Neuroscience Letters 372 (2004) 256–261

Treadmill exercise improves short-term memory by suppressingischemia-induced apoptosis of neuronal cells in gerbils

Young-Je Sima, Sung-Soo Kima, Jee-Youn Kima, Mal-Soon Shinb, Chang-Ju Kimb,∗a Research Institute of Sports Science, Korea University, #15-Ka Anam-dong, Sungbuk-ku, Seoul 136-701, Republic of Korea

b Department of Physiology, College of Medicine, Kyung Hee University, #1 Hoigi-dong, Dongdaemoon-gu, Seoul 130-701, Republic of Korea

Received 7 May 2004; received in revised form 22 September 2004; accepted 22 September 2004

Abstract

In the present study, the effect of treadmill exercise on short-term memory, apoptotic neuronal cell death, and cell proliferation in thehippocampal dentate gyrus following transient global ischemia in gerbils was investigated. Step-down inhibitory avoidance task, terminaldeoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) assay, and immunohistochemistry for caspase-3 and 5-bromo-2′-deoxyuridine (BrdU) were used. Ischemia was induced by the occlusion of both common carotid arteries (CCA) of gerbils for 5 min. Gerbils ine short-termm pressed thei mia-induceda©

K

Cdttdcca

ahsstapd

-lingpop-leathac-diedrs ofof

l roletinueshu-rusechedm-atho-yrusll

0d

xercise groups were forced to run on a treadmill for 30 min once a day for 10 consecutive days. Such treadmill exercise improvedemory by suppressing the ischemia-induced apoptotic neuronal cell death in the dentate gyrus. In addition, treadmill running sup

schemia-induced cell proliferation in the dentate gyrus. The present results suggest that treadmill exercise overcomes the ischepoptotic neuronal cell death and thus facilitates the recovery following ischemic cerebral injury.2004 Elsevier Ireland Ltd. All rights reserved.

eywords: Transient global ischemia; Apoptotic neuronal cell death; Cell proliferation; Short-term memory; Hippocampal dentate gyrus

erebral ischemia is induced by reduced cerebral blood flowue to the transient or permanent occlusion of cerebral ar-

eries[4]. Transient cerebral ischemia for 5 or 10 min leadso neuronal cell damage and eventually causes neuronal celleath[1,20]. Tissue damage following cerebral ischemia isaused by the interaction of complex pathophysiological pro-esses such as excitotoxicity, depolarization, inflammation,nd apoptosis[4].

Apoptosis, also known as programmed cell death, playscrucial role in the development and the maintenance of

omeostasis in all multicellular organisms[25,29]. Thomp-on reported that apoptosis is a form of cell death that con-titutes part of a common mechanism in cell replacement,issue remodeling, and removal of damaged cells[25]. In-ppropriate or excessive apoptosis, however, has been im-licated in many diseases including cancer, autoimmuneiseases, neurodegenerative disorders, acquired immunod-

∗ Corresponding author. Tel.: +82 2 961 0407; fax: +82 2 964 2195.E-mail address:changju@khu.ac.kr (C.-J. Kim).

eficiency syndrome (AIDS), and stroke[25]. Terminal deoxynucleotidyl transferase-mediated dUTP nick end labe(TUNEL) is an assay that detects the characteristic of atotic cell death DNA fragmentation[10]. In ischemic animamodels, TUNEL-positive cells represent apoptotic cell d[20]. Another important characteristic of apoptosis is thetivation of caspase-3. Caspase-3 is the most widely stumember of the caspase family and one of key executoapoptosis[3]. In ischemic animal models, the activationcaspase-3 is implicated in apoptotic neuronal cell death[1].

The neurogenesis in the hippocampus plays a centrain learning and memory process. The neurogenesis conthroughout life in the adult mammalian brain includingmans[6]. Cell proliferation in the hippocampal dentate gyis increased by learning, serotonin,N-methyl-d-aspartat(NMDA) receptor antagonists, and exposure to an enrienvironment[9]. Increased cell proliferation in the hippocapal dentate gyrus has also been observed in some plogical states including seizure, mechanical dentate glesions, and ischemic insult[21]. Such up-regulation of ce

304-3940/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.neulet.2004.09.060

Y.-J. Sim et al. / Neuroscience Letters 372 (2004) 256–261 257

proliferation during ischemic insult is considered as the com-pensatory adaptive response to the increased apoptotic neu-ronal cell death following ischemic injury[21].

It has been reported that physical exercise exertsneuroprotective effect on hippocampal injury that inducesmemory impairment. Physical exercise has been known tofacilitate recovery from injury and improves cognitive func-tion [8,16]. In addition, physical exercise increases the ex-pression of neurotrophic factors such as insulin-like growthfactor-I (IGF-I) and basic fibroblast growth factor (bFGF).These neurotrophic factors are implicated in neuronal sur-vival, differentiation, the alteration in synaptic plasticity, andmemory function[27].

In the present study, the effect of treadmill exercise onshort-term memory, apoptotic neuronal cell death, and cellproliferation in the hippocampal dentate gyrus followingtransient global ischemia in gerbils was investigated. The ef-fect was assessed in step-down inhibitory avoidance task,TUNEL assay, and immunohistochemistry for caspase-3 and5-bromo-2′-deoxyuridine (BrdU).

Adult male Mongolian gerbils (11–13 weeks of age)were used in this experiment. The experimental procedureswere performed in accordance with the animal care guide-lines of the National Institute of Health (NIH) and the Ko-rean Academy of Medical Sciences. Gerbils were housedu ◦( vail-a fourg trol)g emia-i rciseg

icalpI 0%O ex-p witha vedt wasm r-b werem also ceptt .

singt Oned n in-h n the7 t ontg 1 cma mblef tiont revi-o na e la-

tency of step-down inhibitory avoidance task. The latencyover 180 s was counted as 180 s.

From the second day after surgery, gerbils in exercisegroups were forced to run on a motorized treadmill for 30 minonce a day for 10 consecutive days. The exercise workloadconsisted of running at the speed 2 m/min for the first 5 min,5 m/min for the next 5 min, and 8 m/min for the last 20 minwith the 0◦ inclination. Animals in non-exercise groups(sham-operation group and ischemia-induction group) wereleft on a treadmill for 30 min and without forcing to run.One hour prior to the start of treadmill running, all gerbils re-ceived 50 mg/kg BrdU (Sigma Chemical Co., St. Louis, MO)intraperitoneally for 10 consecutive days.

Gerbils were sacrificed after the completion of last tread-mill running on the 11th day of the experiment. Animalswere anesthetized using Zoletil 50® (10 mg/kg, i.p.; VibacLaboratories, Carros, France), transcardially perfused with50 mM phosphate-buffered saline (PBS), and fixed with afreshly prepared solution consisting of 4% paraformaldehydein 100 mM phosphate buffer (PB, pH 7.4). The brains weredissected and postfixed in the same fixative overnight andtransferred into a 30% sucrose solution for cryoprotection.Coronal sections of 40�m thickness were made with a freez-ing microtome (Leica, Nussloch, Germany).

For the visualization of the DNA fragmentation, TUNELs tec-t them di tionsw u-b 0,r ure.T rter-P ’sh forc ontog ighta singP

pase-3 de-s ghtw CruzB 1 hw unds liteA hea izedu ntog ighta singP

tateg eda bi-l in,

nder the controlled temperature (20± 2 C) and lighting07:00–19:00 h) conditions. Food and water were made able ad libitum. Gerbils were randomly assigned intoroups (n= 10 in each group): the sham-operation (conroup, the sham-operation and exercise group, the isch

nduction group, and the ischemia-induction and exeroup.

To induce transient global ischemia in gerbils, the surgrocedure were performed as previously described[15,19].

n brief, gerbils were anesthetized with 3% isoflurane in 22–77% N2. Both common carotid arteries (CCAs) wereosed by bilateral neck incision. CCAs were occludedneurysm clips for 5 min and the clips were then remo

o restore cerebral blood flow. The rectal temperatureaintained at 37± 0.5◦C with a heating lamp until the geil regained consciousness. After recovery, the animalsonitored for additional 2 h to prevent hypothermia. Animf the sham-operation groups were treated identically, ex

hat the CCAs were not occluded after the neck incisionShort-term memory ability was evaluated by asses

he latency of the step-down inhibitory avoidance task.ay after surgery, gerbils were trained in the step-dowibitory avoidance task as follows. Gerbils were placed ocm× 25 cm platform 2.5 cm in height and allowed to res

he platform for 2 min. The platform faced a 42 cm× 25 cmrid of parallel 0.1-cm caliber stainless steel bars spacedpart. In training sessions, animals received 0.3 mA scra

ootshock for 2 s immediately upon stepping down. Retenime was determined at the 10th day after surgery as pusly described[14]. The interval of gerbils stepping downd placing all four paws on the grid was defined as th

taining was performed using In Situ Cell Death Deion Kit® (Roche, Mannheim, Germany) according toanufacturer’s protocol[15,19]. Briefly, sections were fixe

n ethanol–acetic acid (2:1) and rinsed. Then the secere incubated with 100�g/ml proteinase K, rinsed, incated in 3% H2O2, permeabilized with 0.5% Triton X-10insed again, and incubated in the TUNEL reaction mixthe sections were rinsed and visualized using ConveOD with 0.02% 3,3′-diaminobenzidine (DAB). Mayerematoxylin (DAKO, Glostrup, Denmark) was usedounter-staining and the sections were finally mountedelatin-coated slides. The slides were air dried overnt room temperature, and coverslips were mounted uermount®.For the visualization of the caspase-3 expression, casimmunohistochemistry was performed as previously

cribed [15,19]. Brain sections were incubated overniith mouse anti-caspase-3 antibody (1:500; Santaiotechnology, Santa Cruz, CA) and then for anotherith the biotinylated mouse secondary antibody. Boecondary antibody was then amplified with Vector EBC kit® (Vector Laboratories, Burlingame, CA). Tntibody–biotin–avidin–peroxidase complex was visualsing 0.02% DAB. The sections were finally mounted oelatin-coated slides. The slides were air dried overnt room temperature, and coverslips were mounted uermount®.For the detection of newly generated cells in the den

yrus, BrdU-specific immunohistochemistry was performs previously described[19]. Brain sections were permea

ized by incubation in 0.5% Triton X-100 in PBS for 20 m

258 Y.-J. Sim et al. / Neuroscience Letters 372 (2004) 256–261

then pretreated in 50% formamide–2× standard saline cit-rate (SSC) at 65◦C for 2 h, denaturated in 2N HCl at 37◦Cfor 30 min, and rinsed twice in 100 mM sodium borate (pH8.5). Afterwards, the sections were incubated overnight at4◦C with BrdU-specific mouse monoclonal antibody (1:600;Roche, Mannheim). The sections were then washed threetimes with PBS and incubated for 1 h with the biotiny-lated mouse secondary antibody (1:200; Vector Laborato-ries). Then the sections were incubated for another 1 h withavidin–peroxidase complex (1:100; Vector Laboratories). Forvisualization, the sections were incubated for 5 min in 50 mMTris–HCl (pH 7.6) containing 0.02% DAB, 40 mg/ml nickelchloride, and 0.03% hydrogen peroxide. Subsequently, theslides were air dried overnight at room temperature and cov-erslips were mounted using Permount®.

For the evaluation of the data of step-down inhibitoryavoidance task, the difference in the latency among groupswas evaluated by Kruskal–Wallis ANOVA. The individualdifference between groups was evaluated by Mann–WhitneyU tests. The data are presented as the median value (interquar-tile range).

The number of TUNEL-positive and caspase-3-positivecells in the dentate gyrus was counted hemilaterally in ev-ery eighth section throughout the entire extent of the dentategyrus at 400× magnification. The number of BrdU-positivec tateg theg age-P lverS L-p s ex-p are oft ntateg rm -wayAw

ngedf rciseg enceb rationa e la-t from3 roupw -c l ex-e ntly.T d thei

rusoa was3 ofT oupa t sta-t the

Fig. 1. The effect of treadmill exercise on latency of step-down inhibitoryavoidance task after transient global ischemia in gerbils. The data are repre-sented as the median value (interquartile range).* P< 0.05 compared to thesham-operation group.#P< 0.05 compared to the ischemia-induction group.(A) Sham-operation group; (B) sham-operation and exercise group; (C)ischemia-induction group; and (D) ischemia-induction and exercise group.

number of TUNEL-positive cells in the dentate gyrus wasincreased to 67.99 mm−2 ± 5.31 mm−2. In the ischemia-induction and exercise group, on the other hand, the numberof TUNEL-positive cells in the dentate gyrus was reduced to39.95 mm−2 ± 3.92 mm−2 (Fig. 2). The present results showthat ischemic insult-induced apoptotic neuronal cell death inthe dentate gyrus and treadmill exercise significantly sup-pressed the ischemia-induced apoptosis of neuronal cells.

The number of caspase-3-positive cells in the dentategyrus of the sham-operation group was 16.41 mm−2 ±2.78 mm−2 and in the sham-operation and exercise groupwas 29.84 mm−2 ± 6.24 mm−2. The difference of the num-ber of caspase-3-positive cells between the sham-operationgroup and the sham-operation and exercise group was notstatistically significant. In the ischemia-induction group, thenumber of caspase-3-positive cells in the dentate gyrus wasincreased to 91.21 mm−2 ± 23.92 mm−2. In the ischemia-induction and exercise group, on the other hand, the numberof caspase-3-positive cells in the dentate gyrus was re-duced to 35.94 mm−2 ± 7.23 mm−2 (Fig. 3). The presentresults show that ischemic insult induced the caspase-3 expression in the dentate gyrus and treadmill exercisesignificantly suppressed the ischemia-induced caspase-3expression.

The number of BrdU-positive cells in the dentate gyrus oft −2 −2

a was3 edc iong tategi d, then de-c tr s in-c dmill

ells in the subgranular layer of the hippocampal denyrus was counted in a similar fashion. The area ofranular layer of dentate gyrus was traced using Imro®Plus image analyzer (Media Cybernetics Inc., Sipring, MD) at 40× magnification. The number of TUNEositive, caspase-3-positive, and BrdU-positive cells waressed as the mean number of cells per millimetre squ

he cross sectional area of the granular layer of the deyrus. The data are expressed as the mean± standard erroean (S.E.M.). For the comparison between groups, oneNOVA and Duncan’s post hoc test were performed.P< 0.05as considered statistically significant.The latency in the sham-operation group was 175 s, ra

rom 160 to 180 s, and in the sham-operation and exeroup was 180 s, ranged from 175 to 180 s. The differetween the sham-operation group and the sham-opend exercise group was not statistically significant. Th

ency in the ischemia-induction group was 46 s, ranged0 to 66 s, and in the ischemia-induction and exercise gas 117 s, ranged from 70 to 167 s (Fig. 1). The latency in ishemic gerbils was shorter than control gerbils. Treadmilrcise increased the latency in ischemic gerbils significahe present results show that treadmill exercise improve

schemia-induced short-term memory impairment.The number of TUNEL-positive cells in the dentate gy

f the sham-operation group was 34.98 mm−2 ± 2.73 mm−2

nd in the sham-operation and exercise group4.10 mm−2 ± 5.01 mm−2. The difference of the numberUNEL-positive cells between the sham-operation grnd the sham-operation and exercise group was no

istically significant. In the ischemia-induction group,

he sham-operation group was 215.72 mm± 23.66 mmnd in the sham-operation and exercise group23.54 mm−2 ± 11.63 mm−2. Treadmill exercise increasell proliferation in control gerbils. In the ischemia-inductroup, the number of BrdU-positive cells in the denyrus was increased to 693.09 mm−2 ± 47.17 mm−2. In the

schemia-induction and exercise group, on the other hanumber of BrdU-positive cells in the dentate gyru wasreased to 469.82 mm−2 ± 37.63 mm−2 (Fig. 4). The presenesults show that cell proliferation in the dentate gyrus wareased in response to transient global ischemia and trea

Y.-J. Sim et al. / Neuroscience Letters 372 (2004) 256–261 259

Fig. 2. The effect of treadmill exercise on DNA fragmentation in the den-tate gyrus after transient global ischemia. (Upper): A photomicrograph ofterminal deoxynucleotidyl transferase-mediated dUTP nick end labeling(TUNEL)-positive cells (arrows). Gcl, granular cell layer; Hil, hilus. A scalebar represents 25�m. (Lower): The mean number of TUNEL-positive cells.The data are represented as the mean± S.E.M. * P< 0.05 compared to thesham-operation group.#P< 0.05 compared to the ischemia-induction group.(A) Sham-operation group; (B) sham-operation and exercise group; (C)ischemia-induction group; and (D) ischemia-induction and exercise group.

exercise suppressed the ischemia-induced cell proliferationin the dentate gyrus.

Cerebral ischemia induces the deprivation of oxygenand glucose resulting in tissue infarction and neuronal celldeath[1,4,20]. Apoptosis is implicated in ischemia and neu-rodegenerative diseases including Parkinson’s disease andAlzheimer’s disease[25]. The morphological characteristicsof the apoptotic cell death are cell shrinkage, chromatin con-densation, membrane blebbing, and DNA fragmentation[20].In addition, the up-regulation and activation of caspase-3enzyme at the early stage of apoptosis following ischemiahas been reported[1]. In the present study, the number ofTUNEL-positive and caspase-3-positive cells in the dentategyrus was significantly increased following transient globalischemia. This indicates that ischemia induces apoptotic neu-ronal cell death in the dentate gyrus.

Recent studies reported that granular cell proliferation isincreased following transient ischemia in gerbils[21]. Ourdata also show that cell proliferation in the dentate gyrus

Fig. 3. The effect of treadmill exercise on the caspase-3 expression in thedentate gyrus after transient global ischemia. (Upper): A photomicrograph ofcaspase-3-positive cells (arrows). Gcl, granular cell layer; Hil, hilus. A scalebar represents 50�m. (Lower): The mean number of caspase-3-positive.The data are represented as the mean± S.E.M. * P< 0.05 compared to thesham-operation group.#P< 0.05 compared to the ischemia-induction group.(A) Sham-operation group; (B) sham-operation and exercise group; (C)ischemia-induction group; and (D) ischemia-induction and exercise group.

was increased in response to cerebral ischemia. It has beensuggested that enhanced cell proliferation in the dentate gyrusis a compensatory response to excessive apoptotic cell death[19,21].

Running wheel exercise enhances the neurogenesis in thedentate gyrus of adult mice[28]. Trejo et al. showed thatphysical exercise increases cell proliferation and the uptakeof the neurotrophic factor IGF-I from the bloodstream intospecific brain areas including the hippocampus in rats[27]. Inthe present study, treadmill exercise increased the generationof new cells in the dentate gyrus in sham-operation gerbils. Inregard to apoptotic neuronal cell death in sham-operation ger-bils, the difference between exercise group and non-exercisegroup was not statistically significant. Mild-intensity tread-mill exercise was reported to increase cell proliferation with-out inducing apoptosis in the dentate gyrus under normalconditions[19].

Physical exercise facilitates recovery of ischemia-inducedneurological impairment and reduces the expression of

260 Y.-J. Sim et al. / Neuroscience Letters 372 (2004) 256–261

Fig. 4. The effect of treadmill exercise on cell proliferation in the den-tate gyrus after transient global ischemia. (Upper): Photomicrographs of5-bromo-2′-deoxyuridine (BrdU)-positive cells. A scale bar represents100�m. (Lower): The mean number of BrdU-positive. The data are rep-resented as the mean± S.E.M. * P< 0.05 compared to the sham-operationgroup. #P< 0.05 compared to the ischemia-induction group. (A) Sham-operation group; (B) sham-operation and exercise group; (C) ischemia-induction group; and (D) ischemia-induction and exercise group.

mRNA of apoptosis-associated genes including Bcl-x� andDP5[16,26]. Guezennec et al. reported that treadmill runningdown-regulates the expression of glutamate receptors, whichare implicated in excitotoxicity[12]. Carro et al. showedthat treadmill running at low intensity prevents the neu-ronal cell loss as well as the impairment induced by exci-totoxic damage or energy imbalance[2]. The present resultsalso showed that treadmill exercise significantly suppressesthe ischemia-induced increase in DNA fragmentation, thecaspase-3 expression, and cell proliferation in the dentategyrus. The suppression of cell proliferation by treadmill ex-ercise may be attributed to the reduction of apoptotic celldeath by treadmill exercise[19].

It has been suggested that the hippocampal neurogenesismay be affected by memory formation[11]. Ischemic injuryto the hippocampus is known to induce memory impairment[24]. Physical exercise alters neurotransmitter levels and neu-ronal metabolism in the rat hippocampus. It has been reportedthat activation of cholinergic fibers increases acetylcholinerelease and produces vasodilation in the rat hippocampus,resulting in increased cerebral blood flow in the hippocam-

pus during exercise[5,23]. Nakajima et al. demonstrated thatthe increase in cerebral blood flow in the hippocampus dur-ing walking is at least partly responsible for the improvementin spatial learning ability associated with hippocampal acti-vation or hippocampal neurogenesis[23]. Increase in bloodflow in the hippocampus due to physical exercise also resultsin the provision of sufficient oxygen and glucose to the hip-pocampus; this appears to be advantageous for the neurons ofthe hippocampus in the sense that protection against delayedneuronal cell death induced by insufficiency of blood supplyis afforded[13,17]. It was reposted that running wheel en-hances the hippocampal-related spatial learning in rats andaerobic exercise improves cognitive function such as learn-ing and memory in the elderly[7,28]. It has been shown thatphysical exercise prevents cognitive decline during aging andfacilitates functional recovery after central nervous systeminjury [18,22].

We evaluated the effect of treadmill exercise on theischemia-induced short-term memory impairment using la-tency of step-down inhibitory avoidance task. The differencein the latency between exercise group and non-exercise groupin gerbils without ischemia was not statistically significant.The latency shortened by transient global ischemia, on theother hand, was significantly improved by treadmill exercise.

The present study shows that treadmill exercise improvess uceda esentr nt ins duet mille viatec em-o

A

ationG

R

o, D.in the7134.tingicaly, J.

em. J.

mic7.fromtion

hort-term memory by suppressing the ischemia-indpoptotic neuronal cell death in the dentate gyrus. The presults thus provide direct evidence that the improvemehort-term memory associated with treadmill exercise iso the reduction of apoptotic neuronal cell death by treadxercise. This suggests that treadmill exercise may alleentral nervous system complications and short-term mry deficits induced by cerebral ischemia.

cknowledgement

This work was supported by a Korea Research Foundrant (KRF-2003-050-G00005).

eferences

[1] A. Benchoua, C. Guegan, C. Couriaud, H. Hosseini, N. SampaiMorin, B. Onteniente, Specific caspase pathways are activatedtwo stages of cerebral infarction, J. Neurosci. 21 (2001) 7127–

[2] E. Carro, J.L. Trejo, S. Busiguina, I. Torres-Aleman, Circulainsulin-like growth factor I mediates the protective effects of physexercise against brain insults of different etiology and anatomNeurosci. 21 (2001) 5678–5684.

[3] G.M. Cohen, Caspases: the executioners of apoptosis, Bioch326 (1997) 1–16.

[4] U. Dirnagl, C. Iadecola, M.A. Moskowitz, Pathobiology of ischaestroke: an integrated view, Trends Neurosci. 22 (1999) 391–39

[5] J.D. Dudar, I.Q. Whishaw, J.C. Szerb, Release of acetylcholinethe hippocampus of freely moving rats during sensory stimulaand running, Neuropharmacology 18 (1979) 673–678.

Y.-J. Sim et al. / Neuroscience Letters 372 (2004) 256–261 261

[6] P.S. Eriksson, E. Perfilieva, T. Bjok-Eriksson, A.M. Alborn, C. Nord-borg, D.A. Peterson, F.H. Gage, Neurogenesis in the adult humanhippocampus, Nat. Med. 11 (1998) 1313–1317.

[7] C. Fabre, K. Chamari, P. Mucci, J. Masse-Biron, C. Prefaut, Improve-ment of cognitive function by mental and/or individualized aerobictraining in healthy elderly subjects, Int. J. Sports Med. 23 (2002)415–421.

[8] D.E. Fordyce, J.M. Wehner, Physical activity enhances spatial learn-ing performance with an associated alteration in hippocampal pro-tein kinase C activity in C57BL/6 and DBA/2 mice, Brain Res. 619(1993) 111–119.

[9] E. Fuchs, E. Gould, Mini-review: in vivo neurogenesis in the adultbrain: regulation and functional implications, Eur. J. Neurosci. 12(2000) 2211–2214.

[10] Y. Gavrieli, Y. Sherman, S.A. Ben-Sasson, Identification of pro-grammed cell death in situ via specific labeling of nuclear DNAfragmentation, J. Cell Biol. 119 (1992) 493–501.

[11] E. Gould, A. Beylin, P. Tanapat, A. Reeves, T.J. Shors, Learning en-hances adult neurogenesis in the hippocampal formation, Nat. Neu-rosci. 2 (1999) 260–265.

[12] C.Y. Guezennec, A. Abdelmalki, B. Serrurier, D. Merino, X. Bigard,M. Berthelot, C. Pierard, M. Peres, Effects of prolonged exerciseon brain ammonia and amino acids, Int. J. Sports Med. 19 (1998)323–327.

[13] H. Hotta, S. Uchida, F. Kagitani, Effects of stimulating the nucleusbasalis of Meynert on blood flow and delayed neuronal death fol-lowing transient ischemia in the rat cerebral cortex, Jpn. J. Physiol.52 (2002) 383–393.

[14] I. Izquierdo, C. Fin, P.K. Schmitz, R.C. Da Silva, D. Jerusalinsky,J.A. Quillfeldt, M.B. Ferreira, J.H. Medina, N.G. Bazan, Memory

rhinaloid-.

[ H.mia-ocam-

[ , andebral

[17] F. Kagitani, S. Uchida, H. Hotta, A. Sato, Effects of nicotineon blood flow and delayed neuronal death following intermittenttransient ischemia in rat hippocampus, Jpn. J. Physiol. 50 (2000)585–595.

[18] D. Laurin, R. Verreault, J. Lindsay, K. MacPherson, K. Rockwood,Physical activity and risk of cognitive impairment and dementia inelderly persons, Arch. Neurol. 58 (2001) 498–504.

[19] M.H. Lee, H. Kim, S.S. Kim, T.H. Lee, B.V. Lim, H.K. Chang,M.H. Jang, M.C. Shin, M.S. Shin, C.J. Kim, Treadmill exercisesuppresses ischemia-induced increment in apoptosis and cell prolif-eration in hippocampal dentate gyrus of gerbils, Life Sci. 73 (2003)2455–2465.

[20] Y. Li, M. Chopp, N. Jiang, Z.G. Zhang, C. Zaloga, Induction of DNAfragmentation after 10 to 120 minutes of focal cerebral ischemia inrats, Stroke 26 (1995) 1252–1257.

[21] J. Liu, K. Solway, R.O. Messing, F.R. Sharp, Increased neurogenesisin the dentate gyrus after transient global ischemia in gerbils, J.Neurosci. 18 (1998) 7768–7778.

[22] M.P. Mattson, Neuroprotective signaling and the aging brain: takeaway my food and let me run, Brain Res. 886 (2000) 47–53.

[23] K. Nakajima, S. Uchida, A. Suzuki, H. Hotta, Y. Aikawa, The ef-fect of walking on regional blood flow and acetylcholine in thehippocampus in conscious rats, Auton. Neurosci. 103 (2003) 83–92.

[24] L.R. Squire, S.M. Zola, Ischemic brain damage and memory impair-ment: a commentary, Hippocampus 6 (1996) 546–552.

[25] C.B. Thompson, Apoptosis in the pathogenesis and treatment ofdisease, Science 267 (1995) 1456–1462.

[26] L. Tong, H. Shen, V.M. Perreau, R. Balazs, C.W. Cotman, Effectsof exercise on gene-expression profile in the rat hippocampus, Neu-

[ ikember001)

[ s cell, Nat.

[ 407

enhancement by intrahippocampal, intraamygdala, or intraentoinfusion of platelet-activating factor measured in an inhibitory avance task, Proc. Natl. Acad. Sci. U.S.A. 92 (1995) 5047–5051

15] M.H. Jang, M.C. Shin, T.H. Lee, B.V. Lim, M.S. Shin, B.I. Min,Kim, S. Cho, E.H. Kim, C.J. Kim, Acupuncture suppresses ischeinduced increase in c-Fos expression and apoptosis in the hipppal CA1 region in gerbils, Neurosci. Lett. 347 (2003) 5–8.

16] B.B. Johansson, A.L. Ohlsson, Environment, social interactionphysical activity as determinants of functional outcome after cerinfarction in the rat, Exp. Neurol. 139 (1996) 322–327.

robiol. Dis. 8 (2001) 1046–1056.27] J.L. Trejo, E. Carro, I. Torres-Aleman, Circulating insulin-l

growth factor I mediates exercise-induced increases in the nuof new neurons in the adult hippocampus, J. Neurosci. 21 (21628–1634.

28] H. van Praag, G. Kempermann, F.H. Gage, Running increaseproliferation and neurogenesis in the adult mouse dentate gyrusNeurosci. 2 (1999) 266–270.

29] J. Yuan, B.A. Yankner, Apoptosis in the nervous system, Nature(2000) 802–809.