From Food Preference to Craving Behavioural Traits and Molecular Mechanisms Johan Alsiö

! " #$% # & $$'& # ()*)+ )(() #$$# ! "##"##$%##& ' & & (' ')* & +,-' . / ',

!01,#"#,* * (& ,' - ' ,! , 23,$ , ,4 5$678$"822866 8$,

(& & 8 & 9& & .' ' ' ,: .' && ',4 & '; ' & 8 & & . & ,< & 9 ' ' . ' '' & ,,' '.,4 ''. ' '' & & ', -'9 & .' ' ,=' . '989 ' &. '.' & & ''8&)*+ .' ''8 ) + , ' . .'' & & '* , = '' '' & * )* +&& 89' , *' . ' . '.' .& * & && 8 8 ):(+ 8 ):<+ && ,:( ' ) +& & ' &8 ' > .' ., ' && . :( :< ' * .' .,-' ':( ' , 4 ' & & > , = ; & & & & ' ' & & & & ''8&''8 & .' ' , 4 .' ' & & ' ' & & & .'.' .& & ,-'& & ,

:<. * & & ' ! : &8

!" # "$ %&'(" ")*+&,-. "

?1 ' !0#"#

4 5"32"83#3 4 5$678$"822866 8$ % %%% 8""$66$)' %@@ ,9,@ A B % %%% 8""$66$+ From Food Preference to Craving Behavioural Traits and Molecular Mechanisms Johan Alsiö

To tutors and tutees today and in the future

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Alsiö, J., Roman, E., Olszewski, PK., Jonsson, P., Fredriksson, F., Levine, AS., Meyerson, BJ., Hulting, A-L., Lindblom, J., Schiöth, HB. (2009) Inverse association of high-fat diet preference and anxiety-like behavior: a putative role for urocortin 2. Genes, Brain and Behavior 8:193–202

II Alsiö, J., Pickering, C., Stephansson, O., Roman, E., Hulting, A-L., Lindblom, J., Schiöth, HB. (2010) Locomotor adaptation and elevated expression of reward-relevant genes following free-choice high-fat diet exposure. Manuscript

III Alsiö, J., Pickering, P., Roman, E., Hulting, A-L., Lindblom, J., Schiöth, HB. (2009) Motivation for sucrose in sated rats is predicted by low anxiety-like behavior. Neuroscience Letters 454:193–197

IV Pickering, C.*, Alsiö, J.*, Hulting, A-L., Schiöth, HB. (2009) Withdrawal from free-choice high fat high sugar diet induces craving only in obesity-prone animals. Psychopharmacology 204:431–443

V Alsiö, J., Olszewski, PK., Hallsten Norbäck, A., Gunnarsson, Z., Levine, AS., Pickering, C., Schiöth, HB. (2010) Downregu- lation of nucleus accumbens D1 and D2 receptor expression occurs upon exposure to and persists long-term after withdrawal from palatable food: conclusions from diet-induced obesity models. Submitted manuscript

VI Alsiö, J., Stenhammar, C., Benedict, C., Hulting, A-L., Mont- gomery, SM., Edlund, B., Schiöth, HB. (2010) Parental food preferences are associated with body weight disturbance in preschool children. Manuscript

* Equal contribution

Reprints were made with permission from the respective publishers.

Contents

Introduction...... 13 Addictive-like behaviours in obesity...... 13 Feeding for calories...... 15 Hypothalamic appetite control...... 15 Monogenic obesity: severe but not common ...... 16 The thrifty genotype hypothesis ...... 16 Food preferences...... 17 Feeding for pleasure...... 18 The role of dopamine in food-motivated behaviours...... 18 Opioids and food intake...... 22 Diet-induced obesity in the rat ...... 25 Food craving...... 27 The phenomenology of food craving...... 27 Animal models of craving ...... 29 Personality types in obesity...... 31 Emotional eating ...... 32 The growing problem...... 32 Aims...... 34 Material and methods...... 36 Diet-induced obesity and food choices ...... 36 Gene expression measurements...... 37 Exploration of novel environments: rodent models ...... 41 Operant self-administration...... 43 Questionnaires...... 44 Results and discussion ...... 46 General discussion...... 59 Conclusions...... 66 Acknowledgements...... 68 References...... 71

Abbreviations

AAAD Aromatic L-amino acid decarboxylase ACTH Adrenocorticotropic hormone AgRP Agouti-related peptide ARC arcuate nucleus bACT -actin BMI Body mass index bTUB -tubulin CART Cocaine- and amphetamine-regulated transcript

CRF2 Corticotropin-releasing factor receptor subtype 2 COMT Catechol-o-methyl transferase CORT Corticosterone CPu Caudate putamen CRF Corticotropin-releasing factor CYCLO Cyclophilin

D1 Dopamine D1 receptor

D2 Dopamine D2 receptor DAT Dopamine transporter DCR Dark corner room of the MCSF DEBQ Dutch Eating Behavior Questionnaire DIO Diet induced obesity DIST Distance moved in the open field DOR Delta opioid receptor DUR Duration EPM Elevated plus-maze FCI Food Craving Inventory fMRI Functional magnetic resonance imaging FRn Fixed ratio n, e.g. FR1 for fixed ratio 1 FRQ Frequency GALP Galanin-like peptide GAPDH Glyceraldehyde 3-phosphate dehydrogenase GHRF Growth hormone-releasing factor

GRF Gonadotropin-releasing factor H3b Histone protein 3b HCD High-carbohydrate diet HF High-fat HFD High-fat diet HFHS High-fat high-sugar HPD High-protein diet HS High-sugar i.c.v. intracerebroventricular KOR Kappa opioid receptor LH Lateral hypothalamaus MCH Melanin-concentrating hormone MCSF Multivariate Concentric Square Field™ MOR Mu opioid receptor NAcc Nucleus accumbens NPB Neuropeptide B NPFF Neuropeptide FF NPY Neuropeptide Y OP Obesity-prone OR Obesity-resistant PET Positron emission tomography PFC Prefrontal cortex POMC Proopiomelanocortin PR Progressive ratio PrRP Prolactin-releasing peptide PVN Paraventricular nucleus qPCR Quantitative PCR RPL19 Ribosomal protein L19 RT-PCR Reverse transcriptase PCR SAP Stretched attend posture SDCA succinate dehydrogenase complex, subunit A SPSQ Swedish Parental Stress Questionnaire TH Tyrosine hydroxylase ToRC Time of response cessation TRH Thyrotropin-releasing hormone Ucn 2 Urocortin 2 WD Withdrawal VTA Ventral tegmental area

Introduction

We overeat not because of hunger but because of family and friends, packages and plates, names and numbers, labels and lights, colors and candles, shapes and smells, distractions and distances, cupboards and containers. Brian Wansink, Mindless eating

Addictive-like behaviours in obesity For the first time in history, there are presently more people in the world suffering from obesity than from food shortage or starvation. The modern society is extremely obesogenic, with a sedentary lifestyle and easy access to refined food products, high in fat and sugar, high in calories, and highly palatable. However, not all individuals exposed to this environment become obese. Vulnerability to obesity may be caused by factors that are genetic or hormonal, affected by psychological and stressful conditions or pharmacological agents, and strengthened by cultural and familial settings. The vulnerable individual in the modern obesogenic environment gains weight and becomes obese, others remain lean. Once an obesity-prone individual has gained weight, it is exceedingly difficult to lose this weight. A plethora of weight loss programs exists, but even though initial weight loss is often achieved if the subjects adhere to their regimen, the excessive body weight returns as soon as the program is dropped. During abstinence from energy-dense food, craving for food, especially for palatable food, is commonly reported, causing relapse to consumption of energy-dense foods. The subject loses control over food intake. In recent years, studies have shown that there are important parallels between obesity and drug addiction [295,302]. Drug addiction is a chronic relapsing disorder characterized by individual vulnerability and a progres- sion from recreational use, through repeated use, to compulsive use [273,286]. Ultimately, if a drug addict attempts to quit the drug, drug craving ensues that may remain for extended periods of time and, long after the physical signs of withdrawal have subsided, may lead to relapse. The subject loses control over drug intake.

13 One neural pathway common to both drug addiction and obesity is the mesolimbic dopamine system (the ‘reward system’). All drugs with abuse liability release dopamine in the nucleus accumbens (NAcc; [79]), just as natural stimuli such as food do [27]. Genetic variants related to dopamine signaling have been associated with both drug addiction and obesity [180,258]. Dopamine D2 receptor availability is decreased in both drug addicts and the obese [295]. Notable interactions are also present between appetite and drug addiction. First, cocaine and amphetamine are strong anorexigenic agents that are indeed abused by some individuals in order to control appetite and body weight [56]. Second, animal studies show that food restriction expedites acquisition of operant response tasks for psychostimulants [76], increases behavioural sensitization to cocaine [30], stimulates intake of drugs of abuse [48], facilitates conditioned place preference for cocaine [30], and precipitates relapse of drug-seeking in extinction-reinstatement models [241]; see also [46]. In fact, overconsumption of calories has the opposite effect; rats maintained long-term on a high-fat diet, thus becoming obese, had impaired acquisition of an operant response task reinforced by cocaine compared to chow-fed controls [288]. Furthermore, high sucrose preference is associated with elevated operant responding for cocaine [105] and amphetamine [78], while also predicting alcohol responsiveness [303]. Table 1 Comparison of overeating in the obese with drug abuse in addicts Overeating in obese Drug abuse in addicts Homeostatic regulation +++ - Regulation of intake Central and peripheral Central Hedonic impact Sensory and post-oral Pharmacological Reinforcement ++ +++ Craving ++ +++ Personality ++ ++ Stress Promotes appetite Promotes drug use and central obesity and relapse Emotionality +++ ++ Cue-induced relapse ++ +++ Drug addiction and overeating in a wide sense of the terms; no account taken to different drugs or food types. Modified from [302].

Thus, important parallels exist between obesity and drug addiction (Table 1). The differences should also be noted, however. Unlike drugs, food is required for survival and food intake is, hence, meticulously controlled by the organism through the homeostatic food intake regulation. Obesity is not equivalent to addiction, but there are important similarities that can help us understand what has gone awry in the brain of the obese individual.

14 Animal models of obesity such as diet-induced obesity (DIO) in the rodent are valuable methods for understanding both the vulnerability to obesity and the long-term changes in food intake regulation. In this thesis, we used DIO in combination with methods and concepts previously applied in drug addiction research to shed further light on behavioural traits and molecular mechanisms related to both obesity proneness and to craving for food during abstinence from energy-dense foods.

Feeding for calories Unlike drugs, food is essential for the survival of the organism. Thus, evolution has provided us with an intricate system for regulating food intake and energy expenditure. The mechanisms that drive food seeking behaviour and consumption of food during times of negative energy status seem especially powerful. The presence of peripheral signals conveying information about the energy status of the body to the brain was predicted long before the characterization of such signals. The ob/ob and db/db mouse lines were the first clues to this puzzle (for review, see [38]). Both these lines have identical phenotypes: they become extremely obese and develop diabetes as the body weight rises. It was shown that the ob/ob mouse lacks a circulating factor present in the db/db. This factor, later identified as leptin, serves as a hormone produced by the adipose tissue in relation to the amount of fat in the body. Ob/ob mice have a disruption in the gene coding for leptin, while db/db have non-functional leptin receptors. Leptin is released in response to food intake and levels rapidly diminish during food restriction or acute deprivation. Leptin receptors are present throughout the body, but in the arcuate nucleus (ARC) of the hypothalamus, the availability of these receptors is particularly high.

Hypothalamic appetite control The ARC has receptors not only for leptin but also for circulating factors such as insulin, ghrelin, and adiponectin; mechanisms are also present for detecting nutrients such as glucose, fatty acids and amino acids [67]. Indeed, the hypothalamus integrates peripheral signals related to the energy status of the body, and serves as a focal point of appetite regulation in the central nervous system. The hormones and neuropeptides implicated in the homeostatic food intake regulation have been the subject of many reviews [11,39,127,187,234,304], the core message of which will be briefly presented here. Within the ARC, two main populations of neurons exist, co-expressing either 1) neuropeptide Y/agouti-related peptide (NPY/AgRP) or 2) proopiome-

15 lanocortin/cocaine- and amphetamine-related transcript (POMC/CART). The NPY/AgRP neurons are activated by ghrelin and inhibited by leptin and insulin; they have orexigenic effects. POMC/CART neurons mediates anorexigenic effects stimulated by leptin and insulin and inhibited by ghrelin. Targets of these neurons are present mainly in the lateral hypothalamus and in the paraventricular nucleus of the hypothalamus. From these regions, fibers project to diverse parts of the brain with signal molecules such as opioids, orexin, melanin-concentrating hormone, oxytocin, vasopressin, galanin, corticotropin-releasing factor, urocortins, and many more. Afferent fibers to the hypothalamus originate e.g. in the brain stem, the amygdala, and the reward pathways including the NAcc and the ventral tegmental area (VTA).

Monogenic obesity: severe but not common Severe obesity phenotypes have been reported in individuals with disrup- tions in single genes related to appetite control. The first such report indicated that loss of leptin secretion due a mutation in the leptin gene causes early-onset obesity [171]. A loss-of-function mutation in the leptin receptor gene was associated with extreme early-onset obesity and prohibited puber- tal development [53]. Similar phenotypes have been reported in patients with mutations in the POMC gene [135,136] and in the gene encoding melano- cortin receptor 4 [305], a receptor that binds the POMC product MSH. While these genetic variants are rare, they highlight the importance of key elements in energy homeostasis in humans. More recently, a number of genes with common variants that are associated with body weight, e.g. TMEM18 and FTO, have been identified through genome-wide association studies [292]. The role of those genes and their products in energy balance regulation is gradually being unveiled [188]. While the ob/ob and db/db mouse lines have severe phenotypes, mutant mice lacking functional variants of the aforementioned peptides do not always have the expected phenotype. NPY null mice have normal weight gain and food intake [84], while AgRP null mice have unaltered body weight curves and normal food intake patterns during free feeding, food deprivation, diet-induced obesity, and leptin injections [211].

The thrifty genotype hypothesis If monogenic obesity is uncommon, the question arises as to what kind of genotype is more frequent in obesity. The genetic pool changes slowly and novel and widespread mutations cannot possibly explain the rapid emer- gence of the obesity epidemic. Hence, the genotypes underlying vulnerability to obesity predates the modern environment, with which it interacts to produce excess body weight gain in the obesity-prone individual.

16 The concept of a ‘thrifty genotype’ was proposed by Neel to explain the rising incidence of diabetes [176]; the idiom was chosen to describe a genotype apt to conserve energy, ‘exceptionally efficient in the intake and/or utilization of food’ [176]. It was argued that the same thrifty genetic variants that were advantageous in early human history, when starvation was common and energy-dense food scarce, puts the individual at risk for diabetes and obesity in the surplus of modern environment. It has also been speculated that such genetic variants are accumulated in certain ethnic groups that appear to be at increased risk for diabetes, such as the Pima Indians found in Mexico and Southern USA, a population that has been the subject of a vast number of investigations into the genetic basis for diabetes [18]. The thrifty genotype hypothesis has been questioned [251,254], and does not fully explain why some individuals are prone to obesity and others resistant. Nevertheless, vulnerability to obesity is an individual trait and may involve a genetic background associated with energy intake, metabolism, or energy expenditure.

Food preferences The availability of palatable, high-energy food has been proposed to be the main environmental factor driving the prevalence of obesity to epidemic proportions. Preference for high-fat foods has been associated with body weight and body composition in both children [91,217] and adults [167,218] in most but not all studies [97]. Note that the quality of dietary fat may also be important, since it has been shown that the association of dietary fat intake and body weight is particularly strong for saturated fats [111]. Con- sumption of high-sugar foods has been implicated in the development of obesity [282] and may be affected by preference for sweet tastants. It has recently been shown that sweet taste preference has a genetic component related to the reported consumption of high-sugar tastants [128] (see also [214]). In addition, food preferences can be acquired and influenced by culture and social context [218]. In a study by Rissanen et al., monozygotic twins that differed in body weight at adult age were studied with regard to food preferences [218]. Current preference for high-fat foods were three-fold more frequent in the obese subjects than in their lean twin, and both the obese subjects and their siblings reported that the obese twin in each pair had greater preference for high-fat foods in young adulthood, while the lean twin had low preference for such food [218]. Animal experiments on food preferences have identified a number of neuropeptides and other factors that may affect the consumption of individual macronutrients or tastants. For example, neuropeptide Y (NPY) preferentially stimulates carbohydrate intake, while central ghrelin has been suggested to enhance fat ingestion [242]. The impact of other hormones, including leptin [290] and corticosterone [50,209], on food preferences has also

17 been reported. Oxytocin knockout mice overeat sucrose but not lipid emulsions [169,237]. In contrast, the orexigenic effects of galanin are stronger when high-fat diets are offered [184,264]. In addition, opioid peptides were originally suggested to selectively promote intake of dietary fat [160,161,163], but this notion has been contested (see below). In an outbred group of rats, food preferences vary between individuals. For example, when rats are allowed to choose freely between palatable high- fat and high-sucrose diets, some animals eat only 30% whereas others 70% of a given kind of diet [189]. Importantly, such preferences affect food intake to such an extent that some subjects are prone to gain weight in the free- choice situation, whereas others are resistant to such environmental modification [64,278]. Thus, food choices in an outbred group of rats may be investigated in order to gain further insight into the molecular mechanisms behind food and macronutrient preferences.

Feeding for pleasure When the energy demands of the organism are met, palatable food will still stimulate feeding behaviour. Indeed, humans and other mammals will overconsume food if palatable and energy-dense diets are provided ad libitum. Such behaviours are not easily explained by leptin signals or neuropeptides within the hypothalamus. Imaging studies with humans during consumption of palatable food and during expectation of palatable food show increased activity of specific brain regions including the caudal and medial orbitofrontal cortex, amygdala, striatum and midbrain [183,198,248]. Moreover, such brain activity has been shown to differ between obese subjects and lean controls [126,226,259,260]. For example, blunted activation was noted in the caudate nucleus of obese individuals during consumption of a palatable milkshake [259]. While other brain regions have also been implicated in consummatory responses to palatable food, the list above is related to the dopamine pathways originating in the midbrain (VTA, substantia nigra) and terminating in the striatum (NAcc, caudate nucleus, putamen) and prefrontal cortex. It should be noted that the mesolimbic dopamine system, in particular, is known from animal studies to be the pathway activated in reward and reinforcement.

The role of dopamine in food-motivated behaviours Food intake, just as any other natural reward, produces a release of dopamine in the NAcc [27,29]. Dopaminergic neurons sending fibers to the NAcc are present in the VTA, and these projections are denoted the mesolimbic dopamine pathway, or the ‘reward system’, although additional brain structures (e.g. prefrontal cortex, PFC) receiving dopaminergic input are usually

18 included in this construct (Figure 1). Reinforcement, that is the process by which certain behaviours are repeated in order for the organism to either avoid harmful stimuli (negative reinforcement) or seek out pleasurable stimuli (positive reinforcement), is contingent upon the release of dopamine in the NAcc. Indeed, this system is arguably responsible for motivating behaviours relevant to the survival and propagation of the self and the species, such as eating and mating.

(B) (C) A

B C

Figure 1 A) Sagittal schematic view (modified from [193]) of the rodent brain, indicating the regions investigated in this thesis with mesocorticolimbic and nigrostri- atal dopamine pathways labelled in grey; B) D2 receptor gene expression in the striatum [1]; C) Dopamine transporter gene expression [1]. PFC, prefrontal cortex; CPu, caudate putamen; NAcc, nucleus accumbens; HC, hippocampus; HT, hypothalamus; Amy, amygdala; SN, substantia nigra; VTA, ventral tegmental area.

Based on the finding that dopamine antagonists reduce operant responding for food, Wise originally proposed that dopamine encoded the hedonic properties of natural rewards [294], i.e. the pleasure experienced by attaining the reward. Other hypotheses have been posited since, however, that are supported by stronger evidence. Rather than encoding the hedonics of rewards, dopamine seems to relate to the incentive properties thereof. According to Berridge and Robinson, dopamine mediates the incentive salience of rewards and is thus essential for motivation to obtain the reward, but not for the pleasure of actually obtaining it [33]. The basis for this notion is e.g. that depletion of dopamine or blockade of dopamine receptors do not diminish pleasurable responses to palatable foods in animals or humans [33].

19 In agreement with this, preferential release of dopamine occurs when food consumed is novel. For example, palatable food increased dopamine release in the NAcc of rats during the first exposure but not upon subsequent presentations [27,29]. Hollerman and Schultz studied monkeys implanted with electrodes in the NAcc in a learning task [120]. Initially, when the monkey made many errors and the reward (palatable apple juice) was unpredictable, dopamine neuron responses were strong. But as the subjects acquired the task, dopamine firing was reduced and reached baseline as the reward became fully predictable. In addition, changes in dopamine neuron activity was observed in these monkeys when reward was delivered at an unexpected time point, and also when the reward was expected but not delivered [120]. The latter protocol is referred to as ‘error in reward prediction’. Interestingly, Schultz et al. have also shown that monkeys subjected to a protocol where every reward is pre- ceded by a stimulus, this conditioned stimulus activates dopamine neurons even when the reward no longer does [232,233]. These findings indicate that dopamine release in the NAcc displays an inverse association with the pre- dictability of a reward, and that dopamine has an important role in associative learning.

Drugs of abuse tap into the mesolimbic dopamine system As discussed above, dopamine release is under tight temporal control in physiological conditions. Addictive drugs, however, act to release dopamine or to mimic or enhance the effects of dopamine release [79,133]. In addition, drugs of abuse produce higher dopamine levels in the NAcc than natural rewards do [263], and this happens regardless of prior experience, i.e. there is no adaptation [207,293]. The question then arises as to what is the link between such dopaminergic activity and addiction. As previously mentioned, it has been shown that dopamine encodes not only the incentive salience of the reward per se, but also incentive characteristics of stimuli associated with the reward. This concept is important in the incentive-sensitization theory proposed by Robinson and Berridge [219,220]. According to this theory, the dopaminergic stimulation renders the reward system hypersensitive to drugs and drug-associated stimuli. It is argued that this sensitization includes strong ‘wanting’ (i.e. craving) and that this in turn leads to the compulsive drug seeking and loss of control over drug taking seen in addicts [219].

Differential response to palatable and bland food Peciña et al. used a genetic approach to study the role of dopamine in food reward [196]. Mutant mice with a reduced dopamine transporter density and, hence, elevated extracellular dopamine levels were produced; these mice had higher motivation (‘wanting’) for sucrose in a runway task, while hedonic taste reactions (e.g. rhythmic tongue protrusions) were not different from the

20 behaviour of wild type littermates [196]. While this indicates that dopamine is related to ‘wanting’ aspects of food reward, it may be hypothesized that dopamine release and function vary with the perceived palatability of the food. Indeed, it was reported that in animals with intermittent daily access to sucrose, dopamine release was not attenuated over a period of three weeks [212]. Note the lack of this effect in two different control groups, one with intermittent daily access to chow and one with continuous access to sucrose. The repeated release of dopamine may thus be specific for intermittent access to palatability. Note that repeated dopamine release in response to food is seen also during more severe food restrictions [29]. In conclusion, reward produced by both palatable food and drugs of abuse implicate the mesolimbic dopamine pathways, but recruits this system in different ways.

Dopamine and obesity Positron emission tomography (PET) may be used in humans to analyze dopamine release. By means of [11C]raclopride, a radioactively labelled se- 11 lective antagonist for D2 receptors, displacement of [ C]raclopride by released dopamine can be visualized as it occurs. As expected from animal data, dopamine is released during feeding [247,298]. Indeed, dopamine is released in proportion to the degree of pleasure experienced while eating [247]. Interestingly, imaging studies have shown that obese individuals have perturbations in brain regions related to reward [259,260]. Dopamine D2 receptor availability is reduced in the striatum of obese individuals propor- tionally to their BMI [276,301]; i.e., the highest BMI is associated with the lowest D2 availability. In addition, obese individuals have blunted activation in the caudate nucleus during consumption of palatable foods, e.g. a milkshake [259]. Importantly, Stice et al. [258] showed that genetic variants of the D2 gene were associated with activity in the dorsal striatum in response to palatable food. Furthermore, the pattern of D2 receptor binding associated with obesity, i.e. reduced binding potential in the striatum, has been associated also with drug addiction. Volkow and colleagues have shown that cocaine addicts have reduced dopamine release and D2 receptors in the striatum [296]. This indicates that while the mesolimbic dopamine system is differently affected by acute intake of food and drugs of abuse, long-term changes may be similar. Another possible explanation is that the difference in D2 receptor binding occurs already before the development of obesity and drug addiction. This would indicate that vulnerability to both obesity and drug addiction involves D2 receptor phenotype. In fact, such evidence exist [180,258]. It should also be noted that the reductions of striatal dopamine in drug addicts have been associated with decreased activity in the orbitofrontal cortex

21 and the cingulate, a finding that has been speculated to be related to the loss of control over intake of drugs [296]. Volkow et al. also reported that the reduced D2 levels in the striatum of obese subjects were accompanied by low activity in the dorsolateral prefrontal cortex, medial orbitofrontal, cingulate, and somatosensory cortices [295,300,301]. It has been postulated that the low levels of D2 receptors in drug addicts would reduce the rewarding potential of natural stimuli and thus promote further drug use [296]. By analogy, it may also be hypothesized that reduced D2 receptor levels in the obese reduce the rewarding properties of regular food, promoting the intake of palatable energy-dense tastants. This proposi- tion remains to be verified.

Opioids and food intake Endogenous opioids and reward If the dopaminergic system is primarily related to ‘wanting’ rewards, the question arises as to what other neural systems are responsible for the ‘liking’ component. Again, important clues came from the study of addiction. Here, the function of the opioid system in feeding will be discussed. Exogenous opiates such as morphine produce intense euphoria in humans and are liable to abuse, as evidenced by the long track of opium use throughout history. When the endogenous opioids were identified in the 1970’s, it was thus hypothesized that their physiological role would be related to reward or hedonia. Indeed, opioid peptides were self-administered by experimental animals and produced conditioned place preference in rats [6,272], verifying the reinforcing properties of these agents. The endogenous opioids include endorphins, enkephalins, dynorphins, endomorphins, and nociceptin/orphanin FQ (Table 2). These peptides target the opioid receptors: mu (MOR), delta (DOR), kappa (KOR), that bind preferentially enkephalins, endorphins and dynorphins, respectively, and opiate receptor-like 1 (ORL1), that selectively binds nociceptin/orphanin FQ.

22 Table 2 Endogenous opioid peptides Amino acid sequence Receptor -endorphin YGGFMTSEKSQTPLVTLFKNAIIKNVHKKGQ MOR/DOR Leu-enkephalin YGGFL DOR Met-enkephalin YGGFM DOR Dynorphin A YGGFLRRIRPKLKWDNQ KOR Dynorphin B YGGFLRRQFK VVT KOR N/OFQ FGGFTGARKSARKLANQ ORL1 Selected opioid peptides involved in food intake regulation presented with amino acid sequences in rat and receptors for which the peptides have preferential affinity. N/OFQ, nociceptin/orphanin FQ; MOR, mu opioid receptor; DOR, delta opioid receptor; KOR, kappa opioid receptor; ORL1, opiate receptor like-1.

Opioids promote feeding When opioid receptor antagonists, such as naloxone and naltrexone, became available, the role of opioids was established first for feeding in general, and then especially for the feeding of palatable food. These antagonists are more efficient at reducing intake of palatable chow than of standard chow in rats. Indeed, naloxone injections did not affect intake of chow or starch diets in food-restricted rats, but attenuated intake of sweet chow, polycose, and chocolate chip cookies [102,147,287]. This indicates that endogenous opioids are involved in hedonic feeding. It was originally believed that opioids specifically mediate fat preference and intake of HF diets. Indeed, the first macronutrient food-choice studies indicated that morphine preferentially increased the intake of fat in rats [160,161,163] and that naloxone reduced such preference [162]. However, food preferences vary between individuals in an outbred group of rats [243,244] and when correcting for such baseline variance, it was shown that fat consumption was stimulated by morphine injections in fat-preferring animals, while carbohydrate intake was reduced by such treatment in carbohydrate-preferrers [86,106]. Naloxone reduced the intake of preferred foods [104]. Further evidence for the implication of opioids in hedonic feeding comes from studies on the intake of non-caloric sweeteners. Consumption of saccharin solutions and preference for such tastants were reduced by naloxone [146,152,266]. Moreover, mice deficient for the MOR have reduced preference for palatable saccharin solutions [306]. Sham-feeding and sham-drinking are methods that may be used to separate oral reward from post-oral effects of palatable tastants. In the sham- drinking protocol, a gastric fistula is implanted into the stomach of the experimental animal. When the animal is not presently in an experiment, the fistula is closed and the animal feeds as normal. When in the experiment, however, the fistula is opened and the consumed fluid is removed from the

23 stomach before it is absorbed. Rats increase their intake of palatable fluids under these conditions, since there is no negative feedback. However, such overconsumption may be attenuated by naloxone injections [129,221]. In another setup, sucrose was injected directly into the stomach of rats and these injections were paired with a neutral taste, thereby conditioning the animals to prefer the taste. Importantly, such post-oral effects of sucrose were not blocked by naltrexone [17]. These results reveal that opioids are involved in the reinforcement mediated by oral reward but not in the post- oral effects of palatable tastants.

Opioids interact with dopamine signaling It was initially established that opioids act centrally and not peripherally to alter food intake by showing that an antagonist capable of passing the blood- brain-barrier (naloxone) reduced food intake while an antagonist that remains in the periphery (quarternary naloxone) was not efficacious [47]. Since the mesolimbic dopamine pathway had already been identified as the ‘reward system’, it seemed plausible that opioids interact with dopaminergic signaling or overlap with this system. Indeed, peripheral morphine injections activate neurons in the NAcc as shown by experiments using a marker for neuronal activity, the immediate early gene, cFos [36]. Morphine injected into different sites within the ventral striatum (including the NAcc) but not into the dorsal striatum induces feeding in non-deprived rats [20]. The effect was more pronounced for food that was preferred [86]. This strongly indicates that opioid receptor stimulation within the NAcc is implicated in feeding. Injections into the NAcc of selective agonists of the opioid receptors have revealed different roles for different receptors: agonists of the MOR and DOR stimulated food intake while KOR agonists had no effect [157,307- 309]. Noteworthy, MOR and DOR stimulation within the VTA also promotes food intake, while KOR agonists do not affect feeding [42,181,182]. Interactions of opioids and dopamine were also examined. Strikingly, stimulation of the MOR and DOR in the VTA causes a release of dopamine in the NAcc, while the KOR mediates inhibition of this release within the striatum [80,252,253]. The effect of the MOR and DOR stimulation in the VTA is not directly on dopaminergic cells but on GABAergic interneurons [124]. Note that the effects of opioid agonists on dopamine release is in agreement with the effect on feeding. Thus, the receptors that mediate hyperphagic effects (i.e. MOR and DOR) are also the ones that promote dopamine release. KOR acts to reduce dopaminergic activity and agonism of this receptor within the NAcc does not affect feeding.

‘Hedonic hot-spots’ in the ventral striatum The gustatory impact, whether hedonic or aversive, of a tastant may be measured using the taste reactivity test [32,108]. The hedonic response rep-

24 ertoire is largely conserved across species such as rats and primates and include rapid tongue protrusions and, in the rat, licking of the paws [256]. Such behaviours in response to sucrose were indeed increased by morphine injections [192]. The neural mechanisms behind this behaviour has been anatomically localized to regions denoted ‘hedonic hot spots’ within the NAcc shell [194,195,249]. While injections of MOR agonists throughout the NAcc shell promote feeding, only through these subregions are hedonic taste reactions elicited [195]. These findings further consolidate the role for opioids within the NAcc in hedonic feeding.

Opioids: beyond the nucleus accumbens Noteworthy, opioids do not exert their actions on food intake solely within the canonical ‘reward system’. Expression of both opioid peptides and receptors are indeed found in several nuclei commonly associated with homeostatic food intake and energy balance [187]. Morphine injections into either the PVN or the amygdala promote food intake [103,174]. Nonetheless, the VTA and the NAcc appear to be very important regions for the action of opioids on feeding. It should also be noted that the ‘reward system’ and the hypothalamic food intake regulation are not independent from one another but interact through an array of different molecules and connections, e.g. orexin signaling from the lateral hypothalamus to the VTA [12]. Furthermore, important input reaches the NAcc and the VTA from other regions within the brain and affects dopaminergic and opioidergic signaling. Such afferents stem from e.g. the amygdala, prefrontal cortex, hippocampus, and brain stem [127]. In summary, opioids are strongly implicated in the hedonics of feeding, especially oral reward, as shown by several lines of evidence. In particular, seeking and consumption of palatable food appear to be affected by opioid receptor manipulation. MOR and DOR stimulation within the VTA and NAcc promotes dopamine release and consumption of palatable food, while KOR activation in the NAcc attenuates dopamine release and does not affect feeding.

Diet-induced obesity in the rat In the rodent, obesity can be produced by a number of methods [38] such as lesions of key regulatory regions of the brain, by pharmacological interven- tions, by genetic manipulations, and by diet-induced obesity (DIO). DIO has several features that make it an especially attractive model of human obesity (Table 3). Importantly, there is a large variance in the responsiveness to the DIO protocol within an outbred group of rodents [144,278]; this variance may be used to study the etiology of obesity. Individuals that gain weight

25 excessively are considered to be obesity-prone (OP) while animals that do not are resistant to weight gain (OR) on this diet. It was early demonstrated that the vulnerability to DIO has genetic components. Schemmel reported varying body weight gain on a high-fat diet in seven different strains of rats [231]. Levin produced two rat lines that are at the extreme ends of the vulnerability-to-obesity spectrum by bi-directional selective breeding [144]. Noteworthy, both dopaminergic and opioidergic signaling have been implicated in DIO. Examples include the findings that dopamine levels are reduced in the striatum of DIO rats [99], and that chronic antagonism of the MOR in the NAcc attenuates DIO [142].

Table 3 Characteristics of the diet-induced obesity (DIO) model Human obesity DIO in rodents Dietary fat and sugar High levels, variable High levels, controlled1 Sedentary lifestyle Yes Yes Genetic component 70%2 High Excess weight gain Over years Over weeks/months Defense of excess wt Yes Yes3 Food craving Yes Some indications Binge eating Yes/No With intermittent access4 Compulsivity Yes In some protocols5

Several features of DIO make it attractive as an animal model of obesity. 1Variable in ‘cafeteria diet’ protocols; 2Maes et al. [153]; 3Levin et al. [145] 4Avena et al. [16]; 5Heyne et al. [117]

Defining the ‘diets’ in diet-induced obesity In the DIO paradigm, animals are presented with food of higher energy than standard rodent chow to model exposure to the ‘Western diet’ of excess fat and sugar [62]. Such diets are often based on standard rodent chow but with higher levels of fat (e.g. lard or Crisco) and/or sucrose. As an alternative, nutritional supplements such as Intralipid and Ensure have been successfully employed in many studies on food intake and obesity [145,188]. In some experiments, instead of modelling the Western diet, animals are provided with actual food items from the ‘Western menu’. Bassareo and Di Chiara have employed Fonzies/Twisties to study dopamine release during consumption of palatable food [26-29]. However, the ‘cafeteria diet’ is the most striking paradigm in this context. Originally developed by Sclafani and Springer [238], this protocol provides the rodent with access to a range of palatable tastants. Food items have included chocolate chip cookies, salami, cheese, bananas, marshmallows, milk chocolate, bacon, and liver pâté [75,85]. One obvious drawback of the cafeteria setup is that the composition

26 of the diet is uncontrolled, preventing accurate measurements of food and macronutrient intake.

Food craving The phenomenology of food craving Food craving is defined as an intense desire to consume specific kinds of food [118,284]. Craving differs from hunger by being directed at specific foods, and differs from food preference by virtue of the intensity [284]. Such desires to eat may serve physiological functions during e.g. vitamin deficiency, but the food items commonly craved do not readily support this concept. Craving is frequently reported for pizza, ice cream, chips, sweets and desserts, and above all chocolate [227,285,291]. A recent study in a Japanese population reported craving for sushi [132], showing that food craving is strongly influenced by external factors, such as culture [227]. The majority of adults report food craving [118], and the palatable nature of craved foods indicates that such craving is based on sensory (or postoral) reward, rather than nutritional value. Tiggemann and Kemps [265] performed an experiment with a qualitative approach where subjects were asked to recall and describe a craving episode.

The most common theme to emerge from descriptions of the actual experience of the craving was its overwhelming nature. Over half [61.5%] of the respondents described how they ‘couldn't stop thinking about’ the food, how ‘it really played on (their) mind’, and how they ‘could not concentrate on anything else’. They characterized the craving as ‘an obsession’, ‘overpowering’ or ‘really intense’ (e.g. ‘I felt like I would die if I did not have the chocolate’). Most often [32.3%] this was accompanied by physical reactions like mouth watering and stomach grumbling, and occasionally [4.6%] by ambivalence or feel- ings of guilt. [265]

Since the foods that are targets of food craving are generally more energy- dense than the habitual diet [101], it has been speculated that obese individuals experience stronger food craving and that this behaviour may contribute to the vulnerability to obesity or the long-term changes associated with the obese condition. Noteworthy, craving for high-fat and high-sugar foods was positively associated with BMI [40]. In addition, binge eating in obese individuals [107] and in bulimics [283] has been strongly associated with craving. Females consistently report more frequent and stronger food craving than males [40,227]. In addition, craving appears to differ over the menstrual cycle [285], being more frequent during the perimenstrual period [227] or

27 the luteal phase [58]. The role of hormonal changes in such craving has been questioned [122]. Moreover, food craving is frequently reported during pregnancy [284].

Caloric restriction or deprivation from palatability? During dieting, energy-dense foods are commonly excluded from the diet in order to reduce the amounts of calories eaten. It may thus seem logical that food craving is precipitated by this state of negative energy balance. How- ever, reports provide weak or no support for this notion [118,119,265,284]. Surprisingly, total craving scores are reduced after harsh caloric restriction such as very low calorie diets [114,164]. This drastic method for dieting is, however, uncommon. There are two restrictions occurring simultaneously during dieting. First, a caloric restriction that may activate physiological responses such as hunger or slowed metabolism. Second, a restricted access to high-fat high-sugar tastants, and the replacement of these by less preferred, albeit healthy, food items. Most diets deny the dieters the pleasure of eating food they like. Un- surprisingly, dieters thus report higher craving for food items they are cur- rently abstaining from [118]. In agreement with this, Polivy et al. showed that a 7-day period of deprivation from a specific food (chocolate) led to increased ratings of craving for the missing food, and also resulted in a reduced latency to start eating and increased consumption of the food after the deprivation was lifted [206]. In a similar experiment, Coelho et al. showed that 3-day deprivation of carbohydrates provoked craving for carbohydrates and subsequent elevated intake, while protein craving was induced by protein deprivation [57]. Taken together, these results strongly suggest that dieting provokes craving not because of caloric restriction but because it deprives subjects of palatable, preferred food. Reduced calorie consumption may strengthen this behaviour but is not the main cause of it. This notion may have implications for diet regimens, since craving induced by ‘hedonic deprivation’ is reported to limit the adherence to body weight control programs.

Craving for drugs of abuse Drug addiction and alcoholism are chronic relapsing disorders with periods of abuse, abstinence, and relapse to compulsive drug use [273]. Drug craving during abstinence, oftentimes reported as the driving force behind relapse, has been reported for e.g. alcohol [141] and cocaine [205]. While the term is not included in the clinical definition of substance dependence as put forth in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV), it is acknowledged that craving for the drug of abuse is ‘likely to be experienced by most (if not all) individuals with Substance Dependence’ [10]. According to the incentive-sensitization theory of addiction proposed by Robinson and Berridge [219,220], craving and relapse are caused by drug-

28 induced sensitization of the brain circuitry underlying the attribution of incentive salience to stimuli. Thus, as the drug use continues, the stimuli becomes progressively more ‘wanted’ and finally associated with obsessive craving and compulsive use [219]. Note that the incentive-sensitization theory of addiction is attractive in part because it separates between ‘wanting’ and ‘liking’ the drug; indeed, drug addicts continue to crave the drug even when they no longer like it [219], and behaviour is reinforced by drugs at doses much lower than doses reported to be liked [140]. With regard to food craving, however, food that is the target of craving is still considered palatable.

Visualizing craving In drug addicts, brain imaging studies have identified some key components of the neural systems implicated in craving. Anatomical structures activated during drug craving episodes include the NAcc, caudate nucleus, thalamus, amygdala, orbital and dorsolateral prefrontal cortex, anterior cingulate, and insular cortex [121,275]. Notably, these structures are implicated also in food craving. Food craving-induced BOLD fMRI activity was investigated by Pelchat et al. [198] in subjects that were exposed to a monotonous diet for five days and, during the experiment, instructed to visualize consuming either the same diet or a preferred palatable diet. After the scan, subjects were asked whether they had experienced food craving during the experiment. All of the subjects on a monotonous diet reported experiencing craving for the alternative, palatable food and, in addition, had higher fMRI activity in the hippocampus, insula, and caudate nucleus. No differences were observed in the prefrontal cortex in this experiment; however, other findings support the role of this structure in food craving. Since activity in the dorsolateral prefrontal cortex is implicated in drug craving, manipulations of this area might affect the subjective feeling of craving. To this end, a number of experiments have shown that repeated transcranial magnetic stimulation (rTMS) of the dorsolateral prefrontal cortex may alleviate craving compared to sham magnetic resonance in addicts of cocaine [44] and nicotine [7] (see also [23]). Importantly, rTMS also reduce food craving in healthy control women [267] and in women with binge eating disorder [269]. Thus, the dorsolateral prefrontal cortex is arguably implicated in both drug craving and food craving.

Animal models of craving To further investigate the molecular mechanisms behind craving, animal models are required. Several approaches to objectively measure craving in animals have been employed (Table 4). Most of these models were developed in the field of addiction research, and most employ the operant self-

29 administration technique, in which animals are conditioned to respond at a lever or a nosepoke hole to receive a reward such as a food pellet or a drug injection (via an indwelling intravenous catheter). Operant (instrumental) conditioning differs from classical (Pavlovian) conditioning in that it requires a voluntary (i.e. operant) act on the part of the subject. In classical conditioning, a neutral stimulus such as a bell is repeat- edly coupled to a motivational stimulus such as food, so that the neutral stimulus becomes capable of eliciting the same innate response as the motivational stimulus (e.g. salivation). In contrast, operant conditioning is the process by which a behaviour (e.g. lever press or nosepoke) is followed by a stimulus (e.g. food pellet delivery) in order to change the frequency of the behaviour (food pellets increase the likelihood of subsequent responses at the lever, thus being a positive reinforcer). The operant conditioning technique was developed by Skinner [246]. Table 4 Animal models of food craving Setting Food References Progressive ratio Operant box + [51,139] Extinction-Reinstatement Operant box – [90,92,280] Deprivation effect Home cage +++ [14] Response rate Operant box ++ [139] Cue-induced feeding Operant box ++ [52] Persistent food-seeking Operant box – [117] Scheduled feeding Home cage +++ [15,212] Runway task Runway + [196] Behavioural models used to investigate motivation for food reward and food craving in rodents. Setting refers to the locale of the test. Food refers to whether food reward is actually obtained; a minus sign denotes only food-seeking (e.g. responding for food-associated stimuli in the absence of food); + small amount of precision pellets; ++ substantial amount of precision pellets; +++ time-restricted feeding, free access. Selection of references. During acquisition of the operant response in a typical experiment with palatable food as the reinforcer, the animal is once each day placed in an operant chamber programmed to deliver one food pellet each time the animal responds at a lever (fixed ratio 1; FR1). The chamber is fully automated and programmed by the experimenter according to the demands of the study. After a few sessions, the subject will have acquired the lever press response task and self-administer food pellets until satiated or until the program ends (e.g. after an hour). The response requirement is gradually increased to e.g. FR5 to produce a high but stable level of responding during the sessions.

Progressive ratio Consumption of a given reinforcer may be investigated using FR schedules, which provides the animal with free access to the reinforcer; the animal may thus consume the reinforcer until satiated. In contrast, if the animal is on the

30 progressive ratio (PR) schedule of reinforcement, the effort needed to receive each consecutive pellet is increased progressively until finally the animal stops responding. The number of pellets earned on the PR (referred to as break point) is used as a measure of the rewarding potency of the reinforcer. Psychostimulants have high break points [255], sucrose pellets intermediate (Paper III) and alcohol have only low break points [204]. Craving is evoked in the absence of the object of the craving, e.g. palatable food during dieting or drugs of abuse during abstinence. In the PR, animals are not able to consume the reward in the amount they do during FR schedules. Thus, it has been suggested that the PR be used as a model of craving [159]. Break points for sucrose is affected by sucrose concentration [235,236], but operant responding under PR schedules may also be affected by genetic background, sex, environment, emotional reactivity, experience, stress, and concurrent cues [255]. Pharmacological manipulations also affect performance: examples include reduced PR responding for sucrose after opioid and cannabinoid receptor antagonism [250].

The issue of food restriction It should be noted that the vast majority of experiments of self- administration of palatable food employ food restriction in order for the animals to respond more. As discussed above, food craving in humans is not the product of caloric restriction but of the exclusion of certain foods from the diet. Non-deprived animals responding for palatable food items is thus a more valid animal model of food craving. Palatable foods are highly reinforcing in rats even in the sated state and food restriction is not a requirement for robust responding. We used this approach both in Paper III and in Paper IV.

Personality types in obesity Increasing evidence indicates that a set of personality traits may make an individual more vulnerable to develop psychiatric conditions or harmful behaviours later in life. Novelty-seeking in humans is a personality trait characterized by e.g. impulsiveness, monotony intolerance, excitability, and disorderliness [55]. High novelty-seeking has been associated with development of type II alcoholism [109], drug abuse [63], compulsive gambling [70], and attention deficit hyperactivity disorder (ADHD) [82]. More recently, high novelty-seeking has been associated with disinhibition of eating, e.g. failure to stop eating when sated [100], bulimia nervosa [87], obesity [88] and binge eating [110]. Sullivan and co-workers also reported that in a weight loss program, unsuccessful obese subjects had higher scores in novelty-seeking than did successful subjects [261].

31 Response to novelty in rodents High and low novelty-seeking behaviour has been characterized in a number of other species including pigs [116], rainbow trout [271], great tits [289], and rodents [77]. In rodents, the response to novelty, which is oftentimes defined as locomotor activity in a novel environment, is commonly used to differentiate between high and low novelty-seeking. Interestingly, response to novelty has been associated with motivation for drugs of abuse in operant self-administration. Individuals that are high-responders in novel explorative settings also self-administer cocaine [41], alcohol [173], nicotine [262], and amphetamine [201] more readily than low-responders.

Emotional eating Anxiety disorders have been implicated in both obesity and eating disorders such as anorexia nervosa, but the causal relationship is not fully understood. Cross-sectional studies have shown that the prevalence of anxiety disorders is higher in the obese than in the non-obese [24,125,200,239,240], but while some prospective studies indicate that anxiety is a result of obesity [8,98], other reports show that anxiety disorders may be part of the etiology of obesity [9]. Anecdotally, anxious individuals consume palatable energy-dense ‘comfort food’ as a means of relieving anxiety and this behaviour possibly leads to excessive weight gain.

The growing problem More than one in every ten Swedes suffers from obesity (BMI 30 kg/m2); 44% of the population are overweight, thus have a BMI 25 [230]. Al- though this number is still modest compared to figures from other countries such as the US (66% of the adult population being overweight or obese [175], more than 30% obese [185]), the prevalence of obesity doubled in Sweden between 1980 and 2005 and it is still rising. Also, the prevalence is higher in Swedish subpopulations such as males in rural areas, with 58% having a BMI exceeding 25 [177-179,230]. It has been estimated that the annual cost for obesity and diseases related to obesity in Sweden is 3 billion SEK [229]. The indirect cost for obesity is in the same order of magnitude. Although recent data indicate that there is no longer an increase in the number of overweight children [149,170], childhood obesity is still a major health problem world-wide [158,185,268,279]. This is alarming because an increasing number of children develop diabetes and other risk factors for cardiovascular disease [19,93,95,245], but also because excess body weight during the preschool years tends to remain throughout childhood and into adult life [34,94,95,166,213].

32 The level of dietary fat intake is under intense debate as one of the critical factors associated with childhood obesity. The current knowledge on the effect of dietary fat on body weight in children, however, is somewhat controversial. Previous studies have suggested that the level of dietary fat intake in preschool children is associated with higher body-mass index (BMI) values [154-156], to have no predictive effect on obesity [5,13,72], and, surprisingly, to protect against body weight gain [96,115]. Thus, there is a dire need for more information on this complicated relationship.

33 Aims

Zoë: Cap'n'll have a plan... always does. Kaylee: That's good right? Zoë: It's possible you're not recalling some of the cap'n's previous plans...

Firefly: The Series

In the modern environment, hunger is no longer the main driving force for food intake. Rather, the availability of a range of palatable foods allows dietary choices to be made based on individual taste, eating habits, and emotional states. Such choices are also influenced by external cues and cultural factors, and subsequent consumption of energy-dense food may lead to body weight gain and obesity. Therefore, it is important to study the determinants of food choices, such as preference and craving for food. The overarching goal of this thesis was to propose a model for the process driving genetically predisposed individuals to overconsume palatable foods, which subsequently leads to food craving, loss of control over food intake, and excessive weight gain. We designed a number of experiments to investigate different stages of this process. The working hypothesis was that in the environment with ample food sources, individuals’ driving force behind eating behaviour can ‘evolve’ From food preference to craving. Because overeating in obese individuals shares a number of features with drug use in addicts, we chose to apply animal models from the addiction research field to gain further insight into the mechanisms that predispose individuals to overeat and that make abstinence from palatable high-energy food so difficult to maintain. Having this in mind, the aims of the studies were to:

Paper I • Optimize a food choice paradigm in order to identify individual preference for a palatable high-fat diet in rats. • Test the assumption that dietary preference for palatable food would relate to behavioural traits such as anxiety-like behaviour, novelty- seeking, and risk taking.

34 • Test the hypothesis that gene expression of hypothalamic neuropeptides in untreated rats is associated with the food preferences measured in the food choice paradigm.

Paper II • Investigate the effects of palatable diets on anxiety-like behaviour and novelty-induced locomotor activity in rats. • Investigate expression of genes related to food intake and reward in rats exposed to either high-fat or high-sugar diets, or both.

Paper III • Set up the operant self-administration technique in this laboratory and develop a protocol that allows us to measure motivation for palatable food in sated i.e. non-deprived animals. • Test the assumption that the risk-taking trait identified in Paper I would predict the operant behaviour motivated by palatable food in sated rats.

Paper IV • Set up a model for diet-induced obesity with special emphasis on individual vulnerability to such dietary manipulations. • Develop a model of food craving with high face validity in rats, and to employ this model in obesity-prone and obesity-resistant animals.

Paper V • Identify molecular mechanisms involved in the craving behaviour observed in Paper IV in obesity-prone animals. • Test the hypothesis that obesity-prone and obesity-resistant animals differ in the expression of genes implicated in food reward. Such differences may be present before obesity develops, or alternatively be an effect of either excessive weight or excessive consumption of palatable food. We sought to parse these possible contributors from each other.

Paper VI • Analyze food preference data from a large number of parents of young children, in order to test the assumption that parental food preferences and subsequent eating behaviour in the family predispose young children to body weight disturbances.

35 Material and methods

When I was otherwise quite restored to health, the sight of a chemical instrument would renew all the agonies of my nervous symptoms

Mary Shelley, Frankenstein

Diet-induced obesity and food choices Herein, we employed three diet protocols: a three-diet food-choice protocol, a free-choice HFHS diet, and a non-choice HFHS diet.

Food preference diets The diets for the food-choice paradigm in Paper I were modified from [31]. The composition was altered slightly, based on a pilot study, to enhance the preference for the palatable high-fat diet and to produce stable individual food preferences. The rats were simultaneously provided with three different diets: a palatable high-fat diet and two reference diets, viz. a high-protein and a high-carbohydrate (Figure 2). Preference for the high-fat diet was the main outcome and was determined by average kcal% from the palatable high-fat diet over the last three days of the paradigm.

Figure 2 Subjects had access to high-fat, high-protein, and high-carbohydrate diets in the home cage. High-fat diet was provided in a food bowl. Food preference was consistent in the individual but varied between individuals within the group.

36 Free-choice HFHS The high-fat high-sugar (HFHS) diet employed in Paper II, Paper IV, and Paper V had previously been used for diet-induced obesity [139,197]. In addition to ad libitum regular chow and tap water, rats exposed to the HFHS have free access to a bowl of lard and to a bottle of sucrose (30% w/v) solution. la Fleur and colleagues showed that after 7 d of HFHS access, rats had ingested more calories than a group with access to regular chow only and, as a consequence, had increased total fat weight and plasma levels of both leptin and glucose [139]. More recently, it was reported that gene expression of POMC and NPY in the ARC of free-choice HFHS-treated animals was affected by this dietary protocol [138]; NPY expression was upregulated and POMC downregulated, in agreement with the hyperphagic behaviour. The strength of the HFHS diet is that the regular food is supplemented with palatable, calorically dense tastants, as opposed to standard obesogenic diet paradigms that substitute the regular chow for a high-energy diet alto- gether [81,143]. Thus, this paradigm is well suited for the withdrawal procedure, since the palatable food items may easily be added or removed, while the regular chow and tap water remains throughout the experiment. In addition, individual food preferences and the effects thereof on body weight may be analyzed. In our experience, individuals will differ greatly in both macronutrient intake and body weight gain in this paradigm. However, it should be noted that previous investigations have not differentiated between OP and OR animals on this diet.

Non-choice HFHS The van Heek series for DIO comprise three diets with different amounts of energy from fats. The high-fat/high-sugar (D12451; Research Diets, New Brunswick, NJ, USA; 45 kcal% from fat, 20 kcal% from sucrose) diet used in Paper V has been used extensively and effectively to produce DIO, see e.g. [35,83,270]. Note that this diet is commonly referred to as ‘high-fat diet’ in the literature, despite the fact that it also contains a high amount of sucrose. In this thesis, we more adequately refer to this diet as ‘high-fat high- sugar’.

Gene expression measurements Real-time quantitative PCR Kary Mullis developed the polymerase chain reaction in the early 1980s and was awarded the Nobel Prize in chemistry for this achievement (see [25]). While the standard PCR protocol amplifies a DNA product, it cannot be easily used for quantification, since the concentration of amplicon will reach a plateau that is dependent only on the amounts of reagents (nucleotides), not

37 on the amount of starting material. However, if a dye (SYBR green, TaqMan probes, etc.) is added during the actual reaction, the amplicon concentration can be measured at each cycle, in real time. The cycle at which the signal reaches a certain threshold value may then be used to calculate the original concentration. In this thesis, the qPCR method is used to determine the gene expression levels of genes of interest in relevant tissue samples collected from the experimental animals. After RNA extraction, reverse transcriptase (RT) is used to transcribe mRNA into complementary DNA (cDNA). Random hexamers are used as primers for the RT reaction, which means that all RNA in the sample will be transcribed into cDNA. The primers for the qPCR analysis are, however, specific for the genes of interest and will amplify only seg- ments of these genes.

Reference genes To correct for different amount of total cDNA in samples due to variance during dissection, sample handling, RNA extraction or RT reactions, it is necessary to use an internal reference. In this thesis, several reference genes were employed. The ideal reference genes should be expressed in the same amount in all cells within a given tissue type, and should not be affected by the treatment or experiment itself, since this would exaggerate or obscure the potential effects of the intervention on the expression of the genes of interest. The most stable reference genes are commonly found among the ‘house- keeping genes’, i.e. genes coding for proteins involved in basal cellular function and thus vital for cell survival. Reference genes employed in this thesis are listed in Table 5.

Table 5 House-keeping genes used in this thesis Reference gene Cellular function bACT -actin, Structure and kinetics of the cytoskeleton GAPDH Glyceraldehyde 3-phosphate dehydrogenase; glycolytic pathway enzyme bTUB -tubulin; structure and kinetics of the cytoskeleton CYCLO Cyclophilin; protein folding and interactions RPL19 Ribosomal protein L19 H3b Histone protein 3b SDCA succinate dehydrogenase complex, subunit A; mitochondrial protein The most stable group of reference genes was determined for each set of samples using the method proposed by [274].

38 Instead of using a single house-keeping gene as reference, we determined the most stable set of reference genes as proposed by Vandesompele [274]. Thus, we analyzed multiple house-keeping genes for each set of samples (i.e. for each brain region and experiment) and used the geometric means of the most stable genes as reference. The main drawback of the Vandesompele method is that it requires a substantial amount of material for the reference genes; samples that produce a small amount of cDNA cannot be analyzed in this way. Due to this problem, in the analyses of VTA samples in this thesis, a single HKG was used as reference. Conditions for the qPCR reactions are described in Paper I. Primers for the reference genes and for the genes of interest are shown in Table 6.

39 Table 6 Primers for qPCR (see Abbreviations) Transcript Accession Forward primer Reverse primer Reference (house-keeping) genes bACT BC063166 cactgccgcatcctcttcct aaccgctcattgccgatagtg CYCLO M19533 gagcgttttgggtccaggaat aatgcccgcaagtcaaagaaa GAPDH X02231 acatgccgcctggagaaacct gcccaggatgccctttagtgg H3b XM_235304 attcgcaagctcccctttcag tggaagcgcaggtctgttttg RPL19 NM_031103 tcgccaatgccaactctcgtc agcccgggaatggacagtcac SDCA NM_130428 gggagtgccgtggtgtcattg ttcgcccatagcccccagtag bTUB NM_173102 cggaaggaggcggagagc agggtgcccatgccagagc Neuropeptides AgRP XM_226404 tgtaaggctgcacgagtcc ttgaagaagcggcagtagcac CART NM_017110 tgcttgtgaaggggtgacagc ttaaagcggctccagggacaa CRF NM_031019 gttgagaaactgaagagaaaggg actgttgttctgcgaggtac Dynorphin NM_019374 aagaaggctacacggcactg tgagacgctggtaaggagttg Galanin NM_033237 cttccctacgccgctgatc gtccaagtgtctccgtccag GALP NM_022633 gctgtatgccgtgcttttcc agtctggtgtttctgcgagg GHRF NM_031597 ttaagatgggaattttaggg ccttgaccaggagagatc GRF NM_012767 gagtgtgacattgacgctgag cgctgctgggtatagaaatgc MCH NM_012625 tcggttgttgctccttctctg gtggtcttgtcatcgtcatcc Nociceptin NM_013007 cttgcttcccacggctgc ctgctgaacacgctggagag NPB NM_153293 tgccctgtgtgtcaaagatg tgcagcgacaagaagacg NPFF NM_022586 taagggaccccacccatcac tggagcagaacacgcatgag NPY NM_012614 cagaggcgcccagagcag cagccccattcgtttgttacc Orexin NM_013179 ctccttcaggccaacggtaac cagggcagggatatggctcta Oxytocin NM_012996 cggtggatctcggactgaac tagcaggcggaggtcagag POMC AF510391 tgggtcacttccgctggg tcctccgcacgcctctg PRRP NM_022222 cggcaggagaagggcaac acgctgagagaacttggtgc Somatostatin NM_012659 agcccaaccagacagagaac cctcatctcgtcctgctcag TRH NM_013046 caggctttaggtctaggatgatg tctctgaagtcggtaaggaagg Vasopressin NM_016992 tgctcaacactacgctctctg cctcctcttgggcagttctg Ucn 2 NM_133385 ctcctccctcagaactatcc ggctggctttagagttgc Receptors

CRF2 NM_022714 agagaatgcctggagaatgg gtcatacttcctctgcttgtc D1 M35077 cgggctgccagcggagag tgcccaggagagtggacagg D2 NM_012547 agacgatgagccgcagaaag gcagccagcagatgatgaac MOR NM_013071 aaagccctggatttccgtacc gcagaccgatggcagaagag KOR NM_017167 taccagcatattcaccttg aatgacatccacatcttcc Other DAT S76145 gctgcgtcactggctgttgc ctgtccccgctgttgtgaggt AAAD U31884 ctggggaagggggaggagtg gcagctggcggatcattttag TH NM_012740 gcccccacctggagtattttg agacacccgacgcacagagc COMT NM_012531 tgggctggggcttggtga gatgcgctgctcctttgtgtc

40 Exploration of novel environments: rodent models In this thesis, several tests have been used to study anxiety-like behaviour and novelty-induced locomotion in rats. These models all rely on the innate exploratory drive. This drive to explore novel and unknown areas are of considerable value for most species; without it, animals would not find new food sources or mates, and would not identify threats. For rodents in the wild, however, exploration comes with the risk of predation and the evolu- tionarily fit animal accurately balances the risks of exploration with the potential benefits. Importantly, internal and external factors influence this behaviour; external factors include e.g. known presence of predators or light, while internal factors include physiological and emotional states. Since exploratory drive is also preserved in the laboratory animal, and in the laboratory setting, we may use tests of exploration to measure the emotional state of animals, provided that we can control for external and, to some extent, internal factors. This is the basis for the tests of exploration of novel environments used in this thesis. In all tests, the subject is placed in a novel arena and behaviour is recorded for a short period of time (10-20 min).

3 2

Figure 3 The Multivariate Concentric Square Field™. 1) Dark Corner Room; 2) bridge; 3 centre field; 4) hurdle. Picture kindly provided by Dr. Erika Roman.

The Multivariate Concentric Square Field™ (MCSF) The MCSF is an arena used for the evaluation of a range of explorative and risk/safety assessing strategies in rats and mice [168]. The arena consists of a number of compartments with different environmental settings, such as an open centre field (41 × 41 cm), a dark corner room (DCR), a hole-board device accessible by openings slightly elevated above the floor (hurdle), and

41 an illuminated, wire mesh platform (bridge). Three corridors connect these zones. See Figure 3. The bridge and the DCR are contrasting areas with regard to risk and safety. Food deprived male rats remove pellets placed on the bridge and hoard them in the DCR [168]. Lactating females carry their offspring to the DCR if the pups are placed on the bridge [168]. Because the bridge is thus associated with risk, the number of visits to this area is interpreted as risk- taking behaviour [224]. Herein, we evaluated the risk-taking behaviour in the MSCF of rats that differed in their preference for palatable HF food. The test has previously been used to compare strains of animals that differ in their preference for alcohol. AA rats that are alcohol-preferring have elevated risk-taking behaviour in the MCSF compared to ANA (not alcohol-preferring) rats [225]. The AA rats spent more on the bridge and less time in the DCR, compared to ANA rats. In the alcohol-preferring sP rats, however, risk-taking is lower than in their counterparts, the not alcohol-preferring sNP strain [223].

The elevated plus-maze (EPM) The EPM is a test of anxiety-like behaviour utilizing rodents’ aversion for unprotected areas: a high preference for the open arms (associated with risk) of the maze is interpreted as a low level of anxiety-like behaviour, i.e. risk taking is inversely associated with anxiety-like behaviour [199]. The preference for open arms over closed arms (expressed as either %time spent on open arms or %number of entries into open arms) is the classical index of anxiety-like behaviour in the EPM, since this parameter is independent from locomotor activity, but it has been shown that inclusion of additional pa- rameters such as time spent in centre, or number of stretched attend postures and head dips over the edge of any arm may provide further strength to the analysis [222].

Open field (OF) The open field test is commonly used to study locomotor activity in rodents after e.g. drug administration. In the present thesis, we also used it as a measure of anxiety-like behaviour. For analysis, the circular arena (diameter 90 cm) was divided into three areas: a centre area of 30 cm across (A), an intermediate zone of 15 cm (B) and a thigmotaxis zone of 15 cm next to the walls (C). A change in distance moved in the centre with no change in total distance is used as a measure of anxiety [210] since the centre of an open field is associated with risk.

42 Operant self-administration Apparatus Food-reinforced operant responding was studied in MED-PC operant cham- bers (Med Associates Inc., St. Albans, VT, USA) placed inside sound at- tenuating boxes. When the program started, two levers were extended from the same wall and a house-light on the opposite wall was illuminated. Presses on the active (right) lever produced, according to different schedules, deliveries of a single food pellet to a receptacle. Each delivery was coupled with the illumination of a cue light above the active lever. During 5 s after a delivery, presses on the active lever were recorded but had no programmed consequence. Presses on the left (inactive) lever had no programmed consequence at any point but were also recorded. Each session lasted 60 min. See Figure 4.

A B

Figure 4 The rodent subject during a trial in the operant box. A) Responding on the active lever causes illumination of the stimulus light above the lever and the delivery of a reinforcer to the receptacle (food cup). B) The rat collects the reinforcer in the receptacle. Pictures kindly provided by Dr. Chris Pickering.

Acquisition During the acquisition phase, rats are food deprived to increase the motivational value of food pellets. The animal explores the operant chamber for two acquisition sessions on two consecutive days; food is returned to the home cage two hours after the second session. Rats put in the box for the first time approach the lever and activate it as part of their exploration rou- tine; this lever press response is readily reinforced by the delivery of a reinforcer (food pellet). Food deprivation during acquisition ensures that >95% of all animals acquire the lever press response within three days.

43 Fixed ratio and progressive ratio On a FR schedule, the subject needs to press the lever a fixed number of times to receive each delivery. Low FR schedules are needed to maintain self-administration of substances like ethanol [204] while higher FR schedules (e.g. FR5) are more suitable for psychostimulants which have exceptionally high reinforcing ability and can support the execution of several thousand lever press responses within a single session [159]. In this thesis, we chose to use FR5 based on a pilot experiment. It should be noted that using extensive training and an increased number of food pellets awarded (6 pellets of 45 mg), rats can be trained to press on a FR300 schedule [228], a feat which would not be accomplished reproducibly via delivery of a single precision pellet. To measure motivation, it is generally accepted to use the progressive ratio (PR) schedule of reinforcement, which exponentially escalates the lever press response requirements within-session [216]. The response required for the delivery of a pellet equalled 5e(delivery number×0.2)-5 which yields the following sequence: 1, 2, 4, 6, 9, 12, 15, 20, 25, 32, 40, 50, 62, 77, 95, 118, 145, 178, 219, 268, 328, etc. This formula has been frequently used previously for measurement of motivation for sucrose [172,216]. The break point is defined as the maximum number of pellet deliveries in this PR session and is an index of the amount of effort an animal is willing to make in order to earn the reward [216]. Again, reinforcers like ethanol produce small break points [204] while psychostimulants sustain higher break points [255]. The break point on a PR schedule is also thought to measure craving for the reinforcer [159] and, thus, an increase in the PR response provides evidence for food craving. There are important conceptual differences between the FR and PR schedules. To highlight these differences, we chose to set no limit to the number of available rewards during the FR5 sessions in this thesis. In other words, the subjects were able to freely self-administer pellets of the appropriate type. This is in sharp contrast to the PR scenario, where animals are able to receive only a restricted number of rewards, far below the desired amount. As such, the PR is a measure of the reward efficacy, or the motivation towards the reward [216], whereas the FR schedules rather measure consummatory behaviour.

Questionnaires Food Craving Inventory (FCI) Food craving is defined as an intense desire to eat certain kinds of food [284]. By measuring self-reported craving scores on a Likert scale from 1 to 5 for a number of food items, preferences for macronutrients or groups of

44 food items (e.g. high-fat diets) may be measured [285]. White et al. [291] developed the FCI, an instrument that evaluates preferences for foods rich in 1) high-fats, 2) sweets, 3) carbohydrates, and 4) fast food. Such food cravings are strongly correlated to subsequent food intake [165] and are associated with body weight in males and females [40]. Note that there is a qualitative differences between the two high-fat food categories such that ‘Fast Food’ have a higher content of saturated fatty acids (e.g. french fries, chips), whereas the relative amount of proteins is higher in the ‘High-Fats’ category (e.g. fried fish, steak).

Dutch Eating Behavior Questionnaire (DEBQ) Wardle et al. developed and validated this frequently employed questionnaire [281] that contains three different scales: Emotional eating, Restrained eating, and External eating. Representative questions include: “Do you have a desire to eat when you are anxious, worried or tense?” (Emotional eating); “When you have put on weight do you eat less than you normally do?” (Re- strained eating); “If you see others eating do you also want to eat?” (External eating).

Swedish Parenthood Stress Questionnaire (SPSQ) The SPSQ measures aspects of perceived psychosocial stress associated with parenting [191] and it is a revised version of the American Parenting Stress Index [2]. The psychometric qualities for SPSQ are valid for the total score and the subscales. The incompetence subscale focuses on general experiences of care-giving, while role restriction is concerned with life restrictions due to parents’ responsibilities. The social isolation subscale examines feel- ings of loneliness and availability of social contacts. The spouse relationship problems scale concerns partnership issues. The health problems scale measures parental physical health. Responses are on a five-point Likert scale and higher scores indicate higher levels of stress.

45 Results and discussion

Rats, however, are not simply walking 96-well plates

Anxiety-like behaviour and high-fat diet preference (Paper I) Preferences for palatable high-fat food and food choices stemming from these preferences have been linked to weight gain. But what are the molecular and behavioural determinants of preference for such food? In this study, we tested the hypothesis that preference for a palatable high-fat food is associated with 1) anxiety-like behaviour and 2) hypothalamic gene expression of neuropeptides known to affect food intake (Table 7). We found that in outbred Wistar rats, there was high inter-individual variability in anxiety-like behaviour in the elevated plus-maze (EPM). Variance was also observed with regard to preference for a palatable high-fat diet in a food choice paradigm (ranging from 26% to 95% of total food intake). These findings were expected [208,244] and formed the basis for the correlation analysis employed to identify relationships between the behaviours. The main finding of this study was that a risk-taking behavioural phenotype characterized by low anxiety-like behaviour was predictive of a high preference for the HF diet. We identified strong correlations (rs=0.629) of open arm activity (risk taking) in the EPM and preference for the HFD (Fig- ure 5); this indicated an inverse association of anxiety-like behaviour and HFD preference (high open arm activity equals low anxiety-like behaviour). Because intake of high-fat foods has been linked to weight gain and adiposity in humans [91,167,217,218] and laboratory animals [278], this suggests that the risk taking trait may also make the individual more prone to weight gain. Table 7 Experimental set-up for the food preference study (Paper I) Day 1 15-16 18-22 31 Adaptation & MSCF Food-choice Wash-out Tissue handling EPM paradigm (Standard chow) collection Behavioural tests (MCSF, Multivariate Concentric Square Field™; EPM, elevated plus-maze) were conducted prior to the food-choice paradigm. A ‘wash-out phase’ on standard rodent chow was included subsequent to the exposure to the food choice diets.

46 The animals were killed and dissected after a ‘wash-out’ period included to allow the effects of the brief access to the food choice diets to subside and, thus, mRNA levels to return to standard chow-feeding conditions. As- sociations of food preferences and gene expression were analyzed with Spearman rank correlations. Our laboratory has previously employed a similar approach to associate self-administration of ethanol with genes in the hypothalamus [202], prefrontal cortex, hippocampus and amygdala in rats [203]. Expression levels of 21 transcripts within the hypothalamus were analyzed using quantitative real-time PCR. While several genes (AgRP, NPY, MCH, TRH) were associated with intake of standard chow or body weight, the expression of a single gene correlated to HFD preference: urocortin 2 (Ucn 2; Figure 5). A number of neuropeptides have been previously implicated in food or macronutrient preferences. Thus, it was somewhat surpris- ing that we identified only a single neuropeptide whose expression was related to food preference. On the other hand, this makes the implication of Ucn 2 in food preference especially intriguing. Ucn 2 is part of the CRF family of neuropeptides and binds selectively to the type 2 receptor, CRF2 [215]. Ucn 2 injected i.c.v. produces an anorectic response in food-deprived rats [186], reduces nose-poking for food [123] and overall chow intake in non-deprived rats [68]. The present association of Ucn 2 expression levels and preference for dietary fat may implicate Ucn 2 in vulnerability to DIO. Indeed, genetically OP rats bred for susceptibility to DIO are, in a pre-obese chow-fed state, less sensitive to the anorectic effects of Ucn 2 than obesity-resistant rats, indicating differences in the Ucn 2 system between these strains [68]. Also, CRF2- deficient mice consume more than wild-type littermates when kept on a high-fat diet [21].

Figure 5 HFD preference is associated with open arm exploration (% of total entries); expression of Ucn 2 is associated with preference for the palatable high-fat diet (HFD). rs, Spearman rank correlation; *p<0.05; ***p<0.001.

47 High-fat diets affect locomotor activity and gene expression (Paper II) The aim of this study was to investigate the effects of short term (10 days) exposure to palatable diets on behaviour in the open field test (locomotion and anxiety-like behaviour), and on gene expression in a number of sites related to food intake regulation and reward. We exposed animals to either free-choice HFHS diet, or to HF or HS alone. Since it is known that there is a large variance in the anxiety-like behaviour in outbred rat popoulations [208], we screened the subjects for anxiety- like behaviour in the elevated plus-maze prior to the diet intervention and matched the diet groups for baseline behaviour within the two separate studies. By the end of the experiments, the rats were again tested for anxiety-like behaviour. However, we used the open field test for the retest since the EPM is not suited for repeated testing due to carry over effects according to most [3,4,49,69] but not all [45] reports. This general design has been previously used with pharmacological treatments but not with regard to feeding [150,151]. Thus, the outline of this experiment was 1) baseline test for anxiety-like behaviour and novelty-induced locomotion in the elevated plus- maze, 2) access to high-fat diet and/or high-sugar diet for 10 days, 3) retest in the open field, 4) animals were killed and tissue samples collected. There was a hyperphagic response in all diet groups compared to controls. However, the short exposure (10 days) was not sufficient to produce any statistically significant differences in body weight between the groups (Fig- ure 6A). In the HFHS group, there was a positive correlation between HF diet consumption and body weight change (rs=0.590, p=0.021; Figure 6B). In the HF group, this correlation was similar but did not reach statistical significance (rs=0.559, p=0.096).

Figure 6 A) Body weight (mean±SEM) did not differ between high-fat (HF) and/or high-sugar (HS) diet groups after 10 days. B) Consumption of the HF diet was associated with body weight gain during the 10 days in the groups with access to either HF diet alone or both high-fat and high-sugar (HFHS).

48 The experimental set-up was chosen to fully compensate for baseline behavioural profile. Indeed, there was no difference between the diet groups when the unadjusted values were compared. Two regression models were tested, one with locomotor activity (total distance moved in the open field) as the dependent variable and the other with time spent in centre as the dependent variable (this variable is inversely related to anxiety-like behaviour). The main finding of this study was that consumption of the HF diet was associated with both increased locomotion and reduced anxiety-like behaviour. In addition, we observed that tyrosine hydroxylase (TH) expression was upregulated in the VTA of HFHS animals. This is particularly interesting since our laboratory has previously reported the same effect, but in food restricted animals. Furthermore, MOR expression was upregulated in the lateral hypothalamus of HF animals. Taken together, these results show that short-term exposure to a free- choice high-fat diet induces behavioural adaptations such as elevated locomotor activity and attenuated experimental anxiety. The changes observed in gene expression relevant to reward after high-fat diet exposure indicate that these behavioural adaptations are related to reward function.

Anxiety-like behaviour and motivation for palatable food (Paper III) The aim of this study was to investigate the associations of behavioural traits, such as anxiety-like behaviour and novelty-induced locomotion, with the operant self-administration of sucrose and high-fat pellets in non-food deprived rats. This is relevant because anxiety and novelty-seeking have been associated with obesity and binge eating in humans [73,88,110,239]; Sullivan et al. also reported elevated novelty-seeking in obese subjects that were unsuccessful on a weight loss program [261]. Furthermore, we wanted to test whether the association of risk-taking behaviour and preference for palatable food (Paper I) would extend to the operant paradigm.

Figure 7 Active lever press response on progressive ratio schedule (PR) for 95% sucrose pellets is significantly correlated to total number of arm entries (FRQ total), open arm preference (%DUR open arms) and number of unprotected head dips in the EPM. * P <0.05; ** P <0.01.

49 We set up the operant self-administration technique in our laboratory. Im- portantly, we optimized a protocol for self-administration of palatable food in sated, i.e. non-deprived, animals. Male Wistar rats were subjected to the EPM test of anxiety-like behaviour. The rats were tested for FR5 and PR operant responding for different food pellets (50% sucrose food pellets, 99% sucrose pellets, and high-fat food pellets). The main finding was that the PR active lever press response for 95% sucrose, but not the other pellet types, was associated with % time spent on open arms (P=0.019) and total activity (P=0.009) in the EPM (Figure 7). On the FR5 schedule, self-administration of all pellet types was associated with total activity in the EPM (50% sucrose, P=0.009; 95% sucrose, P=0.020; high-fat, P=0.034). This indicated a robust association of novelty- induced activity and self-administration of palatable food in sated rats, as well as a specific association of PR lever press response for 95% sucrose and low anxiety-like behaviour. It has been argued that such active lever press response on PR may be interpreted as craving for the reinforcer; thus, our findings indicate an inverse relationship of experimental anxiety and craving for sucrose. The association of high novelty-seeking [173,201] and low anxiety-like behaviour [41] with self-administration of drugs in experimental animals has been reported. Self-administration of food or sucrose pellets is commonly included as control in such experiments, and indicate weak associations of novelty-seeking with certain features of FR responding for sucrose [43,77,130]. In this experiment, we studied sated (non-deprived) animals in order to use a valid model of food overconsumption in humans, i.e., eating palatable food when the homeostatic demands have already been met. While food deprivation further motivates the subjects to food-reinforced behaviour, it is recognised that rodents, as well as humans and other mammals, are motivated for palatable food even when sated [22]. Indeed, the subjects in our study were held on ad libitum access to standard chow but still consumed substantial amounts of food on the FR5 schedule in the operant chamber. Also, homeostatic signals such as insulin and leptin have previously been shown to affect sucrose-motivated operant responding [89], but we found no correlations of circulating factors to the operant response in the present experiment. Given the weak associations of self-administration of palatable food and response to novelty in earlier studies, it is tempting to speculate that operant responding for palatable food in sated animals is associated with response to novelty, while influence from homeostatic signals occludes this connection in food-restricted animals.

Craving at withdrawal from palatable feeding (IV) In Paper IV, we wished to develop an animal model of craving for palatable food, based on the concept that food craving is produced not by caloric re-

50 striction per se, but by the exclusion of palatable foods from the diet. We used a previously described model of diet-induced obesity [137,139,197] to study the effects of long-term access to a high-fat high-sugar (HFHS) diet on self-administration reinforced by palatable food (sucrose pellets). We tested the assumption that animals that differ in their vulnerability to DIO, i.e. OP and OR animals, would differ also in their behavioural response to HFHS. In addition, we studied the effects of forced withdrawal from HFHS on the self- administration of sucrose. See Table 8. Table 8 Experimental set-up for the food craving experiment (Paper IV & V)

wk: 1 2 3 4 5 6 7 8 9 10 11 12 Diet Chow High-fat/High-sugar Withdrawal (chow) Operant testing & Procedure Acquisition Diet induced obesity Operant testing open field test Animals acquired the lever press response task prior to any dietary manipulations and were subsequently exposed to high-fat high-sugar food during 4 weeks before any further tests. Animals were grouped as obesity-prone or -resistant based on weight gain at 5 weeks. During withdrawal, animals returned to standard rodent chow for 18 days. The animals were divided into OP and OR groups in relation to animals fed a standard diet. Operant self-administration of palatable pellets did not differ between OP and OR animals during access to the HFHS diet. How- ever, during withdrawal, the OP rats displayed elevated break points for sucrose pellets (Figure 8), anxiety-like behaviour, and a delayed re- adaptation to a standard diet compared to OR rats. These findings show that extended access to a HFHS diet causes differential adaptations in OP and OR animals that become apparent at withdrawal from such diets. We interpret the elevated break points in the OP animals as craving for palatable food in the state of withdrawal.

Figure 8 Active lever press response and break point (mean ± SEM) on progressive ratio (PR) schedule motivated by sucrose (A) prior to withdrawal (WD), (B) 1st week of WD, and (C) 2nd week of WD from the high-fat high-sugar diet (HFHS). OP, obesity-prone; OR, obesity-resistant; *p<0.05.

Potential neural circuitry underlying the food craving in rats (Paper V) The observation (in Paper IV) that OP animals exhibited craving and other signs of withdrawal during abstinence from the HFHS diet prompted us to pursue the molecular mechanisms behind these behaviours. Thus, after the behavioural tests were complete, we decapitated the animals and collected the NAcc and CPu for gene expression analysis. At this point, 18 days had elapsed since the animals’ access to the palatable HFHS diet. D1 and D2 receptor gene expression was analyzed, as well as expression of COMT, and the opioid receptors MOR and KOR. In the NAcc, we found a 25% lower

52 gene expression of D1 receptor in the OP compared to OR animals, as shown in Figure 9. These differences in gene expression between OP and OR animals may be present prior to the dietary manipulations, they may be the result of extended access to HFHS diet, or they may be produced by the withdrawal condition. To investigate the effects of extended and continuous access to HFHS (no withdrawal), the second experiment was performed where the animals were killed after 5 weeks of HFHS access. Animals were grouped based on weight gain into OP and OR. We found that D1 receptor expression was lower in the OP group, and that MOR gene expression also was lower in the OP animals under this condition. Thus, these differences between the obesity phenotypes were present during continuous access to HFHS and extended well into the withdrawal phase. Finally, to parse the effects of extended access to HFHS diet from the effects of body weight gain, we analyzed mRNA levels in the NAcc in two groups of rats pair-fed for identical calorie intake, that received either standard chow or HFHS (HFHS pair-fed, these animals are lean i.e. have identical weight curves to chow-fed controls; Figure 9C). We found that both D1 and D2 receptor expression was reduced in HFHS pair-fed animals compared to chow-fed controls. Thus, dopamine receptor downregulation occurred in response to exposure to HFHS even when energy intake did not meet the desired level. Lower striatal D2 levels have been previously reported in obese rats [113] and humans [276,301]. The functional relevance of this receptor downregulation in obese individuals is not presently understood, but it is worth noting that OLETF rats, that have a genetically induced downregulation of D2 receptors in the NAcc, display increased sensitivity to D2 agonists [113] and antagonists [112]. Thus, the present results suggest that after extended access to HFHS diet, the mesolimbic reward system is sensitized to dopaminergic signals even in the non-obese state (lean HFHS pair-fed animals).

53 A

Figure 9 Body weight change and relative gene expression (measured using RT- PCR) in the NAcc (mean±SEM) during the high-fat high-sugar (HFHS) access in obesity prone (OP) and obesity-resistant (OR) animals; A) Extended and continuous HFHS access (no withdrawal); B) Withdrawal after extended access to HFHS diet; C) Extended but pair-fed HFHS access. D1, dopamine D1 receptor; D2, dopamine D2 receptor; COMT, catechol-O-methyl transferase; MOR, mu opioid receptor; KOR, kappa opioid receptor; *p<0.05; **p<0.01; ***p<0.001.

The finding from the mRNA expression analysis that opioid receptors are expressed at different levels before and during withdrawal raises the question whether opioidergic signaling is involved in the observed behavioural effects during withdrawal. In a study by Colantuoni et al. [59], it was illustrated that opioidergic pathways indeed are related to withdrawal from palatable food: in rats that had been subjected to daily sucrose presentations for

54 an extended period of time, injections of the opioid antagonist naloxone induced a state similar to opiate withdrawal, indicating that the animals had adapted to a presumably higher release of endogenous opioids. This effect was not observed in control animals that had not been exposed to sucrose. The present data show that the opioid system is differently affected by withdrawal in OP vs. OR animals. Thus, KOR signaling may be related to the craving and anxiety-like behaviour previously observed in OP but not OR animals at withdrawal from HFHS.

Parental food preferences and body weight in offspring (Paper VI) Parental factors, such as stress and food habits may produce an unhealthful environment and lead to obesity in preschool children. In the present study, we interviewed parents of 915 children (Figure 10) and collected body weights of the children to examine the relationship of parental factors, such as food preferences, to body weights in preschool children. Food preferences were measured with the Food Craving Inventory that records preference for high-fat/high-protein food items (‘High-Fats’), high-sucrose (’Sweets’), high-carbohydrates (‘Carbohydrates’), and ‘Fast food’.

Figure 10 Flow chart of the data collection. Parents who declined to participate did not provide demographic data. Parents who dropped out provided demographic data but did not answer the full questionnaire on eating behaviour, parental stress, and other factors.

The levels of food preferences were compared between mothers and fathers. Overall, both mothers and fathers reported highest preference for Carbohy- drates and there was no difference between the genders in this variable (p=0.14). There were, however, gender differences in the scores for other food types. FCI scores for Sweets were reported to be higher in females (t766=9.0; p=2.9E-18), while craving for High-Fats (t766=5.04; p=5.8E-7) and Fast Food Fats (t766=10.9; p=7.4E-26) were higher in males. Total craving

55 scores were slightly higher in males than in females (t766=3.2; p=0.0013). These results validate and expand on previous findings [40]. The relationship of preferences for High-Fats and Sweets was analyzed using Spearman correlations. Preferences for High-fats and Sweets were not associated in mothers (rs= –0.055, p=0.099). A weak positive correlation was found in fathers (rs=0.10, p=0.005).

Table 9 Food preference by body weight in mothers

Normal-weight mothers Overweight mothers Mean SD Mean SD P-value FCI scores High-Fats 2.53 0.73 2.120.57 1.2E-19 Sweets 2.38 0.59 2.750.69 3.7E-17 Carbohydrates 2.52 0.51 2.69 0.64 2.4E-05 Fast Food 2.08 0.68 2.28 0.77 4.5E-05 Total 2.38 0.41 2.460.42 0.003

Eating behaviors Restrained 2.25 0.90 2.850.89 2.0E-22 Emotional 1.82 0.67 2.681.01 1.1E-44 External 2.81 0.62 3.270.67 5.3E-25 Scores are presented as the mean value and SD of the different factor scores from the Food Craving Inventory. Mothers are divided into normal-weight and overweight categories based on self-reported body weight. In mothers, each score differed between normal-weight and overweight mothers (Table 9). Craving scores for High-Fats were higher in normal- weight than in overweight subjects. All other scores were higher in overweight mothers. In fathers, differences between normal-weight and overweight individuals were less pronounced (data not shown). Restrained, emotional and external eating scores were all higher in overweight than in normal-weight parents. Mothers’ and fathers’ food preferences were associated with body weights of offspring as shown by logistic regression, and the results for mothers are presented in Table 10. In both mothers and fathers, preference for High-Fats was associated with lower odds for children to be over- or underweight. Fast Food preference did not show any associations with body weight of offspring. The associations of paternal craving scores and children’s body weight showed a pattern similar to the results from mothers. Preference for Sweets in fathers was not associated with body weight of offspring.

56 Table 10. Associations of maternal food craving scores with odds ratios for being overweight and underweight in preschool children Overweight children Underweight children OR (95% CI) P-value OR (95% CI) P-value High-Fats 0.34 (0.22-0.53) 0.000 *** 0.43 (0.29-0.63) 0.000 ** Sweets 2.02 (1.40-2.92) 0.000 *** 1.40 (0.95-2.05) 0.085 Carbohydrates 1.56 (1.01-2.41) 0.045 * 2.57 (1.62-4.05) 0.000 ** Fast Food 0.99 (0.72-1.35) 0.939 0.96 (0.69-1.32) 0.785 Associations are presented as odds ratio (95% CI). Parental preferences for various food categories were assessed using the Food Craving Inventory scale [291]. Chil- dren’s body weight was measured during ambulatory medical care visits and classified into underweight, normal-weight, and overweight according to the guidelines of the International Obesity Task Force [60,61]. Data are adjusted for parental educa- tion, age, work status, and self-reported body weight. A p-value less than 0.05 was considered significant. Parental stress was associated with lower preference for ‘High-Fats’ in both mothers and fathers, and higher craving scores for ‘Sweets’ and ‘Car- bohydrates’ in mothers. Overall, the associations were stronger in females, perhaps mirroring the fact that mothers are more prone to parental stress than fathers [190]. Stress has been previously implicated in food intake and obesity in both human and animal studies [71,131]. Although, the level of emotional stress has been associated with the intake of carbohydrates in females [148], this is the first report suggesting increased preference for high- carbohydrate food in association with psychosocial stress. Problems caused by parenting may be very stressful both for the parent and the child. Parental stress has previously been associated with increased body weight of preschool children by us and others [131,257]. The present results show that food preferences and, presumably, food choices are associated with parental stress and, hence, may constitute a link between stress in parents and body weight in offspring.

Figure 11. Association of unadjusted food preference (mean±95% CI) in parents and BMI of their 3-y-old children. A) High-fats preference is highest in the parents of normal weight children. B) Sweets preference is lowest in mothers of normal weight children, compared to mothers of underweight or overweight children.

In conclusion, parents of normal weight children reported higher preference for high-fat/high-protein foods and lower preference for sweet-tasting foods as compared with parents of both overweight and underweight children (Figure 11). Thus, parental attitudes towards food in general and high- fat/high-protein foods in particular may impact body weight development during earlier childhood stages. The present findings indicate that parents’ food preferences and resulting choices of food to be presented at meals in the family are likely implicated in the previously reported positive association between parental stress and children’s body weight [257]. Our findings of increased number of normal-weight children in families where parents show a high preference for high-fat/high-protein foods warrant longitudinal investigations to study the association of parental food preference and body weight of young children over time. These results may be relevant for de- signing health and dietary policy strategies against early childhood obesity.

58 General discussion From food preference to craving Based on our review of the literature, and based on the findings reported here, we propose a theoretical model that delineates how food preference may lead to overeating and weight gain, which, in turn, produces neural adaptations that may propel repeated overeating and ultimately craving for food and loss of control over food intake. Genes and environment may interact at each stage of this process; food preference, overeating, weight gain following excessive food intake, neural adaptations in response to palatable food, and ultimately food craving generated by restrained eating may all be influenced by a genetic predisposition, with different genetic variants inter- acting at different stages. A genetic variant relevant to the dopaminergic system may e.g. determine an individual’s response to long-term palatable food exposure, but have no impact on the propensity to gain weight. Genetic factors that affect any of these steps may be conceived as part of the ‘thrifty genotype’ proposed by Neel [176] to describe vulnerability to obesity in the modern environment. We present the model graphically in the Conclusions section (Figure 13). The main outlines and the conclusions drawn from our own studies are presented here.

1. Preference for high-fat foods is determined by a genetic predisposition (related to risk-taking behaviour in rats) Energy-dense and highly refined foods are constantly available in our modern environment and may lead to obesity if consumed in excessive amounts. However, overconsumption is not seen in all individuals. We showed (Paper I; Paper II; Paper III) that there is a high variability in food preference of outbred rats, and that food preference and motivation for palatable food are linked to a risk-taking behavioural phenotype (Paper I; Paper III). In addition, we identified baseline Ucn 2 expression in the hypothalamus as a co- variate of preference for HF diet (Paper I).

2. Free access to palatable, preferred food leads to overeating and subsequent body weight gain in prone individuals When provided with ad libitum access to high-fat diets, animals with a preference for the food will consume larger amounts than required. Such overconsumption was shown to significantly impact body weight gain both short- term (Paper II) and long-term (Paper V). It should also be noted that while the preference and resulting intake of energy-dense food is associated with body weight gain, other metabolic factors also contribute to variance in body weight. The finding that only some animals gain weight in the DIO paradigm is in agreement with the gene × environment interaction proposed by Neel [176]: the OP animals do not differ from OR until HFHS diets are provided and the OP gain weight. When animals had free access to energy-dense food

59 over extended periods of time, the animals could be divided into OP and OR individuals based on their weight gain (Paper IV; Paper V).

3. Exposure to palatable diets causes behavioural adaptations and perturbations in gene expression relevant to reward Upon both short-term and long-term exposure to HF and HS diets, behavioural adaptations occurred (Paper II, Paper IV). We observed such effects both during access to the palatable diets (Paper II), and after discontinuation of HFHS access (Paper IV). Perturbations in systems relevant to reward may explain these behaviours. Specifically, gene expression of receptors for dopamine and opioids were implicated by our results. Noteworthy, we showed strong effects of HFHS diet on dopamine receptor gene expression after extended pair-fed HFHS access (Paper V). In this experiment, the animals were not able to consume the desired levels of HFHS diet and, thus, the potential confounding factor of body weight was excluded from these results.

4. Adaptations in the reward system drives further consumption of palatable food The aforementioned data allow us to speculate that extended exposure to palatable tastants alter the function of the mesolimbic dopamine system and, hence, processing of reward. We suggest that such alterations increase the preference for palatable food. In contrast to La Fleur et al. [139], we did not observe any differences in operant responding for palatable food when HFHS was available in the homecage. However, when the HFHS diet was removed and only the control chow provided, all animals that previously had access to HFHS responded by a severe and long-term reduction in food intake, indicating reduced preference for or acceptance of the standard diet (Paper IV).

5. Adaptations specific to OP individuals produce signs of withdrawal during discontinued HFHS access At forced abstinence from HFHS we observed a number of behaviours specific to the OP rats and including craving for palatable food as well as anxiety-like behaviour and delayed re-adaptation to the standard rodent chow (Paper IV). We also showed that molecular adaptations differ between OP and OR animals: dopamine and opioid receptors were differentially expressed in the NAcc in the two groups (Paper V). This finding suggests that extended palatable food exposure affects the preference/acceptance for palatable food in all animals, but stimulates craving only in OP rats. Thus, these individuals appear to be at risk for craving and resulting loss of control over food intake.

60 A controversy over terms The term ‘food addiction’ is controversial and provocative, yet compelling enough to spawn commentaries [65], symposia [66], and animal experiments [15]. In lay sources a growing problem of ‘food addiction’ is reflected by popularity of this phrase: the search for this term performed by Corwin and Grigson [66] with the Google search engine produced 334,000 entries in January 2009, while in March 2010 the number reached 12,500,000 items. Although this effect may stem in part from progress in the search engine, it certainly parallels the increasing interest in this issue. In the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) the American Psychiatric Association clinically defines drug addiction as substance dependence using a set of behavioural criteria of which at least three of the following items should be present within one year: 1) tolerance, 2) withdrawal, 3) larger amounts of drug taken than intended, 4) unsuccessful efforts to cut down drug use, 5) large part of life spent trying to obtain the substance, using the substance, or recovering from the substance, 6) time spent with family, friends and recreational activities given up because of the drug, 7) substance use continues despite knowledge of negative conse- quences [10]. It should be noted that compulsive use or loss of control over use is implicated in three of these criteria (3, 4, and 7). Loss of control over intake is observed also in overeating, which thus meets these criteria in a broader sense and may be classified as a ‘substance dependence’ or an addiction. However, the DSM-IV criteria were developed for drugs of abuse such as heroin, ethanol, and cocaine, and not for a ‘substance’ (i.e. food) that is required for the survival of the organism, socially accepted, and legal. Many argue that the term ‘addiction’ should be reserved for drug addiction. In the view of this author, the problem is more semantic than scientific. Compulsive intake does occur in both drug addicts and e.g. in binge eaters. Craving for the substance or food may be implicated in such loss of control over intake. The implications are the same whether we label the behaviour ‘food addiction’, addictive-like behaviour, or compulsive food intake. As we show and as we discuss in this thesis, there are both similarities and differences between drug use/addiction and overeating palatable food. Both the similarities and differences can be of service provided one wants to understand the nature of overeating in obesity or of drug addiction.

Preference for palatable food is associated with a risk-taking phenotype, not with anxiety-like behaviour A risk-taking behavioural profile in rats, characterized by low anxiety-like behaviour, was associated with preference for palatable diets in the food- choice paradigm and motivation for palatable diets in the operant paradigm. We propose that Ucn 2, an anorexigenic neuropeptide, is related to this association, based on the finding that gene expression of Ucn 2 peptide was

61 associated with the preference for a HF diet and that expression of CRF2, the receptor for Ucn 2, was associated with elevated risk-taking behaviour. We found no indications that high levels of anxiety-like behaviour would be predictive of subsequent overconsumption of either HF or HS diets; this negative finding was replicated in three studies. Thus, the association of anxiety in humans with the consumption of palatable ‘comfort foods’ is not replicated in rodents with regard to performance in the standard test for anxiety-like behaviour, the EPM.

DIO may be produced in rats of the OP but not OR phenotype We have used palatable and energy-dense HFHS diets to induce excessive feeding and obesity in rats. Importantly, we have shown that vulnerability to DIO is a highly individual trait; in rats provided with long-term access to HFHS diets, some animals were prone to weight gain (OP) while other seemed resistant (OR). Since vulnerability to weight gain in humans is also an individual trait that may lead to obesity when the person is provided access to HFHS diets, the DIO paradigm has high face validity as a model of human obesity. The importance of acknowledging variance in DIO vulnerability is under- scored by our behavioural (Paper IV) and molecular (Paper V) findings. Indeed, the operant experiment in Paper IV was designed to study differences between these phenotypes; hence, we were able to identify the effects of withdrawal from HFHS in OP animals. This heterogeneity may confound the effects if the phenotypes are not treated as separate groups. Hence, the present results, together with an ample amount of literature [144,278], allows us to propose that vulnerability to DIO should be acknowledged in the analysis of results and, more importantly, in the design of studies. We looked at factors that predispose rats to gain weight of which the first was food preference for a palatable high-fat food. We show that preference for or intake of dietary fat indeed produces excessive weight gain in rats; preference for high-fat food (Paper II), and consumption of the HFHS diet (Paper V), were associated with body weight gain. While this effect was not detected in Paper IV, the present findings add to the body of evidence showing that preference for or acceptability of energy-dense food is a trait that may propel excessive weight gain. Thus, the question arises as to the determinants of such food preference.

Sensitivity to reward is a genetic trait that may be manifested by preference for palatable food Palatability of food and thus the hedonic impact of food consumption affect how much we eat. Therefore, in a group of individuals, it is reasonable to postulate that those who experience the highest level of pleasure from eating palatable food may overconsume and, hence, be liable to gain weight over time. Such sensitivity to reward may be manifested by e.g. preference for

62 high-fat high-sugar food. Indeed, evidence indicates that reward sensitivity is associated with preference for palatable food and to overeating and BMI scores [74]. According to this viewpoint, palatable foods are sought because they are particularly rewarding. Based on the findings from the animal models in this thesis, we suggest that sensitivity to reward is affected by a genetic predisposition that favours preference for palatable food and motivation to obtain palatable food rewards, in effect causing overconsumption of energy-dense diets and subsequent weight gain. Interestingly, such genetic predisposition is also associated with a risk-taking, low-anxiety phenotype, as shown in Paper I and Pa- per III. Some points from alcohol research are of relevance to this discussion: Cloninger proposed that there are two separate types of alcoholism: type I alcoholism that is late onset and where alcohol use precipitates in response to adverse life events, and type II alcoholism that is early onset, has a strong genetic component, and that is associated with a novelty-seeking phenotype [54,55]. Novelty-seeking in humans, characterized by impulsiveness, monotony intolerance, excitability, and disorderliness [55], has been associated with not only alcoholism [54,109], but also drug abuse [63] and compulsive gambling [70]. Interestingly, Piazza and colleagues showed that novelty- induced locomotion, considered an animal model of novelty-seeking, is associated with the self-administration of amphetamine [201], a finding that was subsequently replicated for other drugs of abuse [41,173,262]. Similarly, we found in this thesis that a risk-taking phenotype was associated with elevated preference and motivation for palatable food.

Extended exposure to palatable food may induce reward deficiency that further increases preference and craving for palatable food In contrast, it might be speculated that individuals that are not as responsive to food (i.e. reward deficient) need to consume more palatable food in order to feel pleasure. Reward deficiency has been implicated in alcoholism and drug addiction [37]. According to this hypothesis, genetic variants in e.g. the D2 receptor make individuals less sensitive to natural stimuli, e.g. food, such that the individuals require stronger stimulants (drugs or alcohol) in order to be satisfied. Similarly, food preferences might shift toward those food items that have the strongest hedonic value, i.e. foods with high levels of fat and sugar. The hypothesis is that subjects with reward deficiency seek drugs or palatable food in order to compensate for the dysregulation of reward [277]. A condition similar to reward deficiency syndrome, the allostatic state, has been proposed to occur after repeated drug use [134]. According to this theory, drug addiction may be viewed as a negative hedonic state, a devia- tion from a ‘reward set point’, that motivates the drug addict to take drugs to alleviate the negative mood.

63 Based on the findings from the animal models in Paper II, IV and V, we propose that access to palatable food induces adaptations in the mesolimbic dopamine system that may, in effect, put the animals in a state of reward deficiency. We reported the striking effect (Paper IV) that animals in a state of withdrawal from palatable food reduce their caloric intake to less than half of control levels, even though standard chow is available ad libitum (Figure 12). These animals required two weeks to readjust to the consump-

Figure 12 Caloric intake in controls (grey), obesity-prone (white) and obesity- resistant (black) rats during access to high-fat high-sugar (HFHS) diet and during withdrawal therefrom (Paper IV). Note that both OP and OR display hypophagia and that OP rats are slower to recover from this state. tion of standard rodent chow. After a prolonged hyperphagia on energy- dense diets, increased demands of standard chow would otherwise have been expected; animals had thus adapted to the hedonic value of the tastants, not to the intake of large amounts of food. During this period, there was also elevated motivation for palatable food in the operant paradigm. We interpret this finding as a devaluation of the food reward provided by standard chow, brought about by continuous access to palatable food and hedonic feeding. As illustrated in Paper V, this effect may indeed be associated with dopaminergic signaling and opioid receptor activity.

Perturbations in dopamine signaling in the obese: cause or effect?

Reduced D2 levels have been detected in subjects of both drug addiction and obesity [276,296,301]. This could be due to that long-term exposure of drugs of abuse or palatable food changes receptor presentation, or that individuals with low D2 levels are predisposed to these conditions and, thus, that the dopaminergic disturbances predate the disorder.

64 To approach the implication of dopamine function in the vulnerability to drug abuse, Volkow [297,299] recruited healthy human volunteers that did not have a history of drug use and injected them with the psychostimulant methylphenidate, i.e. Ritalin®. The subjects rated whether their subjective experience of the drug’s effects were pleasant or unpleasant; however, baseline striatal D2 receptor levels had already been measured using PET. Inter- estingly, lower baseline D2 availability was associated with higher subsequent ratings of pleasantness. Thus, response to the psychostimulant action is affected by the dopaminergic phenotype of the subjects. This response (experiencing the drug as pleasant) may predispose individuals to repeated drug use. Importantly, these volunteers were not exposed to drugs before the experiments. Low levels of D2 in the striatum may, hence, be extant without drug addiction. This suggests that low D2 levels in drug addicts may predate the drug addiction. In the context of palatable food, Stice et al. [258] showed that genetic variants of the D2 gene were associated with activity in the dorsal striatum in response to palatable food. The present findings clearly illustrate that HFHS diets are capable of al- tering the dopaminergic system. While we report lower levels of dopamine receptors in the NAcc of OP rats, in agreement with previous studies of rodent and human subjects, we also show that HFHS diet reduces both D1 and D2 receptors in the NAcc upon extended HFHS exposure in lean, pair-fed animals. Such reductions of dopamine receptors have been associated with increased sensitivity to dopaminergic stimulations in rodents [112,113] and reward in humans [297,299]. Thus, we illustrate that perturbations in the dopamine system possibly related to reward sensitivity can be induced by extended access to high-fat high-sugar diets. Taken together, we suggest that perturbations in the reward function may predispose individuals to overeating and weight gain, and that such perturbations may also be evoked or exacerbated by continuous food reward associated with HFHS diets, thus providing a positive feedback leading to even greater overconsumption of palatable food (Figure 13). Both genetic predisposition and – as shown in this thesis – adaptations brought about by extended HFHS exposure may implicate dopaminergic and opioidergic pathways.

65 Conclusions

Humans, however, are not simply big rats

We have provided evidence for a common sequence of events, influenced by both genetic and environmental factors, that leads from food preference to craving. In Figure 13 we propose a model that shows major stages of this transgression. In the Discussion we present how genetic and environmental factors may contribute to this ‘evolution’ of a common food preference to oftentimes undesirable food craving; specific sections of the General Discus- sion are marked with appropriate numbers in our model (Figure 13).

Figure 13 From food preference to craving. Food preference and availability of high-fat high-sugar foods may stimulate overeating. In vulnerable individuals, overeating may lead to adaptations in the reward system that propel further overeating and craving for food. This state is characterized by a loss of control over food intake. Genetic predisposition and environmental factors interact with this progress at each step. Numbers refer to the specific sections in the General Discussion.

66 The specific findings from the studies were that:

• Preference for palatable food is associated with a risk-taking phenotype, but not with anxiety-like behaviour. Such preference may lead to over- sonsumption of food and subsequent weight gain.

• Vulnerability to obesity induced by HFHS diets in rats is an individual trait; animals may be categorized as either OP and OR.

• Exposure to HFHS diet induces behavioural adaptations in rats that indicate perturbations in the reward processing.

• Both OP and OR animals display reduced intake of standard rodent chow after withdrawal from a HFHS diet; this effect persists to the point of weight loss for up to two weeks.

• Extended HFHS diet exposure induces molecular adaptations (downregulation of both D1 and D2 receptor gene expression in the NAcc) that are related to the diets per se and not depending on concomitant weight gain.

• OP rats display distinct behaviour compared to OR at withdrawal from HFHS, including anxiety-like behaviour, slower recovery from hypophagia, and craving for palatable food.

• OP and OR animals differ in gene expression of both dopamine and opioid receptors within the NAcc and CPu. These differences are induced by the HFHS diet but persist even after long-term withdrawal from such diets.

• Reduced preference for high-fat/high-protein foods in parents is associated with body weight disturbances in the offspring.

67 Acknowledgements

Emellertid uppenbarar sig nöden i form av den obetalda hyresräkningen och avbryter mina vetenskapliga arbeten och metafysiska grubblerier genom att kalla mig åter till jorden.

August Strindberg, Inferno

Dear reader, after all this time and text you finally made it to the acknowledgements! You started off a bit hesitantly in the introduction, where I presented the assumption that overeating in obese individuals is similar in many aspects to drug addiction, e.g. implicating the ‘reward system’ in the brain. Encouraged you reached the aims of this thesis, which were to understand food craving in humans by using animal models, the specifics of which in- trigued you in the materials and methods! You then browsed through a couple of animal studies of rats pressing levers and learned, finally, that parental preference for high-fat foods is associated with normal weight development in small children. A couple of conclusions later you arrived here. Excellent!

A lot of people contributed to this thesis and I would like to take this op- portunity to duly thank the following: My brave steps towards a PhD started already when JONAS LINDBLOM was a post-doc at the Department of Neuroscience and without him I would not have enlisted to the PhD program. As an erudite neuropharmacologist with inclinations for scientific methodology, classicism, and Lovecraftian lore, he was the perfect supervisor were it not for his absence from BMC. He also put a lot of brains into my project. To HELGI SCHIÖTH, my main supervisor, I extend my gratitude for these four years of great fun and hard work, and for his excellent tutoring in writing and management. I am in great debt to my three tutors: ERIKA “SLETA” ROMAN, CHRIS “SUGAR DADDY” PICKERING, and PAWEL “THE BLURB” OLSZEWSKI. ERIKA ROMAN, coauthor and coercer of papers, lured me into her maze of animal behaviour from whence I have yet to emerge. CHRIS PICKERING, mentor and psycho pharmacologist, introduced me further into the realm of addiction research than was appropriate during that Grand Pub Crawl. PAWEL OLSZEWSKI’s vast knowledge of food regulations and his highbrow sense of

68 humor are equally important to this thesis, his attractive way of writing papers even more so. Dear tutors, thank you for all the laughs at the lab, for your company at conferences, for encouragement and support, for intellec- tual input, for comic relief, and for being my friends. I would like to acknowledge ROBERT FREDRIKSSON for his broad knowledge of science and scientific methods, ANNA-LENA HULTING for tenderly running the AFA project, OLGA STEPHANSSON for help and company in the lab, CHRISTINA STENHAMMAR for collaborations, and ANNA KINDLUNDH- HÖGBERG, my co-supervisor, for encouragement and input early on and during the writing of this booklet. My pregraduate students (the 11 apostles) were the muscle behind some of the work included in this thesis and some of the work excluded. I wish to thank them all for inspiration, motivation, and diversions: PETRA JONSSON, EMMA FRANS, KRISTIN BROHMÉ, THOMAS SIEVERTSSON, HIROKO SPÅMAN, ZANDRA GUNNARSSON, ELIN QUENNERSTEDT, JUN QIU, ELIN ALMSTEDT, ANNA HALLSTEN NORBÄCK, and HELENA WALLIN. I wish to thank J.J. and KALLE for living the PhD dream with me. And the other PhD students at Helgilab for friendship and scientific discussions, and for company during måndagsfika, parties, concerts, and zombie nights: TATJANA HAITINA, TOBIAS HILL, MARIA HÄGGLUND, MARKUS SÄLLMAN ALMÉN, PÄR HÖGLUND, SMITHA SREEDHARAN, MATHIAS RASK- ANDERSEN, BJÖRN SUNDBERG and JONATHAN CEDERNAES. The rest of the present Helgilab is also acknowledged. MARIA N. SOLVANG for writing-music, for awkward moments in D1, and for the laughs. ANICA KNACKARSCH for enthusiasm, bad jokes and video clips flagged inappropriate by users. EDWIN, KEDAR, SAHAR, CARRO, SOFIE, MADELEINE und CHRISTIAN for fun discussions and for creating a warm atmosphere. Previous members of Helgilab contributed with playfulness, dogs, great discussions, ski trips, table tennis games, and stupid jokes. I would like to thank LINN WALLÉR for late night phone calls and MAJD MIRZA for trying to fit the two largest chairs on Neuroscience into the same office. But I am also indebted to GUNILLA OLSSON, FREDRIK OLSSON, AGNIESZKA OLSZEWSKA, JON EDMAN, SHAHRZAD SHIRAZI, SANDEEP KADEKAR, ALI SHAIK, ANNA KARLSSON, and HUNG TRINH, for great company. The corridor would not be as pleasant without our cousins in the KULLANDER and MACKENZIE LABS, and our lovely building would be a dull place without the rest of the BMC node of Neuroscience. I would especially like to acknowledge GRETA HULTQVIST, LARS ORELAND, TOMAS LARSSON, HANNA-LINN WARGELIUS, and BENGT MEYERSON, and I would also like to thank KASIA ROGOZ and KING COBRA and all the other weekend workers and late night prowlers that make the off hours worthwile. Once upon a time, LENA BERGSTRÖM gave me an ambitious Master project with some hardcore HPLC action; thus commenced my scientific career.

69 The project failed miserably but she somehow still believed in me and has been encouraging and helpful ever since. I also want to thank CAROLINA BIRGNER, my former boss, for being the most loving cynicist I know; BJÖRN REINIUS, for sharing his thoughts on genetically enhanced children; SADIA ORELAND for support; and UWE and ULRIKE and TOBBE. For all the help with my little friends at C4, I would like to sincerely thank MARITA for her advice, and JESSICA, SUSSI, and the rest of SF. Life as a PhD student is also about the couple of hours each day you are not in the lab. I have been blessed with some wonderful friends whose support and perspective I value greatly. I wish to thank: ROBERT BURMAN, my flat-mate, for your competitive mind, for the couch, and for late discussions about the meaning of it all. HENRIK NILSSON for breakfasts at Fågelsången and rock clubs in Berlin. ERIK ALLARD for sushi and Firefly. PETER FRITZON for playing games. JESPER GANTELIUS for inspiration. ANDERS ENGSTRÖM for Bakerstôga. MATS NILSSON for talking and when needed listening. JOHAN BROMAN, JONNIE and JENNY HEDQVIST, NIKLAS WESSLING and others for all the vicious voracious weekend wam- pire fun. ERIK STRIDH, KRISTOFER KLERFALK, AMELIE FONDELL, KLAS PETERSON and GUSTAV CHRISTOFFERSSON for random activities (on ice). PETRA HANSSON for the post-DN beers. During the years I went to a couple of courses and conferences that held both scientific and social rewards. The PENS summer school in Dubrovnik 2009 was a week full of stimulating lectures and discussions, but also nightswimming and multilingual rude remarks. I would especially like to thank FREDERIQUE, ANTONIO and DIANA; it was also great to meet up with EOIN, MARTA, BRIAC, JACQUES, and the Spanish delegation at EBPS in Rome a couple of months later. The visit to CenMet in Lexington, KY, 2009 was another great experience full of Ace of Base, Bourbon, Kings of Leon, and a whole bunch of in vivo electrochemistry. Thank you SARA, CAROLINA, MARC-ANDRÉ, CALVIN, AJ, RAFAEL, RACHAEL, and all others for showing me that Kentucky is more than the fried chicken-thing, which we never found anyways. VGMK for its continuing mission, to seek out the Russian spirit wherever it may roam. Thank you for the music och den goda grabbigheten. Dixi.

KRIDDEN, my best friend, thank you for the significant other four years.

Dear family, CURT-HEINE and ANNE-MARIE and MARTIN and SANNA, thank you for being there always.

70 References

Nullum esse librum tam malum ut non aliqua parte prodesset

Pliny the Younger

1. Allen Mouse Brain Atlas. [Internet] [cited 2010-03-01]; Available from: http://www.brain-map.org/. 2. Abdin R. Parenting Stress Index (PSI) Manual. 1990, Odessa, FL: Psychological Assessment Resources, Inc. 3. Albrechet-Souza L, Borelli KG, and Brandao ML, Activity of the medial prefrontal cortex and amygdala underlies one-trial tolerance of rats in the elevated plus-maze. J Neurosci Methods, 2008. 169(1): p. 109-18. 4. Albrechet-Souza L, Oliveira AR, De Luca MC, Tomazini FM, San- tos NR, and Brandao ML, A comparative study with two types of elevated plus-maze (transparent vs. opaque walls) on the anxiolytic effects of midazolam, one-trial tolerance and fear-induced analge- sia. Prog Neuropsychopharmacol Biol Psychiatry, 2005. 29(4): p. 571-9. 5. Alexy U, Sichert-Hellert W, Kersting M, and Schultze-Pawlitschko V, Pattern of long-term fat intake and BMI during childhood and adolescence--results of the DONALD Study. Int J Obes Relat Metab Disord, 2004. 28(10): p. 1203-9. 6. Amalric M, Cline EJ, Martinez JL, Jr., Bloom FE, and Koob GF, Rewarding properties of beta-endorphin as measured by conditioned place preference. Psychopharmacology (Berl), 1987. 91(1): p. 14-9. 7. Amiaz R, Levy D, Vainiger D, Grunhaus L, and Zangen A, Re- peated high-frequency transcranial magnetic stimulation over the dorsolateral prefrontal cortex reduces cigarette craving and consumption. Addiction, 2009. 104(4): p. 653-60. 8. Anderson SE, Cohen P, Naumova EN, Jacques PF, and Must A, Adolescent obesity and risk for subsequent major depressive disorder and anxiety disorder: prospective evidence. Psychosom Med, 2007. 69(8): p. 740-7. 9. Anderson SE, Cohen P, Naumova EN, and Must A, Association of depression and anxiety disorders with weight change in a prospective community-based study of children followed up into adulthood. Arch Pediatr Adolesc Med, 2006. 160(3): p. 285-91.

71 10. APA. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. 1994, Washington DC, USA: American Psychiatric Press. 11. Arora S and Anubhuti, Role of neuropeptides in appetite regulation and obesity - A review. Neuropeptides, 2006. 40(6): p. 375-401. 12. Aston-Jones G, Smith RJ, Sartor GC, Moorman DE, Massi L, Tahsili-Fahadan P, et al., Lateral hypothalamic orexin/hypocretin neurons: A role in reward-seeking and addiction. Brain Res. 1314: p. 74-90. 13. Atkin LM and Davies PS, Diet composition and body composition in preschool children. Am J Clin Nutr, 2000. 72(1): p. 15-21. 14. Avena NM, Long KA, and Hoebel BG, Sugar-dependent rats show enhanced responding for sugar after abstinence: evidence of a sugar deprivation effect. Physiol Behav, 2005. 84(3): p. 359-62. 15. Avena NM, Rada P, and Hoebel BG, Evidence for sugar addiction: behavioral and neurochemical effects of intermittent, excessive sugar intake. Neurosci Biobehav Rev, 2008. 32(1): p. 20-39. 16. Avena NM, Rada P, and Hoebel BG, Sugar and fat bingeing have notable differences in addictive-like behavior. J Nutr, 2009. 139(3): p. 623-8. 17. Azzara AV, Bodnar RJ, Delamater AR, and Sclafani A, Naltrexone fails to block the acquisition or expression of a flavor preference conditioned by intragastric carbohydrate infusions. Pharmacol Bio- chem Behav, 2000. 67(3): p. 545-57. 18. Baier LJ and Hanson RL, Genetic studies of the etiology of type 2 diabetes in Pima Indians: hunting for pieces to a complicated puzzle. Diabetes, 2004. 53(5): p. 1181-6. 19. Baird J, Fisher D, Lucas P, Kleijnen J, Roberts H, and Law C, Being big or growing fast: systematic review of size and growth in infancy and later obesity. Bmj, 2005. 331(7522): p. 929. 20. Bakshi VP and Kelley AE, Striatal regulation of morphine-induced hyperphagia: an anatomical mapping study. Psychopharmacology (Berl), 1993. 111(2): p. 207-14. 21. Bale TL, Anderson KR, Roberts AJ, Lee KF, Nagy TR, and Vale WW, Corticotropin-releasing factor receptor-2-deficient mice display abnormal homeostatic responses to challenges of increased dietary fat and cold. Endocrinology, 2003. 144(6): p. 2580-7. 22. Barbano MF and Cador M, Various aspects of feeding behavior can be partially dissociated in the rat by the incentive properties of food and the physiological state. Behav Neurosci, 2005. 119(5): p. 1244- 53. 23. Barr MS, Fitzgerald PB, Farzan F, George TP, and Daskalakis ZJ, Transcranial magnetic stimulation to understand the pathophysiol- ogy and treatment of substance use disorders. Curr Drug Abuse Rev, 2008. 1(3): p. 328-39. 24. Barry D, Pietrzak RH, and Petry NM, Gender differences in associations between body mass index and DSM-IV mood and anxiety dis-

72 orders: results from the National Epidemiologic Survey on Alcohol and Related Conditions. Ann Epidemiol, 2008. 18(6): p. 458-66. 25. Bartlett JM and Stirling D, A short history of the polymerase chain reaction. Methods Mol Biol, 2003. 226: p. 3-6. 26. Bassareo V, De Luca MA, and Di Chiara G, Differential Expression of Motivational Stimulus Properties by Dopamine in Nucleus Ac- cumbens Shell versus Core and Prefrontal Cortex. J Neurosci, 2002. 22(11): p. 4709-19. 27. Bassareo V and Di Chiara G, Differential influence of associative and nonassociative learning mechanisms on the responsiveness of prefrontal and accumbal dopamine transmission to food stimuli in rats fed ad libitum. J Neurosci, 1997. 17(2): p. 851-61. 28. Bassareo V and Di Chiara G, Differential responsiveness of dopamine transmission to food-stimuli in nucleus accumbens shell/core compartments. Neuroscience, 1999. 89(3): p. 637-41. 29. Bassareo V and Di Chiara G, Modulation of feeding-induced activation of mesolimbic dopamine transmission by appetitive stimuli and its relation to motivational state. Eur J Neurosci, 1999. 11(12): p. 4389-97. 30. Bell SM, Stewart RB, Thompson SC, and Meisch RA, Food- deprivation increases cocaine-induced conditioned place preference and locomotor activity in rats. Psychopharmacology (Berl), 1997. 131(1): p. 1-8. 31. Bellinger L, Lilley C, and Langley-Evans SC, Prenatal exposure to a maternal low-protein diet programmes a preference for high-fat foods in the young adult rat. Br J Nutr, 2004. 92(3): p. 513-20. 32. Berridge KC, Measuring hedonic impact in animals and infants: microstructure of affective taste reactivity patterns. Neurosci Biobe- hav Rev, 2000. 24(2): p. 173-98. 33. Berridge KC and Robinson TE, What is the role of dopamine in reward: hedonic impact, reward learning, or incentive salience? Brain Res Brain Res Rev, 1998. 28(3): p. 309-69. 34. Bjorge T, Engeland A, Tverdal A, and Smith GD, Body mass index in adolescence in relation to cause-specific mortality: a follow-up of 230,000 Norwegian adolescents. Am J Epidemiol, 2008. 168(1): p. 30-7. 35. Bluher S, Ziotopoulou M, Bullen JW, Jr., Moschos SJ, Ungsunan L, Kokkotou E, et al., Responsiveness to peripherally administered melanocortins in lean and obese mice. Diabetes, 2004. 53(1): p. 82- 90. 36. Bontempi B and Sharp FR, Systemic morphine-induced Fos protein in the rat striatum and nucleus accumbens is regulated by mu opioid receptors in the substantia nigra and ventral tegmental area. J Neu- rosci, 1997. 17(21): p. 8596-612. 37. Bowirrat A and Oscar-Berman M, Relationship between dopaminergic neurotransmission, alcoholism, and Reward Deficiency syn-

73 drome. Am J Med Genet B Neuropsychiatr Genet, 2005. 132B(1): p. 29-37. 38. Bray GA and York DA, Hypothalamic and genetic obesity in experimental animals: an autonomic and endocrine hypothesis. Physiol Rev, 1979. 59(3): p. 719-809. 39. Broberger C, Brain regulation of food intake and appetite: molecules and networks. J Intern Med, 2005. 258(4): p. 301-27. 40. Burton P, Smit HJ, and Lightowler HJ, The influence of restrained and external eating patterns on overeating. Appetite, 2007. 49(1): p. 191-7. 41. Bush DE and Vaccarino FJ, Individual differences in elevated plus- maze exploration predicted progressive-ratio cocaine self- administration break points in Wistar rats. Psychopharmacology (Berl), 2007. 194(2): p. 211-9. 42. Cador M, Kelley AE, Le Moal M, and Stinus L, Ventral tegmental area infusion of substance P, neurotensin and enkephalin: differential effects on feeding behavior. Neuroscience, 1986. 18(3): p. 659- 69. 43. Cain ME, Saucier DA, and Bardo MT, Novelty seeking and drug use: contribution of an animal model. Exp Clin Psychopharmacol, 2005. 13(4): p. 367-75. 44. Camprodon JA, Martinez-Raga J, Alonso-Alonso M, Shih MC, and Pascual-Leone A, One session of high frequency repetitive transcranial magnetic stimulation (rTMS) to the right prefrontal cortex tran- siently reduces cocaine craving. Drug Alcohol Depend, 2007. 86(1): p. 91-4. 45. Carobrez AP and Bertoglio LJ, Ethological and temporal analyses of anxiety-like behavior: the elevated plus-maze model 20 years on. Neurosci Biobehav Rev, 2005. 29(8): p. 1193-205. 46. Carr KD, Augmentation of drug reward by chronic food restriction: behavioral evidence and underlying mechanisms. Physiol Behav, 2002. 76(3): p. 353-64. 47. Carr KD and Simon EJ, Effects of naloxone and its quarternary ana- logue on stimulation-induced feeding. Neuropharmacology, 1983. 22(1): p. 127-30. 48. Carroll ME, France CP, and Meisch RA, Intravenous self- administration of etonitazene, cocaine and phencyclidine in rats during food deprivation and satiation. J Pharmacol Exp Ther, 1981. 217(2): p. 241-7. 49. Carvalho MC, Albrechet-Souza L, Masson S, and Brandao ML, Changes in the biogenic amine content of the prefrontal cortex, amygdala, dorsal hippocampus, and nucleus accumbens of rats submitted to single and repeated sessions of the elevated plus-maze test. Braz J Med Biol Res, 2005. 38(12): p. 1857-66. 50. Castonguay TW, Glucocorticoids as modulators in the control of feeding. Brain Res Bull, 1991. 27(3-4): p. 423-8.

74 51. Choi DL, Davis JF, Fitzgerald ME, and Benoit SC, The role of orexin-A in food motivation, reward-based feeding behavior and food-induced neuronal activation in rats. Neuroscience. 52. Cifani C, Zanoncelli A, Tessari M, Righetti C, Di Francesco C, Cic- cocioppo R, et al., Pre-exposure to environmental cues predictive of food availability elicits hypothalamic-pituitary-adrenal axis activation and increases operant responding for food in female rats. Ad- dict Biol, 2009. 14(4): p. 397-407. 53. Clement K, Vaisse C, Lahlou N, Cabrol S, Pelloux V, Cassuto D, et al., A mutation in the human leptin receptor gene causes obesity and pituitary dysfunction. Nature, 1998. 392(6674): p. 398-401. 54. Cloninger CR, Neurogenetic adaptive mechanisms in alcoholism. Science, 1987. 236(4800): p. 410-6. 55. Cloninger CR, A systematic method for clinical description and classification of personality variants. A proposal. Arch Gen Psy- chiatry, 1987. 44(6): p. 573-88. 56. Cochrane C, Malcolm R, and Brewerton T, The role of weight control as a motivation for cocaine abuse. Addict Behav, 1998. 23(2): p. 201-7. 57. Coelho JS, Polivy J, and Herman CP, Selective carbohydrate or protein restriction: effects on subsequent food intake and cravings. Appetite, 2006. 47(3): p. 352-60. 58. Cohen IT, Sherwin BB, and Fleming AS, Food cravings, mood, and the menstrual cycle. Horm Behav, 1987. 21(4): p. 457-70. 59. Colantuoni C, Rada P, McCarthy J, Patten C, Avena NM, Cha- deayne A, et al., Evidence that intermittent, excessive sugar intake causes endogenous opioid dependence. Obes Res, 2002. 10(6): p. 478-88. 60. Cole TJ, Bellizzi MC, Flegal KM, and Dietz WH, Establishing a standard definition for child overweight and obesity worldwide: international survey. Bmj, 2000. 320(7244): p. 1240-3. 61. Cole TJ, Flegal KM, Nicholls D, and Jackson AA, Body mass index cut offs to define thinness in children and adolescents: international survey. Bmj, 2007. 335(7612): p. 194. 62. Collins S, Martin TL, Surwit RS, and Robidoux J, Genetic vulnerability to diet-induced obesity in the C57BL/6J mouse: physiological and molecular characteristics. Physiol Behav, 2004. 81(2): p. 243-8. 63. Conway KP, Kane RJ, Ball SA, Poling JC, and Rounsaville BJ, Per- sonality, substance of choice, and polysubstance involvement among substance dependent patients. Drug Alcohol Depend, 2003. 71(1): p. 65-75. 64. Cook CB, Shawar L, Thompson H, and Prasad C, Caloric intake and weight gain of rats depends on endogenous fat preference. Physiol Behav, 1997. 61(5): p. 743-8. 65. Corsica JA and Pelchat ML, Food addiction: true or false? Curr Opin Gastroenterol. 26(2): p. 165-9.

75 66. Corwin RL and Grigson PS, Symposium overview--Food addiction: fact or fiction? J Nutr, 2009. 139(3): p. 617-9. 67. Cota D, Proulx K, Smith KA, Kozma SC, Thomas G, Woods SC, et al., Hypothalamic mTOR signaling regulates food intake. Science, 2006. 312(5775): p. 927-30. 68. Cottone P, Sabino V, Nagy TR, Coscina DV, and Zorrilla EP, Feed- ing microstructure in diet-induced obesity susceptible versus resistant rats: central effects of urocortin 2. J Physiol, 2007. 583(Pt 2): p. 487-504. 69. Cruz-Morales SE, Santos NR, and Brandao ML, One-trial tolerance to midazolam is due to enhancement of fear and reduction of anxiolytic-sensitive behaviors in the elevated plus-maze retest in the rat. Pharmacol Biochem Behav, 2002. 72(4): p. 973-8. 70. Cunningham-Williams RM, Grucza RA, Cottler LB, Womack SB, Books SJ, Przybeck TR, et al., Prevalence and predictors of patho- logical gambling: results from the St. Louis personality, health and lifestyle (SLPHL) study. J Psychiatr Res, 2005. 39(4): p. 377-90. 71. Dallman MF, Pecoraro NC, and la Fleur SE, Chronic stress and comfort foods: self-medication and abdominal obesity. Brain Behav Immun, 2005. 19(4): p. 275-80. 72. Davies PS, Diet composition and body mass index in pre-school children. Eur J Clin Nutr, 1997. 51(7): p. 443-8. 73. Davis C, Levitan RD, Carter J, Kaplan AS, Reid C, Curtis C, et al., Personality and eating behaviors: a case-control study of binge eating disorder. Int J Eat Disord, 2008. 41(3): p. 243-50. 74. Davis C, Patte K, Levitan R, Reid C, Tweed S, and Curtis C, From motivation to behaviour: a model of reward sensitivity, overeating, and food preferences in the risk profile for obesity. Appetite, 2007. 48(1): p. 12-9. 75. De Schepper J, Zhou X, De Bock S, Smitz J, Louis O, Hooghe- Peters E, et al., Study of serum leptin in cafeteria-diet-overfed rats. Influence of diet, insulin and corticosterone. Horm Res, 1998. 50(5): p. 271-5. 76. De Vry J, Donselaar I, and Van Ree JM, Food deprivation and acquisition of intravenous cocaine self-administration in rats: effect of naltrexone and haloperidol. J Pharmacol Exp Ther, 1989. 251(2): p. 735-40. 77. Dellu F, Piazza PV, Mayo W, Le Moal M, and Simon H, Novelty- seeking in rats--biobehavioral characteristics and possible relationship with the sensation-seeking trait in man. Neuropsychobiology, 1996. 34(3): p. 136-45. 78. DeSousa NJ, Bush DE, and Vaccarino FJ, Self-administration of intravenous amphetamine is predicted by individual differences in sucrose feeding in rats. Psychopharmacology (Berl), 2000. 148(1): p. 52-8. 79. Di Chiara G and Imperato A, Drugs abused by humans preferentially increase synaptic dopamine concentrations in the mesolimbic

76 system of freely moving rats. Proc Natl Acad Sci U S A, 1988. 85(14): p. 5274-8. 80. Di Chiara G and Imperato A, Opposite effects of mu and kappa opiate agonists on dopamine release in the nucleus accumbens and in the dorsal caudate of freely moving rats. J Pharmacol Exp Ther, 1988. 244(3): p. 1067-80. 81. Dourmashkin JT, Chang GQ, Hill JO, Gayles EC, Fried SK, and Leibowitz SF, Model for predicting and phenotyping at normal weight the long-term propensity for obesity in Sprague-Dawley rats. Physiol Behav, 2006. 87(4): p. 666-78. 82. Downey KK, Stelson FW, Pomerleau OF, and Giordani B, Adult attention deficit hyperactivity disorder: psychological test profiles in a clinical population. J Nerv Ment Dis, 1997. 185(1): p. 32-8. 83. Dube MG, Beretta E, Dhillon H, Ueno N, Kalra PS, and Kalra SP, Central leptin gene therapy blocks high-fat diet-induced weight gain, hyperleptinemia, and hyperinsulinemia: increase in serum ghrelin levels. Diabetes, 2002. 51(6): p. 1729-36. 84. Erickson JC, Clegg KE, and Palmiter RD, Sensitivity to leptin and susceptibility to seizures of mice lacking neuropeptide Y. Nature, 1996. 381(6581): p. 415-21. 85. Esteve M, Rafecas I, Fernandez-Lopez JA, Remesar X, and Ale- many M, Fatty acid utilization by young Wistar rats fed a cafeteria diet. Mol Cell Biochem, 1992. 118(1): p. 67-74. 86. Evans KR and Vaccarino FJ, Amphetamine- and morphine-induced feeding: evidence for involvement of reward mechanisms. Neurosci Biobehav Rev, 1990. 14(1): p. 9-22. 87. Fassino S, Abbate-Daga G, Amianto F, Leombruni P, Boggio S, and Rovera GG, Temperament and character profile of eating disorders: a controlled study with the Temperament and Character Inventory. Int J Eat Disord, 2002. 32(4): p. 412-25. 88. Fassino S, Leombruni P, Piero A, Daga GA, Amianto F, Rovera G, et al., Temperament and character in obese women with and without binge eating disorder. Compr Psychiatry, 2002. 43(6): p. 431-7. 89. Figlewicz DP, Bennett JL, Naleid AM, Davis C, and Grimm JW, Intraventricular insulin and leptin decrease sucrose self- administration in rats. Physiol Behav, 2006. 89(4): p. 611-6. 90. Filip M and Frankowska M, Effects of GABA(B) receptor agents on cocaine priming, discrete contextual cue and food induced relapses. Eur J Pharmacol, 2007. 571(2-3): p. 166-73. 91. Fisher JO and Birch LL, Fat preferences and fat consumption of 3- to 5-year-old children are related to parental adiposity. J Am Diet Assoc, 1995. 95(7): p. 759-64. 92. Floresco SB, McLaughlin RJ, and Haluk DM, Opposing roles for the nucleus accumbens core and shell in cue-induced reinstatement of food-seeking behavior. Neuroscience, 2008. 154(3): p. 877-84. 93. Framme J, Dangardt F, Marild S, Osika W, Wahrborg P, and Friberg P, 24-h Systolic blood pressure and heart rate recordings in lean

77 and obese adolescents. Clin Physiol Funct Imaging, 2006. 26(4): p. 235-9. 94. Freedman DS, Khan LK, Dietz WH, Srinivasan SR, and Berenson GS, Relationship of childhood obesity to coronary heart disease risk factors in adulthood: the Bogalusa Heart Study. Pediatrics, 2001. 108(3): p. 712-8. 95. Freedman DS, Mei Z, Srinivasan SR, Berenson GS, and Dietz WH, Cardiovascular risk factors and excess adiposity among overweight children and adolescents: the Bogalusa Heart Study. J Pediatr, 2007. 150(1): p. 12-17 e2. 96. Garemo M, Lenner RA, and Strandvik B, Swedish pre-school children eat too much junk food and sucrose. Acta Paediatr, 2007. 96(2): p. 266-72. 97. Garemo M, Palsdottir V, and Strandvik B, Metabolic markers in relation to nutrition and growth in healthy 4-y-old children in Swe- den. Am J Clin Nutr, 2006. 84(5): p. 1021-6. 98. Gariepy G, Nitka D, and Schmitz N, The association between obesity and anxiety disorders in the population: a systematic review and meta-analysis. Int J Obes (Lond), 2009. 99. Geiger BM, Haburcak M, Avena NM, Moyer MC, Hoebel BG, and Pothos EN, Deficits of mesolimbic dopamine neurotransmission in rat dietary obesity. Neuroscience, 2009. 159(4): p. 1193-9. 100. Gendall KA, Joyce PR, Sullivan PF, and Bulik CM, Personality and dimensions of dietary restraint. Int J Eat Disord, 1998. 24(4): p. 371- 9. 101. Gilhooly CH, Das SK, Golden JK, McCrory MA, Dallal GE, Saltz- man E, et al., Food cravings and energy regulation: the characteristics of craved foods and their relationship with eating behaviors and weight change during 6 months of dietary energy restriction. Int J Obes (Lond), 2007. 31(12): p. 1849-58. 102. Giraudo SQ, Grace MK, Welch CC, Billington CJ, and Levine AS, Naloxone's anorectic effect is dependent upon the relative palatability of food. Pharmacol Biochem Behav, 1993. 46(4): p. 917-21. 103. Glass MJ, Billington CJ, and Levine AS, Naltrexone administered to central nucleus of amygdala or PVN: neural dissociation of diet and energy. Am J Physiol Regul Integr Comp Physiol, 2000. 279(1): p. R86-92. 104. Glass MJ, Grace M, Cleary JP, Billington CJ, and Levine AS, Po- tency of naloxone's anorectic effect in rats is dependent on diet preference. Am J Physiol, 1996. 271(1 Pt 2): p. R217-21. 105. Gosnell BA, Sucrose intake predicts rate of acquisition of cocaine self-administration. Psychopharmacology (Berl), 2000. 149(3): p. 286-92. 106. Gosnell BA, Krahn DD, and Majchrzak MJ, The effects of morphine on diet selection are dependent upon baseline diet preferences. Pharmacol Biochem Behav, 1990. 37(2): p. 207-12.

78 107. Greeno CG, Wing RR, and Shiffman S, Binge antecedents in obese women with and without binge eating disorder. J Consult Clin Psy- chol, 2000. 68(1): p. 95-102. 108. Grill HJ and Norgren R, The taste reactivity test. I. Mimetic responses to gustatory stimuli in neurologically normal rats. Brain Res, 1978. 143(2): p. 263-79. 109. Grucza RA, Cloninger CR, Bucholz KK, Constantino JN, Schuckit MA, Dick DM, et al., Novelty Seeeking as a Moderator of Familial Risk for Alcohol Dependence. Alcohol Clin Exp Res, 2006. 30(7): p. 1176-1183. 110. Grucza RA, Przybeck TR, and Cloninger CR, Prevalence and correlates of binge eating disorder in a community sample. Compr Psy- chiatry, 2007. 48(2): p. 124-31. 111. Guillaume M, Lapidus L, and Lambert A, Obesity and nutrition in children. The Belgian Luxembourg Child Study IV. Eur J Clin Nutr, 1998. 52(5): p. 323-8. 112. Hajnal A, De Jonghe BC, and Covasa M, Dopamine D2 receptors contribute to increased avidity for sucrose in obese rats lacking CCK-1 receptors. Neuroscience, 2007. 148(2): p. 584-92. 113. Hajnal A, Margas WM, and Covasa M, Altered dopamine D2 receptor function and binding in obese OLETF rat. Brain Res Bull, 2008. 75(1): p. 70-6. 114. Harvey J, Wing RR, and Mullen M, Effects on food cravings of a very low calorie diet or a balanced, low calorie diet. Appetite, 1993. 21(2): p. 105-15. 115. Hassapidou M, Fotiadou E, Maglara E, and Papadopoulou SK, En- ergy intake, diet composition, energy expenditure, and body fatness of adolescents in northern Greece. Obesity (Silver Spring), 2006. 14(5): p. 855-62. 116. Hessing MJ, Hagelso AM, Schouten WG, Wiepkema PR, and van Beek JA, Individual behavioral and physiological strategies in pigs. Physiol Behav, 1994. 55(1): p. 39-46. 117. Heyne A, Kiesselbach C, Sahun I, McDonald J, Gaiffi M, Dierssen M, et al., An animal model of compulsive food-taking behaviour. Addict Biol, 2009. 14(4): p. 373-83. 118. Hill AJ, The psychology of food craving. Proc Nutr Soc, 2007. 66(2): p. 277-85. 119. Hill AJ, Weaver CF, and Blundell JE, Food craving, dietary restraint and mood. Appetite, 1991. 17(3): p. 187-97. 120. Hollerman JR and Schultz W, Dopamine neurons report an error in the temporal prediction of reward during learning. Nat Neurosci, 1998. 1(4): p. 304-9. 121. Hommer DW, Functional imaging of craving. Alcohol Res Health, 1999. 23(3): p. 187-96. 122. Hormes JM and Rozin P, Perimenstrual chocolate craving. What happens after menopause? Appetite, 2009. 53(2): p. 256-9.

79 123. Inoue K, Valdez GR, Reyes TM, Reinhardt LE, Tabarin A, Rivier J, et al., Human urocortin II, a selective agonist for the type 2 corticotropin-releasing factor receptor, decreases feeding and drinking in the rat. J Pharmacol Exp Ther, 2003. 305(1): p. 385-93. 124. Johnson SW and North RA, Opioids excite dopamine neurons by hyperpolarization of local interneurons. J Neurosci, 1992. 12(2): p. 483-8. 125. Jorm AF, Korten AE, Christensen H, Jacomb PA, Rodgers B, and Parslow RA, Association of obesity with anxiety, depression and emotional well-being: a community survey. Aust N Z J Public Health, 2003. 27(4): p. 434-40. 126. Karhunen LJ, Lappalainen RI, Vanninen EJ, Kuikka JT, and Uusi- tupa MI, Regional cerebral blood flow during food exposure in obese and normal-weight women. Brain, 1997. 120 ( Pt 9): p. 1675- 84. 127. Kelley AE, Ventral striatal control of appetitive motivation: role in ingestive behavior and reward-related learning. Neurosci Biobehav Rev, 2004. 27(8): p. 765-76. 128. Keskitalo K, Knaapila A, Kallela M, Palotie A, Wessman M, Sam- malisto S, et al., Sweet taste preferences are partly genetically determined: identification of a trait locus on chromosome 16. Am J Clin Nutr, 2007. 86(1): p. 55-63. 129. Kirkham TC and Cooper SJ, Naloxone attenuation of sham feeding is modified by manipulation of sucrose concentration. Physiol Be- hav, 1988. 44(4-5): p. 491-4. 130. Klebaur JE, Bevins RA, Segar TM, and Bardo MT, Individual differences in behavioral responses to novelty and amphetamine self- administration in male and female rats. Behav Pharmacol, 2001. 12(4): p. 267-75. 131. Koch FS, Sepa A, and Ludvigsson J, Psychological stress and obesity. J Pediatr, 2008. 153(6): p. 839-44. 132. Komatsu S, Rice and sushi cravings: a preliminary study of food craving among Japanese females. Appetite, 2008. 50(2-3): p. 353-8. 133. Koob GF and Bloom FE, Cellular and molecular mechanisms of drug dependence. Science, 1988. 242(4879): p. 715-23. 134. Koob GF and Le Moal M, Drug addiction, dysregulation of reward, and allostasis. Neuropsychopharmacology, 2001. 24(2): p. 97-129. 135. Krude H, Biebermann H, and Gruters A, Mutations in the human proopiomelanocortin gene. Ann N Y Acad Sci, 2003. 994: p. 233-9. 136. Krude H, Biebermann H, Luck W, Horn R, Brabant G, and Gruters A, Severe early-onset obesity, adrenal insufficiency and red hair pigmentation caused by POMC mutations in humans. Nat Genet, 1998. 19(2): p. 155-7. 137. la Fleur SE, Akana SF, Manalo SL, and Dallman MF, Interaction between corticosterone and insulin in obesity: regulation of lard intake and fat stores. Endocrinology, 2004. 145(5): p. 2174-85.

80 138. la Fleur SE, van Rozen AJ, Luijendijk MC, Groeneweg F, and Adan RA, A free-choice high-fat high-sugar diet induces changes in arcuate neuropeptide expression that support hyperphagia. Int J Obes (Lond), 2009. 139. la Fleur SE, Vanderschuren LJ, Luijendijk MC, Kloeze BM, Ties- jema B, and Adan RA, A reciprocal interaction between food- motivated behavior and diet-induced obesity. Int J Obes (Lond), 2007. 31(8): p. 1286-94. 140. Lamb RJ, Preston KL, Schindler CW, Meisch RA, Davis F, Katz JL, et al., The reinforcing and subjective effects of morphine in post- addicts: a dose-response study. J Pharmacol Exp Ther, 1991. 259(3): p. 1165-73. 141. Leggio L, Understanding and treating alcohol craving and dependence: recent pharmacological and neuroendocrinological findings. Alcohol Alcohol, 2009. 44(4): p. 341-52. 142. Lenard NR, Zheng H, and Berthoud HR, Chronic suppression of mu-opioid receptor signaling in the nucleus accumbens attenuates development of diet-induced obesity in rats. Int J Obes (Lond). 143. Levin BE and Dunn-Meynell AA, Differential effects of exercise on body weight gain and adiposity in obesity-prone and -resistant rats. Int J Obes (Lond), 2006. 30(4): p. 722-7. 144. Levin BE, Dunn-Meynell AA, Balkan B, and Keesey RE, Selective breeding for diet-induced obesity and resistance in Sprague-Dawley rats. Am J Physiol, 1997. 273(2 Pt 2): p. R725-30. 145. Levin BE and Keesey RE, Defense of differing body weight set points in diet-induced obese and resistant rats. Am J Physiol, 1998. 274(2 Pt 2): p. R412-9. 146. Levine AS, Murray SS, Kneip J, Grace M, and Morley JE, Flavor enhances the antidipsogenic effect of naloxone. Physiol Behav, 1982. 28(1): p. 23-5. 147. Levine AS, Weldon DT, Grace M, Cleary JP, and Billington CJ, Naloxone blocks that portion of feeding driven by sweet taste in food-restricted rats. Am J Physiol, 1995. 268(1 Pt 2): p. R248-52. 148. Levine MD and Marcus MD, Eating behavior following stress in women with and without bulimic symptoms. Ann Behav Med, 1997. 19(2): p. 132-8. 149. Lissner L, Sohlstrom A, Sundblom E, and Sjoberg A, Trends in overweight and obesity in Swedish schoolchildren 1999-2005: has the epidemic reached a plateau? Obes Rev, 2009. 150. Lofgren M, Johansson IM, Meyerson B, Lundgren P, and Backstrom T, Progesterone withdrawal effects in the open field test can be predicted by elevated plus maze performance. Horm Behav, 2006. 50(2): p. 208-15. 151. Lofgren M, Johansson IM, Meyerson B, Turkmen S, and Backstrom T, Withdrawal effects from progesterone and estradiol relate to individual risk-taking and explorative behavior in female rats. Physiol Behav, 2009. 96(1): p. 91-7.

81 152. Lynch WC, Opiate blockade inhibits saccharin intake and blocks normal preference acquisition. Pharmacol Biochem Behav, 1986. 24(4): p. 833-6. 153. Maes HH, Neale MC, and Eaves LJ, Genetic and environmental factors in relative body weight and human adiposity. Behav Genet, 1997. 27(4): p. 325-51. 154. Maffeis C, Pinelli L, and Schutz Y, Fat intake and adiposity in 8 to 11-year-old obese children. Int J Obes Relat Metab Disord, 1996. 20(2): p. 170-4. 155. Magarey AM, Daniels LA, Boulton TJ, and Cockington RA, Does fat intake predict adiposity in healthy children and adolescents aged 2--15 y? A longitudinal analysis. Eur J Clin Nutr, 2001. 55(6): p. 471-81. 156. Maillard G, Charles MA, Lafay L, Thibult N, Vray M, Borys JM, et al., Macronutrient energy intake and adiposity in non obese prepu- bertal children aged 5-11 y (the Fleurbaix Laventie Ville Sante Study). Int J Obes Relat Metab Disord, 2000. 24(12): p. 1608-17. 157. Majeed NH, Przewlocka B, Wedzony K, and Przewlocki R, Stimula- tion of food intake following opioid microinjection into the nucleus accumbens septi in rats. Peptides, 1986. 7(5): p. 711-6. 158. Marild S, Bondestam M, Bergstrom R, Ehnberg S, Hollsing A, and Albertsson-Wikland K, Prevalence trends of obesity and overweight among 10-year-old children in western Sweden and relationship with parental body mass index. Acta Paediatr, 2004. 93(12): p. 1588-95. 159. Markou A, Weiss F, Gold LH, Caine SB, Schulteis G, and Koob GF, Animal models of drug craving. Psychopharmacology (Berl), 1993. 112(2-3): p. 163-82. 160. Marks-Kaufman R, Increased fat consumption induced by morphine administration in rats. Pharmacol Biochem Behav, 1982. 16(6): p. 949-55. 161. Marks-Kaufman R and Kanarek RB, Morphine selectively influences macronutrient intake in the rat. Pharmacol Biochem Behav, 1980. 12(3): p. 427-30. 162. Marks-Kaufman R and Kanarek RB, Modifications of nutrient selection induced by naloxone in rats. Psychopharmacology (Berl), 1981. 74(4): p. 321-4. 163. Marks-Kaufman R and Kanarek RB, Diet selection following a chronic morphine and naloxone regimen. Pharmacol Biochem Be- hav, 1990. 35(3): p. 665-9. 164. Martin CK, O'Neil PM, and Pawlow L, Changes in food cravings during low-calorie and very-low-calorie diets. Obesity (Silver Spring), 2006. 14(1): p. 115-21. 165. Martin CK, O'Neil PM, Tollefson G, Greenway FL, and White MA, The association between food cravings and consumption of specific foods in a laboratory taste test. Appetite, 2008. 51(2): p. 324-6.

82 166. Mattsson N, Ronnemaa T, Juonala M, Viikari JS, and Raitakari OT, Childhood predictors of the metabolic syndrome in adulthood. The Cardiovascular Risk in Young Finns Study. Ann Med, 2008. 40(7): p. 542-52. 167. Mela DJ and Sacchetti DA, Sensory preferences for fats: relationships with diet and body composition. Am J Clin Nutr, 1991. 53(4): p. 908-15. 168. Meyerson BJ, Augustsson H, Berg M, and Roman E, The Concen- tric Square Field: a multivariate test arena for analysis of explorative strategies. Behav Brain Res, 2006. 168(1): p. 100-13. 169. Miedlar JA, Rinaman L, Vollmer RR, and Amico JA, Oxytocin gene deletion mice overconsume palatable sucrose solution but not palatable lipid emulsions. Am J Physiol Regul Integr Comp Physiol, 2007. 293(3): p. R1063-8. 170. Mitchell RT, McDougall CM, and Crum JE, Decreasing prevalence of obesity in primary schoolchildren. Arch Dis Child, 2007. 92(2): p. 153-4. 171. Montague CT, Farooqi IS, Whitehead JP, Soos MA, Rau H, Ware- ham NJ, et al., Congenital leptin deficiency is associated with severe early-onset obesity in humans. Nature, 1997. 387(6636): p. 903-8. 172. Moreira T, Cebers G, Cebere A, Wagner A, and Liljequist S, Ex- tradural compression of the sensorimotor cortex delays the acquisition but not the recalling of a lever-pressing task in Wistar rats. Be- hav Brain Res, 2005. 164(2): p. 250-65. 173. Nadal R, Armario A, and Janak PH, Positive relationship between activity in a novel environment and operant ethanol self- administration in rats. Psychopharmacology (Berl), 2002. 162(3): p. 333-8. 174. Naleid AM, Grace MK, Chimukangara M, Billington CJ, and Levine AS, Paraventricular opioids alter intake of high-fat but not high- sucrose diet depending on diet preference in a binge model of feeding. Am J Physiol Regul Integr Comp Physiol, 2007. 293(1): p. R99- 105. 175. NCHS. Prevalence of Overweight and Obesity Among Adults: United States, 2003-2004. 2004 [cited 2008 2008-11-21]; Available from: http://www.cdc.gov/nchs/products/pubs/pubd/hestats/overweight/ov erwght_adult_03.htm. 176. Neel JV, Diabetes mellitus: a "thrifty" genotype rendered detrimen- tal by "progress"? Am J Hum Genet, 1962. 14: p. 353-62. 177. Neovius M, Janson A, and Rossner S, Prevalence of obesity in Swe- den. Obes Rev, 2006. 7(1): p. 1-3. 178. Neovius M and Rasmussen F, Place of residence and obesity in 1,578,694 young Swedish men between 1969 and 2005. Obesity (Silver Spring), 2008. 16(3): p. 671-6.

83 179. Neovius M, Teixeira-Pinto A, and Rasmussen F, Shift in the composition of obesity in young adult men in Sweden over a third of a cen- tury. Int J Obes (Lond), 2008. 32(5): p. 832-6. 180. Noble EP, Addiction and its reward process through polymorphisms of the D2 dopamine receptor gene: a review. Eur Psychiatry, 2000. 15(2): p. 79-89. 181. Noel MB and Wise RA, Ventral tegmental injections of morphine but not U-50,488H enhance feeding in food-deprived rats. Brain Res, 1993. 632(1-2): p. 68-73. 182. Noel MB and Wise RA, Ventral tegmental injections of a selective mu or delta opioid enhance feeding in food-deprived rats. Brain Res, 1995. 673(2): p. 304-12. 183. O'Doherty JP, Deichmann R, Critchley HD, and Dolan RJ, Neural responses during anticipation of a primary taste reward. Neuron, 2002. 33(5): p. 815-26. 184. Odorizzi M, Max JP, Tankosic P, Burlet C, and Burlet A, Dietary preferences of Brattleboro rats correlated with an overexpression of galanin in the hypothalamus. Eur J Neurosci, 1999. 11(9): p. 3005- 14. 185. Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, and Flegal KM, Prevalence of overweight and obesity in the United States, 1999-2004. Jama, 2006. 295(13): p. 1549-55. 186. Ohata H and Shibasaki T, Effects of urocortin 2 and 3 on motor activity and food intake in rats. Peptides, 2004. 25(10): p. 1703-9. 187. Olszewski PK, Cedernaes J, Olsson F, Levine AS, and Schioth HB, Analysis of the network of feeding neuroregulators using the Allen Brain Atlas. Neurosci Biobehav Rev, 2008. 32(5): p. 945-56. 188. Olszewski PK, Fredriksson R, Olszewska AM, Stephansson O, Al- sio J, Radomska KJ, et al., Hypothalamic FTO is associated with the regulation of energy intake not feeding reward. BMC Neurosci, 2009. 10: p. 129. 189. Olszewski PK, Grace MK, Sanders JB, Billington CJ, and Levine AS, Effect of nociceptin/orphanin FQ on food intake in rats that differ in diet preference. Pharmacol Biochem Behav, 2002. 73(3): p. 529-35. 190. Ostberg M, Parental stress, psychosocial problems and responsiveness in help-seeking parents with small (2-45 months old) children. Acta Paediatr, 1998. 87(1): p. 69-76. 191. Ostberg M, Hagekull B, and Wettergren S, A measure of parental stress in mothers with small children: dimensionality, stability and validity. Scand J Psychol, 1997. 38(3): p. 199-208. 192. Parker LA, Maier S, Rennie M, and Crebolder J, Morphine- and naltrexone-induced modification of palatability: analysis by the taste reactivity test. Behav Neurosci, 1992. 106(6): p. 999-1010. 193. Paxinos G and Franklin K. The Mouse Brain in Stereotaxic Coordi- nates. 2nd ed. 2004, London: Academic Press.

84 194. Pecina S and Berridge KC, Opioid site in nucleus accumbens shell mediates eating and hedonic 'liking' for food: map based on microinjection Fos plumes. Brain Res, 2000. 863(1-2): p. 71-86. 195. Pecina S and Berridge KC, Hedonic hot spot in nucleus accumbens shell: where do mu-opioids cause increased hedonic impact of sweetness? J Neurosci, 2005. 25(50): p. 11777-86. 196. Pecina S, Cagniard B, Berridge KC, Aldridge JW, and Zhuang X, Hyperdopaminergic mutant mice have higher "wanting" but not "liking" for sweet rewards. J Neurosci, 2003. 23(28): p. 9395-402. 197. Pecoraro N, Reyes F, Gomez F, Bhargava A, and Dallman MF, Chronic stress promotes palatable feeding, which reduces signs of stress: feedforward and feedback effects of chronic stress. Endocri- nology, 2004. 145(8): p. 3754-62. 198. Pelchat ML, Johnson A, Chan R, Valdez J, and Ragland JD, Images of desire: food-craving activation during fMRI. Neuroimage, 2004. 23(4): p. 1486-93. 199. Pellow S, Chopin P, File SE, and Briley M, Validation of open:closed arm entries in an elevated plus-maze as a measure of anxiety in the rat. J Neurosci Methods, 1985. 14(3): p. 149-67. 200. Petry NM, Barry D, Pietrzak RH, and Wagner JA, Overweight and obesity are associated with psychiatric disorders: results from the National Epidemiologic Survey on Alcohol and Related Conditions. Psychosom Med, 2008. 70(3): p. 288-97. 201. Piazza PV, Deminiere JM, Le Moal M, and Simon H, Factors that predict individual vulnerability to amphetamine self-administration. Science, 1989. 245(4925): p. 1511-3. 202. Pickering C, Avesson L, Liljequist S, Lindblom J, and Schioth HB, The role of hypothalamic peptide gene expression in alcohol self- administration behavior. Peptides, 2007. 28(12): p. 2361-71. 203. Pickering C, Avesson L, Lindblom J, Liljequist S, and Schioth HB, Identification of neurotransmitter receptor genes involved in alcohol self-administration in the rat prefrontal cortex, hippocampus and amygdala. Prog Neuropsychopharmacol Biol Psychiatry, 2007. 31(1): p. 53-64. 204. Pickering C, Moreira T, and Liljequist S, Delayed access to alcohol accelerates self-administration of alcohol on a progressive ratio schedule. Basic Clin Pharmacol Toxicol, 2007. 100(2): p. 109-14. 205. Pierce RC and Vanderschuren LJ, Kicking the habit: The neural basis of ingrained behaviors in cocaine addiction. Neurosci Biobe- hav Rev. 206. Polivy J, Coleman J, and Herman CP, The effect of deprivation on food cravings and eating behavior in restrained and unrestrained eaters. Int J Eat Disord, 2005. 38(4): p. 301-9. 207. Pothos E, Rada P, Mark GP, and Hoebel BG, Dopamine microdialy- sis in the nucleus accumbens during acute and chronic morphine, naloxone-precipitated withdrawal and clonidine treatment. Brain Res, 1991. 566(1-2): p. 348-50.

85 208. Prasad A, Henry M, and Prasad C, Heterogeneity in the performance of outbred Sprague-Dawley rats in an elevated-plus maze test: a possible animal model for anxiety disorder. Life Sci, 1996. 59(18): p. 1499-506. 209. Prasad C, delaHoussaye AJ, Prasad A, and Mizuma H, Augmenta- tion of dietary fat preference by chronic, but not acute, hypercorti- costeronemia. Life Sci, 1995. 56(16): p. 1361-71. 210. Prut L and Belzung C, The open field as a paradigm to measure the effects of drugs on anxiety-like behaviors: a review. Eur J Pharma- col, 2003. 463(1-3): p. 3-33. 211. Qian S, Chen H, Weingarth D, Trumbauer ME, Novi DE, Guan X, et al., Neither agouti-related protein nor neuropeptide Y is critically required for the regulation of energy homeostasis in mice. Mol Cell Biol, 2002. 22(14): p. 5027-35. 212. Rada P, Avena NM, and Hoebel BG, Daily bingeing on sugar re- peatedly releases dopamine in the accumbens shell. Neuroscience, 2005. 134(3): p. 737-44. 213. Raitakari OT, Juonala M, and Viikari JS, Obesity in childhood and vascular changes in adulthood: insights into the Cardiovascular Risk in Young Finns Study. Int J Obes (Lond), 2005. 29 Suppl 2: p. S101-4. 214. Reed DR, Bachmanov AA, Beauchamp GK, Tordoff MG, and Price RA, Heritable variation in food preferences and their contribution to obesity. Behav Genet, 1997. 27(4): p. 373-87. 215. Reyes TM, Lewis K, Perrin MH, Kunitake KS, Vaughan J, Arias CA, et al., Urocortin II: a member of the corticotropin-releasing factor (CRF) neuropeptide family that is selectively bound by type 2 CRF receptors. Proc Natl Acad Sci U S A, 2001. 98(5): p. 2843-8. 216. Richardson NR and Roberts DC, Progressive ratio schedules in drug self-administration studies in rats: a method to evaluate reinforcing efficacy. J Neurosci Methods, 1996. 66(1): p. 1-11. 217. Ricketts CD, Fat preferences, dietary fat intake and body composition in children. Eur J Clin Nutr, 1997. 51(11): p. 778-81. 218. Rissanen A, Hakala P, Lissner L, Mattlar CE, Koskenvuo M, and Ronnemaa T, Acquired preference especially for dietary fat and obesity: a study of weight-discordant monozygotic twin pairs. Int J Obes Relat Metab Disord, 2002. 26(7): p. 973-7. 219. Robinson TE and Berridge KC, The neural basis of drug craving: an incentive-sensitization theory of addiction. Brain Res Brain Res Rev, 1993. 18(3): p. 247-91. 220. Robinson TE and Berridge KC, Incentive-sensitization and addiction. Addiction, 2001. 96(1): p. 103-14. 221. Rockwood GA and Reid LD, Naloxone modifies sugar-water intake in rats drinking with open gastric fistulas. Physiol Behav, 1982. 29(6): p. 1175-8.

86 222. Rodgers RJ and Johnson NJ, Factor analysis of spatiotemporal and ethological measures in the murine elevated plus-maze test of anxiety. Pharmacol Biochem Behav, 1995. 52(2): p. 297-303. 223. Roman E and Colombo G, Lower risk taking and exploratory behavior in alcohol-preferring sP rats than in alcohol non-preferring sNP rats in the multivariate concentric square field (MCSF) test. Behav Brain Res, 2009. 205(1): p. 249-58. 224. Roman E, Gustafsson L, Berg M, and Nylander I, Behavioral profiles and stress-induced corticosteroid secretion in male Wistar rats subjected to short and prolonged periods of maternal separation. Horm Behav, 2006. 50(5): p. 736-47. 225. Roman E, Meyerson BJ, Hyytia P, and Nylander I, The multivariate concentric square field test reveals different behavioural profiles in male AA and ANA rats with regard to risk taking and environmental reactivity. Behav Brain Res, 2007. 183(2): p. 195-205. 226. Rothemund Y, Preuschhof C, Bohner G, Bauknecht HC, Klingebiel R, Flor H, et al., Differential activation of the dorsal striatum by high-calorie visual food stimuli in obese individuals. Neuroimage, 2007. 37(2): p. 410-21. 227. Rozin P, Levine E, and Stoess C, Chocolate craving and liking. Ap- petite, 1991. 17(3): p. 199-212. 228. Salamone JD, Wisniecki A, Carlson BB, and Correa M, Nucleus accumbens dopamine depletions make animals highly sensitive to high fixed ratio requirements but do not impair primary food reinforcement. Neuroscience, 2001. 105(4): p. 863-70. 229. SBU, Fetma – problem och åtgärder. 2002, The Swedish Council on Technology Assessment in Health Care. 230. SCB, Undersökningarna av levnadsförhållanden (ULF). 2007. 231. Schemmel R, Mickelsen O, and Gill JL, Dietary obesity in rats: Body weight and body fat accretion in seven strains of rats. J Nutr, 1970. 100(9): p. 1041-8. 232. Schultz W, Dayan P, and Montague PR, A neural substrate of prediction and reward. Science, 1997. 275(5306): p. 1593-9. 233. Schultz W, Tremblay L, and Hollerman JR, Reward processing in primate orbitofrontal cortex and basal ganglia. Cereb Cortex, 2000. 10(3): p. 272-84. 234. Schwartz MW, Woods SC, Porte D, Jr., Seeley RJ, and Baskin DG, Central nervous system control of food intake. Nature, 2000. 404(6778): p. 661-71. 235. Sclafani A, Sucrose motivation in sweet "sensitive" (C57BL/6J) and "subsensitive" (129P3/J) mice measured by progressive ratio licking. Physiol Behav, 2006. 87(4): p. 734-44. 236. Sclafani A and Ackroff K, Reinforcement value of sucrose measured by progressive ratio operant licking in the rat. Physiol Behav, 2003. 79(4-5): p. 663-70. 237. Sclafani A, Rinaman L, Vollmer RR, and Amico JA, Oxytocin knockout mice demonstrate enhanced intake of sweet and nonsweet

87 carbohydrate solutions. Am J Physiol Regul Integr Comp Physiol, 2007. 292(5): p. R1828-33. 238. Sclafani A and Springer D, Dietary obesity in adult rats: similarities to hypothalamic and human obesity syndromes. Physiol Behav, 1976. 17(3): p. 461-71. 239. Scott KM, Bruffaerts R, Simon GE, Alonso J, Angermeyer M, de Girolamo G, et al., Obesity and mental disorders in the general population: results from the world mental health surveys. Int J Obes (Lond), 2008. 32(1): p. 192-200. 240. Scott KM, McGee MA, Wells JE, and Oakley Browne MA, Obesity and mental disorders in the adult general population. J Psychosom Res, 2008. 64(1): p. 97-105. 241. Shalev U, Finnie PS, Quinn T, Tobin S, and Wahi P, A role for corticotropin-releasing factor, but not corticosterone, in acute food- deprivation-induced reinstatement of heroin seeking in rats. Psy- chopharmacology (Berl), 2006. 187(3): p. 376-84. 242. Shimbara T, Mondal MS, Kawagoe T, Toshinai K, Koda S, Yama- guchi H, et al., Central administration of ghrelin preferentially enhances fat ingestion. Neurosci Lett, 2004. 369(1): p. 75-9. 243. Shor-Posner G, Brennan G, Ian C, Jasaitis R, Madhu K, and Lei- bowitz SF, Meal patterns of macronutrient intake in rats with particular dietary preferences. Am J Physiol, 1994. 266(4 Pt 2): p. R1395-402. 244. Shor-Posner G, Ian C, Brennan G, Cohn T, Moy H, Ning A, et al., Self-selecting albino rats exhibit differential preferences for pure macronutrient diets: characterization of three subpopulations. Physiol Behav, 1991. 50(6): p. 1187-95. 245. Sinha R, Fisch G, Teague B, Tamborlane WV, Banyas B, Allen K, et al., Prevalence of impaired glucose tolerance among children and adolescents with marked obesity. N Engl J Med, 2002. 346(11): p. 802-10. 246. Skinner BF. The Behavior of Organisms: An Experimental Analysis. 1938: B F Skinner Foundation. 247. Small DM, Jones-Gotman M, and Dagher A, Feeding-induced dopamine release in dorsal striatum correlates with meal pleasantness ratings in healthy human volunteers. Neuroimage, 2003. 19(4): p. 1709-15. 248. Small DM, Zatorre RJ, Dagher A, Evans AC, and Jones-Gotman M, Changes in brain activity related to eating chocolate: from pleasure to aversion. Brain, 2001. 124(Pt 9): p. 1720-33. 249. Smith KS and Berridge KC, The ventral pallidum and hedonic reward: neurochemical maps of sucrose "liking" and food intake. J Neurosci, 2005. 25(38): p. 8637-49. 250. Solinas M and Goldberg SR, Motivational effects of cannabinoids and opioids on food reinforcement depend on simultaneous activation of cannabinoid and opioid systems. Neuropsychopharmacology, 2005. 30(11): p. 2035-45.

88 251. Southam L, Soranzo N, Montgomery SB, Frayling TM, McCarthy MI, Barroso I, et al., Is the thrifty genotype hypothesis supported by evidence based on confirmed type 2 diabetes- and obesity- susceptibility variants? Diabetologia, 2009. 52(9): p. 1846-51. 252. Spanagel R, Herz A, and Shippenberg TS, The effects of opioid peptides on dopamine release in the nucleus accumbens: an in vivo mi- crodialysis study. J Neurochem, 1990. 55(5): p. 1734-40. 253. Spanagel R, Herz A, and Shippenberg TS, Opposing tonically active endogenous opioid systems modulate the mesolimbic dopaminergic pathway. Proc Natl Acad Sci U S A, 1992. 89(6): p. 2046-50. 254. Speakman JR, Thrifty genes for obesity, an attractive but flawed idea, and an alternative perspective: the 'drifty gene' hypothesis. Int J Obes (Lond), 2008. 32(11): p. 1611-7. 255. Stafford D, LeSage MG, and Glowa JR, Progressive-ratio schedules of drug delivery in the analysis of drug self-administration: a review. Psychopharmacology (Berl), 1998. 139(3): p. 169-84. 256. Steiner JE, Glaser D, Hawilo ME, and Berridge KC, Comparative expression of hedonic impact: affective reactions to taste by human infants and other primates. Neurosci Biobehav Rev, 2001. 25(1): p. 53-74. 257. Stenhammar C, Olsson G, Edlund B, Bahmanyar S, Hulting A, Wet- tergren B, et al., Family stress and BMI in young children. Acta Paediatr, 2010. In press 258. Stice E, Spoor S, Bohon C, and Small DM, Relation between obesity and blunted striatal response to food is moderated by TaqIA A1 al- lele. Science, 2008. 322(5900): p. 449-52. 259. Stice E, Spoor S, Bohon C, Veldhuizen MG, and Small DM, Rela- tion of reward from food intake and anticipated food intake to obesity: a functional magnetic resonance imaging study. J Abnorm Psy- chol, 2008. 117(4): p. 924-35. 260. Stice E, Spoor S, Ng J, and Zald DH, Relation of obesity to consummatory and anticipatory food reward. Physiol Behav, 2009. 97(5): p. 551-60. 261. Sullivan S, Cloninger CR, Przybeck TR, and Klein S, Personality characteristics in obesity and relationship with successful weight loss. Int J Obes (Lond), 2007. 31(4): p. 669-74. 262. Suto N, Austin JD, and Vezina P, Locomotor response to novelty predicts a rat's propensity to self-administer nicotine. Psychophar- macology (Berl), 2001. 158(2): p. 175-80. 263. Tanda G and Di Chiara G, A dopamine-mu1 opioid link in the rat ventral tegmentum shared by palatable food (Fonzies) and non- psychostimulant drugs of abuse. Eur J Neurosci, 1998. 10(3): p. 1179-87. 264. Tempel DL, Leibowitz KJ, and Leibowitz SF, Effects of PVN galanin on macronutrient selection. Peptides, 1988. 9(2): p. 309-14. 265. Tiggemann M and Kemps E, The phenomenology of food cravings: the role of mental imagery. Appetite, 2005. 45(3): p. 305-13.

89 266. Turkish S and Cooper SJ, Fluid consumption in water-deprived rats after administration of naloxone or quaternary naloxone. Prog Neu- ropsychopharmacol Biol Psychiatry, 1983. 7(4-6): p. 835-9. 267. Uher R, Yoganathan D, Mogg A, Eranti SV, Treasure J, Campbell IC, et al., Effect of left prefrontal repetitive transcranial magnetic stimulation on food craving. Biol Psychiatry, 2005. 58(10): p. 840-2. 268. Wake M, Hardy P, Canterford L, Sawyer M, and Carlin JB, Over- weight, obesity and girth of Australian preschoolers: prevalence and socio-economic correlates. Int J Obes (Lond), 2007. 31(7): p. 1044- 51. 269. Van den Eynde F, Claudino AM, Mogg A, Horrell L, Stahl D, Ribeiro W, et al., Repetitive Transcranial Magnetic Stimulation Re- duces Cue-Induced Food Craving in Bulimic Disorders. Biol Psy- chiatry. 270. Van Heek M, Compton DS, France CF, Tedesco RP, Fawzi AB, Graziano MP, et al., Diet-induced obese mice develop peripheral, but not central, resistance to leptin. J Clin Invest, 1997. 99(3): p. 385-90. 271. van Raaij MT, Pit DS, Balm PH, Steffens AB, and van den Thillart GE, Behavioral strategy and the physiological stress response in rainbow trout exposed to severe hypoxia. Horm Behav, 1996. 30(1): p. 85-92. 272. van Ree JM, Smyth DG, and Colpaert FC, Dependence creating properties of lipotropin C-fragment (beta-endorphin): evidence for its internal control of behavior. Life Sci, 1979. 24(6): p. 495-502. 273. Vanderschuren LJ and Everitt BJ, Behavioral and neural mechanisms of compulsive drug seeking. Eur J Pharmacol, 2005. 526(1-3): p. 77-88. 274. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, et al., Accurate normalization of real-time quantitative RT- PCR data by geometric averaging of multiple internal control genes. Genome Biol, 2002. 3(7): p. RESEARCH0034. 275. Wang GJ, Volkow ND, Fowler JS, Cervany P, Hitzemann RJ, Pappas NR, et al., Regional brain metabolic activation during craving elicited by recall of previous drug experiences. Life Sci, 1999. 64(9): p. 775-84. 276. Wang GJ, Volkow ND, Logan J, Pappas NR, Wong CT, Zhu W, et al., Brain dopamine and obesity. Lancet, 2001. 357(9253): p. 354-7. 277. Wang GJ, Volkow ND, Thanos PK, and Fowler JS, Similarity between obesity and drug addiction as assessed by neurofunctional imaging: a concept review. J Addict Dis, 2004. 23(3): p. 39-53. 278. Wang J, Alexander JT, Zheng P, Yu HJ, Dourmashkin J, and Lei- bowitz SF, Behavioral and endocrine traits of obesity-prone and obesity-resistant rats on macronutrient diets. Am J Physiol, 1998. 274(6 Pt 1): p. E1057-66. 279. Wang Y and Lobstein T, Worldwide trends in childhood overweight and obesity. Int J Pediatr Obes, 2006. 1(1): p. 11-25.

90 280. Ward SJ, Walker EA, and Dykstra LA, Effect of cannabinoid CB1 receptor antagonist SR141716A and CB1 receptor knockout on cue- induced reinstatement of Ensure and corn-oil seeking in mice. Neu- ropsychopharmacology, 2007. 32(12): p. 2592-600. 281. Wardle J, Eating style: a validation study of the Dutch Eating Be- haviour Questionnaire in normal subjects and women with eating disorders. J Psychosom Res, 1987. 31(2): p. 161-9. 282. Warner ML, Harley K, Bradman A, Vargas G, and Eskenazi B, Soda consumption and overweight status of 2-year-old mexican-american children in california. Obesity (Silver Spring), 2006. 14(11): p. 1966-74. 283. Waters A, Hill A, and Waller G, Bulimics' responses to food cravings: is binge-eating a product of hunger or emotional state? Behav Res Ther, 2001. 39(8): p. 877-86. 284. Weingarten HP and Elston D, The phenomenology of food cravings. Appetite, 1990. 15(3): p. 231-46. 285. Weingarten HP and Elston D, Food cravings in a college population. Appetite, 1991. 17(3): p. 167-75. 286. Weiss F, Neurobiology of craving, conditioned reward and relapse. Curr Opin Pharmacol, 2005. 5(1): p. 9-19. 287. Weldon DT, O'Hare E, Cleary J, Billington CJ, and Levine AS, Ef- fect of naloxone on intake of cornstarch, sucrose, and polycose diets in restricted and nonrestricted rats. Am J Physiol, 1996. 270(6 Pt 2): p. R1183-8. 288. Wellman PJ, Nation JR, and Davis KW, Impairment of acquisition of cocaine self-administration in rats maintained on a high-fat diet. Pharmacol Biochem Behav, 2007. 88(1): p. 89-93. 289. Verbeek MEM, Drent PJ, and Wiepkema PR, Consistent individual differences in early exploratory behaviour of male great tits. Animal Behaviour, 1994. 48(5): p. 1113-1121. 290. Wetzler S, Dumaz V, Goubern M, Tome D, and Larue-Achagiotis C, Intraperitoneal leptin modifies macronutrient choice in self- selecting rats. Physiol Behav, 2004. 83(1): p. 65-72. 291. White MA, Whisenhunt BL, Williamson DA, Greenway FL, and Netemeyer RG, Development and validation of the food-craving inventory. Obes Res, 2002. 10(2): p. 107-14. 292. Willer CJ, Speliotes EK, Loos RJ, Li S, Lindgren CM, Heid IM, et al., Six new loci associated with body mass index highlight a neuronal influence on body weight regulation. Nat Genet, 2009. 41(1): p. 25-34. 293. Wise RA, Newton P, Leeb K, Burnette B, Pocock D, and Justice JB, Jr., Fluctuations in nucleus accumbens dopamine concentration during intravenous cocaine self-administration in rats. Psychopharma- cology (Berl), 1995. 120(1): p. 10-20. 294. Wise RA, Spindler J, deWit H, and Gerberg GJ, Neuroleptic- induced "anhedonia" in rats: pimozide blocks reward quality of food. Science, 1978. 201(4352): p. 262-4.

91 295. Volkow ND, Fowler JS, Wang GJ, Baler R, and Telang F, Imaging dopamine's role in drug abuse and addiction. Neuropharmacology, 2008. 296. Volkow ND, Fowler JS, Wang GJ, and Goldstein RZ, Role of dopamine, the frontal cortex and memory circuits in drug addiction: insight from imaging studies. Neurobiol Learn Mem, 2002. 78(3): p. 610-24. 297. Volkow ND, Wang GJ, Fowler JS, Logan J, Gatley SJ, Gifford A, et al., Prediction of reinforcing responses to psychostimulants in humans by brain dopamine D2 receptor levels. Am J Psychiatry, 1999. 156(9): p. 1440-3. 298. Volkow ND, Wang GJ, Fowler JS, Logan J, Jayne M, Franceschi D, et al., "Nonhedonic" food motivation in humans involves dopamine in the dorsal striatum and methylphenidate amplifies this effect. Synapse, 2002. 44(3): p. 175-80. 299. Volkow ND, Wang GJ, Fowler JS, Thanos PP, Logan J, Gatley SJ, et al., Brain DA D2 receptors predict reinforcing effects of stimulants in humans: replication study. Synapse, 2002. 46(2): p. 79-82. 300. Volkow ND, Wang GJ, Telang F, Fowler JS, Goldstein RZ, Alia- Klein N, et al., Inverse association between BMI and prefrontal metabolic activity in healthy adults. Obesity (Silver Spring), 2009. 17(1): p. 60-5. 301. Volkow ND, Wang GJ, Telang F, Fowler JS, Thanos PK, Logan J, et al., Low dopamine striatal D2 receptors are associated with prefrontal metabolism in obese subjects: possible contributing factors. Neuroimage, 2008. 42(4): p. 1537-43. 302. Volkow ND and Wise RA, How can drug addiction help us understand obesity? Nat Neurosci, 2005. 8(5): p. 555-60. 303. Woods JE, McKay PF, Masters J, Seyoum R, Chen A, La Duff L, et al., Differential Responding for Brain Stimulation Reward and Su- crose in High–Alcohol-Drinking (HAD) and Low–Alcohol-Drinking (LAD) Rats. Alcohol Clin Exp Res, 2003. 27(6): p. 926-936. 304. Wynne K, Stanley S, McGowan B, and Bloom S, Appetite control. J Endocrinol, 2005. 184(2): p. 291-318. 305. Yeo GS, Farooqi IS, Aminian S, Halsall DJ, Stanhope RG, and O'Rahilly S, A frameshift mutation in MC4R associated with domi- nantly inherited human obesity. Nat Genet, 1998. 20(2): p. 111-2. 306. Yirmiya R, Lieblich I, and Liebeskind JC, Reduced saccharin preference in CXBK (opioid receptor-deficient) mice. Brain Res, 1988. 438(1-2): p. 339-42. 307. Zhang M, Balmadrid C, and Kelley AE, Nucleus accumbens opioid, GABaergic, and dopaminergic modulation of palatable food motivation: contrasting effects revealed by a progressive ratio study in the rat. Behav Neurosci, 2003. 117(2): p. 202-11. 308. Zhang M, Gosnell BA, and Kelley AE, Intake of high-fat food is selectively enhanced by mu opioid receptor stimulation within the nucleus accumbens. J Pharmacol Exp Ther, 1998. 285(2): p. 908-14.

92 309. Zhang M and Kelley AE, Opiate agonists microinjected into the nucleus accumbens enhance sucrose drinking in rats. Psychophar- macology (Berl), 1997. 132(4): p. 350-60.

! " # $ % &# ! # ! ' ' ( & ")* + ,, # ! -( & ."/

! ! "" ! 0112332