One of the most famous effects of marijuana use is the stimulation of appetite, particularly for palatable foods. After the discovery that THC was the psychoactive compound in marijuana, Hollister (1971) verified that a single oral dose of marijuana (containing 0.35 mg/kg THC) increased the intake of milkshakes in healthy volunteers. Subsequently it was demonstrated that smoked marijuana (containing 1.8 percent THC) produced a marked increase in palatable food intake in humans (Foltin, Brady, and Fischman 1986; Foltin, Fischman, and Byrne 1988). Considerable pre-clinical experimental work with rodents demonstrated that the THC and other CB1-receptor agonists act to enhance appetite both by amplifying the rewarding value of food and by reducing the satiety signals (e.g., leptin secretion) that regulate appetite (Cristino and Di Marzo 2015).
The therapeutic effect of cannabis to stimulate appetite has been studied for decades in the treatment of cachexia, a chronic wasting disorder associated with loss of adipose tissue and lean body mass seen in patients with cancer, patients with acquired immunodeficiency syndrome (AIDS), and patients with anorexia nervosa. Synthetic THC (dronabinol) is used in the clinic to combat the reduction in appetite in such patients, with mixed results. Some studies have found that dronabinol (2.5 mg twice a day) enhanced appetite and body weight in AIDS patients suffering from anorexia (Beal et al. 1997); other studies with both AIDS patients and patients with cancer-related cachexia report no effect with a similar dose and regime (Cannabis-In-Cachexia-Study-Group et al. 2006). Orally administered pure THC may not be the optimal treatment for these disorders. One alternative is the use of FAAH inhibitors (Fegley et al. 2005), which have been shown to enhance motivation for food and to promote energy storage in pre-clinical animal studies (Tourino, Oveisi, Lockney, Piomelli, and Maldonado 2010). As we learn more about the effects of cannabinoids and direct manipulations of the endocannabinoid system on appetite, better treatments may be developed.
The potential of the endocannabinoid system to maintain energy balance has been one of its most thoroughly studied effects since the synthetic CB1 antagonist/inverse agonist SR141716A (rimonabant, also known as Acomplia) was developed by Sanofi-Aventis to treat diet-induced obesity (Arnone et al. 1997). Unfortunately, unwanted psychiatric side effects of depression and suicidal ideation have prevented the use of rimonabant as an anti-obesity treatment. Rimonabant was shown to decrease food intake and body weight gain not only in animal models but also in humans. Pre-clinical work (see, e.g., Koch 2001) demonstrated that stimulation of CB1 receptors by THC, by synthetic CB1 agonists, or by endogenous cannabinoids stimulated eating in rats and mice. Genetically modified mice lacking CB1 receptors did not show this effect. Therefore, the feeding stimulation by THC was dependent on an intact CB1 system.
In the brain, the endocannabinoid system interacts with the mesolimbic dopamine system to engage with reward pathways and with the hypothalamus to regulate levels of hormones that mediate the enhancement of feeding. Indeed, endocannabinoid levels in the hypothalamus and in reward pathways are highest during food deprivation, leading to food-seeking behaviors.
Hormonal and nutritional signals in the hypothalamus inform the brain about the free and stored levels of fuel available for the organism. The neural circuitry of the hypothalamus uses this information to regulate caloric intake, energy consumption, and peripheral lipid and glucose metabolism. The actions of endocannabinoids in the hypothalamus serve to quickly fine-tune energy intake. Administration of anandamide (AEA) into the hypothalamus induces eating, and endocannabinoid levels in the hypothalamus vary as a function of nutritional status. The action of endocannabinoids in the hypothalamus regulates the levels of several biochemical compounds that control feeding and body weight, including melanocortin (which reduces feeding) and neuropeptide Y (which stimulates feeding). Within the hypothalamus, modulation of the expression of several pro-feeding and anti-feeding hormones by the endocannabinoid system is counterbalanced by the opposite actions mediated by the adipose-derived hormone leptin (Cristino and Di Marzo 2015).
Leptin originates in adipose tissue and affects a number of appetite-related factors in the hypothalamus; it plays an important role in regulating appetite, hunger, and metabolism. Endocannabinoid levels in the hypothalamus correlate inversely with leptin plasma levels (Di Marzo et al. 2001). Thus, leptin administration (which exerts an anorectic action) suppresses hypothalamic endocannabinoid levels in healthy animals, but in the hypothalamus of obese and hyperphagic rodents lacking leptin—such as ob/ob (obese/obese, leptin mutant) mice (see figure 7.1) and Zucker(fa/fa) (leptin-receptor mutant) rats—endocannabinoid levels are significantly increased (ibid.). Indeed, recent evidence indicates that functional CB1-receptor signaling in the hypothalamus is required for leptin to exert its suppressive effects on food intake; selective genetic knockout of CB1 receptors in the hypothalamus of mice abolished the inhibition of food intake by leptin (Cardinal et al. 2012). This inverse relationship also affects the activity of the brain’s reward system; that is, obese rats with defective leptin signaling show increased CB1 expression and binding activity in the brain’s reward structures (D’Addario et al. 2014).
Figure 7.1 An ob/ob (leptin-receptor mutant) mouse and a mouse with normal body weight.
Besides the hypothalamus and reward system, the endocannabinoid system regulates food intake and energy balance in the vagus nerve, which carries information between the digestive system of the gut and the brainstem. Within the gut, cholesystokinin (CCK) acts to suppress feeding (a satiety signal) and ghrelin acts to stimulate feeding. CB1-receptor expression in vagal afferents is increased in fasting and decreased after refeeding, under the control of CCK (Burdyga et al. 2004). The decrease in CB1-receptor expression after refeeding is prevented by administration of ghrelin. Thus ghrelin opposes the action of CCK on CB1-receptor expression on vagal afferents to the brainstem. In addition, the CB1-receptor antagonist rimonabant abolishes ghrelin-induced feeding by decreasing levels of circulating ghrelin. Ghrelin has also been shown to elevate hypothalamic endocannabinoid content (D’Addario et al. 2014).
Only one clinical study has investigated the regulation of endocannabinoid levels in humans eating food for pleasure (Monteleone et al. 2012). In normal-weight satiated healthy participants, plasma levels of 2-AG and ghrelin, but not AEA, OEA, and PEA, were elevated after hedonic eating. In addition, 2-AG levels in the hypothalamus increased in mice showing a preference for a high-fat diet (Higuchi et al. 2011). Highly palatable foods may evoke alterations in the central nervous system similar to those evoked by drugs of abuse through the regulation of common neurobiological substrates. Indeed, overconsumption of palatable foods is accompanied by the stimulation of the brain’s dopaminergic and opioid reward systems (D’Addario et al. 2014).
As was noted and illustrated in figure 6.3, the mesolimbic dopamine system is the brain’s primary reward system. All drugs of abuse activate this system (Wise 2004). The regulation of reward-related appetite processes may be controlled, in part, by endocannabinoid release in the VTA of the midbrain, which produces inhibition of the release of GABA, thus removing the inhibitory effect of GABA on dopaminergic neurons that project to the NAc (Maldonado et al. 2006). In the NAc, released endocannabinoids act on CB1 receptors on axon terminals of glutamatergic neurons. The resulting reduction in the release of glutamate on GABA neurons projecting to the VTA results in indirect activation of the VTA’s dopamine neurons. Microinjections of THC into the VTA and the NAc serve as rewards for both self-administration and conditioned place preference in rats (Zangen, Solinas, Ikemoto, Goldberg, and Wise 2006).
In pre-clinical rodent studies, the CB1 inverse agonist/antagonist rimonabant reduces the rewarding effects of foods and drugs. Rimonabant leads to a reduction in a conditioned place preference for palatable food (Chaperon, Soubrie, Puech, and Thiebot 1998). It also reduces the motivation for self-administration of beer in rats (Gallate, Saharov, Mallet, and McGregor 1999). CB1 antagonism also counteracts the increase of extracellular DA release induced by highly palatable foods, whereas CB1 agonists do the opposite (Higgs, Barber, Cooper, and Terry 2005; Melis and Pistis 2007). THC increases hedonic reactivity to sucrose (Jarrett, Limebeer, and Parker 2005), and CB1 antagonism does the opposite (Jarrett, Scantlebury, and Parker 2007). In addition, intra-NAc injections of AEA enhance sucrose hedonics (Mahler, Smith, and Berridge 2007). Thus CB1 is an important component of the neural substrate that mediates the reinforcing and motivational properties of a highly palatable food. Indeed, CB1 mRNA expression is downregulated in areas of the limbic forebrain of rats fed a palatable diet (Harrold, Elliott, King, Widdowson, and Williams 2002). This reduction in CB1 expression represents a compensatory mechanism aimed at counteracting increased levels of endocannabinoids resulting from the consumption of fat-rich palatable food. On the other hand, fasting leads to an increase of limbic AEA and 2-AG levels, which return to normal after re-feeding. This effect occurs only in brain areas not involved in the regulation of feeding behavior (Kirkham, Williams, Fezza, and Di Marzo 2002). This over-activation of the endocannabinoid systems during fasting may be part of a physiological mechanism aimed at enhancing the motivation for food.
Endocannabinoids also have important functional relationships with the endogenous opioid system, which also mediates the rewarding value of food (Kirkham 1991). Interestingly, the facilitory effect of THC on food intake is not only reduced by a CB1 antagonist; it is also reduced by the μ-opioid antagonist naloxone (Gallate and McGregor 1999). Indeed, a synergistic effect of very low doses of drugs that block the CB1 receptor and the μ-opioid receptor (Kirkham and Williams 2001) may serve as a promising treatment for obesity.
Because the endocannabinoid system is a modulator of both homeostatic and hedonic aspects of eating, dysfunctions of that system may lead to eating disorders. Anorexia nervosa and bulimia nervosa are characterized by abnormal eating behaviors resulting in severe food restriction, anorexia nervosa with a dramatic loss of body weight and bulimia nervosa with episodes of binge eating and vomiting without significant changes in body weight. The fifth edition of the Diagnostic and Statistical Manual of Mental Disorders includes binge-eating disorder, which is characterized by binge eating but without compensatory vomiting resulting in obesity. Such eating disorders clearly involve both psychosocial and biological factors.
On the basis of the known role of the endocannabinoids in regulation of feeding and energy homeostasis, it is conceivable that dysfunctions in this regulatory system may be involved in the pathophysiology of eating disorders. Monteleone et al. (2005) compared the blood levels of AEA and 2-AG of women with one of these eating disorders with those of healthy controls. They also examined the relationship between the endogenous cannabinoids and circulating leptin (a peripheral fat hormone involved in long-term modulation of body weight and energy balance), nutritional variables, and psychopathological variables. Increased blood levels of AEA (but not of 2-AG) were found in patients with anorexia nervosa or binge-eating disorder, but not in women affected with bulimia nervosa. In addition, anorexic women also showed decreased circulating leptin levels, and binge-eating disorder women showed increased leptin levels relative to healthy controls. In both healthy controls and women with anorexia nervosa, the higher the blood AEA levels the lower were the plasma leptin concentrations. This suggested that the decreased leptin signaling of underweight anorexia nervosa patients could be involved in the increase of AEA levels (Monteleone et al. 2005). Higher blood levels of CB1 mRNA are detected in women suffering from both anorexia nervosa and bulimia nervosa (Frieling et al. 2009), but decreased peripheral CB1 mRNA is found in patients with more severe forms of the disorders. More recently, PET scans demonstrated increased CB1-receptor levels in the insula (a cortical region regulating feeding) and in the inferior frontal and temporal cortex of underweight anorexia nervosa patients and in the insula of women with bulimia nervosa (Gerard, Pieters, Goffin, Bormans, and Van Laere 2011). It is possible that the dysregulated endocannabinoid tone of anorexia nervosa and bulimia nervosa patients may represent an adaptive response aimed at maintaining energy balance by potentiating internal food-seeking signals and hence stimulating food ingestion (Monteleone and Maj 2013). However, most human studies using dronabinol (e.g., Andries, Frystyk, Flyvbjerg, and Stoving 2014; Gross et al. 1983) have reported no improvement of anorexia symptoms.
There appears to be a positive association between obesity and either overproduction of endocannabinoids or increased expression of the CB1-receptor tissues (central and peripheral) involved in energy homeostasis in both animal and human studies. Genetically obese (ob/ob) mice (figure 7.1) were found to have hypothalamic endocannabinoid levels higher than in lean animals. These effects were significantly reduced by the exogenous administration of leptin (Di Marzo et al. 2001). Subsequently, an elevated CB1 expression was observed in the white adipose tissue of mice defective in leptin signaling relative to lean controls. Hyperglycemia and insulinemia may cause an over-activation of the endocannabinoid system in obesity-related type-2 diabetes. Dysregulation of the endocannabinoid system has also been reported in overweight obese women with a binge-eating disorder and in obese postmenopausal women. Obesity-related elevations of endocannabinoids were coupled to decreased activity of FAAH, and consistently specific polymorphisms of the FAAH gene were associated with overweight and obesity or with lower insulin sensitivity. Higher levels of 2-AG in adipose tissue were found in samples from patients with elevated abdominal fat distribution relative to patients with subcutaneous fat or lean controls. In addition, AEA levels are elevated in the saliva of obese subjects and directly correlate with Body Mass Index, waist circumference, and fasting insulin (Matias et al. 2012). In these participants, body-weight loss after a 12-week program decreased salivary AEA levels. Therefore, salivary AEA may be a useful biomarker for obesity.
Since there appears to be pathological over-activation of the endocannabinoid system in overweight and obese subjects, current therapeutic targets for obesity are aimed at the restoration of a normal endocannabinoid tone by means of drugs that interfere with endocannabinoid signaling.
Initial pre-clinical studies conducted with rimonabant support the hypothesis that antagonism of the endocannabinoid system with the CB1 inverse agonist/ antagonist rimonabant may be a promising therapy for obesity. Several clinical trials followed the initial experimental findings from the four Rimonabant in Obesity (RIO) studies (Despres, Golay, and Sjöström 2005; Pi-Sunyer et al. 2006; Scheen et al. 2006; Van Gaal et al. 2005). Rimonabant was marketed as an anti-obesity agent in several European countries. In the United States, the Food and Drug Administration asked for further evidence regarding safety before approving its marketing. A subsequent meta-analysis of the four RIO studies suggested that patients treated with rimonabant were 2.5–3 times as likely to experience psychiatric adverse effects, such as anxiety and depression, as patients receiving placebo (Christensen, Kristensen, Bartels, Bliddal, and Astrup 2007). The European Union Committee for Medicinal Products for Human Use concluded that rimonabant doubled the risk for psychiatric disorders, and the European Medicines Agency (EMA) suspended the license for the drug. Sanofi-Aventis withdrew rimonabant from the worldwide market, and clinical development not only of rimonabant but also of other CB1 antagonists being developed by other companies was halted.
Recent pre-clinical investigation has focused on the evaluation of CB1 antagonists/inverse agonists that do not penetrate the brain as potential anti-obesity treatments. Such treatments may be devoid of the adverse psychiatric side effects produced by brain-penetrate CB1 antagonists/inverse agonists, such as rimonabant. Human adipose tissue has been shown to possess a fully functional endocannabinoid system (Spoto et al. 2006), and a high-fat diet induces an increase of AEA in the liver as a result of reduced degradation by FAAH (Osei-Hyiaman et al. 2005). In addition, 2-AG levels are elevated in the visceral fat of obese patients (Bluher et al. 2006). These findings suggest that the hypophagic effect of rimonabant may be mediated by its action on peripheral CB1 receptors rather than central receptors. (See also Gomez et al. 2002.) Therefore, the beneficial effects of blocking peripheral CB1 receptors to reduce appetite, body weight, hepatic steosis, and insulin resistance are being investigated (Tam et al. 2010; Tam et al. 2012). Tam et al. (2010) found that the CB1-receptor antagonist (without inverse agonist properties) AM6545 was as effective as rimonabant in ameliorating the fatty liver and improving the lipid profile in mice with diet-induced obesity, but that it was less effective than rimonabant in reducing body weight, adiposity, insulin resistance, and hyperleptinemia and had minimal effect on food intake. Because rimonabant is both a brain penetrant and a CB1-receptor inverse agonist, its greater efficacy could be due either to its central action or to its inverse agonist properties. Tam et al. (2012) demonstrated that it is the latter property that makes rimonabant a better treatment for obesity and its related metabolic disorders. That is, with the use of a peripherally restricted CB1-receptor inverse agonist, JD5037, they demonstrated as much reduction in food intake, body weight, and adiposity as was obtained with rimonabant. These results are particularly exciting because they may lead to new treatments for obesity without the possibility of psychiatric side effects that prevented the use of rimonabant as a treatment.
Another approach to reducing the impact of endocannabinoids on feeding is to interfere with their metabolism. Indeed, decreasing 2-AG levels by systemic administration of DAGL inhibitor (the enzyme responsible for the synthesis of 2-AG) reduced intake of palatable or high-fat foods by mice (Bisogno et al. 2009; Bisogno et al. 2013). In addition, mice genetically altered to overexpress MAGL (the enzyme that degrades 2-AG) and put on a high-fat diet showed increased energy expenditure and decreased weight gain (Jung et al. 2012). Another strategy that is being explored is related to the allosteric modulation of the CB1 receptor by changing the binding of the endocannabinoid. PSNCBAM-1, a novel allosteric antagonist of the CB1 receptor, reduces that impact of AEA or that of 2-AG when it binds to the orthosteric receptor site. This allosteric antagonist inhibited appetite and produced weight loss in rats (Horswill et al. 2007). Furthermore, there is some evidence that CB2 antagonism has potential to manage obesity-associated metabolic disorders (Deveaux et al. 2009); this is important because CB2 agonists do not produce central effects.
THC clearly has the potential to enhance appetite through its agonism of the CB1 receptor. However, its psychoactive side effects limit its usefulness as an appetite-stimulating agent in treating anorexia or cachexia. What about the other cannabinoids found in cannabis that are not psychoactive?
In view of the withdrawal from the market of rimonabant, taranabant, and other CB1 inverse agonists because of unwanted psychiatric side effects, safer alternatives are needed. THCV is a CB1 antagonist without inverse agonist effects that reduces food intake in rats (Pertwee 2007; Pertwee 2008; Thomas et al. 2005). The lack of inverse agonism of the CB1 receptor with THCV may render it more tolerable, because pre-clinical research suggests that the psychiatric side effects of rimonabant were produced by its central inverse agonism (not antagonism) of the CB1 receptor (Limebeer et al. 2010; Sink et al. 2008). In addition, CB1 antagonists, unlike inverse agonists, are devoid of potential anhedonic effects (Meye, Trezza, Vanderschuren, Ramakers, and Adan 2013) and nausea-producing effects (Limebeer et al. 2010; Sink et al. 2008). Therefore, it will be interesting to see whether THCV will be developed as a potential appetite-reducing anti-obesity drug.
Early work with CBD indicated that at a high dose (50 mg/kg, ip) it reduced consumption in rats trained to consume their daily food intake in 6 hours; when lab chow was replaced with sucrose, however, suppression of intake was greatly decreased (Scopinho, Guimaraes, Correa, and Resstel 2011; Sofia and Knobloch 1976; Wiley et al. 2005). More recent work has indicated that CBD (3–100 mg/kg, ip) does not modify food intake in mice (Wiley et al. 2005; Scopinho et al. 2011). However, when administered orally, relatively low doses of CBD have been shown to produce a short-term reduction in feeding (Farrimond, Whalley, and Williams 2012) by an unknown mechanism. Since CBD has been shown to ameliorate some unwanted effects of THC (e.g., anxiety and learning deficits), it may be worthwhile to evaluate combined doses of CBD and THC in treatment of anorexia nervosa or other eating disorders.
At the turn of the twenty-first century, new medications based on inverse agonism/antagonism of the CB1 receptor showed great promise in treating obesity. However, shortly after such medications were widely prescribed in several countries, patients reported suicidal thoughts and anxiety, probably attributable to inverse agonism of central CB1 receptors.
It is clear that the endocannabinoid system is an important regulator of appetite, food preference, and body weight. The endocannabinoid system not only regulates metabolic feeding-related hormones in the brain and in the gut, but also regulates the brain reward circuitry involved in palatability-based feeding. One of the primary roles of the endocannabinoid system is in the homeostatic regulation of feeding behavior.
New treatments for obesity being developed by chemists and pharmaceutical companies attempt to harvest the anti-obesity effects of rimonabant, but are devoid of the psychiatric side effects that became clearly known only after rimonabant was widely prescribed.