Is dopamine a physiologically relevant mediator of feeding behavior?

The hypothalamus integrates various hormonal and neuronal signals to regulate appetite and metabolism and thereby serves a homeostatic purpose in the regulation of body weight. Additional neural circuits that are superimposed on this system have the potential to override the homeostatic signals, resulting in either gluttony or anorexia at the extremes. Midbrain dopamine neurons have long been implicated in mediating reward behavior and the motivational aspects of feeding behavior. Recent results reveal that hormones implicated in regulating the homeostatic system also impinge directly on dopamine neurons; for example, leptin and insulin directly inhibit dopamine neurons, whereas ghrelin activates them. Here, I discuss the predictions and implications of these new findings as they relate to dopamine signaling and the physiology of appetite control.

Introduction

Decisions about when to eat, what to eat and when to stop eating are based on a multitude of factors, including perception of energy balance, food palatability and social factors. There has been tremendous progress in deciphering how the homeostatic system in the hypothalamus modulates appetite and metabolism in response to peripheral signals related to energy stores and intestinal tract activity 1, 2, 3 (Figure 1). Nevertheless, it is obvious that this system does not work very well in an environment where little effort is necessary to procure palatable food with high caloric density. The homeostatic system adjusts food intake and energy expenditure over the short run, but the set point (the body weight that the system aims to maintain) gradually increases with persistent overindulgence. Moreover, as obesity progresses, the homeostatic system becomes resistant to signals that should convey that energy stores are sufficient [4]. The motivation to eat or stop eating is clearly more complex than a simple homeostatic system that responds to metabolic and satiety signals from the gut [5]. One idea is that the brain's reward systems respond to the sight, smell and taste of food (or cues that predict food) and override the homeostatic system, which evolved under conditions in which food was never chronically abundant.

What are these reward systems, and how do they become overwhelmed by the cornucopia of tasty food? The dopamine reward system, especially the projection from the ventral tegmental area (VTA) to the nucleus accumbens (NAc), has been studied the most because of its link to drug addiction 6, 7, 8. The endogenous opiate [9] and cannabinoid systems [10] are also important and they interact with the dopamine system. A new concept reviewed here is that hormones (insulin, leptin and ghrelin) that were previously studied as hormonal inputs to the homeostatic system also directly affect the dopamine reward pathway (Box 1). These hormones activate receptors on VTA dopamine neurons, thereby either stimulating (ghrelin) or inhibiting (leptin and insulin) dopamine signaling in the NAc (Figure 1, Figure 2). Thus, by affecting the dynamics of dopamine signaling in the NAc, these polypeptides could influence the subjective reward value of food and hence the motivation to eat.

The overall premise of the studies discussed here is that increases in dopamine signaling in the NAc promote feeding, whereas decreases have the opposite effect. I discuss two fundamental problems related to these recent papers. One problem is that, during a fast, the actions of leptin, insulin and ghrelin on VTA dopamine neurons are predicted to increase dopamine signaling; however, available evidence suggests that dopamine levels do not increase. This problem raises the question of whether signaling by these hormones onto dopamine neurons is physiologically relevant. The second problem, based on our own work, is that dopamine signaling to the dorsal striatum seems to be much more important for feeding than signaling from the VTA to the ventral striatum. After discussing evidence for direct signaling by leptin, insulin and ghrelin on dopamine neurons, and reviewing our data obtained with mice that either lack dopamine or have dopamine signaling restricted to the dorsal striatum, I suggest potential ways to resolve these two problems.