Food reward

A physiological response to ingestion of palatable food, resulting in behaviours directed towards eating.

Introduction
''' Food intake is driven by both 'homeostatic feeding' (energy demands) and ‘non-homeostatic feeding’ (pleasure associated eating or preferred food). The latter is associated with the food reward processes, which is further categorized to’ liking’ (pleasure/palatability) and ‘wanting’ (incentive motivation) according to the Salience theory (Berridge K.C, 2007). Moreover, the salience model describes different brain mechanisms associated with each of the two components. Experiments using mouse model showed that the brain mechanisms attributed to ‘liking’ involve the neurotransmission of mu-opiods in the nucleus accumbens, ventral pallidum, parabrachial nucleus, and nucleus of the solitary tract, while mechanisms attributed to ‘wanting’ implicated the neurotransmitter dopamine secreted in brain areas such as the prefrontal cortex, amygdala, hypothalamus, and projections from the ventral tegmental area to the nucleus accumbens (Berridge K.C, 2007). Dopamine has an important role in both energizing (secondary role) and also reinforcing feeding (primary role). The latter process is associated with various food reward pathways of dopamine and that is where this article focuses the most on. The action of dopamine in dopaminergic systems is analysed. In addition, the interaction of the dopaminergic system with other reward systems is described and finally the effect of other hormones such as insulin, ghrelin, and leptin, on food reward is studied through their association with dopamine.The final part of this article links the aforementioned processes, particularly the dopaminergic pathway, with obesity. A descripttion of what can go wrong, changes in the distribution of dopaminergic receptors, and a genetic approach is given (between the function of dopaminergic system and obesity) by a gene, related in the reinforcing properties of food behaviour.

Motivated behaviour and food as a reinforcer
The underlying pathways in motivating feeding behaviour seem to be far more complex than a simple homeostatic system, which responds to metabolic and satiety signals from the gut. One possible thought is that the brain’s reward systems react to stimuli such as sight, smell and taste, or other cues that predict food (Wise, 2006). However, hunger can not result in unconditioned goal-directed behaviour (Changizi et al. 2002). Chance encounters with various tastes of palatable foods are required before goal-directed behaviour can result from the interaction of the internal needs with the salience of environmental stimuli (Wise, 2006).

For example an infant recognises (Steiner et al. 2001) and learns to seek out (Johanson & Hall 1979)sweet tastants, but the desire for a particular food is controlled by the interaction of peptide levels (related to hunger) with the brain circuitry, coding the animal’s reinforcement history for that specific food. Subsequently, the infant will indiscriminately taste both food and non-food objects, until it has received reinforcing feedback from sufficient stimuli (Wise, 2006). In addition, the monkey’s appetite for a yellow banana requires the prior learning of the relation of the sight of the yellow skin of a banana, with the sweet taste of the white banana meat (Wise, 2004b) plus the consequences resulting from the ingestion of the fruit. Therefore, preference for a specific food, results only when the post-ingestional consequences of that food’ reinforce’ the tendency to eat that food. For the above reasons, food is considered to be a strong reinforcer. Moreover, when the response of a behaviour stimulated by a reinforcer increases the rate of that specific behaviour; that is known as positive reinforcement or reward learning, and the positive events are called rewards (Epstein, 2007). The reinforcing efficacy of food reward is the ability of the reward to maintain rather than to establish behaviour; consequently the stimulus learning contributes to the response learning.

Dopamine is known to play an important role in both. However, evidence from various studies seem to conclude that dopamine’s contribution appears to be chiefly to cause ‘wanting’ (Dopamine signalling in the dorsal striatum/CPu) for hedonic rewards rather than ‘liking’ or learning (mesolimbic dopamine) for those rewards. The first evidence for the implication of dopamine in food reward came from studies in rats, where dopamine antagonists blocked the rewarding effects of brain stimulation (Liebman & Butcher 1974; Fouriezos & Wise 1976) and of psychomotor stimulants.

The role of the Mesolimbic Dopaminergic Reward System
Human eating behaviour is not solely dependent on the homeostatic measures controlled by the hypothalamus but also the dopaminergic system which is activated by various stimuli; auditory, visual, tactile, olfactory and gustatory. The dopaminergic system was first connected with the reward system (de Wit & Wise 1977) in response to pharmacological and genetic approaches. They clearly established that dopamine was involved in motivation, as a deficit results in the starvation and dehydration of the mouse which ultimately results in death. (Palmiter 2007)

In the ‘reward circuit’, projections from the Ventral Tegmental Area (VTA) to the Nucleus Accumbens (NAc) have received the most attention due to the focus of studies on the hedonic impact from drugs and their possible roles in reinforcement, reward and addiction. These results have often led to the conclusion that dopamine action in the NAc is needed for motivation to acquire food or addictive drugs. Most reviews suggest that the projections from the VTA-NAc are needed for the motivation to eat but not for the food consumption. Lesion experiments have shown that even when the VTA-NAc pathway has been destroyed the mice still manage to eat (Wise 2006).

The Dopamine Hypothesis

Dopamine signalling from the VTA to the NAc, hippocampus, amygdale and/or pre-frontal cortex promotes reward-related activities. Dopamine signalling in these brain regions focuses attention to salient environmental events and thereby facilitates behaviour towards directed goals. Also it is thought that dopamine released from the VTA also forms associations to promote learning between food reward and the environment (Palmiter 2007)

However the role of mesolimbic dopamine seems to be controversial. Dopamines’ possible role in relation to reward?

•	Hedonia – Dopamine in the NAc acts as a pleasure neurotransmitter. Proposed due to drug activity. Not all rewards activate the reward system suggesting that the mesolimbic pathway is not solely hedonic.

•	Learning – predictions of future rewards, NAc and VTa lesions do not affect this part but lack the motivation for the reward.

•	Incentive Salience – the ‘wanting’ of the reward, released when there is a stimulus worth working hard for. In absence of DA the environmental stimulus go unnoticed and the animal eventually dies due to starvation and dehydration.

The incentive salience theory seems to best fit the data in this field according to Berridge (2007). Therefore dopamine causes the wanting of the reward after the appropriate stimuli have been processed in the reward system. An elevation of dopaminergic transmission is needed to form these associations. It has been shown that an increase in extracellular dopamine is seen in regard to natural rewards, food, water and sex, during acute administration (Wise & Rompre 1989, Spanagel & Weiss 1999). However it must be noted that novelty is an important factor in the increased release from the NAc.

It has been suggested by Palmiter, 2008 that the role of dopamine in motivation is split between the 2 dopaminergic pathways; the NAc and CPu pathways. The SNpc-CPu pathway is essential for motivation with dopamine signalling from the VTA-NAc needed in regard to modulating the actions of the other dopaminergic pathway.

The substantia nigra pars compacta (SNpc) to the caudate putamen (CPu): A critical dopaminergic pathway
It has been stated that the midbrain dopamine (DA) neurons are the key neural components for reward mechanisms (Satoh, T. et al (2003)). Creation and observation of dopamine deficient (DD) mice implied that DD mice starve because they are not motivated to respond to hunger signals (Palmiter, RD. (2008)). Thus, its been proposed that DA is crucial for mice to engage in the majority of goal-directed or motivated behaviours (Palmiter, RD. (2008)).

However in the literature there is much controversy as to the pathway used; a universal finding is the involvement of the striatum, the input structure of the basal ganglia in a circuit responsible for mediating goal-directed behaviour, with the striatum’s central role being the processing of reward like stimuli (Delgado, MR. (2007)). The two proposed pathways are from the ventral tegmental area (VTA) to nucleus accumbens (NAc) (ventral striatum); or the substantia nigra pars compacta (SNpc) to the caudate putamen (CPu) (dorsal striatum) (Palmiter, RD. (2008)).

Basal ganglia diagram We include in the striatum not only the dorsal region, which encompasses the caudate nucleus and putamen, but also the ventral region that includes the core and shell of the nucleus accumbens (Wickens, JR. et al (2007)).

The bulk of reward information processing comes from animal models in the literature (Delgado, MR. (2007)). One study using nonhuman primates found that striatal neurons responded to the anticipation and delivery of reward (Delgado, MR. (2007)). Another study found reward-related dopamine response specifically in the mouse dorsal striatum, correlated with the delivery of food reward (Natori, S. et al. (2009)). The importance of the DA system in the dorsal striatum is demonstrated in a study using DD mice whose DA signalling is restored by viral rescue (Palmiter, RD. (2008), Darvas, M. & Palmiter, RD.(2009)). These mice learned to lever press for food rewards as quickly as control mice and their motivation to work for food was restored (Darvas, M. & Palmiter, RD.(2009)). An important finding was that in DD deficient mice feeding was never restored after viral transduction in the NAc (Palmiter, RD. (2008)).

Recently, the advancement of neuroimaging techniques has allowed researchers to extend such investigations to the human brain (Delgado, MR. (2007)). DA release increases in dorsal striatum of hungry participants when stimulated with food items, demonstrating its involvement in reward processing (Delgado, MR. (2007)). During the delivery of rewards fMRI signals were higher in the dorsal striatum, particularly the head of the CPu (Delgado, MR. (2007)). These findings strongly suggest the human dorsal striatums involvement in reward processing; with the CPu being an integral structure of a circuit involved in learning and updating current rewards with the aim of maximizing reward consumption (Delgado, MR. (2007)).

The role of DA signalling in the CPu cannot be ignored as viral restoration rescued feeding, whereas in the NAc it did not. It has been proposed that dopamine signaling in the CPu is essential for motivation while dopamine signaling in the NAc modulates this motivation and evaluation of reward like stimuli (Palmiter, RD. (2008)).

The effect of hormones on the dopaminergic reward system
The role of hormones such as Leptin, ghrelin and insulin in the homeostatic control of energy balance has been extensively studied. These hormones reflect the size of energy stores, such as adipose tissue, and feed into the medial hypothalamus to influence feeding behaviours and regulate food intake and energy expenditure in response to metabolic demand. Recent evidence also implicates a role for such hormones in dopamine reward pathways.

, (Figlewicz DP, Benoit SC, 2009).

Receptors for these hormones are located on dopaminergic neurones in the ventral tegmental area (VTA) (Magni P. et al, 2009), and ligand binding results in activation e.g. by ghrelin (Abizaid, A. et al, 2006), or inhibition e.g. by insulin or leptin of dopamine signalling to the nucleus accumbens (NAc) (Magni P. et al, 2009). These alterations in dopamine signalling pathways have complex effects on eating behaviours. Many studies have reinforced the hypothesis that insulin and leptin attenuate the food reward, reducing incentive to eat (Figlewicz DP, Benoit SC, 2009). For example, behavioural studies showed that administering insulin or leptin to rats affected conditional place preference, which assesses the ability to relate a particular food reward to a particular environment (Figlewicz DP, 2003). Rats were fed a high fat diet and underwent a ‘training period’ prior to the test involving intracerebroventricular administration of insulin or leptin. CPP was only abolished in those rats that received insulin or leptin treatment before or during the test as well as in training, whilst those who only received it in training maintained a normal CPP (Figlewicz, D.P. et al, 2004). This suggests that insulin and leptin influence the retrieval of food reward associations rather than the initial formation of these associations (Figlewicz, D.P. et al, 2004); presumably as a consequence of their inhibition of dopamine reward pathways. These results have been reinforced by studies which have shown decreased sucrose self-administration in response to insulin or leptin administration (Figlewicz DP et al, 2006) and decreased sucrose licking following insulin treatment (Sipols AJ et al, 2000). It is apparent that the high levels of insulin and leptin associated with obesity impair dopamine food reward pathways resulting in abnormal eating behaviours (Figlewicz DP, Benoit SC, 2009).

Table 1: Summary of effects of centrally administered insulin and leptin on reward behaviours
 * Adapted from (Figlewicz, D.P. et al, 2004)

Ghrelin has been shown to increase dopamine signalling via ghrelin receptors on VTA neurones, via direct activation and also indirect manipulation of inputs onto the VTA to those of an excitatory nature (Abizaid, A. et al, 2006). However, it remains unclear whether this is a significant part the mechanism by which ghrelin stimulates feeding (Palmiter R, 2007).

A major difficulty in elucidating the specific roles of these hormones in reward systems, and in general regulation of body weight, is that, due to the involvement of multiple hormones, manipulation of one results in activation of compensatory mechanisms, masking the physiological role of the manipulated hormone (Palmiter R, 2007). Furthermore, by manipulating levels of these hormones to abnormal levels, we can suggest potential functions for them, but this may not be relevant when they are found at physiological concentrations (Palmiter R, 2007). It should also be noted that the VTA possesses other neurons such as GABA-projection neurons, which also express receptors for the discussed hormones. Therefore, we can not assume that these hormones affect feeding behaviours solely by their action on dopamine reward pathways. It is also unclear whether these hormones act directly on dopamine reward pathways; insulin and leptin may influence dopamine reward systems by altering the activity of secondary peptide effector pathways, such as orexin A and melanocortins (Figlewicz DP, Benoit SC, 2009).

Much work, therefore, remains to be done to decipher the significance of results from these studies and the specific roles of hormones in food reward pathways (Palmiter R, 2007).

Opioid and cannabinoid systems
Other reward systems, including the endogenous opiate and endocannabinoid systems, also play a role in reward behaviours and interact with dopamine reward pathways (Palmiter R, 2007). Opiates act in the nucleus accumbens to increase ‘wanting’ and ‘liking’ of food rewards (Pecina S, 2008). Opioids also influence mesolimibic dopamine pathways by inhibiting GABAnergic input onto dopamine neurones on the VTA, resulting in increased dopamine release (Spanagel and Weiss, 1999). The endocannabinoid system has also been implicated in reward behaviours. Endocannabinoids act to modulate neurotransmission in several brain areas; and they modulate dopaminergic transmission by inhibiting both excitatory and inhibitory input onto dopamine neurones (Maldonado, R et al 2006).

The relationship between obesity and the dopaminergic system
There is a difference in dopaminergic activity between obese women in comparison to lean women/men in response to food and satiety. It has been shown that the obese have a higher metabolic activity in the parietal somatosensory area of the cortex which is linked to the sensory mouth, lips and tongue. An increased amount of sensory processing in this brain region could increase the reinforcing properties of food (Epstein 2007).

Deficiencies of the D2 receptor have been suggested to increase the likelihood of being obese. Wang et al (2001) showed that obese people have fewer D2 receptors in the striatum and with both the D1 and D2 receptors acting synergistically to decrease feeding, this altered expression causes an increased amount of eating. The DRD2 gene is responsible for the reinforcing properties of food/addictive behaviour ( Noble 2003). Those who have the allelic variant A1 in this gene have fewer D2 receptors and therefore a decreased amount of dopamine signalling within the brain. This has been shown to be higher in obese individuals making the dopamine reward circuits less sensitive. This could possibly explain why obese people possibly overeat in order to compensate for their lack of reward.