Molecular Mechanisms of Fat Preference and Overeating

Dany Gaillard, Patricia Passilly-Degrace, and Philippe Besnard

Physiologie de la Nutrition, UMR INSERM U866, Ecole Nationale Sup erieure de Biologie ` la Nutrition et a ` lAlimentation (ENSBANA), Universit Appliqu ee a e de Bourgogne, Dijon, France

Obesity is recognized as a worldwide health problem. Overconsumption of fatty foods contributes signicantly to this phenomenon. Rodents, like humans, display preferences for lipid-rich foods. Rodents thus provide useful models to explore the mechanisms responsible for this complex feeding behavior resulting from the integration of multiple oral and postoral signals. Over the last decades, the lipid-mediated regulation of food intake has received considerable attention. By contrast, orosensory lipid perception was long thought to involve only textural and olfactory cues. Recent ndings have challenged this limited viewpoint. These recent data strongly suggest that the sense of taste also plays signicant roles in the spontaneous preference for fatty foods. This paper provides a brief overview of postoral regulation of food intake by lipids and then highlights recent data suggesting the existence of a fatty taste which might contribute to lipid overeating and hence to the risk of obesity. Key words: dietary lipid perception; gustation; CD36; addiction; overeating; obesity

Introduction Eating is a complex form of behavior governed by combination of physiological, hedonic, cultural, and perhaps even philosophical factors. The profound technical and economic changes of the 20th century have deeply affected our lifestyle and, consequently, our eating behavior. For the rst time in its history, most of the worlds population no longer has to run after calories. One direct consequence of this fundamental change has been the emergence of plethora of diet-related diseases that provide major public health problems. One of the most prominent examples is obesity. Obesity has reached epidemic proportions worldwide and is a major contributor to the global burden of chronic diseases that include type 2 diabetes, atherosclerosis, hypertension, and cancers. Obesity increasingly affects
Address for correspondence: Prof. Philippe Besnard. Physiologie de la Nutrition, ENSBANA, 1 esplanade Erasme, 21000 Dijon, France. Voice/fax: +33-(0)3-80-39-66-91. pbesnard@u-bourgogne.fr

both adults and children.1 Recent data even suggest that the increasing prevalence of obesity is associated with a decreasing life expectancy in children.2 An abundance of food has obvious consequences: it promotes our specic appetites. Lipids account for about 40% of the calories ingested in Western countries; whereas nutritional recommendations are 5 to 10% lower. This preference for fatty foods is not specic to humans. Rats and mice spontaneously prefer lipid-rich foods if provided with a free choice.3,4 Attraction to lipids is so strong that mice who are provided free access to an oil as optional diet rapidly become obese.5 Relationships between excessive fat consumption and obesity are thus obvious in both rodents and humans. Regulation of food intake is a complex phenomenon that results from integration by the brain of early orosensory (cognitive) cues and delayed postprandial (metabolic) signals (Fig. 1). All these signals converge on specic areas of the central nervous system, in which they are integrated to induce stereotyped physiological
2008 New York Academy of Sciences. 163

Ann. N.Y. Acad. Sci. 1141: 163175 (2008). doi: 10.1196/annals.1441.028

164

Annals of the New York Academy of Sciences

Figure 1. Oral and postoral pathways involved in the regulation of food intake in mice. Eating behavior results from the integration of early (oral) and delayed (postoral) signals by the central nervous system. Early events in fat feeding consist of olfactory, textural, and gustatory cues while delayed events are associated with postingestive and postabsorptive signals. Gustatory information via the chorda tympani (VII), glossopharyngeal (IX), and vagus (X) nerves and postingestive/absorptive information via the X nerve converge on the nucleus of the solitary tract (NST) in the brain stem that connects to central regulatory areas like the nucleus accumbens (NAc) and the hypothalamus (HT), which constitute a pleasure lter and metabolic lter, respectively. The NST also projects X efferents toward the digestive tract, which can account for the cephalic phase of digestion triggered by oral lipid stimulation facilitating fat digestion and absorption. HT activity can also be directly and/or indirectly modulated by plasma factors (hormones, regulatory peptides, lipids) involved in postingestive/absorptive events. All together, these regulatory factors account for the preference for fatty foods and thus eating behavior. I, olfactory nerve; V, trigeminal nerve; ApoA-IV, apolipoprotein A-IV; CCK, cholecystokinin; FFA, free fatty acid; GLP1, glucagon-like protein 1; PYY, peptide YY; TG, triglyceride.

effects that result in food preferences or aversions. Size and frequency of meals depend on this multimodal system, which regulates the bodys energy balance. This homeostatic effect is highly related to the nutrient composition of foods. For instance, regulatory effects of lipids on food intake are less efcient than regulation by carbohydrates or proteins.6 Therefore, chronic high-fat diets promote greater daily intake by eliciting larger and more frequent meals and increase the risks of obesity. To explore origins of this nutrient specicity, considerable at-

tention has been devoted to postoral regulation of food intake by lipids. Postoral Regulation of Food Intake by Dietary Lipids Lipid-mediated regulation of food intake results from integration of multiple short-acting (postingestive) satiety signals and long-acting (postabsorptive) homeostatic signals (Fig. 1). Once ingested, lipids trigger a set of rapid

Gaillard et al.: Molecular Mechanisms of Fat Preference and Overeating

165

regulatory events that limit food intake. Cholecystokinin (CCK) is a key regulatory peptide of this pathway. CCK is released by the proximal intestine in response to dietary lipid loads. CCK has been demonstrated to be a shortterm, meal-reducing signal in humans and many other mammals. Recent data demonstrate that this effect is mainly mediated by paracrine regulation of vagal afferent pathways that ow through the Xth cranial nerves and is relayed to the hypothalamus via the nucleus of the solitary tract (NST) in the brain stem (Fig. 1). Indeed, mesenteric vagal afferent terminals express densities of CCK1R receptors.7 These localizations help to explain why exogenous injections of CCK induce action potentials in vagal afferents and limit food intake.7 Interestingly, chronic high-fat diet is associated with a reduced vagal sensitivity to peripheral CCK injection in rodents.8 This desensitization could be due to dynamic regulation of CCK1R in vagal afferent neurons, since the numbers of these receptors decrease rapidly in response to fat ingestion.7 The vagus nerve also expresses receptors for other shortand longer-term satiety agents. The short-term agents include glucagon-like protein 1 (GLP1) and peptide YY (PYY), which are released by ileal enteroendocrine cells in response to fat. Long-term satiety agents include leptin, which is produced by the adipose tissue (Fig. 1). The expression levels of the receptors for these peptides by vagal neurons seem to be downregulated by lipids, as found for the CCK1R.7 This dynamic regulation might account for some of the reduced ability of lipids to satiate, in comparison to carbohydrates or proteins. Satiation, which largely determines the size of meals, mainly depends on postingestive signals. Postprandial satiety is largely responsible for meal frequency and is essentially related to postabsorptive signals (Fig. 1). The apolipoprotein A-IV (ApoA-IV) is a regulatory protein that inhibits the onset of feeding (that promotes satiety).9 During postprandial period, ApoA-IV is secreted by the small intestine in association with chylomicrons, triglyceride-rich lipopro-

teins responsible for fat absorption. Peripheral or cerebroventricular injections of ApoAIV decrease food intake in a dose-dependent manner in rats.9 As found with other satiating agents, the inhibitory effects of ApoA-IV on feeding becomes less efcient when rats or humans have been subjected to chronic high-fat diets. The origin of this effect has remained elusive. Involvement of the adipocyte hormone leptin has been proposed. Leptin administration does greatly decrease the lipidmediated induction of ApoA-IV expression in rats.10 Blood leptin levels are especially high during obesity. Thus, leptin-mediated inhibition of ApoA-IV release might contribute to the reduced satiety sensitivity observed during chronic exposition to fatty foods.9 Such reduced satiety could help to explain the overfeeding frequently found in obese animals and humans. Surgically parabiotic animals with crossed blood circulations have been used to demonstrate that humoral factors also play a role in the regulation of feeding behavior. The blood brain barrier is discontinuous near the arcuate nucleus of the hypothalamus, providing access for aforementioned lipid-mediated satiating regulators, including CCK, GLP1, PYY3 36, ApoA-IV, and leptin and to nutrients like free fatty acids (FFA), to the central nervous system (Fig. 1). It has been recently shown that FFA can modulate feeding behaviors through direct actions on the brain. Central administration of oleic acid inhibits food intake in rats.11 This event seems to be mediated by FA-sensitive neurons identied in the arcuate nucleus.12 While the molecular mechanisms involved in this regulation are not yet fully established, FFA-mediated control of ion channels and/or FA binding to specic receptors in FA-sensitive hypothalamic neurons provide plausible possibilities. FA metabolites such as malonyl-CoA might also play a signicant role in this neuronal activation.11 Therefore, release of FFAs from plasma triglycerides (TG) and transfer of these FFA through the bloodbrain barrier can provide postprandial satiety signals to the central nervous system.11

166

Annals of the New York Academy of Sciences

This nonexhaustive overview clearly documents some of the many lipid-mediated postoral factors that are likely to interact to regulate food intake. It also highlights the progressive elucidation of molecular mechanisms responsible for the relative weak satiety action of lipids. Spontaneous fat preference observed in rodents and humans depends not only on internal homeostatic mechanisms but also on the hedonic quality of lipids. Until recently, it was thought that palatability of fatty foods was perceived only through somesthesic and olfactory cues (Fig. 1). This restricted view has been challenged by recent observations that suggest that the sense of taste is also involved in the preference for lipid-rich foods (Fig. 1). Rodents Display a Spontaneous Attraction to Fat There is compelling evidence to suggest that laboratory rodents have an orosensory system devoted to long-chain fatty acid (LCFA) detection. The two-bottle preference test is a classical method for studying feeding behaviors of animals in a free-choice situation. This simple paradigm clearly shows that rats3 and mice4 display a strong preference for lipid-rich solutions. However, this observation is not easy to interpret because food preference results from the integration of olfactory, somesthesic, gustatory, and postingestive signals (Fig. 1). Thus, the relative importance of each of these parameters on fat preference has been explored in experiments that provide several key observations. First, spontaneous fat attraction is maintained in anosmic rats and mice, in which olfaction is blocked by chemical means. Therefore, smell does not play a signicant role in this behavior.13,14 Second, the two-bottle preference test clearly reveals that mice prefer vegetable oils to agents that provide similar textures, such as xanthan gum. Texture thus does not appear to provide a major cue in orosensory fat perception.4 This observation is supported by

ndings from conditioned taste aversion (CTA) testing, in which naive animals learn to avoid a newly encountered molecule after suffering temporally conditioned adverse postingestive effects, such as those triggered by intraperitoneal lithium chloride injection. Aversion to a sucrose/corn-oil mixture cannot be induced in rats by replacing the corn oil with mineral oil, which is indigestible but provides a similar texture. The chemical composition of the oil, rather than its textural characteristics, thus plays a key role in the perception of this mixture.15 Interestingly, rats previously conditioned with corn oil display a stronger aversion to the sucrose/corn oil mixture than animals previously conditioned with sucrose alone, indicating that the corn oil provides the salient features of the mixture.15 Third, fat preference is not abolished in very short-term experiments (0.55 min) that are designed to minimize postingestive effects.3,15,16 The persistent attraction of lipids to rats or mice in which textural, olfactory, and postingestive cues have been simultaneously suppressed strongly suggests that taste inuences lipid-preferring feeding behaviors in rodents.13,14 Gustation Plays a Role in Fat Preference In mammals, the sense of taste acts in concert with olfaction and with somatosensory information related to touch, texture, temperature, and pain to indicate whether a food can be ingested (Fig. 1). Taste perception is mediated by taste receptor cells (TRC), which are clustered in specialized onion-shaped structures, the taste buds. Taste buds are found at high densities on the tongue and at low densities on the soft palate, larynx, pharynx, and upper part of the esophagus. In the lingual epithelium, taste buds are located in three types of gustatory papillae with different spatial distributions. Some are located in fungiform papillae, which cover the front two-thirds of the tongue. These papillae are mushroom shaped and have small

Gaillard et al.: Molecular Mechanisms of Fat Preference and Overeating

167

numbers (13) of taste buds on their apical surfaces. The circumvallate and foliate papillae are located on the posterior third tongue, in the central and lateral regions, respectively. The circumvallate papillae display circular depressions whose walls contain several hundred taste buds.17 Humans have about 10 circumvallate papillae, whereas rodents have only one, positioned centrally. Foliate papillae are located at the posterior lateral edge of the tongue and contain hundreds of taste buds. TRC from fungiform papillae and some of the anterior foliate papillae establish synaptic contacts with the chorda tympani (CT) nerve, which forms part of the VIIth cranial nerves. The posterior parts of the foliate and circumvallate papillae are innervated by the glossopharyngeal (GL), IXth cranial nerve. Vagal (Xth cranial nerve) afferents innervate taste buds found outside of the oral cavity. These gustatory nerves transmit information to the brain and digestive tract via the NST in the brainstem. Taste signals to the brain trigger feeding behaviors, while signals to the digestive tract induce digestive secretions (Fig. 1). Which Lipids Are Detected? Although dietary lipids are mainly constituted of TG, LCFA (number of carbons 16) appear to be responsible for much of the oral lipid perception. In a free-choice situation, rats have a weaker preference for TG and mediumchain fatty acids (MCFA, number of carbons from 8 to 14) than for LCFA.3,14 This chemical selectivity is very tight, since LCFA derivatives, such as methyl-LCFA, are not recognized.3 The ability of rodents to specically detect LCFA has also been conrmed with the CTA paradigm. Its noteworthy that both rats and mice can be conditioned to avoid specic LCFA16,18 with remarkable submicromolar detection thresholds.16,19 Lingual lipase, which is responsible for an efcient release of LCFA from TG in rodents, seems to play a signicant role in oral fat perception in the mouse. Pharmacological inhi-

bition of lingual lipase dramatically decreases lipid preference.20 This may explain why mineral oil, which cannot be digested, is not as attractive as vegetable oil in a free-choice situation.19 Interestingly, lingual lipase is released directly into the clefts of foliate and circumvallate papillae from von Ebners glands in rodents.20 This anatomical feature appears to be ideal for efcient TG hydrolysis, generation of high LCFA levels close to the taste buds, and efcient detection of the LCFA by TRC. CD36 as a Lipid Sensor Chemosensitive proteins that are located at the apical sides of TRC, including ion channels, metabotropic, G proteincoupled receptors, and ionotropic receptors, are responsible for initial phases of taste perception. Interactions between a sweet, bitter, sour, salty, or umami tastant with its specic detection system lead to changes in membrane potential and/or intracellular free-calcium concentration ([Ca2+ ] i ). These changes trigger neurotransmitter release and generate ring in gustatory afferent nerve bers.21,22 The gustatory perception of lipids must therefore require the existence of receptors that are expressed by TRC and display high afnities for LCFA. One plausible candidate for this function has recently been identied: the receptor-like glycoprotein CD36. CD36 is a receptor-like protein that binds saturated and unsaturated LCFA with afnities in the nanomolar range.23 It has the structural and functional features required for a putative taste-based lipid receptor. First, CD36 expression appears to be restricted to the gustatory epithelium in rodent tongues.24,25 In mice, CD36 is expressed particularly strongly in circumvallate papillae, to a lesser extent in foliate papillae, and only very weakly in fungiform papillae.25 Immunohistochemical staining has shown that CD36 protein lines the taste pores of the apical sides of some TRC.25 This distribution of a protein with a very high afnity for LCFA is particularly suitable for the generation

168

Annals of the New York Academy of Sciences

Figure 2. Working model for the gustatory perception of LCFA in the mouse. Long-chain fatty acids (LCFA) released from triglycerides (TG) by lingual lipase bind to the CD36, which acts as a gustatory lipid receptor in taste receptor cells ( ). The recognition of LCFA by CD36 induces an increase in intracellular free calcium concentration ([Ca2+ ] i ) ( ), an event known to generate the release of neurotransmitters by taste receptor cells. Lipid taste signal is then transmitted by the gustatory nerves, chorda tympani nerve (VII), and glossopharyngeal nerve (IX) ( ) to the gustatory area in the nucleus of the solitary tract (NST) in the brain stem ( ). The projections of the NST to central nuclei involved in eating behavior (that is, the hypothalamus, HT; and nucleus accumbens, NAc) and to peripheral tissues, including the digestive tract, could account for the CD36-mediated attraction to lipids ( ) and cephalic phase of digestion ( ) reported in mice subjected to an oral lipid stimulation.

of a lipid signal by taste buds. Indeed, CD36positive TRC are directly exposed to a microclimate that would be rich in LCFA following intake of lipids, due to the local release of lingual lipase in the clefts of circumvallate papillae20 (Fig. 2- ). Second, a role of CD36 as a lipid sensor is also supported by the predicted structure of this protein. This plasma membrane protein has a hairpin structure, with a large extracellular hydrophobic pocket located between two short cytoplasmic tails.26 Existence of physi-

cal interactions between the intracellular Cterminal tail of CD36 and Src protein-tyrosine kinases27 results in functional complexes that could facilitate transfer of an exogenous lipid signal into the TRC. Third, CD36 gene knockout abolishes both fat preference25,28 and the cephalic phase of digestion triggered by oral LCFA deposition.25 Taken together, these ndings strongly suggest that CD36 is a lipid receptor involved in the orosensory perception of dietary lipids in rodents.

Gaillard et al.: Molecular Mechanisms of Fat Preference and Overeating

169

Does the Gustatory Pathway Play a Role in CD36-Mediated Fat Perception? A key question at this stage concerned the possible mediation of oral lipid detection by the gustatory pathway. Studies of the lipid transduction signals in TRC and of the afferent nerve route used to transfer the fat signal to the central nervous system were required to address this question.

Mechanisms of Fat Signal Transduction

Tastant-induced release of neurotransmitters onto afferent nerve bers leads to orosensory perceptions of sapid molecules. This event is mediated by changes in [Ca2+ ] i in TRC.21 If lingual CD36 acts as a lipid receptor, LCFA binding to CD36 may also affect [Ca2+ ] i . We tested this hypothesis by determining [Ca2+ ] i in CD36-positive TRC isolated from mouse circumvallate papillae by afnity purication with magnetic beads.18 Saturated and unsaturated LCFA triggered a rapid and robust increase in [Ca2+ ] i in CD36-positive cells. This effect was strictly CD36-dependent, as it was not observed in CD36-negative cells. Moreover, addition of the specic CD36 binding inhibitor, sulfo-Nsuccinimidyl oleic acid ester (SSO)29 to the culture medium completely abolished the linoleic acid-mediated rise in [Ca2+ ] i in CD36-positive cells. These data provided the rst demonstration that LCFA increased [Ca2+ ] i in the taste bud cells, and that this event was CD36dependent18 (Fig. 2- ). The reception of tastes other than salty and sour requires the heterotrimeric gustducin G protein complex.30 CD36-expressing TRC also contain the -subunit of gustducin,18,25 but this G protein complex is not involved in fat preference. Indeed, preference for LCFAenriched solutions is maintained in -gustducin knockout mice.31 This result was expected, because CD36 does not belong to the G protein coupled receptor (GPCR) family, unlike the T1R and T2R receptors responsible for sweet, bitter, and umami taste detection. Since inter-

actions between CD36 and Src kinases Fyn, Lyn, and Yes have been reported in other cell types,27 an alternative mechanism is possible. We have recently checked this hypothesis with success in CD36-positive TRC isolated from mouse circumvallates. Indeed, LCFA binding to CD36 leads to the recruitment and activation of the Src-protein tyrosine kinases Fyn and Yes, inducing an increase in [Ca2+ ] i in TRC via the opening of store-operated calcium (SOC) channels.32 The transient receptor potential protein-5 receptor (TRPM5) is known to play an important role in taste transduction. This Na+ /K+ channel responds to rapid changes in [Ca2+ ] i by inducing transient membrane depolarization33 in a fashion that leads to neuromediator release. The much weaker attraction to LCFA observed in TRPM5 knockout mice than in controls supports involvement of this ion channel in the fat-signaling cascade found in TRC.31 Taken together, these ndings provide the rst evidence that LCFA affect the function of mouse TRC via the activation of a signaling cascade that depends on CD36. Further experiments will determine more of the details of the signaling pathway that are involved.

The Gustatory Nerves Convey the Fat Signal

The possible involvement of gustatory nerves in the LCFA-mediated fat preference has recently been explored in rodents, by studying the impact of bilateral transection of the CT (CTX) and/or GL (GLX) nerves. In rats, CTX decreases fat preference in a free-choice situation.34,35 CTX rats also display much weaker conditioned aversion to linoleic acid than control rats.34,35 Paradoxically, the impact of total denervation of the peripheral gustatory nerves (CTX+GLX) was not investigated in these rat studies, although such an exploration would be required for a full demonstration. In mice, the lack of functional peripheral gustatory nerves (CTX+GLX animals) fully abolishes both spontaneous linoleic acid preference and conditioned aversion.18 These ndings demon-

170

Annals of the New York Academy of Sciences

strate that afferent gustatory nerve bers play crucial roles in the orosensory perception of LCFA in rodents (Fig. 2- ). Although CD36 in TRC probably acts as a lipid receptor in these mammals,25 its involvement in the fat signal conveyed by peripheral gustatory nerves remains to be demonstrated. However, much less pancreatic and biliary secretion was observed after oral linoleic acid deposition in CTX/GLX mice than in control mice18 providing indirect support for this hypothesis. Indeed, the cephalic phase of digestion triggered by the presence of LCFA in the oral cavity is highly CD36-dependent.25 Direct demonstration would be supported by electrophysiological recording of activities in CT and GL nerves in CD36+/+ and CD36/ mice subjected to oral fatty acid stimulation.

Lipid-Induced Neuronal Activity in the Nucleus of the Solitary Tract

Lipid signal therefore appears to be transmitted by a peripheral nerve route that was known to be involved in the transfer of gustatory information to the brain. The NST in the brain stem is the rst synaptic relay in this gustatory neuronal cascade. Regulation of Fos, the protein encoded by the immediate early gene c-fos, has been successfully explored by immunolabeling to identify populations of neurons activated by LCFA in the NST.18 In mice, oral linoleic acid deposition triggers activation of NST areas known to receive CT and GL afferents. This activation appears to be CD36-dependent, as it is not observed in CD36 knockout mice subjected to oral stimulation with linoleic acid.18 There is known involvement of the lateral hypothalamus and nucleus accumbens in food intake and reward, respectively. These areas receive synaptic inputs from the NST36 that may account for the spontaneous preference for LCFA-rich food observed in mice. The digestive projections from the NST36 may also contribute to lipid-mediated reexes. These efferents might control pancreato-biliary secretions directly and/or indirectly, through pro-

duction of intestinal hormones. Axons from the mandibular branch of the trigeminal nerve innervate the anterior tongue and project into taste areas of the NST.37 However, roles for mechanical or textural stimulation that might be mediated by these branches appear unlikely. Indeed, water alone or water mixed with xanthan gum to mimic the texture of lipids does not affect neuronal activity in the NST.18 Thus, the fat signal triggered by the interaction of LCFA with CD36 in the oral cavity is transmitted through the NST (Fig. 2- ). Central and peripheral projections from the NST affect the activities of neurons known to be involved in feeding behavior (Fig. 2- ) and the cephalic phase of digestion (Fig. 2- ). Altogether these data support the existence of a gustatory system devoted to LCFA detection in rodents. Humans are also known to be attracted by lipids. For instance, newborns suck more actively when provided lipidenriched milk than when provided low-fat milk.38 Recent studies have shown that gustation can also account for oral fat perception in humans. LCFA seem to be the stimulus involved in this gustatory fat perception. Indeed, healthy human adult subjects can detect saturated and unsaturated LCFA with a very low detection threshold (0.028% w/v) in ways that do not depend on either olfactory or somatosensory cues.39 As reported in rodents, oral lipid stimulation can activate the cephalic phase of human digestion since it is sufcient to trigger a rapid increase in plasma pancreatic peptide concentration in healthy subjects.40 Moreover sham-feeding experiments in healthy human subjects show that oral exposure to fat enhances postprandial plasma TG levels41 in ways that are independent of textural42 and olfactory cues.43 While the molecular mechanisms at the origin of oral lipid perception and physiological changes, including possible roles for CD36, are not yet elucidated, these parallels with rodent behaviors are consistent with involvement of similar mechanisms in man.

Gaillard et al.: Molecular Mechanisms of Fat Preference and Overeating

171

Is CD36 Multifunctionality Compatible with a Role in Gustation? CD36 is a plasma membrane glycoprotein that is expressed in a wide variety of tissues. It is a multifunctional protein that belongs to the family of class-B scavenger receptors. It increases uptake of LCFA by cardiomyocytes and adipocytes44,45 and uptake of oxidized-lowdensity lipoproteins (LDL) by macrophages.46 CD36 modies platelet aggregation by binding to thrombospondin and collagen,47 facilitates phagocytosis of apoptotic cells by macrophages,48 and increases the cytoadhesion of erythrocytes infected with Plasmodium falciparum.49 In addition, CD36 has recently been shown to play roles in the taste reception of dietary lipids on the tongue.18,25 This new function may seem paradoxical, given the high-tissue and binding specicities generally displayed by other taste receptors, such as T1R and T2R.22 Despite its multifunctionality, CD36 generally plays a specic role in a given cell type. For example, it is involved in collagen-mediated cytoadhesion in platelets, while mediating LCFA uptake in myocytes. This cell specicity of function probably results from both cellular context (genotype and microenvironment) and aspects of CD36 itself. Regulation of the gene encoding this receptor is unusually complex50 and includes alternative promoter usage, alternative splicing, and multiple posttranslational modications.26 Indeed, the human, rat, and mouse CD36 genes contain several alternative and independent promoters and corresponding alternative rst exons.26,51,52 The expression proles of these alternative transcripts differ considerably between tissues,50 resulting in the tissue-specic regulation of CD36 gene expression. Differences in the CD36 promoters expressed in the liver and small intestine account for the differential regulation of the CD36 gene by perixisome proleratoractivated receptors (PPAR) agonists observed in these tissues.51 Diverse posttranslational CD36 modications by glycosylation, phos-

phorylation, acylation, and/or palmitoylation could also contribute to the cell-type specicity of its functions. The degree of glycosylation of the CD36 extracellular domain varies considerably between tissues, for example.53 This phenomenon may result in specic physiological effects due to the selection of a specic ligand in a specic cell. Thus, a role for CD36 in the chemoreception of dietary lipids by TRC remains plausible despite the multifunctional nature of this protein. The recent identication of a CD36-related receptor that is responsible for the olfactory detection of a fatty acid-derived pheromone in Drosophila54 suggests that the chemosensory function of CD36 previously reported in mice may be identied in many branches of the animal kingdom. Are Fatty Foods Addictive? Choosing to eat one specic food rather than another depends on factors that include the hedonic and metabolic satisfaction experienced during and after its intake. Preference for fatty foods must therefore result from the integration of oral and postoral signals. Such integration, for example, can explain why mice rapidly stop eating sorbitol LCFA ester, though they spontaneously prefer this in a free-choice situation, since this lipid is indigestible and not nutritious.55 By contrast, free access to an attractive nutrient that also results in a sense of fullness on digestion may lead to overeating, thereby increasing the risk of obesity. Thus, mice overeat corn oil when exposed to it for long periods of time, and they become obese if this oil is provided as an optional supplement to standard laboratory chow.5 Interestingly, strong preference for corn oil is maintained throughout the 42 days of experimentation, suggesting a possible addiction-like behavior of these rodents to lipids. The voluntary intake of corn oil by mice in tests that are classically used to evaluate drug reward, such as conditioning place preference (CPP), is consistent with this assumption.56

172

Annals of the New York Academy of Sciences

In contrast, no reward effects are observed if corn oil is delivered directly to the stomach 60 min before conditioning.56 Thus, oral signals seem to play a signicant role in this feeding behavior. The molecular mechanisms underlying the strong attraction to fat in rodents are gradually becoming clearer. They involve both the opioidergic system in the arcuate and paraventricular nuclei of the hypothalamus and the dopaminergic system in the accumbens nucleus. In rats conditioned to ingest a corn oil emulsion at the same time each day, proopiomelanocortin (POMC) mRNA levels in the arcuate nucleus increase in anticipation, even 30 min before the programmed period of fat intake.57 This elevation of POMC expression leads to a transient increase in the POMC derivative, the opioid peptide -endorphin in blood and cerebrospinal uid.57 Fat intake in rats is associated with upregulation of dynorphin and enkephalin, two other opioids known to promote lipid intake, in the paraventricular nucleus.58 There may therefore be a positive regulatory loop accounting for the attractive effects of fat, the induction of fat intake by opioids, and the synthesis of these regulatory neuropeptides. Consistent with this assumption, the central injection of a -opioidergic receptor agonist, [D-Ala(2),N-Me-Phe(4),Gly(5)ol]-enkephalin (DAMGO), promotes lipid-rich food intake in the rat.59 Conversely, treatment with a specic -receptor antagonist abolishes not only preference for fat60,61 but also the ability of corn oil to establish a conditioned place preference.60 Thus, brain systems that utilize the OPRM1 -opiate receptor play signicant roles in preference for fat. The dopaminergic system also appears to be implicated in this complex behavior in rodents. Pharmacological inhibition of the D1 and D2 dopamine receptors decreases the preference for a corn oil emulsion in a dose-dependent manner.62,63 In rat, sham feeding with corn oil increases dopamine levels in the nucleus accumbens, a key component of pleasure and reward circuits.64 In this experiment, postoral

feedback was excluded. The results thus highlight the major role played by oral lipid detection in activating reward pathways in ways that drive attraction to fatty foods. Dopamine release by the nucleus accumbens is also regulated by the opioid peptides dynorphin and enkephalin.65 Thus, attraction to lipids in rodents results from connections between several regulatory molecules and areas of the central nervous system that are with specialized involvement in reward. It remains unclear whether lingual CD36 plays a role in activating this regulatory pathway, however. Conclusions and Future Directions Overconsumption of fatty foods participates in the obesity risk in various mammalian species. The association of a high palatability with a high metabolic fullness likely accounts for this tremendous preference for lipids. Recent data show that palatability of fat results not only from olfaction and somatosensation (mainly touch and texture), as it was long thought, but also from gustation. Knocking out the gene that encodes the gustatory lipid-receptor CD36 leads to a dramatic decreases in the spontaneous fat preference in the mouse. Fatty foods also provide metabolic fullness on the basis of the great efciency of absorption, metabolism, and storage of lipids and the relative weakness of the postprandial satiating effects that lipids induce. Because their palatability is only weakly reduced by postoral feedback, high-fat diets promote greater daily intake. This positive regulatory loop that enhances consumption of lipid rich foods would clearly constitute an advantage in times of food scarcity. Indeed, fat-rich foods are an important source of energy, contain essential fatty acids, and carry lipid-soluble vitamins (A, D, E, K) with many fundamental biological functions. Conversely, this system, likely inherited from evolution, would increase the prevalence of obesity during periods of food abundance.

Gaillard et al.: Molecular Mechanisms of Fat Preference and Overeating

173

Further data concerning the molecular mechanisms involved in fat preference are needed to fully understand this complex behavior. It might provide new insights for limiting preference for fatty foods, useful to develop new antiobesity strategies.
Conicts of Interest

The authors declare no conicts of interest. References

1. Malecka-Tendera, E. & A. Mazur. 2006. Childhood obesity: a pandemic of the twenty-rst century. Int. J. Obes. (Lond.) 30(Suppl 2): S1S3. 2. Olshansky, S.J. et al. 2005. A potential decline in life expectancy in the United States in the 21st century. N. Engl. J. Med. 352: 11381145. 3. Tsuruta, M. et al. 1999. The orosensory recognition of long-chain fatty acids in rats. Physiol. Behav. 66: 285288. 4. Takeda, M., M. Imaizumi & T. Fushiki. 2000. Preference for vegetable oils in the two-bottle choice test in mice. Life Sci. 67: 197204. 5. Takeda, M. et al. 2001. Long-term optional ingestion of corn oil induces excessive caloric intake and obesity in mice. Nutrition 17: 117120. 6. Blundell, J.E. & J.I. Macdiarmid. 1997. Passive overconsumption. Fat intake and short-term energy balance. Ann. N. Y. Acad. Sci. 827: 392407. 7. Raybould, H.E. 2007. Mechanisms of CCK signaling from gut to brain. Curr. Opin. Pharmacol. 7: 570 574. 8. Malendowicz, L.K. et al. 2003. Effects of prolonged cholecystokinin administration on rat pituitaryadrenocortical axis: role of the CCK receptor subtypes 1 and 2. Int. J. Mol. Med. 12: 903909. 9. Tso, P. & M. Liu. 2004. Ingested fat and satiety. Physiol. Behav. 81: 275287. 10. Doi, T. et al. 2001. Effect of leptin on intestinal apolipoprotein AIV in response to lipid feeding. Am. J. Physiol. Regul. Integr. Comp. Physiol. 281: R753R759. 11. Migrenne, S., C. Magnan & C. CrucianiGuglielmacci. 2007. Fatty acid sensing and nervous control of energy homeostasis. Diabetes Metab. 33: 177182. 12. Wang, R. et al. 2006. Effects of oleic acid on distinct populations of neurons in the hypothalamic arcuate nucleus are dependent on extracellular glucose levels. J. Neurophysiol. 95: 14911498.

13. Takeda, M. et al. 2001. Preference for corn oil in olfactory-blocked mice in the conditioned place preference test and the two-bottle choice test. Life Sci. 69: 847854. 14. Fukuwatari, T. et al. 2003. Role of gustation in the recognition of oleate and triolein in anosmic rats. Physiol. Behav. 78: 579583. 15. Smith, J.C. et al. 2000. Orosensory factors in the ingestion of corn oil/sucrose mixtures by the rat. Physiol. Behav. 69: 135146. 16. McCormack, D.N., V.L. Clyburn & D.W. Pittman. 2006. Detection of free fatty acids following a conditioned taste aversion in rats. Physiol. Behav. 87: 582 594. 17. Oakley, B. 1993. The gustatory competence of the lingual epithelium requires neonatal innervation. Brain Res Dev. Brain Res. 72: 259264. 18. Gaillard, D. et al. 2008. The gustatory pathway is involved in CD36-mediated orosensory perception of long-chain fatty acids in the mouse. FASEB J. 22: 14581468. 19. Yoneda, T. et al. 2007. The palatability of corn oil and linoleic acid to mice as measured by short-term two-bottle choice and licking tests. Physiol. Behav. 91: 304309. 20. Kawai, T. & T. Fushiki. 2003. Importance of lipolysis in oral cavity for orosensory detection of fat. Am. J. Physiol. Regul. Integr. Comp. Physiol. 285: R447R454. 21. Gilbertson, T.A. & J.D. Boughter, Jr. 2003. Taste transduction: appetizing times in gustation. Neuroreport 14: 905911. 22. Sugita, M. 2006. Taste perception and coding in the periphery. Cell. Mol. Life Sci. 63: 20002015. 23. Baillie, A.G., C.T. Coburn & N.A. Abumrad. 1996. Reversible binding of long-chain fatty acids to puried FAT, the adipose CD36 homolog. J. Membr. Biol. 153: 7581. 24. Fukuwatari, T. et al. 1997. Expression of the putative membrane fatty acid transporter (FAT) in taste buds of the circumvallate papillae in rats. FEBS Lett. 414: 461464. 25. Laugerette, F. et al. 2005. CD36 involvement in orosensory detection of dietary lipids, spontaneous fat preference, and digestive secretions. J. Clin. Invest. 115: 31773184. 26. Rac, M.E., K. Safranow & W. Poncyljusz. 2007. Molecular basis of human CD36 gene mutations. Mol. Med. 13: 288296. 27. Huang, M.M. et al. 1991. Membrane glycoprotein IV (CD36) is physically associated with the Fyn, Lyn, and Yes protein-tyrosine kinases in human platelets. Proc. Natl. Acad. Sci. USA 88: 78447848. 28. Sclafani, A., K. Ackroff & N.A. Abumrad. 2007. CD36 gene deletion reduces fat preference and intake but not post-oral fat conditioning in mice. Am.

174 J. Physiol. Regul. Integr. Comp. Physiol. 293: R1823 R1832. Harmon, C.M. et al. 1991. Labeling of adipocyte membranes by sulfo-N-succinimidyl derivatives of long-chain fatty acids: inhibition of fatty acid transport. J. Membr. Biol. 121: 261268. Wong, G.T., K.S. Gannon & R.F. Margolskee. 1996. Transduction of bitter and sweet taste by gustducin. Nature 381: 796800. Sclafani, A. et al. 2007. Fat and carbohydrate preferences in mice: the contribution of alpha-gustducin and Trpm5 taste-signaling proteins. Am. J. Physiol. Regul. Integr. Comp. Physiol. 293: R1504R1513. El-Yassimi, A. et al. 2008. Linoleic acid induces calcium signaling, SRC-kinase phosphorylation and neurotransmitters release in mouse CD36-positive gustatory cells. J. Biol. Chem. 283: 1294912959. Prawitt, D. et al. 2003. TRPM5 is a transient Ca2+activated cation channel responding to rapid changes in [Ca2+]i. Proc. Natl. Acad. Sci. USA 100: 15166 15171. Stratford, J.M., K.S. Curtis & R.J. Contreras. 2006. Chorda tympani nerve transection alters linoleic acid taste discrimination by male and female rats. Physiol. Behav. 89: 311319. Pittman, D. et al. 2007. Chorda tympani nerve transection impairs the gustatory detection of free fatty acids in male and female rats. Brain Res. 1151: 7483. Berthoud, H.R. 2002. Multiple neural systems controlling food intake and body weight. Neurosci. Biobehav. Rev. 26: 393428. Hamilton, R.B. & R. Norgren. 1984. Central projections of gustatory nerves in the rat. J. Comp. Neurol. 222: 560577. Nysenbaum, A.N. & J.L. Smart. 1982. Sucking behaviour and milk intake of neonates in relation to milk fat content. Early Hum. Dev. 6: 205213. Chale-Rush, A., J.R. Burgess & R.D. Mattes. 2007. Evidence for human orosensory (taste?) sensitivity to free fatty acids. Chem. Senses. 32: 423431. Crystal, S.R. & K.L. Teff. 2006. Tasting fat: cephalic phase hormonal responses and food intake in restrained and unrestrained eaters. Physiol. Behav. 89: 213220. Mattes, R.D. 1996. Oral fat exposure alters postprandial lipid metabolism in humans. Am. J. Clin. Nutr. 63: 911917. Mattes, R.D. 2001. Oral exposure to butter, but not fat replacers elevates postprandial triacylglycerol concentration in humans. J. Nutr. 131: 14911496. Mattes, R.D. 2001. The taste of fat elevates postprandial triacylglycerol. Physiol. Behav. 74: 343 348. Coburn, C.T. et al. 2000. Defective uptake and utilization of long chain fatty acids in muscle and adipose

Annals of the New York Academy of Sciences

tissues of CD36 knockout mice. J. Biol. Chem. 275: 3252332529. Hajri, T. et al. 2001. Defective fatty acid uptake in the spontaneously hypertensive rat is a primary determinant of altered glucose metabolism, hyperinsulinemia, and myocardial hypertrophy. J. Biol. Chem. 276: 2366123666. Febbraio, M., D.P. Hajjar & R.L. Silverstein. 2001. CD36: a class B scavenger receptor involved in angiogenesis, atherosclerosis, inammation, and lipid metabolism. J. Clin. Invest. 108: 785791. Chen, H., M.E. Herndon & J. Lawler. 2000. The cell biology of thrombospondin-1. Matrix Biol. 19: 597614. Ren, Y. et al. 1995. CD36 gene transfer confers capacity for phagocytosis of cells undergoing apoptosis. J. Exp. Med. 181: 18571862. Oquendo, P. et al. 1989. CD36 directly mediates cytoadherence of Plasmodium falciparum parasitized erythrocytes. Cell 58: 95101. Andersen, M. et al. 2006. Alternative promoter usage of the membrane glycoprotein CD36. BMC Mol. Biol. 7: 8. Sato, O., N. Takanashi & K. Motojima. 2007. Third promoter and differential regulation of mouse and human fatty acid translocase/CD36 genes. Mol. Cell. Biochem. 299: 3743. Cheung, L. et al. 2007. Hormonal and nutritional regulation of alternative CD36 transcripts in rat livera role for growth hormone in alternative exon usage. BMC Mol. Biol. 8: 60. Greenwalt, D.E. et al. 1992. Membrane glycoprotein CD36: a review of its roles in adherence, signal transduction, and transfusion medicine. Blood 80: 1105 1115. Benton, R., K.S. Vannice & L.B. Vosshall. 2007. An essential role for a CD36-related receptor in pheromone detection in Drosophila. Nature 450: 289293. Suzuki, A. et al. 2003. Integration of orosensory and postingestive stimuli for the control of excessive fat intake in mice. Nutrition 19: 3640. Imaizumi, M., M. Takeda & T. Fushiki. 2000. Effects of oil intake in the conditioned place preference test in mice. Brain Res. 870: 150156. Mizushige, T. et al. 2006. POORC and orexin mRNA expressions induced by anticipation of a corn-oil emulsion feeding are maintained at the high levels until oil ingestion. Biomed. Res. (Tokyo) 27: 227232. Chang, G.Q. et al. 2007. Dietary fat stimulates endogenous enkephalin and dynorphin in the paraventricular nucleus: role of circulating triglycerides. Am. J. Physiol. Endocrinol. Metab. 292: E561E570. Will, M.J., E.B. Franzblau & A.E. Kelley. 2003. Nucleus accumbens mu-opioids regulate intake of a

29.

45.

30.

46.

31.

47.

32.

48.

33.

49.

50.

34.

51.

35.

52.

36.

37.

53.

38.

54.

39.

55.

40.

56.

41.

57.

42.

58.

43.

44.

59.

Gaillard et al.: Molecular Mechanisms of Fat Preference and Overeating

high-fat diet via activation of a distributed brain network. J. Neurosci. 23: 28822888. 60. Imaizumi, M. et al. 2001. Opioidergic contribution to conditioned place preference induced by corn oil in mice. Behav. Brain Res. 121: 129136. 61. Mizushige, T. et al. 2006. Daily increase of fat ingestion mediated via mu-opioid receptor signaling pathway. Biomed. Res. 27: 259263. 62. Weatherford, S.C., G.P. Smith & L.D. Melville. 1988. D-1 and D-2 receptor antagonists decrease corn oil sham feeding in rats. Physiol. Behav. 44: 569 572.

175

63. Weatherford, S.C. et al. 1990. The potency of D-1 and D-2 receptor antagonists is inversely related to the reward value of sham-fed corn oil and sucrose in rats. Pharmacol. Biochem. Behav. 37: 317323. 64. Liang, N.C., A. Hajnal & R. Norgren. 2006. Sham feeding corn oil increases accumbens dopamine in the rat. Am. J. Physiol. Regul. Integr. Comp. Physiol. 291: R1236R1239. 65. Spanagel, R., A. Herz & T.S. Shippenberg. 1990. The effects of opioid peptides on dopamine release in the nucleus accumbens: an in vivo microdialysis study. J. Neurochem. 55: 17341740.