Central lipid detection and the regulation of feeding behavior

major Abstract: The modern abundance of energy-rich foods combined with a shift to more sedentary lifestyles has led to a thermodynamic imbalance in which excessive caloric intake and reduced energy expenditure account for the prevalence of obesity. In particular, exposure to lipid-rich diet is thought to promote metabolic alteration in peripheral tissue associated with obesity-related diseases. The regulation of energy balance depends on the ability of the brain to provide an adaptive response to change in circulating factors of hunger and satiety. The hypothalamus is particularly regarded as key integrative structure but, aside from hypothalamic-mediated homeostatic control, feeding behavior is also modulated by sensory inputs, such as tastes and odors, as well as by affective or emotional states. The reinforcing and motivational aspects of food are closely tied to the release of the neurotransmitter dopamine by the mesolimbic system, which is stimulated by calorie-dense foods as well as by most other objects of desire. Therefore feeding behavior is regulated by homeostatic as well as non-homeostatic inputs from the hypothalamus and the mesolimbic region. Interestingly, these structures expresses several enzymes involved in the processing of triglyceride and fatty acid and the recent literature provide growing evidence that fatty acid metabolism within discrete brain regions can function as sensor of nutrient availability directly control the hedonic and the homeostatic aspect of feeding.

considered as an epidemic by the French Health system. A recent report from the French SENAT pinpoints the dramatic progression of obesity in France 2 and a ''Plan Ob esit e'' has been launched in 2011 under the highest authorities. According to the WHO, the fundamental cause of obesity and overweight is an energy imbalance between calories consumed and calories expended. And it is a fact that globally, there has been both an increased intake of energy-dense foods that are high in fat, salt and a decrease in physical activity.
In adult mammals energy homeostasis is finely regulated. Blood glucose levels, body weight, and fat content remain within narrow ranges and experimentally-induced perturbations (e.g., food restriction) invariably result in a rapid return to ''set point'' when normal conditions are restored. To accomplish this, circulating peripheral factors such as hormones (e.g., insulin, leptin, ghrelin) and nutrients (e.g., glucose, lipids) activate discrete neural circuits in the brain that trigger changes in the basal metabolic rate and/or feeding behaviour. Disruption of these neural circuits can give rise to life-threatening conditions that include metabolic diseases such as obesity and diabetes in both humans and rodent models. Therefore it is crucial to understand the mechanism that insures the proper equilibrium between energy intake and expenditure.
Several observations led to the identification of the hypothalamus as a major Abstract: The modern abundance of energy-rich foods combined with a shift to more sedentary lifestyles has led to a thermodynamic imbalance in which excessive caloric intake and reduced energy expenditure account for the prevalence of obesity. In particular, exposure to lipid-rich diet is thought to promote metabolic alteration in peripheral tissue associated with obesity-related diseases. The regulation of energy balance depends on the ability of the brain to provide an adaptive response to change in circulating factors of hunger and satiety. The hypothalamus is particularly regarded as key integrative structure but, aside from hypothalamic-mediated homeostatic control, feeding behavior is also modulated by sensory inputs, such as tastes and odors, as well as by affective or emotional states. The reinforcing and motivational aspects of food are closely tied to the release of the neurotransmitter dopamine by the mesolimbic system, which is stimulated by calorie-dense foods as well as by most other objects of desire. Therefore feeding behavior is regulated by homeostatic as well as non-homeostatic inputs from the hypothalamus and the mesolimbic region. Interestingly, these structures expresses several enzymes involved in the processing of triglyceride and fatty acid and the recent literature provide growing evidence that fatty acid metabolism within discrete brain regions can function as sensor of nutrient availability directly control the hedonic and the homeostatic aspect of feeding.
integration site for inputs related to energy homeostasis (Schwartz and Porte, 2005). The hypothalamus contains specialized neurons that monitor circulating factors of hunger and satiety in order to provide an integrated adaptive response at both the metabolic and behavioral levels. This homeostatic control of energy homeostasis relies primarily on the so called melanocortin system which encompasses several molecular actors involved in body weight regulation (Schwartz et al., 2000;Cone, 2005).
However, aside from hypothalamicmediated homeostatic control, feeding behavior is also modulated by sensory inputs, such as tastes and odors, as well as by affective or emotional states. For example, stress or anxiety can stimulate reward seeking and consumption of highly palatable food independent of metabolic demand (Dallman et al., 2005). The reinforcing and motivational aspects of food are closely tied to the release of the neurotransmitter dopamine, which is stimulated by high-fat/ high-sugar foods as well as by most other objects of desire (e.g., sex, drugs) (Gunstad et al., 2007).
In particular, the projection of midbrain dopamine neurons in the ventral tegmental area (VTA) to the nucleus accumbens (NAc) and other limbic brain regions is a crucial neural substrate upon which drugs of abuse (e.g., cocaine, nicotine, morphine) exert their effect; and thus this projection is often referred to as the brain ''reward circuit''. This immediately suggests that compulsive consumption of high-fat/high-sugar foods and the consequent chronic spiking of brain reward circuitry may produce molecular changes similar to those that lead to drug addiction. Thus, as with drug addiction, repeated exposure to high-fat/high-sugar foods may lead to heightened craving for and dependence on these calorie-rich foods resulting in a downward spiral of compulsive over-consumption and ultimately obesity.
In the brain, both the hypothalamus and the brain reward circuit expresses several enzymes involved in the transport, manipulation and metabolism of both free fatty acid and triglyceride rich particles (TG). Elevated circulating levels of long-chain fatty acids (LCFAs), which are often elevated in obese and diabetic individuals, have recently been demon-strated to be an important signaling molecule acting in the hypothalamus to control food intake and energy expenditure (Kim et al., 2002;Ronnett et al., 2005;Ronnett et al., 2006). In addition, component of the reward circuitry also expresses several enzymes involved in the processing of triglyceride and fatty acid including lipoprotein lipase (LPL) (Ben-Zeev et al., 1990;Paradis et al., 2004), an enzyme responsible for triglyceride hydrolysis. Recent studies have provided have highlighted a role for neuronal LPL-mediated hydrolysis of TG particles in the regulation energy balance Wang and Eckel, 2012).
High fat feeding and in general lipid & sugar rich diet in both human and rodent has been associated with the progressive loss in spontaneous locomotor activity, and it has been shown that an acute reduction in locomotor activity in a major contributor to western diet-induced obesity in mice (Bjursell and Gerdin et al., 2008). Moreover high fat/high sucrose diet were also link to the establishment of an addictive-like reward dysfunction and compulsive eating in obese rats (Johnson and Kenny, 2010). Therefore it is tempting to speculate that TG-rich particles, accumulated during high fat feeding and in obese state could directly affect the mesolimbic reward/goal directed dopaminergic circuitry.
This review is an attempt to provide the neuro-anatomic framework together with recent update in the literature supporting the hypothesis in which nutritional lipids could be detected in different area of the brain and could serve as nutritional signal to regulate food intake in its rewarding and homeostatic component.

Hypothalamic regulation of feeding: the Melanocortin-Signaling Pathway
The hypothalamus consists of several nuclei involved in food intake, including the arcuate nucleus (ARC), the paraventricular nucleus (PVN), the lateral hypothalamic area (LHA), the ventromedial nucleus (VMH), and the dorsomedial nucleus (DMH). ARC neurons are located at the bottom of the hypothalamus around the third ventri-cle. They are called ''first order neurons'' because they directly contact the bloodstream and, thus, circulating satiety factors like insulin, ghrelin or leptin. This is due to an anatomical property of the ARC to be surrounded by the Median Eminence (ME) where the Brain Blood Barrier (BBB) is less developed. Two distinct groups of ARC neurons are directly controlling energy balance, the neurons containing orexigenic neuropeptides, agouti-gene-related protein (AgRP) and neuropeptide Y (NPY) (NPY/AgRP Neurons), and the neurons containing anorexigenic neuropeptides, pro-opiomelanocortin (POMC) and cocaine-and amphetamine-related transcript (CART). Both groups also express leptin receptors and insulin receptors. ARC neurons project to 'second order neurons' in the PVN, VMH, DMH and LHA which, in turn, project to other brain areas, essential for the longterm regulation of energy homeostasis, like Nucleus of the Solitary Tract (NTS). The ARC-PVN axis defines the so-called melanocortin-signalling pathway and is a key neuro-circuit for the regulation of feeding, any mutation of the melanocortin system leads to obesity and hyperphagia in both human and rodent (Kalra et al., 1991;Beck, 2001).

Hypothalamic sensing of LCFAs and metabolic syndrome
Plasma fatty acids (FA) as informative molecules acting on brain have been poorly studied for long time, as they were thought unable to cross the blood-brain barrier. However, growing amount of data attest that cerebral lipids come from both local synthesis and plasma origin (Rapoport, 2001). Furthermore, it must be pointed that structure and function of the BBB within hypothalamusespecially the ME and associated circumventricular organs (CVO)is quite different from other part of the brain (review in (Abbott et al., 2006). In particular, CVO are characterized by their small size, high permeability and fenestrated capillaries (Abbott et al., 2006). Thus, substances that do not cross BBB in other part of the brain may be imported in hypothalamus and it has been demonstrated using brain uptake index (BUI) method that palmitate uptake in hypothalamus is about 10 to 15% whereas less than 2% in other brain areas (Rapoport, 2001).
There is now growing amount of evidence that ''central lipid sensing sensing'' is involved in the control of feeding behavior.
Central infusion of LCFAs such as oleate physiologically inhibits both hepatic glucose production and food intake (Obici et al., 2002) and it was demonstrated that neuronal subpopulations of the arcuate nucleus could be either excited or inhibited by LCFAs (Wang et al., 2005). Genetic or pharmacological inhibition of enzymes involved in lipid metabolism like the Fatty Acid Synthase (FAS) or the Carnitine Palmitoyltransferase 1 (CPT-1) leads to a decrease in food intake, hepatic glucose production and are often associated with decreased expression of NPY and AgRP Pocai et al., 2005). The molecular mechanisms relaying fatty acid action involve metabolites such as acylCoA or malonylCoA, which could be major fuel sensors and most of the enzymes involved in lipid metabolism are detected at high level in the hypothalamic nucleus that control feeding behavior like the ARC, DMN and VMN (Kim et al., 2002;Sorensen et al., 2002). Moreover, central administration of C75, a potent inhibitor of FAS, also increases malonyl-CoA concentration in the hypothalamus, suppresses food intake and leads to profound weight loss (Gao and Lane, 2003). It has been proposed that centrally, C75 and ceruleninanother inhibitor of FASalter the expression profiles of feeding-related neuropeptides (such as NPY), often inhibiting the expression of orexigenic peptides (Gao and Lane, 2003). C75 also increases energy consumption, which contributes to weight loss. In vitro and in vivo studies demonstrate that at least part of C75's effects are mediated by the modulation of AMP kinase (AMPK), a known peripheral energy-sensing kinase (Ronnett et al., 2005). Indeed, ICV. Administration of AICAR (5-aminoimidazole-4-carboxamide ribonucleoside), a 5'-AMP kinase activator, rapidly lowers hypothalamic malonyl-CoA concentration and increases food intake (Hu et al., 2005). These effects correlate closely with the phosphorylation and thus inactivation of the Acyl-CoA Carboxylase (ACC), an established target of AMPK. Collectively, these data suggest a role for fatty acid metabolism in the perception and regulation of food intake .
Thus, LCFAs appeared to modulate neuronal activities in hypothalamic areas playing a critical role in the control of energy homeostasis.
However, it must be pointed out that, unlike plasma triglyceride-rich particle which rise after a meal, plasma level of LCFAs actually decrease due to the combined action of lipolysis inhibition and hyperinsulinemia (Ruge et al., 2009). Indeed, during a meal the hydrolysis and absorption by the gut leads to triglycerides packaging and through chylomicrons synthesis. Moreover the liver also produces very low density lipoprotein (VLDL). Thus TG-rich particles, accumulating after a meal, could also pretend to be a physiologically relevant satiety signal acting in the brain to regulate feeding behavior and energy expenditure. One would have to assume that TG particle are entering the brain and can be locally hydrolyzed into areas mentioned above (hippocampus, striatum, and hypothalamus).
The Mesolimbic system: a brain circuits encoding the motivational and hedonic aspects of food intake Palatability and pleasantness are powerful determinant of food intake. Ingestion of food provides subjective pleasure, particularly if the food is rich in sugar or fat, and eating can be a source of comfort in depression or stressful states (Dallman et al., 2003). The positive emotion or pleasure of tasting sugar or fat foods may have evolved to guarantee sufficient intake of varied food and high-energy foods (Kelley et al., 2002). A feeding induced hedonic response is positively reinforcing, and driving an organism to repeat ingestion of calorically dense food. Dopamine (DA) is involved in the reinforcing effects of food (Comings and Blum, 2000). Dopamine antagonist increase appetite, energy intake and weight gain, whereas dopamine agonist reduce energy intake and produce weight loss (Leddy et al., 2004;de Leon et al., 2007). DA binds to several G-protein-coupled dopamine receptors-D1 and D5 coupled to G as ; D2, D3 and D4 coupled to G ai -which activate intracellular signaling pathways in postsynaptic neurons. DA neurons in the Ventral Tegmental Area (VTA) and the Substancia Nigra innervate the Nucleus Accumbens (Nacc), Caudate Putamen (Cpu), cortical regions (prefrontal cortex (PFC), orbitofrontal cortex (OFC)) and limbic regions (amygdala (AG), hippocampus). During regular circumstances, DA neurons are firing continuously causing tonic DA release; salient environmental stimuli induce burst-firing of DA neurons resulting in DA release (Palmiter, 2007). Feeding is associated with DA release in the dorsal striatum (St), and the degree of pleasure from eating correlates with amount of dopamine released (Szczypka et al., 2001;Small et al., 2003). DA is also involved with the motivation to perform the behaviors necessary to procure and consume the food. In fact, Dopamine-deficient mice are hypophagic, fail to engage in goaldirected feeding behaviors, and will die of starvation (Szczypka et al., 1999;Palmiter, 2008). If DA signaling is restored only to the Cpu, mice will eat sufficiently for survival, whereas restoration of DA in the Nacc is not sufficient to sustain long-term feeding (Szczypka et al., 2001;Palmiter, 2008). These results show that Cpu has a role in mediating motivational aspect of feeding whereas Nacc does not. Moreover, it was found that chemical lesions of the DA system (pharmacological antagonists of DA receptor), sufficient to render the animals completely incapable of initiating behavior aimed at obtaining food, did not affect the hedonic-like reactions to sucrose placed in the mouth (Tyrka et al., 1992;Berridge and Robinson, 1998). This finding was interpreted as indicating that the neural systems mediating the hedonic impact of preferred tastes was distinct from that mediating the ability of incentives to elicit goal-directed behavior; these two processes were termed ''liking'' and ''wanting'', respectively. With regard to neurochimical mediation, it was proposed that liking depends upon opioid transmission and wanting is governed by the mesolimbic DA system (Kelley et al., 2005). Opioid peptides that are produced in the body include: endorphins, enkephalins, dynorphins, endomorphins. The striatum contains enkephalin and b-endorphin (Pickel et al., 1980;Sweep et al., 1989). Opioid receptors are found throughout the brain including Striatum, VTA and HT (Mansour et al., 1987). There are also found in neural pathways thought to relay taste information (Nucleus Tractus Solitarius (NTS), the parabrachial nucleus (PB) and the AG) (Mansour et al., 1987;Xia and Haddad, 1991). Opioid receptor agonist enhance food intake by increasing the hedonic valence of food (Kotz et al., 1997;Echo et al., 2002;Wilson et al., 2003;Levine et al., 2004;Kelley et al., 2005). Local injection of the selective m-opioid receptor agonist DAMGO into the nucleus accumbens results in increased intake of both fat and carbohydrate-rich diets when either are presented alone (Zhang and Kelley, 1997;Zhang et al., 1998). Interestingly, if rats were given a choice between both diets, DAMGO treatment strongly increased their intake of the fat diet without affecting carbohydrate consumption, suggesting that the fat diet is inherently more palatable (Kelley et al., 2005).
With sensorial or emotional signals, circulating factors like hormones or nutrients which encoding nutritional status of the organism, influence brain regions that regulate the motivation to consume high-calorie foods.
Circulating factors also provide signals to the mesolimbic system: implication for the regulation of feeding behavior A variety of hormones, such as leptin, insulin and ghrelin can directly modulate DA neuron activity. The role of these hormones in regulating homeostatic circuits in the hypothalamus has been examined extensively, but studying their effects on the DA reward circuit is a new direction.
Leptin, insulin and ghrelin receptors are expressed in the VTA (Figlewicz et al., 2003;Abizaid et al., 2006;Fulton et al., 2006;Hommel et al., 2006;Pardini et al., 2006). Intracerebroventricular (i.c.v.) insulin administration increased the mRNA levels and the functional activity of DA transporter (DAT). Authors suggested that this increase would facilitate clearing DA from synapse and hence reduce DA signaling (Figlewicz et al., 1994;Figlewicz et al., 2003). Moreover, the firing rate of DA neurons in anesthetized rats decreased in response to an intravenous infusion of leptin (Hommel et al., 2006) and i.c.v. leptin infusion reduced extracellular dopamine in the Nacc (Krugel et al., 2003). Inversely, ghrelin increased the firing rate of VTA DA neurons and i.c.v. and systemic administration of ghrelin induced a DA over-flow in the Nacc (Abizaid et al., 2006;Jerlhag et al., 2006;Quarta et al., 2009). In humans, a functional MRI experiment revealed that ghrelin injection activates brain regions associated with reward processing of appetitive behaviors including AG and Striatum (Malik et al., 2008). Behavioral evidences showed that these hormones, by modulating DA system, directly affect feeding behavior. In fact, i.c.v. insulin or leptin administration reduced sucrose self administration, suppressed the conditioned place preference (CPP) for sucrose pellets and reversed CPP for high-fat (HF) food (Figlewicz et al., 2003;Figlewicz et al., 2004;Figlewicz et al., 2006). Moreover, leptin administration to the VTA decreased food intake (Hommel et al., 2006;Morton et al., 2009). Authors proposed that a reduction in CPP by these hormones suggested that they interfere with subjective reward value, and associative learning of food. Moreover, i.c.v. leptin and insulin infusion or more specifically in the VTA decreases the feeding and sucrose intake induced by opioid receptor agonist (m and k) (Sipols et al., 2002;Figlewicz et al., 2007). On the other hand, ghrelin delivery to the VTA or to the Nacc stimulated food intake in a dose dependant manner (Naleid et al., 2005;Abizaid et al., 2006). Thus insulin and leptin provide inhibitory inputs fo DA neurons in the VTA which can decrease dopaminergic tone in postsynaptic areas and ghrelin activates DA neurons activity, stimulates DA turnover and release in the Nacc. Food restriction lower serum levels of insulin and leptin while increasing level of ghrelin; each of these individual effects predicted to enhance DA neuron activity. Thus is consistent with the idea that food restriction enhanced the subjective value of food (Carr, 2002). However there is disconnection between DA signaling and eating behavior, thus it is premature to ascribe feeding and reward behaviors exclusively to DA release (Palmiter, 2007).
We have seen that circulating hormones can modulate mesolimbic system by modulating dopaminergic tonus essentially. Other circulating parameters like plasmatic glucose encoding nutritional status can also influence mesolimbic system. In fact systemic glucose administration suppressed the firing of dopaminergic neurons within the SN (Saller and Chiodo, 1980) and local glucose in the SN increased GABA release (During et al., 1995). Authors suggested that glucose acts as a signaling molecule in DA neurons which played an important role in maintaining motivated feeding behaviors especially in response to abrupt decreases in glucose use (Saller and Chiodo, 1980). However, in this study, insulin was not measured and authors put forward a potential role of insulin in dopaminergic transmission changing. Interestingly, a recent study, succeed to isolated the physiological stimulus glycemic state in humans. By using, functional MRI combined with a stepped hyperinsulinemic euglycemic-hypoglycemic clamp and behavioral measures of interest in food, they shown that hypoglycemia activates limbic striatal brain regions (Nacc, Insula, Ht, Thalamus and Cpu) in response to food cues to produce a greater desire for high-calorie foods.
Thus, hormones and nutrients that regulate energy balance through the hypothalamus also modulate the activity of dopamine cells and their projections into regions involve in the rewarding aspect of food intake. Hypothalamic signaling disruption is now recognized as a major component in metabolic syndrome development (Schwartz et al., 2000;Prodi and Obici, 2006). However, accumulating evidences show that disruptions in brain circuits, other than those regulate hunger and satiety, are involved in feeding behavior disorders and, consequently, obesity development.

Disruptions of mesolimbic system activity associated with obesity and high circulating lipids
Growing evidences show that obesity is associated with deficits in mesolimbic system.
In rats, striatal levels of D2R were inversely related to body weight (Johnson and Kenny, 2010). Importantly, in this study, authors liked the decrease of striatal D2R to compulsive food intake by lentivirus-mediated knockdown of striatal D2R which rapidly accelerated the development of addiction-like reward deficits in rats with extended access to palatable HF food (Johnson and Kenny, 2010). Moreover, another study showed that D2R activation (subcutaneous pellet of bromocriptine/i.p. bromocriptine injection) partly redirected HF diet induced metabolic anomalies in obesity-prone mice (Davis et al., 2009;de Leeuw van Weenen et al., 2011). Conversely, blocking D2R (subcutaneous pellet of haloperidol) induces an adverse metabolic profile in mice that are inherent resistant to the deleterious effect of HF food (de Leeuw van Weenen et al., 2011). Besides D2R alteration, D1R and m-opiod receptor mRNA was down-regulated after longterm (5 weeks) high fat high sucrose (HFHS) access in obesity-prone rats compared with obesity resistant rats (Alsio et al., 2010). In human, striatal D2R availability was significantly lower in obese individuals and the body mass index (BMI) correlated negatively with measures of D2R; the individuals with the lowest D2 values had the largest BMI (Wang et al., 2001;Volkow et al., 2008). Obese versus lean humans showed less activation of striatal DA target regions in response to palatable food intake (Stice et al., 2008) whereas they showed greater activation of regions that process palatability and motivation (Nacc, St, orbitofrontal cortex (OFC), AG. . .) when they anticipated consumption (stimulated with pictures of high-calorie food) (Rothemund et al., 2007;Stice et al., 2008;Stoeckel et al., 2008). Thus suggests that an enhanced activity of regions that process palatability could make obese subjects favor food over other natural reinforcers (Volkow et al., 2011). Taken together, rats and humans data suggest that an attenuated activity of dopaminergic would reduce the hedonic response associated with feeding, a deficit that obese individuals might strive to compensate by overeating (Geiger et al., 2008;Stice et al., 2008;Volkow et al., 2011).
However, there is evidence that consumption of palatable food per se leads to DA signaling alterations. 40 days of cafeteria diet consumption decreased striatal D2R expression and induced brain reward dysfunctions (Johnson and Kenny, 2010). However, at the early stage of obesity, HF diet consumption (20 days) significantly increased D2 receptor (D2R) binding in the Cpu and the Nacc, but decreased DAT binding in the Cpu. Authors suggest that combination of higher levels of D2 and low levels of DAT binding suggest they may have had lower striatal dopaminergic activity compared to a low-fat fed group (South and Huang, 2008). What is consistent with the fact that electrically evoked DA release from Nacc slice preparations and extracellular DA level in the Nacc (microdialysis) was markedly reduced in cafeteria diet induced obesity animals (15 weeks) compared with chow fed animals (Geiger et al., 2009). Moreover, some studies showed that sucrose intake by itself influence dopamine signaling in the brain (Colantuoni et al., 2001;Bello et al., 2002).
As it has been shown in the hypothalamus (Arase et al., 1988), HF diet can also alter mesolimbic insulin action. In fact, HF consumption (28 days) induced impairment in striatal activation of the insulin-activated signaling kinase (Akt). HF-induced Akt impairment reduced DAT cell surface expression and function, thereby decreasing DA homeostasis and DA-related behaviors such as AMPH-induced locomotion and increase caloric intake. Restoration of nigro-striatal Akt phosphorylation using recombinant viral vector induced a rescue of DAT expression in HF fed rats which was associated with a normalization of HF-diet induced hyperphagia (Speed et al., 2011). All these data imply that overeating palatable food per se may contribute to a further attenuation reward circuitry. In humans, a fMRI study indicated that weight gain in overweigh and obese young women was associated with a reduction in striatal activation in response to palatable food intake relative to baseline response six month before. As we discussed above, the attenuation of reward circuitry induced by palatable food consumption could reduce the rewarding and hedonic value of food. Consisting with this hypothesis, it has been shown that animal consuming HF diet, independent of the development of obesity, exhibit decreased dopamine turnover in the OFC and Nacc, what is associated with behavioral defaults like a reduced preference for an amphetamine cue, and an attenuated operant responding for sucrose (Davis et al., 2008).
We have seen that palatable food consumption could directly alters rewarding system, causing inappropriate feeding behavior and leading to obesity. Because the palatability of food is typically a function of its fat and sugar composition (Drewnowski et al., 1989), we can suppose that these nutrients, particularly lipids, could act directly on mesolimbic system to modulate the motivational and rewarding aspect of food intake.

Lipids consumption could induce a disruption of the mesolimbic system
It is important to note that independently of weight gain, at the early stage of obesity development, fat consumption lead to dopamine system dysfunction. In fact, animals consuming high fat diet diet prior to obesity development exhibit a decreased in dopamine turnover (Davis et al., 2008), restricted access to cafeteria diet decreased striatal expression of D2R (Johnson and Kenny, 2010), and HFHS pair fed decreased D1R and D1R Nacc expression, relative to chow fed controls (Alsio et al., 2010). Above, we have discussed the fact that circulating factors like hormones or nutrients (glucose) could modulate mesolimbic system by modulating dopaminergic tonus essentially. We can suppose that nutritional lipids present in HF, HFHS and cafeteria diet could directly modulate mesolimbic system and may contribute to a further attenuation of mesolimbic responsivity to food, as it is suspected in other brain areas. In fact, there is increasing evidence that obesity is associated with impairment on certain cognitive function, such as executive function, attention and memory (Gunstad et al., 2007;Bruce-Keller et al., 2009;Bruehl et al., 2009). Although co-morbid medical conditions (cerebrovascular pathology, hypertension and diabetes) are known to affect cognition adversely, there is also evidence that high BMI, by itself, might impair various cognitive domains particularly executive function (Gunstad et al., 2007) which could be explain by a lower baseline prefrontal (OFC included) metabolism (Volkow et al., 2009). Even though the correlation between lipid levels and cognitive function is still debated, a recent study has shown that impaired cognition in obese mice was improved by selectively lowering triglycerides with gemfibrozil (Farr et al., 2008). Moreover it was shown that i.c.v injection of the triglyceride triolein impaired acquisition in several cognitive paradigms in normal body weight mice. Authors suggested that triglycerides, by itself, mediate cognitive impairment as seen in obesity such as food reward lever press (Farr et al., 2008). In agreement with this hypothesis, when the orosensory reward of fat is bypassed with i.p. injection of triglycerides (TG) the increase in TG levels rats leads to an increase in both DA and and DA metabolites in the Nacc, similar what is observed after the fat consumption (Rada et al., 2010). These result suggest that, as glucose, TG could act as a signaling molecule in dopamine system and could be part of the reinforcing effects of palatable food. Interestingly, obesity prone rats still at normal weight (based on their weight gain during 5 days of access to HF diet) had levels of TG consistently higher in response to high-fat meal or injection of Intralipid compared to obesity-resistant rats. We can suppose that brain overexposure to TG could induce desensitization to TG signaling in the dopamine system. This could in turn reduce DA responsiveness to circulating lipids which may contribute to the failure of obesity-prone rats to properly regulate their intake of a highfat diet. This is consistent with the observation that high-TG responders showed an increased propensity to overeat and gain weight on a chronic HF diet (Karatayev et al., 2009).

Conclusion
Several observations led to the identification of the hypothalamus as a major integration site for inputs related to energy homeostasis (Schwartz and Porte, 2005). The hypothalamus contains specialized neurons that monitor circulating factors of hunger and satiety in order to provide an integrated adaptive response at both the metabolic and behavioral levels. This homeostatic control of energy homeostasis tends to keep body weight within a narrow range.
However, aside from hypothalamic-mediated homeostatic control, feeding behaviour is also modulated by sensory inputs, such as tastes and odors, as well as by affective or emotional states. For example, stress or anxiety can stimulate reward seeking and consumption of highly palatable food independent of metabolic demand (Dallman et al. 2005). The reinforcing and motivational aspects of food are closely tied to the release of the neurotransmitter dopamine, which is stimulated by high-fat/high-sugar foods as well as by most other objects of desire (e.g., sex, drugs) (Wise, 2006). In particular, the projection of midbrain dopamine neurons in the ventral tegmental area (VTA) to the nucleus accumbens (NAc) and other limbic brain regions is a crucial neural substrate upon which drugs of abuse (e.g., cocaine, nicotine, morphine) exert their effect; and thus this projection is often referred to as the brain ''reward circuit''. Strong evidence indicates that chronic exposure to drugs of abuse leads to molecular changes in the neurons of the brain reward circuit which underlie key aspects of the addictive state such as craving and dependence (Wise, 2006).
A widely accepted conceptual framework sees food intake as the result of a complex behavioral sequence that entail both ''homeostatic'' inputs, arising from the hypothalamus and emotionalcognitive inputs encode by the mesolimbic system. The two components integrate several circulating factors of hunger and satiety and communicate with one another to adapt both metabolic demands and the hedonic and motivational aspect of food intake. Although it is generally accepted that the brain does not uses lipids as a primary source of energy, it expresses several enzymes involved in the transport, hydrolysis and manipulation of TG-rich particle and LCFAs (Kim et al., 2002;Ronnett et al., 2005;Ronnett et al., 2006). Central LCFAs sensing in the hypothalamus was recently demonstrated to be a key mechanism for the homeostatic regulation of food intake Migrenne et al., 2011). Beside LCFAs that only increase in the bloodstream during fasting condition, TG-rich particle are found to be elevated in most obese and high fat condition. TG hydrolyses involved lipoprotein lipase activity, and the genetic invalidation of the gene encoding the Lpl, selectively in neurons translate into obesity and feeding alteration Wang and Eckel, 2012). This result strongly suggests a key role for TG-rich particle detection and hydrolysis in brain areas involved in feeding. Both the hypothalamus and the reward neurocircuitry express high level of LPL (Ben-Zeev et al., 1990;Paradis et al., 2004), and could therefore relay the action of TG-rich particle, and in general high fat associate hypertriglycemia, in the long-term alteration of both homeostatic regulation of food intake as well as the cognitive and emotional aspect of feeding