Endocannabinoids: an appetite for fat.

I t has been known for decades that marijuana causes the “munchies,” i.e., a hunger for palatable food, and for more than 10 y that endocannabinoids (eCBs), in some ways marijuana’s counterpart in the organism, are orexigenic mediators (1). When injected in the hypothalamus (HT) or nucleus accumbens (NAc), two key brain areas for the homeostatic and hedonic control of food intake, these compounds stimulate food consumption by acting at the cannabinoid CB1 receptor, one of the two G proteincoupled receptors for marijuana’s psychotropic and appetite-inducing component, Δ-tetrahydrocannabinol (2, 3). Conversely, systemic pharmacological blockade of CB1 receptors causes anorectic effects in rodents exposed to palatable food, or food-deprived for a few hours, and in obese animals (3–6). The two most studied eCBs, anandamide and 2-arachidonoylglycerol (2-AG), are considered local mediators produced by cells only following stimulation, and in this they differ from other signals controlling food intake, which are released into the bloodstream and/or prestored in vesicles (1). What makes the eCBs similar to most orexigenic signals is the fact that their concentrations in the HT and NAc, but also in the proximal intestine, which transmits to the brain the state of “emptiness” or “fullness” of the gut, increase during food deprivation and decrease immediately after food consumption (3, 7, 8). This mechanism is presumably caused by the opposing effects on eCB levels of hormones such as leptin, on the one hand, and ghrelin and corticosterone, on the other hand, the levels of which also vary during food deprivation and refeeding (1, 6). These food intakeand hormonesensitive changes in eCB signaling are thought to play a key role in the regulation of (i) the release of central neurotransmitters and neuropeptides controlling food intake and (ii) the activity of vagal fibers from the duodenum to the brainstem, which signal gastric distension (1). In this scenario, the results of the elegant study by DiPatrizio and coworkers published in PNAS (9), although in full agreement with the general orexigenic function of eCBs and CB1 receptors, might seem somewhat surprising. In fact, the authors report that a fatty meal elevates eCBs selectively in the rat proximal small intestine, and propose that this effect: (i) is induced by the orosensory properties of the meal; (ii) is mediated by vagal afferent and efferent terminals, and (iii) reinforces fat intake via CB1 receptor activation (9). The authors used a “sham-feeding” protocol, which consists of immediately draining liquid foods from the stomach through a chronically implanted gastric cannula (9). In this way, the stomach does not undergo the distension that follows each meal and the ensuing production of gastric anorexigenic signals such as cholecystokinin (CCK), nor the direct activation of vagal fibers that transmit changes in gastric volume to the hindbrain. Thus, sham feeding produces effects on food intake specifically linked to the orosensory properties of food and blunts homeostatic negative feedbacks on food intake, including the previously observed postingestive reduction of small intestinal eCB levels (7, 8). The authors found that liquid meals containing sucrose or proteins do not alter eCB levels in the proximal intestine, which were, instead, enhanced by a fatty meal, although not in sham-fed rats that had underwent subdiaphragmatic vagotomy. The authors suggest that “cephalic signals elicited by sham feeding of fat, but not other nutrients, selectively mobilize eCBs in the upper gut through a mechanism that is mediated by efferent vagal fibers” (9). This phenomenon results in increased liquid fat consumption via activation of duodenal CB1 receptors, as intraduodenal injection of the CB1 antagonist rimonabant, or systemic injection of a more peripherally restricted CB1 antagonist, reduced fat, more than normal chow, intake in sham-fed rats. These findings are striking, as they imply the existence of mechanisms through which fat “sensors” in the oropharyngeal cavity can stimulate cranial afferents, and hence vagal efferents in the proximal small intestine, thereby enhancing eCB levels and acting as a “priming trigger” for further intake of fat versus other nutrients. Among such sensors, CD36, TRPC5, GPR40, and GPR120 are expressed in taste buds (10, 11), and it will be interesting to see if any of them is connected with this phenomenon. However, one caveat of the study is that the authors, possibly to habituate the animals to the stress induced by this procedure, performed sham feeding with the test nutrients for 4 d (30 min per day, followed by ad libitum feeding) before analyzing eCB levels in various peripheral and central tissues after a fifth administration of the test meals. Therefore, it is not possible from these data to understand whether fat-induced stimulation of small intestinal eCB levels was the result of just the last test, or rather an adaptive mechanism induced by the fatty meal over the 5 d of the experiment. Indeed, a recent study showed that a high-fat diet administered for 3 d can enhance hypothalamic 2-AG levels in free-feeding mice (12). It would be interesting to replicate the results by Fig. 1. Effects of CB1 receptor activation by eCBs on the orosensory properties of food. Adapted from Gaillard et al. (10). Blue, red, or purple arrows and text denote CB1-mediated actions activated by fat, sucrose, or both, respectively (9, 17, 18). Blunted arrow denotes inhibition. Roman numbers denote cranial and sympathetic nerves. Green arrow and text depict the recently reported CB1-mediated enhancement of odor sensitivity in Xenopus laevis larvae (19).