Abstract

Introduction

Obesity has been an ever-increasing health problem in developed countries over the past three decades. In the United States, approximately two thirds of the population can be regarded as overweight, with a body mass index ≥ 25 kg/m² (Flegal et al., 2010). In the classic view, the human body regulates feeding behavior homeostatically. Hormonal regulators, such as insulin, leptin, neuropeptide Y, ghrelin, and others, act on specific brain regions, mainly the hypothalamus and hindbrain, to maintain homeostatic food intake and energy balance (Abizaid et al., 2006; Berthoud, 2012; Gao and Horvath, 2007; Havel, 2001; Steffens and Strubbe, 1987; Sumithran and Proietto, 2013; Williams and Elmquist, 2012; Woods et al., 1998). Disturbances in the homeostasis of the system that governs food intake result in an imbalance between the accumulation and expenditure of energy, leading to pathophysiological states. For example, the dysfunction of hormonal regulators, such as leptin deficiency, leads to overeating and positive energy balance, resulting in the development of obesity (Duca and Covasa; Gautron and Elmquist, 2011; Herzog, 2003). However, given the extreme rarity of genetic causes of obesity (Ramachandrappa and Farooqi, 2011), unlikelihood that human genetic changes have substantially accrued over the past three decades, and fact that energy expenditure through physical activities has not dramatically decreased from 1980 to 2005 (Westerterp and Speakman, 2008), a reasonable proposition is that the exponential increase in the prevalence of obesity may be attributable to environmental and psychological factors that are present in our ever-changing societies (Berthoud, 2012; Berthoud and Morrison, 2008).

Stress exerts significant effects on both the energy homeostasis system and the brain regions involved in reward processing such as the mesocorticolimbic dopamine circuit, resulting in changes in feeding behavior and food intake. Human studies have suggested a close relationship between altered eating behavior, obesity, with stress in life (Adam and Epel, 2007; Block et al., 2009; Brunner et al., 2007; Dallman, 2010; Dallman et al., 2005; Kyrou et al., 2006; Lemmens et al., 2011; Maniam and Morris, 2012; Nieuwenhuizen and Rutters, 2008; Sinha and Jastreboff, 2013; Torres and Nowson, 2007). Although people eat less under conditions of extreme and life-threatening stress as a component of “flight or fight” responses, exposure to moderately stressful situations that are increasingly encountered in modern society increases food intake and especially the consumption of foods that have a high content of sugar and fat, even in cases in which the individuals are not hungry (Berthoud, 2004; Born et al., 2010; Epel et al., 2001; Lemmens et al., 2011; Oliver et al., 2000; Rutters et al., 2009; Zellner et al., 2006). Conversely, reports show that the successful maintenance of weight loss is associated with a lower level of stress and better ability to cope with stress (Elfhag and Rossner, 2005; Sarlio-Lahteenkorva et al., 2000). Altogether, results from population-based studies and clinical laboratory studies indicate a significant role for stressful life events in the increasing prevalence of overeating and obesity in developed countries over the past three decades (Berghofer et al., 2008; Caballero, 2007; Flegal et al., 2010).

In animal studies, however, the effects of exposure to stressors on food intake have been inconsistent and depend on the accessibility to food and types of food provided. For example, stress-challenged rodents consumed less ordinary foods (Calvez et al., 2011; Dal-Zotto et al., 2004; Harris et al., 1998; Marti et al., 1994; Rybkin et al., 1997; Solomon et al., 2007; Valles et al., 2000) and exhibited an increase in the intake of energy-dense and palatable foods (Foster et al., 2009; Kinzig et al., 2008; la Fleur et al., 2005). These animal studies were performed using procedures that allowed the subjects to have ad libitum access to food in their home cages. Interestingly, in recent studies that used operant conditioning paradigms, stress consistently reinstated food-seeking responses without food availability during the test sessions in rats with a history of self-administering high-fat foods (Calu et al., 2013; Cifani et al., 2012; Ghitza et al., 2006; Ghitza et al., 2007; Le et al., 2011; Nair et al., 2009; Nair et al., 2008; Nair et al., 2006; Nair et al., 2011; Pickens et al., 2012; Richards et al., 2008). Based on this line of research, stress exposure is hypothesized to increase animals’ motivation for procuring food rewards, especially foods that have high energy contents.

To test this hypothesis, the present study examined the effects of a stress challenge on lever-pressing for food rewards. Several features were embedded in the experimental design. First, a progressive-ratio (PR) schedule of reinforcement was used to measure the motivation for the delivery of rewards (Markou et al., 1993). The PR schedule provided information on the effort that the rats were willingly to expend to earn the delivery of food pellets. Second, an emphasis was placed on comparisons between standard lab chow and high-fat food reward. Third, the animals were satiated in terms of food consumption and energy balance because they had ad libitum access to standard lab chow in their home cages. This allowed examination of the animals’ motivation for gaining excessive food rewards beyond the need to maintain energy homeostasis. Fourth, a pharmacological stressor, yohimbine, was administered before the test sessions. Yohimbine is an α₂ adrenergic receptor antagonist that increases the activity of noradrenergic systems, including the neural structures implicated in stress responses (Abercrombie et al., 1988; Aghajanian and VanderMaelen, 1982; Chopin et al., 1986; Uhde et al., 1984). Yohimbine also produces anxiety- and stress-like states in humans and laboratory animals (Bremner et al., 1996a; b; Charney et al., 1983; Davis et al., 1979; Holmberg and Gershon, 1961; Lang and Gershon, 1963). Therefore, yohimbine has been increasingly used as a stressor in experimental studies, especially in the field of drug addiction research, including our own work (Cippitelli et al., 2010; Feltenstein and See, 2006; Funk et al., 2006; Ghitza et al., 2006; Kupferschmidt et al., 2009; Le et al., 2005; Lee et al., 2004; Liu, 2010; 2012; Nair et al., 2006; Schroeder et al., 2003; Shepard et al., 2004; Zarrindast et al., 2000). Importantly, the behavioral effects of yohimbine-induced stress appear to be more robust than the stress elicited by classically used footshock (Bossert et al., 2005; Le and Shaham, 2002; Lu et al., 2003; Shaham et al., 2000).

In using yohimbine as a pharmacological stressor, it should be acknowledged that central noradrenergic neurotransmission has long been known to participate in the regulation of feeding behavior and food intake (Grossman, 1960; Leibowitz, 1986; Wellman, 2005). Interestingly, however, the role of norepinephrine in food intake seems to be inconclusive and mounting literature has documented norepinephrine to both stimulate and suppress eating behavior, which might be attributable to the activation of distinct adrenoceptor subtypes (Bello et al., 2013; dos Santos et al., 2009; Jackson et al., 1997; Le et al., 2011; Swiergiel and Wieczorek, 2008; Zislis et al., 2007). Moreover, the corticotropin-releasing factor (CRF) systems in both the hypothalamic-pituitary-adrenal (HPA) axis and extrahypothalamic regions such as the extended amygdala circuits participate in the automomic and emotional and behavioral responses to stresses. This study further examined involvement of CRF in mediating yohimbine effects on motivation for food reward by blockade of CRF receptors (CRF₁). Specifically, a CRF₁ antagonist NBI (5-chloro-N-[cyclopropylmethyl]-2-methyl-N-propyl-N’-[2,4,6-trichlorophenyl]-4,6-pyrimidinediamine hydrochloride) was administered prior to yohimbine challenge in rats that responded under the PR schedule for deliveries of either standard or high-fat food pellets.

Materials and methods

Results

Food self-administration: comparison between standard food pellets and high-fat food pellets

The rats successfully learned to press a lever for food self-administration in the daily 1-h sessions under a PR schedule of reinforcement. As shown in Fig. 1a, the rats that worked for the delivery of high-fat food pellets emitted a greater number of lever-press responses compared with their counterparts that responded for standard food pellets. The two-way repeated-measures ANOVA showed significant effects of food pellet type (F_1,14 = 5.15, p < 0.05) and session (F_11,154 = 3.53, p < 0.001) and a significant pellet × session interaction (F_11,154 = 2.32, p < 0.05). Rats that responded for high-fat food pellets exhibited a significantly higher break point for active lever responses (F_1,14 = 5.19, p < 0.05; Fig. 1b) and earned more pellets (F_1,14 = 4.67, p < 0.05; Fig. 1c) compared with the rats that responded for standard food pellets. In contrast, no difference was found in the number of responses on the inactive lever between the rats that self-administered standard vs. high-fat food pellets (data not shown). As shown in Fig. 1d, the rats in these two groups had similar levels of weight gain (F_1,14 = 5.19, p > 0.05).

Open in a separate window

Figure 1

Profiles of active lever responses (a) and breaking points (b), food pellets earned (c), and body weights (d) in rats (n = 8 in each group) that had free access to standard laboratory chow in their home cages and received daily 1-h sessions for food self-administration under a PR schedule of reinforcement.

Effect of yohimbine on the motivation to self-administer food: comparison between standard food pellets and high-fat food pellets

As shown in Fig. 2.a, the pharmacological stressor yohimbine significantly increased responses on the active lever for food self-administration. The two-way repeated-measures ANOVA of active lever responses revealed significant effects of food pellet type (F_1,14 = 4.73, p < 0.05) and yohimbine dose (F_2,28 = 16.47, p < 0.0001) and a significant pellet × dose interaction (F_2,128 = 4.53, p < 0.05). The one-way repeated-measures ANOVA and subsequent Newman-Keuls post hoc test revealed a significant difference (p < 0.01) in the number of active lever responses after pretreatment with 1 and 2.5 mg/kg yohimbine compared with saline controls in rats that self-administered standard food. For the rats that self-administered high-fat food pellets, this statistical analysis revealed a significant difference in the number of active lever responses after yohimbine pretreatment at doses of 1 mg/kg (p < 0.05) and 2.5 mg/kg (p < 0.01) compared with saline controls and a significant difference between the 2.5 and 1 mg/kg yohimbine doses (p < 0.05). Similar differences in the break points for active lever responses and number of pellets self-administered after pretreatment with yohimbine were observed (Fig. 2c, d). However, responses on the inactive lever remained unchanged (Fig. 2b), with no effect of food pellet type (F_1,14 = 0.00, p > 0.05) or yohimbine dose (F_2,28 = 0.19, p > 0.05) and no pellet × dose interaction (F_2,128 = 1.52, p > 0.05).

Open in a separate window

Figure 2

Effects of yohimbine challenge on active lever responses (a) and breaking points (c), food pellets earned (d), and inactive lever responses (b) in rats that under a PR schedule of reinforcement responded for deliveries of standard (n = 8) or high-fat (n=8) food pellets. The data are expressed as mean±SEM. *p < 0.05, **p < 0.01, significant difference from vehicle (0); ⁺p < 0.05, significant difference from the low dose (1 mg/kg).

Effect of NBI on yohimbine-induced enhancement of the motivation for food reward

Pretreatment with the CRF₁ antagonist NBI significantly reduced yohimbine-induced enhancement of the motivation to self-administer food pellets in the rats that had been trained with either the high-fat (Fig. 3a) or the standard (Fig. 3b) food pellets. In the animals on high-fat food pellets, a one-way repeated-measures ANOVA of the number of active lever responses revealed a significant effect of NBI dose (F_2,18 = 14.06, p < 0.001). The subsequent Newman-Keuls post hoc test verified a significant difference (p < 0.01) between 10 mg/kg NBI treatment and both saline and 5 mg/kg NBI treatments. In the animals on standard food pellets, similar statistical analyses demonstrated a significant NBI dose effect (F_2,18 = 5.13, p < 0.05) with a significant difference (p < 0.05) of 10 mg/kg NBI treatment versus saline control (Fig. 3b).

Open in a separate window

Figure 3

Effects of pretreatment with NBI on yohimbine-enhanced lever responses in rats that self-administered either high-fat (a) or standard (b) food pellets under a PR schedule of reinforcement. Panel c shows the number of responses on the active lever after NBI treatment without subsequent yohimbine challenge. The data are expressed as mean±SEM. *p < 0.05, **p < 0.01, significant difference from yohimbine+saline condition.

As shown in Fig. 3c, NBI pretreatment without subsequent yohimbine challenge did not change lever responses in the rats trained with either standard or high-fat food pellets.

In these drug treatment tests, however, responses on the inactive lever were unaltered.

Discussion

The present study examined the issue of whether stress alters rats’ motivation for food reward, with an emphasis on differences between the motivation for standard chow pellets vs. food pellets that have a high-fat content. The results showed three major findings. First, the food-satiated rats willingly emitted lever-press responses under a PR schedule of reinforcement for deliveries of food pellets. Moreover, the rats showed a significantly higher level of responding for the high-fat food pellets than for the standard chow pellets, indicating a stronger motivation for high-fat food even under food satiation conditions. Second, a pharmacological stress challenge, induced by administration of yohimbine, substantially enhanced lever-press responses for the delivery of food pellets. Particularly significant was the observation that the facilitating effect of stress was more robust in rats that responded for high-fat food pellets than in rats that responded for standard chow pellets. Third, pretreatment with the CRF₁ receptor antagonist NBI attenuated the facilitating effect of yohimbine stress on the animals’ lever-press responses, indicating the involvement of CRF₁ receptor activation in the stress-enhanced motivation for food reward.

Although numerous studies have investigated feeding behavior and weight changes within the framework of energy homeostasis, little is known about the psychological factors that contribute to overeating and obesity in subjects without a homeostatic need. Overeating and obese subjects procure and ingest foods beyond the need to maintain energy homeostasis. To closely simulate human overeating, the present study examined operant food self-administration in rats that were given free access to standard laboratory chow in their home cages; therefore, these animals were satiated in terms of food and energy intake. To properly interpret the lever-press response data, two speculations need to be addressed. First, in these rats, the high-fat food pellets might have caused general emotional arousal or elevated activity. The results showed that although the rats that responded for high-fat food pellets emitted significantly more responses on the active lever, they responded on the inactive lever at a very low level, similar to animals that responded for standard food pellets. Therefore, the rats emitted a higher number of lever-press responses specifically for deliveries of the high-fat food pellets rather than responded as a result of generally enhanced arousal or activity. Second, the significantly higher level of motivation for the high-fat pellets may have been attributable to the stimulation of multiple sensory receptors in the oral cavity by the lipid content of the high-fat pellets and not their high caloric value. Although this study did not directly address this issue, a recent study reported that mice readily emitted operant (lick) responses for direct intragastric infusions of lipid emulsions (Tellez et al., 2013), indicating that the stimulation of oral sensory receptors was not a critical factor for the self-administration of high-fat foods in rodents. Thus, oral sensations might have contributed little, if anything, to the self-administration responses observed in the present study.

The pharmacological stress induced by pre-session administration of yohimbine substantially enhanced lever-press responses for the delivery of food reward, regardless of the pellet type. The observation that yohimbine challenge did not change inactive lever responses ruled out the possibility of nonspecific activation of locomotor activity or operant behavior. These findings indicate that exposure to stress enhanced the rats’ motivation for earning food pellets, although these animals were satiated in terms of food intake and energy balance. With regard to understanding the facilitating effects of yohimbine on lever responses, each delivery of food pellets was associated with the presentation of auditory/visual stimuli that consisted of a 5 s tone and illumination of a light above the active lever for 20 s. Because of the repeated association with food reward, the stimuli acquired motivational value and became food-conditioned cues. Yohimbine challenge may also enhance the conditioned incentive effects of the food cue, and such an effect may have at least partially contributed to the increase in lever responses. Such a possibility is supported by previous studies. Shaham and colleagues used response-reinstatement procedures that have been widely employed in research on the reinstatement of drug-seeking behavior (Bossert et al., 2013; Epstein et al., 2006; Shaham et al., 2003). They demonstrated that acute stress induced by yohimbine reliably elicited operant food-seeking responses. The test sessions had response-contingent presentations of food-associated cues without the availability of food rewards (Calu et al., 2013; Cifani et al., 2012; Ghitza et al., 2006; Le et al., 2011; Nair et al., 2009; Nair et al., 2008; Nair et al., 2006; Nair et al., 2011; Pickens et al., 2012).

Notably, the yohimbine-induced enhancement of lever responses was significantly stronger in the rats that responded for high-fat food pellets compared with their counterparts that responded for standard pellets. These findings are consistent with observations in human subjects. For example, under controlled laboratory conditions, acute physical or emotional distress caused subjects to consume food with high-fat and sugar content, even when these individuals were not hungry and had no homeostatic need for calories (Epel et al., 2001; Foster et al., 2009; Lemmens et al., 2011; Oliver et al., 2000; Rutters et al., 2009). Overeating behavior is dissociable from individual physiological states of hunger and satiety but attributable to a substantial degree to exposure to environmental stimuli, such as stressful life events (Carels et al., 2004; Carels et al., 2001; Nishitani and Sakakibara, 2006; Nishitani et al., 2009; Oliver et al., 2001; Ozier et al., 2008; Schachter, 1968; Scott and Johnstone, 2012; Torres and Nowson, 2007). This lends support to earlier work suggesting that stress increased the consumption of and preference for nutrient-dense foods, particularly foods that have high-fat and sugar content (Kaplan and Kaplan, 1957). External and internal stimuli, such as stressful life events and changes in emotional states (either positive or negative), enhance the motivation for high-energy foods, overriding the basic homeostatic control of energy balance (Figlewicz, 2003; Gibson, 2006; Kelley et al., 2005; Zheng et al., 2009). In fact, the increase in stressful experiences in individuals in modern society has been proposed to promote eating foods that have high-fat and sugar content (Lattimore and Maxwell, 2004; O'Connor et al., 2008; Wallis and Hetherington, 2009).

Given that central noradrenergic neurotransmission is involved in the regulation of feeding behavior and food intake (Grossman, 1960; Leibowitz, 1986; Wellman, 2005) and that yohimbine via blocking presynaptic noradrenergic autoreceptors leads to activation of the noradrenergic system, it is reasonable to speculate that the enhancing effect of yohimbine on food reinforced responding might be a result of directly targeting the noradrenergic regulation of feeding behavior rather than a consequence of its interactions with stress-specific mechanisms. However, rat food intake in their home cages within 5 hours (yohimbine half-life is 7-8 hours) after yohimbine test sessions remained unchanged at a group level (data not shown). There was no consistent trend of either increase or decrease in food intake. Majority of animals in the yohimbine treated groups ate chow similar to saline vehicle rats with a couple rats consuming a little more while other a couple a little less chow. This observation is in line with literature showing the inconsistent effect of noradrenergic manipulations on food intake (Bello et al., 2013; dos Santos et al., 2009; Jackson et al., 1997; Le et al., 2011; Swiergiel and Wieczorek, 2008; Zislis et al., 2007). In operant procedures similar to that used in this study, neither activation nor blockade of postsynaptic noradrenergic receptors interfered with lever responses for the reinforcement with food pellets (Le et al., 2011) while a α2-adrenergic receptor agonist clonidine was found slightly decreased food pellet-supported lever responses (Zislis et al., 2007). Taken together, it is suggested that the facilitating effect of yohimbine challenge on lever responses in the present study is attributable to a unique role of yohimbine serving as a pharmacological stressor and thereby the stress-specific mechanisms for enhancing motivation (see below).

Previous animal research on eating behavior and weight changes was typically performed using home cage free-feeding procedures. Studies that use these types of eating paradigms to examine the effects of stress exposure on eating and weight gain have yielded inconsistent results. For example, some studies reported the inhibition of food intake and lower weight gain after exposure to stress (Bhatnagar et al., 2006; Calvez et al., 2011; Diane et al., 2008; Harris et al., 1998; Levin et al., 2000; Pecoraro et al., 2004; Ricart-Jane et al., 2002; Rybkin et al., 1997), which might be related to the effects of glucocorticoids, such as mobilizing energy stores and increasing hepatic gluconeogenesis (Dallman et al., 2005; Torres and Nowson, 2007). Other studies, in contrast, reported the overeating of foods, especially when foods with higher caloric value and palatability were provided, such as sweetened condensed milk (Bertiere et al., 1984; Marti et al., 1994; Miyasaka et al., 2005; Rowland and Antelman, 1976; Wallach et al., 1977). The procedures used in the present study, which measured operant responses for food rewards under a PR schedule of reinforcement and thereby examined the subjects’ motivation for procuring foods (Markou et al., 1993), and home-cage feeding paradigms appear to investigate different physiological processes that govern food intake and maintain energy balance. Such a distinction is supported by work that showed that although manipulations of mesocorticolimbic dopamine system function altered operant responses for food rewards, food intake in free-feeding regimens remained intact (Baldo and Kelley, 2007; Berridge, 2009; Salamone et al., 1994; Wirtshafter and Stratford, 2010).

In the present study, pretreatment with the CRF₁ antagonist NBI prior to yohimbine challenge significantly attenuated lever responses in the rats that had been trained with either the standard or high-fat food pellets. That raises speculations that blockade of CRF₁ receptors might directly reduce the lever responding for pellet reward rather than the yohimbine-enhanced motivation or that the suppressant effect of NBI might be due to the impairment of locomotor activity by the drug. However, such possibilities can be ruled out because NBI did not change lever responses in the tests where rats received NBI without a subsequent yohimbine stress challenge (a vehicle administration instead) and given that NBI at 10 mg/kg did not significantly alter food intake in 4 hours after administration (Giardino et al., 2013). Together, these data demonstrated that NIB pretreatment suppressed specifically the lever responses that were facilitated by yohimbine challenge rather than the pellet-supported responding itself. Moreover, such a suppressant effect of NIB on yohimbine-enhanced responding was observed regardless of the types of the food pellets that rats responded for. Because the levels of lever responses enhanced by yohimbine challenge were substantially different between the two types of pellets (208 ± 48 vs. 509 ± 118), the effect of NBI was independent of the levels of lever responses. Thus, the pharmacological blockade of CRF₁ neurotransmission specifically reduced the yohimbine stress-enhanced motivation for food pellet reward.

The findings with NBI pretreatment in this study are in line with evidence generated in the drug addiction research field. For example, numerous studies have demonstrated the necessity of activation of CRF₁ receptors in the expression of the stress-induced reinstatement of drug-seeking behavior (Bruijnzeel et al., 2009; Erb et al., 1998; Le et al., 2000; Liu and Weiss, 2002; Shaham et al., 1997; Zislis et al., 2007). More directly related to the present data is the finding that systemic antagonism of CRF₁ receptors by antalarmin, a CRF₁ antagonist, effectively suppressed the reinstatement of food-seeking responses induced by yohimbine (Ghitza et al., 2006). CRF is distributed in both the hypothalamus (as part of the HPA axis) and reward circuits, especially the extended amygdala, including the central nucleus of the amygdala, bed nucleus of the stria terminalis, and a transition area in the shell of the nucleus accumbens. Extrahypothalamic CRF plays an important role in behavioral responses to stress (Koob, 2009; Koob et al., 1993). Studies have demonstrated that exposure to stress elevated the level of CRF in the ventral tegmental area, and elevations of CRF facilitated dopamine neuronal activity, which was blocked by CRF₁ antagonists (Rodaros et al., 2007; Wanat et al., 2008). Furthermore, extrahypothalamic CRF rather than HPA axis CRF plays a unique role in stress-induced reward-seeking behavior (Erb et al., 1998; Graf et al., 2011; Le et al., 2000). The present data, together with recent evidence of the overlap of neurobiological substrates for food- and drug-taking/seeking behavior (Kenny, 2011; Lutter and Nestler, 2009; Sinha and Jastreboff, 2013; Volkow et al., 2013 for example reviews), indicate that neurotransmission via the extrahypothalamic CRF system may play a role in heightened desire for food reward and overeating behavior in response to stressful life events, thereby contributing to the epidemic of obesity in an ever-modernizing society.

Acknowledgements

References