Stress-induced endocrine response and anxiety: the effects of comfort food in rats

Abstract

The long-term effects of comfort food in an anxiogenic model of stress have yet to be analyzed. Here, we evaluated behavioral, endocrine and metabolic parameters in rats submitted or not to chronic unpredictable mild stress (CUMS), with access to commercial chow alone or to commercial chow and comfort food. Stress did not alter the preference for comfort food but decreased food intake. In the elevated plus-maze (EPM) test, stressed rats were less likely to enter/remain in the open arms, as well as being more likely to enter/remain in the closed arms, than were control rats, both conditions being more pronounced in the rats given access to comfort food. In the open field test, stress decreased the time spent in the centre, independent of diet; neither stress nor diet affected the number of crossing, rearing or grooming episodes. The stress-induced increase in serum corticosterone was attenuated in rats given access to comfort food. Serum concentration of triglycerides were unaffected by stress or diet, although access to comfort food increased total cholesterol and glucose. It is concluded that CUMS has an anorexigenic effect. Chronic stress and comfort food ingestion induced an anxiogenic profile although comfort food attenuated the endocrine stress response. The present data indicate that the combination of stress and access to comfort food, common aspects of modern life, may constitute a link among stress, feeding behavior and anxiety.

• Anxiety
• body weight
• chronic stress
• comfort food
• corticosterone
• food intake

Introduction

Stress is considered a major risk factor for the development of various chronic diseases, such as metabolic syndrome, anxiety and other affective disorders (Cohen et al., 2012; Pacák & Palkovits, 2001; van Praag, 2004). Although stress has traditionally been defined as a non-specific response of the organism to environmental demands, the so-called “stressors” (Selye, 1974), it is now known that various stressors can trigger specific physiological and behavioral responses (Pacák & Palkovits, 2001; Zafar et al., 1997). Acute stress rapidly activates the sympathetic nervous system (SNS), thereafter activating the Hypothalamic-Pituitary-Adrenal (HPA) axis. Although activation of those two axis constitutes a hallmark of the stress response, the level of activation is stressor-specific.

In healthy individuals, the stress response is short-lived because activation of the SNS is rapidly counter balanced by the parasympathetic nervous system and the HPA axis response is terminated via negative feedback loops. Glucocorticoids act on receptors at the level of the paraventricular nucleus and the pituitary to inhibit the production and release of corticotropin-releasing hormone and adrenocorticotropic hormone. Indirect negative feedback from limbic structures such as the medial prefrontal cortex and hippocampus also contributes to termination of the HPA axis response (McEwen & Gianaros, 2010). Although the stress response is necessary for survival and adaptation, disorders can arise when it is prolonged and, in vulnerable individuals, exposure to chronic stress can adversely affect various aspects of health (McEwen & Gianaros, 2010).

Stress has long been associated with changes in dietary preference and food intake, as well as with weight gain and fat accrual, although the specific mechanisms are poorly understood. Psychological stress is linked to visceral deposition of fat, which is in turn associated with increased health risks (Chrousos, 2000). Accordingly, in animal models stress is known to suppress appetite (Bazhan et al., 2007; De Souza et al., 2000; Harris et al. 2001) and in human studies, it became clear that stress can decrease or increase food intake (Adam & Epel 2007; Epel et al., 2004) with stress-induced obesity being more prevalent (Anderson et al., 2006; Brunner et al., 2007). Indeed, repeated exposure of rats to chronic restraint stress modifies not only the amount of food intake but also eating patterns, particularly concerning the ingestion of sweet food (Ely et al., 1997). It has been demonstrated that humans under stress also show an increased preference for foods rich in sugar and fat (Tomiyama et al., 2011). Furthermore, eating foods rich in fat and carbohydrates during stress reduces the endocrine stress response (Dallman et al., 2003, 2005; Ortolani et al., 2011). It has been proposed that the desire to eat foods that are more satisfying is attributable to the effect of glucocorticoids that are released during stress (Pecoraro et al., 2005), as well as activation of the neural systems involved in the cognitive and reward aspects of ingestive behavior (Torres & Nowson, 2007).

We recently reported that the anorexigenic effect of foot-shock stress in rats is independent of leptin and is restricted to commercial chow intake, since no changes in the intake of comfort food was observed (Ortolani et al., 2011). We found that, in the Elevated Plus-Maze (EPM) and in the open field tests, rats submitted to foot-shock stress (one session per day for three consecutive days; i.e. short-term stress) exhibited less anxiety than did non-stressed rats, and that access to comfort food potentiated that effect. However, there have been no studies of the long-term effects of comfort food intake in a model of stress that is recognized as being anxiogenic.

In the present study, we evaluated behavioral, endocrine and metabolic parameters, including feeding behavior, exhibited by rats submitted to chronic unpredictable mild stress (CUMS) and given access to comfort food. We hypothesized that this chronic stress model, which is recognized as producing anxiety, would alter feeding patterns and increase the serum corticosterone concentration and that the intake of comfort food would attenuate these effects promoted by stress in rats.

Materials and methods

Animals

Forty-nine adult male Wistar rats, weighing 200 to 250 g at the beginning of the experiments, were used. The rats were housed in a temperature-controlled room (22 ± 2 °C), on a 12 h/12 h light/dark cycle (lights on at 07:00 h). The rats were housed in groups of five or four during the acclimation phase. After the first day of CUMS, control and stressed rats were housed individually in Plexiglas cages with sawdust covering the cage floor.

During the experiments, the rats were cared for in accordance with the Federation of American Societies for Experimental Biology Statement of Principles for the Use of Animals in Research & Education. The experimental protocols were approved by the Institutional Committee for Ethics in Animal Experimentation of the Federal University of São Paulo (certificate 1400/11).

The rats were distributed into four groups, according to diet and experimental protocol: those given access to commercial chow (CC) only, submitted to CUMS or not (CC-CUMS group and CC-control group, respectively); and those given access to commercial chow and to comfort food (CF), submitted to CUMS or not (CC/CF-CUMS group and CC/CF-control group, respectively). The diets were started three days before the first stress session and were maintained throughout the experimental period.

Diet composition

All rats were given ad libitum access to commercial chow for rodents (Nuvilab; Nuvital, Curitiba, Brazil) and tap water. The CC/CF-CUMS and CC/CF-control groups were also given ad libitum access to comfort food, which was composed of powdered commercial rodent chow (Nuvilab; Nuvital), peanuts (Hikari, São Paulo, Brazil), milk chocolate (Garoto, Vila Velha, Brazil) and sweet (cornstarch) cookies (Tostines; Nestlé, São Paulo, Brazil) in a proportion of 3:2:2:1. The comfort food diet, which was pelletised, provided the following (Ortolani et al., 2011): 20% protein, 20% fat, 48% carbohydrate and 4% fibre. Caloric densities, determined with an adiabatic calorimeter (C400; IKA Works, São Paulo, Brazil), were found to be 17.03 kJ/g for the commercial chow diet and 21.40 kJ/g (35% of calories as fat) for the comfort food diet (Estadella et al., 2004). To determine the amount of food consumed by each rat, we weighed the food remaining in the feeders at the end of each 24-hour period, using a digital scale.

CUMS protocol

The CUMS protocol was performed as previously described (Moreau et al., 1994a,b), and shown in Table 1. Rats were subjected to six different stressors, which varied from day to day, for 14 consecutive days, as follows: two 30-min periods of restraint, achieved through confinement in a small acrylic cylinder (16 × 7.5 × 7.0 cm); lights on overnight; water and food deprivation overnight (no access to water and food), followed by 2 h of restricted access to food (45 mg of food pellets were placed into a Petri dish which was positioned on the floor of the cage); water deprivation overnight (no access to water), followed by 1 h of contact with an empty bottle; damp sawdust overnight; and inversion of the light/dark cycle from Friday night to Monday morning.

Table 1. Chronic unpredictable mild stress protocol (Moreau et al., 1994).

Period	Monday (Day 1/8)	Tuesday (Day 2/9)	Wednesday (Day 3/10)	Thursday (Day 4/11)	Friday (Day 5/12)	Saturday (Day 6/13)	Sunday (Day 7/14)
Morning	Restraint (60 min)	Restraint (60 min)	Food restriction*	Empty water bottle (60 min)	Restraint (60 min)	Inversion of light/ dark cycle over the weekend	Inversion of light/ dark cycle over the weekend
				Restraint (60 min)
Afternoon	Restraint (60 min)	Restraint (60 min)	Restraint (60 min)	Restraint (60 min)
Night	Light on during the night	Water/food deprivation overnight	Water deprivation overnight	Damp sawdust overnight	Inversion of light/ dark cycle over the weekend**

*after overnight water/food deprivation, on Wednesday (Day 3/10) morning a Petri plate containing 45 mg of food pellets was positioned on the floor of the cage. Two hours later, food was made available ad libitum.

**On Friday night, lights were not turned off and remained on all night. Lights were turned off Saturday morning; they were turned on again Saturday night, and so on during the weekend. On Monday morning the light/dark cycle was restored (i.e. lights were on Monday morning and off Monday night).

EPM test

After the last day of the CUMS protocol, the rats were evaluated in the EPM, under a 60 lux light, as previously described (Melo et al., 1995; Ortolani et al., 2011; Pellow et al., 1985). The rats were placed individually in the centre of the maze, facing one of the closed arms at the junction between open and closed arms. For a period of 5 min, the rat’s behavior was registered on video. A rat was considered to have entered an arm of the maze if all four paws were within that arm. Conventional parameters of anxiety-like behavior were monitored, including the number of entries into closed arms, the number of entries into open arms and the total time spent in each arm. The ratio between time spent in open or closed arms and time spent in the open plus the closed arms was calculated and multiplied by 100 to obtain the percentage of time spent in all arms. These categories were defined as in previous studies (Anseloni & Brandão, 1997; Cruz et al., 1994; Ortolani et al., 2011). After the removal of each rat, the maze was cleaned with a 10% ethanol solution.

Open field test

Immediately after the EPM test, each rat was evaluated in the open field, under a 60 lux light. The apparatus consists of an acrylic cylinder, 72 cm in diameter, and 60 cm high; the base is divided into 12 equal areas (rectangles measuring 13.3 by 15.0 cm). As in the EPM test, the rat’s behavior was registered on video for a period of 5 min. Rats were placed individually in the center of the open field and left for 5 min, and the following parameters were analyzed: latency to the first crossing (a crossing defined as the rat entering another square with all four paws); time spent in the periphery and in the centre of the apparatus; number of crossings; and number of rearing and grooming episodes. After each trial, the apparatus was cleaned with a 10% ethanol solution.

Biochemical analyses

After the last day of chronic stress, one group of rats that were not evaluated in the behavioral tests were euthanized by conscious decapitation between 08:00 h and 11:00 h. Blood was collected in plastic vials and centrifuged for 20 min at 3363 g after which aliquots of serum were removed and stored at −80 °C until assayed for metabolic markers.

Serum corticosterone concentrations were determined by enzyme-linked immunoassay using a commercial kit (Enzo Life Science Inc., Ann Arbor, MI, USA). Samples were run in duplicate, and results are reported as ng/mL. Concentration was determined as percentage bound by using a standard curve ranging from 32 to 20,000 pg/mL. Kit sensitivity was 27 pg/mL and the intrassay coefficient of variation was 8%. For serum glucose, we used Accu-Chek (Roche Diagnostics, São Paulo, Brazil) for glucose test. For serum total cholesterol and triglycerides levels, a standard enzymatic colorimetric kit was used (Labtest Diagnostic S.A., Lagoa Santa, Brazil). Samples were run in duplicate, and the results were reported as mg/dl.

Statistical analysis

Results are expressed as mean ± standard error of mean. Unpaired Student’s t test was used to evaluate significant differences between unstressed and stressed rats. To analyze the effect of stress and the type of diet two-way analysis of variance followed by the Bonferroni test was used. Values of p < 0.05 were considered statistically significant.

Results

Food intake

Food intake was influenced by stress (F_[1,39] = 23.59; p < 0.0001) and diet (F_[1,39] = 2549; p < 0.0001). Rats in the CC-control group ingested less amount of commercial chow during the observation period compared with CC-CUMS group (p = 0.013, Table 2). Figure 1 shows that these groups ate similar amounts in the day before the start of the CUMS protocol (Day 0), and that the difference between groups first became significant after day 10 of the CUMS protocol (F_[1,19] = 241.8; p < 0.05). An apparent difference between CC-control and CC-CUMS on day 1 was not confirmed by statistical analysis. The rats that had access to both diets (CC/CF-CUMS and CC/CF-control groups) showed a significant preference for comfort food (F_[1,39] = 2427; p < 0.0001; Table 2). The rats in the CC/CF-CUMS group ate similar amounts of food compared with the other groups in the day before the CUMS protocol started (Day 0), but this group reduced their intake of commercial chow, as compared to control rats, after day 5 (F_[1,39] = 15.19; p < 0.0001; Figure 2) and reduced their comfort food intake on day 3, after the food restriction period of the CUMS protocol (F_[1,39] = 24.58; p < 0.0001; Figure 2). Therefore, the rats submitted to CUMS decreased their total caloric intake (F_[1,49] = 83330; p < 0.0001) as compared to control rats. Also, an interaction between stress and comfort food was observed (F_[1,49] = 30540; p < 0.0001).

Figure 1. Daily food intake (mean ± sem; n = 10 rats/group) by rats submitted (CUMS) or not (control) to chronic unpredictable mild stress and given access to commercial chow only (CC). Note: On day 0, rats were not submitted to CUMS (the day before the stress protocol), on days 2 and 9, rats submitted to CUMS had access to food and water for only 12 h and restricted access for an additional 2 hours. On days 3 and 10, rats submitted to CUMS had access to water for only 12 h. *p < 0.05 versus controls (two-way ANOVA with Bonferroni test).

Figure 2. Daily intake of commercial chow (CC) and comfort food (CF) shown separately for the control group and for the chronic unpredictable mild stress group (CUMS) given the combined food choice. Hence, total daily food intake (not shown) for the control and CUMS groups can be interpolated by summing daily CC and CF intakes for each group. Data are mean ± sem, n = 10 rats/group. On day 0, rats were not submitted to CUMS (the day before the stress protocol); on days 2 and 9, rats submitted to CUMS had access to food and water for only 12 h and restrict access for an additional 2 hours. On days 3 and 10, rats submitted to CUMS had access to water for only 12 h. *p < 0.05 versus controls (two-way ANOVA with Bonferroni test).’.

Table 2. Daily food and caloric intake, initial body weight and body weight gain by rats submitted (CUMS) or not (control) to chronic unpredictable mild stress and given access either to commercial chow (CC) only or to commercial chow and comfort food (CC/CF).

	Group
Parameter	CC-control	CC-CUMS	CC/CF-control	CC/CF-CUMS
Commercial chow intake (g/day)	24.9 ± 0.3	22.9 ± 0.7^a	4.77 ± 0.02	1.06 ± 0.4^a
Comfort food intake (g/day)	–	–	19.7 ± 0.1^b	19.1 ± 0.5^a,b
Total caloric intake (kJ/day)	423 ± 5	390 ± 12^a	562 ± 2^b	428 ± 10^a,c
Initial body weight (g)	202 ± 5	192 ± 5	196 ± 4	191 ± 6
Body weight gain (g)	79 ± 2	59 ± 4^a	104 ± 4^b	74 ± 5^a

Values are means ± sem; n = 10 rats/group.

^ap < 0.05 versus CC/CF-control (Student’s t-test); ^bp < 0.05 versus the intake of commercial chow; ^cp < 0.05; versus CC-CUMS (two-way analysis of variance followed by Bonferroni test).

Although all of the rats gained weight during the experimental protocol, the stress and the diet both affected the body weight. The weight gain was less pronounced in the CC-CUMS and CC/CF-CUMS group rats than in the CC-control and CC/CF-control group rats (F_[1,48] = 42.02; p < 0.0001). However, the weight gain was greater in the CC/CF-CUMS group rats than in the CC-CUMS group rats (F_[1,48] = 26.41; p < 0.0001; Table 2).

Behavioral analysis

The stress protocol affected behavior in the EPM. In comparison with their respective control groups, the CC-CUMS and CC/CF-CUMS groups showed a lower percentage of entries into and less time spent in the open arms of the EPM (F_[1,37] = 19.80; p < 0.0001 and F_[1,37] = 15.07; p < 0.0005, respectively), whereas the percentage of entries into and the time spent in the closed arms were both higher (F_[1,37] = 19.80; p < 0.0001 and F_[1,37] = 8.74; p = 0.005, respectively), as shown in Figure 3. Comfort food intake also affected the time spent in the arms of the EPM. In comparison with the CC-control rats, the CC/CF-control rats spent less time in the open arms (F_[1,37] = 12.76; p = 0.001) and more time in the closed arms (F_[1,37] = 8.89; p = 0.005). There was no interaction between stress and diet in the entries and the time spent in the arms of the EPM.

Figure 3. Percentage of entries into (A) and time spent in (B) the open and closed arms of the elevated plus-maze by rats submitted (CUMS) or not (control) to chronic unpredictable mild stress and given access either to commercial chow (CC) only or to commercial chow (CC) and comfort food (CF). (n = 10 rats/group). *p < 0.05, stress effect; **p < 0.05, diet effect (two-way ANOVA followed by Bonferroni test).

In the open field tests (Table 3), the stress protocol affected the time spent in the centre and the periphery. The CC-CUMS and CC/CF-CUMS group rats spent less time in the centre of the open field than did the corresponding control rats (F_[1,37] = 10.88; p = 0.002), consequently spending more time in the periphery (F_[1,37] = 10.19; p = 0.003). Access to comfort food did not alter those parameters. Neither stress nor diet affected the latency to the first crossing, the number of crossings, or the number of rearing and grooming episodes.

Table 3. Behavior in the open field test of rats submitted (CUMS) or not (control) to chronic unpredictable mild stress and given access either to commercial chow only (CC) or to commercial chow and comfort food (CC/CF).

	Group
Parameter	CC-control	CC-CUMS	CC/ CF-control	CC/ CF-CUMS
Latency to first crossing (s)	3.3 ± 0.7	3 ± 0.5	2.4 ± 0.4	3.8 ± 0.5
Time spent in centre (s)	51 ± 9.5	27 ± 5^a	42 ± 5.7	24 ± 3.8^a
Time spent in periphery (s)	249 ± 9.5	272 ± 4.5^a	258 ± 5.7	275 ± 3.8^a
Crossing	65 ± 7.4	64 ± 2.8	51 ± 2.6	64 ± 3.7
Rearing	44 ± 5.2	34 ± 3.4	30.6 ± 3.3	32 ± 3.3
Grooming	2.5 ± 0.6	2.9 ± 0.4	2.3 ± 0.5	2.4 ± 0.5

Values are means ± sem; n = 10 rats/group; ^ap < 0.05 versus. CC-control (two-way analysis of variance and Bonferroni test).

Hormone and metabolic parameters

Serum corticosterone concentration was higher in the CC-CUMS group rats than in the CC-control group rats (F_[1,24] = 7.94; p = 0.0005). Serum corticosterone concentration did not differ between the CC-control and CC/CF-control groups. However, the stress-induced increase in serum corticosterone concentration was attenuated in the CC/CF-CUMS rats (Figure 4).

Figure 4. Serum corticosterone concentration (mean ± sem) by rats submitted (CUMS) or not (control) to chronic unpredictable mild stress and given access either to commercial chow (CC) only or to commercial chow (CC) and comfort food (CF). CC-control group: n = 8 rats/group; CC-CUMS group: n = 5 rats/group; CC/CF-control group: n = 8 rats/group; CC/CF-CUMS group: n = 4 rats/group. *p < 0.05, stress effect; **p < 0.05, interaction between stress and diet effects (two-way ANOVA with Bonferroni test).

Table 4 shows that the serum total cholesterol and glucose concentrations were not affected by stress. However, these concentrations were significantly higher in the CC/CF-control rats than in the CC-control rats (F_[1,18] = 5.58; p = 0.032; F_[1,17] = 4.675; p = 0.048, respectively). There was no interaction between stress and diet. Serum triglyceride concentration was unaffected by the CUMS protocol or diet composition.

Table 4. Serum concentrations of glucose, total cholesterol and triglycerides of rats submitted (CUMS) or not (control) to chronic unpredictable mild stress and given access either to commercial chow (CC) only or to commercial chow and comfort food (CC/CF).

	Group
Serum parameter	CC-control	CC-CUMS	CC/ CF-control	CC/ CF-CUMS
Glucose (mg/dl)	111.0 ± 3.5	116.5 ± 2.8	124.8 ± 2.18^a	120.0 ± 6.9
Total cholesterol (mg/dl)	92.0 ± 3.5	104.8 ± 7.6	132.8 ± 25.9^a	122.0 ± 8.2
Triglycerides (mg/dl)	126.1 ± 13.3	108.7 ± 11.5	125.8 ± 13.7	148.7 ± 11.7

Values are means ± sem; n = 5 rats/group; ^ap < 0.05 versus CC-control (two-way analysis of variance and Bonferroni test).

Discussion

Stress is known to suppress appetite in laboratory rodents (Bazhan et al., 2007; De Souza et al., 2000; Harris et al., 2001), whereas in humans stress-induced obesity is more prevalent (Anderson et al., 2006; Brunner et al., 2007). Indeed, data presented here show that rats submitted to CUMS decreased their intake of commercial chow and comfort food, a different behavior compared with foot-shock stressed rats that reduced their intake of commercial chow but did not alter the intake of comfort food (Ortolani et al., 2011). This difference might be attributable to the type, duration, or severity of the stressor (Pecoraro et al., 2004). A stress-induced reduction in food intake has been demonstrated both as an acute response after a single stress and as a maintained decrease in 24-h food intake during and after repeated daily restraint stress (Krahn et al., 1990; Shimizu et al., 1989). Rats submitted to 3 h/day of restraint stress for 5 days increased their intake of fat and sucrose (Dallman et al., 2003; La Fleur et al., 2004; Liang et al., 2008; Pecoraro et al., 2004). However, in rats exposed to the same stressor for only 1 h/day, 5 days/week for 40 days, no such change was observed (Krolow et al., 2010). Glucocorticoids released during the stress response are considered to be orexigenic. Nevertheless, there are other influences in the regulation of food intake (Bazhan & Zelena, 2013) that should be taken into consideration, such as the catecholamines released by the sympathetic nervous system and adrenal medulla (Berthoud & Morrison, 2008); the other HPA axis hormones, including corticotropin-releasing hormone and adrenocorticotropic hormone; leptin and insulin (Adam & Epel, 2007); and the activation of the neural system involved with the cognitive, affective (reward), emotional and behavioral aspects of ingestive behavior (Torres & Nowson, 2007).

In the present study, all of the rats showed a clear preference for comfort food, and that preference was not altered by stress. From data in Table 2, it can be seen that in the normal chow-fed (CC) groups the stress-evoked reduction in caloric intake was more pronounced (about 92%) than it was in the chow-fed groups given comfort food (CF) (∼76%). Considering that the comfort food is more palatable than the commercial chow, it can be suggested that the stressed (CUMS) rats probably did not develop anhedonia. However, Moreau et al., (1994) found that rats submitted to the CUMS protocol decreased responsiveness to a rewarding stimulus, which was a bottle of drinking water with sucrose, and this decrease is considered an indication of anhedonia. However, we did not test the rat preference for sucrose, the gold standard method to test anhedonia.

It has been previously reported that the intake of comfort food reduces the HPA stress response (Dallman et al., 2003; Pecoraro et al., 2004; Ortolani et al., 2011), and this was evidenced by our finding since the stress-induced increase in the serum corticosterone concentration was attenuated in the rats with access to comfort food. It has been proposed that the intake of comfort food activates the reward system in the brain and dampens the stress responses (Dallman et al., 2003). In obese patients, the overconsumption of comfort foods is thought of as a kind of addiction, which is supported by imaging studies that have revealed dysregulation of the dopaminergic reward circuitry, similar to that observed in drug addiction (Volkow et al., 2011). Indeed, the ingestion of high-calorie foods is known to counteract some effects of stress, probably via the same mechanism reported for addiction disorders (Kiefer & Grosshans, 2009). The rewarding system was not addressed in the present work. However, we found that the rats submitted to CUMS showed an increased level of anxiety, evidenced by fewer entries into and less time spent in the open arms of the EPM, as well as less time spent in the centre of the open field. Such an anxiogenic effect of stress has previously been reported by others (Carobrez & Bertoglio, 2005; Krolow et al., 2010; Prut & Belzung, 2003; Teegarden & Bale, 2007). In contrast, in a previous study (Ortolani et al., 2011), we showed that foot-shock stress has an opposite, anxiolytic effect on rats submitted to the EPM test. These conflicting results may be due to the biphasic dose-dependent effects of glucocorticoids on anxiety-related behaviors: anxiolytic effects at low doses and anxiogenic effects at high doses (Vafaei et al., 2008). In addition, it has been hypothesized that the brain categorizes stressors and utilizes neural response pathways that vary in accordance with the assigned stress category (Herman et al., 2003). Categorization proponents (Sawchenko et al., 1996; Herman & Cullinan, 1997; Sawchenko et al., 2000; Herman et al., 2003) generally suggest that the brain discriminates between two major types of stressor: (1) stimuli which produce actual disturbances of physiological status that overwhelm specific homeostatic mechanisms, e.g. haemorrhage or infection; (2) stimuli which threaten the individual’s current or anticipated state, e.g. social conflict, aversive environmental stimuli, predator-related cues, failure to satisfy internal drives. Stressors in the first category have been labeled with terms such as physical or systemic; those of the second category have commonly been labeled as psychological or emotional. If the brain does deal with stressors categorically, one might predict that stressors should elicit category-specific patterns of neuronal activity in the brain. Cespedes et al., (2010) have reported that different brain areas are activated by foot-shock and restraint stress. In this case, the different nature of the stimuli might explain the different behavioral response to foot-shock stress (Ortolani et al., 2011) and to CUMS, as found in the present study.

It is of note that, in the present study, the control rats with access to comfort food for 17 days also spent less time in the open arms of the EPM. Fulton & Sharma, (2013) also reported that rats fed a high-fat diet for 12 weeks had higher levels of anxiety and depression, increased HPA axis responses to stress and biochemical changes in the reward system. These data indicate that, despite reducing the corticosterone response to stress, access to comfort food can affect other systems related or not to anxiety, such as the animal’s metabolic profile.

It is known that there is a relationship between stress, high intake of fat and metabolic disorders (Froy, 2007; Koolhaas et al., 2011). In humans, stress has been known to increase serum triglyceride concentrations (Rosmond et al., 1998). However, in animals, this response is variable, and the main effects are found in chronic stress, with a reduction or increase, or even no significant changes (Chuang et al., 2010; Ricart-Jané et al., 2002). CUMS and access to comfort food did not alter the serum triglyceride concentrations in the present study. However, the comfort food (high-fat diet) intake increased serum concentrations of cholesterol and glucose. There is an extensive literature characterizing responses to high-fat feeding in rodents, and the findings of the present study are in line with previous reports (Buettner et al., 2007; Yang et al., 2007). It is known that high fat diets produce a consistent and significant increase in body weight, hypertrophy and hyperplasia of adipocytes, dyslipidemia, and insulin resistance, depending on the amount of fat in the diet, duration of feeding, and strain of the experimental animal (Buettner et al., 2006; Woods et al., 2003). Therefore, consumption of energy rich high-fat diet may alleviate stress-induced anxiety, reduce the endocrine stress response, but probably will lead to obesity and metabolic diseases.

Conclusions

The present data indicate that CUMS has an anorexigenic effect on the intake of commercial chow and comfort food, in contrast with our previous finding that foot-shock stress altered only the intake of commercial chow (Ortolani et al., 2011). In the present study, the intake of comfort food attenuated the endocrine component of the stress response, as represented by the serum corticosterone concentration. Nevertheless, chronic emotional stress and the intake of comfort food induced an anxiogenic profile. Taken together, these data indicate that the combination of stress and access to comfort food, both of which are elements of modern life, may represent a link between stress and feeding behavior, as well as between stress and anxiety in people.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper. This study received financial support from the Brazilian Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES, Office for the Advancement of Higher Education Personnel, Federal Government, Brasil) and from the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP, Foundation for Research Support of the State of São Paulo, SP, Brasil).