. In this light, the expression of a DA trans-porter (DAT) (Freed et al., 1995) as well as DA receptors (Aizawaet al., 2012) in the LHb suggested to us that microinjections of D-amphetamine into this site would be a more direct way to test the

into the lateral habenula. Reward thresholds and response rates were not altered by D-

V. Duchesne, S.M. Boye / Neuropharmacology 70 (2013) 43e5048

contribution of mesohabenular DA to reward. However, the lack ofclear D-amphetamine effect strongly suggests that mesohabenularDA is not an important contributor to brain stimulation reward. Werecently drew the same conclusion in a study showing that neu-rotoxic lesions of LHb neurons do not alter the reward-enhancingeffect of D-amphetamine in ICSS (Gifuni et al., 2012).

Our knowledge of the neural circuitry linking the LHb to DAneurons of the ventral midbrain is still incomplete, but in general,the main consequence of habenular stimulation is near-completeinhibition of midbrain DA neuronal activity (Christoph et al.,1986; Ji and Shepard, 2007; Matsumoto and Hikosaka, 2007). Atpresent, there is no evidence of a closed feedback loop between theLHb and mesohabenular DA neurons. Direct and indirect inputsfrom the LHb innervate the caudal part of the ventral tegmentalarea whereas mesohabenular DA neurons arise from its anteriorhalf (Gruber et al., 2007). On the other hand, it is well-establishedthat rewarding electrical stimulation increases the activity ofmidbrain DA neurons (Hernandez and Shizgal, 2009; Moisan andRompré, 1998; You et al., 1998, 2001) and that pharmacologicallyincreasing or decreasing DA neurotransmission enhances and at-tenuates brain stimulation reward, respectively (Gallistel andKarras, 1984). Therefore, if mesohabenular DA provides inhibitorycontrol over LHb output neurons, then intra-habenular D-amphetamine should increase this effect and consequently dis-inhibit mesoaccumbens DA activity. In such a case, stimulation-induced reward would be enhanced. Alternatively, if meso-habenular DA excites LHb neurons, then the resulting inhibition ofmidbrain DA activity would attenuate reward. Unfortunately, nei-ther of these two configurations is obvious from our findings.Indeed, a third possibility is that even if mesohabenular DA pro-vides excitatory input to LHb neurons and indirectly inhibits mes-oaccumbens DA transmission, this effect may be countered by theelectrical stimulation. This possibility remains to be studied withsingle unit recording techniques.

Comparison of our findings in nucleus accumbens with pub-lished reports is somewhat limited. Although several studies haveexamined the effect of intra-accumbens D-amphetamine on ICSS,few have used techniques that allow measurement of rewardingeffectiveness independent from the rate of responding. Amongthose that have, most have focused on the contribution of either theshell or the core. Despite this, we can make the following obser-vations. The 16% enhancement in rewarding effectiveness that weobserved with shell infusions of the lowest dose (1 mg/0.25 ml) iscomparable to the 25% enhancement observed by Hayes et al.(2009) using twice the volume (1 mg/0.5 ml) but greater than the5% increase observed by Carr et al. (2009) using a much higher doseof 5 mg/0.5 ml. The discrepancy with the results of Carr et al. (2009)are not evident, since both of these studies determined rewardingeffectiveness over a similar time period post-infusion andemployed the same measure of rewarding effectiveness used here.Intra-shell infusion of nomifensine, a drug with similar mecha-nisms of action to those of D-amphetamine, has also been shown toenhance rewarding effectiveness (Carlezon andWise,1996a). In ourhands, infusion of 20 mg/0.5 ml into the core enhanced rewardingeffectiveness by 23%, an effect comparable to that reported by Colleand Wise (1988) and Ranaldi and Beninger (1994a,b), at compara-ble doses. Lastly, we did not observe a relation between the rostro-caudal location of infusions and the magnitude of changes inrewarding effectiveness as previously reported (Ranaldi andBeninger, 1994a). It is highly probable that the number of animalsin the present study may not have been sufficient to reveal sucha relation.

More generally, our findings are supported by converging evi-dence from studies using different animal models which suggestthat the nucleus accumbens shell plays a key role in mediating the

acute rewarding effect of psychostimulants. Thus, rats will self-administer direct and indirect DA agonists into the shell and lessso, if at all, into the core (Carlezon et al., 1995; Carlezon and Wise,1996b; Chevrette et al., 2002; Hoebel et al., 1983; Ikemoto et al.,1997; Shin et al., 2008), 6-hydroxydopamine lesions of the shellbut not the core abolish D-amphetamine- and cocaine-inducedconditioned place-preferences (Sellings et al., 2006; Sellings andClarke, 2003) and psychostimulants increase extracellular DAlevels preferentially in the shell than in the core (Pontieri et al.,1995).

Contrary to our findings, previous studies have not generallyreported an increase in operant responding following D-amphet-amine infusions. The absence of such an increase may be due todifferent rewarding schedules across studies. In this respect, it ispertinent that responding for the highest frequencies is oftenlimited by the presence of stimulation-bound behavioural con-tamination. In the present study, rewards (400 ms trains) wereadministered on a FI-1s schedule. The long delay (600 ms) imposedbetween successive rewards dissociates maximal response ratesfrom the maximal number of rewards. The uncoupling of responserate from contingent rewards reduces stimulation-bound con-tamination and can unveil increases in maximal rates (Boye andRompré, 1996, 2000). The use of a fixed time delay between thedelivery of rewards also reduces the possibility of artificially con-taminating reward threshold estimates subsequent to changes inresponse rates (Boye and Rompré, 1996). A review of the literatureindicates that with the exception of Ranaldi and Beninger (1994a,1994b) who imposed a 200 ms delay, most studies used a con-tinuous reinforcement schedule.

Comparison of changes in rewarding effectiveness and perfor-mance capacity across shell and core subregions reveals interestingdifferences. First, in both sites, changes in rewarding effectivenessappear to be all or none, whereas operant responding increaseswith dose. This finding strongly suggests that the neural substratethat evaluates the magnitude of the reward is distinct from thatwhich mediates responding for the reward. The finding that dou-bling the volume of the infusion in the shell increased operantresponding but not rewarding effectiveness, whereas the oppositewas true in the core, further suggests distinct functional substrateswith differential contributions in shell and core. Second, althoughon the basis of the minimum required dose our findings suggesta greater sensitivity of reward processes to changes in DA trans-mission in the shell, the effect is actually smaller and shorter lastingthan in the core. It is unlikely that drug diffusion to the shell un-derlies the greater effect seenwith the 0.5 ml coremicroinjections. Ifsuch were the case, one would expect similar or greater enhance-ments and a longer time course in the shell. In this regard, it mightbe pertinent that the density of the dopamine transporter (DAT) issubstantially lower in the shell than in the core (Nirenberg et al.,1997). Differential distribution of DAT, the substrate for D-amphetamine, may underlie the smaller magnitude of the effect inthe shell. Lower DAT density does not easily account for the shortertime-course, however. Paradoxically, despite lower DAT density, DAclearance in the shell is faster than in the core (David et al., 1998). Inthis respect, previous work has shown that in several brain sites thenorepinephrine transporter (NET) can take up extrasynaptic DA(Borgkvist et al., 2012; Carboni et al., 1990). The higher density ofNET in the shell (Berridge et al., 1997) versus the core suggests thatNET-DAT synergism cannot be discounted: a progressively greatercontribution of NET with increasing doses of D-amphetamine inthe shell may account for the short and similar time courses ofchanges in rewarding effectiveness. Finally, despite the fact that D-amphetamine binds to both DAT and NET, it is unlikely thatincreased norepinephrine levels contributed to the increase inrewarding effectiveness. For instance, intra-accumbens infusion of

V. Duchesne, S.M. Boye / Neuropharmacology 70 (2013) 43e50 49

L-amphetamine, which displays a higher affinity for NET than D-amphetamine, does not enhance brain stimulation reward (Colleand Wise, 1988).

Finally, although our results confirm the important role ofaccumbens DA in reward processes, they do not offer insight intothe nature of this contribution. Indeed, the observed changes inrewarding effectiveness may reflect alterations in such variables asthe sensitivity of the electrically-driven reward substrate, in themaximum possible reward sensed by the animal, in the perceivedcost of attaining the reward or in any of several other variables thatplay a role in determining the rat’s propensity to seek out thereward. These variables are not discernible with the curve-shiftparadigm used here (Edmonds and Gallistel, 1974; Miliaressiset al., 1986) and a three-dimensional model can more readilymeasure their respective contributions (Hernandez et al., 2010). Assuch, we do not allude to the underlying mechanism(s) of theobserved changes in reward thresholds and restrict our inter-pretation to an enhancement in the effectiveness of the stimulationin rewarding a given level of responding.

5. Conclusion

In summary, the present study is novel from at least threeperspectives. First, we obtained detailed time courses for severaldoses of D-amphetamine at each of three DA terminal sites.Differences between sites suggest that distinct neural substratesmediate reward function and performance capacity. Second, wedirectly tested the functional significance of mesohabenular DAto brain stimulation reward. Our findings suggest that DAtransmission within the LHb does not contribute significantly tothe circuitry that mediates the rewarding effect of electricalbrain stimulation, an important distinction from meso-accumbens DA. Third, our behavioural measures do not supportthe view of distinct roles of the shell and core in ICSS. Despitetheir different connections within limbic and motor networks,both contributed to enhancements in rewarding effectivenessand responding. What we describe here is a more subtle andpreferential sensitivity of reward processes and motoric capacityto increased DA transmission in nucleus accumbens shell andcore, respectively.

Acknowledgements

This work was supported by a research grant from the NaturalSciences and Engineering Research Council of Canada (NSERC) andby a fellowship from the Fonds de la Recherche en Santé du Québec(FRSQ) to SMB.

References

Aizawa, H., Kobayashi, M., Tanaka, S., Fukai, T., Okamoto, H., 2012. Molecularcharacterization of the subnuclei in rat habenula. J. Comp. Neurol. 520, 4051e4066.

Berridge, C.W., Stratford, T.L., Foote, S.L., Kelley, A.E., 1997. Distribution of dopaminebeta-hydroxylase-like immunoreactive fibers within the shell subregion of thenucleus accumbens. Synapse 27, 230e241.

Borgkvist, A., Malmlof, T., Feltmann, K., Lindskog, M., Schilstrom, B., 2012. Dopaminein the hippocampus is cleared by the norepinephrine transporter. Int. J. Neu-ropsychopharmacol. 15, 531e540.

Boye, S.M., Rompré, P.P., 1996. Effect of pimozide on self-stimulation thresholdunder a continuous and fixed-interval schedule of reinforcement. Behav. BrainRes. 78, 243e245.

Boye, S.M., Rompré, P.P., 2000. Behavioral evidence of depolarization block ofdopamine neurons after chronic treatment with haloperidol and clozapine.J. Neurosci. 20, 1229e1239.

Boye, S.M., Grant, R.J., Clarke, P.B., 2001. Disruption of dopaminergic neuro-transmission in nucleus accumbens core inhibits the locomotor stimulant ef-fects of nicotine and D-amphetamine in rats. Neuropharmacology 40, 792e805.

Bromberg-Martin, E.S., Hikosaka, O., 2011. Lateral habenula neurons signal errors inthe prediction of reward information. Nat. Neurosci. 14, 1209e1216.

Bunney, B.S., Walters, J.R., Roth, R.H., Aghajanian, G.K., 1973. Dopaminergic neurons:effect of antipsychotic drugs and amphetamine on single cell activity.J. Pharmacol. Exp. Ther. 185, 560e571.

Carboni, E., Tanda, G.L., Frau, R., Di, C.G., 1990. Blockade of the noradrenaline carrierincreases extracellular dopamine concentrations in the prefrontal cortex: evi-dence that dopamine is taken up in vivo by noradrenergic terminals.J. Neurochem. 55, 1067e1070.

Carlezon, W.A., Wise, R.A., 1996a. Microinjections of phencyclidine (PCP) andrelated drugs into nucleus accumbens shell potentiate medial forebrain bundlebrain stimulation reward. Psychopharmacology (Berl) 128, 413e420.

Carlezon, W.A., Wise, R.A., 1996b. Rewarding actions of phencyclidine and relateddrugs in nucleus accumbens shell and frontal cortex. J. Neurosci. 16, 3112e3122.

Carlezon, W.A., Devine, D.P., Wise, R.A., 1995. Habit-forming actions of nomifensinein nucleus accumbens. Psychopharmacology (Berl) 122, 194e197.

Carr, K.D., Cabeza, d., V., Sun, Y., Chau, L.S., 2009. Reward-potentiating effects of D-1dopamine receptor agonist and AMPAR GluR1 antagonist in nucleus accumbensshell and their modulation by food restriction. Psychopharmacology (Berl) 202,731e743.

Chevrette, J., Stellar, J.R., Hesse, G.W., Markou, A., 2002. Both the shell of the nucleusaccumbens and the central nucleus of the amygdala support amphetamine self-administration in rats. Pharmacol. Biochem. Behav. 71, 501e507.

Christoph, G.R., Leonzio, R.J., Wilcox, K.S., 1986. Stimulation of the lateral habenulainhibits dopamine-containing neurons in the substantia nigra and ventraltegmental area of the rat. J. Neurosci. 6, 613e619.

Colle, L.M., Wise, R.A., 1988. Effects of nucleus accumbens amphetamine on lateralhypothalamic brain stimulation reward. Brain Res. 459, 361e368.

David, D.J., Zahniser, N.R., Hoffer, B.J., Gerhardt, G.A., 1998. In vivo electrochemicalstudies of dopamine clearance in subregions of rat nucleus accumbens: dif-ferential properties of the core and shell. Exp. Neurol. 153, 277e286.

Edmonds, D.E., Gallistel, C.R., 1974. Parametric analysis of brain stimulation rewardin the rat: III. Effect of performance variables on the reward summation func-tion. J. Comp. Physiol. Psychol. 87, 876e883.

Freed, C., Revay, R., Vaughan, R.A., Kriek, E., Grant, S., Uhl, G.R., Kuhar, M.J., 1995. Dop-amine transporter immunoreactivity in rat brain. J. Comp. Neurol. 359, 340e349.

Gallistel, C.R., Karras, D., 1984. Pimozide and amphetamine have opposing effects onthe reward summation function. Pharmacol. Biochem. Behav. 20, 73e77.

Gallistel, C.R., Gomita, Y., Yadin, E., Campbell, K.A., 1985. Forebrain origins and ter-minations of the medial forebrain bundle metabolically activated by rewardingstimulation or by reward-blocking doses of pimozide. J. Neurosci. 5, 1246e1261.

Gifuni, A.J., Jozaghi, S., Gauthier-Lamer, A.C., Boye, S.M., 2012. Lesions of the lateralhabenula dissociate the reward-enhancing and locomotor-stimulant effects ofamphetamine. Neuropharmacology 63, 945e957.

Gomita, Y., Gallistel, C.R., 1982. Effects of reinforcement-blocking doses of pimozideon neural systems driven by rewarding stimulation of the MFB: a 14C-2-deoxyglucose analysis. Pharmacol. Biochem. Behav. 17, 841e845.

Gruber, C., Kahl, A., Lebenheim, L., Kowski, A., Dittgen, A., Veh, R.W., 2007. Dop-aminergic projections from the VTA substantially contribute to the meso-habenular pathway in the rat. Neurosci. Lett. 427, 165e170.

Hayes, D.J., Clements, R., Greenshaw, A.J., 2009. Effects of systemic and intra-nucleus accumbens 5-HT2C receptor compounds on ventral tegmentalarea self-stimulation thresholds in rats. Psychopharmacology (Berl) 203,579e588.

Heimer, L., Zahm,D.S., Churchill, L., Kalivas, P.W.,Wohltmann,C.,1991. Specificity in theprojectionpatterns of accumbal core and shell in the rat.Neuroscience 41, 89e125.

Herkenham, M., Nauta, W.J., 1977. Afferent connections of the habenular nuclei inthe rat. A horseradish peroxidase study, with a note on the fiber-of-passageproblem. J. Comp. Neurol. 173, 123e146.

Hernandez, G., Shizgal, P., 2009. Dynamic changes in dopamine tone during self-stimulation of the ventral tegmental area in rats. Behav. Brain Res. 198, 91e97.

Hernandez, G., Breton, Y.A., Conover, K., Shizgal, P., 2010. At what stage of neuralprocessing does cocaine act to boost pursuit of rewards? PLoS. One 5,e15081.

Hodos, W., Valenstein, E.S., 1962. An evaluation of response rate as a measure ofrewarding intracranial stimulation. J. Comp. Physiol. Psychol. 55, 80e84.

Hoebel, B.G., Monaco, A.P., Hernandez, L., Aulisi, E.F., Stanley, B.G., Lenard, L., 1983.Self-injection of amphetamine directly into the brain. Psychopharmacology(Berl) 81, 158e163.

Ikemoto, S., Panksepp, J., 1999. The role of nucleus accumbens dopamine in moti-vated behavior: a unifying interpretation with special reference to reward-seeking. Brain Res. Brain Res. Rev. 31, 6e41.

Ikemoto, S., Glazier, B.S., Murphy, J.M., McBride, W.J., 1997. Role of dopamine D1 andD2 receptors in the nucleus accumbens in mediating reward. J. Neurosci. 17,8580e8587.

Ikemoto, S., 2007. Dopamine reward circuitry: two projection systems from theventral midbrain to the nucleus accumbens-olfactory tubercle complex. BrainRes. Rev. 56, 27e78.

Ji, H., Shepard, P.D., 2007. Lateral habenula stimulation inhibits rat midbrain dop-amine neurons through a GABA(A) receptor-mediated mechanism. J. Neurosci.27, 6923e6930.

Matsumoto, M., Hikosaka, O., 2007. Lateral habenula as a source of negative rewardsignals in dopamine neurons. Nature 447, 1111e1115.

McCulloch, J., Savaki, H.E., Sokoloff, L., 1980. Influence of dopaminergic systems onthe lateral habenular nucleus of the rat. Brain Res. 194, 117e124.

V. Duchesne, S.M. Boye / Neuropharmacology 70 (2013) 43e5050

Miliaressis, E., Rompré, P.P., Laviolette, P., Philippe, L., Coulombe, D., 1986. The curve-shift paradigm in self-stimulation. Physiol. Behav. 37, 85e91.

Moisan, J., Rompré, P.P., 1998. Electrophysiological evidence that a subset of mid-brain dopamine neurons integrate the reward signal induced by electricalstimulation of the posterior mesencephalon. Brain Res. 786, 143e152.

Nakajima, S., Patterson, R.L., 1997. The involvement of dopamine D2 receptors, butnot D3 or D4 receptors, in the rewarding effect of brain stimulation in the rat.Brain Res. 760, 74e79.

Nirenberg, M.J., Chan, J., Pohorille, A., Vaughan, R.A., Uhl, G.R., Kuhar, M.J.,Pickel, V.M., 1997. The dopamine transporter: comparative ultrastructure ofdopaminergic axons in limbic and motor compartments of the nucleusaccumbens. J. Neurosci. 17, 6899e6907.

Paxinos, G., Watson, C., 1997. The Rat Brain in Stereotaxic Coordinates, third ed.Academic Press, San Diego, CA.

Phillipson, O.T., Griffith, A.C., 1980. The neurones of origin for the mesohabenulardopamine pathway. Brain Res. 197, 213e218.

Pontieri, F.E., Tanda, G., Di, C.G., 1995. Intravenous cocaine, morphine, andamphetamine preferentially increase extracellular dopamine in the “shell” ascompared with the “core” of the rat nucleus accumbens. Proc. Natl. Acad. Sci. U.S. A 92, 12304e12308.

Ranaldi, R., Beninger, R.J., 1994a. Rostral-caudal differences in effects of nucleusaccumbens amphetamine on VTA ICSS. Brain Res. 642, 251e258.

Ranaldi, R., Beninger, R.J., 1994b. The effects of systemic and intracerebral in-jections of D1 and D2 agonists on brain stimulation reward. Brain Res. 651,283e292.

Sellings, L.H., Clarke, P.B., 2003. Segregation of amphetamine reward and locomotorstimulation between nucleus accumbens medial shell and core. J. Neurosci. 23,6295e6303.

Sellings, L.H., McQuade, L.E., Clarke, P.B., 2006. Evidence for multiple sites within ratventral striatum mediating cocaine-conditioned place preference and loco-motor activation. J. Pharmacol. Exp. Ther. 317, 1178e1187.

Sellings, L.H., Baharnouri, G., McQuade, L.E., Clarke, P.B., 2008. Rewarding andaversive effects of nicotine are segregated within the nucleus accumbens. Eur. J.Neurosci. 28, 342e352.

Shin, R., Qin, M., Liu, Z.H., Ikemoto, S., 2008. Intracranial self-administration ofMDMA into the ventral striatum of the rat: differential roles of the nucleusaccumbens shell, core, and olfactory tubercle. Psychopharmacology (Berl) 198,261e270.

Stellar, J.R., Corbett, D., 1989. Regional neuroleptic microinjections indicate a role fornucleus accumbens in lateral hypothalamic self-stimulation reward. Brain Res.477, 126e143.

Stellar, J.R., Kelley, A.E., Corbett, D., 1983. Effects of peripheral and central dopamineblockade on lateral hypothalamic self-stimulation: evidence for both rewardand motor deficits. Pharmacol. Biochem. Behav. 18, 433e442.

Ullsperger, M., von Cramon, D.Y., 2003. Error monitoring using external feedback:specific roles of the habenular complex, the reward system, and the cingulatemotor area revealed by functional magnetic resonance imaging. J. Neurosci. 23,4308e4314.

Waraczynski, M., Salemme, J., Farral, B., 2010. Brain stimulation reward is affectedby D2 dopamine receptor manipulations in the extended amygdala but not thenucleus accumbens. Behav. Brain Res. 208, 626e635.

Wechsler, L.R., Savaki, H.E., Sokoloff, L., 1979. Effects of D- and L-amphetamine onlocal cerebral glucose utilization in the conscious rat. J. Neurochem. 32, 15e22.

You, Z.B., Tzschentke, T.M., Brodin, E., Wise, R.A., 1998. Electrical stimulation of theprefrontal cortex increases cholecystokinin, glutamate, and dopamine release inthe nucleus accumbens: an in vivo microdialysis study in freely moving rats.J. Neurosci. 18, 6492e6500.

You, Z.B., Chen, Y.Q., Wise, R.A., 2001. Dopamine and glutamate release in the nu-cleus accumbens and ventral tegmental area of rat following lateral hypotha-lamic self-stimulation. Neuroscience 107, 629e639.

Zahm, D.S., Brog, J.S., 1992. On the significance of subterritories in the “accumbens”part of the rat ventral striatum. Neuroscience 50, 751e767.