Male C57BL6 mice at 7 weeks of age were purchased from Daehan BioLink (Eumsung, Chungbuk, Republic of Korea). After their arrival, the mice were randomly divided into experimental groups as indicated and were housed in pairs in standard plastic cages in a temperature (23~24℃)- and humidity (50~60%)-controlled animal room. All animals were handled in accordance with the animal care guidelines of Ewha Womans University. All the proposed experimental procedures were approved by the Ewha Womans University IACUC (IACUC 16-018).

Scheduled repeated restraints were performed as described previously [31,32]. In brief, mice were individually placed in a 50-ml conical tube with many holes for good ventilation and were restrained within this tube for 2 h daily beginning at 10 a.m. After each session of restraint, the mice were returned to their home cages and allowed to have free access to food and water. This procedure was repeated each day for the indicated number of days.

Imipramine was purchased from Sigma-Aldrich (St. Louis, MO, USA), and directly diluted in 0.9% saline. SNAP94847, an MCH receptor 1 antagonist, was purchased from Tocris Bioscience (#3347; Bristol, U.K), and was first dissolved in dimethylsulfoxide (DMSO), and then diluted in 0.9% saline. Imipramine (20 mg/kg/d) and SNAP94847 (30 mg/kg/d) were intraperitoneally (i.p) injected at a volume of 120 µl per injection for the indicated periods.

Passive wheel-running exercise was performed as described previously [31,32]. In brief, the running wheel consisted of a rotating drum with two circular ventilated plastic walls (20 cm, in inner diameter) with a rim (7 cm, in width) composed of evenly spaced aluminum bars (each 2 mm in width, spaced by 0.65 cm). Mice were placed on the running wheel for 20 min daily for 5 days so that they become familiar with the equipment and wheel running. This procedure was applied to animals prior to the start of the restraint session and was regarded as habituation to the running wheel. Each mouse was placed on a running wheel rotating at 9 m/min for 60 min daily, starting at 10:00 am, and this treatment was repeated for 21 days. All exercise performances for the 5 days of prehabituation and 21 days of scheduled exercise were videotaped. Any mice that run on the running wheel for less than 80% of the training time for more than two consecutive days was regarded as having failed to perform the exercise properly and therefore was excluded from the final analysis.

Immunohistochemical analyses were carried out as previously described [31,33]. Briefly, mice were anesthetized with a mixture (3.5:1) of ketamine hydrochloride (50 mg/ml) and xylazine hydrochloride (23.3 mg/ml) at a dose of 2.5 µl/g body weight. They were then perfused with 4% paraformaldehyde by a trans-cardiac method, and brains were isolated. Brains were post-fixed further in the same solution overnight at 4℃. Brains were coronally cut into 40-µm-thick sections with a vibratome (Leica VT 1000S; Leica Instruments, Nussloch, Germany). These sections in a floating state were incubated with 4% bovine serum albumin in PBS with Tween-20 (PBST) for 1 h, then reacted with the primary antibody at 4℃ overnight. The sections were washed with PBST, reacted with the secondary antibody, and visualized with an ABC Elite kit (PK-6200, Vector Laboratories, Burlingame, CA, USA) for DAB staining.

Anti-MCH was purchased from Phoenix Pharmaceuticals (#H-070-47; 1:500; Germany), while the secondary antibody, biotinylated anti-rabbit IgG, was purchased from Vector Laboratories (#BA-1000; 1:200, Burlingame, CA, USA). DAB-stained images were analyzed with an Olympus BX 51 microscope equipped with a DP71 camera and MetaMorph Microscopy Automation & Image Analysis software (Molecular Devices, Sunnyvale, CA, USA).

Real-time PCR for the quantification of mRNA levels was carried out as described previously [31,32]. Total RNA from tissue samples or cell cultures was purified with TRI reagent (15596-018, Invitrogen, CA, USA). Isolated RNA from each group was treated with DNase I to avoid genomic contamination. A reverse transcription system (Promega, MO, USA) was used to convert 1 µg of total RNA to cDNA. Real-time PCR was performed with 10 µl of 2X iQ™ SYBR Green Supermix (Bio-Rad Laboratories, Foster City, CA, USA), 1 µl each of 5 pmol/µl primers, and 4 µl of the cDNA (1/8 dilution of the cDNA) in a volume of 20 µl using the CFX 96 Real-Time PCR System Detector (Bio-Rad Laboratories; Foster City, CA, USA). The primer sequences used were 5'-TGAACGATGATGACAATAAGAA-3' and 5'-TCAGAGCGAGGTAAGGTT-3' for MCH; 5'-AGAAGGTGGTGAAGCAGGCATC-3' and 5'-CGAAGGTGGAAGAGTGGGAGTTG-3' for GAPDH; and 5'-GCTGCCATCTGTTTTACGG-3' and 5'-TGACTGGTGCCTGATGAACT-3' for L32. GAPDH and L32 were used as controls.

HT22 mouse hippocampal cells were cultured as described previously [34]. Briefly, HT22 cells were cultured in DMEM (Hyclone, Logan, UT, USA; SH30243.01) supplemented with 10% heat-inactivated FBS (CellNest, Charlotte, NC, USA; CNF303-0500), penicillin (20 U/ml), and streptomycin (20 mg/ml) at 37℃ in a humidified incubator gassed with 95% air and 5% CO₂. For corticosterone treatment, HT22 cells at 40~50% confluence were incubated with corticosterone at 400 or 800 ng/ml in DMEM containing 1% FBS in 6-well plates. After 24 h, cells were washed in PBS and harvested.

Corticosterone levels in serum were measured as described previously [34]. Briefly, blood was collected from the hearts of sacrificed mice, centrifuged at 1,500 g for 15 min to obtain serum, and stored at -80℃ until use. Sera were diluted 1: 50 in the EIA buffer provided in the EIA kit (Cayman Chemicals, MI, USA). Each diluted serum sample (50 µl) was mixed with an equal volume of a specific antibody to corticosterone (corticosterone EIA antiserum) and acetylcholinesterase-linked corticosterone (AChE Tracer) from the EIA kit on a 96-well plate. The reaction was incubated for 2 h at room temperature on an orbital shaker rotating at 150 rpm. Then, the reaction mixture in each well was discarded and rinsed with the washing buffer provided in the EIA kit. The Ellman's Develop Reagent in the kit was added at 200 µl/well, and the plate was incubated for 90 min at room temperature in the dark on an orbital shaker (150 rpm). Finally, the absorbance at 405 nm was measured using a spectrofluorometer (SpectraMax® M5; Molecular Devices, Sunnyvale, CA, USA).

Behavioral tests were performed as described previously [31,32,35]. Mice were brought to the testing room 30 min prior to the start of each behavioral test. At all times non-specific background sounds were masked with 65-dB white noise. The behavior testing room was lit with indirect illumination of 20~30 lux for the social interaction test, elevated plus maze test, and novelty-suppressed feeding test and at 250 lux for the marble-burying test. Behavioral tests were carried out during the light cycle (9 A.M.~3 P.M.). The performance of each animal in the behavioral tests was recorded with a computerized video-tracking system (SMART; Panlab S.I., Barcelona, Spain) and/or webcam recording system (HD Webcam C210, Logitech, USA). All parts of the apparatus were cleaned with 70% ethanol between tests.

The social interaction test was performed as described previously [32,34]. In brief, a U-shaped field was prepared by partitioning an open field (45×45×40 cm) to the central point with a wall (20×40 cm, width by height), so that each field had one enclosed and one open square. On the test day, subject mice were individually placed in the empty U-shaped field for 5 min of exploration, during which the locomotive trajectory and the time spent in each field were recorded. This procedure was regarded as habituation to the test field. While the subject mice were returned to their home cages for 2~3 min, an active social target (10~12 weeks old male B6 mouse) was loaded in a circular grid cage (12 cm in diameter × 33 cm in height) and was positioned in the corner of the enclosed square (called the target zone), while an unanimated grid cage was placed on the opposite side (called the non-target zone). Then, each subject mouse was tested for its trajectory in the U-field for 10 min, during which the locomotive trajectory and the time spent in each zone were recorded.

The elevated plus maze test was performed as described previously [5,35]. Briefly, the elevated plus maze apparatus consisted of four arms (30×7 cm each) made of gray FOAMEX (Expanded PVC; LG Ltd., Korea), which were elevated 50 cm above the floor and placed at right angles to each other. Two of the arms had 20-cm-high walls (enclosed arms), while the other two had no walls (open arms). The illumination at the open center was adjusted to 40 lux. For the test, subject mice were individually placed at the center of the two open arms and left to explore all arms for 5 min. The number of entries into the open and enclosed arms and the time spent in each arm were recorded. Entry into each arm was scored as an event if the animal crossed over the line of the cross square (7×7 cm) with all paws.

The light-dark box test was performed as described previously [11]. Briefly, the light box (36×25×18 cm) and dark box (12×25×18 cm) were made from a clear plastic rat cage partitioned with a separating wall with a shuttle door (5×5 cm) at the center of the two compartments at floor level. The light box was open at the top and illuminated to 700 lux at the bottom. The dark box was painted black and covered with a removable black lid. For the test, mice were individually placed into the light box with the shuttle door closed. After 10 sec, the shuttle door was opened, and the total number of transitions between the dark and light compartments, latency to entry into the dark box, and time spent in the light/dark boxes were recorded for 5 min.

The open field test was performed as described previously [11,35]. Locomotor activity was measured in an open field (45 cm×45 cm×40 cm). The open field was lit with indirect illumination of 70 lux. Mice were individually placed at the center of the open field, and the locomotive trajectory, distance traveled, and times spent in the center and on the periphery were recorded for 30 min. The center area of the open field was defined as the inner rectangular area that constituted 30% of the open field.

The novelty-suppressed feeding test was performed as described previously [35]. In brief, mice were deprived of food, but not water, in their normal home cages for 24 h. On the test day, mice were individually transferred to a holding cage without food and water for 30 min. Then, each mouse was placed in a corner of the open field (45×45×40 cm), and a single food pellet (2×2×2.5 cm) placed on a circular piece of white filter paper (15 cm in diameter) in the center of the open field. The latency to the first bite of the food pellet was recorded. Immediately after the mouse began to eat the lab chow, the subject animal was removed and individually placed in its home cage with a weighed piece of chow. After 5 min, the amount of food consumed was determined based on the weight of the piece of chow.

The marble-burying test was carried out as described previously [35]. Empty normal plastic cages were filled with smooth bedding (JRS 3-4; J. Rettenmaier & Söhne, Rosenberg, Germany) up to 5 cm from the cage floor, and 12 marbles (glass balls with diameters of 1.5 cm) were placed evenly throughout the cage. Mice were individually allowed to freely explore the cage for 30 min, and the number of successfully buried marbles was counted. Marble "burying" was defined when less than 25% of a marble was visible.

Two-sample comparisons were carried out with Student's t-test, while multiple comparisons were conducted using one-way ANOVA followed by the Newman-Keuls multiple comparison test. Statistical significance analyses were performed with GraphPad PRISM 6 software (GraphPad Software. Inc, CA, USA). All data are presented as means±S.E.M. and statistical significance was established at the 5% level unless otherwise indicated.

Previously we reported that chronic restraints treatment for 2 h daily for 14 days (2 h×14 d RST) in mice produced anxiety-like behaviors [5]. In the present study, we investigated the mechanisms underlying stress-induced anxiety and its reversal by exercise using this stress-induced anxiety model. Mice treated with 2 h×14 d RST showed reduced dwelling times and entry numbers in the open arms in the elevated plus maze test, along with increased times in the dark box in the light-dark box test (Fig. 1A~F). Stress-treated mice also spent less time spent in the center area in the open field test, were more latent to eat a food pellet placed at the center of the open field in the novelty-suppressed feeding test, and buried fewer marbles in the marble-burying test (Fig. 1G~K). Stress-treated mice also exhibited decreased social interaction in the social interaction test (Fig. 1L~N). Thus, mice treated with repeated restraints displayed anxiety-like behaviors in various types of behavioral tests.

Toward understanding the mechanisms by which exercise produced anxiolytic effects, mice with stress-induced anxiety were treated with physical exercise for 1 h daily for 21 days (1 h×21 d EXE). Post-stress treatment with the 1 h×21 d EXE in mice with stress-induced anxiety reversed the reduced dwelling times in the open arms in the elevated plus maze trst and suppressed the increased times in the dark box in the the light-dark box test (Fig. 1A~F). Post-stress treatment with the 1 h×21 d EXE also increased the reduced times in the center in the open field test, reduced the increased latency to eating a food pellet placed at the center in the open field test, and increased the reduced marble burying in the marble-burying test (Fig. 1G~K). Post-stress treatment with the 1 h×21 d EXE increased the reduced social interaction in the social interaction test (Fig. 1L~N). Thus, post-stress treatment with exercise, although it was forcefully imposed, effectively blocked stress-induced anxiety-like behaviors in several types of behavioral tests. We included an imipramine group because this drug has shown to produce long-lasting effects in recovering impaired mood behaviors in the same stress-treated animal model [32,34]. Post-stress treatment with imipramine also produced anxiolytic effects in this animal model (Fig. 1A~N).

Next, we examined whether the therapeutic effects of exercise on stress-induced anxiety were transient or long lasting. To address this, four groups of mice were prepared; untreated control mice (CON group), mice treated with the repeated stress (2 h×14 d restraints: RST group), mice subjected to the 2 h×14 d restraints, followed by exercise or imipramine treatments (RST+EXE and RST+IMI groups). Post-stress treatments with exercise or imipramine were terminated on post-stress day 21 (Fig. 2A). When examined 7 days after the last exercise or imipramine treatment (i.e., post-stress day 28), mice treated with repeated stress (RST group) displayed reduced entries and visiting times in the open arms in the elevated plus maze test, and reduced marble burying in the marble-burying test (Fig. 2A~C). In contrast, mice subjected to repeated stress followed by exercise (RST+EXE group) or imipramine (RST+IMI group) exhibited significantly or partially increased entries into the open arms of the elevated plus maze and control levels of marble burying in the marble-burying test (Fig. 2A~C). When the same mouse groups were subjected to the elevated plus maze test and marble-burying test on post-stress day 84 (that is, 12 weeks after the 2 h×14 d RST), the anxiolytic effects of exercise were sustained (Fig. 2D and E). These results suggest that repeated stress caused by the 2 h×14 d RST produces long-lasting anxiety-like behaviors and the therapeutic effects of exercise on stress-induced anxiety are also persisting.

Previously, we demonstrated that the neuropeptide MCH was upregulated in the amygdala in mice treated with 2 h~14 d RST [31]. Consistent with this report, real-time PCR analysis indicated that the expression of MCH increased in both the amygdala and the hippocampus in mice treated with 2 h~14 d RST (Fig. 3A and B).

Mice exposed to restraint for 2 h had markedly higher serum corticosterone levels than controls (Fig. 3C and D). We wondered whether the stress-dependent increase of MCH in the brain was induced by this stress hormone. To address this question, we performed an in vitro experiment using HT22 cells, a murine hippocampal cell line. When HT22 cells were treated with corticosterone at 400 ng/ml or 800 ng/ml for 24 h, MCH expression increased significantly (Fig. 3E and F). This result supports the possibility that the stress hormone corticosterone triggers the upregulation of MCH in the brain.

Next, we examined whether stress-induced up-regulation of MCH in the brain was reversed by post-stress treatment with scheduled exercise. Immunohistochemical analysis of mice treated with repeated restraint indicated that MCH expression was upregulated in the hippocampus, basolateral amygdala (BLA), and lateral hypothalamus. The increased expression of MCH in each of these brain regions was reversed in animals that were treated with the scheduled exercise (Fig. 4A~E).

We examined whether pharmacological inhibition of MCH receptors in the brain in the post-stress period blocked stress-induced anxiety-like behaviors. Mice treated with repeated restraints (2 h×14 d RST) exhibited reduced social interaction in the social interaction test (Fig. 5A and B) and reduced marble burying in the marble-burying test (Fig. 5C). On the contrary, intraperitoneal administration of SNAP-94847 (an antagonist for MCH receptor 1) to mice subjected to the 2 h×14 d RST reversed the reduced social interaction (Fig. 5B) and increased the reduced marble burying (Fig. 5C) in the behavioral tests.