Neuroscience 202 (2012) 384395

CHRONIC CAFFEINE CONSUMPTION PREVENTS COGNITIVE DECLINE FROM YOUNG TO MIDDLE AGE IN RATS, AND IS ASSOCIATED WITH INCREASED LENGTH, BRANCHING, AND SPINE DENSITY OF BASAL DENDRITES IN CA1 HIPPOCAMPAL NEURONS
S. VILA-LUNA,a S. CABRERA-ISIDORO,b L. VILA-LUNA,a I. JUREZ-DAZ,b,c J. L. BATA-GARCA,a F. J. ALVAREZ-CERVERA,a R. E. ZAPATA-VZQUEZ,d,e G. ARANKOWSKY-SANDOVAL,a F. HEREDIA-LPEZ,a G. FLORESb AND J. L. GNGORA-ALFAROa,d*
a

Departamento de Neurociencias, Centro de Investigaciones Regionales (CIR) Dr. Hideyo Noguchi, Universidad Autnoma de Yucatn (UADY), Avenida Itzes 490 59, Mrida, Yucatn 97000, Mexico Laboratorio de Neuropsiquiatra, Instituto de Fisiologa, Universidad Autnoma de Puebla, 14 Sur 6301, Puebla 72570, Mxico Laboratorio de Fisiologa, Facultad de Estomatologa, Universidad Autnoma de Puebla, 31 Poniente 1304, Puebla 72410, Mxico

Facultad de Medicina, Universidad Autnoma de Yucatn, Avenida Itzes 498 59-A, Mrida, Yucatn 97000, Mexico Unidad Mdica de Alta Especialidad, Instituto Mexicano del Seguro Social, Ex-terrenos El Fnix, Mrida, Yucatn 97150, Mexico

increase the anxiety indexes assessed by measuring the time spent in the open arms of the elevated plus maze. In addition, morphometric analysis of hippocampal neurons revealed that dendritic branching (90 140 m from the soma), length of 4th and 5th order branches, total dendritic length, and spine density in distal dendritic branches were greater in the basal but not the apical dendrites of CA1 pyramidal neurons from rats chronically treated with caffeine, in comparison with their age- and littermate-matched controls. Altogether, the present ndings strengthen the epidemiological observations suggesting that prolonged caffeine intake prevents the cognitive decline associated with aging, and open the possibility that this process could be mediated by promoting the growth of dendrites and spines in neurons of the adult mammalian brain. 2011 IBRO. Published by Elsevier Ltd. All rights reserved. Key words: caffeine, methylxanthines, aging, memory, neuroprotection, dendritic growth.

AbstractChronic caffeine consumption has been inversely associated with the risk of developing dementia and Alzheimers disease. Here we assessed whether chronic caffeine treatment prevents the behavioral and cognitive decline that male Wistar rats experience from young ( 3 months) to middle age ( 10 months). When animals were young they were evaluated at weekly intervals in three tests: motor activity habituation in the open eld (30-min sessions at the same time on consecutive days), continuous spontaneous alternation in the Y-maze (8 min), and elevated plus-maze (5 min). Afterward, rats from the same litter were randomly assigned either to a caffeine-treated group (n 13) or a control group (n 11), which received only tap water. Caffeine treatment (5 mg/kg/day) began when animals were 4 months old, and lasted for 6 months. Behavioral tests were repeated from day 14 to day 28 after caffeine withdrawal, a time period that is far in excess for the full excretion of a caffeine dose in this species. Thirty days after caffeine discontinuation brains were processed for Golgi-Cox staining. Compared with controls, we found that middle-aged rats that had chronically consumed low doses of caffeine (1) maintained their locomotor habituation during the second consecutive day exposure to the open eld (an index of non-associative learning), (2) maintained their exploratory drive to complete the conventional minimum of nine arm visits required to calculate the alternation performance in the Y-maze in a greater proportion, (3) maintained their alternation percentage above chance level (an index of working memory), and (4) did not
*Corresponding author. Tel: 52 999-924-6412; fax: 52 999-923-6120. E-mail address: jlgongoralf@gmail.com (J. L. Gngora-Alfaro). Abbreviations: AD, Alzheimers disease; A1R, adenosine A1 receptor; A2AR, adenosine A2A receptor; A , amyloid- ; DG, dentate gyrus; EPM, elevated plus maze; LTP, long-term potentiation; OF, open eld; RM-ANOVA, repeated measures analysis of variance; SAB, spontaneous alternation behavior.
0306-4522/12 $36.00 2011 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2011.11.053

During the past decade evidence that moderate consumption of coffee and other caffeinated beverages delays the onset of dementia (Eskelinen and Kivipelto, 2010) and reduces the risk of developing Alzheimers disease (AD) in elderly people (Maia and de Mendona, 2002; Lindsay et al., 2002; Eskelinen and Kivipelto, 2010) has accumulated. In addition, epidemiological surveys performed in elderly populations have found that habitual caffeine consumption shows a signicant association with a lower cognitive decline in men (van Gelder et al., 2007) and women (Santos et al., 2010b), a better capacity for word recall and verbal retrieval in women (Johnson-Kozlow et al., 2002; Ritchie et al., 2007), and a better long-term memory for word recall, and faster motor speed in both sexes (Hameleers et al., 2000). However, not all published reports agree with the above ndings, an incongruity that has been attributed to the lack of a standard methodology among studies (Rosso et al., 2008; Santos et al., 2010a). These discrepancies have given impetus to animal studies designed to assess the neuroprotective actions of chronic caffeine administration under controlled experimental conditions. Thus, the possibility that long-term caffeine consumption could prevent AD has been supported by experiments with transgenic mice (APPsw) bearing a mutated form of human amyloid- (A ) linked to familial AD, in which longterm treatment with caffeine at a dose equivalent to that contained in ve cups of coffee per day, afforded protection against cognitive impairment during aging and prevented A deposition in the hippocampus (Arendash et al.,

384

S. Vila-Luna et al. / Neuroscience 202 (2012) 384395

385

2006). Moreover, when aged APPsw mice already demonstrating cognitive impairment were treated with a chronic caffeine regime, they showed a signicant improvement in working memory that was associated with a signicant reduction of A deposits in the hippocampus and entorhinal cortex (Arendash et al., 2009). In addition, other studies have demonstrated that wild-type mice pretreated with caffeine (30 mg/kg/day) during four days become resistant to the memory decits caused by the i.c.v. injection of A peptide fragment 2535 (DallIgna et al., 2007). However, studies aimed at evaluating whether chronic caffeine treatment prevents the cognitive deterioration that occurs during aging in healthy animals have given inconsistent results. Thus, in one report 18-month old male mice were evaluated after 12 months of caffeine ingestion through their drinking water ( 125 mg/kg/day), and it was found that their working memory in an object recognition task was better than that of their age-matched controls, and similar to the performance of 6-month old animals (Costa et al., 2008). In contrast, another study that evaluated 16-month old male mice following 10 months of caffeine treatment ( 60 mg/kg/day) failed to nd improved performance in a battery of cognitive and motor tests, in comparison with animals that consumed only water (Arendash et al., 2009). These last ndings agree with those of an earlier study performed in 10-month old male rats that consumed a caffeine solution (0.1 mg/ml) during 12 months, and showed no difference in the maintenance of their acquired spatial memory in the Lashley III maze when compared with control animals (Espinola et al., 1997). It should be noted that in all these studies behavioral testing was performed while the animals were still under caffeine treatment. We are not aware of any study aimed at assessing whether the putative cognitive benets afforded by chronic caffeine treatment extend far beyond its consumption period. Such experimental design would be necessary to disclose whether any memory improvement seen after prolonged periods of caffeine ingestion are caused by a true neuroprotective action, and not by its acute cognitive normalizing effects, which have been consistently observed in animal models of memory dysfunction (Takahashi et al., 2008). Another point that deserves consideration is the fact that high doses have been used in the studies that have tested the impact of chronic caffeine intake in healthy animals (Costa et al., 2008; Arendash et al., 2009). This has likely produced brain concentrations affecting multiple molecular targets other than adenosine receptors (Yu et al., 2009), whose blockade by caffeine has been linked to its protective effects against memory dysfunction (Takahashi et al., 2008; Cunha and Agostinho, 2010). However, some reports have supported the possibility that lower doses of caffeine can produce benecial effects on brain functioning. Thus, APPsw transgenic mice that received a single dose of caffeine, as low as 5 mg/kg, had a signicant decrease of A outow measured by microdialysis in the interstitial space of the hippocampus (Cao et al., 2009). Another study found that healthy male rats that consumed a daily caffeine dose of 5 mg/kg during 6 months, followed

by a withdrawal period of at least 2 weeks, developed a greater resistance to the catalepsy induced with the dopaminergic antagonist haloperidol, suggesting that this chronic caffeine schedule produced perdurable changes in brain functioning, not attributable to the presence of caffeine or its metabolites in the cerebral tissue (GngoraAlfaro et al., 2009). Consequently, the present study was designed to assess whether chronic intake of a low caffeine dose by healthy adult rats, followed by a 2-week withdrawal period (a) prevents the progressive motor impairment that occurs during normal aging (Willig et al., 1987; Altun et al., 2007); (b) preserves the habituation of motor activity during repeated exposure to the open eld (OF), which is a form of non-associative learning (Leussis and Bolivar, 2006) that deteriorates during the aging process (Fraley and Springer, 1981); (c) prevents the aging-associated working memory decline in the Y-maze (Willig et al., 1987; Stone et al., 1992); and (d) favors the development of anxiety-like states as those produced by either acute (Pellow et al., 1985) or chronic (El Yacoubi et al., 2000) high doses of caffeine. In addition, the morphometric analysis of GolgiCox stained hippocampal neurons from control and caffeine-treated rats was made because caffeine induces spontaneous ring oscillations in the hippocampus (Pietersen et al., 2009), a nucleus involved in the control of a variety of motor and exploratory behaviors, and previous studies have shown that chronic caffeine treatment causes morphological changes in vertebrate pyramidal neurons (Burgess and Monachello, 1983; Jurez-Mndez et al., 2006).

EXPERIMENTAL PROCEDURES
Animals
The present study was performed on 24 male Wistar rats from ve litters bred in our facilities. Only males were used because there is evidence that estrogens may counteract some neuroprotective actions of caffeine in rodents (Xu et al., 2006). They were acclimatized to the laboratory environment for at least 1 week before any experimental manipulation took place, with 12:12 h light/dark cycles (lights on at 07:00), room temperature of 23 2 C, and food and water ad libitum. Groups of two to four animals were housed in acrylic cages (length, 42 cm; width, 32 cm; height, 17.5 cm) with the same partners throughout the experimental period. All behavioral tests were carried out during light hours (09:00 15:00 h), in a room with controlled temperature (23 2 C), and a xed light intensity of approximately 100 lx. When testing began rats had an average weight of 312 8 g, and at the moment of the last experiment their weights averaged 489 14 g. This study was approved by the Institutional Bioethics Committee of the CIRUADY, and all efforts were made to minimize animal discomfort according to the recommendations of the Guide for the Care and Use of Laboratory Animals of the USA, 1996 revised version.

Treatments
A similar number of animals from the same litter was assigned (see Table 1) to either a caffeine-treated group (5 mg/kg/day, n 13) or a control group (n 11) that received only tap water. The caffeine solution was freshly prepared every day and administered ad libitum through the drinking water. In order to ensure that the animals ingested the expected 5 mg/kg daily dose of caffeine its

386

S. Vila-Luna et al. / Neuroscience 202 (2012) 384395 tions served to determine the horizontal movements of the rat. The occurrence of vertical exploration (rat standing on two hind paws, or rearing) was measured with an additional set of beams (zcoordinate) situated 15 cm above the oor. The data le generated with the xyz motion sequence of each rat was recorded with a personal computer using a virtual instrument developed with LabVIEW software (National Instruments, Austin, TX, USA). The data were analyzed off-line using a program in the same graphical programming environment. We used the paradigm of habituation in the OF, where the decrease of exploratory activity as a function of repeated exposure to the same environment is taken as an index of memory (Leussis and Bolivar, 2006). Consequently, the OF test was performed twice in 30-min sessions at the same hour on consecutive days (herein referred to as day 1 and day 2). The parameters that were measured and compared among experimental groups were distance traveled (cm), time spent in vertical exploration (s), and time in center (min). This last variable was dened as the sum of the time spent in those x-y coordinates situated at more than 5 cm from the box walls. The removable aluminum tray placed at the box bottom was thoroughly cleaned between trials.

Table 1. Assignment of rats from the same litter to treatments Treatment rats Litter A B C D E Total Water 3 1 1 2 4 11 Caffeine 3 2 2 2 4 13

concentration was periodically adjusted according to the body weight gain of all the rats housed in a given cage and to their average daily liquid ingestion in the preceding 7 days. Following this procedure the concentration of caffeine in the solution varied between 0.03 and 0.08 mg/ml. A human would receive an equivalent dose of caffeine by drinking two to three cups of regular coffee, which is in the range of normal daily consumption (Fredholm et al., 1999). Caffeine administration started when the rats were young adults ( 3.9 months old) and lasted for 6 months. This period was the same that was previously used for delivering an equivalent amount of caffeine that induced perdurable resistance to haloperidol-induced catalepsy in this rat strain (GngoraAlfaro et al., 2009).

Spontaneous alternation behavior

Spatial working memory performance was assessed by recording continuous spontaneous alternation behavior (SAB) during a single session in a Y-maze (Stone et al., 1992; Hooper et al., 1996). The rst test was performed when animals were 3.4 months old (100 104 days from birthdate), and the second one 3 weeks after caffeine withdrawal, when rats were 10.6 months old (314 321 days from birthdate). The device was constructed of black acrylic, with three identical arms (45 cm long, 14 cm wide, and 16 cm high) converging on a triangular area (the neutral zone). The sequence of arm entries performed by the rat in the maze was automatically determined by means of three sets of infrared beams positioned at 3.5, 10, and 26 cm from the neutral zone, and 5 cm above the maze oor. Consecutive entries to the same arm (visiting the neutral zone between entries, but not the other arms) were considered as a single visit. A record of the order of arm visits was kept in an electronic memory for subsequent transfer to a personal computer. Rats were placed in the starting arm facing the closed end and allowed to freely explore the maze for 8 min. It was considered that a minimum of nine arm visits was needed to calculate a reliable alternation performance value. Thus, animals unable to meet this criterion were classied as non-alternating. Every three consecutive arm visits made up a triplet set. An alternation was dened as the consecutive entry into three different arms on overlapping triplet sets. Percentage alternation was calculated with the formula: % Alternation {(Number of alternations)/(Total arm entries 2)} 100 (Stone et al., 1992). In addition, the total number of arm entries was used as an index of locomotor activity. The maze was thoroughly cleaned between trials.

Experimental design
Before starting treatments each rat was evaluated in three behavioral tests, at weekly intervals, in the following order: (1) motor activity in the OF, (2) spontaneous alternation in the Y-maze, and (3) elevated plus-maze. In order to minimize the impact of stressors on behavioral outcomes, the animals of both treatment groups were housed under similar conditions, they were moved to the testing room 24 h before evaluation, and all of them were consistently habituated to handling by the experimenter during the 3 4 days preceding the experiment. In addition, to guarantee that the environmental conditions were comparable for both treatment groups, animals scheduled to caffeine and water were alternatively evaluated with the test performed on a given day. Caffeine administration started 1 week after the plus-maze test, when the rats were 3.9 months old (113119 days from birthdate), and continued until they were 9.9 months old (294 299 days from birthdate), when treatment was discontinued. As 6 months represent about 20% of the mean survival age of male Wistar rats (Altun et al., 2007), this was considered a sufciently long period to uncover perdurable behavioral and morphological brain changes induced by chronic caffeine treatment. The control animals received only tap water during this same period. Two weeks after caffeine withdrawal the animals were evaluated in the battery of behavioral tests, which were applied in the same order and intervals as before. As previously discussed (Gngora-Alfaro et al., 2009), the 2-week washout period was judged adequate to allow for a complete elimination of caffeine or its metabolites in the brain tissue of rats.

Elevated plus maze

The elevated plus maze (EPM) (50 cm height) was made of black acrylic, and consisted of two open arms (50 cm 10 cm), and two enclosed arms (50 cm 10 cm 40 cm) (Pellow et al., 1985). The four arms converged on a square central area (neutral zone), and were arranged in such a way that the two open arms were opposite to each other, as were the enclosed arms. The rst test was performed when animals were 3.7 months old (107111 days from birthdate), and the second one 4 weeks after caffeine withdrawal, when rats were 10.8 months old (321328 days from birthdate). Initially rats were placed in the neutral zone always facing the open arm opposite to the location of the experimenter. Animals were allowed to freely explore the maze for 5 min, and their behavior was videotaped for posterior analysis on a

Open eld test

The rst OF test was performed when animals were 3.2 months old (9398 days from birthdate), and the second one 2 weeks after caffeine withdrawal, when rats were 10.4 months old (307315 days from birthdate). The spontaneous motor activity was automatically recorded with a custom-made device consisting of an uncovered transparent acrylic cube (50 cm 50 cm 50 cm), equipped with two sets of nine equidistant (5 cm) infrared beams located 5 cm above the oor, and traversing the area in perpendicular directions, for a total of 81 x-y coordinates. Beam obstruc-

S. Vila-Luna et al. / Neuroscience 202 (2012) 384395 screen by a trained observer. Time spent in the open or closed arms was counted only when the rat introduced its four paws into an arm, whereas time in the central square was counted every time the rat placed any paw in this zone. The measures recorded from all the subjects were time spent in open arms, closed arms, and neutral zone, and number of entries in open and closed arms. The maze was thoroughly cleaned between trials.

387

Golgi-cox staining
Thirty days after caffeine withdrawal, animals were deeply anesthetized with pentobarbital (60 mg/kg) and perfused intracardially with 200 ml of 0.9% saline. The brains were removed and processed with the modied Golgi-Cox staining procedure described by Gibb and Kolb (1998). The brains were rst stored in the dark for 14 days in Golgi-Cox solution, then 3 days in 15% sucrose. Coronal sections (200 m) at the level of the hippocampus were cut with a vibroslicer (Camden Instruments, Leicester, England). Sections were collected onto gelatin-coated microscope slides. The mounted tissue was rinsed in distilled water and then placed in a bath of ammonium hydroxide for 30 min in the dark. After rinsing the sections in water they were immersed in Kodak lm xer for 30 min in the dark and subsequently washed with distilled water and dehydrated in successive baths of 50% (1 min), 70% (1 min), 95% (1 min), and 100% (2 5 min) alcohol, followed by 15 min in xylene. The slides were covered with balsam resinous medium. Dendritic morphology was evaluated in neurons from three areas of the dorsal hippocampus, CA1, CA3, and dentate gyrus (DG), as previously described (Jurez-Mndez et al., 2006). The Golgi-impregnated pyramidal neurons from the hippocampus [Plates 2732 of Paxinos and Watson (1986)] were readily identied by their characteristic triangular soma shape, apical dendritic extension toward the pial surface, and numerous dendritic spines. Granule cells of the DG were recognized by their distinctive oval or spherical soma shape and numerous dendritic spines. Only complete, fully impregnated pyramidal neurons located in the CA1, CA3, or granule cells of the DG elds of hippocampus, with no apparent truncation of the basal dendritic tree were included in our analyses; the ends of the dendrites were positively identied by their characteristic conical shape. The following criteria were used to select pyramidal neurons for analysis: (1) localization of the cell soma within the middle of the thickness of the section; (2) full impregnation of the dendritic longitude and spine neurons; (3) presence of at least three primary basilar dendritic shafts, each of which branched at least once; (4) no morphological artifacts attributable to Golgi-Cox staining. For each hippocampal region, ve neurons in each hemisphere were drawn using a camera lucida at a magnication of 250 by a person blind to treatment conditions. For each neuron, the three-dimensional apical and basal dendritic tree, including all branches, was reconstructed in a two-dimensional plane and the dendritic tracing was quantied by Sholl analysis. Using the center of the soma as the reference point, a transparent grid with concentric rings spaced 10 m was placed over the drawing and the total number of intersections between rings and each dendritic branch was counted separately for apical and basal dendrites with the purpose of estimating the branch order of the dendritic arborization and the total dendritic length. The density of dendritic spines was estimated by drawing segments of 10 m from the terminal tips at high magnication (1000 ) and then counting the number of spines.

before and after caffeine treatment. Fishers exact test (one-tailed) was used to analyze the change in the proportion of animals that reached the criteria of nine arm visits in the Y-maze necessary to assess percent alternation. In order to determine whether rats alternated above chance level, one-sample t-tests on the percent alternations minus 50% (i.e. chance level) were performed per treatment group, and the resulting difference score was compared with zero (Hooper et al., 1996). A signicance level of 0.05 was chosen for all tests. Values are reported as mean SEM.

RESULTS
Fig. 1A shows that the actual caffeine intake of rats sharing the same home cage was close to the intended dose of 5 mg/kg/day during the 6 months of its free ingestion through drinking water. Consistent with the mild diuretic action of caffeine (Rieg et al., 2005), the bedding of caffeine-treated rats usually needed more frequent changes because it became damp with urine faster than that of control animals. The enhanced urine output, likely induced by caffeine, probably rendered the rats thirsty, which might explain why they drank more liquid on average than those receiving only water, although this effect was not signicant (Fig. 1B). This last observation can be taken as indirect evidence that animals consuming caffeine at a mean dose of 5 mg/kg/day experienced its pharmacological effects continuously. This caffeine regime did not affect the weight gain of animals (Fig. 1C). Locomotion in the open eld The values of the distance traveled in the OF are shown in Table 2. RM-ANOVA revealed signicant effects for treatment (F1,66 5.4, P 0.05) and session (F3,66 17.8, P 0.01), but not for treatment session interaction (F3,66 1.0). At young age, before treatments, the rats assigned to the water regime walked longer distances during day 1 than those allocated to the caffeine schedule (Table 2). Besides, both groups walked signicantly less during day 1 at middle age than at young age. The difference between the distances walked during day 1 and day 2 was taken as an index of locomotion habituation (higher differences representing greater habituation). The mean habituation distance values at young age were 801 344 cm for the water group and 493 167 cm for the caffeine group; whereas at middle age the means were 49 178 cm for water and 449 186 cm for caffeine. RM-ANOVA disclosed a marginally signicant effect for age (F1,22 3.8, P 0.06), but not for treatment (F1,22 0.04, P 0.85), or treatment age interaction (F1,22 3.0, P 0.10). Bonferroni tests indicated that only the rats assigned to the water regime did not present locomotion habituation when they were middle-aged (P 0.05). Vertical exploration in the open eld

Data analysis
As the rats assigned to both treatments were tested four times in the OF, the obtained data were analyzed with two-way repeatedmeasures ANOVA (RM-ANOVA), with OF session and treatment as factors. Paired comparisons were made with Bonferronis test. The same statistical tests were applied to analyze the results of the times spent in the open and closed arms of the EPM, recorded

The values of the time spent exploring the OF are shown in Table 2. Two-way RM-ANOVA only revealed a signicant effect for session (F3,66 16.0, P 0.01), but not for treatment (F1,66 0.3, P 0.61) or treatment session interaction (F3,66 0.2, P 0.92). With respect to the session effect, Bonferroni test showed that both treatment groups

388

S. Vila-Luna et al. / Neuroscience 202 (2012) 384395

spent signicantly less time in vertical exploration during day 1 or day 2 at middle age than on day 1 at young age. The difference between the vertical exploration times during day 1 and day 2 was taken as an index of exploration habituation. At young age the mean exploration habituation time values were 64 19 s for water and 48 13 s for caffeine; whereas at middle age the means were 0 12 s for water and 5 15 s for caffeine. RM-ANOVA disclosed a signicant effect for age (F1,22 14.5, P 0.01), indicating that both treatment groups had exploration habituation only at young age. However, no signicant effects for treatment or treatment age interaction were found. Time in the center of the open eld The values of the time spent in the center of the OF are shown in Table 2. RM-ANOVA revealed a signicantly lower mean for caffeine than for water (F1,66 4.6, P 0.05). However, this result should be interpreted with caution because the mean times spent in the OF center were lower for the caffeine group even before treatments were applied. A signicant effect for session was also found (F3,66 6.7, P 0.01), given that both groups spent more time in the center at middle age than at young age. There was no signicant treatment session interaction (F3,66 0.4, P 0.73). Spontaneous alternation in the Y-maze At young age, all rats assigned either to the water (n 11) or caffeine (n 13) treatments completed the minimum of nine arm visits required to calculate the alternation performance in the Y-maze. However, when these animals were reevaluated at middle age, only 6 out of 11 rats in the water group met this criterion, representing a signicant change in proportion (P 0.018, Fishers test), whereas 12 out of 13 rats of the caffeine group still reached the mark (P 0.5). Half of the caffeine-treated rats that met the SAB criterion scored higher at middle age than at young age (increments of 4, 6, 9, 20, 22, and 30%). In contrast, only two out of six control animals increased their alternation percentage at middle age vs. young age (increments of 4, and 16%). One-sample t-test analysis of SAB performance of those rats that reached the criterion at both ages revealed that the water group performed signicantly above chance level at young age (68.1 5.3%; t5 3.44, P 0.02), but not at middle age (59.8 4.8%; t5 2.03, P 0.10). In contrast, the caffeine group performed signicantly above chance level at both young age (67.9 3.7%; t11 4.88, P 0.001) and middle age (66.4 4.4%; t11 3.69, P 0.01). Regarding the number of arm entries in the Y-maze, no signicant differences were found between treatment groups either at middle age (water, 10.9 1.7; caffeine, 12.3 1.1) or at young age (water, 15.0 1.4; caffeine, 15.9 1.1). However, for both treatment groups, the values were signicantly lower at middle age vs. young age (P 0.05). Elevated plus maze test The values of the parameters measured in the EPM are shown in Table 3. RM-ANOVA of the time spent in the

Fig. 1. Average caffeine intake, uid consumption, and body weight gain during 6 mon. (A) Caffeine intake by four groups of rats housed in separate cages. Each symbol represents the average caffeine dose collectively consumed at weekly intervals by all the animals (n) sharing a given cage. The actual daily caffeine consumption was calculated according to the formula: (concentration of caffeine solution supplied in the preceding week [mg/ml]) (average daily uid ingestion by all rats in a given cage in the preceding week [ml])/(sum of body weights of all rats in a given cage [kg]). (B) Fluid consumption. The daily liquid intake was calculated with the formula: (average daily uid ingestion by all rats in a given cage in the preceding week [ml])/(sum of body weights of all rats in a given cage [kg]). Each symbol represents the mean SEM values for animals housed in four cages, receiving either caffeine or water. Two-way RM-ANOVA indicated a signicant effect for time (F25,150 5.1, P 0.01), but not treatment (F1,150 2.0, P 0.21), or their interaction (F25,150 0.1, P 1.0) (C) Time course of body weight increase of rats treated with caffeine (n 13) or water (n 11). Arrows point at the average weight values of animals when behavioral tests were performed: OF, open eld; Y, alternation in the Y-maze; , elevated plus maze.

S. Vila-Luna et al. / Neuroscience 202 (2012) 384395

Table 2. Spontaneous motor activity parameters displayed by rats during 30 min in the open eld Parameter Treatment Age 3.2 mon Day 1 Distance traveled (cm) Vertical exploration (s) Time in center (min) Water Caffeine Water Caffeine Water Caffeine 4059 3112 155 157 8.3 7.0 345*, , 204 , 18 ,, 22 ,, 0.9** 1.0
,

389

10.4 mon Day 2 3258 2620 91 102 8.5 5.0 247** 236 16*** 18** 1.1 0.8*** Day 1 2683 2376 78 79 12.5 10.6 320*** 159** 16 17 1.6 , 1.6*** Day 2 2634 1927 84 89 11.3 8.0 352 189, 10 10 1.9 1.2

For each variable, paired comparisons with Bonferroni test of the means ( SEM) arranged in columns showed that the only signicant difference occurred at young age, when the rats assigned to the water regime traveled a signicantly longer distance on day 1 in the OF (* P 0.05) than those allocated to caffeine. Paired comparisons of the means arranged in rows yielded signicant differences among those marked with the same symbols. ** P 0.05; *** P 0.01; P 0.01; P 0.05; P 0.01.

open arms revealed a signicant effect for age (F1,22 18.6, P 0.01), given that both treatment groups spent signicantly less time exploring the open arms at middle age vs. young age (Table 3). The time spent in the open arms showed no signicant effect for treatment (F1,22 0.9, P 0.35) or treatment age interaction (F1,22 0.2, P 0.68). Further, no signicant changes were found in the time spent exploring the closed arms or the neutral zone of the EPM. Regarding the number of open arms entries in the EPM, no signicant differences were found between treatment groups at both ages (Table 3). However, both groups performed less open arms entries at middle age than at young age (P 0.05).
Table 3. Time measurements (in seconds and as percentages) and arm entries for the elevated plus maze test Zone Treatment Age 3.7 mon Open arms Water Time % Time Entries Caffeine Time % Time Entries Water Time % Time Entries Caffeine Time % Time Entries Water Time % Time Caffeine Time % Time 10.8 mon

Morphometry of hippocampal neurons The Golgi-Cox impregnation procedure clearly lled the dendritic shafts and spines of hippocampal neurons (Fig. 2A). The morphological analysis presented here is based on a total of 710 neurons from 24 animals (11 vehicletreated and 13 caffeine-treated). As one brain of the caffeine-treated rats had insufcient number of stained neurons in the CA1 eld for a reliable morphometric analysis, only the data of 12 brains are reported for this group. Thus, estimates of spine number, total dendritic length, and length for each dendritic branching order were obtained from 230 CA1 neurons, 240 CA3 neurons, and 240 DG neurons. Sholl analysis revealed that the dendritic branching (90 140 m from the soma), the length of the 4th and 5th order branches (Fig. 3BG), and the total dendritic length (Table 4) were greater in the basal, but not the apical, dendrites of CA1 pyramidal neurons from rats chronically treated with caffeine, in comparison with their age and littermate-matched controls (Fig. 2B). The density of spines in distal dendritic branches showed a marginally signicant increase only in the basal dendrites of CA1 neurons (Table 4). In contrast, no signicant differences were found between caffeine and water-treated rats for the same morphometric parameters of CA3 and DG neurons.

36.0 8.3 12.0 2.8 2.0 0.3 27.3 6.8 9.1 2.3 1.3 0.3 182.8 12.7 60.9 4.2 10.5 1.4 197.5 9.7 65.8 3.2 9.6 1.1 81.2 10.7 27.1 3.6 75.2 8.1 25.1 2.7

11.4 5.0* 3.8 1.7* 0.6 0.2* 7.0 3.6** 2.3 1.2** 0.4 0.2* 204.6 12.7 68.2 4.2 8.2 1.0 211.6 16.7 70.5 5.6 6.6 1.3 84.0 11.3 28.0 3.8 81.4 16.3 27.1 5.4

DISCUSSION
The main nding of the present study was that prolonged caffeine treatment (from 4 to 10 months of age) given to adult male rats caused a signicant attenuation of some indexes of behavioral decline associated with aging: (1) preserving their locomotor habituation in the OF, (2) maintaining their exploratory drive for completing the minimum of nine arm visits required to calculate the alternation performance in the Y-maze in a greater proportion than control animals, and (3) maintaining their alternation percentage above chance level. As these behavioral outcomes were observed 2 weeks after caffeine discontinuation, it is unlikely that they were a manifestation of withdrawal symptoms (e.g. hypolocomotion), which in rats does not last more than 5 days after ceasing the chronic administration of high doses ( 160 mg/kg/d) (Holtzman, 1983). However, as the time course of mood or mem-

Closed arms

Central square

* P 0.01,** P 0.05; vs. 3.7 mon.

390

S. Vila-Luna et al. / Neuroscience 202 (2012) 384395

Fig. 2. Chronic caffeine treatment selectively modies the morphology of basal dendrites of hippocampal CA1 pyramidal neurons. (A) Photomicrograph of a representative Golgi-Cox-impregnated CA1 pyramidal neuron of the right hippocampus from a vehicle-treated rat. Bar 50 m. (B) Representative schematic drawings of the dendritic arbor of the CA1 pyramidal neurons of the right and left hippocampus. Observe that both neurons from a caffeine-treated rat have longer basal dendrites than those from a control animal.

ory-related performance after caffeine withdrawal in rats has never been tested, the possibility that the lower behavioral decline observed in caffeine-treated rats was a manifestation of drug withdrawal cannot be completely discarded. Besides the behavioral effects, histological analysis of brains processed for Golgi-Cox staining 1 month after treatment withdrawal revealed that caffeine-exposed rats developed a persistent increase in the length, branching, and spine density of the basal dendrites of CA1 hippocampal pyramidal neurons. The present experimental design does not permit discerning whether these morphologic changes occurred during chronic caffeine administration or after its withdrawal. However, there is evidence that the changes in the number of spines measured in dendrites of accumbal neurons 2 days after discontinuing the chronic administration of morphine (Diana et al., 2006) or cocaine (Dobi et al., 2011) are no longer observed 2 4 weeks later. Altogether, these evidences argue against the possibility that the morphologic changes observed in CA1 hippocampal neurons had occurred during the month elapsed after caffeine withdrawal. It should be noted that chronic caffeine treatment did not prevent the aging-associated decline in all recorded behaviors. Specically, the caffeine-treated rats and their

littermate controls performed similarly at middle age in various parameters measured in the OF (traveled distance, exploration time), the Y-maze (% SAB), and the elevated-plus maze (time spent in open and closed arms). These ndings agree with those of an earlier study showing that healthy mice chronically treated with caffeine for 10 months (from 5 to 15 months of age) did not improve their cognitive performance in a battery of memory tests, neither were sensorimotor or anxiety measures affected when compared with control animals consuming only water (Arendash et al., 2009). Unfortunately, these authors did not measure the behavioral performance of mice before starting the caffeine or vehicle treatments, when animals were middle-aged (5 months), making it impossible to assess whether caffeine treatment prevented any behavioral decline that might have occurred at advanced age (15 months). Other experiments with male Wistar rats have shown that their acquired spatial memory in a Lashley III maze remained unaffected during 1 year of chronic caffeine treatment in comparison with control animals (Espinola et al., 1997). In the only study that has reported preservation of working memory in aged mice after chronic caffeine treatment, very high doses ( 125 mg/kg/day) were used, and were still delivered at the mo-

S. Vila-Luna et al. / Neuroscience 202 (2012) 384395

391

Fig. 3. Effect of chronic caffeine treatment on dendritic morphometry of hippocampal neurons. Graphs (AE): Sholl analysis of dendrite ramications. Graphs (FJ): length of dendritic branches according to their order. Long-term caffeine consumption caused a signicant effect only in the basal

392

S. Vila-Luna et al. / Neuroscience 202 (2012) 384395

Table 4. Dendritic length and spine density in hippocampal neurons Dendrites CA1 apical Total length ( Spines/10 m CA1 basal Total length ( Spines/10 m CA3 apical Total length ( Spines/10 m CA3 basal Total length ( Spines/10 m DG Total length ( Spines/10 m Water Caffeine

m)

1752 88 7.8 0.3 1562 74 7.2 0.3 1466 43 6.4 0.3 2003 75 6.0 0.3 1425 75 6.9 0.3

1749 70 7.8 0.4 1782 68* 8.0 0.3** 1383 73 6.6 0.3 2156 78 6.6 0.2 1466 90 7.4 0.3

m)

m)

m)

m)

riorates during the aging process (Fraley and Springer, 1981), and the hippocampus is critically involved in the acquisition of motor habituation to the OF (Leussis and Bolivar, 2006). Using this paradigm we found that, at young age, both the water- and caffeine-treated rats had a signicant reduction in the distance walked during the second consecutive day of testing in the OF, which is consistent with previous studies (Miyagawa et al., 1998; Bert et al., 2002). However, when animals were evaluated again at middle age, those that had been under chronic caffeine treatment maintained their locomotion habituation at values that were closer to those measured when they were young, whereas control animals did not. These observations could be considered as evidence that long-term consumption of caffeine at low doses prevents the decline of locomotion habituation to a novel environment associated with aging (Fraley and Springer, 1981). Effects of chronic caffeine treatment on anxiety indexes As caffeine is a drug that has been associated with the development of anxious states (El Yacoubi et al., 2000), it was important to assess whether its chronic consumption could enhance the anxiety levels of rats. In the OF test, any experimental manipulation that decreases the time spent in the central part of the arena is interpreted as anxiogenic, whereas the opposite is considered anxiolytic (Prut and Belzung, 2003). The observation that control and caffeinetreated rats spent similar amounts of time in the OF center when they reached middle age suggests that chronic consumption of low amounts of caffeine does not predispose the subjects to the development of anxiety to open spaces. In the EPM, the amount of time spent in the open arms is considered the most reliable procedure to evaluate emotionality in rodents (Walf and Frye, 2007). Using this paradigm no signicant differences were detected between control animals and those that had consumed caffeine. Although an earlier study reported that chronic caffeine treatment reduced the time that mice spent in the open arms of an EPM in comparison with their controls, this effect occurred at very high doses (50 mg/kg) (El Yacoubi et al., 2000). Collectively, the above ndings argue against the possibility that chronic caffeine consumption at the dose of 5 mg/kg/day could favor the development of anxious states. Effects of chronic caffeine treatment on spontaneous alternation behavior SAB in a Y-maze is considered a test of working spatial memory because animals need to remember the arm that was previously visited in order to sequentially explore the three arms of the maze (Hooper et al., 1996). The expression of SAB requires an intact hippocampal function (Ger-

* P 0.02,** P 0.05, vs. water; unpaired Students t-test (one-tailed).

ment of behavioral testing (Costa et al., 2008), thus making it difcult to distinguish whether the effect was because of a neuroprotective action or because of the acute cognitive normalizing effects of caffeine (Takahashi et al., 2008; Cunha and Agostinho, 2010). The present results show that, even after several weeks of caffeine withdrawal, the rats that consumed low doses of caffeine during 6 months exhibited improved performance in some memory tests than their controls. These observations indicate that chronic caffeine intake prevents the decline of some brain functions that occur during the aging process. Effects of chronic caffeine treatment on open eld behavior In agreement with previous studies showing that aging is accompanied by a gradual decrease in locomotion and rearing (Willig et al., 1987; Miyagawa et al., 1998; Altun et al., 2007), here we found that rats traveled shorter distances and spent less time in vertical exploration at middle age than at young age. The activity decline associated with aging was also evident as a reduced number of arm visits made by rats from both treatment groups in the Y-maze and the EPM at middle age. Altogether, these ndings suggest that chronic caffeine treatment during the period spanning from young to middle age does not prevent the normal decline of locomotion and exploration that occurs during the aging process. Long-term habituation to a novel environment in rodents is conceptualized as a form of non-associative learning in which the decline of locomotion after repeated exposure to the same environment is taken as an index of memory (Leussis and Bolivar, 2006). This behavior dete-

dendrites of CA1 pyramidal neurons. (B) Two-way ANOVA of Sholl analysis revealed a signicant effect of treatment (F1,483 4.8, P 0.05) and circle radius (F23,483 595.4, P 0.0001) on the number of intersections per shell, as well as a signicant interaction between the main factors (F23,483 3.7, P 0.0001). Bonferroni test indicated that the number of intersections per concentric ring was signicantly greater at distances of 90 140 m from the soma. (G) Two-way ANOVA of dendritic length per branch order revealed a signicant effect of treatment (F1,126 4.9, P 0.05) and branch order (F6,126 156.2, P 0.001) on branch length, as well as a signicant interaction between the main factors (F6,126 4.1, P 0.0001). Bonferroni test indicated that 4th and 5th order dendritic branches were signicantly longer in caffeine-treated rats: * P 0.05, and P 0.01, vs. control animals.

S. Vila-Luna et al. / Neuroscience 202 (2012) 384395

393

lai, 1998), and it is a behavior that deteriorates during the aging process (Willig et al., 1987; Stone et al., 1992). Here we evaluated SAB performance of the same animals at two different ages, and found that when they were young all of them completed the minimum of nine arm visits required to reliably calculate the alternation performance, irrespective of the assigned treatment. However, when they were retested at middle age, about half of the animals that consumed only water failed to meet this criterion, which might reect the loss of exploratory drive usually observed during the aging process (Willig et al., 1987). In contrast, all but one of the rats that consumed caffeine for 6 months reached the norm, suggesting that this treatment prevented the deterioration of the brain processes that drive exploration of the environment. In addition, the mean percent alternation at young age was practically the same for animals assigned to the water or caffeine regimes, which in both cases occurred above chance level. However, when they were reevaluated at middle age, only the rats that had consumed caffeine for 6 months maintained their percent alternation above chance level, but not the control animals. It is important to emphasize that the improved alternation execution recorded in middle-aged rats that had been under caffeine treatment was observed 3 weeks after its withdrawal, which implies that the previous prolonged caffeine consumption induced persistent brain changes (see next section) that preserved the normal functioning of the neural circuits involved in short-term spatial memory, and also in the neural processes that drive the animals to explore their surroundings, effects that could be considered as neuroprotective. Effects of chronic caffeine treatment on neuronal morphology in the hippocampus The most remarkable nding of the present study was that in the group of rats chronically treated with caffeine the neurons of the CA1 eld of the hippocampus had longer and more branched basal dendrites, and with greater numbers of spines, compared with the same neurons of the control group. As the size of neurites in CA1 pyramidal neurons has reached the adult levels by postnatal age P15 (Pokorn and Yamamoto, 1981), this nding can be explained in two possible ways. First, that chronic caffeine consumption had prevented the initial steps of atrophy in the basal dendrites of CA1 pyramidal neurons during the transition from young to middle age. However, although it has been reported that the apical dendrites of CA1 neurons experience degeneration at advanced ages (Lolova, 1989), at present there is no evidence that this degenerative process occurs at earlier ages. An alternative explanation is that the prolonged caffeine consumption had triggered some kind of trophic mechanism that selectively promoted the growth of the basal dendritic tree and spines in CA1 pyramidal neurons. There is evidence that the release of neurotrophic factors in the hippocampus is directly proportional to the ring rate of neurons, with the maximal release occurring at frequencies that induce long-term potentiation (LTP) (Balkowiec and Katz, 2002). As low caffeine concentrations ( 30 M)

induce spontaneous ring oscillations in the hippocampus (Pietersen et al., 2009) and facilitate the development of LTP (Costenla et al., 2010; Simons et al., in press), the possibility exists that during its chronic consumption caffeine had promoted the activity-dependent release of neurotrophic factors that stimulated the growth of more branches and spines in the basal dendrites of adult CA1 pyramidal neurons. Caffeine reaches an average concentration of 22 M in the hippocampus of rats that consumed high doses during 4 weeks (Costenla et al., 2010). As the animals used in the present study consumed smaller amounts of caffeine, it is conceivable that its brain levels attained values in the low micromolar range, at which caffeine binds preferentially to A1 and A2A adenosine receptors (A1Rs, A2ARs) (Fredholm et al., 1999). Then, it is plausible that the morphologic changes found in CA1 neurons of caffeine-treated rats were mediated through blockade of either of these receptors, or both. In the hippocampus, the blockade of A1Rs facilitates the induction of LTP in young animals, but not in those that have reached either middle or old age, which is explained by the greater number of A1Rs in the glutamatergic nerve endings at young age in comparison with the older age groups (Costenla et al., 2011). Remarkably, the greater expression of A1Rs occurs in the pyramidal neurons of the CA2 eld (Ochiishi et al., 1999), which are refractory to the development of LTP (Zhao et al., 2007), possibly through a strong inhibitory tone imposed by endogenous adenosine (Ochiishi et al., 1999). However, following in vivo dosing and in vitro application of caffeine, a robust and longlasting LTP can be elicited in CA2 neurons (Simons et al., in press). In the present study, chronic caffeine administration to rats started at 4 months of age, when hippocampal A1Rs are more abundant (Costenla et al., 2011). As the majority of synapses made by CA2 neurons occur on the proximal branches of basal dendrites of CA1 neurons (Ropireddy and Ascoli, 2011), we suggest that sustained blockade of A1Rs in CA2 pyramidal neurons during chronic caffeine treatment increased their basal ring rate, thus favoring the occurrence of LTP (Costenla et al., 2010; Simons et al., in press), and the release of neurotrophic factors (Balkowiec and Katz, 2002) that promoted the growth of 4th and 5th order branches and spines in the basal dendrites of CA1 pyramidal neurons. LTP is considered the functional substrate of memory, and both of these processes deteriorate during aging, starting at middle age (Kumar, 2011). The observation that mice carrying a mutation that selectively unmasks a robust LTP in CA2 neurons exhibit marked enhancement in spatial learning (Lee et al., 2010) leaves open the possibility of a causal relationship between the enlargement of the basal dendritic tree in CA1 neurons and the preservation of motor habituation and SAB in caffeine-treated rats. It is unlikely that the morphologic changes seen in CA1 neurons of caffeine-treated rats was mediated through blockade of A2ARs, as this suppresses the ring oscillations induced by adenosine (Pietersen et al., 2009), and inhibits the induction of LTP in the hippocampus (Costenla

394

S. Vila-Luna et al. / Neuroscience 202 (2012) 384395

et al., 2011). Although many studies have demonstrated that acute or chronic caffeine treatment attenuates memory impairment through A2AR blockade, these experiments have been performed in animal models of pathological conditions (e.g. emotional stress, diabetes, seizures, Parkinsons or Alzheimers diseases) that produce noxious brain insults which disrupt the inhibitory/facilitatory balance of adenosine in the hippocampus through down-regulation of A1Rs and up-regulation of A2ARs (Costenla et al., 2010; Cunha and Agostinho, 2010). However, a different picture would be expected in healthy young animals (as those used here), in which adenosinergic tone would be stronger in the more abundant A1Rs expressed in hippocampus at this age (Costenla et al., 2011). Hypothetically, by promoting the activity-dependent growth of dendrites and spines of hippocampal CA1 neurons in young subjects, caffeine would compensate the atrophy and loss of dendrites that occur at advanced ages (Lolova, 1989), thus providing an alternative mechanism by which chronic caffeine consumption could provide protection against aging-dependent cognitive decline. At present it is an enigma whether the morphological changes seen in the basal dendrites of CA1 neurons from caffeine-treated rats would have persisted beyond the 30day period of caffeine withdrawal. However, previous studies suggest that this could be the case, as juvenile sh raised during 50 days in water with a low caffeine concentration ( 7 M) show a greater number of dendrites and spines in their neurons when they reach the age of 21 months, in comparison with their controls of the same age (Burgess and Monachello, 1983). Similarly, neonatal rats chronically treated with high doses of caffeine develop prefrontal cortical neurons with longer and more branched dendrites at puberal and postpuberal ages (Jurez-Mndez et al., 2006).

REFERENCES
Altun M, Bergman E, Edstrm E, Johnson H, Ulfhake B (2007) Behavioral impairments of the aging rat. Physiol Behav 92:911923. Arendash GW, Mori T, Cao C, Mamcarz M, Runfeldt M, Dickson A, Rezai-Zadeh K, Tan J, Citron BA, Lin X, Echeverria V, Potter H (2009) Caffeine reverses cognitive impairment and decreases brain amyloid- levels in aged Alzheimers disease mice. J Alzheimers Dis 17:661 680. Arendash GW, Schleif W, Rezai-Zadeh K, Jackson EK, Zacharia LC, Cracchiolo JR, Shippy D, Tan J (2006) Caffeine protects Alzheimers mice against cognitive impairment and reduces brain -amyloid production. Neuroscience 142:941952. Balkowiec A, Katz DM (2002) Cellular mechanisms regulating activitydependent release of native brain-derived neurotrophic factor from hippocampal neurons. J Neurosci 22:10399 10407. Bert B, Fink H, Huston JP, Voits M (2002) Fischer 344 and Wistar rats differ in anxiety and habituation but not in water maze performance. Neurobiol Learn Mem 78:1122. Burgess JW, Monachello MP (1983) Chronic exposure to caffeine during early development increases dendritic spine and branch formation in midbrain optic tectum. Brain Res 282:123129. Cao C, Cirrito JR, Lin X, Wang L, Verges DK, Dickson A, Mamcarz M, Zhang C, Mori T, Arendash GW, Holtzman DM, Potter H (2009) Caffeine suppresses amyloid- levels in plasma and brain of Alzheimers disease transgenic mice. J Alzheimers Dis 17:681 697. Costa MS, Botton PH, Mioranzza S, Souza DO, Porcincula LO (2008) Caffeine prevents age-associated recognition memory decline and changes brain-derived neurotrophic factor and tirosine kinase receptor (TrkB) content in mice. Neuroscience 153:10711078. Costenla AR, Cunha RA, de Mendona A (2010) Caffeine, adenosine receptors, and synaptic plasticity. J Alzheimers Dis 20 (Suppl 1):S25S34. Costenla AR, Digenes MJ, Canas PM, Rodrigues RJ, Nogueira C, Maroco J, Agostinho PM, Ribeiro JA, Cunha RA, de Mendona A (2011) Enhanced role of adenosine A2A receptors in the modulation of LTP in the rat hippocampus upon ageing. Eur J Neurosci 34:1221. Cunha RA, Agostinho PM (2010) Chronic caffeine consumption prevents memory disturbance in different animal models of memory decline. J Alzheimers Dis 20 (Suppl 1):S95116. DallIgna OP, Fett P, Gomes MW, Souza DO, Cunha RA, Lara DR (2007) Caffeine and adenosine A2a receptor antagonists prevent beta-amyloid (2535)-induced cognitive decits in mice. Exp Neurol 203:241245. Diana M, Spiga S, Acquas E (2006) Persistent and reversible morphine withdrawal-induced morphological changes in the nucleus accumbens. Ann N Y Acad Sci 1074:446 457. Dobi A, Seabold GK, Christensen CH, Bock R, Alvarez VA (2011) Cocaine-induced plasticity in the nucleus accumbens is cell specic and develops without prolonged withdrawal. J Neurosci 31: 18951904. El Yacoubi M, Ledent C, Parmentier M, Costentin J, Vaugeois JM (2000) The anxiogenic-like effect of caffeine in two experimental procedures measuring anxiety in the mouse is not shared by selective A2A adenosine receptor antagonists. Psychopharmacology (Berl) 148:153163. Eskelinen MH, Kivipelto M (2010) Caffeine as a protective factor in dementia and Alzheimers disease. J Alzheimers Dis 20 (Suppl 1):S167S174. Espinola EB, Dias RF, Mattei R, Carlini EA (1997) Pharmacological activity of guarana (Paullinia cupana Mart.) in laboratory animals. J Ethnopharmacol 55:223229. Fraley SM, Springer AD (1981) Memory of simple learning in young, middle-aged, and aged C57/BL6 mice. Behav Neural Biol 31:17. Fredholm BB, Bttig K, Holmn J, Nehlig A, Zvartau EE (1999) Actions of caffeine in the brain with special reference to factors that contribute to its widespread use. Pharmacol Rev 51:83133.

CONCLUSION
These results of the present study agree with previous observations indicating that chronic caffeine treatment does not produce cognitive enhancement but prevents the cognitive decline associated with aging (Cunha and Agostinho, 2010). They also strengthen the ndings of epidemiological surveys suggesting that prolonged caffeine intake is associated with lower cognitive decline and a reduced risk of developing AD in elderly people. It is suggested that the longer basal dendrites and higher number of spines seen in CA1 hippocampal neurons of animals that consumed caffeine from young to middle age could be another mechanism by which caffeine helps to prevent the cognitive decline associated with aging.
AcknowledgmentsS.V.-L. and S.C.-I. contributed equally to the experimental part of this report. A scholarship to S.V.-L. and the behavioral part of this work were supported by CONACYT-Mexico grant 47763 to J.L.G.-A. The histological part was supported by CONACYT-Mexico grant 138663 to G.F.

S. Vila-Luna et al. / Neuroscience 202 (2012) 384395

Gerlai R (1998) A new continuous alternation task in T-maze detects hippocampal dysfunction in mice: a strain comparison and lesion study. Behav Brain Res 95:91101. Gibb R, Kolb B (1998) A method for vibratome sectioning of Golgi-cox stained whole rat brain. J Neurosci Methods 79:1 4. Gngora-Alfaro JL, Moo-Puc RE, Villanueva-Toledo JR, Alvarez-Cervera FJ, Bata-Garca JL, Heredia-Lpez FJ, Pineda JC (2009) Long-lasting resistance to haloperidol-induced catalepsy in male rats chronically treated with caffeine. Neurosci Lett 463:210 214. Hameleers PA, Van Boxtel MP, Hogervorst E, Riedel WJ, Houx PJ, Buntinx F, Jolles J (2000) Habitual caffeine consumption and its relation to memory, attention, planning capacity and psychomotor performance across multiple age groups. Hum Psychopharmacol 15:573581. Holtzman SG (1983) Complete, reversible, drug-specic tolerance to stimulation of locomotor activity by caffeine. Life Sci 33:779 787. Hooper N, Fraser C, Stone TW (1996) Effects of purine analogues on spontaneous alternation in mice. Psychopharmacology (Berl) 123:250 257. Johnson-Kozlow M, Kritz-Silverstein D, Barrett-Connor E, Morton D (2002) Coffee consumption and cognitive function among older adults. Am J Epidemiol 156:842 850. Jurez-Mndez S, Carretero R, Martnez-Tellez R, Silva-Gmez AB, Flores G (2006) Neonatal caffeine administration causes a permanent increase in the dendritic length of prefrontal cortical neurons of rats. Synapse 60:450 455. Kumar A (2011) Long-term potentiation at CA3-CA1 hippocampal synapses with special emphasis on aging, disease, and stress. Front Aging Neurosci 3:7. Lee SE, Simons SB, Heldt SA, Zhao M, Schroeder JP, Vellano CP, Cowan DP, Ramineni S, Yates CK, Feng Y, Smith Y, Sweatt JD, Weinshenker D, Ressler KJ, Dudek SM, Hepler JR (2010) RGS14 is a natural suppressor of both synaptic plasticity in CA2 neurons and hippocampal-based learning and memory. Proc Natl Acad Sci U S A 107:16994 16998. Leussis MP, Bolivar VJ (2006) Habituation in rodents: a review of behavior, neurobiology, and genetics. Neurosci Biobehav Rev 30: 10451064. Lindsay J, Laurin D, Verreault R, Hbert R, Helliwell B, Hill GB, McDowell I (2002) Risk factors for Alzheimers disease: a prospective analysis from the Canadian study of health and aging. Am J Epidemiol 156:445 453. Lolova I (1989) Dendritic changes in the hippocampus of aged rats. Acta Morphol Hung 37:310. Maia L, de Mendona A (2002) Does caffeine intake protect from Alzheimers disease? Eur J Neurol 9:377382. Miyagawa H, Hasegawa M, Fukuta T, Amano M, Yamada K, Nabeshima T (1998) Dissociation of impairment between spatial memory, and motor function and emotional behavior in aged rats. Behav Brain Res 91:73 81. Ochiishi T, Saitoh Y, Yukawa A, Saji M, Ren Y, Shirao T, Miyamoto H, Nakata H, Sekino Y (1999) High level of adenosine A1 receptorlike immunoreactivity in the CA2/CA3a region of the adult rat hippocampus. Neuroscience 93:955967. Paxinos G, Watson W (1986) The rat brain in stereotaxic coordinates, 2nd ed. San Diego: Academic Press. Pellow S, Chopin P, File SE, Briley M (1985) Validation of open:closed arm entries in an elevated plus-maze as a measure of anxiety in the rat. J Neurosci Methods 14:149 167. Pietersen AN, Lancaster DM, Patel N, Hamilton JB, Vreugdenhil M (2009) Modulation of gamma oscillations by endogenous adeno-

395

sine through A1 and A2A receptors in the mouse hippocampus. Neuropharmacology 56:481 492. Pokorn J, Yamamoto T (1981) Postnatal ontogenesis of hippocampal CA1 area in rats: part I. Development of dendritic arborisation in pyramidal neurons. Brain Res Bull 7:113120. Prut L, Belzung C (2003) The open eld as a paradigm to measure the effects of drugs on anxiety-like behaviors: a review. Eur J Pharmacol 463:333. Rieg T, Steigele H, Schnermann J, Richter K, Osswald H, Vallon V (2005) Requirement of intact adenosine A1 receptors for the diuretic and natriuretic action of the methylxanthines theophylline and caffeine. J Pharmacol Exp Ther 313:403 409. Ritchie K, Carrire I, de Mendonca A, Portet F, Dartigues JF, Rouaud O, Barberger-Gateau P, Ancelin ML (2007) The neuroprotective effects of caffeine: a prospective population study (the three city study). Neurology 69:536 545. Ropireddy D, Ascoli GA (2011) Potential synaptic connectivity of different neurons onto pyramidal cells in a 3D reconstruction of the rat hippocampus. Front Neuroinform 5:5. Rosso A, Mossey J, Lippa CF (2008) Caffeine: neuroprotective functions in cognition and Alzheimers disease. Am J Alzheimers Dis Other Demen 23:417 422. Santos C, Costa J, Santos J, Vaz-Carneiro A, Lunet N (2010a)Caffeine intake and dementia: systematic review and meta-analysis. J Alzheimers Dis 20 (Suppl 1):S187S204. Santos C, Lunet N, Azevedo A, de Mendona A, Ritchie K, Barros H (2010b) Caffeine intake is associated with a lower risk of cognitive decline: a cohort study from Portugal. J Alzheimers Dis 20 (Suppl 1):S175S185. Simons SB, Caruana DA, Zhao M, Dudek SM (in press) Caffeineinduced synaptic potentiation in hippocampal CA2 neurons. Nat Neurosci, doi:10.1038/nn.2962. Stone WS, Rudd RJ, Gold PE (1992) Glucose attenuation of scopolamine- and age-induced decits in spontaneous alternation behavior and regional brain [3H]2-deoxyglucose uptake in mice. Psychobiology 20:270 279. Takahashi RN, Pamplona FA, Prediger RD (2008) Adenosine receptor antagonists for cognitive dysfunction: a review of animal studies. Front Biosci 13:2614 2632. van Gelder BM, Buijsse B, Tijhuis M, Kalmijn S, Giampaoli S, Nissinen A, Kromhout D (2007) Coffee consumption is inversely associated with cognitive decline in elderly European men: the FINE study. Eur J Clin Nutr 61:226 232. Walf AA, Frye CA (2007) The use of the elevated plus maze as an assay of anxiety-related behavior in rodents. Nat Protoc 2:322328. Willig F, Palacios A, Monmaur P, MHarzi M, Laurent J, Delacour J (1987) Short-term memory, exploration and locomotor activity in aged rats. Neurobiol Aging 8:393 402. Xu K, Xu Y, Brown-Jermyn D, Chen JF, Ascherio A, Dluzen DE, Schwarzschild MA (2006) Estrogen prevents neuroprotection by caffeine in the mouse 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine model of Parkinsons disease. J Neurosci 26:535541. Yu L, Coelho JE, Zhang X, Fu Y, Tillman A, Karaoz U, Fredholm BB, Weng Z, Chen JF (2009) Uncovering multiple molecular targets for caffeine using a drug target validation strategy combining A2A receptor knockout mice with microarray proling. Physiol Genomics 37:199 210. Zhao M, Choi YS, Obrietan K, Dudek SM (2007) Synaptic plasticity (and the lack thereof) in hippocampal CA2 neurons. J Neurosci 27:1202512032.

(Accepted 24 November 2011) (Available online 2 December 2011)