Animal Models

Adaptations in brain reward circuitry underlie palatable food cravings and anxiety induced by high-fat diet withdrawal



To identify the emotional and motivational processes that reinstate palatable food intake following removal of high-fat diet (HFD) and associated neuroadaptations tied to neurochemical and behavioural changes underlying dopaminergic function.


Adult male C57Bl6 mice were placed on a HFD (58% kcal fat) or ingredient-matched, low-fat diet (LFD; 11% kcal fat) for 6 weeks. At the end of diet-regimen mice were either maintained on their respective diets, or HFD and LFD were replaced with normal chow (withdrawal). Effort-based operant responding for sucrose and high-fat food rewards was measured along with basal and stress-induced corticosterone levels and anxiety (elevated-plus maze). Protein levels for tyrosine hydroxylase (TH), corticosterone releasing factor type 1 receptor (CRF-R1), brain-derived neurotrophic factor (BDNF), phospho-CREB (pCREB) and ΔFosB (truncated splice variant of FosB) were assessed in the amygdala, nucleus accumbens (NAc) and ventral tegmental area (VTA) via western immunoblotting.


Six weeks of HFD resulting in significant weight gain elicited sucrose anhedonia, anxiety-like behaviour and hypothalamic-pituitary-adrenocortical axis (HPA) hypersensitivity to stress. Withdrawal from HFD but not LFD-potentiated anxiety and basal corticosterone levels and enhanced motivation for sucrose and high-fat food rewards. Chronic high-fat feeding reduced CRF-R1 and increased BDNF and pCREB protein levels in the amygdala and reduced TH and increased ΔFosB protein in NAc and VTA. Heightened palatable food reward in mice withdrawn from HFD coincided with increased BDNF protein levels in NAc and decreased TH and pCREB expression in the amygdala.


Anhedonia, anxiety and sensitivity to stressors develops during the course of HFD and may have a key role in a vicious cycle that perpetuates high-fat feeding and the development of obesity. Removal of HFD enhances stress responses and heightens vulnerability for palatable foods by increasing food-motivated behaviour. Lasting changes in dopamine and plasticity-related signals in reward circuitry may promote negative emotional states, overeating and palatable food relapse.


Overconsumption of calorically-dense food contributes to increasing rates of obesity and associated comorbidities including mood disorders. Palatable high-fat and high-sugar foods are rewarding and their intake can recruit neural pathways and mechanisms linked to addiction. Although the concept of food addiction has gained much attention recently, whether or not clinical features of addiction can be applied to feeding is still being debated.1 Explaining the failure of many weight-loss programs, dieting is often accompanied by heightened desire for high-fat and -sugar foods and reverting to unhealthy eating habits. High-fat diet (HFD) withdrawal can elicit palatable food cravings that along with reexposure to these foods, associated cues and/or stressful experiences can cause relapse.2 Physical, psychosocial and pharmacological stressors can stimulate or reinstate palatable food seeking2, 3, 4 and clinical reports show that food addicts use food to self-medicate; they often eat in order to escape negative emotional states.5

The neural circuitry underlying reward, emotion and addictive processes includes mesolimbic dopamine (DA) neurons originating in the ventral tegmental area (VTA) of the midbrain that innervate limbic sites including the nucleus accumbens (NAc) and amygdala.6 The VTA, NAc and amygdala are posited to form an interconnected network important for reward processing,7 while the amygdala also has a fundamental role in the control of anxiety.8, 9 Repeated intake of drugs and natural rewards leads to adaptations in DA neurons and their targets that escalate intake and can promote compulsive behaviour.10, 11 Important neural adaptations within mesolimbic reward-associated areas include increase in transcriptional and neurotrophic signals such as phospho-CREB (pCREB), ΔFosB (truncated splice variant of FosB) and brain-derived neurotrophic factor (BDNF). Previous studies report that chronic HFD consumption reduces DA tone and increases BDNF, pCREB and ΔFosB protein levels in midbrain and limbic reward sites.12, 13, 14 We recently found that chronic intake of HFD-induced obesity (DIO) enhances hypothalamic-pituitary-adrenocortical axis (HPA) stress responsiveness and promotes anxiety- and depressive-like behaviour along with reductions in tyrosine hydroxylase (TH), the rate-limiting enzyme in DA biosynthesis, and increased BDNF, pCREB and ΔFosB protein levels in the NAc.15 It is nonetheless important to determine whether or not such changes are secondary to deleterious metabolic effects of obesity including hypercorticosteronemia.

Little is known about the evolution of cravings and motivation for palatable foods following withdrawal from chronic HFD and the reward-relevant neurobehavioural changes involved. Cottone et al.16, 17, 18 found that withdrawal from extended, intermittent access to palatable food increases anxiety and decreases effort-based responding for less palatable chow, an effect linked to elevated corticotrophin-releasing factor receptor-1 (CRF-R1) signalling in the amygdala.19 HFD removal is reported to increase operant responding for sucrose rewards in obesity-prone but not obesity-resistant rats,20 however, it is not known if withdrawal from chronic high-fat feeding increases anxiety and palatable food reward, independent of genetic factors that enhance susceptibility to weight gain and which neurobiological changes are involved.

In the present study we tested the hypothesis that HFD withdrawal enhances effort-based responding for sucrose and high-fat foods, increases behavioural and biochemical anxiety measures and modulates dopamine-, stress- and plasticity-related proteins in components of mesolimbic reward circuitry. To dissociate the effects of chronic HFD from metabolic disturbances caused by severe DIO, we investigated the impact of HFD for 6 weeks. Accordingly, a second objective was to test the hypothesis that chronic high-fat feeding and weight-gain before the development of severe obesity produces anxiety, anhedonia and associated neuroadaptations in brain reward circuitry.


Animals and diets

Experiments were performed on 90 adult male C57Bl/6 mice received at 6−7 weeks of age from Jackson Laboratory (Bar Harbour, Maine, USA). All mice were maintained in environmentally controlled rooms (22–24 °C) with ad libitum access to standard chow and water. Mice were pseudo-randomly assigned to one of two diet groups such that mean-body weights were initially equal between groups. One group received a HFD (HFD; D12231; Research Diets, Inc., New Brunswick, NJ, USA) containing 58% kcal from fat; 16.4% kcal from protein and 25.5% kcal from carbohydrates. To control the novelty of the diet the other group of mice were placed on a control, ingredient-matched low-fat diet (LFD; D12328; Research Diets, Inc., New Brunswick, NJ, USA) containing 10.5% kcal from fat; 16.4% kcal from protein and 73.1% kcal from carbohydrates. Three separate cohorts of mice (Groups 1, 2 and 3) were used to investigate the effect of diet and diet withdrawal on various behavioural and biochemical measures. All procedures involving the use of animals were approved by the CRCHUM Animal Care Committee.

Diet withdrawal

Following 6 weeks on either HFD or LFD all mice were transferred to new cages. Mice from the withdrawal groups were given ad libitum access to standard chow, whereas mice from the non-withdrawal group were maintained on their respective diets.

Food-motivated operant responding

We used a well-validated measure of food reward known as the progressive ratio (PR) operant task for which response requirements for food escalate over the course of the session. An increase in PR responding is inferred as increased motivation for food.21 Experiments were performed in mouse operant chambers (Med Associates Inc., St. Albans, VT, USA) as described previously.22 Briefly, chambers were equipped with two ultra-light response levers and a food receptacle. Only one lever was designated as ‘active’ (triggering food reward) and the allocation of right and left levers were counterbalanced between mice. Responding on the active lever was reinforced by the delivery of a 20-mg sucrose pellet or a high-fat pellet containing 48.9% kcal as fat (Bio-Serv, Frenchtown, NJ, USA). Mice (Group 1, N=12) were trained in a PR schedule where response demands increased according to r=(5 × e0.2n) −5, rounded to the nearest integer and with n as the position in the sequence of ratios (1, 2, 4, 9, 12, 15, 20, etc.).21 The last ratio completed in the PR session is considered as the breakpoint. After obtaining basal breakpoints in the PR, task mice were divided pseudo-randomly to either the HFD or LFD group with the added constraint that there basal breakpoints were not different during baseline. During the fifth week of the diet-regimen, mice were introduced back into the operant chambers for retraining. After the 6-week diet period all mice were subjected to diet withdrawal and tested once a day for 5 days following withdrawal; the first 3 days with sucrose and the last 2 days using high-fat rewards. Caloric intake and body weight were measured at each test day.

Elevated plus maze (EPM)

One day following diet withdrawal (or maintenance), mice from Group 2 (N=30) were tested in the EPM, the most commonly used rodent model of anxiety,23 as described previously.15 Each experiment was recorded and tracked for 5 min using an overhead video camera and software (Ethovision XT, Med Associates Inc., St. Albans, VT, USA).

Basal and stress-induced corticosterone measures

Basal corticosterone was measured in mice from Group 3 (counterbalanced across conditions) on blood collected during sacrifice. The day following the EPM test each mouse from Group 2 (counterbalanced across conditions) was restrained for 30 min in decapicones and tail-blood samples were collected. Plasma corticosterone was measured by an enzyme-linked immunosorbent assay corticosterone kit (Enzo Life Sciences, Farmingdale, NY, USA).

Western immunoblotting

Mice (Group 3; N=48) were sacrificed under isoflurane anaesthesia 3 days following the diet withdrawal manipulation. Brains were rapidly dissected and frozen in isopentane. Frozen brains were sliced into 200 μm coronal sections using a cryostat and mounted onto slides maintained on dry ice. Nuclei were microdissected using brain tissue punches (Stoelting, Inc., Kiel, WI, USA). As shown in Figure 1, bilateral tissue punches of 0.75 mm diameter were obtained from the VTA and 1.0 mm diameter punches from the NAc and basolateral amygdala and central nucleus of amygdala. Samples were homogenized on ice in 100 μl of cell lysis buffer (20 mM Tris, pH 7.5; 150 mM NaCl; 1 mM Na2EDTA; 1 mM EGTA; 1% Triton; 2.5 mM sodium pyrophosphate; 1 mM β-glycerophosphate; 1 mM Na3VO4; 1 μg ml−1 leupeptin) with added protease (phenylmethylsulphonyl fluoride 100 μM) and phosphatase inhibitors (Sigma phosphatase inhibitor cocktails I and II) using a motorized pestle. Homogenates were centrifuged for 15 min at 14 000 Hg. Protein samples (20 μg) were separated by electrophoresis on a 10% polyacrylamide gel and electro-transferred to a polyvinylidene difluoride membrane (Millipore, Bedford, MA, USA). Membranes were blocked in tris-buffered saline with 5% low-fat milk and 0.1% Tween-20 or 5% bovine serum albumin. Membranes were rinsed in buffer (0.1% Tween-20 in tris-buffered saline) and then incubated with anti-BDNF, anti-CRF-R1 (1:500; Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-pCREB (Ser133), anti-Delta FosB, (1:1000; Cell Signalling Technology Inc, Danvers, MA, USA), anti-TH (1:1000; Millipore, Bedford, MA, USA), followed by antirabbit or antigoat or antimouse Immunoglobulin G horseradish peroxidase-conjugate (1:5 000). After rinsing with buffer, immunocomplexes were visualized by chemiluminescence using western lighting plus ECL (PerkinElmer, Waltham, MA, USA). Film signals were digitally scanned and then density quantified using ImageJ software (free software available from NIH at All protein data were normalized according to GAPDH control values.

Figure 1

Location of brain dissections for immunoblotting experiments. Illustration depicting brain-punch sites (light grey) collected bilaterally from the NAc, amygdala and VTA. BLA: basolateral amygdala; CeA: central nucleus of the amygdala; NAc-C: nucleus accumbens core; NAc-Sh: nucleus accumbens shell; VTA: ventral tegmental area.

Statistical analysis

Data were analysed using GraphPad Prism 5 software ( Data are presented as means±s.e. of the mean (s.e.m.). A two-way analysis of variance with Bonferonni post-tests was used to calculate data collected from the operant conditioning task, food intake and body weight following withdrawal, EPM test, plasma corticosterone and western blotting. Unpaired t-tests were used to compare food intake, body weight and sucrose anhedonia measures.


Food intake and body weight

Cumulative food intake was significantly increased by 5 weeks in mice consuming HFD as compared with control LFD mice (Figure 2a). By 6 weeks HFD mice had an 11.8% increase in body weight relative to LFD mice (Figure 2b).

Figure 2

Food intake and body weight. (a) Cumulative food intake of mice on LFD or HFD for 6 weeks. As compared with LFD mice, HFD mice consumed significantly more calories as of 5 weeks on diets. (b) HFD mice gained significantly more weight by week 4 as compared with LFD mice. Means±s.e.m. *P0.05, **P0.01, ***P0.001.

HFD-induced sucrose anhedonia and elevated food reward following HFD withdrawal

To measure motivation and craving for food we used a progressive-ratio task that assesses effort-based responding for food rewards. Mice fed HFD for 6 weeks displayed reduced sucrose intake and breakpoint response thresholds for sucrose relative to LFD mice (Figure 3a). Decreased motivation for sucrose is also known as sucrose anhedonia, a phenomenon linked to negative emotional states and reduced reward sensitivity.24 Withdrawal from the HFD increased breakpoint responding (Figure 3b) and total lever responses (Figure 3c) for sucrose by Day 3 of withdrawal, a finding not observed in LFD withdrawal mice. We next sought to determine if the increased motivation for sucrose also applied to high-fat food by using a reward similar in composition to the HFD on Day 4 and 5 postwithdrawal. Mice withdrawn from HFD showed a similar increase in food-motivated responding for high-fat food such that breakpoint thresholds (Figure 3b) and number of total lever responses (Figure 3c) were significantly increased. We did not observe any changes in the percentage of correct lever responses in HFD and LFD mice over the course of the testing (Figure 3d). Caloric intake was significantly increased on days 1, 2 and 3 postwithdrawal, (Figure 3e) whereas body weights of HFD mice remained higher than LFD control counterparts (Figure 3f).

Figure 3

Food-motivated operant behaviour. (a) Six weeks of HFD decreased response thresholds for sucrose relative to LFD mice in the PR operant task. (b) Withdrawal from HFD, but not LFD, significantly increased breakpoint thresholds for sucrose rewards as of day 3 of withdrawal (‘WD D3’). Breakpoints remained elevated on day 4 and 5 of withdrawal when sucrose rewards were replaced by a high-fat food rewards that were similar in macronutrient composition to the HFD used. (c) A similar increase in lever responses following HFD withdrawal was observed such that HFD mice responded more on days 3, 4 and 5 of withdrawal. (d) No differences were observed in the proportion of correct lever responses between HFD and LFD mice. (e) Caloric intake was significantly greater in HFD mice as compared with LFD mice on postwithdrawal days 1, 2 and 3. (f) Body weights remained significantly elevated for HFD mice as compared with LFD controls on each day postwithdrawal. Means±s.e.m. *P0.05, **P0.01, ***P0.001.

HFD and HFD withdrawal increases anxiety-like behaviour

There were significant main effects of HFD to decrease EPM open arm time (Figure 4a) and open arm entries, both indices of anxiety-like behaviour (Figure 4b). Withdrawal from the HFD decreased the proportion of time spent in the open arms by 45% relative to mice withdrawn from LFD (Figure 4a). As shown in Figures 4c and d, there was no difference in closed arm entries and distance travelled between HFD and LFD mice indicating similar locomotor activity. However, distance travelled was elevated by HFD withdrawal (Figure 4d).

Figure 4

Anxiogenic effects of HFD and its withdrawal. (a) There was a main effect of HFD to reduce the percentage of time spent in open arms of the EPM. In addition, open arm time was significantly reduced in mice withdrawn from HFD as compared with mice withdrawn from LFD. (b) There was a main effect of HFD to reduce the number of open arm entries; however, there was no difference between mice HFD and LFD withdrawal mice. As indices of locomotor activity, no differences were observed between HFD and LFD mice for (c) closed arm entries and (d) total distance travelled in the EPM. There was however, an increase in locomotor activity in mice withdrawn from HFD relative to mice maintained on HFD. Means±s.e.m. *P0.05, **P0.01.

HFD and HFD withdrawal increase physiological measures of stress

Mice in the HFD withdrawal group showed a small, but reliable decrease in body weight 24 h postwithdrawal that was not observed in the LFD group (Figure 5a). Decreased body weights following HFD withdrawal likely reflect increased stress and locomotor activity and are consistent with the behavioural and biochemical results we and others obtained.25 HFD did not elevate basal corticosterone levels, but withdrawal from HFD increased corticosterone levels relative to those obtained in LFD withdrawal mice (Figure 5b). To assess whether or not HPA stress reactivity is elevated by HFD and its withdrawal, we measured plasma corticosterone levels following 30-min restraint stress. There was a significant main effect of HFD but no effect of withdrawal to increase corticosterone levels (Figure 5c).

Figure 5

Physiological consequences of HFD withdrawal. (a) Body weights were significantly decreased 24 h postwithdrawal in mice subjected to HFD withdrawal. (b) Elevated basal plasma corticosterone in HFD withdrawal mice as compared with LFD withdrawal mice. (c) Restraint stress potentiated plasma corticosterone levels in HFD mice relative to control LFD mice. In addition, plasma corticosterone levels under stress conditions were elevated in HFD withdrawal mice as compared with LFD withdrawal mice. Means±s.e.m. *P0.05, **P0.01.

HFD and withdrawal from HFD alter dopamine & plasticity-related proteins in brain reward circuitry


To determine if HFD and its withdrawal give rise to adaptations in brain reward circuitry, we measured the expression of proteins relevant to mesolimbic DA function and plasticity at a time point following withdrawal that corresponded to the increased sucrose cravings (Day 3). We first focused our attention on the amygdala where DA and CRF-R1 signalling have a well-defined role in the aversive effects of the drug26, 27, 28, 29, 30 and food19 withdrawal. In addition to a main effect of withdrawal to decrease levels of TH, the rate-limiting enzyme for DA biosynthesis, we found a significant interaction between withdrawal and diet on amygdalar TH protein levels: HFD withdrawal decreased TH expression by 78%, which was significantly different that the 54% reduction observed in LFD withdrawal mice (Figure 6a) (t-test on delta-TH; P=0.04). There was a main effect of HFD to decrease CRF-R1 levels in the amygdala as compared with the LFD mice, but no effect of withdrawal. The expression of pCREB was increased in the amygdala by HFD, whereas HFD withdrawal reduced pCREB levels (Figure 6a). Finally, there was a near-significant increase in BDNF levels in HFD mice to LFD controls (P=0.06).

Figure 6

Plasticity- and stress-related protein changes in amygdala, NAc and VTA. (a) Amygdala. The protein levels of amygdalar TH, a key enzyme in DA metabolism, decreased after withdrawal. HFD-reduced CRF-R1 and increased BDNF and pCREB expression, whereas withdrawal from HFD significantly reduced the levels of pCREB. (b) NAc. HFD-reduced TH and increased ΔFosB levels in the NAc. Withdrawal from HFD increased BDNF and decreased ΔFosB expression in the NAc. (c) VTA. As in the NAc, HFD decreased TH and increased ΔFosB levels in the VTA, however withdrawal did not alter the expression on any of the measured proteins in the VTA. Means±s.e.m. *P0.05, **P0.01, ***P0.001.


In the NAc, TH levels were significantly reduced by HFD but unaltered by HFD withdrawal (Figure 6b). There was no effect of HFD on BDNF levels, however, withdrawal from HFD-increased NAc−BDNF expression (Figure 6b). ΔFosB protein in the NAc was increased by HFD, whereas withdrawal from HFD supressed NAc ΔFosB expression (Figure 6b). Decrease in TH and increase in BDNF and ΔFosB protein levels in the NAc are similar to what has been observed previously with DIO.15


Similar to the results obtained in the NAc, there was a main effect of HFD to decrease TH levels and increase ΔFosB levels in the VTA. Decreased VTA−TH and increased ΔFosB expression are consistent with previous observations in DIO mice (15).


Food cravings and negative emotions play a key role in the failure of many dieting and weight-loss programs. Chronic high-fat feeding is comparable to an addictive-like process such that rewarding, yet unhealthy eating habits can persist and even strengthen despite clear deleterious consequences. The present study sought to determine the impact of long-term high-fat feeding and its withdrawal on the evolution of food-motivated behaviour, negative mood states and molecular adaptations in mesolimbic reward circuitry. We found that chronic high-fat feeding and weight gain before the onset of severe obesity produces a hypo-affective state characterized by sucrose anhedonia, anxiety and increased HPA reactivity. Linked to these changes, we identified several molecular modifications in the amygdala, NAc and VTA similar to those observed in states of DIO12, 15 and chronic drug use.31, 32, 33 Withdrawing palatable HFD but not LFD heightened craving and motivation for sucrose and high-fat food and increased physiological and behavioural indices of anxiety. Accompanying enhanced food cravings and anxiety in mice withdrawn from HFD, were reduced TH and pCREB expression in amygdala and decreased ΔFosB and increased BDNF protein levels in the NAc. These biochemical changes are tied to lasting neurochemical and behavioural changes related to dopaminergic function and parallel outcomes obtained during chronic administration of drugs of abuse and their withdrawal.34, 35

We previously reported on the effects of DIO to produce anxiety, despair and neuroadaptations in brain reward circuitry.15 Here we found that a shorter period of high-fat feeding resulting in about half the weight gain observed in DIO mice in our previous study also elicits negative emotional states and hyper-reactivity to stressors, albeit to lesser degree. Motivation for sucrose was decreased following 6 weeks of high-fat feeding as was the amount of time spent in open arms of the EPM implicating anhedonia and anxiety-like behaviour, respectively. As in our DIO study, the negative emotional outcome of high-fat feeding in the present study may rely on increased weight gain and associated metabolic changes. However, unlike our DIO study basal corticosterone levels were not increased in mice fed HFD, rather corticosterone was elevated in HFD mice only following the restraint stress manipulation. Collectively, these findings suggest that negative mood and sensitivity to stressors develop over the course of HFD and may contribute to a vicious cycle that perpetuates high-fat feeding to promote weight gain and the development of obesity.

Stress and the presence or removal of rewarding stimuli can modulate motivation and affective behaviour.36, 37 Operant conditioning experiments permitted evaluation of the effects of HFD withdrawal on food-motivated behaviour and craving by determining changes in the amount of effort mice were willing to engage in for palatable food rewards. Mice subjected to HFD but not LFD withdrawal displayed increased effort-based responding for sucrose rewards by the third day following withdrawal. That food-motivated behaviour was not elevated during the first 2 days suggests the presence of an incubation period for sucrose craving following HFD withdrawal. Increased food cravings also translated to fatty food rewards, as mice withdrawn from HFD worked harder to obtain high-fat food pellets relative to mice withdrawn from the LFD. Importantly, as we implemented a control LFD different than chow we can rule out any effects of novelty on food reward, and similarities in the proportion of correct responses (active versus inactive levers) between prediet baseline, postdiet and withdrawal conditions excludes diet-induced changes in learning and memory as potential confounds. Heightened sucrose reward in mice withdrawn from HFD relative to LFD withdrawal mice was evident despite the fact that sucrose responding was attenuated before withdrawal in HFD mice. As described above, decreased sucrose intake is commonly used as a marker of anhedonia,24, 38, 39 thus we interpret decreased response thresholds of HFD mice as a reflection of a hypo-affective state. Taken together, these observations suggest that although there are susceptibility factors that contribute to food cravings following HFD withdrawal,20 removal of high-fat food following long-term intake is sufficient to enhance motivation and cravings for palatable foods.

In addition to increasing motivation for palatable food, HFD removal potentiated anxiety-like behaviour. Consistent with this data, Teegarden et al.25 found significant anxiety-like effects using an open field-test 24 h after HFD withdrawal. In addition, Cottone et al.19 reported that rats subjected to withdrawal from intermittent HFD display increased anxiety in the EPM task as early as 5–9 h postwithdrawal. Similarly, withdrawal from intermittent sugar exposure decreases time spent in the open arms of the EPM40 and can elicit classic symptoms of withdrawal including teeth chattering, forepaw tremor and head shakes.41 Our observation of elevated corticosterone levels in mice subjected to withdrawal from palatable HFD underlines the stress-provoking effects of HFD withdrawal. Moreover, we found that mice withdrawn from a palatable HFD showed a small, yet consistent drop in body weight, which is a reported physiological response to a heightened-stress-state associated with reductions in rewarding stimuli.25, 42 Several lines of evidence indicate that, increased stress and anxiety triggers relapse to drug43, 44, 45 and food seeking,46, 47 and thus we posit that the stressful effects of diet withdrawal contribute to the increased motivation for sucrose and fat we observed.

Chronic high-fat feeding and withdrawal produced several changes in DA and plasticity-related proteins in the amygdala and NAc, and to lesser extent in the VTA, that are tied to lasting neurochemical and behavioural changes related to dopaminergic function. First, HFD-reduced protein levels for the DA biosynthetic enzyme TH in the VTA and the NAc, a major projection site of VTA−DA neurons. This observation is consistent with several reports that obesity15, 48, 49 and high-fat feeding independent of obesity50 reduces mesolimbic DA tone. As there were no changes in TH in the amygdala, our findings suggest that decreases in DA biosynthesis by high-fat feeding are specific to the mesoaccumbens DA pathway. Decreases in DA signalling by chronic high-fat feeding are speculated to reduce reward sensitivity and thereby promote compensatory overeating of palatable foods.51, 52, 53, 54 Diet withdrawal did not affect mesoaccumbens TH levels but decreased TH expression in the amygdala and to a significantly greater extent in mice withdrawn from HFD. Interestingly, elevated anxiety following a stressor has been shown to reduce DA biosynthesis in the amygdala of rats.55 In view of the role of the amygdala in behavioural stress responses, decreased DA biosynthesis in the amygdala following withdrawal may have a role in the anxiogenic effects of HFD withdrawal. In opposition to TH levels, amygdalar CRF-R1 protein expression in the amygdala was reduced by HFD and unchanged by withdrawal. Reduced CRF-R1 levels in the amygdala may be a compensatory response to increased amygdalar CRF levels that are reportedly elevated by long-term high-fat feeding.56 Indeed, increased CRF mRNA and protein levels are associated with loss of CRF-R1 function.57, 58 Although amygdalar CRF-R1 is implicated in the anxiogenic effects of intermittent HFD withdrawal,19 we did not see any effect of withdrawal on CRF-R1 protein levels and speculate that this may be due to the later time point (Day 3) tissue was analyzed following withdrawal.

Chronic cocaine administration is known to upregulate pCREB in the NAc59 and amygdala60 and is associated with tolerance-like increase in cocaine self-administration (that is, reduced reward sensitivity). With this in mind, we speculate that the increased pCREB expression in the amygdala may contribute to HFD-induced sucrose anhedonia we observed. Consistent with such a possibility, amygdalar pCREB levels during HFD withdrawal decreased in a manner that coincided with enhanced food-motivated responding. Similar reductions in pCREB protein levels in the amygdala are reported during withdrawal from nicotine following prolonged exposure.61 These findings suggest that decreased CREB transcriptional activity in the amygdala may be an important adaptation to high-fat food withdrawal that orients behaviour towards restoring palatable food intake.

BDNF has an important role in behavioural plasticity and its signalling actions in the NAc are well-implicated in the modulation of reward.31, 62, 63 Unlike DIO conditions,15 the expression of BDNF in the NAc was not elevated by high-fat feeding whereas withdrawal from HFD-increased BDNF protein levels in NAc. Therefore, neurotrophic signalling in the NAc may contribute to heightened food cravings and/or anxiety triggered by HFD withdrawal. In agreement with previous reports,15, 25 ΔFosB protein levels in NAc and VTA were elevated by chronic high-fat feeding and, as a novel finding, were attenuated in NAc by HFD withdrawal. NAc ΔFosB levels are also positively associated with an antidepressant action and coping with stressful experiences64 and thus perhaps decreases in NAc−ΔFosB by HFD withdrawal, may underlie the anxiogenic effects of abstaining from HFD.

In conclusion, chronic high-fat feeding and subsequent withdrawal from HFD reflects human conditions, in which individuals attempt to replace a calorically-dense diet with healthier, low-fat and low-sugar food options, and therefore may provide a clinically-relevant model to explore the neurobiological mechanisms reinstating palatable food intake during diet regimens. The negative emotional outcomes elicited by chronic intake of high-fat food in the present study were milder than those we previously observed following longer periods of high-fat feeding leading to DIO. However, as in our previous study, the observed emotional impairments and biochemical changes by HFD, may depend on increase in body weight. Switching HFD to normal chow potentiated anxiety responses and HPA reactivity and gave rise to increased motivation for sucrose and high-fat food rewards following increase in NAc−BDNF and decrease in amygdalar pCREB. Functional studies will be required to determine the exact contribution of these signalling molecules to palatable food relapse. Heightened motivation for sugary and high-fat foods during HFD withdrawal or intermittent access may contribute to an addiction-like process in which repeated cycles of access, deprivation and resumption promote cravings and dependence.


  1. 1

    Gold MS, Graham NA, Cocores JA, Nixon SJ . Food addiction? J Addict Med 2009; 3: 42–45.

    Article  Google Scholar 

  2. 2

    Nair SG, Adams-Deutsch T, Epstein DH, Shaham Y . The neuropharmacology of relapse to food seeking: methodology, main findings, and comparison with relapse to drug seeking. Prog Neurobiol 2009; 89: 18–45.

    CAS  Article  Google Scholar 

  3. 3

    Ghitza UE, Gray SM, Epstein DH, Rice KC, Shaham Y . The anxiogenic drug yohimbine reinstates palatable food seeking in a rat relapse model: a role of CRF1 receptors. Neuropsychopharmacology 2006; 31: 2188–2196.

    CAS  Article  Google Scholar 

  4. 4

    Nair SG, Navarre BM, Cifani C, Pickens CL, Bossert JM, Shaham Y . Role of dorsal medial prefrontal cortex dopamine D1-family receptors in relapse to high-fat food seeking induced by the anxiogenic drug yohimbine. Neuropsychopharmacology 2011; 36: 497–510.

    CAS  Article  Google Scholar 

  5. 5

    Ifland JR, Preuss HG, Marcus MT, Rourke KM, Taylor WC, Burau K et al. Refined food addiction: a classic substance use disorder. Med Hypotheses 2009; 72: 518–526.

    CAS  Article  Google Scholar 

  6. 6

    Fulton S . Appetite and reward. Front Neuroendocrinol 2010; 31: 85–103.

    Article  Google Scholar 

  7. 7

    Carlezon WA, Thomas MJ . Biological substrates of reward and aversion: a nucleus accumbens activity hypothesis. Neuropharmacology 2009; 56 (Suppl 1): 122–132.

    CAS  Article  Google Scholar 

  8. 8

    Etkin A, Prater KE, Schatzberg AF, Menon V, Greicius MD . Disrupted amygdalar subregion functional connectivity and evidence of a compensatory network in generalized anxiety disorder. Arch Gen Psychiatry 2009; 66: 1361–1372.

    Article  Google Scholar 

  9. 9

    Tye KM, Prakash R, Kim SY, Fenno LE, Grosenick L, Zarabi H et al. Amygdala circuitry mediating reversible and bidirectional control of anxiety. Nature 2011; 471: 358–362.

    CAS  Article  Google Scholar 

  10. 10

    Hyman SE, Malenka RC, Nestler EJ . Neural mechanisms of addiction: the role of reward-related learning and memory. Annu Rev Neurosci 2006; 29: 565–598.

    CAS  Article  Google Scholar 

  11. 11

    Russo SJ, Bolanos CA, Theobald DE, DeCarolis NA, Renthal W, Kumar A et al. IRS2-Akt pathway in midbrain dopamine neurons regulates behavioral and cellular responses to opiates. Nat Neurosci 2007; 10: 93–99.

    CAS  Article  Google Scholar 

  12. 12

    Li Y, South T, Han M, Chen J, Wang R, Huang XF . High-fat diet decreases tyrosine hydroxylase mRNA expression irrespective of obesity susceptibility in mice. Brain Res 2009; 1268: 181–189.

    CAS  Article  Google Scholar 

  13. 13

    South T, Westbrook F, Morris MJ . Neurological and stress related effects of shifting obese rats from a palatable diet to chow and lean rats from chow to a palatable diet. Physiol Behav 2012; 105: 1052–1057.

    CAS  Article  Google Scholar 

  14. 14

    Vucetic Z, Carlin J, Totoki K, Reyes TM . Epigenetic dysregulation of the dopamine system in diet-induced obesity. J Neurochem 2012; 120: 891–898.

    CAS  PubMed  PubMed Central  Google Scholar 

  15. 15

    Sharma S, Fulton S . Diet-induced obesity promotes depressive-like behaviour that is associated with neural adaptations in brain reward circuitry. Int J Obes 2012. e-pub ahead of print 17 April 2012; doi:10:1038/ijo2012.48.

  16. 16

    Cottone P, Sabino V, Steardo L, Zorrilla EP . Opioid-dependent anticipatory negative contrast and binge-like eating in rats with limited access to highly preferred food. Neuropsychopharmacology 2008; 33: 524–535.

    CAS  Article  Google Scholar 

  17. 17

    Cottone P, Sabino V, Steardo L, Zorrilla EP . Intermittent access to preferred food reduces the reinforcing efficacy of chow in rats. Am J Physiol Regul Integr Comp Physiol 2008; 295: R1066–R1076.

    CAS  Article  Google Scholar 

  18. 18

    Cottone P, Sabino V, Steardo L, Zorrilla EP . Consummatory, anxiety-related and metabolic adaptations in female rats with alternating access to preferred food. Psychoneuroendocrinology 2009; 34: 38–49.

    CAS  Article  Google Scholar 

  19. 19

    Cottone P, Sabino V, Roberto M, Bajo M, Pockros L, Frihauf JB et al. CRF system recruitment mediates dark side of compulsive eating. Proc Natl Acad Sci USA 2009; 106: 20016–20020.

    CAS  Article  Google Scholar 

  20. 20

    Pickering C, Alsiö J, Hulting A-L, Schiöth H . Withdrawal from free-choice high-fat high-sugar diet induces craving only in obesity-prone animals. Psychopharmacology (Berl) 2009; 204: 431–443.

    CAS  Article  Google Scholar 

  21. 21

    Richardson NR, Roberts DC . Progressive ratio schedules in drug self-administration studies in rats: a method to evaluate reinforcing efficacy. J Neurosci Methods 1996; 66: 1–11.

    CAS  Article  Google Scholar 

  22. 22

    Sharma S, Hryhorczuk C, Fulton S . Progressive-ratio Responding for Palatable High-fat and High-sugar Food inMice. J Vis Exp 2012; 63: e3754.

    Google Scholar 

  23. 23

    Carobrez AP, Bertoglio LJ . Ethological and temporal analyses of anxiety-like behavior: The elevated plus-maze model 20 years on. Neuroscience & Biobehavioral Reviews 2005; 29: 1193–1205.

    CAS  Article  Google Scholar 

  24. 24

    Papp M, Willner P, Muscat R . An animal model of anhedonia: attenuation of sucrose consumption and place preference conditioning by chronic unpredictable mild stress. Psychopharmacology (Berl) 1991; 104: 255–259.

    CAS  Article  Google Scholar 

  25. 25

    Teegarden SL, Bale TL . Decreases in Dietary Preference Produce Increased Emotionality and Risk for Dietary Relapse. Biol Psychiatry 2007; 61: 1021–1029.

    Article  Google Scholar 

  26. 26

    George O, Ghozland S, Azar MR, Cottone P, Zorrilla EP, Parsons LH et al. CRF-CRF1 system activation mediates withdrawal-induced increases in nicotine self-administration in nicotine-dependent rats. Proc Natl Acad Sci U S A 2007; 104: 17198–17203.

    CAS  Article  Google Scholar 

  27. 27

    Goeders NE, Guerin GF . Effects of the CRH receptor antagonist CP-154,526 on intravenous cocaine self-administration in rats. Neuropsychopharmacology 2000; 23: 577–586.

    CAS  Article  Google Scholar 

  28. 28

    Koob GF . Brain stress systems in the amygdala and addiction. Brain Res 2009; 1293: 61–75.

    CAS  Article  Google Scholar 

  29. 29

    Richter RM, Weiss F . In vivo CRF release in rat amygdala is increased during cocaine withdrawal in self-administering rats. Synapse 1999; 32: 254–261.

    CAS  Article  Google Scholar 

  30. 30

    Roberto M, Cruz MT, Gilpin NW, Sabino V, Schweitzer P, Bajo M et al. Corticotropin releasing factor-induced amygdala gamma-aminobutyric Acid release plays a key role in alcohol dependence. Biol Psychiatry 2010; 67: 831–839.

    CAS  Article  Google Scholar 

  31. 31

    Graham DL, Edwards S, Bachtell RK, DiLeone RJ, Rios M, Self DW . Dynamic BDNF activity in nucleus accumbens with cocaine use increases self-administration and relapse. Nat Neurosci 2007; 10: 1029–1037.

    CAS  Article  Google Scholar 

  32. 32

    Krishnan B, Centeno M, Pollandt S, Fu Y, Genzer K, Liu J et al. Dopamine receptor mechanisms mediate corticotropin-releasing factor-induced long-term potentiation in the rat amygdala following cocaine withdrawal. Eur J Neurosci 2010; 31: 1027–1042.

    Article  Google Scholar 

  33. 33

    Nestler EJ, Kelz MB, Chen J . DeltaFosB: a molecular mediator of long-term neural and behavioral plasticity. Brain Res 1999; 835: 10–17.

    CAS  Article  Google Scholar 

  34. 34

    Russo SJ, Mazei-Robison MS, Ables JL, Nestler EJ . Neurotrophic factors and structural plasticity in addiction. Neuropharmacology 2009; 56 (Suppl 1): 73–82.

    CAS  Article  Google Scholar 

  35. 35

    Robison AJ, Nestler EJ . Transcriptional and epigenetic mechanisms of addiction. Nat Rev Neurosci 2011; 12: 623–637.

    CAS  Article  Google Scholar 

  36. 36

    Koob GF . Stress, corticotropin-releasing factor, and drug addiction. Ann N Y Acad Sci 1999; 897: 27–45.

    CAS  Article  Google Scholar 

  37. 37

    Sarnyai Z . Neurobiology of stress and cocaine addiction. Studies on corticotropin-releasing factor in rats, monkeys, and humans. Ann N Y Acad Sci 1998; 851: 371–387.

    CAS  Article  Google Scholar 

  38. 38

    Muscat R, Papp M, Willner P . Reversal of stress-induced anhedonia by the atypical antidepressants, fluoxetine and maprotiline. Psychopharmacology (Berl) 1992; 109: 433–438.

    CAS  Article  Google Scholar 

  39. 39

    Papp M, Willner P, Muscat R . Behavioural sensitization to a dopamine agonist is associated with reversal of stress-induced anhedonia. Psychopharmacology (Berl) 1993; 110: 159–164.

    CAS  Article  Google Scholar 

  40. 40

    Avena NM, Bocarsly ME, Rada P, Kim A, Hoebel BG . After daily bingeing on a sucrose solution, food deprivation induces anxiety and accumbens dopamine/acetylcholine imbalance. Physiol Behav 2008; 94: 309–315.

    CAS  Article  Google Scholar 

  41. 41

    Colantuoni C, Rada P, McCarthy J, Patten C, Avena NM, Chadeayne A et al. Evidence that intermittent, excessive sugar intake causes endogenous opioid dependence. Obesity 2002; 10: 478–488.

    CAS  Article  Google Scholar 

  42. 42

    Bailey A, Davis L, Lesscher HMB, Kelly MDW, Ledent C, Hourani SMO et al. Enhanced morphine withdrawal and μ-opioid receptor G-protein coupling in A2A adenosine receptor knockout mice. J Neurochem 2004; 88: 827–834.

    CAS  Article  Google Scholar 

  43. 43

    Shepard JD, Bossert JM, Liu SY, Shaham Y . The anxiogenic drug yohimbine reinstates methamphetamine seeking in a rat model of drug relapse. Biol Psychiatry 2004; 55: 1082–1089.

    CAS  Article  Google Scholar 

  44. 44

    Sinha R, Fox HC, Hong KA, Bergquist K, Bhagwagar Z, Siedlarz KM . Enhanced negative emotion and alcohol craving, and altered physiological responses following stress and cue exposure in alcohol dependent individuals. Neuropsychopharmacology 2009; 34: 1198–1208.

    CAS  Article  Google Scholar 

  45. 45

    Zhao LY, Shi J, Zhang XL, Epstein DH, Zhang XY, Liu Y et al. Stress enhances retrieval of drug-related memories in abstinent heroin addicts. Neuropsychopharmacology 2010; 35: 720–726.

    CAS  Article  Google Scholar 

  46. 46

    Ghitza UE, Nair SG, Golden SA, Gray SM, Uejima JL, Bossert JM et al. Peptide YY3-36 decreases reinstatement of high-fat food seeking during dieting in a rat relapse model. J Neurosci 2007; 27: 11522–11532.

    CAS  Article  Google Scholar 

  47. 47

    Nair SG, Adams-Deutsch T, Pickens CL, Smith DG, Shaham Y . Effects of the MCH1 receptor antagonist SNAP 94847 on high-fat food-reinforced operant responding and reinstatement of food seeking in rats. Psychopharmacology (Berl) 2009; 205: 129–140.

    CAS  Article  Google Scholar 

  48. 48

    Fulton S, Pissios P, Manchon RP, Stiles L, Frank L, Pothos EN et al. Leptin regulation of the mesoaccumbens dopamine pathway. Neuron 2006; 51: 811–822.

    CAS  Article  Google Scholar 

  49. 49

    Geiger BM, Haburcak M, Avena NM, Moyer MC, Hoebel BG, Pothos EN . Deficits of mesolimbic dopamine neurotransmission in rat dietary obesity. Neuroscience 2009; 159: 1193–1199.

    CAS  Article  Google Scholar 

  50. 50

    Davis JF, Tracy AL, Schurdak JD, Tschop MH, Lipton JW, Clegg DJ et al. Exposure to elevated levels of dietary fat attenuates psychostimulant reward and mesolimbic dopamine turnover in the rat. Behav Neurosci 2008; 122: 1257–1263.

    Article  Google Scholar 

  51. 51

    Davis C, Patte K, Levitan R, Reid C, Tweed S, Curtis C . From motivation to behaviour: a model of reward sensitivity, overeating, and food preferences in the risk profile for obesity. Appetite 2007; 48: 12–19.

    Article  Google Scholar 

  52. 52

    Stice E, Yokum S, Blum K, Bohon C . Weight gain is associated with reduced striatal response to palatable food. J Neurosci 2010; 30: 13105–13109.

    CAS  Article  Google Scholar 

  53. 53

    Stice E, Yokum S, Burger KS, Epstein LH, Small DM . Youth at risk for obesity show greater activation of striatal and somatosensory regions to food. J Neurosci 2011; 31: 4360–4366.

    CAS  Article  Google Scholar 

  54. 54

    Volkow ND, Wang GJ, Baler RD . Reward, dopamine and the control of food intake: implications for obesity. Trends Cogn Sci 2011; 15: 37–46.

    CAS  Article  Google Scholar 

  55. 55

    Jang S, Kim D, Lee Y, Moon S, Oh S . Modulation of Sphingosine 1-Phosphate and Tyrosine Hydroxylase in the Stress-Induced Anxiety. Neurochem Res 2011; 36: 258–267.

    CAS  Article  Google Scholar 

  56. 56

    Auvinen HE, Romijn JA, Biermasz NR, Pijl H, Havekes L, Smit JW et al. The Effects of high fat diet on the basal activity of the hypothalamus-pituitary-adrenal axis in mice. J Endocrinol 2012; 214: 191–197.

    CAS  Article  Google Scholar 

  57. 57

    Bale TL, Picetti R, Contarino A, Koob GF, Vale WW, Lee K-F . Mice deficient for both corticotropin-releasing factor receptor 1 (crfr1) and crfr2 have an impaired stress response and display sexually dichotomous anxiety-like behavior. The Journal of Neuroscience 2002; 22: 193–199.

    CAS  Article  Google Scholar 

  58. 58

    Smith GW, Aubry J-M, Dellu F, Contarino A, Bilezikjian LM, Gold LH et al. Corticotropin releasing factor receptor 1–deficient mice display decreased anxiety, impaired stress response, and aberrant neuroendocrine development. Neuron 1998; 20: 1093–1102.

    CAS  Article  Google Scholar 

  59. 59

    Carlezon WA, Thome J, Olson VG, Lane-Ladd SB, Brodkin ES, Hiroi N et al. Regulation of cocaine reward by CREB. Science 1998; 282: 2272–2275.

    CAS  Article  Google Scholar 

  60. 60

    Edwards S, Graham DL, Bachtell RK, Self DW . Region-specific tolerance to cocaine-regulated cAMP-dependent protein phosphorylation following chronic self-administration. Eur J Neurosci 2007; 25: 2201–2213.

    Article  Google Scholar 

  61. 61

    Pandey SC, Roy A, Xu T, Mittal N . Effects of protracted nicotine exposure and withdrawal on the expression and phosphorylation of the CREB gene transcription factor in rat brain. J Neurochem 2001; 77: 943–952.

    CAS  Article  Google Scholar 

  62. 62

    Cordeira JW, Frank L, Sena-Esteves M, Pothos EN, Rios M . Brain-derived neurotrophic factor regulates hedonic feeding by acting on the mesolimbic dopamine system. J Neurosci 2010; 30: 2533–2541.

    CAS  Article  Google Scholar 

  63. 63

    Lobo MK, Covington HE, Chaudhury D, Friedman AK, Sun H, Damez-Werno D et al. Cell type-specific loss of BDNF signaling mimics optogenetic control of cocaine reward. Science 2010; 330: 385–390.

    CAS  Article  Google Scholar 

  64. 64

    Vialou V, Robison AJ, LaPlant QC, Covington HE, Dietz DM, Ohnishi YN et al. [Delta]FosB in brain reward circuits mediates resilience to stress and antidepressant responses. Nat Neurosci 2010; 13: 745–752.

    CAS  Article  Google Scholar 

Download references


This project was supported by grants to SF from the Natural Sciences and Engineering Research Council of Canada (No.355881), Canadian Diabetes Association (OG-2-09-2835-SF) and Canadian Foundation for Innovation. SS is supported by a postdoctoral fellowship from the CIHR Neuroinflammation Training Program and MFF by a PhD fellowship from the Canadian Diabetes Association.

Author information



Corresponding author

Correspondence to S Fulton.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Sharma, S., Fernandes, M. & Fulton, S. Adaptations in brain reward circuitry underlie palatable food cravings and anxiety induced by high-fat diet withdrawal. Int J Obes 37, 1183–1191 (2013).

Download citation


  • high-fat food
  • reward
  • anxiety
  • stress
  • mesolimbic
  • dopamine

Further reading