Regulatory interactions of stress and reward on rat forebrain opioidergic and GABAergic circuitry

Abstract

Palatable food intake reduces stress responses, suggesting that individuals may consume such “comfort” food as self-medication for stress relief. The mechanism by which palatable foods provide stress relief is not known, but likely lies at the intersection of forebrain reward and stress regulatory circuits. Forebrain opioidergic and gamma-aminobutyric acid ergic signaling is critical for both reward and stress regulation, suggesting that these systems are prime candidates for mediating stress relief by palatable foods. Thus, the present study (1) determines how palatable “comfort” food alters stress-induced changes in the mRNA expression of inhibitory neurotransmitters in reward and stress neurocircuitry and (2) identifies candidate brain regions that may underlie comfort food-mediated stress reduction. We used a model of palatable “snacking” in combination with a model of chronic variable stress followed by in situ hybridization to determine forebrain levels of pro-opioid and glutamic acid decarboxylase (GAD) mRNA. The data identify regions within the extended amygdala, striatum, and hypothalamus as potential regions for mediating hypothalamic–pituitary–adrenal axis buffering following palatable snacking. Specifically, palatable snacking alone decreased pro-enkephalin-A (ENK) mRNA expression in the anterior bed nucleus of the stria terminalis (BST) and the nucleus accumbens, and decreased GAD65 mRNA in the posterior BST. Chronic stress alone increased ENK mRNA in the hypothalamus, nucleus accumbens, amygdala, and hippocampus; increased dynorphin mRNA in the nucleus accumbens; increased GAD65 mRNA in the anterior hypothalamus and BST; and decreased GAD65 mRNA in the dorsal hypothalamus. Importantly, palatable food intake prevented stress-induced gene expression changes in subregions of the hypothalamus, BST, and nucleus accumbens. Overall, these data suggest that complex interactions exist between brain reward and stress pathways and that palatable snacking can mitigate many of the neurochemical alterations induced by chronic stress.

Keywords::

• Dynorphin
• enkephalin
• GAD65/67
• hypothalamus
• nucleus accumbens
• sucrose

Introduction

Stress, defined as a real or perceived threat to homeostasis, evokes a number of physiological responses. Responses to stress include hormone release from the hypothalamic–pituitary–adrenal (HPA) axis, activation of the autonomic nervous system, and changes in behavior (Dallman 1993; Strack et al. 1995, 1997; Herman and Cullinan 1997; Denver 2009; Ulrich-Lai and Herman 2009). These stress response systems are managed, in large part, by the forebrain (Herman and Cullinan 1997; Herman et al. 2003, 2005). In particular, forebrain regions important in stress modulation often employ inhibitory neurotransmitters such as gamma-aminobutyric acid (GABA) and opioids (Bowers et al. 1998; Lucas et al. 2007; Poulin et al. 2009). While these stress responses can be protective during bouts of acute stress, repeated stimulation of these systems by chronic stress can be maladaptive, with implications for a number of serious health problems, including depression, anxiety disorders, and obesity (Tsigos and Chrousos 2002; Simon et al. 2006; Scott et al. 2008). For example, chronic stress elicits facilitation of HPA responses, increases autonomic tone, and increases anxiety-like and depression-like behavior (Herman et al. 1995; Paskitti et al. 2000; Ostrander et al. 2006). In addition, chronic stress alters activity in stress regulatory brain regions (Herman and Cullinan 1997; Herman et al. 2003, 2005). For example, chronic variable stress (CVS) increases expression of the GABA-synthesizing enzyme glutamic acid decarboxylase (GAD65/67) mRNA (Bowers et al. 1998) and pre-proenkephalin-A (ENK) and pre-prodynorphin (DYN) mRNA (Dumont et al. 2000; Chen et al. 2004) in key stress regulatory regions. This suggests that these neurotransmitter systems may be important for adaptation to chronic stress.

“Comfort food” is a colloquial term used to denote highly palatable, calorically dense foods sought out to provide relief from stress. Individuals appear to engage in consumption of palatable food to limit HPA, autonomic, and behavioral responses to stress (Dallman 1993; Strack et al. 1995, 1997; Prasad and Prasad 1996; Markus et al. 2000; la Fleur et al. 2005; Schiltz et al. 2007; Ulrich-Lai et al. 2007). Like chronic stress, repeated palatable food intake can elicit changes in HPA function and brain regulation in forebrain regions important in processing emotionally salient information (Prasad and Prasad 1996; Strack et al. 1997; Markus et al. 2000; Kelley et al. 2003; Schiltz et al. 2007; Ulrich-Lai et al. 2007). Opioidergic and GABAergic circuitry regulate brain regions activated by palatable food intake (Will et al. 2003; Ikemoto and Wise 2004; Kelley et al. 2005), and palatable food can modulate GABA and opioid brain signaling (Kelley et al. 2003; Schiltz et al. 2007).

Forebrain opioidergic and GABAergic signaling are ideal candidates for mediating the stress-dampening properties of palatable foods. Thus, the goal of the present study was to determine how chronic palatable “snacking” affects CVS-induced changes in opioid and GAD gene expression in stress and reward brain regions. We hypothesized that the rewarding and/or metabolic effects resulting from regular access to palatable comfort food would reverse the effects of chronic stress on GAD, DYN, and ENK mRNA expression in brain regions important in stress and/or reward regulation. To test this hypothesis, the present study used a model of palatable “snacking” in combination with a model of CVS to assess stress–reward interaction in the regulation of opioid and GAD mRNA expression in forebrain circuitry.

Methods

Animals and experimental design

To directly compare neuroanatomical findings with the hormonal stress response and to conserve animal usage, the brain samples analyzed in this study were from adult, male Long-Evans rats used in experiments reported by Ulrich-Lai et al. (2007). All protocols were approved by the University of Cincinnati Institutional Animal Care and Use Committee and were consistent with NIH guidelines. Rats were given ad libitum access to water and chow (LM-485 mouse/rat sterilizable diet; Harlan-Teklad, Madison, WI, USA) for the duration of the study. Rats were given additional intermittent (up to 30 min) access twice daily to 4 ml sucrose (30% Sigma Aldrich Co., St Louis, MO, USA), sodium saccharin (0.1%, Sigma Aldrich Co.), or water at 09:30 and 15:30 h for 28 days. Saccharin, a noncaloric sweetener, was included to determine whether palatable food effects were due to sweet taste alone (saccharin) vs. sweet taste with calories. After 2 weeks of intermittent drink treatment, half of each group of rats was concurrently exposed to 14 days of chronic stress (for a total of 6 treatment groups). The chronic stress paradigm used was CVS, a well-characterized chronic stress protocol (Herman et al. 1995; Paskitti et al. 2000; Ostrander et al. 2006). The stressors have all been shown to promote activation of the stress axis without substantial injury, morbidity, or mortality. CVS consisted of twice daily exposure (at unpredictable times, spaced a minimum of 4 h apart) to one of several different stressors in an unpredictable order. Stressors included: 20 min hypoxia (8% oxygen, 92% nitrogen), 20 min warm (26–30°C) swim, 10 min cold (17–18°C) swim, 1 h in cold (4°C) room (housed two per cage with no bedding), 5 min novel environment, and 1 h in a cage atop an orbital platform shaker (90 rpm). In addition to the twice daily stressors, CVS rats were individually housed overnight in large (guinea pig) cages on experiment days 17, 19, 22, and 25 and in small (mouse) cages on experiment days 16, 21, 24, and 27. Approximately 15 h after the last stress exposure, rats did not receive their respective drink solution and all rats, both control and CVS, were given a novel restraint stress challenge. Rats were placed into well-ventilated restraint tubes (∼21 cm in length, with ∼5 cm opening) for 20 min. Rats were killed by rapid decapitation without anesthesia at 60 min after the initiation of restraint. Brains were removed and snap frozen in isopentane on dry ice, then stored at − 80°C. Brains were sectioned coronally at a thickness of 14 μm on a cryostat (Microm, Waldorf, Germany), thaw mounted onto slides (Gold Seal Ultrastick Slides, Portsmouth, NH, USA), and stored at − 20°C until being pre-treated for in situ hybridization.

As shown previously, the rats that did not receive chronic stress drank sucrose and saccharin in amounts approaching the maximum permitted, whereas rats receiving chronic stress showed markedly reduced saccharin intake (Figure 1) (Ulrich-Lai et al. 2007). Additionally, the chronic stress treatment groups demonstrated thymic involution, a decrease in body weight gain, and the appropriate hormonal responses indicative of a chronic stress condition (Ulrich-Lai et al. 2007).

Figure 1.  Time course of mean daily drink intake of (a) sucrose or saccharin solution, or water (8 ml/day max) in control and (b) CVS rats receiving twice daily access to these drink solutions in addition to ad libitum food and water. Abscissa numbers are days. Palatable drink intake was reduced during chronic stress. All values for sucrose and saccharin intake were significantly different from water intake (p < 0.05). ^#p < 0.05 vs. nonCVS control; ^†p < 0.05 vs. the previous timepoint by post hoc analysis. Abbreviations: CVS, chronic variable stress. Graph reproduced from Ulrich-Lai et al. (2007), Copyright 2007 The Endocrine Society.

In situ hybridization

A one-in-ten series of brain sections was fixed in phosphate-buffered paraformaldehyde (4%) for 10 min, then was rinsed twice for 5 min in diethylpyrocarbonate (DEPC)-treated 5 mM potassium phosphate-buffered saline (KPBS), followed by two 5 min rinses in KPBS plus 0.2% glycine, and an additional two 5-min rinses in KPBS. The sections were then treated with 0.25% acetic anhydride (in 0.1 M triethanolamine, pH8) for 10 min and rinsed twice in 2 × saline-sodium citrate (SSC) for 5 min before being dehydrated in a graded alcohol series.

Antisense cRNA probes for DYN (Curran and Watson 1995), ENK (Curran and Watson 1995), GAD65 (Bowers et al. 1998), and GAD67 (Bowers et al. 1998) mRNAs were in vitro transcribed using ³⁵S-UTP. The transcription reactions consisted of the following: 1 × transcription buffer, 62.5 μCi ³⁵S-UTP, 330 μM ATP, 330 μM GTP, 330 μM CTP, 10 μM cold UTP, 66.6 mM dithiothreitol (DTT), 40 U ribonuclease inhibitor, 20 U of the appropriate RNA polymerase (T3 or T7), and 2.5 μg of the appropriate linearized DNA. The reaction was incubated for 1 h at 37°C, then was DNAse treated, and had tRNA added. Finally, the labeled probe was separated from the free nucleotide by ammonium acetate precipitation.

³⁵S-UTP-labeled probes were re-suspended in DEPC water and diluted in hybridization buffer (50% formamide, 20 mM Tris–HCl pH 7.5, 1 mM EDTA, 335 mM NaCl, 1 × Denhardt's solution, 200 μg/ml herring sperm DNA, 100 μg/ml yeast tRNA, 20 mM DTT, and 10% dextran sulfate) to generate a solution of 100,000 cpm per 50 μl of buffer. Each slide was treated with 50 μl of diluted probe, coverslipped, and incubated overnight at 55°C in chambers humidified with 50% formamide. The next day, cover slips were carefully removed in 2 × SSC and slides were rinsed in 2 × SSC for 5 min. Slides were transferred into 100 μg/ml ribonuclease A solution and incubated at 37°C for 30 min. Slides were rinsed twice in 2 × SSC for 5 min, three times in 0.2 × SSC for 10 min, and once in 0.2 × SSC at 65°C for 1 h. Slides were dehydrated through a graded alcohol series and exposed to Kodak Biomax MR-2 film. Sense probes and RNase-treated tissue were run for all probes as negative controls. On each radiograph, an ARC 146-14C standard slide (American Radiolabeled Chemicals, Inc., St Louis, MO, USA) was included as an internal control to verify that the film exposure was constant between films and was not saturated.

Due to the complexity of the experimental design, all rats in this study received an acute restraint stress 1 h prior to tissue collection. Basal levels of ENK, DYN, GAD65, and GAD67 mRNAs are rather high in all brain regions examined, and show either no increase or a slight increase at 60 min after an acute stimulus (Bowers et al. 1998; Dumont et al. 2000; Wang et al. 2003; Lucas et al. 2007). The authors acknowledge that the rats in the present study were not naïve; however, these studies indicate that the mRNA levels assessed in the present study were predominantly reflective of expression prior to the acute stress challenge, thereby representing primarily the effects of the chronic palatable drink and stress treatments.

Image analysis

Semi-quantitative gray-scale densitometric analysis of the X-ray autoradiographs was performed using the Scion Image 1.62 software (Scion, Frederick, MD, USA). Anatomical regions of interest were targeted on the basis of known involvement in stress and reward brain circuits, and were delineated on the basis of the Swanson (1998) and the Paxinos and Watson (1986) rat brain atlases. Because recent data indicate differential involvement of subareas of many brain regions in stress or reward (e.g. bed nucleus of the stria terminalis (BST) and nucleus accumbens), areas were parcellated into known subdivisions on the basis of anatomical landmarks defined in the atlases. A region was required to have more than twice the background to be included for analysis. Labeled mRNA was measured in the amygdala (central (CeA), anterior medial (anterior MeA), posterior medial (posterior MeA), and basolateral nucleus (BLA) regions); the NAc (medial shell (NAc medial shell), lateral shell (NAc lateral shell), and core (NAc core) subregions); dorsal striatum; dentate gyrus of hippocampus (DG); hypothalamus (dorsomedial nucleus (DMH), ventrolateral part of the ventromedial nucleus, medial preoptic area (mPOA), anterior hypothalamic nucleus (AHN), lateral hypothalamic area (LHA), and paraventricular hypothalamic nucleus (PVN) subregions); the zona incerta (ZI); and the bed nucleus of the stria terminalis (anterodorsal subregion (BSTad), anteroventral subregion (BSTav), posteriordorsal subregion (BSTpd), and posterioventral subregion (BSTpv)). Background signal was determined from a nonhybridized portion of the tissue in each section, and then subtracted from the total gray level to give corrected gray level. The mean corrected gray level was determined bilaterally from two to four sections per region. All image analysis was done blind to treatment and the “n” for each group ranged from five to nine rats depending on variations in tissue integrity and availability.

Statistical analysis

Data are presented as mean corrected gray level ± SE of the mean (SEM). ENK, DYN, GAD65, and GAD67 mRNA expression were analyzed as described by Ulrich-Lai et al. (2007). Briefly, data from each brain region were analyzed by two-way ANOVA. If there were group differences, Fisher's least significant comparisons procedure was used to determine specific planned pairwise comparisons; no further adjustments were made to control for the experiment-wise error rate. Outliers were removed only if they differed from the mean by more than 1.96 times the SD and they were outside the lower or upper quartiles by more than 1.5 times the interquartile range (McClave and Dietrich 1994). Statistical significance is reported at p < 0.05.

Results

ENK mRNA expression

There was extensive ENK mRNA labeling in regions of the hypothalamus, striatum, hippocampus, and amygdala as reported previously (Harlan et al. 1987) (see Figure 2a). Sucrose drink reduced ENK mRNA labeling in the anterior BST (Figure 3, BSTad (CVS × drink interaction F_(2,48) = 3.3, p = 0.048); BSTav (main effect of drink F_(2,45) = 3.4, p = 0.044, CVS × drink interaction F_(2,47) = 3.8, p = 0.031)) and the lateral shell of the NAc (Table I, main effect of CVS F_(1,45) = 10.8, p = 0.002; main effect of drink F_(2,45) = 3.4, p = 0.044), and these effects were prevented by chronic stress (Figure 3, Table I, p < 0.05). In contrast, chronic stress induced an overall increase in ENK mRNA in numerous brain regions (Table I, NAc core (main effect of CVS F_(1,42) = 24.2, p = 0.0001), CeA (main effect of CVS F_(1,48) = 4.7, p = 0.036), PVN (main effect of CVS F_(1,44) = 4.5, p = 0.04), and DG (main effect of CVS F_(1,47) = 7.5, p = 0.009)). Post hoc analysis revealed that for the NAc lateral shell, CeA, DG, and ventromedial hypothalamic nucleus (VMH), the stress-induced increase in labeled ENK mRNA occurred primarily in groups with sucrose or saccharin drink (p < 0.05). Conversely, palatable drink (sucrose and saccharin) blunted the effects of chronic stress in the DMH (Figure 4, main effect of CVS F_(1,45) = 6.5, p = 0.014; main effect of drink F_(2,45) = 3.8, p = 0.031; CVS × drink interaction F_(2,45) = 3.7, p = 0.033), and NAc medial shell (Table I, main effect of CVS F_(1,42) = 16.1, p = 0.0003). ENK mRNA labeling was not affected by palatable snacking or chronic stress in the other examined brain regions (Table I).

Figure 2.  Representative images and regions analyzed for ENK (a), DYN (b), GAD65 (c), and GAD67 (d) mRNA expression by in situ hybridization. Abbreviations: anterior hypothalamic nucleus, AHN; basolateral amygdala, BLA; the bed nucleus of the stria terminalis, BST; (anterodorsal (adBST), anteroventral (avBST), posterior dorsal (pdBST), posterior ventral (pvBST), central amygdala (CeA), dentate gyrus (DG), dorsal striatum (DSt), dorsomedial hypothalamic nucleus (DMH), lateral hypothalamic area (LHA), medial amygdala (MeA), medial preoptic area (mPOA), nucleus accumbens NAc (medial shell (NAcM) and lateral shell (NAcL), and core (NAcC)), paraventricular hypothalamic nucleus, PVN; ventromedial hypothalamic nucleus, VMH; and zona incerta, ZI. Images for each probe are displayed rostral to caudal from left to right with the most rostral image at the top left.

Figure 3.  ENK mRNA in the anterior BST. Sucrose drinking reduced ENK mRNA expression in the anterior BST and this reduction was prevented by CVS in both (a) the anterodorsal area and (b) the anteroventral area. *Indicates significant difference (p < 0.05) from nonCVS control and ^#indicates significant difference (p < 0.05) from water drinking; post hoc analysis. Data are shown as mean ± SEM with n = 5–9 per group.

Table I.  Forebrain expression of ENK mRNA (corrected gray levels).

	Control			CVS
	Water	Saccharin	Sucrose	Water	Saccharin	Sucrose
mPOA	38.66 ± 3.0	39.3 ± 2.5	36.0 ± 2.2	40.6 ± 3.4	39.6 ± 2.4	41.8 ± 3.1
AHN	84.6 ± 5.8	90.5 ± 3.6	84.7 ± 2.5	91.9 ± 4.4	87.4 ± 4.1	92.8 ± 5.2
Posterior BST	31.0 ± 1.1	31.2 ± 2.8	27.8 ± 2.4	31.5 ± 3.0	29.7 ± 2.5	31.0 ± 2.9
Dorsal striatum	114.9 ± 2.5	123.5 ± 4.1	115.6 ± 3.3	118.1 ± 4.3	120.2 ± 3.1	118.8 ± 2.3
Anterior MeA	79.3 ± 3.2	80.2 ± 3.8	72.2 ± 2.3	78.0 ± 6.3	76.1 ± 3.1	82.5 ± 3.5
Posterior MeA	64.4 ± 3.4	61.45 ± 1.4	57.9 ± 4.0	61.7 ± 4.8	58.9 ± 2.2	64.2 ± 4.6
BLA	55.8 ± 2.4	53.9 ± 2.0	56.2 ± 3.1	60.0 ± 2.9	60.5 ± 3.8	58.5 ± 3.9
NAc Core*	102.4 ± 2.5	101.4 ± 2.1	97.8 ± 4.8	112.5 ± 3.0^†	116.4 ± 2.4^†	110.7 ± 2.8^†
NAc Med shell*	95.1 ± 4.9	95.9 ± 3.2	95.8 ± 1.8	106.1 ± 3.6^†	109.6 ± 3.0^†	105.3 ± 2.8
NAc Lat shell*^‡	118.1 ± 4.0	123.8 ± 5.0	108.2 ± 3.2^¶	126.1 ± 2.5	126.5 ± 1.0	124.7 ± 3.7^†
CeA*	125.6 ± 4.9	123.2 ± 5.5	117.4 ± 6.3	129.3 ± 3.3	130.3 ± 3.5	131.1 ± 2.7^†
PVN*	87.1 ± 2.2	83.2 ± 3.2	85.2 ± 4.7	91.0 ± 5.1	92.1 ± 4.2	96.7 ± 6.9
DG*	55.8 ± 2.4	53.9 ± 2	56.2 ± 3.1	60.0 ± 2.9	60.5 ± 3.8^†	58.5 ± 3.9
VMH^§	118.8 ± 2.5	125.4 ± 3.7	115.2 ± 4.2	123.9 ± 2.6	119.2 ± 1.8	124.2 ± 1.6^†

Additional forebrain regions expressing ENK mRNA. Data are shown as mean ± SEM with n = 5–9 per group. List of abbreviations: chronic variable stress, CVS; medial preoptic area, mPOA; anterior hypothalamic nucleus, AHN; posterior bed nucleus of the stria terminalis, posterior BST; anterior medial amygdala, anterior MeA; posterior medial amygdala, posterior MeA; basolateral nucleus of the amygdala, BLA; nucleus accumbens core, NAc core; nucleus accumbens medial shell, NAc Med shell; nucleus accumbens lateral shell, NAc Lat shell; central nucleus of the amygdala, CeA; paraventricular nucleus of the hypothalamus, PVN; dentate gyrus of hippocampus, DG; and ventromedial hypothalamic nucleus, VMH. * indicates regions with a main effect of stress; ^† indicates a significant difference, p < 0.05, from nonCVS control; ^‡ indicates regions with a main effect of drink; ^¶ indicates a significant difference, p < 0.05, from water control by post hoc analysis; ^§ indicates regions with stress–drink interaction, p < 0.05.

Figure 4.  ENK mRNA in the dorsomedial hypothalamus (DMH). Chronic stress increased ENK mRNA level in the DMH and this increase was prevented by palatable drink. *Indicates significant difference (p < 0.05) from nonCVS control; and ^#indicates significant difference, p < 0.05, from water, post hoc analysis. Data are shown as mean ± SEM with n = 5–9 per group.

DYN mRNA expression

DYN mRNA was labeled in regions of the hypothalamus, striatum, hippocampus, and amygdala as reported previously (Morris et al. 1986; Sato et al. 1991) (Figure 2b). Chronic stress increased DYN mRNA labeling in the NAc core (main effect of CVS F_(1,44) = 8.2, p = 0.007), and this CVS-induced increase was prevented by sucrose or saccharin drink (Figure 5, p < 0.05). In contrast, palatable drink and chronic stress did not affect DYN mRNA labeling in any of the other examined brain regions (Table II).

Figure 5.  DYN mRNA in the nucleus accumbens (NAc) core. Chronic stress increased DYN mRNA expression in the NAc core and this increase was prevented by sucrose or saccharin drink. *Indicates significant difference, p < 0.05, from nonCVS control; and ^#indicates significant difference, p < 0.05, from water; post hoc analysis. Data are shown as mean ± SEM with n = 5–9 per group.

Table II.  Forebrain expression of pro-dynorphin (DYN) mRNA (corrected gray levels).

	Control			CVS
	Water	Saccharin	Sucrose	Water	Saccharin	Sucrose
VMH	54.1 ± 1.7	66.0 ± 4.4	61.8 ± 3.3	60.4 ± 4.5	62.2 ± 3.7	62.2 ± 4.5
LHA	39.9 ± 1.5	40.0 ± 3.7	36.0 ± 1.8	37.1 ± 2.5	33.6 ± 2.3	40.0 ± 2.5
PVN	54.7 ± 2.7	57.0 ± 4.2	54.4 ± 2.6	54.1 ± 4.6	49.6 ± 1.6	55.5 ± 4.2
SON	102.3 ± 6.5	87.0 ± 6.5	92.9 ± 5.5	98.3 ± 6.7	98.7 ± 5.2	98.6 ± 6.3
NAc Med shell	38.8 ± 2.4	44.0 ± 3.1	43.1 ± 2.0	49.3 ± 2.0	44.8 ± 4.0	43.7 ± 3.0
NAc Lat shell	34.8 ± 4.3	38.8 ± 2.9	39.9 ± 3.2	45.9 ± 3.6	41.7 ± 3.5	40.3 ± 3.5
Dorsal striatum	24.5 ± 1.3	30.8 ± 2.1	26.3 ± 2.5	30.4 ± 2.4	28.0 ± 2.1	29.4 ± 2.8
DG	37.1 ± 5.0	41.5 ± 4.9	42.9 ± 2.2	43.8 ± 2.8	37.5 ± 2.8	42.9 ± 1.9
DMH	40.6 ± 5.9	53.2 ± 5.1	50.0 ± 4.4	53.3 ± 3.9	46.4 ± 3.0	50.6 ± 5.1
mCeA	51.0 ± 2.1	44.8 ± 3.9	43.2 ± 2.9	48.6 ± 2.6	47.5 ± 3.2	54.7 ± 3.3
adBST	23.0 ± 2.1	23.6 ± 3.2	26.2 ± 3.7	27.8 ± 2.0	26.9 ± 3.2	32.0 ± 4.1

Additional forebrain regions expressing DYN mRNA showed no significant differences with stress or access to palatable drink. Data are shown as mean ± SEM with n = 5–9 per group. List of abbreviations: chronic variable stress, CVS; ventromedial hypothalamic nucleus, VMH; lateral hypothalamic area, LHA; paraventricular nucleus of the hypothalamus, PVN; supraoptic nucleus, SON; nucleus accumbens medial shell, NAc Med shell; nucleus accumbens Lat shell, NAc lateral shell; dentate gyrus of hippocampus, DG; dorsomedial hypothalamus, DMH; medial central nucleus of the amygdala, mCeA; and bed nucleus of the stria terminalis anterodorsal, BSTad.

GAD65 mRNA expression

The pattern of GAD65 mRNA labeling observed in hypothalamic, hippocampal, and amygdalar subregions (Figure 2c) matched expression patterns reported previously (Bowers et al. 1998). Palatable drink caused a reduction in labeled GAD65 mRNA levels in the BSTpd that was reversed by chronic stress (Table III, main effect of CVS F_(1,44) = 7.7, p = 0.009; main effect of drink F_(2,44) = 3.7, p = 0.033). In contrast, chronic stress differentially altered GAD65 mRNA levels in a region-specific manner. Chronic stress decreased GAD65 mRNA in the DMH and this effect was prevented by palatable drink (Figure 6, main effect of CVS F_(1,46) = 10.8, p = 0.002). Conversely, chronic stress increased labeled GAD65 mRNA in the AHN (CVS × drink interaction F_(2,47) = 4.2, p = 0.023), BSTpd (main effect of CVS F_(1,44) = 7.7, p = 0.009; main effect of drink F_(2,44) = 3.7, p = 0.0326), and BSTpv (main effect of CVS F_(1,47) = 4.2, p = 0.047), with palatable drink preventing these effects in the AHN (Figure 7, Table III, p < 0.05). GAD65 mRNA labeling was not affected by palatable snacking or chronic stress in the other examined brain regions (Table III).

Table III.  Forebrain expression of GAD65 mRNA (corrected gray levels).

	Control			CVS
	Water	Saccharin	Sucrose	Water	Saccharin	Sucrose
mPOA	38.6 ± 3.3	45.0 ± 2.9	40.2 ± 1.3	47.5 ± 3.7	40.9 ± 3.1	42.3 ± 3.4
ZI	36.3 ± 2.9	39.5 ± 1.0	33.2 ± 2.1	30.9 ± 2.1	35.7 ± 1.9	35.5 ± 2.6
BSTpd*^†	43.1 ± 1.5	42.0 ± 2.0	36.9 ± 1.9^‡	48.2 ± 2.2	43.9 ± 1.6	43.5 ± 2.9^¶
BSTpv*	34.0 ± 1.8	34.6 ± 1.9	31.0 ± 1.7	36.8 ± 1.9	34.7 ± 1.6	38.4 ± 3.4^¶
BSTad	26.0 ± 1.2	26.8 ± 1.6	26.8 ± 1.6	26.0 ± 1.6	24.3 ± 1.1	26.1 ± 1.2
BSTav	28.1 ± 1.5	28.3 ± 1.4	27.6 ± 2.3	29.3 ± 1.2	27.3 ± 1.3	31.0 ± 1.1
NAc core	21.6 ± 1.2	24.1 ± 1.1	22.1 ± 1.4	22.3 ± 1.6	22.7 ± 1.5	23.1 ± 1.8
NAc shell	27.2 ± 1.7	31.3 ± 1.4	29.8 ± 2.6	30.5 ± 2.5	30.0 ± 1.9	31.0 ± 2.5
DG	25.2 ± 1.7	20.5 ± 1.4	21.4 ± 2.6	23.5 ± 1.2	23.1 ± 1.9	23.2 ± 1.1

Additional forebrain regions expressing GAD65 mRNA. Data are shown as mean ± SEM with n = 5–9 per group. List of abbreviations: chronic variable stress, CVS; medial preoptic area, mPOA; zona incerta, ZI; bed nucleus of the stria terminalis posteriordorsal subregion, BSTpd; bed nucleus of the stria terminalis posterioventral subregion, BSTpv; bed nucleus of the stria terminalis anterodorsal subregion, BSTad; bed nucleus of the stria terminalis anteroventral subregion, BSTav; nucleus accumbens core, NAc core; nucleus accumbens shell, NAc shell; and dentate gyrus of hippocampus, DG.; * indicates regions with a main effect of stress; ^† Indicates regions with a main effect of drink, p < 0.05; ^‡ Indicates a significant difference, p < 0.05, from water control by post hoc analysis; ^¶ Indicates a significant difference, p < 0.05, from nonCVS control.

Figure 6.  GAD65 mRNA in the DMH. Chronic stress decreased GAD65 mRNA level in the DMH and this effect was prevented by palatable drink. *Indicates significant difference, p < 0.05, from nonCVS control and ^#indicates significant difference, p < 0.05, from water; post hoc analysis. Data are shown as mean ± SEM with n = 5–9 per group.

Figure 7.  GAD65 mRNA in the AHN. Chronic stress increased GAD65 mRNA level in the AHN and palatable drink prevented these effects. *Indicates significant difference (p < 0.05) from nonCVS control and ^#indicates significant difference (p < 0.05) from water; post hoc analysis. Data are shown as mean ± SEM with n = 5–9 per group.

GAD67 mRNA expression

GAD67 mRNA labeling occurred in limited hypothalamic, hippocampal, and amygdalar subregions (Figure 2d) which matched expression patterns reported previously (Bowers et al. 1998). Neither palatable drink nor CVS altered the labeling of GAD67 in any of the examined brain regions (Table IV).

Table IV.  Forebrain expression of GAD67 mRNA (corrected gray levels).

	Control			CVS
	Water	Saccharin	Sucrose	Water	Saccharin	Sucrose
mPOA	40.0 ± 5.9	38.3 ± 3.8	38.5 ± 7.7	36.8 ± 5.9	38.0 ± 5.4	42.2 ± 5.2
AHN	28.8 ± 1.5	32.9 ± 2.7	33.2 ± 1.8	29.6 ± 2.5	28.4 ± 3.4	25.2 ± 1.6
DMH	27.9 ± 2.8	35.6 ± 3.8	28.0 ± 1.8	30.5 ± 4.1	30.4 ± 1.3	27.9 ± 1.8
DG	23.9 ± 1..3	23.3 ± 2.2	23.3 ± 1.4	19.7 ± 2.0	29.4 ± 4.3	25.4 ± 2.3
ZI	42.7 ± 2.9	38.6 ± 3.5	40.3 ± 4.7	41.2 ± 4.1	35.6 ± 3.8	38.7 ± 2.2
BST	38.7 ± 4.5	35.5 ± 3.9	29.7 ± 3.8	21.3 ± 5.8	28.9 ± 2.1	39.6 ± 3.7

There were no significant differences in GAD67 mRNA expression after palatable drink access or stress. Data are shown as mean ± SEM with n = 5–9 per group. List of abbreviations: chronic variable stress, CVS; medial preoptic area, mPOA; anterior hypothalamic nucleus, AHN; dorsomedial nucleus of the hypothalamus, DMH; dentate gyrus of hippocampus, DG; the zona incerta, ZI; and bed nucleus of the stria terminalis, BST.

Discussion

The primary aim of the present study was to assess changes in GABA-producing enzyme and opioid gene transcription in specific reward and stress-regulatory brain regions following comfort food intake in rats with or without a history of chronic stress. Importantly, these data enabled identification of potential neurochemical mechanisms for comfort food-mediated stress buffering.

Palatable food

Palatable food intake, on its own, altered mRNA expression of inhibitory neurotransmitters in the NAc and the BST. Sucrose decreased ENK mRNA in the NAc lateral shell, in agreement with previous data (Kelley et al. 2003). Previous studies link increased ENK and DYN transcription with increased neuronal signaling (Giraud et al. 1991; Simpson and McGinty 1995; Isola et al. 2008). The present data indicating decreased activity in enkephalinergic neurons in the NAc are commensurate with palatable food-induced neuroadaptation in reward brain circuitry (Kelley et al. 2003).

Sucrose solution drinking decreased GAD65 mRNA in the dorsal portion of the posterior BST. GAD mRNA levels are regulated by GABA agonists in an inverse manner, in that high levels of GABA reduce GAD expression and activity (Rimvall and Martin 1994; Sheikh and Martin 1998; Raol et al. 2005; Martyniuk et al. 2007). Taken with the present results, this suggests that sucrose intake increases GABA release in the BSTpd. The BST is an important integrator of limbic stress input (Choi et al. 2007), and the BSTpd is considered to be stress inhibitory, particularly to acute stress responses (Choi et al. 2007, 2008). Collectively, the data indicate that sucrose-dampening of HPA responses to acute stress (Ulrich-Lai et al. 2007) may be mediated, at least in part, via altered GABAergic tone in the posterior BST.

In some cases, gene expression effects were more pronounced with sucrose than saccharin intake, indicating that regulatory changes may be related to either metabolic and/or postingestional properties of sucrose or an enhanced hedonic value of the sucrose solution relative to saccharin (Sclafani and Abrams 1986) (e.g. calories themselves can increase reward value (Lucas and Sclafani 1989; Ackroff et al. 2010)).

Chronic stress

Chronic stress altered expression of inhibitory neurotransmitters in numerous stress and reward brain regions. Specifically, chronic stress increased ENK mRNA level in the DMH, NAc shell, CeA, PVN, and DG; increased DYN mRNA in the NAc core; increased GAD65 mRNA in the AHN and BST; and decreased GAD65 mRNA in the DMH. Chronic stress increased ENK mRNA in the hippocampus and the PVN as described previously (Dumont et al. 2000; Chen et al. 2004). Glucocorticoids can drive ENK mRNA expression in the forebrain (Ahima et al. 1992), suggesting stress-induced signaling in the DMH, NAc shell, CeA, PVN, and DG. Chronic stress increased DYN mRNA expression in the NAc core, agreeing with previous results and indicating increased signaling in dynorphinergic neurons (Isola et al. 2008). These data indicate that opioidergic neuropeptides may be important in the organization of stress responses (Dumont et al. 2000; Palkovits 2000; Isola et al. 2008). Chronic stress-induced increases in GAD65 mRNA expression in the BST and AHN agree with previous studies (Bowers et al. 1998). However, unlike previous studies (Bowers et al. 1998), GAD65 mRNA was decreased in the DMH. This discrepancy may be due to differences in the strain of rats or duration of stress exposure. Overall, the changes in GAD65 mRNA indicate that chronic stress may have decreased GABAergic tone in the AHN and BST, and increased it in the DMH.

Importantly, gene expression (mRNA level) changes induced by chronic stress were prevented by comfort food intake in the DMH (ENK and GAD65), AHN (GAD65), BSTpd (GAD65), and the NAc core (DYN). The DMH is a known stress regulatory region, and DMH efferents directly target the PVN (Thompson et al. 1996; DiMicco et al. 2002). The DMH mediates changes in HPA axis and autonomic responses to stress (DiMicco et al. 2002) and is held under tonic inhibition by GABAergic neurotransmission (Brandao et al. 1986; Milani and Graeff 1987, Canteras 2002; DiMicco et al. 2002). In the DMH, chronic stress decreased GAD65 mRNA expression, which may indicate a chronic stress-induced increase in GABAergic tone (Rimvall and Martin 1994; Sheikh and Martin 1998; Raol et al. 2005; Martyniuk et al. 2007). In contrast, ENK mRNA was increased in the DMH. Increases in ENK mRNA are generally considered to underlie a compensatory increase in synthesis that occurs after periods of high neuronal activity and high peptide release (Giraud et al. 1991; Simpson and McGinty 1995; Isola et al. 2008), indicating that chronic stress may increase activity of both GABA and ENK neurons in the DMH. Conversely, in the AHN, chronic stress increases GAD65 mRNA level, consistent with a decrease in GABAergic tone in AHN neurons. Like the DMH, the AHN is also a central component of defensive response circuitry. However, neurons in the AHN respond similarly to appetitive and aversive stimuli, indicating a role in attentional tuning to emotionally salient encounters (Ono and Nakamura 1985). The ability of palatable drink intake to reverse the changes in gene expression in the DMH and AHN, brain regions critical for the HPA, behavioral, and autonomic responses to stress, implicates these regions as likely mediators for comfort food-induced stress buffering.

Sucrose intake decreased GAD65 mRNA expression in the BSTpd in the absence of chronic stress. However, during chronic stress GAD65 mRNA expression levels of rats receiving sucrose were not different from the water controls. The posterior BST is stress inhibitory, particularly in response to acute stressors (Choi et al. 2007, 2008). This indicates that chronic stress overcomes the effects of the palatable drink on gene expression, possibly decreasing GABAergic tone and increasing GAD65 mRNA expression to levels similar to levels in controls drinking water.

Chronic stress also increased DYN mRNA expression in the NAc core in water-drinking rats, but not in rats drinking sucrose or saccharin solution, indicating an increase in NAc activity during chronic stress that is prevented by palatable drink. Stress activates the NAc (Perrotti et al. 2004) and dynorphin-dependent opioid receptor phosphorylation increases in the NAc after stress (Land et al. 2008). In addition, the dynorphin system generally promotes anxiety-like and dysphoric behaviors (Bruchas et al. 2010). Together, the data indicate that the NAc may be another key region in palatable comfort food-mediated stress reduction.

Although palatable drink did not prevent CVS-induced facilitation of the hormonal stress response in these rats (Ulrich-Lai et al. 2007), a history of sucrose decreased the hormonal response to acute stress regardless of chronic stress history, suggesting that altered inhibitory neurotransmission in striatal and hypothalamic brain regions may mediate acute stress-dampening by comfort food. In the chronically stressed rats, the diminished stress buffering by the saccharin drink may be due, at least in part, to the decrease in saccharin intake after the onset of the CVS paradigm (Ulrich-Lai et al. 2007). However, 2 weeks of intake prior to CVS, together with a small intake during CVS, is sufficient to induce changes in opioid and GAD mRNA expression in many of the same stress and reward brain regions as sucrose.

Perspectives

The regulation of stress and reward pathways in the brain is extremely complex, therefore requiring thorough examination of stress and reward brain regions, which, not surprisingly, reveals multiple brain regions affected by chronic stress or palatable snacking. Importantly, there is substantial overlap in the stress and reward pathways and the present study demonstrates that palatable “snacks” can prevent neurochemical alterations induced by chronic stress (summary diagram: Figure 8). These data indicate that engagement of forebrain reward circuits may play a role in inhibiting (buffering) responses of physiological or behavioral effector systems following chronic stress exposure.

Figure 8.  Summary diagram of regional neurochemical changes in response to palatable snacking and chronic stress. In a number of brain regions, shown in the overlapping circles, neurochemical changes (expression of ENK, GAD65, and DYN mRNAs) induced by CVS were prevented by palatable snacking, thus identifying these regions as potential mediators of comfort food-induced stress buffering. Abbreviations: bed nucleus of the stria terminalis, BST; nucleus accumbens, NAc; dentate gyrus of hippocampus, DG; dorsomedial hypothalamus, DMH; anterior hypothalamic nucleus, AHN; paraventricular nucleus of the hypothalamus, PVN; and central amygdala, CeA.

In summary, the present data identify the BST, NAc, DMH, and AHN as brain reward and stress regions that may mediate HPA-dampening following chronic snacking, and suggest that there are clear interactions between brain reward and stress pathways that serve to determine the eventual response to physical or psychological adversity. These findings are consistent with the hypothesis that palatable food induced alterations in stress-inhibitory neurocircuitry result in a decreased stress response. These observations are in agreement with the “comfort food hypothesis”, suggesting engagement of inhibitory signaling mechanisms in stress regulatory circuits. Understanding the endogenous neural mechanisms by which comfort food buffers the stress response could help improve strategies for the prevention and/or treatment of stress-related disorders and obesity.

Acknowledgements

The authors would like to thank Amanda Jones and Ben Packard for their technical assistance and Colin Kennard for his assistance in preparing the figures. This work was supported by DK059803 (A.M.C.), MH069725 (J.P.H.), MH049698 (J.P.H.), DK067820 (Y.M.U.), and DK078906 (Y.M.U).

Declaration of interest:The authors have no conflict of interest. The authors alone are responsible for the content and writing of the paper.