Voluntary exercise and increased food intake after mild chronic stress improve social avoidance behavior in mice

Physiology & Behavior 151 (2015) 264–271

Contents lists available at ScienceDirect

Physiology & Behavior journal homepage: www.elsevier.com/locate/phb

Voluntary exercise and increased food intake after mild chronic stress improve social avoidance behavior in mice Airi Otsuka a,b, Tetsuya Shiuchi a,c,⁎, Sachiko Chikahisa a, Noriyuki Shimizu a, Hiroyoshi Séi a a b c

Department of Integrative Physiology, Institute of Biomedical Science, Tokushima University Graduate School, Tokushima 770-8503, Japan Department of Food Science, Institute of Biomedical Science, Tokushima University Graduate School, Tokushima 770-8503, Japan. JST, PRESTO, Kawaguchi 332-0012, Japan

H I G H L I G H T S • • • •

Voluntary wheel running after social defeat stress reduced social avoidance. Voluntary exercise did not clearly affect gene expression and monoamine levels. Food intake was greatest in the stress + exercise group. Increases in food intake influenced energy metabolism and social avoidance.

a r t i c l e

i n f o

Article history: Received 19 March 2015 Received in revised form 5 July 2015 Accepted 17 July 2015 Available online 29 July 2015 Keywords: Voluntary exercise Social defeat stress Social avoidance Food intake

a b s t r a c t It is well-established that exercise can influence psychological conditions, cognitive function, and energy metabolism in peripheral tissues including the skeletal muscle. However, it is not clear whether exercise can influence social interaction with others and alleviate defeat stress. This study investigated the effect of voluntary wheel running on impaired social interaction induced by chronic social defeat stress (SDS) using the resident– intruder social defeat model. Mice were divided into three groups: control, stress alone, and stress + exercise. SDS was performed by exposing C57BL/6 mice to retired ICR mice for 2.5 min. The C57BL/6 mice were continuously defeated by these resident (aggressor) mice and, following 5 days of SDS, experienced 2 days of rest with no SDS. Mice in the stress + exercise group were allowed to voluntarily run on a wheel for 2 h after every SDS exposure. Two weeks later, compared to the control group, the stress group showed a higher ratio of time spent in the corner zone of a social interaction paradigm even though SDS did not elicit depressive- and anxiety-like behaviors. We also observed that voluntary exercise, which did not affect muscle weight and gene expression, decreased social avoidance behavior of stressed mice without clear changes in brain monoamine levels. Interestingly, food intake in the stress + exercise group was the greatest among the three groups. To test the effect of the exercise-induced increase in food intake on social behavior, we set up a pair-fed group where food intake was restricted. We then compared these mice to mice in the stress alone group. We found that the ratio of time spent in the corner zone of the social interaction test was not different between ad libitum- and pair-fed groups, although pair-fed mice spent more time in the corner zone when an aggressor mouse was present than when it was absent. In addition, pair-feeding did not show exercise-induced reductions of adrenal gland weight and enhanced the loss of body fat. Our findings indicate that voluntary exercise reduces social avoidance behavior induced by SDS. Further, we determined that SDS and exercise-induced increases in food intake partially influence energy metabolism and social avoidance behavior. © 2015 Elsevier Inc. All rights reserved.

1. Introduction Psychological stress is often influenced by our relationships with others. For example, in human society, social defeat stress (SDS) is a ⁎ Corresponding author at: Department of Integrative Physiology, Institute of Medical Biosciences, Tokushima University Graduate School, 3-18-15, Kuramoto, Tokushima, Tokushima 770-8503, Japan. E-mail address: [email protected] (T. Shiuchi).

http://dx.doi.org/10.1016/j.physbeh.2015.07.024 0031-9384/© 2015 Elsevier Inc. All rights reserved.

critical cause of psychiatric disorders such as depression [1]. Therefore, to prevent the development of depression, it is important to establish provisions against SDS. Several studies have established that exercise can influence not only the energy metabolism in peripheral tissues including the skeletal muscle, but also psychological conditions and cognitive function. A recent study in humans indicated that a certain amount of exercise is an important factor in overcoming the development of mood disorders such as depression and anxiety [2]. In addition, in rodents, physical

A. Otsuka et al. / Physiology & Behavior 151 (2015) 264–271

training, using a treadmill or running wheel, has been reported to improve depressive-like behavior in various disorders such as diabetes, Parkinson's disease, Rett syndrome, Huntington's disease, and ethanol withdrawal [3,4,5,6,7,8,9]. It is thought that chronic exercise promotes neurogenesis in the hippocampus, which is damaged by stressinduced elevations of glucocorticoids, thereby improving emotional behavior [10,11]. However, whether exercise can affect social interactions with others and protect against defeat stress has not been elucidated. The SDS model in rodents is one of the most ubiquitously used paradigms for psychological stress [12]. This model is also called “the resident-intruder paradigm” because an “intruder” (experimental) animal is placed in the cage of a “resident” (aggressor) animal. Rodents exposed to SDS show a variety of behavioral changes, including social avoidance [12,13], which has been shown to improve with the administration of anti-depression medication [14,15,16,17]. Therefore, social avoidance induced by SDS is used as a measure of depressive-like behavior and sociality. Generally, stress-induced elevations in glucocorticoids stimulate appetite via the central nervous system [18]. It is relatively wellestablished that certain types of stress can promote excess calorie intake, which is known as an allostatic behavior that can relieve psychological stress [19,20]. It has also been shown that voluntary wheel running induces food consumption [21]. Thus, food intake may mediate behavioral changes that are induced by SDS. However, the above hypothesis has been challenged since some opposing reports have suggested that SDS reduces body weight and food intake [22,23]. In this paper, we examined the effect of voluntary exercise on social interactions after mild SDS. Our study revealed that voluntary exercise enhances social interaction with others, and that exercise-induced increases in calorie intake after social stress are partially necessary for improving emotional behavior and maintaining body weight.

265

The SDS paradigm was performed for 5 days followed by 2 days of no SDS. This paradigm was repeated 2 times. After finishing all SDS paradigms, we evaluated behavioral aspects using the following tests (Fig. 1A). 2.3. Open field (OF) test To assess the effect of SDS on anxiety-like behavior, we measured locomotor activity of experimental mice in an open field chamber that consisted of an acrylic box (50 cm × 50 cm × 30 cm). The behavior of mice, total distance traveled, and time spent in the central area (25% of the box) were monitored for 10 min, and were analyzed using the OF program (O'Hara, Tokyo, Japan) derived from ImageJ (National Institutes of Health, Bethesda, MD, USA). 2.4. Social interaction (SI) test To assess social activity, experimental mice were exposed to a SI test using the open field chamber. The chamber was separated into three positions: the interaction zone (25% of the center area), corner zone (9 cm × 9 cm; four positions in the corner zone), and others. The SI test was performed two times, and behavior of the mice was monitored for 2.5 min. For the first test, we put an empty gauze cage into the central position and monitored the behavior of experimental mice. The experimental animal was then placed back in its home cage for 1 min. For the second test, we put the same gauze cage into the central position, however, this time it contained the aggressor mouse. We then monitored behavior of experimental mice (Fig. 1B), which included total distance traveled and time spent in the interaction zone (25%). This behavior was then analyzed using the OF program (O'Hara, Tokyo, Japan) derived from ImageJ (National Institutes of Health, Bethesda, MD, USA). Time spent in the corner zone was analyzed using a recorded video.

2. Materials and methods 2.5. Tail suspension test 2.1. Animals Experimental mice consisted of C57BL/6 male mice (Japan SLC, Shizuoka, Japan) that were 10-weeks-old at the beginning of the experiment. Aggressor mice consisted of retired male ICR mice (Japan SLC, Shizuoka, Japan). All mice were individually housed, and maintained on a 12-hour light/dark cycle (light on at 9:00 am) with food and water available ad libitum for 2 weeks. Experimental animals of similar body weight were randomly assigned to the following groups: Control, Stress, Stress + Exercise, and Exercise [Control, 27.2 ± 1.1 g; Stress, 27.8 ± 0.8 g; Stress + Exercise, 26.0 ± 0.7 g, Exercise, 29.3 ± 0.3 g]. Mice in the exercise group were individually housed in cages equipped with a running wheel and habituated for 2 days before experimental procedures began. The Animal Study Committee of Tokushima University approved these experiments, and we performed them in accordance with Guidelines for the Care and Use of Animals approved by the Council of the Physiological Society of Japan. 2.2. SDS and wheel running paradigms The SDS paradigm was conducted based on a previous report with slight modifications [12]. In this paradigm, experimental mice were put into an aggressor's home cage for 2.5 min every day. After the physical interaction, aggressor and experimental mice were maintained in sensory contact for 1 h using an acryl plate that divided the aggressor's home cage in half. After sensory contact, experimental mice were transferred to another cage with or without free access to an activity wheel for 2 h. Experimental mice were then housed in the same cage as the resident aggressor mouse, using the same acrylic plate that had been employed for sensory contact following SDS. Control and exercise groups that were not exposed to SDS were also developed.

To assess depressive-like behavior, experimental mice were exposed to the tail suspension test in which the tips of their tails were fixed with adhesive tape to wires that dangled from the ceiling. The percentage of time that mice spent immobile was measured for 5 min, and decreased immobility time indicated reduced depressive-like behavior. Data were recorded using the Image FST program (O'Hara), which is also based on ImageJ. 2.6. Forced swim test To assess depressive-like behavior using another paradigm, experimental mice were exposed to the forced swim test. Mice were placed individually in an acrylic cylinder (22 cm high and 11.5 cm in diameter) that contained a 15-cm-deep water column (temperature 25 °C ± 1 °C). As with the tail suspension test, the percentage of time that mice spent immobile was measured for 5 min. A mouse was considered immobile when it was observed to be floating in the water in an upright position without moving. Decreased immobility indicates reduced depressivelike behavior. These data were recorded using the Image FST program (O'Hara). 2.7. Monoamine concentration in the brain Experimental mice were decapitated, and the whole brain was removed and dissected into five brain regions (striatum, cortex, hypothalamus, hippocampus, and amygdala). All tissue was extirpated from the section by tweezers, frozen rapidly in liquid nitrogen, and stored at − 80 °C. Monoamine levels (dopamine, 3,4-dihydroxyphenylacetic acid [DOPAC], homovanillic acid [HVA], serotonin [5-HT], and 5hydroxyindoleacetic acid [5-HIAA]) were quantified by highperformance liquid chromatography (HPLC) according to previously

266

A. Otsuka et al. / Physiology & Behavior 151 (2015) 264–271

A 1st Social-defeat stress

2nd Social-defeat stress

8

5

Day 1

Stress, Stress + Exercise

12

OF

SI

TS

FS

15

16

17

18

Sampling

19

Stress group

Experimental mouse: C57BL/6

Experimental mouse’s home cage (2 hrs) Stress + Exercise group Aggressor: retired ICR mouse

Physical contact (2.5 min)

Sensory stimulation (until next SDS) in Aggressor’s home cage

Sensory stimulation (1 hr)

Aggressor’s home cage Wheel running In Experimental mouse’s home cage (2 hrs) Control group

Control, Exercise

Experimental mouse’s home cage

Experimental mouse’s home cage Experimental mouse’s home cage Exercise group

Wheel running In Experimental mouse’s home cage (2 hrs)

B Interaction zone

Corner zone

Corner zone

Fig. 1. Schematic chart of the social defeat stress (SDS) paradigm (A). Experimental mice were separated into four groups: control, stress, stress + exercise, and exercise. Stress + exercise and stress groups mice were placed into an aggressor's home cage and experienced physical contact with aggressive ICR mice mouse for 2.5 min (upper part). In contrast, whereas mice in other groups were placed into their own cages (bottom part). Stress + exercise and exercise groups were placed into cages with running wheels after each SDS exposure whereas all other groups were placed into cages without running wheels. OF: open field test, SI: social interaction test, TS: tail suspension test, FS: forced swim test. Social interaction test paradigm (B). The chamber was separated into three positions: the interaction zone, corner zone, and other zones.

published methods with slight modifications [24]. DOPAC/dopamine, HVA/dopamine, and 5-HIAA/5-HT ratios were used to estimate the metabolic ratio. 2.8. Real-time reverse transcription-polymerase chain reaction RNA was prepared from the skeletal muscle of mice using a commercially available isolation protocol (RNAiso Plus; Takara Bio,

Shiga, Japan). cDNA was generated using a Gene Amp RNA polymerase chain reaction (PCR) kit (Applied Biosystems, Foster, CA). We used predesigned, gene-specific SYBR Green probes and primer sets to assess the expression of the following genes: PPARα (PPARΑ), CPT1b (CPT1B), myoD (MYOD1), GLUT4 (SLC2A4), myostatin (MSTN), UCP3 (UCP3), PGC1α (PPARGC1A), and 18S (18S). Primers for the genes are shown in Supplementary Table 1. The real-time (RT)-PCR reaction was carried out with an Applied Biosystems 7900HT real-time RT-PCR system using

A. Otsuka et al. / Physiology & Behavior 151 (2015) 264–271

SYBR Green PCR Master Mix (Roche Diagnostics, Indianapolis, USA) according to the manufacturer's instructions. For endogenous quantity control, we used 18S and normalized values to 18S mRNA expression. 2.9. Statistical analysis Values are expressed as mean ± standard error (SE) in the text and figures. Data were analyzed by a one-way analysis of variance. If a statistically significant effect was found, post hoc analysis (Bonferroni/ Dunn) was performed to detect differences between groups. A value of p b 0.05 was considered to be statistically significant. 3. Results

267

11.6; Exercise, 80.4 ± 16.2 (time in the center area)] (Fig. 3). As described in our Materials and methods section, the tail suspension and forced swim tests are used to assess depressive-like behaviors. In the present study, no significant difference was observed between 3 groups in terms of the amount of time the mice were immobile (Fig. 4). Taken together, our findings indicate that exposure to SDS did not induce anxiety- and depressive-like behaviors. 3.3. Food intake Food intake on the day was significantly increased in the stress + exercise group compared with the control group [Control, 3.33 ± 0.12; Stress, 3.48 ± 0.14; Stress & Exercise, 4.03 ± 0.20; Exercise, 3.45 ± 0.09 (food intake g/day)] (Fig. 5).

3.1. Body composition 3.4. Pair-fed mice and general parameters Weight gain was similar between the control and exercise groups, while mice in the stress only group gained less body weight following SDS. The amount of epididymal white adipose tissues (eWAT) was significantly decreased in both SDS-exposed groups compared with the control group. Moreover, adrenal gland weight was significantly increased in the stress only group compared with the control group; however, corticosterone level did not significantly differ between the control and stress groups [Control, 19.8 ± 9.3; Stress, 23.1 ± 9.6 (ng/ml)]. Weight of the musculus tibialis anterior was not different between any of the groups (Table 1). We next screened for the mRNA of interest to examine in skeletal muscles so that we might investigate the influence of exercise; however, we found no difference between any of the groups (Supplementary Fig. 1).

We hypothesized that food intake was associated with reductions in stress-related social avoidance. To test the hypothesis, we generated a group of pair-fed mice that were exposed to similar conditions as the stress + exercise group [Stress + Exercise (pair-fed)]. We found that the pair-fed group reduced their food intake to near the average of stress mice (3.5 g/day). Moreover, the pair-fed group exhibited obvious reductions in body weight and eWAT compared with the control group. Running distance was not different between stress + exercise mice that were ad libitum-fed and those that were pair-fed. Additionally, adrenal gland weight between the two groups was not significantly different although adrenal gland weight in pair-fed was greater compared to the control group (Table 1). 3.5. Pair-fed mice and behavioral testing

3.2. Behavioral tests The SI test was performed to assess the social preference of experimental mice. Experimental mice were exposed to a novel, non-social object and an aggressor mouse, and time spent in delineated zones of the testing box was measured. The interaction zone was identified as the area proximal to the novel object (first exposure) and the aggressor (second exposure). We also investigated time spent in the corner zones away from the interaction zone. The ratio of time spent with the novel object to time spent with the aggressor mouse indicates social preference. We found that mice in the stressed group spent more time (i.e., higher ratio) in corner zones compared with the control group. Interestingly, this effect was not seen in the stressed + exercise group [Control, 0.95 ± 0.14; Stress, 2.29 ± 0.30; Stress + Exercise, 1.51 ± 0.22, Exercise, 0.83 ± 0.21] (Fig. 2). In order to assess anxiety-like behavior, we employed the OF in which experimental mice were put into a novel environment that contained an open field. We found that there was no difference in the total distance and time spent in center areas of the field among all groups [Control, 3161.5 ± 159.2; Stress, 2657.0 ± 266.2; Stress + Exercise, 3171.1 ± 189.2; Exercise 3284.5 ± 284.9 (total distance)] [Control, 78.8 ± 10.8; Stress, 74.4 ± 13.3; Stress + Exercise, 83.9 ±

There was no difference in time in interaction zone between stress + exercise ad libitum- and pair-fed mice when an aggressor mouse was present [Stress + Exercise, 47.4 ± 5.2; Stress + Exercise (pair-fed), 50.0 ± 14.6 (s)]. However, pair-fed mice spent longer times in the corner zone when an aggressor mouse was present compared to the presence of a novel, non-social object [Stress + Exercise (pair-fed), 90.3 ± 15.0 (s, aggressor mouse); 67.3 ± 12.6 (s, novel object)]. As for the aggressor mouse/novel object ratio, there was no difference between the Stress + Exercise ad libitum-fed and pair-fed groups both for time spent in interaction and corner zones [Stress + Exercise, 1.07 ± 0.24; Stress & Exercise (pair-fed), 0.91 ± 0.22 (interaction zone ratio)] [Stress + Exercise, 1.51 ± 0.22; Stress + Exercise (pairfed), 1.53 ± 0.32 (corner zone ratio)] (Fig. 2). The OF showed no difference between the ad libitum- and pair-fed mice in the Stress + Exercise condition (Fig. 3). 3.6. Dopamine and its metabolite concentration in the brain In the striatum, the exercise group showed significantly higher HVA concentrations than other groups, and the dopamine metabolic ratio (HVA/DA) was higher when compared with the stress alone group. In

Table 1 Weight gain, running wheel routines, and tissue weight. Data are presented as mean ± SE, n = 6. Significant differences were detected using a Bonferroni/Dunn test for each pair of experimental groups; different superscript letters denote p b 0.05.

Body weight gain (g) Running distance TA (mg)/BW (g) eWAT (mg)/BW (g) Adrenal gland (mg)/BW (g)

Stress (−)

Stress (+)

Stress (+)

Stress (−)

Stress (+)

Exercise (−)

Exercise (−)

Exercise (+)

Exercise (+)

Exercise (+) (pair-fed)

0.83 ± 0.10ac – 1.6 ± 0.07 10.7 ± 0.78a 0.15 ± 0.008a

−0.30 ± 0.28ab – 1.7 ± 0.06 6.4 ± 0.42b 0.23 ± 0.013b

1.48 ± 0.20c 757 ± 233.3a 1.7 ± 0.06 6.5 ± 0.28b 0.19 ± 0.010ab

0.43 ± 0.30abc 2542 ± 483.9b 1.8 ± 0.05 10.4 ± 0.88a 0.16 ± 0.009a

−0.67 ± 0.44b 893 ± 187.1a 1.8 ± 0.03 4.3 ± 0.31 0.22 ± 0.012b

268

A. Otsuka et al. / Physiology & Behavior 151 (2015) 264–271

B

A

15

# 60

# 30

No aggressor Aggressor

Ratio of time in interaction zone (aggressor / no aggressor)

Time in interaction zone (s)

90

0

12

9

6

3

0

D

C #

4

#

90

60

30

No aggressor Aggressor

Ratio of time in corner zone (aggressor / no aggressor)

120

0

p < 0.01

3

** 2

1

0

Fig. 2. Examination of social behavior using the social interaction test. Time spent in the interaction zone (A) and the ratio of time spent in the interaction zone (B). Time spent in the corner zone (C) and the ratio of time spent in the corner zone (D). A white bar indicates that there was no aggressor present while a black bar indicates that the aggressor mouse was present (A, C). It is calculated the time in interaction zone when aggressor present/no aggressor (B, D). Data are presented as mean ± SE, n = 6–8. **: p b 0.01 versus the control group. #: p b 0.05 versus the respective no- aggressor trial. A Bonferroni/Dunn post hoc analysis was done to detect differences between groups.

the hypothalamus, the exercise group showed higher concentrations of DA and DOPAC, while the stress + exercise group showed low DOPAC and HVA levels; the DA metabolic ratio in the hypothalamus was not different between any of the groups. In the hippocampus, there was a significant difference in the dopamine metabolic ratio (DOPAC/DA) although DA, DOPAC, and HVA concentrations were not different between any of the groups. In the amygdala, SDS was found to increase the dopamine metabolite, HVA, and voluntary exercise was found to decrease those levels (Table 2).

3.7. Serotonin and its metabolite concentration in the brain The exercise group showed higher 5-HIAA and serotonin metabolic ratio in the striatum and hypothalamus than other groups. In the hypothalamus, groups exposed to SDS showed lower 5-HIAA concentrations compared with groups that were not exposed to SDS. In the amygdala, the stress + exercise group decreased concentrations of 5-

HT and 5-HIAA, while the exercise only group increased concentrations of 5-HT (Table 3). 4. Discussion Previous studies have demonstrated that chronic SDS can lead to several psychological deficits such as depression, anxiety, and cognitive impairment. Stress-induced reduction in social interaction is thought to be an important diagnostic phenomenon of these psychological deficits. In our current study, we found that SDS-induced avoidance behavior can be ameliorated by small amounts of voluntary exercise. Moreover, we found that both stress and exercise induce a concomitant increase in food intake, and that this partially influences social behavior. Chronic social stress not only induces avoidant social behavior, but is also frequently accompanied by anxiety- and depressive-like behaviors in rodents. In experiments, these behaviors often complicate the interpretation of data. In our study, we tried to employ a social interaction-specific stress model where experimental mice were

A. Otsuka et al. / Physiology & Behavior 151 (2015) 264–271

A

269

A Tail suspension test 100 Total immobility (%)

Time in center area (s)

120

80

40

50

0 Control

0

Stress

Stress + Exercise

B Forced swim test

B

Total immobility (%)

100

50

0 Control

Stress

Stress + Exercise

2000

Fig. 4. Examination of depressive-like behaviors using the tail suspension (A) and forced swim (B) tests. Data are presented as mean ± SE, n = 6. Bonferroni/Dunn was used to assess differences between groups.

0

Fig. 3. Examination of anxiety-like behavior using the open field test. Time spent in the center (A) and total distance (B). Data are presented as mean ± SE, n = 8 for control, stress, stress + exercise, and stress + exercise (pair fed) groups, and n = 6 for the exercise alone group. A Bonferroni/Dunn test was performed to detect differences between groups.

physically attacked by aggressive ICR mice for 2.5 min a day (2 times for 5 consecutive days with 2 interval days of no physical contact) (Fig. 1A). Experimental mice were housed in the same cage as aggressor ICR mice; however, mice were separated from each other by an acrylic plate. Our results showed that social defeat did not increase anxiety- and depressive-like behaviors as assessed by the OF, forced swim, and tail suspension tests. On the other hand, we did observe significant differences in the SI test. When an aggressor mouse was present in the interaction zone, stressed mice spent significantly lesser time in the interaction zone and spent significantly more time in the corner zone. Moreover, the stressed group showed a significantly lower ratio of time spent in the interaction zone compared to mice in the exercise group (Fig. 2B). Although the adrenal gland appeared to enlarge after SDS, plasma corticosterone levels were comparable to those in the control group, suggesting that the social stress employed in this study was milder than that of previous reports. Exercise is known to play beneficial roles in psychological conditioning and brain plasticity. However, at the point this study was conducted, it was unclear whether spontaneous exercise could affect social behavior following SDS. As discussed, SDS elicited avoidant social behavior, and this could be reversed by spontaneous wheel running. Intriguingly, the spontaneous exercise that mice performed for 2 h after SDS was significantly shorter (in distance) than that of the exercise without stress group, which exhibited positive interactive behaviors with

aggressor mice. Furthermore, the stressed + exercised mice seemed to exhibit even less immobility time, but this observation was not statistically significant. These findings are consistent with previous reports that exercise can improve depressive-like behavior. In addition, we found no significant differences between groups regarding gene expression related to lipid metabolism, glucose metabolism, and cell differentiation in skeletal muscles. Thus, these results suggest that very low intensity and small amounts of exercise can prevent an SDSinduced reduction in social interaction without changing metabolic function in skeletal muscles. Feeding behavior is thought to be an allostatic action that can control an excited and/or depressed psychological condition [25]. In the current study, we found that social stress and exercise themselves did not alter food intake; however, social stress followed by spontaneous exercise for 2 h was found to increase food intake. Therefore, we prepared another group so that we could compare social behavior in stressed + exercised mice with free access to food to stressed + exercised mice that were food restricted. Our results showed that, without a concomitant increase in food intake, the exercise-induced reduction in avoidant social behavior with an aggressor mouse was diminished in SDSexposed mice, whereas the ratio of social avoidance remained 5

Food intake (g/day)

Total distance (cm)

4000

* 4

3

2 Control

Stress

Stress + Exercise

Exercise

Fig. 5. Average daily chow intake. Data are presented as mean ± SE, n = 6. *Significant difference (p b 0.05) using the Bonferroni/Dunn test.

270

A. Otsuka et al. / Physiology & Behavior 151 (2015) 264–271

Table 2 Dopamine, dopamine metabolite concentrations, and dopamine turnover in the brain. Data are presented as mean ± SE, n = 4–6. Significant differences were detected using a Bonferroni/ Dunn test for each pair of experimental groups; different superscript letters denote p b 0.05.

Striatum

Control Stress Stress + Exercise Exercise Control Stress Stress + Exercise Exercise Control Stress Stress + Exercise Exercise Control Stress Stress + Exercise Exercise

Hypothalamus

Hippocampus

Amygdala

DA

DOPAC

HVA

DOPAC/DA

7984 ± 916 10,678 ± 550 9724 ± 451 7670 ± 1781 108.3 ± 30.7ab 78.6 ± 23.5a 92.8 ± 11.6a 231 ± 35.3b 16.1 ± 4.3 14.2 ± 2.9 47.7 ± 36.4 279.2 ± 179.0 218.8 ± 45 657.0 ± 284 85.7 ± 29 700.8 ± 281.6

1286 ± 87 1262 ± 84 1469 ± 124 2060 ± 342 124.8 ± 11.3a 88.7 ± 6.2ab 64.6 ± 5.4b 187.3 ± 21.2c 60.2 ± 12.8 34.9 ± 7.9 40.4 ± 24.3 50.1 ± 24.8 495.4 ± 190.7 842.0 ± 172.9 146.6 ± 61.8 123.3 ± 50.1

551 ± 31a 617 ± 23a 628 ± 25a 1454 ± 267b 130.1 ± 6.0a 125.6 ± 8.0ab 96.3 ± 5.3b 134.3 ± 6.7a 72.5 ± 5.7 53.5 ± 10.8 42.9 ± 7.5 59.7 ± 18.6 202.1 ± 51.7ab 290.5 ± 56.6a 66.7 ± 22.1b 130.2 ± 31.2ab

0.18 ± 0.04ab 0.12 ± 0.01a 0.15 ± 0.01ab 0.33 ± 0.07b 3.07 ± 1.31 2.43 ± 1.23 0.78 ± 0.14 0.88 ± 0.15 4.60 ± 0.59a 2.51 ± 0.21b 1.58 ± 0.27bc 0.45 ± 0.12c 1.96 ± 0.33 3.37 ± 1.45 1.98 ± 0.38 0.65 ± 0.42

unchanged (Fig. 2). These results suggest that spontaneous exercise, along with an associated increase in food intake, can partially improve stress-induced deficits in social interaction. Interestingly, compared to the non-stressed groups, chronic social defeat tended to increase the time the mice spent in the interaction zone when a target aggressor mouse was absent; however, this increase was not statistically significant. Furthermore, this phenomenon was not reversed by exercise. The social interaction test was performed in the same field as the open field test (Fig. 1B), and the interaction zone in the social interaction test corresponded to the center area of the open field test. The social interaction test was conducted a day after the open field test was administered (Fig. 1A); therefore, mice had been exposed to the same area twice. We, therefore, speculated that stressed mice had a greater amount of vigilance when seeking out novel objects rather than just increased curiosity; we also determined that exercise did not inhibit this behavior. Additional experiments are necessary to confirm this speculation. A recent study demonstrated that chronic social stress can induce anxiety-like behavior through epigenetic mechanisms in the hippocampus and that these changes can be normalized by treadmill exercise [26]. Therefore, we hypothesized that stress and/or exercise might cause epigenetic changes in neuronal networks of specific brain regions. Table 3 Serotonin, serotonin metabolite concentrations, and serotonin turnover in the brain. Data are presented as mean ± SE, n = 4–6. Significant differences were detected using a Bonferroni/Dunn test for each pair of experimental groups; different superscript letters denote p b 0.05. 5-HT

5-HIAA

Control Stress Stress +

377.0 ± 30.0 391.3 ± 3.96 344.3 ± 8.18

0.57 ± 0.04a 207.9 ± 5.84a 226.4 ± 11.76a 0.58 ± 0.03a 207.6 ± 8.53a 0.60 ± 0.02a

Exercise Exercise Hypothalamus Control Stress Stress +

401.8 ± 28.8 493.4 ± 91.4 416.1 ± 84.4 504.3 ± 36.3

379.2 ± 53.74b 680.0 ± 14.1a 576.7 ± 41.4ab 419.3 ± 27.2b

0.93 ± 0.06b 2.01 ± 0.57 1.91 ± 0.63 0.85 ± 0.07

Hippocampus

Exercise Exercise Control Stress Stress +

708.5 ± 73.8 279.6 ± 14.0 324.6 ± 25.7 277.3 ± 25.6

876.3 ± 58.4c 385.7 ± 32.5 340.0 ± 25.3 325.5 ± 11.9

1.29 ± 0.17 1.38 ± 0.08 1.06 ± 0.08 1.20 ± 0.07

Amygdala

Exercise Exercise Control Stress Stress +

326.4 ± 34.5 349.3 ± 28.6a 299.4 ± 41.03b 107.7 ± 24.9b

377.2 ± 26.6 604.6 ± 19.6a 575.1 ± 36.7a 194.8 ± 53.5b

1.18 ± 0.08 1.80 ± 0.15ab 2.28 ± 0.54a 1.68 ± 0.14ab

Exercise Exercise

880.9 ± 98.6C

480.3 ± 25.1a

0.56 ± 0.04b

Striatum

5-HIAA/5-HT

0.07 ± 0.007a 0.06 ± 0.004a 0.06 ± 0.002a 0.22 ± 0.030b 3.49 ± 1.63 3.48 ± 1.80 1.15 ± 0.20 0.64 ± 0.11 7.74 ± 2.31 3.92 ± 0.30 4.85 ± 1.43 2.67 ± 1.32 0.89 ± 0.08 1.12 ± 0.43 0.86 ± 0.09 1.76 ± 1.35

However, as we have described, the intensity of the defeat stress used in this study was probably lower than that used in other studies because the mice that experienced SDS did not show anxiety-like behaviors or increases in plasma corticosterone levels. Moreover, the spontaneous exercise that mice performed in the stress + exercise group was of low intensity. Therefore, we consider that there were quite few, if any, epigenetic changes in the brains of mice in this study. In the current study, we observed that SDS with or without exercise decreased the body fat of mice, but that spontaneous wheel running in mice that were not stressed only slightly decreased body fat when compared to controls (Table 1). In addition, spontaneous wheel running for 2 h after SDS showed similar reductions in body fat despite the fact that food intake in this group had increased. Finally, the body fat of pair-fed mice further decreased when compared to the ad libitum-fed group. These results suggest that SDS (both with and without spontaneous exercise) can enhance stress-induced energy expenditure; however, the relationship between social interaction and energy expenditure is unknown. The adrenal gland plays an important role in responding to stress as it is part of the hypothalamus–pituitary–adrenal gland axis. Moreover, hypertrophy of the adrenal gland is an indicator of the body's stress response. In the current study, we found that SDS increased adrenal gland weight, and that 2 h of spontaneous wheel running after SDS could reduce this hypertrophy. Interestingly, pair-feeding did not reverse the increase in adrenal gland weight. This finding indicates that excess food intake might alleviate adrenal gland hypertrophy. However, it is not clear whether stress-induced hypertrophy of the adrenal gland induces social avoidance. It has been shown that imbalanced monoamine levels in the brain can lead to certain mood disorders and psychiatric diseases, such as depression and anxiety [27,28,29]. For example, dopaminergic neuronal activity is implicated in fear and anxiety, and it has been reported that activation of the D2 receptor in the amygdala potentiates fear memory and anxiety-like behaviors [30]. On the other hand, it is still unclear whether exercise-induced anti-depressive effects are dependent on changes in neural monoamine levels [31,32]. Therefore, in the current study, we measured monoamine levels in several brain regions following SDS, but we failed to find clear results. In terms of DA, we found that spontaneous exercise after chronic social stress decreased DA metabolites in the amygdala compared to mice that were exposed to stress without exercise. Numerous studies have demonstrated that DA neurons are excited by aversive stimuli [33], and that stressful events are associated with DA release [34,35]. Thus, spontaneous exercise following SDS may partly decrease DA tone in the amygdala, thereby alleviating stress-induced social avoidance. Our methods, however, could not reflect direct DA release and/or dopaminergic neuronal activity during the SI test. Therefore, we calculated DA metabolic ratio

A. Otsuka et al. / Physiology & Behavior 151 (2015) 264–271

(DOPAC/DA or HVA/DA) to estimate DA release. Contrary to our expectation, we did not observe a statistically significant difference in DA metabolic ratio in the amygdala between any of the groups investigated. In contrast, the metabolic ratio of DOPAC/DA in the hippocampus of experimental mice was decreased by chronic social stress, and spontaneous exercise did not reverse this ratio. This latter finding suggests that DA metabolic ratio in the hippocampus is not involved in the exercise-induced improvement in social interaction. Serotonin is thought to be an important molecule in emotional stability [1,36], and there are some reports concerning the effect of exercise on changing 5-HT transporter expression [37]. Our present experiments showed that the 5-HT metabolite, 5-HIAA, was significantly decreased in the amygdala following SDS + exercise. However, 5-HT metabolic ratio in the amygdala did not exhibit similar results. Taken together, our monoamine studies suggest that further experiments are needed to fully understand the impact of spontaneous exercise on the brain following social stress. In conclusion, our study demonstrated the effects of voluntary exercise on social interaction after SDS. Further, we showed that voluntary exercise can enhance social interaction with others, and that an exercise-induced increase in calorie intake after social stress is partially necessary for improving emotional behavior and maintaining body weight.

[9]

[10]

[11]

[12]

[13] [14]

[15]

[16]

[17]

[18]

Disclosures The authors have no conflicts of interest, financial or otherwise to declare.

[19]

[20]

Acknowledgments This work was supported by Grant-in-aid for Scientific Research on Innovative Areas (26560398, to T.S.) from the Japan Society for the Promotion of Science. This work was also supported by the Uehara Memorial Foundation (to T.S.). We thank the Support Center for Advanced Medical Sciences at the Institute of Biomedical Sciences at Tokushima University Graduate School for their aid in our real-time PCR experiments. We also would like to thank Editage for English language editing.

[21]

[22]

[23] [24]

[25]

Appendix A. Supplementary data

[26]

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.physbeh.2015.07.024.

[27] [28]

References [1] A. Caspi, K. Sugden, T.E. Moffitt, A. Taylor, I.W. Craig, H. Harrington, J. McClay, J. Mill, J. Martin, A. Braithwaite, R. Poulton, Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene, Science 301 (2003) 386–389. [2] C. Laske, S. Banschbach, E. Stransky, S. Bosch, G. Straten, J. Machann, A. Fritsche, A. Hipp, A. Niess, G.W. Eschweiler, Exercise-induced normalization of decreased BDNF serum concentration in elderly women with remitted major depression, Int. J. Neuropsychopharmacol. 13 (2010) 595–602. [3] T. Renoir, C. Chevarin, L. Lanfumey, A.J. Hannan, Effect of enhanced voluntary physical exercise on brain levels of monoamines in Huntington disease mice, PLoS Curr 3 (2011) RRN1281. [4] T. Renoir, T.Y. Pang, M.S. Zajac, G. Chan, X. Du, L. Leang, C. Chevarin, L. Lanfumey, A.J. Hannan, Treatment of depressive-like behaviour in Huntington's disease mice by chronic sertraline and exercise, Br. J. Pharmacol. 165 (5) (2012) 1375–1389. [5] T.Y. Pang, T. Renoir, X. Du, A.J. Lawrence, A.J. Hannan, Depression-related behaviours displayed by female C57BL/6J mice during abstinence from chronic ethanol consumption are rescued by wheel-running, Eur. J. Neurosci. 37 (11) (2013) 1803–1810. [6] M.H. Kim, Y.H. Leem, Chronic exercise improves repeated restraint stress-induced anxiety and depression through 5HT1A receptor and cAMP signaling in hippocampus, J Exerc Nutrition Biochem 18 (1) (2014) 97–104. [7] W. Liu, X. Zhai, H. Li, L. Ji, Depression-like behaviors in mice subjected to cotreatment of high-fat diet and corticosterone are ameliorated by AICAR and exercise, J. Affect. Disord. 156 (2014) 171–177. [8] T. Tuon, S.S. Valvassori, G.C. Dal Pont, C.S. Paganini, B.G. Pozzi, T.F. Luciano, P.S. Souza, J. Quevedo, C.T. Souza, R.A. Pinho, Physical training prevents depressive

[29] [30]

[31] [32]

[33]

[34]

[35] [36]

[37]

271

symptoms and a decrease in brain-derived neurotrophic factor in Parkinson's disease, Brain Res. Bull. 108 (2014) 106–112. M.A. Kondo, L.J. Gray, G.J. Pelka, S.K. Leang, J. Christodoulou, P.P. Tam, A.J. Hannan, Affective dysfunction in a mouse model of Rett syndrome: therapeutic effects of environmental stimulation and physical activity, Dev. Neurobiol. 27 (2015 May). H. van Praag, B.R. Christie, T.J. Sejnowski, F.H. Gage, Running enhances neurogenesis, learning, and long-term potentiation in mice, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 13427–13431. S.Y. Yau, B.W. Lau, J.B. Tong, R. Wong, Y.P. Ching, G. Qiu, S.W. Tang, T.M. Lee, K.F. So, Hippocampal neurogenesis and dendritic plasticity support running-improved spatial learning and depression-like behaviour in stressed rats, PLoS One 6 (2011) e24263. O. Berton, C.A. McClung, R.J. Dileone, V. Krishnan, W. Rentha, S.J. Russo, D. Graham, N.M. Tsankova, C.A. Bolanos, M. Rios, L.M. Monteggia, D.W. Self, E.J. Nestler, Essential role of BDNF in the mesolimbic dopamine pathway in social defeat stress, Science 311 (2006) 864–868. E.J. Nestler, S.E. Hyman, Animal models of neuropsychiatric disorders, Nat. Neurosci. 13 (2010) 1161–1169. M. Razzoli, L. Carboni, M. Andreoli, F. Michielin, A. Ballottari, R. Arban, Strain-specific outcomes of repeated social defeat and chronic fluoxetine treatment in the mouse, Pharmacol. Biochem. Behav. 97 (2011) 566–576. S.D. Iñiguez, L.F. Alcantara, B.L. Warren, L.M. Riggs, E.M. Parise, V. Vialou, K.N. Wright, G. Dayrit, S.J. Nieto, M.B. Wilkinson, M.K. Lobo, R.L. Neve, E.J. Nestler, C.A. Bolaños-Guzmán, Fluoxetine exposure during adolescence alters responses to aversive stimuli in adulthood, J. Neurosci. 34 (2014) 1007–1021. C.H. Bourke, E.R. Glasper, G.N. Neigh, SSRI or CRF antagonism partially ameliorate depressive-like behavior after adolescent social defeat, Behav. Brain Res. 270 (2014) 295–299. K. Ramirez, D.T. Shea, D.B. McKim, B.F. Reader, J.F. Sheridan, Imipramine attenuates neuroinflammatory signaling and reverses stress-induced social avoidance, Brain Behav. Immun. ii (2015) S0889-1591(15)00024-0. K.E. Zakrzewska, I. Cusin, A. Stricker-Krongrad, O. Boss, D. Ricquier, B. Jeanrenaud, F. Rohner-Jeanrenaud, Induction of obesity and hyperleptinemia by central glucocorticoid infusion in the rat, Diabetes 48 (1999) 365–370. J.C. Chuang, M. Perello, I. Sakata, S. Osborne-Lawrence, J.M. Savitt, M. Lutter, J.M. Zigman, Ghrelin mediates stress-induced food-reward behavior in mice, J. Clin. Invest. 121 (2011) 2684–2692. B.C. Finger, T.G. Dinan, J.F. Cryan, The temporal impact of chronic intermittent psychosocial stress on high-fat diet-induced alterations in body weight, Psychoneuroendocrinology 37 (2012) 729–741. Y. Yasumoto, R. Nakao, K. Oishi, Free access to a running-wheel advances the phase of behavioral and physiological circadian rhythms and peripheral molecular clocks in mice, PLoS One 10 (2015) e0116476. W. Iio, Y. Tokutake, N. Matsukawa, T. Tsukahara, S. Chohnan, A. Toyoda, Anorexic behavior and elevation of hypothalamic malonyl-CoA in socially defeated rats, Biochem. Biophys. Res. Commun. 421 (2012) 301–304. W. Iio, N. Matsukawa, T. Tsukahara, D. Kohari, A. Toyoda, Effects of chronic social defeat stress on MAP kinase cascade, Neurosci. Lett. 504 (2011) 281–284. K. Kitaoka, A. Hattori, S. Chikahisa, K. Miyamoto, Y. Nakaya, H. Sei, Vitamin A deficiency induces a decrease in EEG delta power during sleep in mice, Brain Res. 1150 (2007) 121–130. J. Maniam, M.J. Morris, The link between stress and feeding behavior, Neuropharm 63 (2012) 97–110. G. Patki, L. Li, F. Allam, N. Solanki, A.T. Dao, K. Alkadhi, S. Salim, Moderate treadmill exercise rescues anxiety and depression-like behavior as well as memory impairment in a rat model of posttraumatic stress disorder, Physiol. Behav. 130 (2014) 47–53. A. Coppen, The biochemistry of affective disorders, Br. J. Psychiatry 113 (1967) 1237–1264. N. Nordquist, L. Oreland, Serotonin, genetic variability, behaviour, and psychiatric disorders—a review, Ups. J. Med. Sci. 115 (2010) 2–10. A. Kleinridders, W. Cai, L. Cappellucci, A. Ghazarian, W.R. Collins, S.G. Vienberg, E.N. Pothos, C.R. Kahn, Proc. Natl. Acad. Sci. U. S. A. 2 (2015) pii: 201500877. A.R. de Oliveira, A.E. Reimer, C.E. de Macedo, M.C. de Carvalho, M.A. Silva, M.L. Brandão, Conditioned fear is modulated by D2 receptor pathway connecting the ventral tegmental area and basolateral amygdala, Neurobiol. Learn. Mem. 95 (2011) 37–45. S. Brené, A. Bjørnebekk, E. Aberg, A.A. Mathé, L. Olson, M. Werme, Running is rewarding and antidepressive, Physiol. Behav. 92 (1–2) (2007) 136–140. L.M. Gorton, M.G. Vuckovic, N. Vertelkina, G.M. Petzinger, M.W. Jakowec, R.I. Wood, Exercise effects on motor and affective behavior and catecholamine neurochemistry in the MPTP-lesioned mouse, Brain Res. 213 (2) (2010) 253–262. F. Brischoux, S. Chakraborty, D.I. Brierley, M.A. Ungless, Phasic excitation of dopamine neurons in ventral VTA by noxious stimuli, Proc. Natl. Acad. Sci. U. S. A. 106 (2009) 4894–4899. S. Puglisi-Allegra, A. Imperato, L. Angelucci, S. Cabib, Acute stress induces timedependent responses in dopamine mesolimbic system, Brain Res. 554 (1991) 217–222. J.W. Tidey, K.A. Miczek, Social defeat stress selectively alters mesocorticolimbic dopamine release: an in vivo microdialysis study, Brain Res. 721 (1996) 140–149. B. Mo, N. Feng, K. Renner, G. Forster, Restraint stress increases serotonin release in the central nucleus of the amygdala via activation of corticotropin-releasing factor receptors, Brain Res. Bull. 76 (2008) 493–498. B.N. Greenwood, P.V. Strong, A.B. Loughridge, H.E. Day, P.J. Clark, A. Mika, J.E. Hellwinkel, K.G. Spence, M. Fleshner, 5-HT2C receptors in the basolateral amygdala and dorsal striatum are a novel target for the anxiolytic and antidepressant effects of exercise, PLoS One 7 (2012) e46118.