Chronic stress-induced memory deficits are reversed by regular exercise via AMPK-mediated BDNF induction

Accepted Manuscript Chronic stress-induced memory deficits is reversed by regular exercise via AMPK-mediated BDNF induction Dong-Moon Kim, Yea-Hyun Leem PII: DOI: Reference:

S0306-4522(16)00242-6 http://dx.doi.org/10.1016/j.neuroscience.2016.03.019 NSC 16983

To appear in:

Neuroscience

Accepted Date:

7 March 2016

Please cite this article as: D-M. Kim, Y-H. Leem, Chronic stress-induced memory deficits is reversed by regular exercise via AMPK-mediated BDNF induction, Neuroscience (2016), doi: http://dx.doi.org/10.1016/j.neuroscience. 2016.03.019

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

Chronic stress-induced memory deficits is reversed by regular exercise via AMPKmediated BDNF induction

Dong-Moon Kim1, Yea-Hyun Leem2,*

1

Department of Society of Sports & Leisure Studies, Wonkwang University, 460 Iksandea-ro,

Iksan, Jeonbuk, Republic of Korea 2

Department of Molecular Medicine and TIDRC, School of Medicine, Ewha Women’s

University, Seoul 158-710, Republic of Korea

Running title: Exercise restores cognitive function in chronic stress through AMPK

Grant sponsor: Wonkwang University in 2014 and National Research Foundation of Korea funded by the Korean Government, Grant number: NRF-2013R1A1A2062984

*

Corresponding author: Yea-Hyun Leem, Ph.D.,

Department of Neuroscience and TIDRC, Ewha Womans University, Mokdong Hospital, 911-1 Mok-Dong, Yangcheon-Ku, Seoul 158-710, Republic of Korea. Tel: +82-2-2650-5749, Fax: +82-2-2650-5850, Email: [email protected]

2

ABSTRACT

Chronic stress has a detrimental effect on neurological insults, psychiatric deficits, and cognitive impairment. In the current study, chronic stress was shown to impair learning and memory functions, in addition to reducing in hippocampal Adenosine monophosphateactivated protein kinase (AMPK) activity. Similar reductions were also observed for brainderived neurotrophic factor (BDNF), synaptophysin, and post-synaptic density-95 (PSD-95) levels, all of which was counter-regulated by a regime of regular and prolonged exercise. A 21-day restraint stress regimen (6 h/day) produced learning and memory deficits, including reduced alternation in the Y-maze and decreased memory retention in the water maze test. These effects were reversed post-administration by a 3-week regime of treadmill running (19 m/min, 1 h/day, 6 days/week). In hippocampal primary culture, phosphorylated-AMPK (phospho-AMPK) and BDNF levels were enhanced in a dose-dependent manner by 5amimoimidazole-4-carboxamide riboside (AICAR) treatment, and AICAR-treated increase was blocked by Compound C. A 7-day period of AICAR intraperitoneal injections enhanced alternation in the Y-maze test and reduced escape latency in water maze test, along with enhanced phospho-AMPK and BDNF levels in the hippocampus. The intraperitoneal injection of Compound C every 4 days during exercise intervention diminished exerciseinduced enhancement of memory improvement during the water maze test in chronically stressed mice. Also, chronic stress reduced hippocampal neurogenesis (lower Ki-67- and doublecortin-positive cells) and mRNA levels of BDNF, synaptophysin, and PSD-95. Our results suggest that regular and prolonged exercise can alleviate chronic stress-induced hippocampal dependent memory deficits. Hippocampal AMPK-engaged BDNF induction is at least in part required for exercise-induced protection against chronic stress.

3

Key words: chronic restraint stress, treadmill running, learning and memory, AMPK, BDNF

1 INTRODUCTION

Stress causes widespread changes to the neurochemical, neurobiological, and behavioral responses of the brain. Accumulating evidence also suggests that chronic stress negatively affects neural plasticity to produce deficits in memory and learning processes (Sandi and Pinelo-Nava, 2007; Krishnan and Nestler, 2008). The detrimental effect of chronic stress on cognitive function is suggested to be modulated by corticosterone, neurotrophins, oxidative stress, and various neurotransmitters (McGaugh and Roozendaal, 2002; Yamada and Nabeshima, 2003; Sandi and Pinelo-Nava, 2007; Calabrese et al., 2012; Kwon et al., 2013). Among the brain structures affected, the hippocampus is a region commonly implicated in repeated or chronic stress-triggered abnormalities of neural plasticity. Such abnormalities include hippocampal atrophy, decreased neurogenesis, and impaired synaptic plasticity (Watanabe et al., 1992; Sousa et al., 2000; Pham et al., 2003; Han et al., 2015). Since a large amount of energy is required to fulfill the physiological demands of neurons in the central nervous system (CNS), dysregulation of energy metabolism will deleteriously affect their survival and function. Adenosine monophosphate-activated protein kinase (AMPK) is an energy metabolite-sensing protein kinase that contributes to regulating cellular energy homeostasis (Spasic et al., 2009; Steinberg and Kemp, 2009). The phosphorylation of AMPK on threonine 172, producing phospho-AMPK, stimulates catabolic processes such as glucose

uptake,

glycolysis,

and

fatty

acid

oxidation.

Correspondingly,

AMPK

4

phosphorylation also suppresses anabolic process, including the synthesis of fatty acid, cholesterol, and protein, to restore cellular energy levels (Hadad et al., 2008; Ronnett et al., 2009). The activation of AMPK through phosphorylation is triggered by ATP depletion (increased AMP/ATP ratio), metabolic stresses (hypoxia, glucose deprivation, oxidative stress), and exercise (Culmsee et al., 2001; McCullough et al., 2005; Hardie, 2007). Furthermore, AMPK is also phosphorylated by Ca2+/calmodulin-dependent protein kinase β, suggesting that the kinase activity of this molecule is regulated indirectly by intracellular Ca2+ levels (Woods et al., 2005). As addressed above, AMPK is considered to play a crucial role in CNS neuronal responses to various physiological and pathological stimuli, particularly within the hippocampus. Supporting this, modest activation of AMPK in the hippocampus by diet restriction improves cognitive function and enhances hippocampal neurogenesis (Dagon et al., 2005). Similarly, Resveratrol treatment elicits activation of hippocampal AMPK and alleviates prenatal stress-induced memory impairment in pups (Cao et al., 2014). These articles suggest a potential role for hippocampal AMPK activation in the modulation of cognitive function. Furthermore, a recent study showed that the induction of unpredictable chronic mild stress for a 4-week period results in the inactivation of AMPK and the emergence of abnormal mood-related behaviors (Zhu et al., 2014). The anti-depressive actions of ketamine appear to require both availability of brain-derived neurotrophic factor (BDNF) and AMPK activation, with the activation of AMPK causing induction of BDNF expression (Yoon et al., 2008; Autry et al., 2011; Xu et al., 2013). The neuroprotective role of BDNF and its link to cognitive function is well established. BDNF contributes to the promotion of long-term potentiation, enhanced synaptic plasticity, and improved cognitive function (Conner et al., 1997; Duman and Monteggia, 2006). Judging from these studies, alterations in hippocampal AMPK activity may be linked to stress-induced memory impairment.

5

Physical exercise is renowned for its ability to improve brain function, influencing both cognitive function and mood (Hillman et al., 2008). In particular, the effect of exercise on performance in hippocampal dependent memory tasks is believed to be associated with hippocampal neurogenesis, synaptic plasticity and neurotrophins (Eadie et al., 2005; GomezPinilla et al., 2008; Hillman et al., 2008; van Praag, 2008). In addition, AMPK is highly expressed in brain regions such as the hippocampus and plays a crucial role in exercise physiology (Hadie, 2004; Spasic et al., 2009). However, the involvement of hippocampal AMPK in chronic stress-induced memory impairment and its subsequent reversal with exercise is still poorly understood. To unravel this issue, we explored whether exercise could alleviate chronic stress-induced memory impairment using pharmacological activation and/or inhibition of AMPK.

2 EXPERIMENTAL PROCEDURES

2.1 Experimental subjects Male 7-week-old C57BL/6 mice were obtained from Daehan Biolink, Inc. (Eumsung, Chungbuk, Korea) and housed in clear plastic cages in specified pathogen-free conditions under a 12:12-h light-dark cycle (lights on at 0800 and off at 2000). Mice had free access to standard irradiated chow (Purina Mills, Seoul, Korea). Ewha Womans University Animal Care and Use Committee granted the approval for all experimental procedures involving animals.

2.2 Experimental design In experimental 1 (Fig.1A), mice were subjected to the 21 consecutive days of restraint stress

6

with various restraint durations (2-6 h/day). Water maze test was performed 27 days after the exposure to stress. In experiment 2 (Fig. 1B), mice were subjected to the 21 consecutive days of restraint stress. Water maze test was performed 3 or 27 days after the exposure to stress in independent experiment. In experiment 3 (Fig. 3A), mice were divided into three groups (control: CON, restraint stress: RST, exercise combined with restraint stress group: RST+Ex) with each group containing 10 mice. To induce chronic stress by restraint, 8-week-old mice were individually placed into well-ventilated 50-mL conical tubes, which prevented forward or backward movement. Control mice remained undisturbed in their home cages during restraint exposure. Restraint stress was delivered at set times from 1000 to 1600 for 6 h for 21 days. After the stress exposure period, all mice were pre-exercised to acclimate them to treadmill-running (Myung Jin Instruments Co., Seoul, Korea) from 1000 (Pre-Ex; 12 m/min, 20min/day) for 3 days. Subsequently treadmill exercise was performed at 19 m/min for 60 min/day, 6 days/week for 21 days. Non-exercised mice were placed on the treadmill that was turned off for 60 min once a day. Two days after the last treadmill session, Y-maze and water maze tests were performed. In experiment 4 (Fig. 5A), mice were subjected to the 21 consecutive days of restraint stress. Serum CORT and ACTH levels were measured 1 day after the last exposure to restraint. In experiment 6 (Fig. 5B), mice were subjected to a 21-day (19 m/min, 1 h/day, 6 days/week) of treadmill running. Serum CORT and ACTH levels were measured 1 day after the last administration to treadmill running. In experiment 5 (Fig. 7), the Y-maze test and water maze test was performed 21 days after 7 consecutive days of intraperitoneal injection with the AMPK agonist 5-amimoimidazole-4-carboxamide riboside (AICAR, 500 mg/kg, once a day; Tocris Bioscience, Bristol, UK; eight mice per group). In experimental 6 (Fig. 8), mice were intraperitoneally injected with AICAR (0-500 mg/kg) for 7. Water maze test was performed 14 days after the last treatment with AICAR. In experiment 7 (Fig.9A), mice were divided into four groups (control, restraint stress, exercise combined

7

with restraint stress group, exercise combined with restraint stress and Compound C; eight mice per group). The experimental procedure of experiment 9 was equal to that of experiment 3, except on Compound C (10 mg/kg; EMD Chemicals, Gibbstown, NJ) treatment. Compound C was intraperitoneally injected during exercise intervention every 4 days. Mice without drugs (AICAR and Compound C) treatment were injected with saline (1% DMSO).

2.3 Y-maze test The Y-maze consisted of three equal-sized arms made of white PVC. The arms measured 38.5-cm long, 3-cm wide, and 13-cm high, and were oriented at 60° angles from each other (JEUNG DO Bio & Plant Co. LTD, Seoul, Korea). The Y-maze test was performed under moderate lighting conditions (200 Lux) with moderately loud background white noise (40 dB). Mice began a single trial at the end of one arm and were allowed to explore the Y-maze freely for 8 minutes. The number and sequence of arm visits were recorded manually by an observer. Alternation was defined as a consecutive entry in three different arms. The alternation percentage was calculated

with the

following formula: (number of

alternations/total number of arm visits) − 2.

2.4 Water maze test The test was performed using the SMART-CS (Panlab, Barcelona, Spain) program in an airconditioned room. The water maze experiment was carried out in a 1.5m diameter plastic circular pool with 22°C water containing powdered milk to obstruct the platform visibility. Escape latency was monitored by a computer, using the SMART-LD program, which was connected to a ceiling-mounted camera directly above the pool. The training schedule consisted of two trials per day over 4 days of testing, and each trial assessed the ability of the

8

mouse to reach the platform within 60s. On day 5, the mice were subjected to three probe trials, in which they were required to swim for 60s without a platform. The time required to reach the previous platform location (escape latency) was recorded for each animal. Each trial was stored on videotape for subsequent analysis.

2.5 Western blot analysis Hippocampal tissue was homogenized with lysis buffer (50 mM HEPES pH 7.5; 150 mM NaCl; 10% glycerol; 1% TritonX-100; 1 mM PMSF; 1 mM EGTA; 1.5 mM MgCl2· 6H2O; 1 mM sodium orthovanadate; 100 mM sodium fluoride). Protein samples (20 µg) were electrophoretically separated on 10% polyacrylamide gels, transferred to nitrocellulose membranes (Amersham Bioscience, Buckinghamshire, UK), and incubated with a primary antibody in a blocking buffer at room temperature overnight. The next day, the samples were rinsed with a washing buffer and incubated with horseradish peroxidase-conjugated secondary antibody for 2 hours at room temperature. The optical density of each band was measured using the SCION program (NIH Image Engineering, Bethesda, MD, USA). Anti-phospho-AMPK and anti-AMPK antibodies were obtained from Cell Signaling Tech. Inc. (Danvers, MA, USA), anti-BDNF was from Santa Cruz Biotechnology (Santa Cruz, CA, USA), and anti-synaptophysin and anti-PSD-95 were obtained from Abcam (Cambridge, UK).

2.6 Primary hippocampal culture Primary hippocampal cell cultures were prepared from E17 ICR mice. Dissociated single cells were plated in a solution containing RF media, DMEM with 10% FBS, 1× penicillin/streptomycin, 1.4 mM L-glutamine, and 0.6% glucose, in 12-well plates for 1 day.

9

On day in vitro (DIV) 1, the cells were incubated in Neurobasal Medium with 1× B27, 1× penicillin/streptomycin, and 1× L-glutaMax. The medium was changed every 2 days. Cultures taken from DIV 7-9 were used in experiments. In experiment 1, cells were cultured with AICAR (1, 0.5, 1 mM) for 1 hour, and then BDNF, phospho-AMPK, and AMPK levels were measured using western blot analysis. In experiment 2, cells were cultured with Compound C (1, 3, 10 µM) for 2 hours before undergoing treatment with AICAR (0, 0.5 mM) for 1 hour.

2.7 Immunohistochemistry

Anaesthetized mice were perfused with 100 mM phosphate buffer (PBS; pH 7.4), followed by cold 4% paraformaldehyde in PBS. After perfusion, the brains were removed, fixed for another 18 h,, and transferred to 10–30% sucrose solution. Finally, 40-µm-thick sections were prepared using a vibratome (Leica, Wetzlar, Germany). Every eleventh section was taken from the region between bregma −1.46 mm and −2.80 mm. Free-floating sections were incubated with 0.3% hydrogen peroxide (H2O2), permeabilized with 0.3% Triton X-100, and nonspecific protein binding was blocked by incubation with 3% normal goat serum. Sections were incubated overnight at 4°C with anti-Ki-67 and anti-doublecortin (DCX) primary antibodies, respectively (Abcam, Cambridge, MA, USA; rabbit polyclonal, 1: 2,000) and subsequently with biotinylated secondary antibodies (Vector Laboratories; Burlingame, CA, USA 1: 200, respectively), and then visualized using the ABC method (ABC Elite kit, Vector Laboratories; Burlingame, CA, USA). The sections were mounted and assessed in digital images (captured at 100× magnification) using Image J (NIH Image Engineering, Bethesda, MD).

10

2.8 CORT and ACTH measurements Blood samples were centrifuged at 1000 × g for 15 min to obtain serum. CORT and ACTH levels were measured from serum using CORT enzyme immunoassay kits (Cayman Chemical, Ann Arbor, MI, USA), ACTH ELISA kit (Sigma-Aldrich, MO. USA). 2.9 RT-PCR Total RNA was extracted using Trizol reagent kit (Invitrogen, CA, USA). The RNA was reverse transcribed with reverse transcriptase and a random hexamer primer (Promega, CA, USA). cDNA was amplified with the following sense and antisense primers (5′→3′): for BDNF sense 5′-TGG CTG ACA CTT TTG AGC AC-3′ and antisense 5′-GTT TGC GGC ATC CAG GTA AT-3′; for synaptophysin sense, 5’-TAA CCC GAG TAA GAA TGT C-3’ and antisense 5’-CCC TAC ATT CAC CCA CTT CTC C-3’; for PSD-95 sense 5’-TGC ACT CTT GAT GTA TCA GC-3’ and antisense 5’-ACG GAT GAA GAT GGC GAT AG-3’;for GAPDH for sense 5’-TCC ATG ACA ACT TTG GCA TT-3’ and antisense 5’-GTT GCT GTT GAA GTC GCA GG-3’. PCR products were analyzed on 1.5% agarose gel, and the intensity of each band was measured using Image J (NIH Image Engineering, Bethesda, MD).

2.10 Statistical analysis Statistical analysis was performed with SPSS (SPSS for Windows, version 18.0, Chicago, IL, USA), using one-way ANOVA, two-way repeated measured ANOVA, and independent t-tests to assess for significance. Post-hoc comparisons were made using Newman-Keuls tests. All values are reported as mean ± standard error (SE). Statistical significance was set at p < 0.05.

11

3 RESULTS

3.1 Regular and prolonged treadmill running alleviated memory impairments produced by repeated restraint. First, we explored chronic stress-induced memory defect according to various durations of restraint exposure (2-6 hours/day). The exposure to restraint for 6 hours per day resulted in the decreased escape latency on the final day of water maze testing 4 weeks after the last exposure to restraint, but not 2-3 hours per day (Fig.1A; for 2h/21d t14 = -0.87, p > 0.05; for 3h/21d t14 = -1.90, p > 0.05; for 6h/21d t14 = -5.71.90, p > 0.01). Next, the escape latency of stressed mice was enhanced on the final day of water maze testing compared with that of control mice 1 week after the last exposure to restraint and this change lasted up to 4 weeks, suggesting that the 21 consecutive days of restraint stress-induced memory deficit at least lasted until 4 weeks in this experimental paradigm (Fig. 1B; for on day 28 t14 = -11.20, p < 0.01; 53 (t14 = -9.98, p < 0.01 ). The rate of alternation in Y-maze tasks was significantly enhanced 2 and 9 days after the last exposure to treadmill running relative to that of nontreadmill-run mice, but not 7-day treadmill running (Fig. 2A; for 7-day Ex, t12 = 0.13, p > 0.05; for 21-day Ex, post-day 2, t12 = -3.74, p < 0.01; for 21-day Ex, post-day 9, t12 = -2.82, p < 0.05). Furthermore, pre-exercise for the acclimation to treadmill running did not enhance memory performance (Fig. 2A; t10 = 0.01, p > 0.05). Memory function was evaluated using water maze and Y-maze tests to elucidate whether repeated and regular exercise might alleviate memory impairments induced by 21 consecutive days of restraint (Fig.3A). In the Ymaze test, the alternation of restrained mice was markedly reduced compared with that of control mice. This was reversed by treadmill exercise (Fig.3Ba; F2, 27 = 5.637, p < 0.01). In the water maze test, the escape latency of restrained mice was significantly enhanced compared with that of control mice, which was reversed by exercise treadmill running (the

12

interaction effect of group x day F8, 108 = 18.13, p < 0.01; the main effect of group F1, 27 = 18218.74, p < 0.01; the main effect of day F4, 108 = 235.49, p < 0.01; Fig. 3Bb-c).

3.2 The expression of BDNF, phospho-AMPK, and synaptic proteins such as synaptophysin and PSD-95 were reduced by chronic restraint, and this change was counter-regulated by regular and prolonged treadmill running. Chronic stress in the form of repeated restraint was found to down-regulate hippocampal mature BDNF and phospho-AMPK expression. This effect was reversed by regular and prolonged treadmill running (Fig. 4A; for BDNF F2, 21 = 70.90, p < 0.01, for phospho-AMPK F2, 21 = 85.30, p < 0.01). Additionally, the expression of two synaptic proteins, synaptophysin (a presynaptic marker) and PSD-95 (a postsynaptic marker), was reduced by chronic stress, which was similarly counter-regulated by regular and prolonged exercise (Fig. 4B; for synaptophysin F2, 21 = 184.06, p < 0.01; for PSD-95 F2, 21 = 49.59, p < 0.01).

3.3 Chronic stress enhanced serum corticosterone (CORT) and adrenocorticotrophic hormone (ACTH) Serum corticosterone (CORT) and adrenocorticotrophic hormone (ACTH) levels was in chronic stressed mice significantly higher than in control mice at the rest, when CORT and ACTH were measured 2 days after the 21 consecutive day of restraint stress (Fig. 5A; for CORT, t8 = -4.74, p < 0.01; for ACTH, t8 = -3.70, p < 0.01). Regular exercise didn’t alter basal levels of both CORT and ACTH (Fig. 5B; for CORT, t8 = 0.17, p > 0.05; for ACTH, t8 = -0.24, p > 0.05).

3.4 AICAR treatment enhanced BDNF expression, and Compound C treatment reduced BDNF expression in the primary hippocampal culture.

13

Since the expression pattern of phospho-AMPK corresponded well to that of mature BDNF, we assessed mature BDNF expression level in the primary hippocampal culture using the AMPK agonist, AICAR and AMPK antagonist, Compound C. BDNF expression was enhanced in a dose-dependent manner and phospho-AMPK expression was simultaneously increased by AICAR treatment (Fig. 6A; for BDNF F2, 9 = 126.7, p < 0.01; for phosphoAMPK F2,

9

= 165.42, p < 0.01). Such effects were suppressed by prior exposure to

Compound C (Fig. 6B; for BDNF F4, 15 = 88.11, p < 0.01; for phospho-AMPK F4, 15 = 106.84, p < 0.01).

3.5 AICAR treatment induced hippocampal BDNF and phospho-AMPK expression, enhanced alternation in the Y-maze test, and reduced escape latency. Since hippocampal AMPK activity mediated the induction of mature BDNF expression in vitro, the ensuing experiments explored whether AICAR would exert a similar effect on memory function in vivo (Fig. 7A). Seven consecutive days of AICAR treatment enhanced the alternation rate in the Y-maze test up to 2 weeks after the last treatment (Fig. 7Ac; t14 = 3.68, p < 0.05). Similarly, AICAR treatment up regulated the expression of hippocampal mature BDNF and phospho-AMPK (Fig. 7Ab; t6 = -11.71, p < 0.01). Also, in water maze test, the escape latency decreased by ALCAR treatment on day 4-5 (Fig. 7B; the interaction effect of group x day F4, 72 = 1.74, p > 0.05; the main effect of group F1, 18 = 18092.12, p < 0.01; the main effect of day F4, 72 = 158.97, p < 0.01). The escape latency was reduced on day 4-5 in mice treated with ALCAR at 500 mg/kg, but not 100-300 mg/kg (Fig. 8; the interaction effect of group x day F12, 144 = 0.93, p > 0.05; the main effect of group F1, 36 = 40774.08, p < 0.01; the main effect of day F4, 144 = 333.89, p < 0.01).

3.6 Compound C treatment during exercise intervention blocked exercise-induced

14

memory improvement in mice subjected to restraint. We investigated whether AMPK inactivation by treatment with Compound C blocked exercise-induced memory improvement in stressed mice. Whilst exercise reduced the escape latency of the water maze task in stressed mice, this effect was blocked by Compound C treatment (Fig. 9; the interaction effect of group x day F12,144 = 9.48, p < 0.01; the main effect of group F1, 36 = 35766.26, p < 0.01; the main effect of day F4, 144 = 245.29, p < 0.01).

3.7 Compound C treatment during exercise intervention suppressed exercise-produced hippocampal neurogenesis enhancement in mice subjected to restraint. We investigated whether AMPK inactivation by treatment with Compound C suppressed exercise-produced enhancement of Ki-67-and doublecortin-positive cells in chronically stressed mice. Compound C treatment reduced exercise-induced increase in hippocampal Ki67- and doublecortin-positive cells in chronically restrained mice (Fig. 10; for Ki-67 F3, 28 = 16.86, p < 0.01; for doublecortin F3, 28 = 16.86, p < 0.01).

3.8 Compound C treatment during exercise intervention blocked exercise-induced memory improvement in mice subjected to restraint. We investigated whether AMPK inactivation by treatment with Compound C suppressed exercise-produced enhancement of BDNF, synaptophysin, and PSD-95 mRNA in chronically stressed hippocampus (Fig. 11). Compound C treatment reduced exercise-induced increase in hippocampal BDNF, synaptophysin, and PSD-95 mRNA levels in chronically restrained mice (Fig. 7; for BDNF F3, 16 = 62.97, p < 0.01; for synaptophysin F3, 16 = 63.21, p < 0.01; for PSD95 F3, 16 = 57.61, p < 0.01).

4 DISCUSSION

15

The current study demonstrates that chronic restraint stress induces hippocampal-dependent memory deficits, which can be counteracted with regular and prolonged treadmill exercise. Furthermore, the protective effect of exercise against chronic stress-induced memory impairment is at least in part dependent on hippocampal AMPK-mediated BDNF induction. To the best of our knowledge, this is the first study investigating the role of hippocampal AMPK in this capacity, at least with regard to its up regulation by prolonged exercise, and subsequent induction of BDNF expression. To investigate whether chronic stress caused memory defect, mice were subjected to restraint stress for 21 days with various durations (2-6 hours/day) and subsequently memory function was measured by water maze test (Fig. 1A). In the first instance, the exposure to restraint for 6 hours per day resulted in the decreased escape latency on the final day of water maze testing 4 weeks after the last exposure to restraint, but not 2-3 hours per day. The escape latency of stressed mice was enhanced on the final day of water maze testing compared with that of control mice 1 week after the last exposure to restraint and this change lasted up to 4 weeks, suggesting that the 21 consecutive days of restraint stress-induced memory deficit at least lasted until 4 weeks in this experimental paradigm (Fig. 1B). Furthermore, although species and testing method were different from those of this study, the 21 consecutive days of restraint stress (6h/day) reduced latency of entrance to the dark chamber in passive avoidance test until 21 day after the last stress exposure (Radahmadi et al., 2015; Radahmadi et al., 2013). For this reason, the chronic stress regime with a 6h/21d of restraint was adopted in the current study. Next, we investigated whether a 21-day regime of treadmill running would affect memory function. Analysis of memory function suggested that the rate of alternation in Y-maze tasks was significantly enhanced 2 and 9 days after the last exposure to treadmill running relative to that of non-treadmill-run mice, but not 7-day treadmill running (Fig. 2A).

16

Furthermore, pre-exercise for the acclimation to treadmill running did not enhance memory performance (Fig. 2B). Based on this data, the exercise paradigms that mice were subjected to during treadmill running after pre-exercise were adopted in the current study. We found that treadmill running for three weeks alleviated memory deficits induced by the 21 consecutive days of restraint stress. Chronic restraint stress elicited a decreased rate of alternation in the Y-maze test, declined retention of memory of water maze testing. These effects were reversed by exercise intervention over three weeks (Fig. 3), suggesting that this experimental paradigm was appropriate for exploring the potential role of exercise in alleviating chronic stress-induced memory deficits. This result was consistent with our previous findings that prolonged exercise elicited memory improvement in chronically restrained mice, although the exercise regimen used was different between both of studies (Kwon et al., 2013). Chronic physical or psychological stress can lead to abnormal morphology and function of hippocampal neurons, which may be attributed to altered synthesis of proteins involved in synaptic plasticity (Kasai et al., 2010; Surget et al., 2011; Zhu et al., 2014). In the current study, synaptophysin and PSD-95 levels were profoundly reduced by chronic stress and this decrease was reversed by regular and prolonged exercise (Fig. 4B). Synaptophysin and PSD95 are mainly localized in pre- and post-synaptic regions, playing a crucial role in regulating synaptic transmission and organizing post-synaptic densities. Reduced object novelty recognition and impaired spatial learning performance are evident in synaptophysin knockout mice (Schmitt et al., 2009). PSD-95 contributes to synaptic formation and stabilization, and participates in the synaptic trafficking of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors that modulate excitatory synaptic transmission (El-Husseini et al., 2000; Henley JM and Wilkinson KA, 2013). The aforementioned results support our theory that chronic stress causes reduced hippocampal synaptophysin and PSD-95 levels. Since

17

these proteins regulate synaptic transmission, strength, and plasticity, a reduction in levels of synaptophysin and PSD-95 could likely impair hippocampal dependent learning and memory function. The results of this study suggest that stress-induced memory impairments can be reversed by exercise, which implies that regular, prolonged physical activity can restore the reduced expression of both synaptic proteins, resulting in the improved efficacy of neuronal information transfer and affecting cognitive behavior. Neurons consume a large amount of energy to perform their physiological functions. Abnormalities

of

hippocampal energy metabolism

are

strongly associated

with

neurocognitive deficits that result from β-amyloid accumulation and chronic stress (Park et al., 2013; Cao et al., 2014). AMPK modulates cellular energy homeostasis as a transcriptional regulator in response to ATP depletion, metabolic stresses, intracellular Ca2+ levels, and exercise (Culmsee et al., 2001; McCullough et al., 2005; Woods et al., 2005; Hardie, 2007; Kobilo et al., 2015a, 2015b). Our data showed that hippocampal phospho-AMPK levels were markedly reduced by chronic stress and this decrease was subsequently reversed by exercise intervention (Fig. 4A). The role of AMPK in cognitive functions has been previously described by a number of researchers (Dagon et al., 2005; Cao et al., 2014; Zhu et al., 2014). The above-addressed findings supported our result that chronic stress inactivates hippocampal AMPK and that regular and prolonged exercise counter-regulates chronic stressinduced deficits. This chronic stress-provoked hippocampal reduction of AMPK activity may be associated with HPA axis abnormality. Serum corticosterone (CORT) and adrenocorticotrophic hormone (ACTH) levels was in chronic stressed mice significantly higher than in control mice at the rest, when CORT and ACTH were measured 2 days after the 21 consecutive day of restraint stress (Fig. 5A). This result suggests that chronic stress induces the sustained hypothalamicpituitary-adrenal (HPX) axis activation, which the persisted increase in CORT may affect

18

hippocampal AMPK activity. On the other hand, regular exercise didn’t alter basal levels of both CORT and ACTH, suggesting that regular exercise didn’t affect basal HPX axis activity (Fig. 5B). In our study, AMPK phosphorylation behavior corresponds well to hippocampal mature BDNF expression. BDNF is a neurotrophic molecule contributing to neuronal growth, development, plasticity, survival, neuroprotection, and repair, which when stimulated, may act to improve cognitive ability (Conner et al., 1997; Duman and Monteggia, 2006). To confirm AMPK-mediated BDNF induction, we analyzed mature BDNF protein levels in primary hippocampal cell cultures treated with AMPK agonist AICAR and antagonist Compound C. The mature form of BDNF and phospho-AMPK protein levels were enhanced in dose-dependent manner by AICAR treatment (Fig. 6A), which was reversed by Compound C treatment (Fig. 6B). Moreover, 7 days of treatment with AICAR enhanced hippocampal mature BDNF and phospho-AMPK expression, concomitant with the increase in alternation in the Y-maze test and the decrease in escape latency in water maze test (Fig. 7). The escape latency was reduced on day 4-5 in mice treated with ALCAR at 500 mg/kg, but not 100-300 mg/kg (Fig. 8). Although AICAR has low permeability across the blood-brain barrier (BBB), our study demonstrated an increase in hippocampal AMPK activation. There are two possible reasons for this change. First, in spite of low permeability across the BBB, a small quantity of permeable AICAR might linger in the extracellular space due to a limited capacity for reuptake and degradation. Second, enhanced memory function following intraperitoneal injection of AICAR both in this study and another (Kobilo et al., 2015b) may be attributed to release of myokines in the muscles. These secretory molecules may pass across the BBB and indirectly affect cognitive function (Sakuma and Yamaguchi, 2011; Pedersen and Febbraio, 2012). Judging from our results and other findings, systematically circulating AICAR likely influences hippocampal AMPK activity, whether directly or indirectly. Several studies have demonstrated that AMPK activation induces BDNF expression (Yoon et al., 2008; Autry et al.,

19

2011; Xu et al., 2013). AMPK activation suppresses 3-hydroxy-3-methylgutaryl coenzyme A reductase (HMG-CoA reductase) and acetyl-CoA carboxylase, thereby restoring ATP levels by reducing fatty acid and cholesterol synthesis (Hardie, 2003). More recently, a study demonstrated that statin, a selective inhibitor for HMG-CoA reductase, facilitates CREBmediated BDNF induction via its binding to PPARα, and therefore causes an improvement in memory function (Roy et al., 2015). These articles support the results of this study, suggesting an important role for hippocampal AMPK-mediated BDNF induction in the improvement of memory function. This suggests that regular and prolonged exercise induces hippocampal AMPK activation and exerts a neurotrophic effect on surrounding tissue, thereby promoting memory function in chronically stressed mice. To further explore the role of hippocampal AMPK with regard to exercise-induced changes in memory function, mice were intraperitoneally administrated with Compound C, an AMPK antagonist during treadmill running in chronically stressed mice (Fig. 9A). Enhanced memory in exercised mice relative to control mice was reversed by Compound C treatment (Fig. 9B). Furthermore, to investigate the role of AMPK in hippocampal neurogenesis as well as BDNF, synaptophysin, and PSD-95 mRNA levels in chronically stressed mice, we measured hippocampal neurogenesis (Ki-67: a proliferating marker, doublecortin: a differentiating marker) and the mRNA levels of BDNF, synaptophysin, and PSD-95 in hippocampus (Fig. 10-11). Chronic stress-produced decrease in Ki-67- and doublecortin-positive cells, BDNF, synaptophysin, and PSD-95 mRNA levels were reversed by exercise intervention. However, exercise-exerted these changes were suppressed by Compound C treatment. This result suggests that AMPK activity may contribute to hippocampal neurogenesis as well as BDNF, synaptophysin, PSD95 expression in hippocampus of chronically restrained mice, which exercise-modulated hippocampal AMPK activity may play crucial role in hippocampal neurogenesis and the expression of BDNF and synaptic proteins. Collectively, pharmacological manipulation of

20

AMPK activity suggests that AMPK is heavily involved in the memory side effects of exercise such as enhancement of memory function. Additionally, hippocampal AMPK activation is at least partly required for the induction of exercise-induced neurotrophic effects such as BDNF expression.

CONFLICT OF INTEREST

The authors declare no financial and non-financial competing interests.

ACKNOWLEDGMENTS

This study was supported by Wonkwang University in 2014 and grants from the National Research

Foundation

of

Korea

funded

by

the

Korean

Government

(NRF-

2013R1A1A2062984)..

REFERENCES

Autry AE, Adachi M, Nosyreva E, Na ES, Los MF, Cheng PF, Kavalali ET, Monteggia LM (2011) NMDA receptor blockade at rest trigger rapid behavioral antidepressant responses. Nature 475:91-5. Calabrese F, Guidotti G, Molteni R, Racagni G, Mancini M, Riva MA (2012) Stress-induced changes of hippocampal NMDA receptors: modulation by duloxetine treatment. PLoS One 7:e37916-37924. Cao K, Zheng A, Xu J, Li H, Liu J, Peng Y, Long J, Zou X, Li Y, Chen C, Liu J, Feng Z

21

(2014) AMPK activation prevents prenatal stress-induced cognitive impairment: Modulation of mitochondrial content and oxidative stress. Free Rad Biol Med 75:156-166. Conner JM, Lauterborn JC, Yan Q, Gall CM, Varon S (1997) Distribution of brain-derived neurotrophic factor (BDNF) protein and mRNA in the normal adult rat CNS: evidence for anterograde axonal transport. J Neurosci 17: 2295−2313. Culmsee C, Monnig J, Kemp BE, Mattson MP (2001) AMP-activated protein kinase is highly expressed in neurons in the developing rat brain and promotes neuronal survival following glucose deprivation. J Mol Neurosci 17:45-58. Dagon Y, Avraham Y, Magen I, Gertler A, Ben-Hur T, Berry EM (2005) Nutritional status, cognition, and survival: A new role for leptin and AMP kinase. J Biol Chem 280(51):4214242148. Duman RS, Monteggia LM (2006) A neurotrophic model for stress-related mood disorders. Biol Psychiatry 59:1116−1127. Eadie BD, Redila VA, Christie BR (2005) Voluntary exercise alters the cytoarchitecture of the adult dentate gyrus by increasing cellular proliferation, dendritic complexity, and spine density. J Comp Neurol 486:39-47. El-Husseini AE, Schnell E, Chetkovich DM, Nicoll RA, Bredt DS (2000) PSD-95 involvement in maturation of exciatory synapses. Science 290:134-1368. Gomez-Pinilla F, Vaynmas S, Ying Z (2008) Brain-derived neurotrophic factor functions as a metabotrophin to mediate the effects of exercise on cognition. Eur J Neurosci 28:2278-2287. Hadad SM, Fleming S, Thompson AM (2008) Targeting AMPK: a new therapheutic opportunity in breast cancer. Crit Rev Oncol Hematol 67:1-7.

22

Han TK, Lee JK, Leem YH (2015) Chronic exercise prevents repeated restraint stressprovoked enhancement of immobility in forced swimming test in ovariectomized mice. Metab Brain Dis 30(3):711-8. Hardie DG (2004) The AMP-activated protein kinase pathway-new players upstream and downstream. J Cell Sci 117:5479-5487. Hardie DG (2007) AMP-activated/SNF1 protein kinases: conserved guardiance of cellular energy. Nat Rev Mol Cell Biol 8:774-785. Henley JM, Wilkinson KA (2013) AMPA receptor trafficking and the mechanism underlying synaptic plasticity and cognitive aging. Dialogues Clin Neurosci 15:11-27. Hillman CH, Erickson KI, Kramer AF (2008) Be smart, exercise your heart: Exercise effects on brain and cognition. Nat Rev Neurosci 9:58-65. Kasai H, Fukuda M, Watanabe S, Hayashi-Takagi A, Noguchi J (2010) Structural dynamics of dendritic spines in memory and cognition. Trends Neurosci 33:121-129. Kobilo T, Guerrieri D, Zhang Y, Collica SC, Becker KG, van Praag H (2015a) AMPK agoinst AICAR improves cognition and motor coordination in young and aged mice. Learning and Memory 21:119-126. Kobilo T, Yuan C, van Praag H (2015b) Endurance factors improve hippocampal neurogenesis and spatial memory in mice. Learning and Memory 18:103-107. Krishnan V, Nestler EJ (2008) The molecular neurobiology of depression. Nature 455:894902. Kwon D, Kim B, Chang H, Kim Y, Ahn Jo S, Leem YH (2013) Exercise overcame impaired cognition by restraint stress-induced oxidative insult and BDNF abnormality. BBRC

23

434:245-251. Magarinos AM, McEwen BS (1995) Stress-induced atrophy of apical dendrites of hippocampal CA3c neurons: involvement of glucocorticoid secretion and excitatory amino acid receptors. Neuroscience 69:89-98. McGaugh JL, Roozendaal B (2002) Role of adrenal stress hormones in forming lasting memories in the brain. Curr Opin Neurobiol 12:205-210. McLaughlin KJ, Gomez JL, Baran SE, Conrad CD (2007) The effects of chronic stress on hippocampal morphology and function: an evaluation of chronic restraint paradigms. Brain Res 1161:56-64. Park S, Kim S, Kang S, Moon NR (2013) β-Amyloid-induced cognitive dysregulation impairs glucose homeostasis by increasing insulin resistance and decreasing β-cell mass in non-diabetic and diabetic rats. Metabolism 62:1749-1760. Pedersen BK, Febbraio MA (2012) Muscle, exercise and obesity: skeletal muscle as a secretory organ. Nat Rev Endocrinol 8:4574-4565. Pham K, Nacher J, Hof PR, McEwen BS (2003) Repeated restraint stress suppresses neurogenesis and induces biphasic PSA-NCAM expression in the adult rat dentate gyrus. Eur J Neurosci. 17:879-86. Radahmadi M, Alaei H, Sharifi MR, Hosseini N (2015) Preventive and therapeutic effect of treadmill running on chronic stress-induced memory deficit in rats. J Bodyw Mov Ther. 19:238-45. Radahmadi M, Alaei H, Sharifi MR, Hosseini N (2013) The effect of synchronized forced running with chronic stress on shoet, mid and long-term memory in rats. Asian J Sports Med. 4:54-62.

24

Ronnettt GV, Ramamurthy S, Kleman AM, Landree LE, Aja S (2009) AMPK in the brain: its roles in energy balance and neuroprotection. J Neurochem 109 Suppl. 1:17-23. Roy A, Jana M, Kundu M, Corbett GT, Rangaswamy SB, Mishra RK, Luan C, Gonzalez FJ, Pahan K (2015) HMG-CoA reductase inhibitors bind to PPARα to upregulate neurotrophin expression in the brain and improve memory in mice. Cell Metabolism 22:253-265. Sakuma K, Yamguchi A (2011) The recent understanding of the neurotrophin;s role in skeletal muscle adaptation. J Biomed Biotechnol 2011:201696, doi: 1155/2011/201396. Sandi C, Pinelo-Nava MT (2007) Stress and memory: behavioral effects and neurobiological mechanisms. Neural Plast. 2007:1-20. Schmitt U, Tanimoto N, Seeliger M, Schaeffel F, Leube RE (2009) Detection of behavioral alterations and learning deficits in mice lacking synaptophysin. Neuroscience 162:234-243. Sousa N, Lukoyanov NV, Madeira MD, Almeida OF, Paula-Barbosa MM (2000) Reorganization of the morphology of hippocampal neurites and synapse after stress-induced damage correlates with behavioral improvement. Neuroscience 97:253-66. Spasic MR, Callaerts P, Norga KK (2009) AMP-activated protein kinase (AMPK) molecular crossroad for metabolic control and survival of neurons. Neuroscientist 15:309-316. Steinberg GR, Kemp BE. 2009. AMPK in health and disease. Physiol Rev 89:1025-78. van Praag H (2008) Neurogenesis and exercise: Past and future directions. Neuromolecular Med 10:128-140. Watanabe Y, Gould E, McEwen BS (1992) Stress induces atrophy of apical densrites of hippocampal CA3 pyramidal neurons. Brain Res 588:341-345. Woods A,. Dickerson K, Heath R, Hong SP, Momcilovic M, Johnstone SR, Carlson M,

25

Carling D (2005) Ca2+/calmodulin-dependent protein kinase-beta acts upstream of AMPactivated protein kinase in mammalian cells. Cell Metab 2:21-33. Xu S, Zhou Z, Li X, Ji M, Zhang G, Yang J (2013) The activation of adenosine monophosphate-activated protein kinase in rat hippocampus contributes to the rapid antidepressant effect of ketamine. Behav Brain Res 253:305-309. Yamada K, Nabeshima T (2003) Brain-derived neurotrophic factor/TrkB signaling in memory processes. J Pharmacol Sci 91:267-270. Yoon H, Oh YT, Lee JY, Baik HH, Kim SS, Choe W, Yoon KS, Ha J, Kang I (2008) Activation of AMP-activated protein kinase by kainic acid mediates brain-derived neurotrophic factor expression through a NF-kappaB dependent mechanism in C6 glioma cells. BBRC 371:495-500. Zhu S, Wang J, Zhang Y, Li V, Kong J, He J, Li X (2014) Unpredictable chronic mild stress induces anxiety and depression-like behaviors and inactivates AMP-activated protein kinase in mice. Brain Res 1576:81-90. Figure Legends

Figure 1. A 21-day of restraint stress (6h/day) caused memory decline and this effect continued up to 4 weeks, but not 2-3 h/day. A. Quantitative analysis of the escape latency on the final day of water maze testing. The escape latency of stressed mice was enhanced compared with that of control mice on day 28 (for 2h/21d t14 = -0.87, p > 0.05; for 3h/21d t14 = -1.90, p > 0.05; for 6h/21d t14 = -5.71.90, p > 0.01). B. Quantitative analysis of the escape latency on the final day of water maze testing. The

26

escape latency of stressed mice was enhanced compared with that of control mice on day 28 (t14 = -11.20, p < 0.01) and 53 (t14 = -9.98, p < 0.01). Data are presented as the mean ± SEM. ** denote differences at p < 0.01.

Figure 2. Pre-exercise for the acclimation to treadmill running did not affect memory ability measured by Y-maze test. A. 21-days of exercise intervention improved memory function, but not 7-days of exercise. a. Experimental design. Mice were subjected to 7-day (19 m/min, 1 h/day) or 21-day (19 m/min, 1 h/day, 6 day/week) treadmill running 2 days after pre-exercise. Y-maze test was assessed 2 or 9 days after the last exercise intervention, independently. b. Quantitative analysis of Y-maze test. A 7-day exercise intervention did not change rates of alternation in the Y-maze test. However, the 21-day exercise regimen enhanced alternation of Y-maze test 2 and 9 days after the last exercise intervention (for 7-day Ex, t12 = 0.13, p > 0.05; for 21-day Ex, post-day 2, t12 = -3.74, p < 0.01; for 21-day Ex, post-day 9, t12 = -2.82, p < 0.05) B. Experimental design and alternation of Y-maze test between sedentary and pre-exercised mice. a. Experimental design. Mice were subjected to treadmill running for 3 days (12 m/min, 20 min/day), and subsequently Y-maze test was assessed 21 days after the last exercise administration. b. Quantitative analysis of Y-maze test. Alternative rate of Y-maze test in 3-day of treadmillrun mice was comparable to that of sedentary mice (t10 = 0.01, p > 0.05).

27

Data are presented as the mean ± SEM. Data are presented as the mean ± SEM. * and ** denote differences at p < 0.05 and p < 0.01, respectively.

Figure 3. Regular and prolonged treadmill running ameliorated stress-induced learning and memory impairment. A. Experimental design B. Quantitative analysis of Y-maze and water maze test data a. Quantitative analysis of Y-maze test data. b. Quantitative analysis of the escape latency of water maze test. c. Photomicrograph showing swimming patterns on the final day of water maze testing. Data are presented as the mean ± SEM. ** denote differences at p < 0.01, and †† denote difference at p < 0.01 from CON.

Figure 4. Regular and prolonged treadmill running reversed chronic stress-elicited reduction of hippocampal phospho-AMPK, BDNF, synaptophysin and PSD-95 levels. A. Quantitative analysis of hippocampal phospho-AMPK and BDNF. a. Photomicrograph showing western blot data for AMPK and BDNF. b. Quantitative analysis of phospho-AMPK and BDNF. B. Quantitative analysis of hippocampal phospho-AMPK and BDNF. a. Photomicrograph showing western blot data for synaptophysin and PSD-95. b. Quantitative analysis for synaptophysin and PSD-95. Data are presented as the mean ± SEM. ** denote differences at p < 0.01.

28

Figure 5. Chronic stress caused the sustained serum CORT and ACTH levels, but not the regular and prolonged exercise. A. Quantitative analysis of serum CORT and ACTH levels following chronic stress. a. Experimental design. Mice were subjected to restraint for 21 days (6h/day), followed by the assess of serum CORT and ACTH levels. b. Quantitative analysis of serum CORT and ACTH levels. Serum CORT and ACTH levels were significantly enhanced by chronic restraint stress. B. Quantitative analysis of serum CORT and ACTH levels following chronic stress. a. Experimental design. Mice were subjected to treadmill running for 21 days (19 m/min, 1 h/day, 6 day/week), followed by the assess of serum CORT and ACTH levels. b. Quantitative analysis of serum CORT and ACTH levels. Serum CORT and ACTH levels were no significantly different between groups.

Figure 6. AICAR treatment upregulated phospho-AMPK and BDNF, which was reversed by Compound C in hippocampal primary culture. A. Quantitative analysis of hippocampal phospho-AMPK and BDNF in hippocampal primary culture treated with AICAR. a. Photomicrograph showing western blot data for AMPK and BDNF. b. Quantitative analysis of phospho-AMPK and BDNF. B. Quantitative analysis of hippocampal phospho-AMPK and BDNF in hippocampal primary culture treated with AICAR or/and Compound C. a. Photomicrograph showing western blot data for AMPK and BDNF. b. Quantitative analysis of phospho-AMPK and BDNF.

29

Data are presented as the mean ± SEM. * and ** denote differences at p < 0.05 and p < 0.01, respectively.

Figure 7. Intraperitoneal injection of AICAR enhanced alternation in the Y-maze test, concomitant with hippocampal phospho-AMPK and BDNF levels. A. Quantitative analysis of hippocampal phospho-AMPK and BDNF levels, and Y-maze test data a. Experimental design b. Photomicrograph showing western blot data and quantitative analysis for hippocampal phospho-AMPK and BDNF in mice treated with AICAR. c. Quantitative analysis of Y-maze test data. B. Quantitative analysis of water maze data a. Experimental design b. Quantitative analysis of water maze test Data are presented as the mean ± SEM. * and ** denote differences at p < 0.05 and p < 0.01, respectively. † denote difference from 0 mg/kg at p < 0.01 from 0 mg/kg.

Figure 8. AICAR treatment with 500 mg/kg improved memory function, but not 0-200 mg/kg. A. Experimental design. Mice were treated with AICAR (0-500 mg/kg) for 7 days, and subsequently memory function were measured by water maze test 14 days after the last treatment with AICAR. B. Quantitative analysis of the escape latency of water maze testing. The escape latency of mice treated with 500 mg/kg AICAR decreased on day 4-5, but not 100-200 mg/kg.

30

Data are presented as the mean ± SEM. ** denote differences from 0 mg/ kg at p < 0.01.

Figure 9. Compound C treatment during exercise intervention blocked exercise-induced memory improvement in mice subjected to restraint. A. Experimental design B. Quantitative analysis of water maze test data a. Quantitative analysis of escape latency of water maze test. b. Photomicrograph showing swimming patterns on the final day of water maze testing.

Data are presented as the mean ± SEM. ** denote differences at p < 0.01. †† and ǂǂ denote difference from CON and RST+Ex, respectively at p < 0.01 from

Figure 10. Compound C treatment during exercise intervention blocked exerciseinduced enhancement of neurogenesis in mice subjected to restraint. A. Quantitative analysis of Ki-67-positive cells. a. Photomicrograph showing immunohistochemical data for Ki-67. b. Quantitative analysis of Ki-67-positive cells. A. Quantitative analysis of doublecortin-positive cells. a. Photomicrograph showing immunohistochemical data for doublecortin. b. Quantitative analysis of doublecortin-positive cells. Data are presented as the mean ± SEM. ** denote differences at p < 0.01.

Figure 11. Compound C treatment during exercise intervention blocked exerciseinduced enhancement of BDNF, synaptophysin, and PSD-95 mRNA in mice subjected to

31

restraint. a. Photomicrograph showing RT-PCR data for BDNF, synaptophysin, and PSD-95 mRNA. b. Quantitative analysis of BDNF, synaptophysin, and PSD-95 mRNA. Data are presented as the mean ± SEM. ** denote differences at p < 0.01.

32

RESEARCH HIGHLIGHTS

Chronic stress causes hippocampal-dependent memory deficit. Exercise improves memory in chronically stressed mice. Chronic stress reduces hippocampal AMPK activity, BDNF, and neurogenesis. Exercise enhances hippocampal AMPK activity in chronically stressed mice. AMPK is required for exercise-exerted memory and neurogenesis enhancement.