at SciVerse ScienceDirect

Neuropharmacology 73 (2013) 320e326

Contents lists available

Neuropharmacology

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

Inactivation of muscarinic receptors impairs place and responselearning: Implications for multiple memory systems

Juliana Carlota Kramer Soares a, Maria Gabriela Menezes Oliveira a,*,Tatiana Lima Ferreira a,b

aDepartamento de Psicobiologia, Universidade Federal de São Paulo e UNIFESP, Rua Napoleão de Barros, 950, São Paulo, SP 04024-002, BrazilbCentro de Matemática, Computação e Cognição, Universidade Federal do ABC e UFABC, Santo André, SP, Brazil

a r t i c l e i n f o

Article history:Received 6 December 2012Received in revised form3 June 2013Accepted 7 June 2013

Keywords:ScopolamineMuscarinic receptorsHippocampusStriatumPlace learningResponse learningMaze

* Corresponding author. Tel.: þ55 11 2149 0155; faE-mail addresses: gabriela.oliveira@unifesp.b

(M.G.M. Oliveira).

0028-3908/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.neuropharm.2013.06.009

a b s t r a c t

Extensive research has shown that the hippocampus and striatum have dissociable roles in memory andare necessary for place and response learning, respectively. Additional evidence indicates that muscariniccholinergic receptors in the hippocampus and striatum exert an important role in the modulation ofthese memory systems. In our experiments, we assessed whether intact hippocampal and striatalmuscarinic cholinergic transmission may be essential and/or necessary for place and response learning.We addressed these questions using administration of the muscarinic receptor antagonist, scopolamine,on both place and response learning in a food-rewarded T-maze task. The administration of scopolamine(15 mg or 30 mg) directly into the dorsal hippocampus impaired the performance of rats subjected to bothplace and cue-rich response version of the task, but did not affect the response version, when the taskwas performed under cue-poor conditions. However, the administration of scopolamine in the dorso-lateral striatum impaired the cue-poor response version of the T-maze task without interfering with theplace version or cue-rich response version. Taken together, these results indicate that activation ofmuscarinic cholinergic receptors in the hippocampus and striatum facilitate the use of different strate-gies of learning, thus strengthening the hypothesis of multiple memory systems. Additionally, theseresults emphasize the importance of the environmental conditions under which tasks are performed.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Extensive evidence from human and other animal studies in-dicates the existence of multiple memory systems (Squire and Zola-Morgan, 1988; White and Mcdonald, 2002; Doeller et al., 2008). Inrats, lesions of the hippocampus or anatomically related structuresimpair the learning of tasks that require information about theplace or are based on the use of extra-maze cues (i.e., placelearning). However, striatal damage generally impairs the perfor-mance of tasks that involve associations between discrete cues andbehavioral responses (i.e., cued or response learning) (Morris et al.,1982; Packard and McGaugh, 1996; Xavier et al., 1999; Lee et al.,2008; Miyoshi et al., 2012; but see Oliveira et al., 1997; Changand Gold, 2004).

Several studies support the view that acetylcholine (ACh)modulates learning andmemory processes in thesemultiple neural

x: þ55 11 5572 5092.r, gabi.oliveira@gmail.com

All rights reserved.

systems (for review, see Gold, 2003; Hasselmo, 2006; Havekeset al., 2011; Deiana et al., 2011). In the hippocampus and stria-tum, as well as in other brain areas, the effects of ACh are mediatedprimarily by activation of different subtypes of muscarinicreceptors (Hersch et al., 1994; Levey et al., 1995; Yan et al., 2001). Ingeneral, studies using muscarinic antagonists injected directly intothe hippocampus (Riekkinen and Riekkinen,1997; Herrera-Moraleset al., 2007; Mikami et al., 2007; Olson and Cero, 2010) or striatum(Prado-Alcala et al., 1985; Diaz del Guante et al., 1991; Ragozzinoet al., 2002; Legault et al., 2006) impair learning and memorytasks related to the particular neural system. Furthermore, someexperiments indicate that the activation of muscarinic receptors inthese regions is required for the induction of LTP (long-termpotentiation), a form of synaptic plasticity that is widely thought tounderlie learning and memory processes (Segal and Auerbach,1997; Suzuki et al., 2001; Ghiglieri et al., 2011).

The hippocampus and the striatum have high concentrations ofACh. While hippocampal cholinergic inputs arise from basal fore-brain structures (Lewis et al., 1967; Mesulam et al., 1983; Dutaret al., 1995), the high amount of ACh in the striatum is due to thepresence of cholinergic interneurons (Lynch et al., 1972; Bolam

J.C.K. Soares et al. / Neuropharmacology 73 (2013) 320e326 321

et al., 1984; Calabresi et al., 2000). This distinct pattern of cholin-ergic innervation suggests that the cholinergic system may playdifferent roles in modulating behavioral tasks mediated by thesetwo structures. Studies using in vivo microdialysis methods tomeasure ACh release in these brain areas have shown that ratstrained in a dual-solution task e a food-rewarded T-maze that canbe learned using either place or response strategies e showdifferent ACh efflux between the hippocampus and striatum(Chang and Gold, 2003b). Furthermore, the use of a place strategyat the beginning of training was accompanied by early increases ofACh release in the hippocampus, and the use of a response strategylater in training was associated with increases in ACh in the stria-tum. Similar results were observed in experiments using otherversions of the task, designed to require the use of only one of thesetwo strategies. In the place version of the T-maze task, rats weretrained to find food at a particular spatial location (e.g., the armpointingwest). In the response version of the task, the goal armwasalways the arm to the right (or left) of the start arm, regardless ofthe start position. For the response version, rats were trained to findfood by repeating the same body turns. Of related interest, the AChrelease in the hippocampus increased during training for both placeand response versions of the task. These findings might reflect theuse of spatial information to solve the place task as well as theresponse task, possibly because of the availability of extra-mazecues. To test this hypothesis, rats were trained in a responseversion of a maze under either “cue-rich” or “cue-poor” environ-mental conditions. The results indicate that a similar increase ofACh release in the hippocampus was present for both cue condi-tions, but a decrease of ACh release was observed during training inthe cue-poor condition (Pych et al., 2005). These findings suggestthat the hippocampus remained activated throughout trainingwhen extra-maze cues were available but not when the cues wereminimized. Consequently, if intact hippocampal ACh function isnecessary for place and response learning, it is possible thatblocking cholinergic function in the hippocampus will inducesimilar effects on both tasks; however, blocking cholinergic func-tion in the striatum may only impair a response learning task.

Therefore, the purpose of the experiments presented here is tocompare the effects of intrahippocampal and intrastriatal admin-istration of the muscarinic receptor antagonist scopolamine onboth place and response learning in a food-rewarded T-maze.Additionally, we sought to evaluate the environmental conditionsunder which scopolamine administration in the dorsal hippocam-pus does or does not impair response learning.

2. Materials and methods

2.1. Animals

Wistar male rats, 3e4 months old, were used as animal models. The animalswere bred and raised in the animal facility of the Department of Psychobiology andof the Centro de Desenvolvimento de Modelos Experimentais (CEDEME), both of theUniversidade Federal de São Paulo (UNIFESP), and were maintained undercontrolled temperature (23 � 2 �C) and 12:12-h lightedark cycle conditions (lightson between 7:00 h and 19:00 h). Food and water were provided ad libitum untilusage of a food restriction protocol. All procedures followed the local ethical com-mittee of UNIFESP, under number 2000/07, in accordance with international rulesfor animal use and care.

2.2. Surgery

The animals were anesthetized with ketamine (90 mg/kg, ip) and xylazine(10 mg/kg, ip) and placed in a stereotaxic apparatus. Stainless steel guide cannulae(23 gauge, 8 mm) were implanted bilaterally in the dorsal hippocampus or thedorsal striatum and then fixed with dental cement and micro screws. The followingcoordinates were used within the dorsal hippocampus: 3.8 mm posterior to bregma,2.4 mm from the midline and 2.6 mm ventral to the skull surface. Coordinates usedfor the dorsal striatum were 0.3 mm anterior to bregma, 4.0 mm from the midlineand 4.5 mm ventral to the skull surface (Paxinos and Watson, 1997a).

2.3. Apparatus

An elevated plusemaze was used. The maze had four identical arms, with eacharm 60 cm long and 10 cm wide, with walls 2 cm high. A removable wooden blockwas used to block the entry to one of the arms, turning the plus-maze into a T-maze,and another block was used to force the animal to remain in a given segment of themaze, when necessary. Small food wells were positioned at the end of each arm. Themaze remained suspended 1 m from the ground and was located in a lighted roomthat contained a door, windows, and shelves that could serve as extra-maze cues. Asthe maze walls were low, these extra-maze cues were visible to the animal.

2.4. Handling and maze habituation

After surgery, the animals were individually housed in polypropylene plasticcages and were handled approximately 3 min/day, every day for 7 days before mazehabituation. On the first and second days of habituation, the animals were placedindividually in the center of the maze without reward. Each animal was allowed toexplore the maze for 5 min. On the third, fourth and fifth days, each animal wasplaced into the arms, with food inside the food well, and confined for 1 min in eacharm.

2.5. Food deprivation

On the same day as the first day of the habituation phase, the animals wereweighed and submitted to a restricted food regimen. Body weights were graduallyreduced and maintained at 85% of their free-feeding weights. During food depri-vation, 10 g of feed per day were given to each animal until each animal reached thedesired weight. Thereafter, the available food was maintained to keep the animal’sbody weight constant. From the first day of adaptation to the food deprivation,sucrose balls (50 mg used in the preparation of homeopathic medicines) wereplaced in the animal’s cage, in addition to normal dietary feed for the animal tobecome accustomed to the type of food that would be used as a training reward.

2.6. Preference test

On the sixth day of habituation, the preferred arm was established in a prefer-ence test. The animal was placed at the end of the start-arm and allowed to choosefreely between the two other arms of the maze (the arm located in front of the startarmwas blocked). A response was recorded when the rat had four feet placed 10 cminside an arm; no reward was given to the animal in this case.

2.7. Tasks

Separate groups of rats were trained in either a place or a response version of theT-maze. In the place version, the reward was always in the same arm, located in thesame position relative to the room cues (e.g., the arm pointing east). In the responseversion, the animal’s performance depended on the use of an egocentric or responsestrategy based on proprioceptive stimulus (consistently make the same body turn e

e.g., turn to the righte at the choice-point for food reward). In all versions, the northand south arms were pseudo randomly selected as start arms for each trial. The eastand west arms were considered to be goal arms. When a rat started from the south,the north arm was blocked and vice-versa. The correct (rewarded) arm was theopposite of that chosen by the rat on the preference test. Thus, the animals wererewarded in the arm opposite to their initial preference.

In each trial, the animal was placed at end of the start-arm and was allowed torun through themaze until it entered one of the two arms. Upon entering the correctarm, the rat was allowed to eat the reward. If a rat entered an incorrect arm, the foodwell at the end of the non-rewarded arm remained empty; the rat remained ineither arm for 30 s, and no retracting was allowed. When the rats failed to leave thestart arm within 120 s, such trials were considered to be omission errors. Aftereating the reward or after reaching the 120 s time limit, the rats were submitted tothe next trial. The training ended when the rats reached 90 consecutive trials.Training was completed within a single day session. Arm selections and the criterionof 10 consecutive correct choices were recorded as measures of learning. Betweeneach trial, the four arms were cleaned with alcohol (30%) to avoid olfactory cues andto avoid intra-maze cues from the maze. Additionally, the maze was rotated 90�

counterclockwise after every five trials or after any three successive correct choices.

2.8. Drugs and microinjections

Scopolamine hydrobromide (Sigma Chemical Co) was dissolved in 0.9% saline(vehicle) and kept frozen until use. The doses used were 15 and 30 mg/0.5 ml. Thedrug was kept at room temperature on the day of the experiment. The control an-imals received 0.5 ml of saline per side.

Solutions were injected bilaterally through microinjection needle (30 gauges)that extended 1mmbeyond the tip of the guide cannula. Eachmicroinjection needlewas attached to a 10 ml Hamiltonmicro syringe through polyethylene tubing (PE-10).Infusions were controlled by an infusion pump (Model Bi2000 e Insight Equi-pment�, Sao Paulo, Brazil), programmed to deliver solution at a constant speed of

Fig. 1. A representative photography (left) showing the cannula tip and dispersion ofmethylene blue dye and schematic representation (right) of coronal sections of thehippocampus (A) and striatum (B). The dot depicts the area within which the cannulaetips were positioned. Numbers beside each section indicate the distance (mm) pos-terior to Bregma. Only data from animals with correct cannulae implants wereincluded in statistical analyses. Behavioral data were discarded from those animalswith one or both injection sites outside dorsal hippocampus or dorsolateral striatum.Adapted from Paxinos and Watson (1997b). (For interpretation of the references tocolor in this figure legend, the reader is referred to the web version of this article.)

J.C.K. Soares et al. / Neuropharmacology 73 (2013) 320e326322

0.5 ml per min. The microinjection needle was kept in place for an additional minuteto allow drug diffusion. The rats were allowed to move freely during drug admin-istration. The training task began 10 min after microinjection.

2.9. Experimental design

2.9.1. Experiment 1: effects of intrahippocampal scopolamine administration in theacquisition of the place and response version of the T-maze task

Different groups of animals received bilateral microinjections of vehicle (saline0.9%, 0.5 ml) or scopolamine (15 or 30 mg/0.5 ml) directly into the dorsal hippocam-pus, 10 min before training in the place version (Experiment 1A), in the “cue-rich”response version (1B) or in the “cue-poor” response version (Experiment 1C) of theT-maze task.

In Experiment 1C, the general procedures were similar to those used in Exper-iment 1B, except that in this “cue-poor” version of the task, the maze was sur-rounded by black curtains during habituation phase and training trails to completelyremove visual extra-maze cues and to force response learning.

2.9.2. Experiment 2: effects of intrastriatal scopolamine administration in theacquisition of the place and response version of the T-maze task

Different groups of animals received bilateral microinjections of vehicle (saline0.9%, 0.5 ml) or scopolamine (15 or 30 mg/0.5 ml) directly into the dorsal striatum,10min before training in the place version (Experiment 2A), in the cue-rich responseversion (2B) or in the cue-poor response version (Experiment 2C) of the T-maze task.

2.10. Histology

After the behavioral tests, the rats were euthanized with a lethal dose of 20%chloral hydrate. In order to verify cannulae placements and get an estimate of drugdiffusion, a bilateral infusion of 4% methylene blue (0.5 ml) was performed using thesame procedure of microinjection mentioned earlier. Ten minutes after the injectionthe brains were removed, frozen at�80 �C and sliced in coronal sections of 40 mm ina cryostat. The sections were stained with cresyl violet and examined in light mi-croscope to determine the localization of the cannula track. It should be noted,however, that the diffusion of the dye may not be the same of the drug’s due to thedifferent molecular weights and their absorption. The locations of the track and tipsof the cannulae in representative brain rats are shown in Fig. 1. For the rats in whichour aim was to target the dorsal hippocampus, the dye appeared to spread throughthe dorsal hippocampus and dentate gyrus (Fig. 1A). The Fig. 1B shows an example ofdiffusion restricted to the dorsolateral striatum.

2.11. Statistical analysis

Data obtained during training (the number of correct choices by each animalin each block of 10 trials) were analyzed by a two-way ANOVA with repeatedmeasures (training or trials blocks) and treatment (saline or scopolamine) as mainfactors. The analysis was followed by a post hoc Tukey test. The trials to reachcriterion (10 consecutives trials) were analyzed by KruskaleWallis test followedby ManneWhitney test when necessary. We also assessed the number of omis-sions (the number of trials not completed within 120 s). This parameter wasanalyzed by a one-way ANOVA (data not shown). A p-value �0.05 was consideredsignificant.

3. Results

3.1. Experiment 1A

Animals were trained in the place version of the T-maze andreceived bilaterally 15 mg or 30 mg of scopolamine (SCP15 and SCP30groups, respectively) or bilateral saline infusions into the dorsalhippocampus. The number of correct responses is shown in Fig. 2A. Inthe place version, visual extra-maze cues were readily available foranimals to forma spatial locationof the armmaze containing the foodreward. Both doses of scopolamine impaired place learning. A two-way ANOVA showed significant effects of training [F(8,104) ¼ 4.09,p < 0.01] and treatment [F(2,13)¼ 15.28, p < 0.01]. A post-hoc Tukeytest revealed that both groups that received scopolamine (15 mg and30 mg) showed significantly fewer correct choices than the salinetreated controls (p < 0.01). The training � treatment interactionwasnot significant [F(16,104) ¼ 1.82, p ¼ 0.10]. A one-way ANOVA withrepeatedmeasures conducted separately for each group showed thatonly saline group learned the task [F(8,40) ¼ 12.28, p < 0.01]; SCP15group: [F(8,24) ¼ 0.60, p ¼ 0.77]; SCP30 group [F(8,40) ¼ 1.19,p ¼ 0.33].

In place task, saline, SCP15 and SCP30 groups required mediantrials to reach criterion of 53.5; 90.0 and 90.0, respectively. Thedifference in trials to criterion among the groups was significant(KruskaleWallis, H ¼ 6.27; p ¼ 0.043). ManneWhitney U-testindicated that the group that received scopolamine 15 mg requiredsignificantly more trials to reach criterion compared to that of thegroups that received saline during place learning (U ¼ 3.00;p ¼ 0.015). The animals in the scopolamine 30 mg group reachedcriterion somewhat later than did those in the saline group, but thisdifference was not statistically significant (U ¼ 5.000; p ¼ 0.171).

3.2. Experiment 1B

Surprisingly, similar results were obtained when rats weretrained on the cue-rich response version of the T-maze (as shown inFig. 2B). In this version of the task, the animals needed to repeat thesame body turn response to find the food reward, regardless ofextra-maze cues. As in Experiment 1A, rats in the scopolaminetreated groups exhibited learning rates that were significantlyslower than those of the saline group [F(2,10) ¼ 11.66; p < 0.01]. A

Fig. 2. Effects of scopolamine intrahippocampal administration under learning curvesduring place version (A); cue-rich response (B) and cue-poor response (C) versions ofthe T-maze. As compared with saline controls, acquisition was impaired inscopolamine-treated rats if the rats were trained under available extra-maze cuesindependently of versions of the task. SAL, saline; SCP15, scopolamine 15 mg/0.5 ml;SCP30, scopolamine 30 mg/0.5 ml. *p < 0.05 when compared to saline control group(two-way ANOVA followed by Tukey test). The number of animals per group is shownin parentheses.

J.C.K. Soares et al. / Neuropharmacology 73 (2013) 320e326 323

two-way ANOVA also showed significant training effect[F(8,80) ¼ 3.38; p < 0.01]. A post-hoc Tukey test indicated that bothgroups that received scopolamine showed a significantly decreasednumber of correct choices than the saline control group (p < 0.05).The training � treatment interaction was not significant[F(16,80) ¼ 0.77; p ¼ 0.69]. A one-way ANOVA with repeatedmeasures conducted separately for each group showed that onlysaline group learned the task [F(8,32)¼ 5.29, p< 0.01 SCP15 group:[F(8,24) ¼ 0.81, p ¼ 0.60]; SCP30 group [F(8,24) ¼ 0.72, p ¼ 0.67].

The median number of trials to reach criterion in the cue-richresponse version was 42.0 for saline group, 90.0 for SCP15 and90.0 for SCP30 treated rats. The difference in trials to criterionamong the groups was significant (KruskaleWallis, H ¼ 5.898;p ¼ 0.052). ManneWhitney U-test indicated that the group thatreceived scopolamine 15 mg required significantly more trials toreach criterion compared to that of the groups that received salineduring place learning (U ¼ 0.500; p ¼ 0.016). The animals in thescopolamine 30 mg group reached criterion somewhat later thandid those in the saline group, but this difference was not statisti-cally significant (U ¼ 4.000; p ¼ 0.190).

3.3. Experiment 1C

Fig. 2C illustrates the learning curves for the animals thatreceived saline or scopolamine (15 mg or 30 mg) administered in thehippocampus before the cue-poor response version of the T-mazetask. In this version of the task, the availability of room cues waseliminated by curtains placed around the maze. There was no dif-ference between the groups, which is in contrast to the resultsobtained from the response version of the T-maze task that hadextra-maze cues available (Experiment 1B). The treatment effectwas not significant [F(2,17) ¼ 1.74, p ¼ 0.20], nor was thetraining � treatment interaction [F(16,136) ¼ 1.25, p ¼ 0.28]. Onlythe training effect was statistically significant [F(8,136) ¼ 16.83,p < 0.01]. A post hoc Tukey test revealed that all of the trainingblocks differed compared to block 1, indicating that all of the ani-mals showed a performance increase during the cue-poor condi-tion response task. A one-way ANOVA with repeated measuresconducted separately for each group showed that all groupslearned the task equally during the training. Saline group:[F(8,48) ¼ 17.54, p < 0.01]; SCP15 group: [F(8,48) ¼ 5.06, p < 0.01];SCP30 group [F(8,40) ¼ 2.78, p ¼ 0.015].

The saline group, as well as both groups that received scopol-amine (15 mg and 30 mg) achieved criterion in the cue-poorresponse version in a comparable manner. The difference in trialsto reach criterion among the groups was not significant (KruskaleWallis, H ¼ 0.75; p ¼ 0.688). The saline, SCP15 and SCP30 groupsrequired median trials to reach criterion of 36.0, 41.0 and 41.0,respectively.

3.4. Experiment 2A

As observed in Fig. 3A, there was no difference in the number ofcorrect choices made by animals trained in the place version of theT-maze that received scopolamine (15 mg and 30 mg) or saline in-fusions into the striatum. A two-way ANOVA showed that therewasno significant treatment effect [F(2,27) ¼ 2.21, p ¼ 0.13], nor wasthere any interaction between training � treatment factors[F(16,216)¼ 1.04, p¼ 0.41]. Therewas a significant effect of trainingfactor [F(8,216) ¼ 32.70, p < 0.01]. A post-hoc Tukey test revealedthat from the third training block onward, all of the blocks differedfrom the first indicating that the saline, 15 mg scopolamine and30 mg scopolamine groups improved performance during the placeversion of the T-maze task. The one-way ANOVA with repeatedmeasures conducted separately for each group showed that all

Fig. 3. Effects of scopolamine intrastriatal administration under learning curves to sa-line and scopolamine-treated groups trained in the place version (A); cue-rich response(B) and cue-poor response (C) versions of the T-maze. Note that scopolamine treatmentsignificantly impaired the rate of acquisition only cue-poor response version relative tothe saline control group. *p < 0.05 when compared to saline control group (two-wayANOVA followed by Tukey test). SAL, saline; SCP15, scopolamine 15 mg/0.5 ml; SCP30,scopolamine 30 mg/0.5 ml. The number of animals per group is shown in parentheses.

J.C.K. Soares et al. / Neuropharmacology 73 (2013) 320e326324

groups learned the task equally during the training trials. Salinegroup: [F(8,96) ¼ 31.18, p < 0.01]; SCP15 group: [F(8,64) ¼ 12.81,p < 0.01]; SCP30 group [F(8,56) ¼ 3.63, p < 0.01].

In this experiment, the difference in trials to reach criterionamong the groups was not significant (KruskaleWallis, H ¼ 4.08;p ¼ 0.130). The median number of trials to reach criterion in theplace version was 46.0 for the saline group, 76.0 for the SCP15 and76.0 for the SCP30 treated rats.

3.5. Experiment 2B

Fig. 3B illustrates the learning curves for the animals that receivedsaline or scopolamine (15 mg or 30 mg) administered in the striatumbefore cue-rich response version of the T-maze task. The treatmenteffect [F(2,15) ¼ 0.65, p ¼ 0.53] and the training � treatment inter-action effectwerenot significant [F(16,120)¼1.26,p¼0.24].Only thetraining effect was statistically significant [F(8,120)¼ 9.94, p< 0.01].A post hoc Tukey test revealed that all of the training blocks differedcompared to block 1, indicating that all of the animals showed aperformance increase during the response task. Beside this, the one-way ANOVAwith repeated measures conducted separately for eachgroup showed that all groups learned the task equally during thetraining. Saline group: [F(8,48) ¼ 9.88 p < 0.01]; SCP15 group:[F(8,40) ¼ 1.86, p < 0.01]; SCP30 group [F(8,32) ¼ 2.90, p < 0.01].

The difference in trials to reach criterion among the groups incue-rich response task was not significant (KruskaleWallis,H ¼ 2.53; p ¼ 0.28). The saline, SCP15 and SCP30 groups requiredmedian trials to reach criterion of 40.0; 69.5 and 52.0, respectively.

3.6. Experiment 2C

The learning curves for the saline and scopolamine groupsduring the cue-poor response version of the T-maze task are shownin Fig. 3C. In this version, the availability of room cueswas eliminateby curtains placed around the maze. A two-way ANOVA indicated asignificant effect of training [F(8,160) ¼ 14.59; p < 0.01] andtreatment factors[F(2,20) ¼ 10.24; p < 0.01]. A post-hoc Tukey testrevealed that both the 15 mg scopolamine and 30 mg scopolaminegroups displayed significantly fewer correct choices than the salinetreated controls (p < 0.01). The training � treatment interactionfactor was not significant [F(16,160) ¼ 1.23; p ¼ 0.29]. The one-wayANOVA with repeated measures conducted separately for eachgroup showed that all groups learned the task. However, the post-hoc Tukey test revealed that saline group differs significantly fromthe second block, while SCP15 and SCP30 groups differ from thefifth and the third block of trials, respectively. In summary, the taskacquisition was significantly slower in the scopolamine treatedgroups than in the saline group. Saline group: [F(8,56) ¼ 11.58,p < 0.01]; SCP15 group: [F(8,56) ¼ 3.21, p < 0.01]; SCP30 group[F(8,48) ¼ 4.76, p < 0.01].

In cue-poor response version, saline, SCP15 and SCP30 groupsrequired median trials to reach criterion of 28.0; 52.0 and 72.0,respectively. The difference in trials to criterion among the groupswas significant (KruskaleWallis, H ¼ 11.94; p < 0.01). ManneWhitney U-test indicated that the group that received scopolamine30 mg required significantly more trials to reach criterion comparedto that of the groups that received saline during place learning(U ¼ 0.00; p < 0.01). The difference between saline and SCP15groups approached but did not reach statistical significance(U ¼ 14.000; p ¼ 0.058).

Additional measures were registered to determine whether atreatment with scopolamine could also interfere with non-mnemonic factors (e.g., motor or motivational) that may influ-ence task performance. A putative interference on non-mnemonicfactors was excluded based on the results of number of

J.C.K. Soares et al. / Neuropharmacology 73 (2013) 320e326 325

omissions, inwhich the scopolamine infusions had no effects in anyof the accomplished experiments.

4. Discussion

The experiments in this study assessed the effects of intra-hippocampal and intrastriatal scopolamine administration on bothplace and response learning in a food-rewarded T-maze task. Inaddition, our experiments uncovered the conditions in whichscopolamine administration in the hippocampus impairs responselearning.

Scopolamine administered directly into the dorsal hippocampusimpaired the acquisition of the reward in a place version of T-mazetask. Both doses of scopolamine (15 and 30 mg in 0.5 ml) decreasedthe number of correct responses. These results are consistent withthe involvement of the hippocampal cholinergic system in spatialmemory tasks. Several studies have assessed spatial learning andmemory using radial and water mazes. The administration ofscopolamine, at doses similar to the ones used in the present study,impaired the performance in those tasks (Riekkinen and Riekkinen,1997; Herrera-Morales et al., 2007; Mikami et al., 2007; Olson andCero, 2010).

In the present study, the blockade of hippocampal cholinergictransmission with scopolamine administration also impairs thecue-rich response version of the T-maze task. Importantly, theanimals treated with saline or scopolamine did not differ in thenumber of omissions (data not shown), thus suggesting that thedecreased performance was not due to changes in the animals’locomotor activity. Our results are, apparently, inconsistent fromthose of past reports in which the administration of scopolamineimpairs the spatial water maze task, but did not affect the non-spatial task (Riekkinen and Riekkinen, 1997; Herrera-Moraleset al. 2007). However, cholinergic activation in the hippocampusduring response learning has already been observed. As describedpreviously, Chang and Gold (2003b) demonstrated that afterextensive training in a dual solution T-maze task, most rats even-tually switch from place to response strategies, but the ACh releasein the hippocampus remained high throughout training. A similarresult was observed using the training schedule adopted in thepresent study designed to presumptively require the use of onlyone of these two learning strategies. Notably, Pych et al. (2005)showed that the ACh release in the hippocampus increased dur-ing training for both the place and cue-rich response versions of thetask. The absence of a significant difference might reflect the use ofspatial information while the animals were performing the task,irrespective of whether the animal was in the place or cue-richresponse version of the T-maze task. In other words, the hippo-campal cholinergic system is required when extra-maze cues canbe used to solve the maze independently of the presumed learningstrategy used by animals.

In this regard it is important to note that in the cue-rich con-dition all animals had opportunity to learn about extra-maze cuesand consequently to form a cognitive map of the environment. Asthe training task requires an egocentric strategy we could expect animprovement of the animals treated with scopolamine in com-parison to the control animals. In that case, the place strategywould be concurrent to egocentric one (Chang and Gold, 2003a).However, we observed the opposite result. The difference betweenChang and Gold (2003a) and our behavioral procedure relaysmainly on the habituation phase (present in our study). It ispossible that the control animals are solving the task relying on aconditional spatial strategy (for example, when the animals left thesouth arm, they found reward as they turned to one side of theroom, and when they left the north arm, they found reward as theyturned to the opposite side). Animals treated with scopolamine did

not learn the task, at all. As they were also habituated to the roomand maze before any drug was administered, it is probable that thehabituation phase had an impact on the training phase, impairing,thus, the use of the egocentric strategy.

When the availability of room cues was completely removed bythe use of curtains around the maze, a cue-poor response version,intrahippocampal scopolamine had no effect on the number ofcorrect choices or omissions (Experiment 1C). Our finding isconsistent with results obtained by Pych et al. (2005) whichdemonstrated that the ACh released in the hippocampus duringtraining of the response task varies with cue availability. Thisrelease increased in a similar manner during early trials under cue-rich and cue-poor conditions. However, the ACh released in thehippocampus was sustained only in rats trained in the cue-richcondition and declined significantly across trials in the cue-poorcondition.

Other studies showed that intrahippocampal injections ofscopolamine can increase anxiety like (Smythe et al., 1998); andanxiety itself shifts rats from preferred place to response strategiesin a dual-solution task (Packard and Wingard, 2004; Elliott andPackard, 2008). If this was the case for the single-solution tasksused in the present experiment, the animals that received scopol-amine into hippocampus should have learned the cue-richresponse task, what did not happened. Therefore, although anxi-ety effects cannot be ruled out as a contributing factor in the cur-rent results, this interpretation seems unlikely.

The infusion of scopolamine into the dorsolateral striatum didnot affect place (experiment 2A) or cue-rich response (experi-ment 2B) version of T-maze task. In both experiments, there wasno significant difference in the number of correct choices, neitherin the number of trails to learning criterion, or number ofomissions. Initially, these findings suggest that the striatalmuscarinic receptors had no apparent involvement on responselearning. Previously, Chang and Gold (2004) showed thatresponse learning is impaired by inactivation of the dorsolateralstriatum with lidocaine when rats are trained under cue-poorconditions, but not when trained with available visual extra-maze cues. Similarly, in the present study, intrastriatal scopol-amine impaired the learning response of animals when the mazewas placed in an environment without extra-maze cues (exper-iment 2C). Taken together this data suggest that decreasedavailability of extra-maze cues appears to be a variable importantfor the demonstration of striatal muscarinic involvement inresponse learning.

Beside this, based on dissociation effects observed in the placeand cue-poor response version, it seems unlikely that these resultswere due to an alteration in locomotion or motivational factors.Generally, the results obtained in this study are consistent withfindings of other authors showing the involvement of the dorsalstriatum in egocentric or response learning (Potegal, 1972; Cookand Kesner, 1988; Packard and McGaugh, 1996) and extend thesefindings showing the involvement of the muscarinic receptor.

In regards to the striatal cholinergic system, Kitabatake et al.(2003) observed that selective ablation of cholinergic neurons inthe striatum impaired cue learning in the tone-cued T-maze task. Inthis task, animals were trained to respond to auditory instructioncues, which indicated whether the reward was in the left or rightarm and implied a stimuluseresponse association. Subsequently,others studies have also observed an increased release of ACh in thedorsolateral striatum of rats during a similar version of theresponse task that we used in our study (Chang and Gold, 2003b;Pych et al., 2005). This result suggests that ACh striatal is impor-tant for the performance of a cue-poor response version of the T-maze task. Taken together, our findings indicate the involvement ofthe striatal cholinergic system in response learning.

J.C.K. Soares et al. / Neuropharmacology 73 (2013) 320e326326

In summary, our results suggest that the cholinergic systemmodulates the central role of the hippocampus and striatum duringlearning a T-maze discrimination task. This finding do not precludethe involvement of other neurotransmitters in this same functionand reinforce the idea that hippocampus and striatum dorsal arecritical structures for place and response learning, respectively. Wealso demonstrated the important role of the hippocampal cholin-ergic system when extra-maze cues are present in a response task.Although this task could be performed using a strategy indepen-dent of the hippocampus, it is important to consider the availabilityof extra-maze cues to force the animal to use a learning strategythat is truly independent of the hippocampus.

Acknowledgments

This work was supported by grants from AFIP, CAPES and CNPq.The authors thank Jose Bernardo da Costa for technical assistanceand Karina Possa Abrahão for text revision and helpful comments.

References

Bolam, J.P., Wainer, B.H., Smith, A.D., 1984. Characterization of cholinergic neuronsin the rat neostriatum. A combination of choline acetyltransferase immuno-cytochemistry, Golgi-impregnation and electron microscopy. Neuroscience 12,711e718.

Calabresi, P., Centonze, D., Gubellini, P., Pisani, A., Bernardi, G., 2000. Acetylcholine-mediated modulation of striatal function. Trends Neurosci. 23, 120e126.

Chang, Q., Gold, P.E., 2003a. Intra-hippocampal lidocaine injections impair acqui-sition of a place task and facilitate acquisition of a response task in rats. Behav.Brain Res. 144, 19e24.

Chang, Q., Gold, P.E., 2003b. Switching memory systems during learning: changes inpatterns of brain acetylcholine release in the hippocampus and striatum in rats.J. Neurosci. 23, 3001e3005.

Chang, Q., Gold, P.E., 2004. Inactivation of dorsolateral striatum impairs acquisitionof response learning in cue-deficient, but not cue-available, conditions. Behav.Neurosci. 118, 383e388.

Cook, D., Kesner, R.P., 1988. Caudate nucleus and memory for egocentric localiza-tion. Behav. Neural Biol. 49, 332e343.

Deiana, S., Platt, B., Riedel, G., 2011. The cholinergic system and spatial learning.Behav. Brain Res. 221, 389e411.

Diaz del Guante, M.A., Cruz-Morales, S.E., Prado-Alcala, R.A., 1991. Time-dependenteffects of cholinergic blockade of the striatum on memory. Neurosci. Lett. 122,79e82.

Doeller, C.F., King, J.A., Burgess, N., 2008. Parallel striatal and hippocampal systemsfor landmarks and boundaries in spatial memory. Proc. Natl. Acad. Sci. U. S. A.105 (15), 5915e5920.

Dutar, P., Bassant, M.H., Senut, M.C., Lamour, Y., 1995. The septohippocampalpathway: structure and function of a central cholinergic system. Physiol. Rev.75, 393e427.

Elliott, A.E., Packard, M.G., 2008. Intra-amygdala anxiogenic drug infusion prior toretrieval biases rats towards the use of habit memory. Neurobiol. Learn. Mem.90, 616e623.

Ghiglieri, V., Sgobio, C., Costa, C., Picconi, B., Calabresi, P., 2011. Striatum-hippo-campus balance: from physiological behavior to interneuronal pathology. Prog.Neurobiol. 94 (2), 102e114.

Gold, P.E., 2003. Acetylcholine modulation of neural systems involved in learningand memory. Neurobiol. Learn. Mem. 80, 194e210.

Hasselmo, M.E., 2006. The role of acetylcholine in learning and memory. Curr. Opin.Neurobiol. 16, 710e715.

Havekes, R., Abel, T., Van der Zee, E.A., 2011. The cholinergic system and neostriatalmemory functions. Behav. Brain Res. 221, 412e423.

Herrera-Morales, W., Mar, I., Serrano, B., Bermudez-Rattoni, F., 2007. Activation ofhippocampal postsynaptic muscarinic receptors is involved in long-term spatialmemory formation. Eur. J. Neurosci. 25, 1581e1588.

Hersch, S.M., Gutekunst, C.A., Rees, H.D., Heilman, C.J., Levey, A.I., 1994. Distributionof m1em4 muscarinic receptor proteins in the rat striatum: light and electronmicroscopic immunocytochemistry using subtype-specific antibodies.J. Neurosci. 14, 3351e3363.

Kitabatake, Y., Hikida, T., Watanabe, D., Pastan, I., Nakanishi, S., 2003. Impairment ofreward-related learning by cholinergic cell ablation in the striatum. Proc. Natl.Acad. Sci. U. S. A. 100, 7965e7970.

Lee, A.S., Duman, R.S., Pittenger, C., 2008. A double dissociation revealing bidirec-tional competition between striatum and hippocampus during learning. Proc.Natl. Acad. Sci. U. S. A. 105 (44), 17163e17168.

Legault, G., Smith, C.T., Beninger, R.J., 2006. Post-training intra-striatal scopolamine orflupenthixol impairs radial maze learning in rats. Behav. Brain Res.170,148e155.

Levey, A.I., Edmunds, S.M., Koliatsos, V., Wiley, R.G., Heilman, C.J., 1995. Expressionof m1em4 muscarinic acetylcholine receptor proteins in rat hippocampus andregulation by cholinergic innervation. J. Neurosci. 15, 4077e4092.

Lewis, P.R., Shute, C.C., Silver, A., 1967. Confirmation from choline acetylase analysesof a massive cholinergic innervation to the rat hippocampus. J. Physiol. 191,215e224.

Lynch, G.S., Lucas, P.A., Deadwyler, S.A., 1972. The demonstration of acetylcholin-esterase containing neurones within the caudate nucleus of the rat. Brain Res.45, 617e621.

Mesulam, M.M., Mufson, E.J., Wainer, B.H., Levey, A.I., 1983. Central cholinergicpathways in the rat: an overview based on an alternative nomenclature (Ch1eCh6). Neuroscience 10, 1185e1201.

Mikami, A., Masuoka, T., Yasuda, M., Yamamoto, Y., Kamei, C., 2007. Participation ofcholinergic system in memory deficits induced by blockade of hippocampalmGlu(1) receptors. Eur. J. Pharmacol. 575, 82e86.

Miyoshi, E., Wietzikoski, E.C., Bortolanza, M., Boschen, S.L., Canteras, N.S.,Izquierdo, I., Da Cunha, C., 2012. Both the dorsal hippocampus and the dorso-lateral striatum are needed for rat navigation in the Morris water maze. Behav.Brain Res. 226, 171e178.

Morris, R.G., Garrud, P., Rawlins, J.N., O’Keefe, J., 1982. Place navigation impaired inrats with hippocampal lesions. Nature 297, 681e683.

Oliveira, M.G., Bueno, O.F., Pomarico, A.C., Gugliano, E.B., 1997. Strategies used byhippocampal- and caudate-putamen-lesioned rats in a learning task. Neurobiol.Learn. Mem. 68, 32e41.

Olson, M.L., Cero, I.J., 2010. Intrahippocampal Norleucine(1)-Angiotensin IV miti-gates scopolamine-induced spatial working memory deficits. Peptides 31,2209e2215.

Packard, M.G., McGaugh, J.L., 1996. Inactivation of hippocampus or caudate nucleuswith lidocaine differentially affects expression of place and response learning.Neurobiol. Learn. Mem. 65, 65e72.

Packard, M.G., Wingard, J.C., 2004. Amygdala and “emotional” modulation of therelative use of multiple memory systems. Neurobiol. Learn. Mem. 82, 243e252.

Paxinos, G., Watson, C., 1997a. The Rat Brain in Stereotaxic Coordinates. AcademicPress, San Diego.

Paxinos, G., Watson, C., 1997b. The Rat Brain in Stereotaxic Coordinates: ComputerGraphic Files. Academic Publisher, San Diego.

Potegal, M., 1972. The caudate nucleus egocentric localization system. Acta Neu-robiol. Exp. (Wars) 32, 479e494.

Prado-Alcala, R.A., Fernandez-Samblancat, M., Solodkin-Herrera, M., 1985. In-jections of atropine into the caudate nucleus impair the acquisition and themaintenance of passive avoidance. Pharmacol. Biochem. Behav. 22, 243e247.

Pych, J.C., Chang, Q., Colon-Rivera, C., Haag, R., Gold, P.E., 2005. Acetylcholine releasein the hippocampus and striatum during place and response training. Learn.Mem. 12, 564e572.

Ragozzino, M.E., Jih, J., Tzavos, A., 2002. Involvement of the dorsomedial striatum inbehavioral flexibility: role of muscarinic cholinergic receptors. Brain Res. 953,205e214.

Riekkinen, M., Riekkinen Jr., P., 1997. Dorsal hippocampal muscarinic acetylcholineand NMDA receptors disrupt water maze navigation. Neuroreport 8, 645e648.

Segal, M., Auerbach, J.M., 1997. Muscarinic receptors involved in hippocampalplasticity. Life Sci. 60, 1085e1091.

Smythe, J.W., Bhatnagar, S., Murphy, D., Timothy, C., Costall, B., 1998. The effects ofintrahippocampal scopolamine infusions on anxiety in rats as measured by theblackewhite box test. Brain Res. Bull. 45, 89e93.

Squire, L.R., Zola-Morgan, S., 1988. Memory: brain systems and behavior. TrendsNeurosci. 11, 170e175.

Suzuki, T., Miura, M., Nishimura, K., Aosaki, T., 2001. Dopamine-dependent synapticplasticity in the striatal cholinergic interneurons. J. Neurosci. 21, 6492e6501.

White, N.M., Mcdonald, R.J., 2002. Multiple parallel memory systems in the brain ofthe rat. Neurobiol. Learn. Mem. 77, 125e184.

Xavier, G.F., Oliveira-Filho, F.J., Santos, A.M., 1999. Dentate gyrus-selective colchicinelesion and disruption of performance in spatial tasks: difficulties in “placestrategy” because of a lack of flexibility in the use of environmental cues?Hippocampus 9, 668e681.

Yan, Z., Flores-Hernandez, J., Surmeier, D.J., 2001. Coordinated expression ofmuscarinic receptor messenger RNAs in striatal medium spiny neurons.Neuroscience 103 (4), 1017e1024.