This article was originally published in a journal published byElsevier, and the attached copy is provided by Elsevier for the

author’s benefit and for the benefit of the author’s institution, fornon-commercial research and educational use including without

limitation use in instruction at your institution, sending it to specificcolleagues that you know, and providing a copy to your institution’s

administrator.

All other uses, reproduction and distribution, including withoutlimitation commercial reprints, selling or licensing copies or access,

or posting on open internet sites, your personal or institution’swebsite or repository, are prohibited. For exceptions, permission

may be sought for such use through Elsevier’s permissions site at:

http://www.elsevier.com/locate/permissionusematerial

Autho

r's

pers

onal

co

py

Neurobiology of Learning and Memory 87 (2007) 42–56

www.elsevier.com/locate/ynlme

1074-7427/$ - see front matter © 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.nlm.2006.05.004

Working memory deWcits in transgenic rats overexpressing human adenosine A2A receptors in the brain

Lydia Giménez-Llort a, Serge N. SchiVmann b, Tanja Shmidt c, Laia Canela d, Lluïsa Camón e, Monica Wassholm f, Meritxell Canals f, Anton Terasmaa f, Albert Fernández-Teruel a,

Adolf Tobeña a, Elena Popova c, Sergi Ferré g, Luigi Agnati h, Francisco Ciruela d, Emili Martínez e, Jörgen Scheel-Kruger i, Carmen Lluis d, Rafael Franco d,

Kjell Fuxe f,¤, Michael Bader c,¤

a Medical Psychology Unit, Department of Psychiatry and Forensic Medicine, School of Medicine, Institute of Neuroscience, Autonomous University of Barcelona, Barcelona, Spain

b Laboratory of Neurophysiology CP601, Université Libre de Bruxelles, Brussels, Belgiumc Max-Delbrück-Center for Molecular Medicine (MDC), D-13125 Berlin-Buch, Germany

d Institute of Biomedical Investigations of Barcelona, CSIC, IDIBAPS, Barcelona, Catalonia, Spaine Department of Biochemistry and Molecular Biology, University of Barcelona, Barcelona, Catalonia, Spain

f Department of Neuroscience, Karolinska Institute, Stockholm, Swedeng Behavioural Neuroscience Branch, National Institutes of Health, National Institute on Drug Abuse, Baltimore, MD, USA

h Department of Biomedical Sciences, University of Modena, Modena, Italyi Department of Behavioural Pharmacology, NeuroSearch A/S, Ballerup, Denmark

Received 29 March 2006; revised 23 May 2006; accepted 25 May 2006Available online 7 July 2006

Abstract

Adenosine receptors in the central nervous system have been implicated in the modulation of diVerent behavioural patterns and cogni-tive functions although the speciWc role of A2A receptor (A2AR) subtype in learning and memory is still unclear. In the present work weestablish a novel transgenic rat strain, TGR(NSEhA2A), overexpressing adenosine A2ARs mainly in the cerebral cortex, the hippocampalformation, and the cerebellum. Thereafter, we explore the relevance of this A2ARs overexpression for learning and memory function. Ani-mals were behaviourally assessed in several learning and memory tasks (6-arms radial tunnel maze, T-maze, object recognition, and sev-eral Morris water maze paradigms) and other tests for spontaneous motor activity (open Weld, hexagonal tunnel maze) and anxiety (plusmaze) as modiWcation of these behaviours may interfere with the assessment of cognitive function. Neither motor performance and emo-tional/anxious-like behaviours were altered by overexpression of A2ARs. TGR(NSEhA2A) showed normal hippocampal-dependentlearning of spatial reference memory. However, they presented working memory deWcits as detected by performance of constant errors inthe blind arms of the 6 arm radial tunnel maze, reduced recognition of a novel object and a lack of learning improvement over four trialson the same day which was not observed over consecutive days in a repeated acquisition paradigm in the Morris water maze. Given theinterdependence between adenosinic and dopaminergic function, the present results render the novel TGR(NSEhA2A) as a putative ani-mal model for the working memory deWcits and cognitive disruptions related to overstimulation of cortical A2ARs or to dopaminergicprefrontal dysfunction as seen in schizophrenic or Parkinson’s disease patients.© 2006 Elsevier Inc. All rights reserved.

Keywords: Adenosine receptors; Dopamine receptors; Transgenic rat; Memory; Prefrontal cortex; Heterodimerization

* Corresponding authors. Fax: +46 8 337941 (K. Fuxe), +49 30 9406 2110 (M. Bader).E-mail addresses: Kjell.Fuxe@neuro.ki.se (K. Fuxe), mbader@mdc-berlin.de (M. Bader).

Autho

r's

pers

onal

co

py

L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56 43

1. Introduction

Adenosine A2A receptors are thought to play a role in anumber of physiological responses and pathological condi-tions (reviewed by Moreau & Huber, 1999; Popoli et al.,2003). Their high expression in the striatum and antagonis-tic interaction with dopamine receptors are consistent witha key role for A2A receptors in motor activity control(Ferré, Fuxe, von Euler, Johansson, & Fredholm, 1992,1997; Fuxe, Ferré, Zoli, & Agnati, 1998). Previous work onmice lacking A2A receptors (Chen et al., 1999; Ledent et al.,1997) indicates an involvement of these receptors in neuro-repair mechanisms, anxiety and aggressive behaviours. Inthose mice, the lack of A2A receptors also resulted inreduced startle habituation and prepulse inhibition (Wang,Short, Ledent, Lawrence, & van den Buuse, 2003) which areimpairments in sensorimotor gating also seen in schizo-phrenia (BraV et al., 1978) and after treatment of rats withdopamine receptor agonists (Swerdlow et al., 1995). Phar-macological studies have indicated that combined A1–A2Areceptor blockade exerts facilitative eVects on spatial mem-ory performance in rats, and both receptor subtypes mightalso be involved in hippocampal long-term potentiation(Arai, Kessler, & Lynch, 1990; Kessey, Trommer, Over-street, Ji, & Mogul, 1997; Rebola et al., 2003). Furthermore,there is evidence for the involvement of A2A receptors instriatal long-term potentiation (d’Alcantara, Ledent, Swil-lens, & SchiVmann, 2001). However, the speciWc role of A2Areceptors in learning and memory is still unclear, with A1receptors apparently having a more dominant role (Hauber& Bareiss, 2001; Moreau & Huber, 1999; Suzuki et al.,1993).

Short-term or ‘working’ memory is a cognitive processrelevant for keeping track of important information andideas that becomes critical for human reasoning andjudgement. Working memory depends on the integrity ofprefrontal function although hippocampus, inferior parie-tal cortex, caudate nucleus and dorsomedial nucleus ofthe thalamus are brain areas also known to play a rele-vant role in its neural circuitry (reviewed by Castner,Goldman-Rakic, & Williams, 2004). At the neuropsychi-atric level, working memory is the core and most consis-tently observed cognitive deWcit exhibited by patients withschizophrenia (Park & Holzman, 1992). Prefrontal dys-function linked to altered dopaminergic and glutamater-gic transmission or to disruption of mesencephalocorticalpathways is also suggested to cause working memory deW-cits in Parkinson’s disease patients (García-Munoz,Young, & Groves, 1991). Studies in both human and sub-human primates have also identiWed the prefrontal cortexas a key node in the functional neural networks of work-ing memory and its deWcits (Goldman-Rakic, 1994). Inrodents, experimentally induced models of prefrontal dys-function including amphetamine sensitisation, subchronicphencyclidine and neurodevelopmental insults exhibitspatial and object working memory deWcits in delayedresponse procedures or within-day learning in several

kind of mazes (Morris water maze, T-maze, Y-maze,Radial maze, etc), and social and object recognition tests.where working memory can be distinguished from long-term or ‘reference’ memory associated with day-to-daylearning (reviewed by Castner et al., 2004).

It is of substantial interest that stimulation of A2A recep-tors, like dopamine D1 receptors, increases cAMP signal-ling and neuronal excitability, especially in view of the factthat not only reductions but also increases in D1 receptor-mediated dopaminergic transmission lead to working mem-ory deWcits (Goldman-Rakic, Muly, & Williams, 2000).Therefore, in order to study the consequences of increasedcentral A2A receptor signalling, we established a noveltransgenic rat model TGR(NSEhA2A) overexpressing thisreceptor. Preferential overexpression of A2A receptors waslocated in some of these neuroanatomical substrates, wherealso dysregulation of dopamine signalling is associatedwith working memory impairments and schizophrenia.Thereafter, we submitted the animals to a behavioural bat-tery to ascertain motor activity, anxiety, and learning–memory functions. The results revealed a marked neuronalA2A overexpression especially in cortical regions that maybe critical for the working memory deWcits observed. Inview of the existence of A2A/D2 and A2A/mGlu5 heteromericcomplexes in the striatum (Canals et al., 2003; Ferré et al.,2002; Fuxe et al., 2005; Hillion et al., 2002) also D2 andmGlu5 proteins were analysed in immunoblotting experi-ments on the striatum of the TGR(NSEhA2A) over-expressing the A2A receptors to further understand theneurobiological basis for the working memory deWcitsobserved.

2. Methods and materials

2.1. Generation of transgenic animals

Transgenic rats for the overexpression of A2A receptors in the cen-tral nervous system were generated by microinjection of a DNA con-struct into the male pronucleus of Sprague–Dawley rat zygotes withestablished methods (Popova, Krivokharchenko, Ganten, & Bader,2002). The construct contained a full-length human A2A cDNA clonedinto an expression vector 3� of the 1.8 kb rat neuron-speciWc enolase(NSE) promoter and 5� of an intron/polyadenylation cassette of SV40virus (see Fig. 1A).

2.2. Genotyping of rats

Transgenic rats were identiWed by PCR (30 cycles, 54 °C annealingtemperature) on their genomic DNA isolated from tail biopsies by the useof the following transgene-speciWc primers: SV40ipa5: 5�-GAAGGAACCTTACTTCTGTGG-3� and SV40ipa3: 5�-TCTTGTATAGCAGTGCAGC-3� (see Fig. 1A).

2.3. RNase protection assay

Total RNA was isolated from rat tissues by TRIZOL reagent (Invitro-gen) according to the manufacturer’s instructions. With this RNA, trans-gene expression was measured by RNase protection assay (RPA) asdescribed previously (Silva et al., 2000) using a commercially available kit(RPAII, Ambion) and a probe covering the 5�-terminal 125 nucleotides ofthe transgenic mRNA (see Fig. 1A).

Autho

r's

pers

onal

co

py

44 L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56

2.4. In situ hybridization

The in situ hybridization technique was adapted from previouslydescribed methods (SchiVmann & Vanderhaeghen, 1993). The sectionsmounted on RNAse free poly-L-lysine coated slides were Wxed in freshlyprepared 4% paraformaldehyde solution for 30 min and rinsed in 0.1MPBS. All sections were dehydrated and dipped for 3 min in chloroform.After air drying, the sections were incubated overnight at 42 °C with0.35 £ 106 cpm per section of 35S-labelled probes diluted in hybridiza-tion buVer, which consisted of 50% formamide, 4£ SSC (1£ SSC:0.15 M NaCl, and 0.015 M sodium citrate, pH 7.4), 1£ Denhardt’s solu-tion (0.02% each of polyvinylpyrrolidone, bovine serum albumin,Wcoll), 1% sarcosyl, 0.02 M sodium phosphate at pH 7.4, 10% dextransulfate, yeast tRNA at 500 �g/ml, salmon sperm DNA at 100 �g/ml, and60 mM dithiothreitol. Compounds were provided by Sigma Chemicals(Belgium). After hybridization, the sections were rinsed for 4 £ 15 minin 1£ SSC at 55 °C, dehydrated and covered with HyperWlm-� max Wlm(Amersham, Belgium) for two or three weeks. The oligonucleotideprobes were synthesized on an Applied Biosystems 381A DNA synthe-sizer or Eurogentec (Belgium) with a GC to AT ratio between 45 and65%. The human A2A receptor oligonucleotide probe (CAGCCCTGGGAGTGGTTCTTGCCCTCCTTT-GGCTGACCGCA) is comple-mentary to nucleotides 123–166 in a partial human cDNA sequence(Libert et al., 1989) and has been previously successfully used on humanbrain sections (SchiVmann, Libert, Vassart, & Vanderhaeghen, 1991).The rat A2A receptor oligonucleotide probe (CCGCTCCCCTGGCAGGGGCTGGCTCTCCATCTGCTTCAGCTG) is complementaryto nucleotides 604–645 of the rat cDNA sequence (Fink et al., 1992).Oligonucleotides were labelled with �-35S dATP (DuPont-NEN, Bel-gium) at their 3� end by terminal DNA deoxynucleotidylexotransferase(Gibco, Belgium) and puriWed with a G50 column (Pharmacia, Belgium)according to the manufacturer’s instructions. QuantiWcation wasperformed using the public domain NIH image J 1.33 program

(National Institute of Health, USA) on digitalized images capturedfrom autoradiograms for areas displaying moderate to high expressionlevels (striatum, frontal and prefrontal cerebral cortex, thalamus,Ammon’s horn of the hippocampus, dentate gyrus, and cerebellum).For each area, analysis was performed for 3 rats of each genotype (3slices per area and animal) and optical density of the background wassubtracted.

2.5. Membrane preparation and radioligand binding

Tissue pieces were homogenized by sonication in 20 volumes of ice-cold preparation buVer (PB, 20 mM Tris–HCl, 7 mM MgCl2, and 1 mMEDTA, pH 7.4). Membranes were collected by centrifugation at 46,000gfor 20 min at 4 °C. The pellet was then resuspended in PB and centri-fuged twice as before. The Wnal pellet was homogenized in incubationbuVer 1 (20 mM Tris–HCl, 100 mM NaCl, 7 mM MgCl2, 1 mM EDTA,and 1 mM DTT, pH 7.4) containing an EDTA free protease inhibitorcocktail (Roche Diagnostics GmbH, Mannheim, Germany) and theobtained suspension was used directly for radioligand binding experi-ments. The protein concentration of the samples was determined usinga BCA protein assay kit (Pierce Biotechnology, Rockford, IL) andusing BSA dilutions as a standard. Binding of the A2A antagonist radi-oligand [3H]ZM241385 (17 Ci/mmol; Tocris) was performed in incuba-tion buVer 2 (20 mM Tris, 100 mM NaCl, 5 mM MgCl2, and 1 mMDTT, pH 7.5). Membranes were incubated with diVerent concentrationsof [3H]ZM241385 (0.02–10 nM) for 90 minutes at 30 °C in a total vol-ume of 500 �l. The reaction was terminated by rapid Wltration throughglass-Wber Wlters (GF/B, Whatman Int. Ltd., Maidstone, UK) and theWlters were washed three times with 5 ml of ice-cold buVer containing20 mM Tris and 100 mM NaCl (pH 7.4). Non-speciWc binding wasdeWned as the binding in presence of 10 �M SCH58261,a potent andselective A2A antagonist with 50-fold selectivity for A2A vs. A1 recep-tors in the rat (Zocchi et al., 1996).

Fig. 1. Generation of the transgenic rats, TGR(NSEhA2A). (A) DNA construct used to generate transgenic rats. A full-length human A2A receptor(hA2A) cDNA was cloned into an expression vector 3� of the 1.8 kb rat NSE promoter and 5� of an intron/polyadenylation cassette of SV40 virus(SV40intron/pA). The probe used for RPA is shown and the primers used for genotyping are marked by arrow heads. (B) Genotyping of transgenic rats.PCR with the primers shown in (A) revealed the presence of the transgene in the genome of positive animals (+) but not in wild-type (WT) littermates (¡).�: DNA marker �X174/HaeIII. (C) RNase protection assay (RPA) detecting human A2A receptor (hA2A) mRNA only in brain and testis of transgenicrats (TGR(NSEhA2A)) but not in other tissues or in wild-type Sprague–Dawley animals (WT). Y¡, yeast RNA without RNase treatment, Y+, yeastRNA with RNase treatment.

Autho

r's

pers

onal

co

py

L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56 45

2.6. Radioligand binding and autoradiography

Adenosine A2A receptor binding autoradiography was performed aspreviously described (Ledent et al., 1997) using a single concentration ofthe radioligand. The gelatine-coated slides, stored at ¡20 °C until use, werebrought to room temperature 30 min before the experiments. The sectionswere incubated for 90 min at room temperature in a buVer containing50 mM Tris–HCl, pH 7.4, 10 mM MgCl2, 2.5 UI/ml adenosine deaminase(ADA; Roche, Belgium),and 10 nM of the prototypical A2A agonist[3H]CGS21680 (47.0 Ci/mmol; Dupont NEN, Belgium) with 140-foldselectivity for A2A vs. A1 receptors (Hutchison et al., 1989; Jarvis et al.,1989). Non-speciWc binding of the 3H-labelled ligand was assessed by theaddition of 1 �M CGS21680. Slides were washed four times for 5 min inice-cold 50 mM Tris–HCl, pH 7.4 buVer, dipped in ice-cold distilled water,dried under a stream of cold air, and exposed to 3H-HyperWlm (Amer-sham, Belgium) for 6 weeks.

2.7. Western blot and immunohistochemistry

2.7.1. Western blottingPrefrontal cortex (rostral to Bregma 2.7 mm), dorsal (mainly frontopa-

rietal) cortex or striatum from both hemispheres from three wild-type ratsand three transgenic rats were homogenized in 10 volumes of Tris–HCl(50 mM, pH 7.4) with a protease inhibitor cocktail (Sigma–Aldrich, CO,St. Louis, MS) and centrifuged at 109,000g for 45 min at 4 °C to removedebris. The pellet was washed once more as described above and resus-pended in the same buVer solution for protein determination by BCA(Pierce Biotechnology, Rockford, IL). Proteins were solubilized in 2£ ureabuVer (urea 8 M, SDS 2%, DTT 200 mM) and incubated at 37 °C for 2 h.Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS/PAGE)was performed using 10% or 6.5% (only for detection of mGlu5) polyacryl-amide gels. Proteins were immunoblotted to PVDF membranes (Immobi-lon-P; Millipore) using a semidry transfer system and developed with theenhanced SuperSignal chemiluminescence detection kit (Pierce Biotech-nology). Receptors were detected with the following primary antibodies:rabbit anti-adenosine A1 receptor (1:1000; ABR, Golden, CO), mouseanti-A2A receptor (1:1000; UpState, Lake Placid, NY) rabbit anti-gluta-mate mGlu5 receptor (0.5 �g/mL; UpState), rat anti-dopamine D1 receptor(1:500; Sigma–Aldrich, CO), rabbit anti-dopamine D2 receptor (1 �g/mL;Chemicon, Temecula, CA) and mouse anti-�-tubulin (1:6000; Sigma–Aldrich, CO) and the following horseradish peroxidase secondary anti-bodies: goat anti-rat IgG, goat anti-rabbit IgG (1:60000; Pierce Biotech-nology) and rabbit anti-mouse IgG (1:2000; Dako A/S, Denmark). ThespeciWcity of the monoclonal anti-A2AR antibody has been previouslydescribed (Rosin, Robeva, Woodard, Guyenet, & Linden, 1998). Amongother species and as indicated by the supplier (Upstate catalogue number05-717), the antibody recognizes human and rat A2AR. In the absence ofthe primary D1 antibody and at 1:60000 dilution of the goat anti-rat IgGsecondary antibody we do not observe any band in the immunoblotting.

2.7.2. ImmunohistochemistryAt the end of the behavioural experiments, wild-type or

TGR(NSEhA2A) were deeply anaesthetized with sodium pentobarbitaland perfused transcardially with heparinized 0.9% saline followed by 4%paraformaldehyde (w/v) in 0.01 M phosphate buVer saline (PBS), pH 7.4.The brains were removed, postWxed in the same Wxative at 4 °C and, after adehydration process, embedded in paraYn. Coronal sections (5 �m)through prefrontal cortex and striatum at several rostro-caudal levels werecut and processed with monoclonal antibodies against A2A receptors(Rosin et al., 1998; Upstate, Lake Placid; see above) using the avidin-bio-tin-peroxidase technique. BrieXy, endogenous peroxidase activity wasblocked with 0.3% H2O2 and 30% methanol in PBS and unspeciWc proteinbinding sites were blocked with 3% normal goat serum in 0.01 M PBS con-taining 0.2% gelatin and 0.2% Triton X-100. Sections were then incubatedwith a mouse monoclonal antibody against the A2A receptor (Upstate, 05-717), diluted 1:600 with 1% goat serum and 0.1% Triton X-100 for 24 h at4 °C in a humidiWed chamber. Biotinylated goat anti-mouse IgG (1:200;Vector Labs, USA), as the secondary antibody, was applied for 1 h at

room temperature followed by avidin–biotin–horseradish peroxidase com-plex (1:100; Vectastain Elite ABC kit, Vector Labs.). Peroxidase labellingwas visualized with a solution of 3,3�-diaminobenzidine containing 0.03%H2O2. In each experiment, a tissue section was also processed in parallelwithout the primary antibody as a control for non-speciWc staining.

2.8. Behavioural testing

Adult male TGR(NSEhA2A) (n D 11) and WT rats (n D 14, the Spra-gue–Dawley strain used to generate the transgenic rats) were maintainedin groups of two or three (Macrolon, 57£ 35 £ 19 cm) under standard lab-oratory conditions (food and water ad lib, 22§ 2 °C, 60 § 10% relativehumidity, 12 h light/dark cycles beginning at 7:00 h). The animals weresubjected to a series of tests used to screen for behavioural abnormalitiesin mutant rodents (Giménez-Llort et al., 2002) assessing spontaneousmotor activity, learning and memory and anxiety. Behaviour was evalu-ated by both direct observation and a video-computerized tracking system(SMART, Panlab S.A., Barcelona, Spain) by two independent observersunaware of the animal’s genotype. Animals were weighed before each test.A few animals of each group [two TGR(NSEhA2A) and one WT rat] werenot assessed in the tunnel maze tests because their body size was not suit-able. The experiments were performed from 10:00 to 13:00 h under dimwhite light and in accordance with the Spanish legislation on “Protectionof Animals Used for Experimental and Other ScientiWc Purposes” and theEuropean Communities Council Directive (86/609/EEC) on this subject.

2.8.1. Hexagonal tunnel mazeThe tunnel maze (extension 1.4 m, tunnels 8 cm wide and 9 cm high) con-

sisted of two concentrical long angled alleys interconnected to a central area(Bättig, 1983; Fitzgerald, Berres, & Schaeppi, 1988). Forty-two infrared pho-tocell units were uniformly distributed throughout the tunnels and interfacedto an IBM XT computer to measure the locomotor activity (total number ofphotobeams interrupted). Animals were introduced into the maze by a ceilingdoor at the centre and 5 min later the trial was automatically terminated.

2.8.2. 6-arm radial tunnel maze testBarriers in the hexagonal tunnel maze conWgured a 6-arm radial maze.

Each arm lead to a T-shaped choice-point with a blind alley on the right anda long angled alley on the left. Entries into each of the 6 radial arms and tothe short, blind alley were measured during 5 min by means of the infraredphotocell units described above. This procedure (Bättig, 1983; Fitzgeraldet al., 1988) was repeated during 3 consecutive days. From each session, thefollowing behavioural parameters were obtained from individual animals:total number of photobeams interrupted (locomotor activity); the number ofdiVerent arms explored; the number of short, blind alleys explored; the num-ber of re-entries (repetitions) into arms before all six arms had been visited;the number of re-entries (repetitions) into short, blind alleys already explored.

2.8.3. T-maze testThe apparatus was an enclosed T-maze (woodwork; walls: 30 cm

high; starting box: 20 £ 15 cm; stem: 20 £ 15 cm; arms: 30 £ 15 cm). Therats were given three sessions, one per day with three trials each. Eachtrial involved one forced choice and one free choice. with a 15 s intratrialinterval and a 75–90 s intertrial interval. In the forced choice, only one ofthe arms according to a random order and contrabalanced in eachgroup, was accessible. The rat was placed in the starting box and 15 slater it was allowed to explore the maze. After spending 30 s in the arm,the rat was put back into the starting box. Fifteen seconds later, the ratwas again allowed to explore the maze in a free choice trial where botharms were accessible. The arm chosen by the rats during the free choicewas recorded. The olfactory trails were removed by cleaning the surfaceof the maze during the intratrial and intertrial intervals. This task(Olton, Becker, & Handelmann, 1979; Squire, 1969) was based on thespontaneous tendency of rats to explore places; no reinforcer was used.

2.8.4. Open-Weld testRats were placed in a corner of the open-Weld apparatus (wood-work

white box, 32 £ 32 £ 30 cm) divided into 25 equal sectors. Exploratory

Autho

r's

pers

onal

co

py

46 L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56

behaviour was measured as the number of crossings and rearings for a5 min recording period.

2.8.5. Object recognition testThe animals were assessed twice (1 week apart) for their ability to rec-

ognize a familiar object (S, sample) from a new one (N). Each of these ses-sions consisted of a “sample trial” followed by a “test trial.” In the sampletrial, the animals were placed in the open-Weld (a known environment) andlet explore (nose directed to the object not less than 2 cm) two identicalobjects, S1 and S2 (rectangular aluminium cans, 15£ 12 £ 5 cm), equallyspaced in the Xoor of the apparatus until they reached the criteria of explo-ration of both for 20 s. The time required to reach the criteria was named“time 1.” One minute later, animals were reintroduced in the apparatus for5 min (test trial) where two diVerent non-explored objects were located: anidentical copy of the sample objects (S3) and a completely new object (N,cylindrical aluminium can, 15£ 10 cm, diameter 10 cm). Preference for thenew object was measured by means of the index TN¡TS/TN+TS (Ennac-eur & Delacour, 1988; Willig et al., 1987) where TS and TN are the timespent exploring “S3” and “N”, respectively.

2.8.6. Morris water maze testsThe tests (Morris, 1984) consisted of three sessions of place learning

for reference memory (days 1–3), one session of cue learning (day 4) and 3sessions of “repeated acquisition paradigm” (Whishaw, 1985) for workingmemory (days 5–7). The circular pool (diameter: 140 cm, height: 60 cm)was Wlled to a depth of 29 cm with 24 °C water and rendered opaque withmilk. In all the tasks, four trials were administered per day. The rats weregently lowered into the water with the head facing the wall from one ran-domly selected cardinal starting point and allowed to swim until theylocated the platform (16 cm diameter, 28 cm height). Several external cueswere visible from the pool. Rats failing to Wnd the platform within 60 swere placed on it for 20 s, the same period as the successful animals. Thedistance travelled by the animal during the tests was measured with thevideo-computerized tracking system which also enabled the calculation ofthe average swimming speed. Learning improvement over four trials of thesame day was measured by the index “trial n/trial n ¡ 1” deWned as theratio of distance covered on trial “n” divided by the distance covered ontrial “n¡ 1” of a same day. In the similar way, the index ‘trial 1 /trial n’ forperformances between the Wrst and each of the following trials on the sameday was also considered.

In the Place learning task the platform was submerged 2 cm in a Wxedposition (north-west quadrant and 18 cm away from the wall) and the fourtrials administered per day were spaced 15 min apart. In the cue learningthe platform was placed in the east, 18 cm away from the wall and was ele-vated 1 cm above the water level with its position indicated by a visible Xag(10 cm high, 5 £ 8 cm black and white striped panel) and two black panelspreventing subjects using most of the extra-maze cues. The ‘repeated acqui-sition paradigm’ for working memory (Whishaw, 1985) consisted of 4 con-secutive trials per day during 3 consecutive days. The animals were placedat the same starting point in all the 4 consecutive trials of one session andallowed to swim until they located the platform submerged in a Wxed posi-tion which was randomly changed every day. Fixed room cues were con-stantly visible of all these from the pool.

2.8.7. Elevated plus maze testThe plus-maze (woodwork) consisted of two enclosed arms (EA,

50£ 10£ 40 cm) and two open arms (OA, 50 £ 10 cm) forming a squarecross with an open 10£ 10 cm square centrepiece, the whole being set50 cm above the ground. The animal was placed in the centre of the plus-maze facing one of the open arms. The number of entries (all four paws)into, the time spent, and the distance covered in each arm were recordedfor 5 min.

2.9. Statistics

Results are expressed as means§ SEM. DiVerences between genotypes,within days and genotype£ day interaction eVects in the 6-arms radialtunnel maze, the T-maze and the Morris water maze were analysed by

two-way ANOVA. SpeciWc diVerences between the two genotypes wereanalysed with Student’s t-test comparisons, and diVerences between twosessions of the same test with paired t-test. In all cases, statistical signiW-cance was considered at P < 0.05. All radioligand binding data were ana-lysed using GraphPad Prism 3.0 (GraphPad, San Diego, CA, USA).Statistical analysis was performed using a nonparametric one wayANOVA followed by Newman–Keuls multiple comparison tests.

3. Results

3.1. Generation of transgenic rats

The DNA construct containing the human A2A cDNAunder the control of the 1.8 kb rat (NSE) promoter(Fig. 1A) (Forss-Petter et al., 1990) was Wrst tested in a neu-roglioma cell line for functionality and shown to expressA2A mRNA (data not shown). Then it was used to generatetransgenic rats by microinjection into the male pronucleusof Sprague–Dawley rat zygotes (Popova et al., 2002).Transgene speciWc PCR conWrmed the successful integra-tion of the construct into the genomic DNA of the resultingrat strain, TGR(NSEhA2A) (Fig. 1B).

3.2. Transgene expression in TGR(NSEhA2A)

RNase protection assay (RPA). Transgene expressionwas analysed by RPA in TGR(NSEhA2A) rats (Fig. 1C).Transgenic A2A receptor mRNA was only detected in brainand testis of the transgenic rats but in no other organ testedproving the speciWcity of the promoter.

In situ hybridization. Radiolabelled oligonucleotideprobes speciWc for the human and rat A2A receptor cDNAwere used for in situ hybridization on adjacent brain sec-tions. As shown in the Fig. 2, human A2A receptor mRNAwas detected in the cerebral cortex, the hippocampal for-mation (hippocampus and dentate gyrus), the cerebellum,and certain thalamic nuclei and, at a lower level in the brainstem. However, only a very low level of human A2A recep-tor mRNA expression was detected in the striatum. As alsoseen in Fig. 2, there is a certain cross hybridization of therat probe to the human A2A receptor mRNA conWrming aclear-cut overproduction of A2A receptor mRNA in theseextrastriatal brain regions.

3.2.1. Western blotDespite the fact that human mRNA was hardly detected

in striatum, the level of striatal A2A receptor immunoreac-tivity detected by Western blot was signiWcantly increasedin TGR(NSEhA2A) (Fig. 3A). Immunoblots also indicatedthat the level of A2A receptor expression in both frontopari-etal and prefrontal cortex of transgenic animals was signiW-cantly and substantially higher than that found in WTanimals, which showed a very low level of receptor expres-sion in these regions (Fig. 3A). As seen in this Wgure there isa marked increase in A2A receptor monomers and dimers inthe frontoparietal cortex and the prefrontal cortex of theTGR(NSEhA2A) vs. WT rats. In the transgenic animals, itseems that the formation of A2A receptor dimers is

Autho

r's

pers

onal

co

py

L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56 47

favoured since they were not found in the striatum of con-trol rats. Alternatively the A2A antibody may not recognizethe rat A2A receptor dimers but this seems less likely.

3.2.2. ImmunohistochemistryThe localization of A2A receptor in prefrontal cortex and

striatum of WT and TGR(NSEhA2A) was further studiedby means of immunohistochemistry. The localization ofA2A receptor observed in striatum and prefrontal areas ofWT animals is in agreement with that described by Rosinet al. (1998). Prominent labeling was observed in the neuro-pil of the caudate-putamen in both WT andTGR(NSEhA2A) but the intensity of labeling was similar(Fig. 4). In contrast, TGR(NSEhA2A) showed strong label-ling of many pyramidal and non-pyramidal nerve cellbodies in some areas of prefrontal cortex, mainly cingulatecortex, and medial and lateral orbital cortex and also in thehippocampal formation, where the labelling in WT animalswas weak (Fig. 4). No staining was found after the omissionof the primary A2A receptor antibody.

3.3. A2A receptor binding in TGR(NSEhA2A)

Biochemical radioligand binding. Using an A2A receptorantagonist radioligand ([3H]ZM241385) a clear-cut increasein the density of A2A receptor antagonist binding sites inthe prefrontal cortex was detected in TGR(NSEhA2A)while very few A2A receptors were found in WT rats in thisarea (Fig. 5B). No changes in the KD values of A2A receptorantagonist binding sites were found in theTGR(NSEhA2A) vs. the WT rats (Fig. 5B, legend). In thestriatum only a small but signiWcant increase in the density

of A2A receptors was found in the TGR(NSEhA2A) withno change in the KD values of the A2A receptor antagonistbinding sites (Fig. 5A, legend).

3.3.1. Receptor autoradiographyUsing the A2A receptor agonist radioligand [3H]CGS

21680 and a single point analysis, a higher speciWc bindingwas observed in extrastriatal areas such as the cerebral cor-tex, the cerebellum and the thalamus in theTGR(NSEhA2A) vs. WT rats (Fig. 2F–G).

3.4. Expression of adenosine A1, dopamine D1 and D2 glutamate mGlu5 receptors

Detection of adenosine A1, dopamine D1, and D2, andglutamate mGlu5 receptors in samples from cortex (fronto-parietal or prefrontal) and striatum was performed byimmunoblotting. The expression of these receptors did notchange in the cerebral cortex of TGR(NSEhA2A) whencompared to the levels found in WT animals (Fig. 3B). Instriatum however D2 and mGlu5 receptors were signiW-cantly less abundant in TGR(NSEhA2A) (Fig. 3B). Theexpression of A1 and D1 receptors did not signiWcantlychange in the striatum of transgenic animals with respect towild-type rats.

3.5. Behavioural testing of TGR(NSEA2A)

Table 1 summarizes the results of behavioural assess-ment of TGR(NSEA2A) and the WT group. Both groupsof animals displayed similar spontaneous activity scores inthe two modalities of the tunnel mazes (total activity counts

Fig. 2. Localisation of human A2A receptor mRNA expression and [3H]CGS21680 binding. In situ hybridization autoradiograms using probes comple-mentary to the human A2A receptor cDNA (A and B) or to the rat A2A receptor cDNA (D and E) and quantiWcation (C) show human A2A receptormRNA expression in several brain areas including the cerebral cortex, the hippocampus and the cerebellum of TGR(NSEhA2A) (B and E) as comparedto wild-type rats (A and D). [3H]CGS21680 binding autoradiograms in control (F) and transgenic (G and H) rats show a substantial increase in bindingdensity in these extrastriatal regions (total (F and G) and non-speciWc (H) binding).

Autho

r's

pers

onal

co

py

48 L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56

in the hexagonal maze and in the 6-arms radial maze) aswell as in the open-Weld test (crossings and rearings). Anxi-ety-like behaviour measured in the plus maze was similar inboth groups of rats with also equal number of entries inboth arms.

Upon initial testing in the 6-arms radial tunnel maze, nogenotype diVerences were found in their ability to exploreall the 6 arms, in the latency to explore and in the numberof reentries. In both groups of rats repeated testing resultedin a decrease of latency and reentries to explored arms overthe sessions of the test revealed by a “day” eVect[F (2, 40)D9.07, P < .001 and F (2,40)D 5.19, P < .01, respec-tively]. ‘Day’ and ‘genotype£day’ interaction eVects werefound in the performances in the short-blind arms, both inthe number of visits to these alleys [Fig. 6A, day eVect:F (2,40)D 6.25, P < .01; genotype£day eVect:F (2,40)D 6.44, P < .01] and the re-entries into them[Fig. 6B, day eVect: F (2, 40)D 4.76, P < .05; genotype£dayeVect: F (2, 40)D5.58, P < .01] with TGR(NSEhA2A) mak-ing more visits and errors on the Wrst and second days thanthe other group (Student’s t-test, P < .05).

In the T-maze, the same number of correct choices wasmade by animals of both genotypes and a ‘day’ eVect[F (2, 46)D5.90, P < .01] showed an improvement over thesessions in both groups of rats.

When assessed in the object recognition test equalexploratory behaviour was shown during the sample trialas both groups of rats required the same time to reach crite-ria (time 1). In the test trials, the ability to recognize thenew object as compared to the sample (index) was lower inTGR(NSEhA2A) than in WT rats (Student’s t-test,P < .05).

Analyses of swimming speed over the diVerent sessionsand paradigms in the Morris water maze indicated nodiVerences between genotypes. On the Wrst task for spatialreference memory, the two groups of rats showed similaracquisition patterns with distances (averaged for the fourtrials of each session) decreasing at the same rates over thethree sessions of the task (Fig. 7A, ‘day’ eVect:F (2,46)D 2.46 P < .001). Additionally, both groups of ratsdisplayed a maximum eYciency to reach the platform in thesingle cue learning task (Fig. 7A). On the following days

Fig. 3. Western blot for A2A, A1, D1, D2, and mGlu5 receptors. Expression pattern of (A) A2A and (B) D1, D2, A1, and mGlu5 receptors in striatal, cortexand prefrontal cortex membranes of wild-type (WT) and TGR(NSEhA2A) (A2A) rats. Tissue membranes were analysed by Western blot as indicated inSection 2. In A, 20 �g of protein were applied in each lane and blots were treated with mouse anti-�-tubulin to conWrm equal loading of protein. In B, 15 �g(mGlu5), 20 �g (A1 and D2), and 40 �g (D1) of protein were applied in each lane and blots were also treated with mouse anti-�-tubulin (data not shown)conWrming equal loading. The position of the molecular mass markers in kDa are indicated on the left. A representative image and the quantiWcation (barcharts) corresponding to the average data of four independent experiments are shown in the lower panels. The intensity of the immunoreactive bands wasmeasured by densitometric scanning and the results are presented as % of the immunoreactivity detected in striatum of wild-type rats (means § SEM).*P < 0.05, **P < 0.01 compared with wild-type of the same speciWc tissue.

A2A receptor D2 receptorD1 receptor

PrefrontalStriatum

WT A2A

Cortex

WT A2A WT A2A

80

A2A monomer

MW(kDa)

100

40

120

A2A dimer

50

Tubulin

PrefrontalStriatum

WT A2A

Cortex

WT A2A WT A2A

80 D2 dimer

MW(kDa)

60

50

40

D2 monomer

100

PrefrontalStriatum

WT A2A

Cortex

WT A2A WT A2A100

D1

MW(kDa)

80

60

50

Multimericform

0%

20%

40%

60%

80%

100%

*

WTS A2AS WTC A2AC WTPF A2APF

% o

fW

T s

tria

tum

0%

20%

40%

60%

80%

100%

*

WTS A2AS WTC A2AC WTPF A2APF

% o

fW

T s

tria

tum

0%

20%

40%

60%

80%

100%

WTS A2AS WTC A2AC WTPF A2APF

% o

fWT

str

iatu

m

0%

20%

40%

60%

80%

100%

120%

140%

160%

** *

*

WTS A2AS WTC A2AC WTPF A2APF

% o

fW

T s

tria

tum

0%

20%

40%

60%

80%

100%

120%

140%

160%

** *

*

WTS A2AS WTC A2AC WTPF A2APF

% o

fW

T s

tria

tum A1 receptor

mGlu5 receptor

PrefrontalStriatum

WT A2A

Cortex

WT A2A WT A2A

35

A1 monomer

MW(kDa)

30

A1 dimer

75

PrefrontalStriatum

WT A2A

Cortex

WT A2A WT A2A

MW(kDa)

160mGluR5 monomer

Multimeric form

0%

20%

40%

60%

80%

100%

120%

140%

160%

180%

WTS A2AS WTC A2AC WTPF A2APF

% o

fWT

str

iatu

m

0%

20%

40%

60%

80%

100%

*

WTS A2AS WTC A2AC WTPF A2APF

% o

fWT

str

iatu

m

A B

Autho

r's

pers

onal

co

py

L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56 49

5–7, when animals where assessed in a series of place tasksfor working memory, the averaged performance of controlrats to Wnd the new platform positions on each of thesedays was as good as that achieved on the previous placetask (Paired t-test, day 5, 6, or 7 vs. day 3: All t’s <¡.90,dfD10, n.s.). In contrast, TGR(NSEhA2A) showed worse

performance on days 5 and 7 when compared to the levelachieved on day 3 (Fig. 7A, Paired t-test, day 5 vs. day 3:tD¡3.96, dfD 10, P < .01; day 7 vs. day 3: tD¡3.51,dfD10, P < .01).

Each ‘repeated acquisition paradigm’ was realized in a veryshort interval of time (averaged maximum about 3min), sincelatencies to reach the platform were short and the four trialswere consecutively administered (i.e., means§SEM latenciesday 1, control rats: trial 1, 26.07§4.26; trial 2, 15.50§3.22;trial 3, 14.29§2.97; and trial 4, 13.93§2.37; transgenic rats:trial 1, 24.91§7.04; trial 2, 20.55§4.97; trial 3, 16.27§5.38;and trial 4, 22.36§5.93). When ‘genotype’ and ‘trial’ factorswhere analysed on each of the three repeated acquisition par-adigms for working memory only a ‘trial’ eVect on the dis-tance travelled to reach the platform could be found on day 7[Fig. 7B, F (3,69)D5.77, P<.001]. No diVerences were foundin the comparisons between each consecutive trial (index trialn/trial n¡1) on each day. The comparisons of the Wrst andeach of the following trials (index trial 2/trial 1, trial 3/trial 1,or trial 4/trial 1) were also considered. When the index trial4/ trial 1 during the three days was analysed it revealed a‘genotype£day’ eVect [Fig. 7B, F (2,46)D4.86; P<0.05] asthe fourth trial was worse than expected for theTGR(NSEhA2A) animals on day 5 (post hoc analysis, indextrial 4/ trial 1, P<0.05).

Fig. 4. Immunohistochemistry for A2A receptors. A2A receptor immunoreactivity in nerve cell bodies and neuropil in the diVerent areas of the rat prefron-tal cortex (PFCx; coronal section, Bregma 4.2 mm) and caudate-putamen (C-Pu; coronal section, Bregma 0.2 mm) in wild-type (SD) and TGR(NSEhA2A)(A2A) rats. Fr, frontal cortex, area 2; Cg, cingulate cortex; MO, medial orbital cortex; VLO, ventrolateral orbital cortex; LO, lateral orbital cortex; AI,agranular insular cortex (Paxinos & Watson, 1986). Bars: PFCx areas D 40 �m, C-Pu D 150 �m.

Fig. 5. Radioligand binding. Saturation binding analysis of adenosine A2A

receptor antagonist [3H]ZM-241385 in wild-type (SD) andTGR(NSEhA2A) (A2A) rats in the striatum (A) and the prefrontal cortex(B). Results represent the means § SEM of the Bmax of four independentexperiments. *P < 0.05 by Student-t-test. The KD values of the A2A recep-tor did not vary from control to transgenic rats, being 0.83 § 0.3 and0.87 § 0.2 nM in the striatum, 3.03§ 0.2, and 2.20§ 1 nM in the prefrontalcortex (means § SEM).

Autho

r's

pers

onal

co

py

50 L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56

4. Discussion

In the present work, we establish TGR(NSEhA2A), anovel transgenic rat strain with overexpression of A2Areceptors mainly in the cerebral cortex including the hippo-campal formation, and the cerebellum which exhibits work-ing memory deWcits. It must be noted that the homology inamino acids between hA2A and rA2A receptors is about84–85%. Receptors are functionally similar since they areactivated by endogenous adenosine with a similar althoughnot identical aYnity and their coupling to a main G protein

(Gs) is identical. Although this does not exclude the exis-tence of subtle diVerences that have been suggested byslightly diVerent potencies of some selective A2A ligands onthese two receptors, altogether, this strongly supports thatoverexpressing human A2A receptor is of functional signiW-cance in the rat.

Since the areas with A2A receptor overexpression areknown to play a relevant role in the neural circuitry forworking memory (see introduction) special interest waspaid to behavioural screening of putative cognitive deWcitsof TGR(NSEhA2A) in learning and memory paradigms.

Table 1Behavioral assessment of TGR(NSEhA2A)

G D GxD: genotype, day and genotype £ day interaction eVects, P < .05, respectively (MANOVA or Student’s t-test, see Section 2).

Age (weeks) Wild-type (means § SEM) TGR(NSEhA2A) (means § SEM) Statistics

G D G £D

A. Hexagonal tunnel maze 4–5 w n D 12 n D 10Total activity counts 274.75 § 12.68 293.10 § 15.78 n.s. — —

B. 6 Arms radial tunnel maze 4–5w n D 12 n D 10Total activity counts n.s. n.s. n.s.Day 1 204.42 § 12.14 232.80 § 13.10Day 2 219.33 § 10.14 237.20 § 15.45Day 3 237.25 § 16.57 228.30 § 10.48Number of explored arms n.s. n.s. n.s.Day 1 5.33 § 0.40 5.90 § 0.10Day 2 6.00 § 0.00 6.00 § 0.00Day 3 5.83 § 0.17 6.00 § 0.00Latency to explore the 6 arms (s) n.s * n.s.Day 1 216.08 § 19.63 200.40 § 19.97Day 2 143.50 § 18.98 156.90 § 18.39Day 3 119.92 § 25.73 134.80 § 20.40Reentries to explored arms n.s. * n.s.Day 1 5.98 § 1.00 6.18 § 1.03Day 2 3.78 § 1.07 4.06 § 0.74Day 3 3.05 § 1.00 3.31 § 0.75

Number of blind arms See Fig. 6A n.s. ¤ ¤Reentries to blind arms See Fig. 6B n.s. ¤ ¤

C. T-maze 5–6 w n D 14 n D 11Number of correct choices n.s. * n.s.Day 1 2.42 § 0.20 2.27 § 0.24Day 2 2.71 § 0.13 2.73 § 0.14Day 3 2.86 § 0.10 2.91 § 0.09

D. Open-Weld 6–7 w n D 14 n D 11 n.s. — —Total number of crossings 133.47 § 7.19 135.80 § 6.50Total number of rearings 26.13 § 2.66 19.90 § 1.79

E. Object recognition test 6–7 w n D 14 n D 11Time 1 163.9 § 16.31 158.00 § 9.22 n.s. — —TS (time sample object)(s) 15.50 § 1.46 19.10 § 1.38 n.s. — —TN (time new object)(s) 22.54 § 2.21 21.47 § 1.71 n.s. — —Index (TN ¡ TS)/(TN + TS) 0.18 § 0.04 0.05 § 0.04 * — —

F. Morris water maze 7–8 w n D 14Place task for spatial learning (days 1-3) See Fig. 7A n.s. * n.s.Cue learning (day 4) See Fig. 7A n.s. — —Place task for working memory (days 5–7) See Fig. 7A and B n.s. ¤ ¤

G. Elevated plus maze 8–9 w n D 14 n D 11Latency to enter into the open arms 61.43 § 27.35 56.27 § 27.12 n.s. — —Number of entries into the open arms 3.36 § 0.51 2.73 § 0.49 n.s. — —Number of entries into the enclosed arms 10.07 § 0.87 9.18 § 0.50 n.s. — —% time in open arms TOA/(TOA+TEA)£ 100 15.48 § 2.91 11.72 § 3.07 n.s. — —

Autho

r's

pers

onal

co

py

L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56 51

Thus, the battery of tests was designed to assess spatial ref-erence memory and working memory. In particular, fourtests for working memory were used: three spatial ones, 6-arms radial tunnel maze (Bättig, 1983; Fitzgerald et al.,1988), spontaneous alternation in a T-maze (Olton et al.,1979; Squire, 1969) and the ‘repeated acquisition paradigm’in the Morris water maze (Whishaw, 1985) and a non-spa-tial one, the object recognition test (Ennaceur & Delacour,1988; Willig et al., 1987). Other tests for spontaneous motoractivity and anxiety were included as modiWcation of thesebehaviours may interfere in the assessment of cognitivefunction.

The major result obtained in these transgenic rats wasthe discovery of working memory deWcits as measured inthe 6-arms radial tunnel maze when blind arms perfor-

mance was analysed; in the object recognition test withreduced recognition of the novel object; and in the‘repeated acquisition paradigm’ on the Morris water maze

Fig. 6. Learning and memory in a 6-arms radial tunnel maze repeated test(days 1–3). Results are means § SEM. (A) Number of explored blindalleys. (B) Number of re-entries into already explored blind alleys. Wild-type (WT) group, n D 12; TGR(NSEhA2A), n D 10. Two-wayANOVA,‘day’ eVect: aP < 0.01, bP < .05; ‘genotype £ day’ interactioneVects: cP < .01 (see Section 3). Student’s t-test, *P < .05 versus the WTgroup.

Fig. 7. Learning and memory in several tasks in the Morris water maze.Results are means § SEM. (A) Mean distance to reach the platform in thefour trials per session during the place task for spatial learning (days 1–3),the cue learning (day 4) and the ‘repeated acquisition paradigm’ for work-ing memory (days 5–7). (B) Upper panel: distance to reach the platform ineach of the four consecutive trials per session during the place task forworking memory (days 5–7). Lower panel: Index “trial 4/trial 1”: ratio ofdistance (cm) covered on trial 4 divided by the distance (cm) covered ontrial 1, on the same day in this working memory task. Wild-type (WT)group, nD 12; TGR(NSEhA2A), n D 10. Two-way ANOVA, ‘day’ eVect:aP < 0.001; ‘trial’ eVect: bP < .001; ‘genotype £ day’ interaction eVects:cP < 0.05 (see Results). Student’s t-test: *P < .05 versus the WT group;Paired t-test: #P < 0.05 versus day 3 of the same group.

Autho

r's

pers

onal

co

py

52 L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56

test with performance deWcit in the 4th trial vs. 1st trialcomparisons. Certain spatial long-term reference memorydeWcits involving deWcits in a slower learning process werealso observed in the 6-arms radial maze or in the acquisi-tion of a diVerent task in the Morris water maze when theplatform was located in a new position. These memory deW-cits were correlated with evidence of increased A2A mRNAlevels and A2A receptor protein levels as well as of increasedA2A receptor binding function especially in regions of thecerebral cortex.

Spontaneous motor activity, measured by total activitycounts in the two modalities of the tunnel maze and bymeans of crossing and rearings in the open-Weld test, wasnot modiWed in TGR(NSEhA2A). Similar exploratorybehaviours were also exhibited in the sample trial of theobject recognition tests when the ‘time 1’ to reach the crite-ria was analysed or in the number of entries in the plusmaze test. Also, motor abilities to navigate (swim speeds)were equal in the Morris water maze. No genotype diVer-ences were also found in the emotional/anxious-like behav-iour recorded in the plus maze test nor in motivation andsensory motor performance assessed in the cue-learningtask. These are important results since performance of ani-mals in learning and memory tasks may be modiWed bygenotype diVerences in such behavioural features and sincethe overexpression of A2A receptors in areas such as thestriatum could have an eVect on such measures or, forinstance, indirect actions of adenosine on attention andwakefulness could contribute to diVerences in learning andmemory. In this sense, reproducing the main features of ratburrows in the wild provide a more ethological contextthan other devices when assessing locomotor activity,exploratory behaviour and learning and memory, namelythose with interference of motivation or emotionality byusing food-deprivation or punishment (Bättig, 1983).

In the 6-arms radial tunnel maze the number ofre-entries into previously explored arms are consideredequivalent to the retracing errors in more conventional 8-or 16-arms mazes (Fitzgerald et al., 1988). During successivesessions the rats are known to increase their exploratoryeYciency as measured by a lower probability of re-enteringa sector of the maze until all sectors are visited, with areduced time needed to visit all sectors without increasinglocomotor activity. These decreases have been proposed asmeasures of long-term memory formation because there isa successive decrease over repetitive trials (Fitzgerald et al.,1988; Sarter & Steckler, 1989; Welzl, Alessandri, Oettinger,& Bättig, 1988). In our study, between-sessions (long-term)spatial reference memory, mainly dependent on hippocampus(Kesner & Rolls, 2001), was normal in TGR(NSEhA2A) asseveral measures (latency, re-entries into explored arms)including number and re-entries into blind arms decreasedwith repeated testing while total activity counts and num-ber of explored arms were maintained constant over thesessions. In contrast, in the choice between the long-angledarms and blind alleys, considered to provide a measure ofwithin-session working memory similar to that of standard

T-maze (Fitzgerald et al., 1988), the TGR(NSEhA2A)showed a high number of blind arm entries while a singleexperience (on average less than twice) was suYcient forwild-type rats to realize that the alley on the right was shortand blind. Moreover, only a few wild-type animals re-entered an explored blind arm while TGR(NSEhA2A)repeated the same choice several times. The repeated testingoVered TGR(NSEhA2A) a chance to learn about this factand they slowly achieved the exploratory eYciency shownby wild-type rats already from the Wrst day. In other studieswhere lesions induce a transient increase of the number ofblind arm entries the results are also discussed in terms oforientation impairments or even thigmotaxia (Sarter &Steckler, 1989).

Interestingly, our transgenic animals did not diVer fromcontrols in the T-maze. In other animal models for workingmemory, such as the subchronic phencyclidine model, it hasbeen shown that deWcits produced in this test (Jentsch et al.,1999) fail to be reproduced in the radial arm maze (Li, Cul-len, Anwyl, & Rowan, 2003). In our 6-arms radial maze testthe T-intersections oVer a free-choice while in the T-mazethe Wrst trial is always a forced choice and the test trial is afree choice. Measured eVects may also depend on to whichextent tasks purported to assess working memory involveepisodic memory and recruit hippocampal function as well(reviewed by Castner et al., 2004). In this respect, it is inter-esting to note that, as before, the analysis of performance inthe T-maze over three consecutive days provided evidencethat other forms of memory consolidation (i.e., long-termmemory) were more or less preserved resulting in animprovement although achieved more slowly than in wild-type rats.

In the Morris water maze, the acquisition of the placetask for spatial reference memory, which is mainly depen-dent on hippocampus, was as good in TGR(NSEhA2A) asin wild-type rats with similar within-session performanceand a fast improvement with repeated testing over threesessions. The ‘repeated acquisition paradigms’ with four tri-als consecutively administered were realized very fast andeach session was completed within 2 or 3 min. However,when the performance was analysed, the 4th trial wasshown to be worse than expected in TGR(NSEhA2A) sug-gesting deWcits in this working memory task. On the otherhand, in this test, animals daily confronted with an unex-pected modiWcation of the platform position needed a cer-tain level of learning plasticity to inhibit the trend to lookfor previous platform positions while searching its new one.Here, the wild-type rats made it with an average distance onday 5, 6, or 7 similar to that achieved at the end of the pre-vious place task whereas TGR(NSEhA2A) spent twice thattime to reach the platform on days 5 and 7. In a similarplace-alternation task, an essential role of posterior cingu-late areas in the use of topographical information has beendemonstrated (Sutherland, Whishaw, & Kolb, 1988) proba-bly by transmitting and elaborating information passingbetween the hippocampal system and neocortical associa-tion areas.

Autho

r's

pers

onal

co

py

L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56 53

Finally, working memory deWcits shown byTGR(NSEhA2A) were not only restricted to some spatiallearning tasks studied in the mazes. These animals alsoexhibited lower ability than wild-type rats to recognize anew object in a non-spatial learning task assessing workingmemory even when the acquisition of the task, the sponta-neous exploratory behaviour measured by ‘time 1’ duringthe sample trial, was not modiWed.

It seems possible that the deWcits in working memoryobserved in the TGR(NSEhA2A) are mainly caused by theneuronal overexpression of A2A receptors in the corticalregions and specially the prefrontal cortex. Other receptorsputatively involved in working memory, dopamine D1,dopamine D2, adenosine A1, and mGlu5 receptors were notdiVerentially expressed in the cortex of transgenic versuswild-type rats (Fig. 3B). It is true however, that workingmemory deWcits are associated with a disturbed networkactivity involving not only prefrontal cortex but also parie-tal cortex, hippocampus, basal ganglia, and thalamus (seee.g., Levy, Friedman, Davachi, & Goldman-Rakic, 1997,reviewed by Castner et al., 2004) and therefore, the neuro-nal A2A receptor overexpression demonstrated also in thesebrain regions could be of relevance for the observed deWcitsin working memory. With respect to striatum, the increasein striatal A2A receptor density was weak and the presenceof the human A2A mRNA level could hardly be demon-strated in this region. However, this striatal A2A receptoroverexpression may have contributed to the signiWcantdownregulation of D2 and mGlu5 receptors in this region(Fig. 3B), which may be involved in mediating the observedalterations in working memory (see Miyoshi et al., 2002).The selective D2 and mGlu5 receptor downregulation, with-out changes in A1 or D1 receptor expression, is most proba-bly related to the selective ability of A2A receptors to formheteromeric complexes with D2 and mGlu5 receptors(Canals et al., 2003; Ciruela et al., 2004; Ferré et al., 2002;Hillion et al., 2002). It has recently been demonstrated thatA2A receptors form homodimers, which may represent thefunctional form of the A2A receptors expressed in theplasma membrane (Canals et al., 2004; Kamiya, Saitoh,Yoshioka, & Nakata, 2003; Thevenin, Roberts, Lazarova,& Robinson, 2005). In the TGR(NSEhA2A) it seems thatthe formation of A2A receptor dimers is favoured since theywere not found in the striatum of WT rats. Alternativelythe A2A antibody may not recognize the rat A2A receptordimers but this seems less likely. Both A2A receptor homo-mers and heteromers are preformed before traYcking tothe plasma membrane (Canals et al., 2003, 2004) and over-expression of A2A receptors seems to favor the formation ofA2A receptor homodimers versus heteromers, with the con-sequent reduced expression of D2 and mGlu5 receptors.Oligomerization of the 5th transmembrane domain of theA2A receptor may play a role in the A2A dimerization andare highly resistant to temperature and chemical denatur-ation (Thevenin et al., 2005). GPCR heteromerization is animportant event that has pharmacological consequences.We have provided a mechanistic framework to explain

GPCR operation by means of the “two-state dimer model”,a very useful tool for understanding the pharmacology ofmany receptor homo-heterodimers (Franco et al., 2005,2006).

It is known that A2A receptors are positively coupled toadenylate cyclase and increase cAMP accumulation andprotein kinase A activity (Kull et al., 1999) leading toincreased neuronal excitability as seen from increasedGABA release in the ventral striatopallidal GABA path-way after A2A receptor agonist treatment (Ferré, O’Connor,Snaprud, Ungerstedt, & Fuxe, 1994) and A2A receptor ago-nist-induced increases in the striatal glutamate release fromcorticostriatal glutamate terminals (Popoli et al., 2003;Quarta et al., 2004). Therefore, enhanced A2A receptor-mediated transmission in the cortex cerebri may lead to adysbalance in the activity of pyramidal glutamate nervecells and non-pyramidal GABA nerve cells depending onthe degree of A2A receptor overexpression in pyramidal vs.non-pyramidal nerve cells in the cortex cerebri, which maycause the deWcits in the working memory described in thepresent paper. A2A receptors are functionally very similar toD1 receptors, since their activation increases cAMP andneuronal excitability. The role of D1 receptors in the cortexand in working memory has been thoroughly studied. It iswell known that both an increased and a reduced D1 recep-tor-mediated transmission will lead to a deWcit in workingmemory performance (Goldman-Rakic et al., 2000). Thissupports the role of cortical A2A receptor overexpression inthe memory deWcits found in the present study. The A2Aoverexpression in the hippocampus may not be of relevancesince learning of a spatial reference memory tasks (i.e., theWrst place task in the Morris water maze) and other formsof memory dependent on hippocampal function (i.e.,improvement by repeated testing) were not aVected in theTGR(NSEhA2A) in spite of the A2A mRNA and proteinoverexpression found in this region with in situ hybridiza-tion and immunohistochemistry, respectively. For this rea-son Western blot of A2A receptors in the hippocampalformation was not performed.

The present model oVers an advantage over traditionalpharmacology/local injection studies with A2A agonists.Thus, the A2A agonist CGS 21680 in cortical regions bindsalso to a site diVerent from A2A receptors (Johansson &Fredholm, 1989), which has complicated previous observa-tions that e.g., CGS 21680 injected into the posterior cinn-gulate cortex impairs memory retrieval for an inhibitoryavoidance task (Pereira et al., 2005). The results obtained inthe present transgenic model with rats overexpressinghuman A2A receptors and showing working memory deW-cits complement these previous studies. They indicate thatcortical A2A receptors in WT animals even if present in lowdensities in the cerebral cortex (Johansson, Georgiev, Par-kinson, & Fredholm, 1993) may in fact be involved in caus-ing the deWcits found in memory processes afterintracortical A2A agonist injections.

In conclusion, we have established TGR(NSEhA2A) anovel transgenic rat strain overexpressing adenosine A2A

Autho

r's

pers

onal

co

py

54 L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56

receptors mainly in cortical regions of the brain associatedwith a selective failure of working memory performance.The failure of working memory performance is a key deWcitin cognitive malfunction related to prefrontal cortex anom-alies (Goldberg & Weinberger, 1988; Kolb & Whishaw,1983; Weinberger, Berman, & Zec, 1986). Therefore, ratsoverexpressing adenosine A2A receptors could provide anuseful animal model for some of the cognitive disruptionsrelated to altered adenosine/dopamine interactions at thelevel of prefrontal cortex, in view of the existence of antago-nistic A2A/D2 receptor interactions in the brain (Fuxe et al.,2005) and speciWc stimulation of D2 receptors may improvelearning and memory in schizophrenia (Tamminga &Carlsson, 2002) and is involved in modulating workingmemory processes (Wang, Vijayraghavan, & Goldman-Rakic, 2004). Finally, it has been shown that cyclic AMP-mediated signaling components are increased in the pre-frontal cortex of depressed suicide victims (Odagaki et al.,2001). Memory disturbances are known to exist in depres-sion (see Antikainen et al., 2001) and an A2A upregulationin the prefrontal cortex may therefore contribute to suchmemory dysfunctions in depression based on the presentWndings.

Acknowledgments

Thanks to Gloria Blázquez and Laetitia Cuvelier fortechnical assistance. This work was supported by an ECgrant (QLG3-CT-2001-01056), the Swedish ResearchCouncil, Programa Ramon y Cajal and Fundació Marató-TV3 núm. 014110, grants from Ministerio de Ciencia y Tec-nología SAF2002-03293 to R.F., Grant 02/056-00 fromFundacio la Caixa for R.F., Grants 01/012710 from Fun-dació Marató TV3 for R.F., and grants of the FondsNational de la Recherche ScientiWque (SNS), Queen Elisa-beth Medical Foundation (SNS), Van Buuren Foundation(SNS), and Action de Recherche Concertée (SNS).

References

Antikainen, R., Hanninen, T., Honkalampi, K., Hintikka, J., Koivumaa-Honkanen, H., Tanskanen, A., et al. (2001). Mood improvementreduces memory complaints in depressed patients. European Archivesin Psychiatry and Clinical Neuroscience, 251, 6–11.

Arai, A., Kessler, M., & Lynch, G. (1990). The eVects of adenosine on thedevelopment long-term potentiation. Neuroscience Letters, 119, 41–44.

Bättig, K. (1983). Spontaneous tunnel maze locomotion in rats. In G.Zbinden, V. Cuomo, G. Racagni, & B. Weiss (Eds.), Application ofbehavioural pharmacology in toxicology (pp. 15–26). New York: RavenPress.

BraV, D., Stone, C., Callaway, E., Geyer, M., Glick, I., & Bali, L. (1978).Prestimulus eVects on human startle reXex in normals and schizo-phrenics. Psychophysiology, 15, 339–343.

Canals, M., Marcellino, D., Fanelli, F., Ciruela, F., de Benedetti, P., Gold-berg, S. R., et al. (2003). Adenosine A2A-dopamine D2 receptor-recep-tor heteromerization: qualitative and quantitative assessment byXuorescence and bioluminescence energy transfer. Journal of BiologicalChemistry, 278, 46741–46749.

Canals, M., Burgueno, J., Marcellino, D., Cabello, N., Canela, E. I., Mallol,J., et al. (2004). Homodimerization of adenosine A2A receptors: quali-

tative and quantitative assessment by Xuorescence and biolumines-cence energy transfer. Journal of Neurochemistry, 88, 726–734.

Castner, S. A., Goldman-Rakic, P. S., & Williams, G. V. (2004). Animalmodels of working memory: insights for targeting cognitive dysfunc-tion in schizophrenia. Psychopharmacology, 174, 111–125.

Chen, J. F., Huang, Z., Ma, J., Zhu, J., Moratalla, R., Standaert, D., et al.(1999). A(2A) adenosine receptor deWciency attenuates brain injuryinduced by transient focal ischemia in mice. Journal of Neuroscience,19, 9192–9200.

Ciruela, F., Burgueno, J., Casado, V. V., Canals, M., Marcellino, D., Gold-berg, S. R., et al. (2004). Combining mass spectrometry and pull-downtechniques for the study of receptor heteromerization. Direct epitope–epitope electrostatic interactions between adenosine A(2A) and dopa-mine D(2) receptors. Analytical Chemistry, 76, 5354–5363.

d’Alcantara, P., Ledent, C., Swillens, S., & SchiVmann, S. N. (2001). Inacti-vation of adenosine A2A receptor impairs long term potentiation inthe accumbens nucleus without altering basal synaptic transmission.Neuroscience, 107, 455–464.

Ennaceur, A., & Delacour, J. (1988). A new one-trial test for neurobiologi-cal studies of memory in rats. 1, Behavioural data. Behavioural BrainResearch, 31, 47–59.

Ferré, S., Fuxe, K., von Euler, G., Johansson, B., & Fredholm, B. B. (1992).Adenosine–dopamine interactions in the brain. Neuroscience, 51, 501–512.

Ferré, S., O’Connor, W. T., Snaprud, P., Ungerstedt, U., & Fuxe, K. (1994).Antagonistic interaction between adenosine A2A receptors and dopa-mine D2 receptors in the ventral striopallidal system. Implications forthe treatment of schizophrenia. Neuroscience, 63, 765–773.

Ferré, S., Fredholm, B. B., Morelli, M., Popoli, P., & Fuxe, K. (1997). Aden-osine–dopamine receptor–receptor interactions as an integrative mech-anism in the basal ganglia. Trends in Neurosciences, 20, 482–487.

Ferré, S., Karcz-Kubicha, M., Hope, B. T., Popoli, P., Burgueno, J., Gut-ierrez, M. A., Casado, V., Fuxe, K., Goldberg, S. R., Lluis, C., Franco,R., & Ciruela, F. (2002). Synergistic interaction between adenosineA2A and glutamate mGlu5 receptors: implications for striatal neuro-nal function. Proceedings of the National Academy of Science of USA,99, 11940–11945.

Fink, J. S., Weaver, D. R., Rivkees, S. A., Peterfreund, R. A., Pollack, A. E.,Adler, E. M., et al. (1992). Molecular cloning of the rat A2 adenosinereceptor: selective co-expression with D2 dopamine receptors in ratstriatum. Molecular Brain Research, 14, 186–195.

Fitzgerald, R. E., Berres, M., & Schaeppi, U. (1988). Validation of a radialmaze test for assessment of learning and memory in rats. Toxicology,49, 425–432.

Forss-Petter, S., Danielson, P. E., Catsicas, S., Battenberg, E., Price, J.,Nerenberg, M., & SutcliVe, J. G. (1990). Transgenic mice expressing �-galactosidase in mature neurons under neuron-speciWc enolase pro-moter control. Neuron, 5, 187–197.

Franco, R., Casado, V., Mallol, J., Ferre, S., Fuxe, K., Cortes, A., et al.(2005). Dimer-based model for heptaspanning membrane receptors.Trends in Biochemical Sciences, 30, 360–366.

Franco, R., Casado, V., Mallol, J., Ferrada, C., Ferre, S., Fuxe, K., et al.(2006). The “two state dimer receptor model”. A general model forreceptor dimers. Molecular Pharmacology, 69, 1905–1912.

Fuxe, K., Ferré, S., Zoli, M., & Agnati, L. F. (1998). Integrated events incentral dopamine transmission as analyzed at multiple levels. Evidencefor intramembrane adenosine A2A/dopamine D2 and adenosine A1/dopamine D1 receptor interactions in the basal ganglia. Brain ResearchReviews, 26, 258–273.

Fuxe, K., Ferre, S., Canals, M., Torvinen, M., Terasmaa, A., Marcellino,D., et al. (2005). Adenosine A2A and dopamine D2 heteromeric recep-tor complexes and their function. Journal of Molecular Neuroscience,26, 209–219.

García-Munoz, M., Young, S. J., & Groves, P. M. (1991). Terminal excit-ability of the corticostriatal pathway. 1. Regulation by dopaminereceptor stimulation. Brain Research, 551, 195–206.

Giménez-Llort, L., Fernández-Teruel, A., Escorihuela, R. M., Fredholm, B.B., Tobeña, A., Pekín, M., et al. (2002). Mice lacking the adenosine A1

Autho

r's

pers

onal

co

py

L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56 55

receptor are anxious and aggressive, but are normal learners withreduced muscle strength and survival rate. European Journal of Neuro-science, 16, 547–550.

Goldberg, T. E., & Weinberger, D. R. (1988). Probing prefrontal functionin schizophrenia with neuropsychological paradigms. SchizophreniaBulletin, 14, 179–183.

Goldman-Rakic, P. S. (1994). Working memory dysfunction in schizo-phrenia. Journal of Neuropsychiatric Clinical Neuroscience, 6, 348–357.

Goldman-Rakic, P. S., Muly, E. C., III, & Williams, G. V. (2000). D1 recep-tors in prefrontal cells and circuits. Brain Research Reviews, 31, 295–301.

Hauber, W., & Bareiss, A. (2001). Facilitative eVects of an adenosine A1/A2 receptor blockade on spatial memory performance of rats: selectiveenhancement of reference memory retention during the light period.Behavioural Brain Research, 118, 43–52.

Hillion, J., Canals, M., Torvinen, M., Casado, V., Scott, R., Terasmaa, A.,et al. (2002). Coaggregation., cointernalization., and codesensitizationof adenosine A2A receptors and dopamine D2 receptors. Journal of Bio-logical Chemistry, 277, 18091–18097.

Hutchison, A. J., Webb, R. L., Oei, H. H., Ghai, G. R., Zimmerman, M. B.,& Williams, M. (1989). CGS 21680C, an A2 selective adenosine recep-tor agonist with preferential hypotensive activity. Journal of Pharma-cology and Experimental Therapeutics, 251, 47–55.

Kamiya, T., Saitoh, O., Yoshioka, K., & Nakata, H. (2003). Oligomeriza-tion of adenosine A2A and dopamine D2 receptors in living cells. Bio-chemical and Biophysical Research Communications, 306, 544–549.

Jarvis, M. F., Schulz, R., Hutchison, A. J., Do, U. H., Sills, M. A., & Wil-liams, M. (1989). [3H]CGS 21680, a selective A2 adenosine receptoragonist directly labels A2 receptors in rat brain. Journal of Pharmacol-ogy and Experimental Therapeutics, 251, 888–893.

Jentsch, J. D., Taylor, J. R., Redmon, D. E., Jr., Elsworth, J. D., Youngren,K. D., & Roth, R. H. (1999). Dopamine D4 receptor antagonist reversalof subchronic phencyclidine-induced object retrieval/detour deWcits inmonkeys. Psychopharmacology, 142, 78–84.

Johansson, B., & Fredholm, B. B. (1989). Further characterization of thebinding of the adenosine receptor agonist [3H]CGS 21680 in rat brainusing autoradiography. Neuropharmacology, 34, 393–403.

Johansson, B., Georgiev, V., Parkinson, F. E., & Fredholm, B. B. (1993).The binding of the adenosine A2 receptor selective agonist [3H]CGS21680 to rat cortex diVers from its binding to rat striatum. EuropeanJournal of Pharmacology, 247, 103–110.

Kesner, R. P., & Rolls, E. T. (2001). Role of long-term synaptic modiWca-tion in short-term memory. Hippocampus, 11, 240–250.

Kessey, K., Trommer, B. L., Overstreet, L. S., Ji, T., & Mogul, D. J. (1997).A role for adenosine A2 receptors in the induction of long-term poten-tiation in the CA1 region of rat hippocampus. Brain Research, 756,184–190.

Kolb, B., & Whishaw, I. Q. (1983). Performance of schizophrenic patientson tests sensitive to left or right frontal, temporal, or parietal functionin neurological patients. Journal of Nervous and Mental Disorders, 171,435–443.

Kull, B., Arslan, G., Nilsson, C., Owman, C., Lorenzen, A., Schwabe, U., etal. (1999). DiVerences in the order of potency for agonists but notantagonists at human and rat adenosine A2A receptors. BiochemicalPharmacology, 57, 65–67.

Ledent, C., Vaugeois, J-M., SchiVmann, S. N., Pedrazzini, T., El Yacoubi,M., Vanderhaeghe, J-J., et al. (1997). Aggressiveness., hypoalgesia andhigh blood pressure in mice lacking the adenosine A2a receptor.Nature, 388, 674–678.

Levy, R., Friedman, H. R., Davachi, L., & Goldman-Rakic, P. S. (1997).DiVerential activation of the caudate nucleus in primates performingspatial and nonspatial working memory tasks. Journal of Neuroscience,17, 3870–3882.

Li, S., Cullen, W. K., Anwyl, R., & Rowan, M. J. (2003). Dopamine-depen-dent facilitation of LTP induction in hippocampal CA1 by exposure tospatial novelty. Nature Neuroscience, 6, 526–531.

Libert, F., Parmentier, M., Lefort, A., Dinsart, C., Van Sande, J., Maen-haut, C., et al. (1989). Selective ampliWcation and cloning of four new

members of the G protein-coupled receptor family. Science, 244, 569–572.

Miyoshi, E., Wietzikoski, S., Camplessei, M., Silveira, R., Takahashi, R. N.,& Da Cunha, C. (2002). Impaired learning in a spatial working mem-ory version and in a cued version of the water maze in rats with MPTP-induced mesencephalic dopaminergic lesions. Brain Research Bulletin,58, 41–47.

Moreau, J-L., & Huber, G. (1999). Central adenosine A2A receptors: anoverview. Brain Research Reviews, 31, 65–82.

Morris, R. (1984). Developments of a water-maze procedure for studyingspatial learning in the rat. Journal of Neuroscience Methods, 11, 47–60.

Odagaki, Y., Garcia-Sevilla, J. A., Huguelet, P., La Harpe, R., Koyama, T.,& Guimon, J. (2001). Cyclic AMP-mediated signaling components areupregulated in the prefrontal cortex of depressed suicide victims. BrainResearch, 898, 224–231.

Olton, D. S., Becker, J. T., & Handelmann, G. E. (1979). Hippocampus,space, and memory. Behavioural Brain Science, 2, 313–365.

Park, S., & Holzman, P. S. (1992). Schizophrenics show spatial workingmemory deWcits. Archives of General Psychiatry, 49, 975–982.

Paxinos, G., & Watson, C. (1986). The rat brain in stereotaxic coordinates(second ed.). San Diego: Academic Press.

Pereira, G. S., Rossato, J. I., Sarkis, J. J., Cammarota, M., Bonan, C. D., &Izquierdo, I. (2005). Activation of adenosine receptors in the posteriorcingulate cortex impairs memory retrieval in the rat. Neurobiology ofLearing and Memory, 83, 217–223.

Popoli, P., Frank, C., Tebano, M. T., Potenza, R. L., Pintor, A., Domenici,M. R., et al. (2003). Modulation of glutamate release and excitotoxicityby adenosine A2A receptors. Neurology, 61, S69–S77.

Popova, E., Krivokharchenko, A., Ganten, D., & Bader, M. (2002). Com-parison between PMSG and FSH induced superovulation for the gen-eration of transgenic rats. Molecular Reproduction and Development,63, 177–182.

Quarta, D., Borycz, J., Solinas, M., Patkar, K., Hockemeyer, J., Ciruela, F.,et al. (2004). Adenosine receptor-mediated modulation of dopaminerelease in the nucleus accumbens depends on glutamate neurotransmis-sion and N-methyl-D-aspartate receptor stimulation. Journal of Neuro-chemistry, 91, 873–880.

Rebola, N., Sebastiao, A. M., de Mendonca, A., Oliveira, C. R., Ribeiro, J.A., & Cunha, R. A. (2003). Enhanced adenosine A2A receptor facilita-tion of synaptic transmission in the hippocampus of aged rats. Journalof Neurophysiology, 90, 1295–1303.

Rosin, D. L., Robeva, A., Woodard, R. L., Guyenet, P. G., & Linden, J.(1998). Immunohistochemical localization of adenosine A2A receptorsin the rat central nervous system. Journal of Comparative Neurology,401, 163–186.

Sarter, M., & Steckler, T. (1989). Spontaneous exploration of a 6-armradial tunnel maze by basal forebrain lesioned rats: eVects of the ben-zodiazepine receptor antagonist �-carboline ZK 93 426. Psychophar-macology, 98, 193–202.

SchiVmann, S. N., Libert, F., Vassart, G., & Vanderhaeghen, J.-J. (1991).Distribution of adenosine A2 receptor mRNA in the human brain.Neuroscience Letters, 130, 177–181.

SchiVmann, S. N., & Vanderhaeghen, J.-J. (1993). Adenosine A2 receptorsregulate the gene expression of striatopallidal and striatonigral neu-rons. Journal of Neuroscience, 13, 1080–1087.

Silva, J.-A., Jr., Araujo, R. C., Baltatu, O., Oliveira, S. M., Tschöpe, C.,Fink, E., et al. (2000). Reduced cardiac hypertrophy and altered bloodpressure control in transgenic rats with the human tissue kallikreingene. FASEB Journal, 14, 1858–1860.

Squire, L. R. (1969). EVects of pre-trial and post-trial administration ofcholinergic and anticholinergic drugs on spontaneous alternation.Journal of Comparative Physiology and Psychology, 74, 41–45.

Sutherland, R. J., Whishaw, I. Q., & Kolb, B. (1988). Contributions of cin-gulate cortex to two forms of spatial learning and memory. Journal ofNeuroscience, 8, 1863–1872.

Suzuki, F., Shimada, J., Shiozaki, S., Ichikawa, S., Ishii, A., Nakamura, J.,et al. (1993). Adenosine A1 antagonists. 3. Structure–activity relation-ships on amelioration against scopolamine- or N6-((R)-phenylisopro-

Autho

r's

pers

onal

co

py

56 L. Giménez-Llort et al. / Neurobiology of Learning and Memory 87 (2007) 42–56

pyl)adenosine-induced cognitive disturbance. Journal of MedicinalChemistry, 36, 2508–2518.

Swerdlow, N. R., Lipska, B. K., Weinberger, D. R., BraV, D. L., Jaskiw, G. E.,& Geyer, M. A. (1995). Increased sensitivity to the sensorimotor gating-disruptive eVects of apomorphine after lesions of medial prefrontal cortexor ventral hippocampus in adult rats. Psychopharmacology, 122, 27–34.

Tamminga, C. A., & Carlsson, A. (2002). Partial dopamine agonists anddopaminergic stabilizers, in the treatment of psychosis. Current DrugTargets in CNS and Neurological Disorders, 1, 141–147.

Thevenin, D., Roberts, M. F., Lazarova, T., & Robinson, C. R. (2005).Identifying interactions between transmembrane helices from theadenosine A2A receptor. Biochemistry, 44, 16239–16245.

Wang, J. H., Short, J., Ledent, C., Lawrence, A. J., & van den Buuse, M. (2003).Reduced startle habituation and prepulse inhibition in mice lacking theadenosine A2A receptor. Behavioural Brain Research, 143, 201–207.

Wang, M., Vijayraghavan, S., & Goldman-Rakic, P. S. (2004). Selective D2receptor actions on the functional circuitry of working memory. Sci-ence, 303, 853–855.

Weinberger, D. R., Berman, K. F., & Zec, R. F. (1986). Physiologic dys-function of dorsolateral prefrontal cortex in schizophrenia. I.Regional cerebral blood Xow evidence. Archives of General Psychia-try, 43, 114–124.

Welzl, H., Alessandri, B., Oettinger, R., & Bättig, K. (1988). The eVects oflong-term nicotine treatment on locomotion., exploration and memoryin young and old rats. Psychopharmacolpgy, 96, 317–323.

Willig, F., Palacios, A., Monmaur, P., M’Harzi, M., Laurent, J., & Dela-cour, J. (1987). Short-term memory, exploration and locomotor activ-ity in aged rats. Neurobiology of Aging, 8, 393–402.

Whishaw, I. Q. (1985). Formation of a place learning-set by the rat: a newparadigm for neurobehavioural studies. Physiology & Behaviour, 35,139–143.

Zocchi, C., Ongini, E., Conti, A., Monopoli, A., Negretti, A., Baraldi, P.G., et al. (1996). The non-xanthine heterocyclic compound SCH58261 is a new potent and selective A2a adenosine receptor antago-nist. Journal of Pharmacology and Experimental Therapeutics, 276,398–404.