Influence of caloric restriction history, intermittent exposure to palatable food and sex in a binge eating model in C57BL/6J mice Authors: Nuria Morales-Puerto, Clémence Fayt, Chloé Tezenas du Montcel, Amandine Everard Published: 2025 ABSTRACT AbstractBackgroundSeveral animal protocols have been designed to model binge eating disorder (BED) and elucidate its neurological basis. A wide variety of factors, including the frequency of access to palatable food and a previous history of caloric restriction, have been modulated to induce binge eating intake. The aims of this work were to compare the effects of these two factors on the development of binge eating behaviour and to analyse possible alterations in neurotransmitter markers in the striatum.MethodsWe performed a binge eating test in male and female C57BL/6J mice, which consisted of 2 h of limited access to a high-fat high-sucrose (HFHS) diet in a sated state with ad libitum access to a control diet and water. Mice were subjected to either daily or intermittent (every other day) 2-h limited access to the HFHS, with or without previous episodes of caloric restriction. Dopaminergic, glutamatergic and GABAergic markers were analysed in the striatum by RT–qPCR. Behavioural and biomolecular differences between groups were analysed by one-way ANOVA for multiple comparisons.ResultsBoth male and female mice developed binge eating behaviour following a 2-h HFHS access paradigm, while the frequency of exposure and history of caloric restriction did not affect the BE response. However, the combination of caloric restriction and intermittent access to HFHS induced an early BE response in females, with no compensatory decrease in chow intake and no weight changes in either sex. Moreover, we observed differences in the modulation of Drd2 expression between animals subjected to binge eating and those with ad libitum access to HFHS, independent of sex.ConclusionsThis work is the first to compare the effects of the frequency of exposure to HFHS and a history of caloric restriction on binge eating behaviour in both male and female mice within a single experimental setting. The results do not reveal any major effects on binge eating behaviour but highlight the combination of both conditions for translational applications since any compensatory process is observed. Moreover, the results underscore a role of striatal Drd2 expression as a key factor involved in the course of BED.Supplementary InformationThe online version contains supplementary material available at 10.1186/s40337-025-01503-8.Keywords: Binge eating disorder (BED), High-fat high-sucrose, Limited access, Intermittent access, Caloric restriction, Dopamine, Drd2 Background Binge eating disorder (BED) is a psychiatric disorder classified in the “Feeding and Eating Disorders” section of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5). It is characterized by repetitive episodes of eating a large amount of food in a discrete period of time, associated with a sense of loss of control over eating during the episodes. Unlike bulimia nervosa patients, patients affected with BED do not engage in compensatory behaviours (i.e. excessive exercise, purging, laxative use) but report psychological distress, such as disgust, embarrassment or guilt over the binge eating (BE) condition [1]. Compared with anorexia nervosa or bulimia nervosa, BED is the most prevalent eating disorder, affecting 1.5% of women and 0.3% of men in the general population, with a more balanced female/male sex ratio [2]. Importantly, it is associated with a high relapse rate and a low probability of full recovery, increasing the risk of chronic disease [3]. Moreover, BED is frequently associated with both psychiatric comorbidities, especially anxiety and depressive disorders, as well as co-occurring medical conditions such as obesity, type 2 diabetes and hypertension due to excessive energy intake [4]. Access to adequate treatment for BED is hampered by multiple factors, such as social stigma, frequently increasing delays in seeking treatment but also the lack of knowledge on the mechanistic aspects of the disease and adequate validation of drug efficacy [3, 5, 6]. Recently, accumulating evidence has implicated the reward system in the pathophysiology of BED [7]. Dopamine is among the main neurotransmitters involved in reward signalling during the anticipation and consumption of palatable food, as well as in decision-making. Interestingly, the only approved pharmacotherapy for the treatment of BED is lisdexamfetamine, a drug that targets the recapture of dopamine by blocking its transporter [8]. Findings regarding imbalances in dopaminergic markers in BED patients have been divergent, as both hyperactivation and hypoactivation have been described [9]. These discrepancies in findings can be attributed to variabilities between behavioural tasks (i.e. exposure to palatable food cues or inhibitory processes) [10–14], the type of control group (i.e. healthy controls or weight-paired controls within an obese population) [15], and the duration of illness with early and late stages that could be associated with opposite neurobiological modulations in BED, supporting the concept of staging in eating disorders [16]. Moreover, other neurotransmitters, such as excitatory glutamatergic and inhibitory GABAergic projections, are crucial for the control of feeding behaviours and the reward response [17]. Given the poor understanding of the synaptic alterations in the reward circuitry in the context of eating disorders, further exploration of the role of these neurotransmitters in BED is imperative. Animal models have been developed to mimic different core characteristics of BED with the goal of identifying the molecular and neuronal prints associated with this disorder. The current BE models are predominantly based on time-restricted access to palatable food, which has been demonstrated to induce BE intake (i.e. significantly higher consumption than unrestricted access). Several studies have explored stress-induced BE using various stressors, such as social defeat, forced-swim tests or foot shocks prior to exposure to palatable food [18–20]. Emotional stress, characterized by frustration from exposure to inaccessible palatable food, induces an increase in BE behaviour in female mice [21, 22]. Although relevant in terms of translational potential, stress-induced BE models involve neural substrates that are implicated in emotional and memory processes distinct from reward-related circuits. Among the diverging factors between BE protocols, the nature of the food chosen to induce BE intake is crucial. Food enriched in sugar, fat or a combination of both is classically used as a palatable food in preclinical models [23]. The combination of fat and sugar should, however, be prioritized, as patients frequently display BE episodes with high-energy palatable food [24]. The timeframe and period of the circadian cycle of exposure to palatable food also varies in BE models that are available in the literature. Most paradigms use a restricted access period between 1 and 2 h, which is close to the clinical definition of a BE episode. Protocols using shorter (less than 1 h) or longer (12–24 h) durations [25] have less translational validity because of poor comparability with clinical settings [26]. Exposure during the light phase has been described as sated feeding (i.e. food intake without feeling hungry), which is a crucial component for evaluating non-homeostatic feeding similar to bingeing in humans (criteria B3 of DSM-5: eating large amounts of food when not feeling physically hungry). In a recent meta-analysis of animal models of BE, more than 80% of the studies assessed binge-like intake in sated animals and concluded that this type of access should be prioritized [25]. Two other factors have been classically targeted to modulate BE. First, the frequency of access to palatable food refers to the time between periods of access to the target diet. Early studies by Corwin and colleagues revealed that the more episodic the access, the more marked the development of BE intake will be [27]. This intermittent access may also induce behavioural changes related to food intake, such as an escalation in the consumption of palatable food (i.e. an increase in BE intake over time) [28, 29]. That said, the use of escalation as a criterion required to define BE in rodents remains questionable since it is not a criterion for the diagnosis of BED, as specified in the DMS-5. The second main factor in overeating protocols is a history of food restriction, which refers to a limitation on the availability of standard chow, imitating periods of dietary restraint that patients undergo to lose weight and that have been shown to predict BE episodes [30–32]. Generally, these protocols alternate periods of caloric restriction with periods of ad libitum access to a standard diet and limited access to a palatable diet. Indeed, several studies in recent decades have acknowledged that BE patterns are invigorated by episodes of caloric restriction [33, 34]. Others have reported that a history of caloric restriction could be necessary to induce BE [35] or that caloric restriction potentiated with stress [36–39] increases BE intake. Finally, the type of experimental subject, including the species and sex of the animals, seems to modulate the response to BE paradigms. Most preclinical research on eating disorders has been conducted on rodents, and studies have shown that females tend to exhibit a more pronounced BE response. However, few studies have included both sexes in their experiments, making it difficult to compare responses between females and males [10]. Taken together, these protocols differ greatly in terms of the key parameters that induce BE intake. Although extensive reviews of the literature have provided valuable information on the influence of these features [25, 40], comparisons of the effects of the different factors used to induce BE in rodent models performed within a single experimental setting are lacking to provide a uniformized BE protocol for future research. Therefore, the objective of this study was to compare, in the same experiment, the effects of two main binge-triggering factors, namely, the frequency of exposure to palatable food and a history of caloric restriction, and potential sex differences on the development of BE. For this purpose, we exposed male and female mice to BE protocols involving different schedules of 2-h access to a HFHS diet with or without episodes of caloric restriction. We then evaluated the gene expression of markers of the dopaminergic, glutamatergic and GABAergic systems in the striatum to determine whether BE-inducing factors could differentially modulate the signalling of key neurotransmitters in reward-associated areas. Methods Experimental animals All experimental procedures were conducted according to the European legislation EU Directive 2010/63 and were approved by the ethical committee for animal care of the Health Sector of the Université Catholique de Louvain, under the specific number A21 MNUT 2023-A21 (2023-41), in accordance with the Belgian Law of May 29, 2013, regarding the protection of laboratory animals (agreement number LA1230314). A cohort of 9-week-old specific-opportunistic and pathogen-free (SOPF) male and female C57BL/6J mice (64 mice, n = 4–6/group, Janvier laboratories, France) were housed in groups of two same-sex mice per cage, with free access to sterile (irradiated) food and sterile (autoclaved) water in a controlled environment (room temperature of 22 ± 2 °C, 45–60% humidity, 12-h daylight cycle). Upon delivery, the mice underwent a 1-week acclimatization period, during which time they had ad libitum access to a control diet (CT, 3.85 kcal/g with 9% fat and 76% carbohydrates (kcal/100 g), AIN93Mi, Research Diets Inc., New Brunswick, NJ, USA). During the last 3 days of acclimatization, CT food intake was measured to calculate the 24-h intake and determine the caloric restriction (66% of daily CT intake), which was used later in some conditions of the experiments before the BE test. Finally, on the last day of habituation, the CT food was removed from the cages, and the mice were exposed overnight to a palatable high-fat high-sucrose diet (HFHS, 4.7 kcal/g with 45% fat and 27.8% sucrose (kcal/100 g), D17110301i, Research Diets Inc., New Brunswick, NJ, USA) to prevent neophobia to the palatable food that was later used during the BE test. The diets used in this study were subjected to a single gamma-irradiation cycle (10–20 kGy) to ensure sterility and to cause only minimal nutrient loss without affecting palatability or intake. Afterwards, the mice were randomly divided into groups of similar body weight with different binge schedules. During the entire duration of the procedure, CT food was available ad libitum, except during periods of food restriction for the corresponding experimental groups, as described in the next section. Experimental design The BE test consisted of limited periods of 2-h access to HFHS repeated within a 9-day cycle and with ad libitum access to CT, as described in Fig. 1a. During this 2-h period, the animals were separated into individual cages to assess individual food consumption (Fig. 1b). After completion of the test, we measured 2-h HFHS and CT intake. Ad libitum 24-h CT intake was also measured throughout the entire duration of the experiment. The animal placed in the new cage was counterbalanced in each session, and the weight of the mice was monitored daily after the BE test. As the mice are nocturnal, the 2-h binge period always took place during the light phase of the circadian cycle to ensure that the mice were in a sated state; thus, the consumption of HFHS was motivated by pleasure and reward rather than physiological hunger. The 9-day cycle was repeated for a total of 4 BE cycles (Fig. 1a). Experimental design. a The BE test was performed in cycles of 9 days, for a total of 4 cycles. The daily binge group had access to HFHS every day, while the intermittent binge group had access every 2 days. In the caloric restriction groups, CT food restriction occurred during the first 3 days of each cycle and was reduced to 66% of normal CT intake. After caloric restriction and an overnight refeeding period, either a daily or an intermittent schedule was followed. The control groups consisted of a continuous access group, with ad libitum access to HFHS and CT food, and a no-access group, which was only exposed to CT food. Days of access to HFHS are represented in black and bold. b The BE test consisted of 2-h access to HFHS during the light phase. After 2 h, the mice were returned to their home cages and housed in pairs. Water and CT food were available ad libitum. CA continuous access, CR caloric restriction, CT control food, DB daily binge, HFHS high-fat high-sucrose, IB intermittent binge, NA no access To elucidate the effects of the frequency of palatable food exposure and caloric restriction on BE intake, the following BE groups were established (Fig. 1a): Daily binge (DB): daily 2-h access to HFHS, with no history of caloric restriction. Daily binge (DB): daily 2-h access to HFHS, with no history of caloric restriction. Intermittent binge (IB): 2-h access to HFHS every 48 h, with no history of caloric restriction. Intermittent binge (IB): 2-h access to HFHS every 48 h, with no history of caloric restriction. Caloric restriction + daily binge (CR + DB): 3 days of restricted access to CT food (66% of normal daily CT intake) + daily 2-h access to HFHS (6 BE test). Caloric restriction + daily binge (CR + DB): 3 days of restricted access to CT food (66% of normal daily CT intake) + daily 2-h access to HFHS (6 BE test). Caloric restriction + intermittent binge (CR + IB): 3 days of restricted access to CT food (66% of normal daily CT intake) + 2-h access to HFHS every 48 h (3 BE test). Caloric restriction + intermittent binge (CR + IB): 3 days of restricted access to CT food (66% of normal daily CT intake) + 2-h access to HFHS every 48 h (3 BE test). These experimental groups were compared to two control groups: Continuous access (CA): ad libitum access to both CT and HFHS food (24 h/9 d access). Continuous access (CA): ad libitum access to both CT and HFHS food (24 h/9 d access). No access (NA): ad libitum access to CT food (never exposed to HFHS). No access (NA): ad libitum access to CT food (never exposed to HFHS). For the caloric-restricted groups, during the 3-day caloric restriction period, the mice had access only to CT food and never to HFHS. Importantly, after the 3rd day of food limitation, caloric-restricted animals were refed overnight with ad libitum CT food to avoid hunger-driven HFHS intake on the 1st day of exposure to HFHS in each cycle. In this sense, HFHS intake in caloric-restricted mice took place under a sated state; hence, it was not simply a product of physiological hunger. In any experimental group, the operational definition of BE behaviour was established as a caloric intake of HFHS that was significantly higher than that of the continuous access (CA) group during the 2 h during which the BE test took place. Tissue sampling At the end of each experiment, the mice were euthanized for tissue collection. First, the mice were exposed to the HFHS diet in their cages for 1 h to activate brain reward signalling. Afterwards, the mice were anaesthetized with 2.7% isoflurane (Forene, Abbott, England). Finally, the mice were euthanized by exsanguination and cervical dislocation. The brain region of the dorsal striatum was precisely dissected (from bregma AP = + 0.86, ML = + 1.5, and DV = -3.25) using a mouse brain matrix with 1-mm-thick coronal sections and immediately immersed in liquid nitrogen and then stored at − 80 °C for further analysis. RNA Preparation and real-time RT–qPCR analysis Total RNA was extracted from the dorsal striatum using TriPure reagent (Roche, Bale, Switzerland). Complementary DNA was prepared by reverse transcription of 1 µg of total RNA using a GoScript Reverse Transcriptase kit (Promega, Madison, WI, USA). Real-time PCR was performed with a QuantStudio 3 real-time PCR system (Thermo Fisher, Waltham, MA, USA). Rpl19 was chosen as the housekeeping gene. All samples were examined in duplicate, and the data were analysed according to the 2−ΔΔCt method. The identity and purity of the amplified product were assessed by melting curve analysis. The sequences of the primers used for real-time PCR are shown in Table 1. Table 1Sequences of the primers used for qPCRSystemGeneForwardReverseDopaminergic Drd1 CCAAGAACGTGAGGGCTAAGTGAGGATGCGAAAGGAGAAG Drd2 GAGCCAACCTGAAGACACCTGACAGCATCTCCATTTCCAG Slc6a3 AAATGCTCCGTGGGACCAATGGTCTCCCGCTCTTGAACCTC Th GCCAAGGACAAGCTCAGGAACATCAATGGCCAGGGTGTACGGlutamatergic Gria1 ACCACTACATCCTCGCCAACTCACTTGTCCTCCACTGCTG Gria2 TGCTTGAAGGGAATGAGCGTTCTGCATACTTGGAAGACATCAG Grin2a CGCTCTGCTCCAGTTTGTTGGCTGCTCATCACCTCATTCTTC Grin2b TGGTATCACGCAGCAATGGCAGAGACAATGAGCAGCATCAC Slc1a2 GCCAACAATATGCCCAAGCAGGACACCAAACACAGTCAGTGAGABAergic Gabra1 GCCCTCCCAAGATGAACTTACTCTCCCAAACCTGGTCTCA Gabrb1 TCGGGTGGTATGCTGACAACGGAAATGCTTCGGGTGTTGG Gabrb2 TGTGCATCCTGATCTGCTGGGGGTGTTTTCGCAGTGTTCC Slc6a11 ATTATGTCTCCTGGCGGCCTAAAAGATCTGCGTCCCAGCG Gad1 GAACAACCATGGTGGGCTACGCCGATGATTCTGGTTCTGTHouse-keeping Rpl19 GAAGGTCAAAGGGAATGTGTTCACCTTGTCTGCCTTCAGCTTGT Sequences of the primers used for qPCR Statistical analysis Statistical analyses were performed using GraphPad Prism version 10.2.3 for Windows (GraphPad Software, San Diego, CA, USA). The data are expressed as the mean ± SEM. Differences between groups were assessed using one-way ANOVA, followed by Tukey’s post hoc test for multiple comparisons, in the case of the behavioural analysis. For qPCR analysis, the data were tested for a normal distribution with the Shapiro-Wilk test. If the data did not follow a normal distribution, the nonparametric Kruskal–Wallis test was performed, followed by Dunn’s post hoc test for multiple comparisons against the continuous access group. When the data were normally distributed, one-way ANOVA followed by Dunnett’s post hoc test was used for multiple comparisons against the continuous access group. Statistical analysis was performed by comparing each experimental group against the CA control group to control for the effect of HFHS exposure on neurotransmission in reward-related brain regions. Two-way ANOVA (repeated measures or not), followed by Tukey’s post hoc test, was used for the escalation of HFHS consumption across the different BE cycles. For all the experiments, outlier values were excluded after the ROUT test, with a significance level of Q = 1%. The threshold for statistical significance was set at p < 0.05. Results Limited 2-h HFHS access induces binge eating intake in both male and female mice To assess the response of the animals to BE protocols, we compared the average 2-h HFHS intake during the last cycle (cycle 4) in all experimental groups against the continuous access (CA) to HFHS as a control group, given that ad libitum availability of the palatable diet did not induce a binge-like phenotype. We observed greater HFHS intake in all experimental groups exposed to limited access to HFHS food compared with the CA group over the same 2-h period in both male and female mice (Fig. 2a, males: p = 0.0002, F(4, 23) = 8.519; Fig. 2d, females: p = 0.0003, F(4, 25) = 7.868). However, the frequency of limited exposure to HFHS did not affect BE intake, as no differences were observed among the daily (DB) and intermittent bingeing (IB) mice. Similarly, previous caloric restriction did not affect BE intake, since compared with non-restricted animals (DB and IB), animals exposed to chow restriction (CR + DB and CR + IB) had similar levels of HFHS intake (Fig. 2a and d). Moreover, the CT diet intake was close to zero for all groups during the BE tests (data not shown), which indicated a strong tropism towards palatable food. Binge eating test. Behavioural analysis of 2-h HFHS intake (a, d), 24-h CT food intake (b, e) and body weight gain (c, f) in male (a–c) and female (d–f) mice subjected to the BE test. a, d Average 2-h consumption of HFHS (kcal/g) during cycle 4. b, e Average 24-h consumption of CT food (kcal/g) during cycle 4 (Days 6 to 9). c, f Total body weight gain (g) from the beginning of cycle 1 to the end of cycle 4 of the BE test. The data are expressed as the mean ± SEM; n = 4–6/group. The results were analysed by one-way ANOVA, followed by Tukey’s post hoc test for multiple comparisons. *p < 0.05, ** p < 0.01, *** p < 0.001. CA continuous access, CR + DB caloric restriction + daily binge, CR + IB caloric restriction + intermittent binge, CT control diet, DB daily binge, HFHS high-fat high-sucrose diet, IB intermittent binge, NA no access To explore potential temporal patterns in the BE response, we compared 2 h-HFHS intake in experimental groups separately in each BE cycle in males and females. Interestingly, we found no major difference between groups in the first cycle in males (Supp. Fig. S1a), whereas in females, significantly higher HFHS intake was already observed in the CR + IB group compared with the CA group (Supp. Fig. S1d). The BE response appeared to be significantly greater in all the experimental groups compared with the CA group in both males and females during cycles 2 and 3 (Supp. Fig. S1b, c, e, f). To study escalation on palatable food intake following repeated exposure, we compared the average consumption of HFHS on the 1st cycle with that on the 4th cycle. In male mice, we observed a positive escalation for all BE schedules (DB, IB, DB + CR and IB + CR), which revealed an increase in HFHS consumption over the four BE cycles (p < 0.0001, F(3, 60) = 28.13) (Supp. Fig. S2a-c). Interestingly, in females, only the DB group had significantly higher HFHS intake in the 4th cycle than in the 1st cycle, but no clear escalation was observed throughout the different cycles for the other groups (p = 0.047, F(3, 59) = 2.813) (Supp. Fig. S2d–f). Furthermore, to investigate potential compensatory behaviours, we monitored possible changes in CT diet intake as a decrease in chow consumption to balance the excessive energy intake from a palatable diet. We calculated the average 24-h CT food intake from the final days of cycle 4 (Days 6 to 9), expressed per gram of body weight, compared with that of animals exposed only to a CT diet and hence with a prototypical CT intake (no access to HFHS, NA). In male mice, only animals that were not exposed to a previous history of caloric restriction decreased their CT intake compared with that in the NA group (p = 0.0758 NA vs. DB, p = 0.0112 NA vs. IB) (Fig. 2b), whereas in females, a significant decrease in CT food intake was observed in groups exposed daily to HFHS (DB and CR + DB) compared with that in the NA group (p = 0.0084 NA vs. DB, p = 0.0084 NA vs. CR + DB) (Fig. 2e). Finally, to determine the impact of the different BE protocols on body weight, we calculated the weight gain of the animals from the beginning of cycle 1 to the last day of cycle 4. In general, at the end of the entire experiment, total body weight gain did not significantly differ within male (Fig. 2c) and female mice (Fig. 2f), with the exception of DB male mice, which had a significantly greater weight gain than that observed in the intermittent binge groups and control groups. Drd2 expression is differentially modulated in the striatum after binge eating exposure compared with continuous HFHS access The second objective was to analyse potential alterations in the expression of neurotransmission markers in the striatum of mice subjected to different BE protocols. We studied the gene expression of receptors, transporters and enzymes involved in dopaminergic, glutamatergic and GABAergic neurotransmission by RT–qPCR. To delineate the specific effects related to BE behaviour, we used the NA group as the reference for relative expression levels. In general, no statistically significant differences were observed between the groups for any of the dopaminergic markers analysed in the striatum. That said, some tendencies could be observed. With respect to dopamine receptor 1 (Drd1), the modulation pattern differed by sex, although it also displayed high inter-individual variability. In male mice, compared with those in the CA group, Drd1 expression in the BE-subjected groups tended to increase, with the exception of that in the CR + DB group, which tended to be more similar to that in the CA group, exhibiting a trend towards a reduction in Drd1 expression compared with that in the NA group (Fig. 3a), whereas in female mice, no apparent trend towards changes was observed (Fig. 3e). Interestingly, the expression of dopamine receptor 2 (Drd2) tended to differ between BE animals and mice with ad libitum access to palatable food (i.e. the CA group). Indeed, the Drd2 transcript level seemed to increase in both male and female mice in all the experimental groups subjected to BE tests (DB, IB, CR + DB, and CR + IB) compared with that in the CA group (Fig. 3b, f). That said, this effect seems to be due mainly to the reduction in the Drd2 transcript in the CA group compared with that in the NA group, which did not occur in the BE groups. The dopamine transporter Slc6a3 differed between males and females. In males, daily binge groups were similar to the CA control group, while intermittent bingers tended to increase Slc6a3 expression (Fig. 3c). In females, all groups presented similar levels of Slc6a3 (Fig. 3g). Finally, regarding the limiting enzyme for dopamine synthesis, tyrosine hydroxylase Th, again, the male and female profiles were not similar, with a significant increase in the expression of the enzyme transcript in IB male mice compared with the CA group, but no change in female animals (Fig. 3d, h). To explain the variability in gene expression, we performed correlation analysis between the gene expression of dopaminergic markers in the striatum and BE response or body weight, but none of these correlations were significant (Supp. Table 1). Expression of dopaminergic system markers in the striatum. Relative mRNA levels of striatal dopaminergic markers in male (a–d) and female (e–h) mice subjected to the BE test. a, e Dopamine receptor 1 (Drd1), b, f dopamine receptor 2 (Drd2), c, g dopamine transporter (Slc6a3), and (d, h) tyrosine hydroxylase (Th) mRNA expression in the dorsal striatum. Data are expressed as the mean ± SEM relative to the expression of the no access group (NA) (not shown, represented as a dotted line); n = 4–6/group. The results were analysed by the Kruskal–Wallis test for multiple comparisons followed by Dunn’s post hoc test against the continuous access group. *p < 0.05. CA continuous access, CR + DB caloric restriction + daily binge, CR + IB caloric restriction + intermittent binge, DB daily binge, IB intermittent binge Glutamatergic and GABAergic systems in the striatum were also analysed by RT–qPCR (Supp. Figs. S3 and S4). Unlike Drd2, none of the studied genes showed a similar trend between male and female mice. Changes in the glutamatergic system were observed only in male mice. Compared with that in the CA group, the expression of the subunit 2a of the NMDA receptor (Grin2a) in the IB and CR + DBsubjected groups significantly decreased and followed the same trend in the DB group but remained similar to that in the NA group (Supp. Fig. S3c). In a similar pattern, the expression of the excitatory amino acid transporter 2 (Slc1a2) was also diminished in the CR + DB group compared with the CA group (Supp. Fig. S3e). With respect to the GABAergic system, only the level of the GABA transporter 3 (Slc6a11) was significantly greater in the CR + DB female mice than in the CA group (Supp. Fig. S4i). Discussion Given its prevalence, severity and chronicity, it is imperative to pursue a deeper understanding of BED. However, diversity in preclinical experimental designs to induce BE behaviour makes it difficult to extract conclusions that are valid across all paradigms. The aim of the current study was to explore the specific effects of specific features of BE protocols on behavioural responses and neurobiological adaptations. To the best of our knowledge, this is the first study to jointly examine the binge-triggering effects of these two parameters (i.e. frequency of exposure to HFHS or history of caloric restriction) under the same experimental settings in both sexes. Our results revealed that time-limited access to palatable food induced BE intake in both male and female mice, which was not significantly altered by either the frequency of exposure to HFHS or a history of caloric restriction. However, the combination of caloric restriction and intermittent exposure to HFHS (i.e. the CR + IB group) induced the earliest response in females, starting in cycle 1. Moreover, our results revealed no compensatory decrease in CT chow intake for either male or female mice in this paradigm, preferentially supporting the use of such conditions to induce an overeating phenotype in BE models. Features influencing binge eating models We used a protocol in which mice were exposed to a HFHS diet for 2 h during their inactive circadian phase. This time frame reflected two diagnostic criteria for a binge episode as offered by the DSM-5: eating in a discrete time period (i.e. within any 2 h period) and when not feeling physically hungry [1]. Moreover, the experimental groups were compared against a group of mice that had ad libitum access to both the CT and the HFHS (CA group), controlling for the consumption that an animal would typically eat if the diet was accessible without any restriction in time [25]. As expected, compared with the mice in the CA group, the mice in the BE groups displayed much higher values of HFHS intake in a sated state, and this pattern of excessive intake of palatable food was observed in both male and female mice. Thus, this model is in agreement with the previous literature, confirming the binge-inducing effect of limited access to palatable food [27, 40, 41]. To assess the impact of the frequency of exposure, access to HFHS was proposed daily (DB) or every 48 h (IB). In the literature, several studies have reported that intermittent (nondaily) models increase the amount of food eaten by rodents during the test period, and the more spaced in time the access to palatable food is, the more pronounced is the BE intake compared with that of animals with daily access or ad libitum access [29, 42–44]. However, we did not observe any differences in HFHS consumption between animals of either sex that were subjected to daily (DB) or intermittent (IB) BE tests. Although surprising, several aspects could explain this outcome. First, some studies allowed access to the target food closer to or during the dark phase, which could increase the drive to eat. Nonetheless, consistent with our protocol design, a recent meta-analysis revealed that more than 80% of BE models exposed the experimental subjects to the target diet while in a sated state, reinforcing our methodological choice [25]. From our perspective, performing the BE test during the inactive phase highlights two additional strengths of our study: first, HFHS intake is driven mainly by pleasure and not physiological needs because the mice are in a sated state, as confirmed by the absence of CT food intake during the BE access period in all groups; second, the comparison of HFHS intake during the light phase between the BE groups and the CA group allows us to confidently assume that our experimental groups developed a BE response. Additionally, other features of eating behaviour, such as intake speed, could be explored to address the apparent lack of differences between DB and IB. It has been reported that during 2-h periods, most of the intake takes place during the first hour; therefore, we speculate that a different frequency of exposure to palatable food (DB vs. IB) could lead to changes in the microstructure of the feeding patterns. Another interesting possibility would be to modulate the predictability instead of the frequency of access at similar timepoints for every session, as the unpredictability of access has been reported to increase the BE response [45]. The second factor we evaluated was the history of dietary restraint modelled by episodes of caloric restriction. Importantly, to avoid a confounding factor on hunger-driven BE intake, the animals were re-fed ad libitum prior to the BE test. In our study, compared with non-restricted mice, animals of both sexes subjected to caloric restriction did not display significantly greater HFHS consumption. We hypothesize that under our experimental conditions, compared with non-restricted conditions, 3 days of caloric restriction was not sufficient to induce higher BE intake compared with non-restricted conditions because the animals were fully re-fed before the BE test days. However, we found that previous exposure or not to caloric restriction influenced the 24-h chow intake in male mice. Conversely, the 24-h chow intake was not modulated by the history of previous caloric restriction but by the frequency of BE exposure in female mice, with a decrease in mice exposed to daily BE. The escalation of palatable food consumption has been frequently used to characterize the development or worsening of BE intake. Interestingly, under our experimental conditions, males showed a positive escalation, with a significantly greater HFHS intake in cycle 4 than in cycle 1, but such a change was not as clear in female mice. Indeed, only female mice exposed to daily binge exhibited an escalation pattern between cycles, but mice exposed to caloric restriction showed a negative trend towards escalation. This phenomenon might be interpreted as a habituation to the binge-triggering effects of caloric restriction or, conversely, as a rapid installation of intense BE intake in females. Another possibility for the general lack of escalation in female mice is the effect of the hormonal balance of the female cohort. The phase of the oestrous cycle in female mice was not monitored throughout the experiment. Although some studies have shown an effect of the oestrous phase on palatable food binge-like intake in female animals [46], others have reported contradictory results, with no effect of ovariectomy on binge-like eating in females [21]. Therefore, we cannot rule out an effect of the hormonal state of female mice on the behavioural outcomes observed in the present study. That said, while escalation of palatable food intake has been used in some studies to assess the development of binge-like behaviour in rodent models, it does not provide a strong translational trait and does not resemble any core characteristic of human pathology. Indeed, even if escalation has classically been used in studies on addictive behaviour, mimicking the tolerance pattern in substance use disorder (i.e. decreased responses to the same amount of substance and a gradual increase in consumption), the escalation criterion has been suggested to be more appropriate for exploring neurobiological substrates in addictive behaviour than in BE[47]. Moreover, the strength of our model relies on the comparison with a control group where exposure to the palatable food occurs ad libitum (CA group) that does not develop binge-like intake. Indeed, it has been proposed in the literature [25] that assessing escalation in intake as the main binge-defining criterion without a non-binge control does not provide a robust binge behaviour analysis. Taken together, palatable food intake seems to be a more robust measure of binge-like intake than escalation, which has methodological and translational limitations. Dopaminergic substrates implicated in binge eating disorder In the brain, food intake is regulated by the interaction of two major interconnected systems: the homeostatic system, which is controlled mainly by hypothalamic circuits and responds to energy need/expenditure balance, and the hedonic system, which is integrated mainly by the mesocorticolimbic pathways and driven by the palatability and rewarding value of food [48], acting through projections from the ventral tegmental area (VTA) to the nucleus accumbens (NAc), the dorsal striatum and the prefrontal cortex [17]. High-fat and/or high-sugar food affects neurotransmission signalling systems in reward-related brain regions, but the activated pathways may differ depending on the type of food [49, 50]. Exposure to a sweet-fat diet may promote synergic effects on brain reward circuits that lead to both behavioural alterations (i.e., increased palatable food intake) and changes in neural transmission, such as dopamine release [49, 51]. In obesity, hypofunction of the dopaminergic system in the reward circuit seems to play a relevant role in the development of eating habits, where an overconsumption of palatable food is promoted to compensate for a diminished dopamine response [52–54]. In patients with BED, alterations in dopaminergic markers within the reward circuit have been described but remain inconsistent [9]. For instance, a genetic study in patients with obesity and comorbid BED revealed an increased frequency of the A2 allele of the Taq1A single-nucleotide polymorphism of the ankyrin-containing kinase 1 gene (a neighbouring gene of Drd2), which was associated with increased striatal DRD2 expression compared with that in patients without BED [55]. However, another genetic study revealed an association between the development of BED after bariatric surgery and the presence of allelic variants of the gene encoding DRD2, leading to decreased expression of the receptor in the mesolimbic pathway [56]. Using magnetic resonance imaging, a recent study of patients with BED demonstrated a decreased binding potential of D2/3 dopaminergic receptors and plastic alterations in the grey matter microstructure in the putamen, a striatal subregion [57]. Certainly, some preclinical studies have reported changes in dopamine receptor 2 expression in reward-related brain regions after exposure to a BE paradigm, which may be region specific. For example, the expression of both genetic and proteic DRD2 in the NAc has been shown to decrease in binge-subjected rats [58–60]. In contrast, the expression of this receptor was elevated in the caudate putamen and in the VTA after BE [60]. Thus, evidence is accumulating regarding an altered dopaminergic state associated with the development of BE conditions, but the nature of these changes may vary depending on the assessment technique (genetic, imaging, or pharmacological), the brain area investigated or the coexistence of comorbidities such as obesity or other mental diseases. In our study, although not significantly different, the expression of Drd2 in the striatum remained higher in all BE groups than in the CA group in both male and female mice. This difference was due mainly to a reduction in Drd2 mRNA expression in the CA group compared with that in the NA group. Indeed, exposure to HFHS provided ad libitum appeared to promote a reduction in Drd2 expression, much like it did in obesity, whereas this decrease was absent when access to palatable food occurred in a binge-like pattern. Hence, we speculate that unmodulated Drd2 expression following limited palatable food exposure might increase HFHS intake and promote reward-driven BE behaviour since it has been described that striatal dopaminergic release in anticipation of a reward is crucial for modulating reward-oriented behaviour [61]. Notably, in addition to its postsynaptic location, DRD2 also functions as an autoreceptor that regulates dopamine release from the presynaptic terminal [62]. Therefore, increased Drd2 expression in the striatum may also result in lower levels of extracellular dopamine, which could promote overeating to compensate for potentially diminished food-induced reward signalling. This hypothesis aligns with the mechanism of lisdexamfetamine, the only approved treatment for BED. By blocking dopamine reuptake, lisdexamfetamine increases extracellular dopamine levels [63], which is consistent with a possible hypodopaminergic state in BED. Nonetheless, we observed high variability within subjects, both in gene expression in the striatum and in the BE response. Interestingly, Chawla and colleagues [59] reported variability in Drd2 expression in BE protocols, with a decrease in Drd2 expression in binge-prone rats when compared with resistant animals. This finding indicates that specific alterations in the Drd2 gene may vary between subpopulations in terms of their vulnerability to BE. However, we did not identify such a correlation between behavioural and neurobiological changes in our study. Conversely, although genetic expression provides valuable information, we cannot directly infer any alterations in the dopaminergic tone, for which we would need to evaluate the dopamine release or electrical activity of dopaminergic neurons in the striatum. Finally, we did not identify any robust changes in the glutamatergic or GABAergic systems, but continuous exposure to palatable food induced a change in the expression of subunit 2a of the NMDA receptor that was not induced by BE protocols, similar to the effects observed for dopaminergic markers. Therefore, the impact of BE on reward signalling and alterations in these neurotransmission pathways seems to rely not only on the type of food but also on the alternation between exposure, anticipation and non-exposure. Limitations of the study Several limitations should be addressed. First, the small sample size of animals used in this study may mask more conclusive results and does not allow the distinction of behavioural phenotypes regarding proneness/resistance to BE intake. We are confident that our behavioural observations are strong enough to draw our main conclusions, as animals develop clear BE behaviour. Nonetheless, interpretation of our results regarding the role of dopaminergic markers should be undertaken with caution because of the high variability. Second, because of the variety of schedules of access to palatable food, experimental animals were not equally exposed to HFHS for the same total duration between experimental groups, which may result in different behavioural and systemic alterations. Third, this study may benefit from the analysis of other eating parameters, such as the latency to eat. Finally, the characterization of BE is limited to palatable food intake over a short period of time but does not assess any other behavioural traits linked to BED, such as anxiety-like, compulsive or impulsive behaviours. Despite these limitations, this work provides a comparison of the same experimental protocol with different combinations of binge-inducing factors in a limited-access model, namely, the intermittency of access to palatable food and prior episodes of caloric restriction. We showed that 2 h of limited access to HFHS could induce BE intake in male and female mice compared with animals with ad libitum access to palatable food and that the combination of caloric restriction and intermittent access best mirrors the clinical features of the disease since it did not induce a compensatory decrease in CT diet intake. In addition, compared with ad libitum HFHS access, BE intake may be associated with different modulations of Drd2 expression in animals subjected to BE, opening the way for further research on the characterization of dopaminergic involvement in the development of BED. Conclusions Male and female mice develop a BE behavioural phenotype when exposed to HFHS in a 2-h-limited manner. In contrast to expectations, neither intermittency of access to the palatable diet nor a history of caloric restriction accentuated the binge-like intake of HFHS, but the combination of caloric restriction and intermittent exposure most closely reflected the clinical situation in terms of early response and the absence of compensatory mechanisms. That said, we highlighted that binge exposure to HFHS might entail differential regulation of the dopaminergic system in areas of the reward system of the brain, potentially suggesting a key role of this system in the neural imprint of BED. Supplementary Information Supplementary material 1. Acknowledgements We thank Bouazza Es Saadi, Anthony Castellano and Marine Olivier (UCLouvain, Brussels) for their excellent support and assistance. Abbreviations Binge eating Binge eating disorder Continuous access Caloric restriction + daily binge Caloric restriction + intermittent binge Control diet Daily binge Dopamine receptor 1 Dopamine receptor 2 Diagnostic and Statistical Manual of Mental Disorders Subunit alpha-1 of GABAA receptor Subunit beta-1 of GABAA receptor Subunit beta-2 of GABAA receptor Glutamate decarboxylase Subunit 1 of AMPA receptor Subunit 2 of AMPA receptor Subunit 2 A of NMDA receptor Subunit 2B of NMDA receptor High-fat high-sucrose Intermittent binge No access Nucleus accumbens Ribosomal protein L19 Excitatory amino acid transporter 2 Dopamine transporter GABA transporter 3 Tyrosine hydroxylase Ventral tegmental area Author contributions AE: conceptualization, funding acquisition, methodology, project administration, resources, supervision, validation, visualization, writing–original draft preparation, writing–review and editing; CF: formal analysis, investigation, methodology, writing–review and editing; CTM: writing–review and editing; NMP: formal analysis, investigation, methodology, validation, visualization, writing–original draft preparation, writing–review and editing. All the authors read and approved the final manuscript. Funding Amandine Everard is a senior research associate at Fonds de la Recherche Scientifique - FNRS and a recipient of grants from FNRS under Grant n° T.0115.24 and FNRS FRFS-WELBIO under the Grants n° WELBIO-CR-2019S-03E and WELBIO X.1517.24. Data availability The datasets generated and/or analysed during the current study are available from the corresponding author upon reasonable request. Declarations Competing interests Amandine Everard is inventor on patent applications dealing with the use of specific bacteria and components in the treatment of different diseases. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary material 1. Data Availability Statement The datasets generated and/or analysed during the current study are available from the corresponding author upon reasonable request.