cbd oil for treating binge eating disorder

Cannabis for Eating Disorders

Can medical marijuana help people with eating disorders like anorexia and bulimia? Cannabis is often used to increase appetite in people suffering from conditions such as cancer or AIDS/HIV. The idea of using cannabis for those with eating disorders is not exactly new, and in many ways the logic is entirely sound. However, eating disorders have several key differences to wasting developing from other chronic illnesses, meaning that treating them requires slightly different approaches.

Using cannabis for eating disorders is a subject we have written about before here at Leafwell. Today we’ll learn all about the potential of cannabis as a medication for a variety of eating disorders.

What is an eating disorder?

One of the key defining symptoms of an eating disorder is an unhealthy attitude towards food and eating either too much or too little. Other symptoms include an unhealthy obsession with weight and body shape, over-exercising, obsessive dieting, binge-eating (sometimes followed by intentional vomiting, or "purging"), extreme dissatisfaction with one’s own appearance (Body Dysmorphic Disorder, or BDD) depression, anxiety and extreme feelings of guilt, regret and/or worthlessness.

In some instances, an eating disorder may lead to "refeeding syndrome", which is when malnourished or starved people take in food too quickly after a fasting period and develop electrolyte disorders. This leads to further pulmonary, cardiac, neuromuscular and blood complications. Refeeding syndrome can be potentially fatal. Other long-term complications include increased likelihood of stress fractures and Raynaud syndrome.

There are various types of eating disorders, including:

    • Anorexia nervosa – keeping your weight as low as possible by purposefully not eating enough food, exercising too much or both.
    • Binge eating disorder (BED) – losing control of your eating and eating too much at once, until you are uncomfortably full. Often followed by feelings of guilt and regret.
    • Bulimia – Binge eating in a small amount of time, then deliberately feeling sick, using laxatives or exercising too much in order to prevent weight gain.
    • Obesity – While not always considered an "eating disorder", obesity does follow many of the same patterns as other eating disorders, including binge eating and an unhealthy relationship with food. Indeed, it is not unheard of for a person to swing between anorexia and obesity.
    • Other specified feeding or eating disorder (OSFED) – an eating disorder that doesn’t necessarily match all the symptoms of one of the above, and/or has "mixed" symptoms from one or more of the above. OSFED can include atypical anorexia, avoidant/restrictive food intake beyond that of "picky eating", night eating syndrome, anorexia athletica and eating disorders related to type-I diabetes (e.g. deliberate insulin under use in order to prevent weight gain).

    Some statistics on Eating Disorders

    Eating disorders affect approximately 30 million people in the US. They have the highest mortality rate of any mental illness, and are often comorbid with mood disorders, anxiety disorders and substance misuse disorders (especially alcohol).

    Eating disorders affect a wide variety of people. Women aged 50 or over, girls aged between 13 – 17 and women in high-pressured environments such as athletics are the highest risk groups for anorexia and bulimia. Restrictive eating is more likely to be found in boys and men. A 2015-2016 study by the Center for Disease Control and Prevention (CDC) showed that 39.6% of US adults age 20 and older were obese as of 2015-2016 (37.9% for men and 41.1% for women). Other risk factors include:

      • to under- and overfeeding of the fetus during pregnancy. Maternal obesity and malnutrition play a huge role in the development of eating disorders among offspring.
      • Adiposity rebound – the "adiposity rebound" refers to the age when the second rise in body-mass index (BMI) occurs, which is between 3 – 7 years old. An early age adiposity rebound is correlated with obesity in later life.
      • Early life malnutrition and/or lack of breastfeeding – early nutrient deprivation can lead to a change in the body’s metabolism, leading to fat storage. This can make people vulnerable to obesity as adolescents and adults. Those who are not breastfed may suffer from stunted growth or grow too fast, leading to an earlier-onset adipose rebound. This is one reason why malnutrition, a lack of access to food and obesity are often linked and found together in impoverished parts of the world.
      • Steroid-based medications such as prednisone can lead to weight-gain.

      What is Cachexia?

      While cachexia (which means, "weakness and wasting of the body due to severe chronic illness") is often associated with conditions such as anorexia, a person who is suffering from cachexia is not necessarily suffering from an eating disorder. Cachexia can be caused by many illnesses and conditions, as well as treatments and medications. Many of those with cachexia may well have a perfectly fine relationship with food, but are unfortunate enough to suffer from a condition that causes them to lose weight and muscle.

      Anorexia Athletica

      Those in highly competitive environments such as sports and athletics, where extreme fastidiousness is practiced with regards to diet and exercise, eating disorders are not uncommon. Many athletes also need a high intake of calories, meaning they need to learn portion control when training slows down or ceases. Athletes of all types can potentially suffer from eating disorders.

      Even boxers and wrestlers, who are considered some of the strongest athletes in the world, often dehydrate, starve and over-exert themselves in order to make weight, which can lead to all sorts of health problems. Gymnasts, dancers, figure skaters, weightlifters, bodybuilders, synchronized swimmers, and endurance runners are other examples of athletes who may suffer from eating disorders due to the emphasis on weight and appearance.

      How Does Cannabis Help Eating Disorders?

      When it comes to using cannabis for conditions such as anorexia, people see the logic quite easily. However, when it comes to obesity (as well as diabetes), people find the concept of using cannabinoid-based medications to help treat it unusual. Yet, regular use of cannabis is actually linked to lower BMI, even when controlling for diet, exercise and alcohol consumption. While these studies do not prove for sure that cannabis use can help people maintain a healthy weight, there are several sound theories as to why cannabinoids may be used to help maintain a healthy appetite for both over- and under- eaters. These include:

      The endocannabinoid system (ECS) plays a role in regulating appetite. Cannabinoids such as tetrahydrocannabinol (THC) stimulate appetite and food intake.

      Download Free Guide to the ECS
        • There is some suggestion that those who suffer from eating disorders have a disruption and/or dysregulation in the production of the hormones leptin (which can regulate energy balance by inhibiting hunger) and ghrelin (the "hunger hormone", which stimulates appetite).
        • Cannabis use in HIV-infected men leads to an increase in plasma levels of ghrelin and leptin. THC in particular seems to have this effect.

        Repeated exposure to THC may initially stimulate appetite, but use over the long-term could dampen CB1 receptor sensitivity, thus dampening hunger signals.

        Some suggest that cannabis "supercharges" the body’s metabolism, meaning that fat is burnt off faster and levels of fasting insulin are lower. The body may be more sensitive to the effects of sugar while using cannabinoids, meaning that the brain sends signals to stop eating sooner than it usually would. So, while cannabis users may get the "munchies", they may also have a tendency to stop eating sooner and only until they are full, rather than over-full.

        There is much interest in the cannabinoid tetrahydrocannabivarin (THCV) for obesity and diabetes. THCV is a CB1 receptor antagonist, meaning that it has the opposite effect as THC when in low doses (THCV is a CB1 receptor agonist in high doses) and curbs hunger. In studies on mice, researchers found that THCV did not significantly affect food intake or body weight gain. THCV did, however, reduce glucose intolerance and improve insulin sensitivity. Such studies could offer hope to diabetics, but research on humans is necessary before making any assertions.

        Cannabidiol (CBD) can also help control blood-sugar levels and reduce the production of fat while also reducing inflammation caused by insulin resistance.
        Cannabis can potentially help with the depression and anxiety often associated with eating disorders. In turn, this may lead to an easier, less stressful relationship with food.

        Are There Any Potential Negatives with Using Cannabinoids for Eating Disorders?

        While cannabis can help improve the mood for many, for some using too much THC may lead to increased anxiety or paranoia. Also, if a person has been starving themselves for too long, care must be taken not to binge on food, lest refeeding syndrome occurs. Some may also be attracted to the idea that cannabis can help lose weight, which is beneficial for some but not necessarily others. Therefore, care must be taken to prevent misuse.

        Those suffering from eating disorders such as anorexia or bulimia may be interested in low doses of THC and CBD, whereas those who are obese (or just plain overweight) may look into a combination of low doses of THC and THCV, combined with CBD. However, this is only theoretical, and has not been tested clinically. As there are few effective medications for eating disorders, cannabinoids represent an extremely promising avenue to look at as a potential therapeutic target.

        There has been a look into other cannabis-based treatments for obesity in the past, namely Rimonabant. However, Rimonabant was not approved for usage due to its psychiatric side effects. Rimonabant has also been reported to cause partial seizures in those who suffer from epilepsy. It must also be noted that Rimonabant is a synthetic cannabinoid. We here at Leafwell have looked at the pros and cons of synthetic cannabinoids before, and as such we recommend being highly cautious of using them.

        Remember: the endocannabinoid system is very powerful, and our efforts to replicate the safety margins of phytocannabinoids have generally not been successful so far. In short, the natural form of the cannabis plant is probably best for eating disorders and other conditions.

        If you are suffering from an eating disorder and think you may be helped by cannabinoid-based medications, feel free to check out our medical card page and set up an appointment with one of our physicians.

        Article written by

        Tina Magrabi Senior Content Writer

        Tina Magrabi is a writer and editor specializing in holistic health. She has written hundreds of articles for Weedmaps where she spearheaded the Ailments series on cannabis medicine. In addition, she has written extensively for the women's health blog, SafeBirthProject, as well as print publications including Destinations Magazine and Vero's Voice. Tina is a Yale University alumna and certified yoga instructor with a passion for the outdoors.

        Keep updated with our social media

        Leafwell HQ
        Phone: +1 (800) 660-9085

        ©2022 Leafwell. Note: Information on this site does not constitute medical advice or legal advice.

        Impact of Marijuana Legalization for Eating Disorder Sufferers

        For those struggling with anorexia and bulimia, there is some evidence that there may be an imbalance in brain chemistry that relates to there internal marijuana neurotransmitter system. This system in the brain is called the endocannabinoid system.

        Somehow, the functionality of the endocannabinoid system is adversely affected or impaired in those with anorexia and bulimia [1].

        Marijuana is a species of the Genus – Cannabis L which is the same Genus as hemp. [8]. It contains more than 400 chemicals, but researchers tend to know the more about THC and Cannabidiol. THC is responsible for stimulating the cannabinoid receptors in the brain.

        The stimulation then triggers other chemical reactions that underlie marijuana’s psychological and physical effects [5]. Cannabidiol is not as well known or understood as THC, but research suggests that it interacts with THC to produce sedation. However, independently, it may have anti-inflammatory, neuroprotective, or antipsychotic effects.

        Impact of Marijuana on the Brain

        The regulation of appetite and feeding behaviors involves many processes, but there is extensive evidence that suggests the endogenous endocannabinoid system plays an essential role in signaling rewarding events, such as eating or restricting [1].

        There was a study conducted at the Katholieke Universiteit Leuven in Belgium where researchers used PET (positron emission tomography) imaging to look at the endocannabinoid system in the brains of 30 women with anorexia and bulimia.

        Results showed that the brain’s marijuana-like neurotransmitter system was significantly underactive in participants with anorexia and bulimia. What was observed was the underactive neurotransmitter was in the area of the brain known as the insula. The insula is responsible for the integration of the taste of food with our emotional response to eating [1].

        The physical characteristics of food, such as taste, flavor, and oral texture, as well as how hungry we feel is all integrated with the insula, and the insula also influences food’s rewarding properties.

        There is a wide range of sensations that play a role in how we feel. Physical information such as pain, temperature, sensual touch, stomach pH, and intestinal tension (such as constipation) is processed in the insula.

        Integration of these internal feelings provides a blended sense of the state of the entire body. The insula is the place in the brain where our sensory experience (from eating), our emotions (in response to eating) and thoughts (about why should not be eating!) come together [1].

        The amount of pleasure we derive from sensory experiences is usually controlled by the endocannabinoid system in our brain. The pleasure that is felt then motivates us to repeat the experience again and again.

        An obsessive interest in food coupled with an inappropriate emotional response is consistent with a dysfunction in the brain’s endocannabinoid system. This new information might help identify new targets for medications that may help reverse the symptoms of anorexia and bulimia [1].

        A study conducted by Tamas Horvath of New Haven’s Yale University discovered that our CB1 receptors might hold the answer. CB1 receptors are located in all of our body’s cells, and when they are activated with dronabinol (an anti-nausea drug and a component of cannabis), the CB1 receptors stimulated the release of hormones promoting hunger. Suppressing this activation may also result in weight loss [3].

        Eating Disorders and Co-Occurring Addiction

        The effect of marijuana legalization could be profound for those struggling with eating disorders and co-occurring issues. Observational studies suggest that one in nine people who smoke marijuana regularly become dependant on it [5].

        For many people who smoke marijuana, the THC often increases anxiety and panic attacks. Studies have reported that 20-30% of recreational users experience problems with increased anxiety. This is especially true for those who have not smoked marijuana before. [5]

        Individuals who have bipolar disorder and use marijuana seem to induce manic episodes and rapid cycling between manic and depressive moods increases. Marijuana use can also exacerbate psychotic symptoms and worsen outcomes in patients previously diagnosed with psychotic disorders. [5]

        Medical marijuana could aid in the refeeding process of an anorexic patient due to it being less traumatizing than a feeding tube. This allows the patient to chose to eat rather than being forced to eat.

        Marijuana could also play a role in the later stages of eating disorder recovery through use in relation therapy and exploring new ideas and insights.

        In a longitudinal study of 94 AIDS patients, the use of THC in doses ranging from 5-20 mg, confirmed that appetite was enhanced and patients tended to retain stable body weight over the course of the seven-month study. [2]

        Marijuana Legalization Across the United States

        The number of areas in the United States where marijuana is being legalized for medicinal use is rapidly growing. Rarely are eating disorders seen as a valid reason to issue or prescribe marijuana for therapeutic use.

        At least 24 million Americans have an eating disorder, with at least half of them meeting the criteria for depression and anxiety [4].

        In 2011, a study in Biological Psychiatry found a link between anorexia and bulimia with that of a potential brain malfunction that leads to a loss of the endocannabinoids.

        Marijuana is legal in Canada, and a review by a team of researchers found that only 31 studies focused on the medical benefits of the drug [5]. The American Medical Association also concluded that the research in the area of marijuana medical benefits seems to be limited.

        The Institute of Medicine (IOM) stated that with the widespread availability of FDA approved medications to help treat pain relief (particularly nerve pain), appetite stimulation for people with AIDS wasting syndrome or eating disorders, and control of chemotherapy-related nausea and vomiting, marijuana should only be considered for treatment when patients do not get relief from currently available medicines [5, 7].

        New studies evaluating the use of marijuana as a treatment for psychiatric disorders are inconclusive because the drug may have contradictory effects in the brain depending on the dose of the drug and inherent genetic vulnerability.

        In conclusion, marijuana can have various benefits to those struggling with eating disorders, but further research and clinical studies need to be done to see what effect it can have on the treatment of eating disorders.

        About the Author: Libby Lyons, MSW, LCSW, CEDS is a specialist in the eating disorder field. Libby has been treating eating disorders for 10 years within the St. Louis area, and enjoys working with individuals of all ages.

        [1] https://www.psychologytoday.com/blog/your-brain-food/201204/the-connection-between-anorexia-bulimia-and-marijuana
        [2] https://www.medicalmarijuana.com/medical-marijuana-treatments-cannabis-uses/eating-disorders-anorexia-and-medical-marijuana/
        [3] https://cannabis.net/blog/how-to/can-cannabis-help-treat-eating-disorders
        [4] http://www.psyweb.com/lifestyle/eating-disorders/marijuana-and-eating-disorders
        [5] http://www.health.harvard.edu/mind-and-mood/medical-marijuana-and-the-mind
        [6] http://thescienceexplorer.com/brain-and-body/smoking-weed-may-help-treat-eating-disorders
        [7] http://www.healthline.com/health/eating-disorders/binge-eating-disorder-and-marijuana#2
        [8] https://plants.usda.gov/java/ClassificationServlet?source=display&classid=CASA3

        The opinions and views of our guest contributors are shared to provide a broad perspective of eating disorders. These are not necessarily the views of Eating Disorder Hope, but an effort to offer discussion of various issues by different concerned individuals.

        We at Eating Disorder Hope understand that eating disorders result from a combination of environmental and genetic factors. If you or a loved one are suffering from an eating disorder, please know that there is hope for you, and seek immediate professional help.

        Pharmacological modulation of the endocannabinoid signalling alters binge-type eating behaviour in female rats

        4 Preclinical Pharmacology Section, Behavioral Neuroscience Research Branch, Intramural Research Program, National Institute on Drug Abuse, National Institutes of Health, Department of Health and Human Services, Baltimore, MD, USA

        C Dessi

        1 Department of Biomedical Science, Section of Neuroscience and Clinical Pharmacology, University of Cagliari, Monserrato (Cagliari), Italy

        W Fratta

        1 Department of Biomedical Science, Section of Neuroscience and Clinical Pharmacology, University of Cagliari, Monserrato (Cagliari), Italy

        2 Center of Excellence ‘Neurobiology of Addiction’, Monserrato (Cagliari), Italy

        P Fadda

        1 Department of Biomedical Science, Section of Neuroscience and Clinical Pharmacology, University of Cagliari, Monserrato (Cagliari), Italy

        2 Center of Excellence ‘Neurobiology of Addiction’, Monserrato (Cagliari), Italy

        Abstract

        Background and Purpose

        Binge eating disorder (BED) is characterized by excessive food intake during short periods of time. Recent evidence suggests that alterations in the endocannabinoid signalling could be involved in the pathophysiology of BED. In this study, we investigated whether pharmacological manipulation of endocannabinoid transmission may be effective in modulating the aberrant eating behaviour present in a validated rat model of BED.

        Experimental Approach

        Binge-type eating was induced in female rats by providing limited access to an optional source of dietary fat (margarine). Rats were divided into three groups, all with ad libitum access to chow and water: control (C), with no access to margarine; low restriction (LR), with 2 h margarine access 7 days a week; high restriction (HR), with 2 h margarine access 3 days a week.

        Key Results

        Compared with the LR group, the HR group consumed more margarine and this was accompanied by an increase in body weight. The cannabinoid CB1/CB2 receptor agonist Δ 9 -tetrahydrocannabinol significantly increased margarine intake selectively in LR rats, while the fatty acid amide hydrolase inhibitor URB597 showed no effect. The CB1 receptor inverse agonist/antagonist rimonabant dose-dependently reduced margarine intake in HR rats. Notably, in HR rats, chronic treatment with a low dose of rimonabant induced a selective long-lasting reduction in margarine intake that did not develop tolerance, and a significant and persistent reduction in body weight.

        Conclusions and Implications

        Chronic pharmacological blockade of CB1 receptors reduces binge eating behaviour in female rats and may prove effective in treating BED, with an associated significant reduction in body weight.

        Linked Articles

        This article is part of a themed section on Cannabinoids. To view the other articles in this section visit http://dx.doi.org/10.1111/bph.2013.169.issue-4 & http://dx.doi.org/10.1111/bph.2012.167.issue-8

        Introduction

        The endocannabinoid system, which comprises two cannabinoid receptor subtypes CB1 and CB2, their endogenous ligands (endocannabinoids, ECs) anandamide (AEA) and 2-arachidonoylglycerol, and corresponding biosynthesis and degradation pathways, is a homeostatic system involved in the regulation of several physiological functions, including cognition, reward, emotion, pain sensitivity and motor activity (Ameri, 1999; Solinas et al., 2008; Zanettini et al., 2011). There is a large body of evidence supporting the involvement of the endocannabinoid system in the modulation of energy balance by controlling food intake through central and peripheral mechanisms (Di Marzo and Matias, 2005). Systemic and local administrations of ECs and CB1 receptor agonists increase food intake in both animals and humans (Williams and Kirkham, 1999; Jamshidi and Taylor, 2001; Hart et al., 2002), whereas levels of ECs change during fasting and feeding in the rat hypothalamus and limbic brain areas (Kirkham et al., 2002). These hyperphagic actions are mediated by the CB1 receptor (Williams and Kirkham, 1999), as they can be selectively blocked by CB1 receptor antagonists, such as rimonabant and AM251 (Soria-Gòmez et al., 2007). In keeping with this, rimonabant and AM251 reduce food intake and feeding-associated behaviours in several rodent models (McLaughlin et al., 2003). Rimonabant also induces a significant decrease in food intake and body weight when chronically administered to either normal or obese rodents (Carai et al., 2006). Moreover, mice lacking the CB1 receptor are lean and hypophagic (Wiley et al., 2005), indicating that food intake and body weight gain depend upon the functional expression and activity of CB1 receptors. Conversely, an overactivity of the endocannabinoid system seems to be a key component in the pathophysiology of obesity (Di Marzo and Matias, 2005).

        For these reasons, CB1 receptor blockade has long been considered a potential pharmacological tool to restore the normal endocannabinoid system tone under this pathological condition (Harrold and Williams, 2003). Rimonabant was the first selective CB1 receptor ligand to be used clinically and was approved worldwide as a treatment for obese and overweight individuals with metabolic complications (Scheen, 2008; Despres et al., 2009). However, due to the occurrence of psychiatric side effects after prolonged use (European Medicines Agency, 2009), in 2009 rimonabant was withdrawn from the European market, but still represents a reference drug in the search for new therapies for obesity (Cervino et al., 2009).

        There is also converging evidence indicating that defects in the endocannabinoid system might be implicated not only in obesity but also in other eating disorders (Marco et al., 2011), such as binge eating disorder (BED). In the Diagnostic and Statistical Manual of Mental Disorders (fourth edition, text revision), BED is categorized as an ‘Eating Disorder Not Otherwise Specified’, and only listed in the appendix (American Psychiatric Association, 2000). Experts define ‘binge eating behaviour’ as an intermittent and uncontrollable consuming of an unusual amount of food (typically highly palatable foods rich in calories and with a high fat content) larger than that normally eaten by non-bingeing people in comparable periods of time. Binge eating episodes are not inevitably motivated by hunger or metabolic needs, and may or may not be followed by regular use of inappropriate compensatory behaviours, such as laxative use, vomiting, fasting, excessive exercise training, to purge the body of excess calories (American Psychiatric Association, 2000). Like other eating disorders, BED is more common in young people (Neumark-Sztainer et al., 2002) with a lifetime prevalence of 1.9 and 0.3% for women and men in six European countries (Preti et al., 2009) and of 3.5 and 2.0% for women and men in the USA (Hudson et al., 2007).

        BED is of public and clinical interest due to its association with emotional distress and physic health problems (Johnson et al., 2001) and with the risk of excessive weight gain and obesity (Yanovski et al., 1993). Co-morbidity of obesity or abnormal eating behaviours marked by binge eating (Hudson et al., 2007; Grilo et al., 2009) and substance abuse disorders is well known (Wiederman and Pryor, 1996; Davis and Claridge, 1998). Compelling parallelisms exist between compulsive overeating and drug abuse, among which are the presence of craving and the loss of the inhibitory control on food intake and drug use. Interestingly, there is neuroanatomical and neurochemical overlapping between food and drug craving, with similar factors triggering relapse to overeating and drug use. Indeed, if drug-taking behaviour can be re-instated in abstinent rats and humans by a single exposure to the previously abused drug (Schmidt et al., 2005), a sustained intermittent exposure to sugar solution induces patterns of food intake and neurotransmitter changes comparable to those observed in animal models of drug abuse (Kelley et al., 2005; Avena et al., 2008). It is widely recognized that the endocannabinoid system is greatly involved in food intake (Li et al., 2011), drug addiction (Serrano and Parsons, 2011) and relapse (Fattore et al., 2007).

        Overweight and obese subjects show increased frequency of the naturally occurring missense polymorphism of the gene that encodes the AEA hydrolyzing enzyme fatty acid amide hydrolase (FAAH), FAAH cDNA 385 A/A, which might potentiate the drive to eat by increasing endocannabinoid signalling (Sipe et al., 2005; Monteleone et al., 2008). On the other hand, a polymorphism of the CNR1 gene (encoding the human CB1 receptor) is thought to contribute to the vulnerability to anorexia nervosa (Siegfried et al., 2004). Moreover, women with anorexia nervosa or BED have elevated plasma levels of AEA (Monteleone et al., 2005), which could affect the rewarding aspect of eating that is compromised in these patients (Monteleone et al., 2008).

        Although clinical studies indicate some improvement in the pharmacological management of bingeing-related eating disorders (Marazziti et al., 2011), treatment options are quite restricted and relapse incidences are still high. Limited progress in the development of strategies for the treatment of BED was mainly due to the fact that the neurobiological bases for repeatedly engaging in binge-type behaviour are not fully understood, and are not easy to study in humans. Examination of what is known about analogous behaviours in animals would help to elucidate the core components of BED.

        In humans, accumulating data suggest that different situations may trigger compulsive eating leading to binge behaviour, psychological distress and dysphoria (Polivy, 1996). People typically binge on high palatable foods to which they have limited access (Kales, 1990; Fisher and Birch, 1999), after starvation and dietary restriction (American Psychiatric Association, 2000) or under stress (Oliver and Wardle, 1999).

        Binge eating behaviour can be modelled in animal protocols to investigate neurobiological substrates and pharmacological determinants of human bingeing disorders (Berner et al., 2011). All animal models of BED available to date are isomorphic, and as such they mimic the human symptomatology by inducing similar behavioural states, that is, compulsive overeating (Hancock and Olmstead, 2010).

        Based on previous studies showing that CB1 receptor blockade reduces feeding behaviour and the hedonic response to food, probably through the modulation of the mesolimbic dopaminergic pathway (Melis et al., 2007), this study was undertaken to investigate whether the CB1 receptor inverse agonist/antagonist rimonabant may be effective in modifying the aberrant eating behaviour present in a validated rat model of binge eating (Corwin et al., 1998). The effects of the natural CB1/CB2 receptor agonist Δ9-tetrahydrocannabinol (THC) were evaluated to compare the effects of CB1 receptor activation with that of CB1 receptor blockade. Furthermore, we tested the indirectly acting CB1 receptor synthetic agonist URB597 (a FAAH inhibitor) because it possesses anti-craving properties (Scherma et al., 2008b) and exerts anxiolytic-like and antidepressant-like effects in rodents (Piomelli et al., 2006; Gaetani et al., 2008; Scherma et al., 2008a) without evoking classical cannabinoid agonist-like effects (e.g. catalepsy, hypothermia, hyperphagia). Because craving as well as anxiety or depressive-like status may represent relevant behavioural traits related to BED, it was important to verify whether binge eating behaviour was affected by URB597 pretreatment.

        Several studies have evaluated the effect of chronic blockade of CB1 receptors on eating behaviour and have reported a significant reduction in food consumption exclusively at the beginning of the treatment, thus demonstrating a transient reducing effect (Colombo et al., 1998; Carai et al., 2006; Martín-García et al., 2010). Therefore, it was important to verify in our experimental conditions whether rimonabant was able to maintain its acute effect on binge eating behaviour over time or whether such an effect underwent tolerance development.

        Methods

        Animals

        Sprague–Dawley young adult female rats (Harlan Nossan, Udine, Italy) weighing 185–200 g at the start of the study (60–65 days old) were used. In our experiments, we used young female rats because in humans BED is more frequent in young females than in young males (Hudson et al., 2007). Following arrival, animals were individually housed in a climate-controlled animal room (21 ± 2°C temperature; 60% humidity) under a reversed 12 h light/dark cycle (lights on 24:00 h) with standard rat chow and water ad libitum. All experiments were approved by the local Animal Care Committee and carried out in strict accordance with the E.C. Regulations for Animal Use in Research (CEE No. 86/609). All studies involving animals are reported in accordance with the ARRIVE guidelines for reporting experiments involving animals (Kilkenny et al., 2010; McGrath et al., 2010).

        Diets

        Standard rat chow (Safe, France): 3% kcal from fat, 61% kcal from carbohydrate, 16% kcal from protein, 20% moisture, containing 2.9 kcal·g −1 . High-fat diet (Margarine, Gradina Unilever Italia Mkt.): 70% kcal from fat, <1% kcal from carbohydrate, containing 6.5 kcal·g −1 ).

        Experimental procedure

        Our protocol is based on previous work showing that limiting access to an optional source of dietary fat induces a binge-type eating behaviour in rats that can be maintained for long periods of time, even though animals are never deprived of food (Corwin et al., 1998; Corwin and Buda-Levin, 2004; Hancock and Olmstead, 2010). Animals were given access to high-fat food during their low activity period, that is, close to the time of switching lights off, as this does not interfere with their normal circadian rhythm, as previously described by Corwin and Wojnicki (2006).

        As illustrated in Figure 1 , following 1 week of acclimatization rats were divided into three groups matched for body weight, which correspond to the following diet conditions (that were maintained for the entire period of the experimental study):

        Schematic representation of the experimental design.

        Low restriction (LR): had continuous access to standard chow and water. In addition, animals were given 2 h access to a separate bowl of margarine introduced into the home cage every day of the week, 3 h before the start of the dark cycle (Corwin and Wojnicki, 2006).

        High restriction (HR): had continuous access to standard chow and water. In addition, animals were given 2 h access to a separate bowl of margarine introduced into the home cage every Monday, Wednesday and Friday, 3 h before the start of the dark cycle.

        Control (C): had continuous access to standard chow and water. Margarine was not provided at any time of the study.

        Margarine and/or standard chow were measured in all diet groups on Mondays, Wednesdays and Fridays (MWF) throughout the study by weighing them before and after the 2 h margarine access period ( Figure 1 ). Before the start of the experiments, both LR and HR groups were given overnight access to a bowl of margarine to prevent neophobia.

        Drug treatment commenced only once binge eating behaviour was firmly established (induction phase: 3–4 weeks). In our experiments, we included only animals that, at the end of the induction phase, displayed an intake of margarine within ±25% variation of the mean intake. Two separate batches of animals were used for acute and chronic treatments.

        Acute treatment

        Animals from each diet group (C, LR and HR, n = 24 per diet group) were randomly allocated into three different groups according to the pharmacological treatment assigned on the test day (Friday). Drug treatments (n = 8 per drug treatment group) were administered in a random sequence at weekly intervals. In keeping with previous studies (Koch JE, 2001; Fegley et al., 2005; Orio et al., 2009), a 1 week interval between each drug treatment proved to be a sufficient washout period. Treatments were the following: (i) vehicle, THC 0.125 and 0.250 mg·kg −1 ; (ii) vehicle, URB597 0.3 and 3 mg·kg −1 ; (iii) vehicle, rimonabant 0.3 and 3 mg·kg −1 . Drug injections were given i.p. 30 min before the margarine access period, with the only exception of URB597, in which pretreatment time was 40 min based on previous studies (Solinas et al., 2006; Scherma et al., 2008a). In all diet groups, margarine and/or chow were weighed before and after the 2 h access period on the test day.

        Chronic treatment

        Animals from each diet group (n = 20) were randomly assigned to two different groups, which received either rimonabant 0.3 mg·kg −1 (n = 10) or vehicle i.p. (n = 10). Drugs were administered once a day for 21 consecutive days, 30 min before the margarine access period. In both groups, margarine and/or chow were weighed on MWF before and after the 2 h access period. Body weight was recorded once a week on Fridays.

        Materials

        THC (RTI International, Research Triangle Park, NC, USA), 50 mg·mL −1 in ethanol, and rimonabant (National Institute on Drug Abuse, NIH, Baltimore, MD, USA) were dissolved in 2% Tween 80, 2% ethanol, and saline. URB597 (Cayman Chemical Company, Ann Arbor, MI, USA) was dissolved in 20% DMSO and saline. All drugs were injected i.p. in a volume of 1 mL·kg −1 .

        Data analysis

        Data from the induction of binge-type eating are expressed as mean kcal of margarine, chow and margarine + chow (total intake) (1-block week: MWF) ± SEM during the 2 h access period. Data were analysed by two-way anova for repeated measures with diet group and week as factors, and week as a repeated factor.

        Data from each acute treatment (margarine, chow and total intake) are expressed as mean kcal ± SEM during the 2 h access period on the test day, and were analysed by two-way anova with diet group and treatment as factors. The effects of treatment within each diet group were analysed by one-way anova as treatment between-subjects factor.

        Data from chronic treatment (margarine, chow and total intake) are expressed as mean kcal (1-block week: MWF) ± SEM during the 2 h access period, and were analysed by three-way anova with diet group, treatment and week as main factors, and week as a repeated factor. Significant differences within the diet group were further analysed by two-way anova with treatment and week as main factors and week as a repeated factor.

        Data from body weight during the induction phase of binge eating are expressed as mean in g ± SEM and were analysed by two-way anova with diet groups and week as main factors and week as a repeated factor.

        Data from body weight during chronic treatment were analysed by three-way anova with groups, treatment and week as main factors and week as a repeated factor. Significant differences within diet groups were further analysed by two-way anova with treatment and week as main factors and week as a repeated factor.

        Post hoc comparisons, when appropriate, were performed by Newman–Keuls multiple comparison test or by Bonferroni test. In all cases, differences with a P < 0.05 were considered significant. Student’s t-test was used when indicated and used to compare two conditions.

        Results

        Experiment 1. Acute drug treatments

        Induction of binge-type eating

        Consistent with previous findings (Corwin et al., 1998; Dimitriou et al., 2000), HR rats invariably consumed more margarine than LR rats during the 2 h limited access ( Figure 2 A). Two-way anova revealed a significant effect of diet group [F(1,138) = 12.77, P = 0.0008] and week [F(3,138) = 6.22, P = 0.0005], and a diet group × week significant interaction [F(3,138) = 3.78, P = 0.0121]. Post hoc analysis showed that margarine consumption was significantly greater in the HR than in the LR group by the third week of the study (3rd week: P < 0.001; 4th week: P < 0.01).

        Induction of binge-type eating. All data are presented as mean kcal (1-block week: MWF) ± SEM during the limited (2 h) access. (A) Margarine intake: HR group with limited access to margarine 3 days a week consumed more margarine than LR group with daily access to margarine (3rd week: $ P < 0.001; 4th week: # P < 0.01, Bonferroni post test). (B) Chow intake: LR group consumed less chow than C group (1st week: # P < 0.01; 2nd and 3rd weeks: $ P < 0.001, Student’s t-test) and HR group (2nd week: $ P < 0.001, Student’s t-test). HR group consumed less chow than C group only during the third week (*P < 0.05, Student’s t-test). (C) Total intake: during the limited access HR group displayed higher total intake than LR group (2nd week: # P < 0.01; 3rd and 4th weeks: $ P < 0.001, Bonferroni post test). In the C group, the total intake during the limited access was significantly less than that of both the LR and HR groups ( $ P < 0.001, Bonferroni post test).

        Notably, chow consumption during the 2 h limited access period was affected by the schedule of margarine availability ( Figure 2 B), since two-way anova revealed a significant effect of diet group [F(2,207) = 11.80, P < 0.0001] and week [F(3,207) = 7.46, P < 0.0001]. LR group consumed significantly less chow than C group during the first 3 weeks of the study [Student’s t-test: 1st week: t(46) = 2.814, P = 0.0072; 2nd week: t(46) = 4.387, P < 0.0001; 3rd week: t(46) = 4.234, P = 0.0001], but not on the last (4th) week. Moreover, LR rats consumed significantly less chow than HR rats on the second week only [Student’s t-test: t(46) = 3.614, P = 0.0007], while a significant difference between C and HR groups was found on the third week [Student’s t-test: t(46) = 2.459, P = 0.0178].

        When looking at the cumulative amount of margarine and/or chow consumed by each diet group during the 2 h limited access ( Figure 2 C), two-way anova detected a main effect of diet group [F(2,207) = 82.92, P < 0.0001] and week [F(3,207) = 4.33, P = 0.0055] and a diet group × week significant interaction [F(6,207) = 3.81, P = 0.0013]. Post hoc analysis showed that HR group displayed higher total intake during the limited access than LR group by the second week of the study (2nd week: P < 0.01; 3rd and 4th weeks: P < 0.001). In both LR and HR groups, the total intake during the limited access was significantly higher than that of C group (P < 0.001).

        Effect of THC on binge-type eating

        On the test day, margarine consumption during the 2 h limited access period was affected by treatment with THC (0.125 and 0.250 mg·kg −1 ) ( Figure 3 A). Two-way anova revealed a significant effect of diet group [F(1,42) = 5.66, P = 0.0220] and treatment [F(2,42) = 4.60, P = 0.0156]. Subsequent individual one-way anova showed that THC significantly increased margarine intake in LR group [F(2,21) = 7.112, P = 0.0044]. Compared with vehicle-treated rats, post hoc analysis revealed that both doses of THC were effective in increasing margarine intake in LR rats (+101% and +121%, respectively, P < 0.01). THC did not affect margarine intake in the HR group [F(2,21) = 0.367, P = 0.697].

        Effect of THC on binge-type eating. All data are presented as mean kcal ± SEM during the limited (2 h) access on the test day. (A) Margarine intake: THC doses of 0.125 and 0.250 mg·kg –1 increased margarine intake in the LR group ( # P < 0.01 vs. Veh, Newman–Keuls multiple comparison test). No effect was found in the HR group. (B) Chow intake: THC was effective in increasing chow intake in all groups (C group: *P < 0.05 vs. Veh; LR group: *P < 0.05 vs. Veh; HR group: # P < 0.01 and *P < 0.05 vs. Veh, Newman–Keuls multiple comparison test). (C) Total intake: THC was effective in increasing the total intake in both C and LR groups (*P < 0.05 vs. Veh, # P < 0.01 vs. Veh, Newman–Keuls multiple comparison test).

        On the other hand, THC also affected chow consumption during the 2 h limited access period, as two-way anova detected a significant effect of diet group [F(2,63) = 4.68, P = 0.0128] and treatment [F(2,63) = 14.41, P < 0.0001] ( Figure 3 B). Individual one-way anova revealed that THC significantly increased chow intake in all diet groups as compared with vehicle-treated rats [C group: F(2,21) = 4.258, P = 0.028; LR group: F(2,21) = 4.772, P = 0.0196; HR group: F(2,21) = 6.715, P = 0.0056, one-way anova ]. Post hoc analysis revealed that both doses of THC were effective in increasing chow intake by 73 and 65%, respectively, in C rats (P < 0.05), and by 96 and 72%, respectively, in the HR group (P < 0.01 and P < 0.05). Conversely, only the highest dose of THC (0.250 mg·kg −1 ) was effective in increasing chow intake in LR rats by 102% (P < 0.05).

        Finally, as shown in Figure 3 C, THC affected the total food intake during the 2 h limited access period (margarine + chow for LR and HR groups, and only chow for the C group) in both C [F(2,21) = 4.258, P = 0.028, one-way anova ] and LR [F(2,21) = 18.89, P < 0.0001, one-way anova ] groups. Post hoc analysis showed that both doses of THC were effective in increasing total intake as compared with vehicle-treated rats (P < 0.05 and P < 0.01 respectively).

        Effect of URB597 on binge-type eating

        On the test day, treatment with URB597 (0.3 and 3 mg·kg −1 ) did not produce significant changes on margarine intake during the 2 h limited access ( Figure 4 A). Two-way anova revealed a significant effect of diet group [F(1,42) = 11.83, P = 0.00013], but not treatment [F(2,42) = 0.15, P = 0.8628]. Although chow intake was affected by URB597 treatment [F(2,63) = 4.46, P = 0.0155, two-way anova ], one-way anova within each group did not show significant differences with respect to vehicle-treated rats during the 2 h limited access period [C group: F(2,21) = 2.218, P = 0.1337; LR group: F(2,21) = 1.546, P = 0.2365; HR group: F(2,21) = 1.629, P = 0.219, one-way anova ] ( Figure 4 B).

        Effect of URB597 on binge-type eating. All data are presented as mean kcal ± SEM during the limited (2 h) access on the test day. (A) Margarine intake: URB597 (0.3 and 3 mg·kg −1 ) did not alter margarine intake in both LR and HR groups as compared with vehicle-treated rats. (B) Chow intake: no effect was found on chow intake in all experimental groups. (C) Total intake: URB597 treatment did not modify the total intake.

        URB597 treatment did not modify the total food intake during the 2 h limited access period (margarine + chow for the LR and HR groups, only chow for the C group) either ( Figure 4 C), as revealed by individual one-way anova within each diet group [C: F(2,21) = 2.218, P = 0.133; LR: F(2,21) = 2.173, P = 0.138; HR: F(2,21) = 0.145, P = 0.865].

        Effect of rimonabant on binge-type eating

        Treatment with rimonabant (0.3 and 3 mg·kg −1 ) significantly affected margarine intake in both LR and HR groups on the test day ( Figure 5 A). Two-way anova revealed a significant effect of diet group [F(1,42) = 5.72, P = 0.0213] and treatment [F(2,42) = 14.18, P < 0.0001]. Individual one-way anova showed that both doses of rimonabant tested were effective in reducing margarine intake by 47 and 76%, respectively, as compared with vehicle-treated rats in HR group [F(2,21) = 10.82, P = 0.0006, (P < 0.01 and P < 0.001, post hoc test)]. Notably, the highest dose of rimonabant (3 mg·kg −1 ) also reduced (−41%) the intake of margarine in LR group as compared with vehicle-treated rats [F(2,21) = 3.709, P = 0.0418, (P < 0.05, post hoc test), one-way anova ]. Treatment with rimonabant also affected chow consumption during the 2 h limited access period [two-way anova , significant effect of diet group (F(2,63) = 4.38, P = 0.0166) and treatment (F(2,63) = 28.89, P < 0.0001)]. As shown in Figure 5 B, rimonabant decreased chow intake in all groups [C group: F(2,21) = 11.48, P = 0.0004; LR group: F(2,21) = 8.584, P = 0.0019; HR group: F(2,21) = 9.679, P = 0.001, one-way anova ] at both doses tested as compared with vehicle-treated rats (C group: −73 and −83% with P < 0.001; LR group: −65 and −95% with P < 0.05 and P < 0.01 respectively; HR group: −64 and −83% with P < 0.01, post hoc test).

        Effect of rimonabant on binge-type eating. All data are presented as mean kcal ± SEM during the limited (2 h) access on the test day. (A) Margarine intake: rimonabant doses of 0.3 and 3 mg·kg −1 decreased margarine intake in the HR group ( # P < 0.01 and $ P < 0.001 vs. Veh, Newman–Keuls multiple comparison test). The higher dose of 3 mg·kg −1 was also found to be effective in the LR group (*P < 0.05 vs. Veh, Newman–Keuls multiple comparison test). (B) Chow intake: both doses were effective in decreasing chow intake in all groups (C group: $ P < 0.001 vs. Veh; LR group: *P < 0.05 and # P < 0.01 vs. Veh; HR group: # P < 0.01 vs. Veh, Newman–Keuls multiple comparison test). (C) Total intake: rimonabant was effective in decreasing the total intake in all groups (C group: $ P < 0.001 vs. Veh; LR group: *P < 0.05 and # P < 0.01 vs. Veh; HR group: # P < 0.01 and $ P < 0.001 vs. Veh, Newman–Keuls multiple comparison test).

        Finally, rimonabant significantly decreased the total intake of food during the 2 h limited access period (margarine + chow for the LR and HR groups, and only chow for the C group) in all three diet groups as compared with vehicle-treated rats [C group: F(2,21) = 11.48, P = 0.0004; LR group: F(2,21) = 6.118, P = 0.0081; HR group: F(2,21) = 15.11, P < 0.0001, one-way anova ] ( Figure 5 C). Post hoc analysis revealed a significant drug effect at both the 0.3 and 3 mg·kg −1 doses (C group: P < 0.001; LR group: P < 0.05 and P < 0.01 respectively; HR group: P < 0.01 and P < 0.001 respectively).

        Experiment 2. Chronic rimonabant treatment

        Effect of chronic rimonabant on binge-type eating

        The effects of a chronic treatment (21 consecutive days) with rimonabant or its vehicle on margarine and/or chow intake during the 2 h limited access period were studied in a different set of animals showing binge-type eating behaviour that was induced in a similar manner to animals in Figure 2 (data not shown). Figure 6 A shows the effect of chronic rimonabant and vehicle on margarine intake. Three-way anova showed a significant effect of diet group [F(1,36) = 71.630, P < 0.0001] and treatment [F(1,36) = 71.630, P < 0.0001], but not week [F(2,36) = 0.7592, P = 0.4713], nor a significant interaction among these three factors. Two-way anova performed within each diet group also revealed a significant effect of treatment [LR: F(1,36) = 18.50, P = 0.0004; HR: F(1,36) = 21.75, P = 0.0002]. Both LR and HR groups treated with rimonabant consumed significantly less margarine compared with vehicle-treated rats during the 3 weeks of treatment [Student’s t-test: LR 1st week: t(18) = 4.518, P = 0.0003; 2nd week: t(18) = 4.816, P = 0.0001; 3rd week: t(18) = 2.622, P = 0.0173; HR 1st week: t(18) = 3.775, P = 0.0014; 2nd week: t(18) = 4.009, P = 0.0007; 3rd week: t(18) = 3.256, P = 0.0044].

        Effect of chronic rimonabant on binge-type eating. All data are presented as group mean kcal (1-block week: MWF) ± SEM during the limited (2 h) access. (A) Margarine intake: rimonabant (0.3 mg·kg −1 ) decreased margarine intake in both LR (1st and 2nd weeks: $ P < 0.001; 3rd: *P < 0.05, Student’s t-test) and HR (1st and 3rd weeks: # P < 0.01; 2nd week: $ P < 0.001, Student’s t-test) groups as compared with vehicle-treated rats. (B) Chow intake: chronic treatment with rimonabant significantly decreased chow intake only in C group during the first 2 weeks as compared with vehicle-treated rats (1st week: *P < 0.05; 2nd week: # P < 0.01, Student’s t-test). (C) Total intake: chronic rimonabant was effective in decreasing the total intake in all groups as compared with vehicle-treated rats (C group: 1st week: *P < 0.05; 2nd week: # P < 0.01, Student’s t-test; LR group: 1st and 2nd weeks: $ P < 0.001; 3rd: *P < 0.05, Student’s t-test; HR group: 1st and 3rd weeks: *P < 0.05; 2nd week: # P < 0.01, Student’s t-test).

        On the other hand, three-way anova showed that chow consumption was affected by treatment only [F(1,54) = 5.1504, P = 0.027], as no significant effect of diet group [F(2,54) = 2.5399, P = 0.088] or week [F(2,54) = 1.613, P = 0.203] was found, nor a significant interaction among these three factors ( Figure 6 B). Two-way anova performed within each diet group revealed a significant effect of treatment in the C group only [F(1,36) = 11.66, P = 0.0031]: rimonabant-treated animals consumed significantly less chow compared with vehicle-treated rats during the first 2 weeks of the treatment [Student’s t-test: 1st week: t(18) = 2.716, P = 0.0142; 2nd week: t(18) = 3.369, P = 0.0034], but not on the last (3rd) week.

        Finally, as shown in Figure 6 C, rimonabant also affected the total intake (margarine + chow for the LR and HR groups, and only chow for the C group) in all three diet groups; two-way anova performed within each diet group detected a main effect of treatment [C: F(1,36) = 11.66, P = 0.0031; LR: F(1,36) = 20.97, P = 0.0002; HR: F(1,36) = 9.78, P = 0.0058]. The total food intake in both LR and HR groups treated with rimonabant was less than in vehicle-treated rats during all the 3 weeks of treatment [Student’s t-test: LR 1st week: t(18) = 5.224, P < 0.0001; 2nd week: t(18) = 4.897, P = 0.0001; 3rd week: t(18) = 2.599, P = 0.0181; HR 1st week: t(18) = 2.826, P = 0.0112; 2nd week: t(18) = 3.247, P = 0.0045; 3rd week: t(18) = 2.748, P = 0.0132]. As mentioned above, rimonabant affected the total intake of the C group only during the first 2 weeks of treatment.

        Effect of chronic rimonabant on body weight

        As shown in Figure 7 A, significant changes in the mean body weight were detected during the induction phase. Two-way anova revealed a significant effect of diet group [F(2,171) = 4.13, P = 0.0212] and week [F(3,171) = 529.55, P < 0.0001], and a significant diet group × week interaction [F(6,171) = 4.13, P = 0.0007]. Post hoc analysis indicated that HR rats weighed more than C rats by the third week of the study (3rd week: P < 0.05; 4th week P < 0.01). No significant differences were found between HR and LR rats, nor between C and LR rats.

        All data are presented as group means ± SEM weekly change in body weight. (A) Induction of binge-type eating: HR group weighed more than the C group (3rd week: *P < 0.05; 4th week: # P < 0.01, Bonferroni post test). Not significant differences were found between HR and LR groups, nor between C and LR groups. (B) Effect of chronic rimonabant on body weight: HR rats chronically treated with rimonabant had a significantly decreased body weight as compared with corresponding vehicle-treated rats (*P < 0.05, Student’s t-test). Chronic treatment with rimonabant did not affect the body weight in either theC or LR groups when compared with corresponding vehicle-treated rats.

        The effects of chronic treatment with rimonabant or its vehicle are shown in Figure 7 B. Three-way anova showed a significant effect of diet group [F(2,54) = 4.72, P = 0.01294] and treatment [F(1,54) = 5.15, P = 0.0276] and week [F(2,54) = 51.38, P < 0.0001], but not a significant interaction among these three factors. Two-way anova performed within each group detected a main effect of treatment in the HR group only [F(1,36) = 5.58, P = 0.0296]. HR rats chronically treated with rimonabant significantly decreased body weight as compared with corresponding vehicle-treated rats [Student’s t-test: 1st week: t(18) = 2.612, P = 0.0176; 2nd week: t(18) = 2.372, P = 0.0290; 3rd week: t(18) = 2.566, P = 0.0194].

        Discussion

        The aim of our study was to verify if the pharmacological manipulation of the endocannabinoid system could be effective in the modulation of abnormal eating behaviour developed by female rats in a confirmed rat model of BED, in which binge eating behaviour is induced in animals by giving them a sporadic (3 days week -1 ) and limited (2 h) access to a high-fat diet (margarine) in addition to a continuous access to chow and water (HR group). In these animals, the intake of margarine becomes significantly greater than those of animals with limited daily access to margarine (LR group), and remains stable over prolonged periods of time (Corwin and Buda-Levin, 2004; Corwin and Wojnicki, 2006).

        As in other animal models of BED, in our limited access model binge eating was characterized by behavioural patterns similar to those seen in humans, as rats consumed a large quantity of food in a brief, defined period of time, and these quantities exceeded the amount typically consumed by control animals. Notably, our rats were never deprived of food, which is similar to bingeing humans who eat in the absence of hunger (Marcus and Kalarchian, 2003). Unfortunately, human subjective feelings of distress or loss of control found in some (Engel et al., 2007) but not in all bingeing subjects (Wegner et al., 2002) cannot be assessed easily in animals (Corwin and Buda-Levin, 2004). However, binge eating animal models exploit the precursory circumstances leading to binge eating in humans (e.g. dieting, exposure to palatable foods and fluids, stress) (Corwin and Buda-Levin, 2004).

        In this study, treatment with the natural CB1/CB2 receptor agonist THC proved to be effective in increasing margarine intake exclusively in the LR group, in line with the orexigenic effects of THC in humans and rodents (Williams and Kirkham, 1999; Hart et al., 2002), and with the finding that our control (C) animals, in which margarine was not provided at any time, ate more standard chow after THC injection.

        In LR rats, THC increased the total food intake with a specific effect on palatable food (Koch, 2001), where it was effective even at the lowest dose tested. However, THC did not affect margarine consumption in HR animals, although it significantly stimulated chow intake. We assume that the consumption of margarine was already the highest achievable in our bingeing animals (HR group), so that the CB1 receptor agonist was not able to increase it further, suggesting the possibility of an enhanced endocannabinoid tone in HR rats compared with the LR and C groups.

        Treatment with URB597, which prevents intracellular inactivation of AEA by FAAH inhibition and prolongs its behavioural and neurochemical effects (Kathuria et al., 2003), did not induce a significant increase in the amount of margarine or for chow consumed, although we observed a positive trend. In contrast to our findings, Soria-Gòmez et al. (2007) reported that local infusion of N-arachidonoyl-serotonin (AA5-HT), another FAAH inhibitor (Bisogno et al., 1998), into the nucleus accumbens shell markedly stimulated the ingestion of standard chow. This effect was prevented by AM251, suggesting the involvement of the CB1 receptor in the orexigenic responses to AA5-HT administration. The FAAH inhibitor AA5-HT, infused in the parabrachial nucleus, also increased the consumption of a palatable high-fat/sucrose diet, an effect mediated by CB1 receptors (Dipatrizio and Simansky, 2008). After injection of doses similar to those used in this study (0.3 and 3 mg·kg −1 ), FAAH inhibition was reported to be rapid (<15 min), persistent (>16 h) and correlated with a threefold increase in brain AEA levels (Kathuria et al., 2003; Fegley et al., 2005). The lack of response in our animals could be ascribed to the different route of administration used (i.p. rather than direct brain infusion), or to the fact that URB597 is unnable to further enhance the pre-existing endocannabinoid tone of our bingeing rats. Alternatively, the finding that the FAAH inhibitor had no effect in our bingeing animals can imply the involvement of other neurotransmitter systems. It should be noted that FAAH inhibition increases brain levels and magnifies and prolongs the effects of the non-cannabinoid fatty acid ethanolamides oleoylethanolamide (OEA) and palmitoylethanolamide, which are endogenous ligands for the PPARα (Fegley et al., 2005; O’Sullivan, 2007). In contrast to AEA, OEA decreases food intake and body weight gain in lean and obese rats and mice through a CB1 receptor-independent mechanism (Fu et al., 2003; Lo Verme et al., 2005).

        In contrast, the inverse agonist/antagonist rimonabant, when given acutely significantly and dose-dependently decreased margarine intake in HR rats; yet, it only reduced margarine intake in LR animals at high doses (3 mg·kg −1 ). In line with this, Parylak et al. (2012) showed that administration of a different CB1 receptor antagonist, SR147778, dose-dependently attenuated binge-like intake of a sweet-fat diet in rats. In addition, both doses of rimonabant reduced consumption of standard chow in all three diet groups.

        The most intriguing outcome of our study comes from the chronic treatment experiment with the low dose (0.3 mg·kg −1 ) of rimonabant, which showed that rimonabant preserves its selective reducing effect on fat food over time, as demonstrated by the finding that both HR and LR rats treated with the CB1 receptor inverse agonist/antagonist consumed less margarine but their intake of standard chow was unaltered. At present, as reviewed by Berner et al. (2011), the anorectic effect of CB1 receptor antagonists on palatable food consumed in a binge-like manner is not completely elucidated. Our results show for the first time that chronic treatment with rimonabant in rats that consumed the high-fat diet in a binge-type pattern resulted in a selective decrease in the consumption of palatable food, an effect that was maintained throughout the entire treatment period without the occurrence of tolerance, that instead develops towards standard chow intake (Colombo et al., 1998; Carai et al., 2006; Martín-García et al., 2010) or in other animal models of food intake (Mathes et al., 2008). Indeed, chronic rimonabant reduced food intake in the control rats with only access to standard chow, showing the development of tolerance from the second week of treatment. In line with this, preceding studies showed that CB1 receptor antagonists decrease energy intake by selectively reducing the consumption of palatable diets in normal rats (Arnone et al., 1997; Simiand et al., 1998; Mathes et al., 2008), and that the antagonistic effect might not be limited to palatable food (McLaughlin et al., 2003; Foltin and Haney, 2007). In our study, chronic treatment with rimonabant also reduced body weight in the HR group (which was increased at the end of the induction phase), but not in the C and LR groups. This finding is consistent with previous reports showing that CB1 receptor antagonism preferentially reduces the body weight of obese rats or rats with access to a sugar fat whip dessert (Mathes et al., 2008; Martín-García et al., 2010).

        Although the mechanism through which rimonabant exerts its effects on binge-type eating behaviour is still to be clarified, previous findings suggest that it may be effective in suppressing the hedonic response to food rather than just hunger (Cota et al., 2006). Several lines of evidence point to the endocannabinoid system as an important constituent of neuronal substrates involved in brain reinforcement/reward processes implicated in both food consumption and the homeostatic and hedonic regulation of eating (Di Marzo and Matias, 2005). In keeping with this, CB1 receptors are expressed particularly in brain areas (nucleus accumbens, hippocampus and entopeduncular nucleus) that are either directly involved in the hedonic aspects of eating or are connected to reward-related brain areas (Herkenham et al., 1990). Administration of ECs into the nucleus accumbens exerts a potent CB1 receptor-selective hyperphagic action (Kirkham et al., 2002; Soria-Gòmez et al., 2007) and enhances sucrose hedonic impact with hotspot focus in the dorsal shell part of the nucleus (Mahler et al., 2007). It is well recognized that palatable food stimulates the mesocorticolimbic dopamine system in a way similar to that of drugs of abuse, by increasing dopamine release in the shell of the nucleus accumbens (Martel and Fantino, 1996). Importantly, the increase in dopamine induced by presentation of palatable foods is blocked by administration of rimonabant (Melis et al., 2007), which suggests that the hedonic response to food might depend on the endocannabinoid system, probably through modulation of the mesocorticolimbic system. It should be noted that the release of dopamine disappears with repeated access to palatable food (Bassareo and Di Chiara, 1999), an effect not found when the subject is exposed to drugs of abuse (Wise et al., 1995). On the other hand, it has been demonstrated that animals that binge to a 10% sucrose solution repeatedly release dopamine in the nucleus accumbens shell each time they binge (Rada et al., 2005). The same effect has been shown when animals binge high-fat diets (Liang et al., 2006). In rats, prolonged binge-like intake of sugar solution leads to increased D1 receptor binding in the accumbens core and shell and decreased D2 receptor binding in the dorsal striatum (Colantuoni et al., 2001). Also, restricted feeding with scheduled sucrose access is associated with increased dopamine membrane transporter protein density in the nucleus accumbens and in the ventral tegmental area (Bello et al., 2003) in addition to an increased dopamine turnover exclusively in the nucleus accumbens (Hajnal and Norgren, 2002).

        In our study, the effect of rimonabant on the bingeing group could be related to its capacity to block dopamine release in the nucleus accumbens shell that might be induced by the consumption of margarine, and by a possible enhancement in the tone of the endocannabinoid system. Chronic exposure to high-fat palatable diet was found to decrease the expression of CB1 receptors in the nucleus accumbens (Harrold et al., 2002). Accordingly, Bello et al. (2012) reported a reduction in CB1 receptor density in the same central area in an animal model of BED. The reduction in CB1 receptor expression could be interpreted as the resulting effect of increased EC levels that in turn could induce dopamine release in this area (Bermudez-Silva et al., 2012). As mentioned before, CB1 receptors are expressed in the nucleus accumbens and are mainly located at the presynaptic level, and an important functional consequence of their activation is the inhibition of the release of other neurotransmitters (Schlicker and Kathmann, 2001). Activation of CB1 receptors on axon terminals of glutamatergic neurons in the nucleus accumbens would inhibit glutamate release, thus inhibiting the GABAergic neurotransmission in the ventral tegmental area, consequently disinhibiting VTA dopamine neurons (Melis et al., 2004; Riegel and Lupica, 2004). Thus, it is possible that rimonabant, by blocking the inhibitory effect of ECs, elicits stimulation of GABA release, resulting in reduced firing activity of dopaminergic neurons and reduced release of dopamine in terminal areas. On the other hand, clinical studies have found elevated EC plasma levels in women with BED, and this increase may drive the binge episodes and reinforce the rewarding effects of palatable foods, promoting the cycle of binge eating (Monteleone et al., 2005). However, our results do not exclude the possibility that rimonabant can produce its effect acting as an inverse agonist but also through one or more CB1 receptor-independent mechanisms (Pertwee, 2010).

        An overlap between compulsive overeating and drug abuse has been established, including the occurrence of craving and the loss of the inhibitory control over food intake and drug use. Notably, the choice of using the low dose (0.3 mg·kg −1 ) of rimonabant for chronic treatment was due to the fact that the same dose is able to antagonize i.v. self-administration of the CB1 receptor agonist WIN 55 212-2 as well as the reinstatement of drug-seeking behaviour in abstinent rats (Fattore et al., 2001; 2005), including female rats (Fattore et al., 2010).

        In conclusion, our results provide the first evidence that a chronic low dose of a CB1 receptor antagonist reduces fat intake in female rats showing binge-type eating behaviour, with a concomitant significant weight loss without development of any tolerance. In this respect, our data suggest the potential therapeutic utility of CB1 receptor antagonists in the treatment of binge-like eating disorders. Further studies are needed to characterize the exact mechanisms and brain areas mediating such effects.

        Acknowledgments

        This work was partially supported by grant from the Italian Ministry of University and Scientific Research (FAR DM28141 del 21/11/2005) and by the Intramural Research Program of the National Institute on Drug Abuse, National Institutes of Health, Department of Health and Human Services, Baltimore, MD, USA.