Regulation of energy homeostasis by peripheral signals

Best Practice & Research Clinical Endocrinology & Metabolism Vol. 18, No. 4, pp. 497–515, 2004 doi:10.1016/j.beem.2004.08.004 available online at http://www.sciencedirect.com

4 Regulation of energy homeostasis by peripheral signals Stephen C. Woods* Stephen C. Benoit1 Deborah J. Clegg2 Randy J. Seeley3 Department of Psychiatry, University of Cincinnati, 2170 East Galbraith Road, Cincinnati, OH 45237, USA

The increased incidence of obesity makes it imperative to understand the regulation of food intake and body weight. We review the signals that interact with the brain to control energy homeostasis, i.e. energy intake and expenditure. Three broad categories can be distinguished. Signals generated in the gastrointestinal tract during meals (‘satiety’ signals, e.g. cholecystokinin) elicit satiation and contribute to stopping the meal. The potency of these acutely acting signals must be increased if they are to be used therapeutically. Hormonal signals whose secretion is proportional to body fat (adiposity signals, leptin and insulin) robustly reduce food intake and body weight by directly stimulating receptors locally in the brain. Therapeutic applications will have to find ways to circumvent the systemic actions of these hormones, targeting only the brain. Satiety and adiposity signals interact with neuronal circuits in the brain that utilize myriad neurotransmitters to cause net catabolic or anabolic responses. Considerable effort is being directed towards finding ways to intervene in specific circuits to help accomplish weight loss. Key words: obesity; satiety; body fat distribution; melanocortins; neuropeptide Y; adiponectin.

* Corresponding author. Tel.: C1-513-558-2699; Fax: C1-513-558-8990. E-mail addresses: [email protected] (S.C. Woods), [email protected] (S.C. Benoit), [email protected] (D.J. Clegg), [email protected] (R.J. Seeley). 1 Tel.: C1-513-558-4313. 2 Tel.: C1-513-558-3429. 3 Tel.: C1-513-558-6664. 1521-690X/$ - see front matter Q 2004 Elsevier Ltd. All rights reserved.

498 S. C. Woods et al

THE ENERGY EQUATION AND THE OBESITY EPIDEMIC Food intake is one of the commonest behaviors we undertake on a daily basis, and one with profound consequences because food intake also comprises one side of the energy equation. The energy equation holds that if body weight is to remain stable, food intake (i.e. energy intake) must equal energy expenditure (i.e. metabolism plus the effects of exercise) over long intervals. If the energy equation is not in balance, body weight will drift upwards or downwards over time.1 In point of fact, there are two common sentiments related to the energy equation and the regulation of body weight, which appear to be self-contradictory. The first is that the US, Europe and much of the world are currently experiencing an ‘epidemic of obesity’.2–4 That is, both average body weight and the incidence of overweight and obese individuals within these populations are increasing at an alarming rate. The second sentiment is that the physiological system that regulates body weight is robust and very precise.5–8 Stated another way, a control system exists in the body that monitors energy stores and perhaps energy expenditure, and uses this information to influence both food intake and body weight. This is a negative feedback-type system (Figure 1) that acts to maintain relatively constant body fat stores over long intervals. An important question therefore concerns how and why body weight is increasing on a population basis in the face of a regulatory system whose function is to prevent deviations on a chronic basis. While there are no easy answers to this question5,9, the solution may lie in a common misconception about body weight. It should be noted that we are referring to ‘body weight’ as being synonymous with body adiposity; i.e. we assume that a change of body weight reflects an underlying change of body fat, and that body fat, rather than weight per se, is the regulated variable. The common misconception is that there is a set point for body weight, i.e. a genetically or ontogenetically determined ‘ideal’ amount of weight that an individual should carry, and that deviations from this ideal weight are detected and trigger compensatory reflexes via the regulatory system that function to restore body weight to ideal whenever possible.10–12 We believe that a better characterization is that rather than there being a set point, there is a range of body weight that an individual is willing to accept and defend.1,13,14 The upper and lower limits of the range are probably genetically determined and under independent control. The actual weight that is defended by the regulatory system will lie somewhere within the range, the exact location being a function of environmental factors. Many of these factors are well known. For example, an individual living with chronic stress, or exercising regularly, tends to defend a lower body weight. Conversely, an individual with ample, easy-to-obtain and highly palatable food available tends to defend a higher body weight. Viewed in this way, the epidemic of obesity can be considered to be an appropriate response on the part of a population existing in an environment with increased access to and availability of palatable, energydense foods and with less average physical energy needed to be exerted to obtain it. In this schema, the regulatory system is working well, and it is the environment that is changing. Another consideration is the magnitude of the ‘epidemic’. A prototypical adult man of 70–75 kg consumes around 1 000 000 kcal per year. If that individual were to gain one pound (approximately 0.45 kg) per year, it would represent an error of regulation of only around 11 kcal/day, or less than 0.5%. This obviously represents only a very small slippage of the regulatory system, and in actuality the ‘epidemic’ of obesity comprises an average gain of less than one pound per year, the point being that

Regulation of energy homeostasis by peripheral signals 499

Figure 1. Model summarizing different levels of control over energy homeostasis. During meals, signals such as cholecystokinin (CCK) or distension of the stomach that arise from the gut (stomach and intestine) trigger nerve impulses in sensory nerves traveling to the hindbrain. These satiety signals synapse with neurons in the nucleus of the solitary tract (NTS), where they influence meal size. Signals related to body fat content such as leptin and insulin, collectively called adiposity signals, circulate in the blood to the brain. They pass through the blood-brain barrier in the region of the arcuate nucleus (ARC) and interact with neurons that synthesize proopiomelanocorticotropin (POMC) or neuropeptide Y (NPY) and Agouti-related protein (AgRP) (NPY). These neurons in turn project to other hypothalamic areas, including the paraventricular nuclei (PVN) and the lateral hypothalamic area (LHA). The net output of the PVN is catabolic and enhances the potency of satiety signals in the hindbrain. The net output of the LHA, on the other hand, is anabolic, suppressing the activity of the satiety signals. In this way, body fat content tends to remain relatively constant over long intervals by means of changes of meal size. GI, gastrointestinal; SNS, sympathetic nervous system. (Reproduced from Ref. 7).

the regulatory system is quite precise in spite of the recent upward drift of mean body weight.

THE CONTROL OF ENERGY INTAKE Meals Food intake in mammals, including humans, occurs in distinct bouts or meals, and the number and size of meals over the course of a day comprises the meal pattern. Most humans are quite habitual in that they eat approximately the same number of meals, and at the same time of day, on a daily basis.15,16 There is, however, a considerable variation among individuals, with some having three meals a day, others four, others with small snacks interspersed, and so on. It is generally accepted that the factors that control

500 S. C. Woods et al

when meals occur are different from those that control when they end; i.e. different factors control meal onset and meal size.16,17 Historically, meal onset was thought to be a reflexive response to a reduction in the amount or availability of some parameter related to energy. The most popular such position, the glucostatic theory, posited that a reduction of glucose utilization by sensor cells in the hypothalamus of the brain was the cause of hunger and a tendency to start a meal.18,19 Other hypotheses have been based upon body heat, upon fat utilization by the liver, upon the generation of adenosine triphosphate (ATP) and other energy-rich molecules by cells in the liver and/or brain, and so on.20–23 It is generally accepted today that whereas a sudden decrease in one or another parameter of available energy can cause an individual to begin eating at a time it ordinarily would not, the threshold decrease is far greater than what would occur before normal or spontaneous meals, and the eating in that instance is considered to be an emergency response.22,24 The alternative position is that most meals are initiated at times that are convenient or habitual and are thus based upon social or learned factors as opposed to fluxes of energy within the body.24 In this schema, the regulatory control over food intake is exerted on how much food is consumed once a meal is started rather than on when the meal occurs.25,26 This allows individuals considerable flexibility as they adapt their individualized meal patterns to their environment and lifestyle while still maintaining control over the amount of food consumed. Hence, regulatory controls determine meal size, and this is generally equated with the phenomenon of satiety or fullness.17 Satiety If meal size is a regulated parameter, the individual must have a means of reliably measuring how much food has been eaten; i.e. the number of calories consumed, or perhaps the precise mix of carbohydrates, lipids and proteins, and/or other foodrelated parameters, must be monitored as the meal progresses so that the person knows when to say ‘I am full’ and put down her fork.17 In principle, any number of parameters might provide the important feedback during an ongoing meal. The individual might use vision, or smell, or taste to gauge the amount of energy consumed. However, several types of experiment have found that any such input is at best minimal. Perhaps the most persuasive are experiments in which animals have an implanted gastric fistula.27 When the fistula is closed, swallowed food enters the stomach, is processed normally and moves into the duodenum. When the fistula is open, swallowed food enters the stomach and then exits the body via the fistula in a process called sham eating. In both instances, the visual, olfactory and taste inputs are the same, but the amount eaten varies considerably. When the fistula is closed (real eating), animals eat normal-sized meals; when the fistula is open (sham eating), animals continue eating for long intervals and consume very large meals.27–29 Hence, whatever signals an individual uses to gauge how many calories have been consumed must arise no more proximally than the distal stomach and/or small intestine. As ingested food interacts with the stomach and intestine, it elicits the secretion of an array of gut peptides and other signals that function to coordinate and optimize the digestive process. In 1973, Gibbs and Smith and their colleagues reported that the gut peptide cholecystokinin (CCK) acts as a satiety signal. When purified or synthetic CCK was administered to rats prior to a meal, it dose-dependently reduced the size of that meal.30 Dozens of experiments have since documented the ability of exogenous CCK to reduce meal size in numerous species including humans.31–34 A role of endogenous

Regulation of energy homeostasis by peripheral signals 501

CCK in eliciting satiety is indicated by the observation that the administration of specific CCK-1 receptor antagonists prior to a meal causes an increased meal size in animals and humans35–38 and reduces the subjective feeling of satiety in humans.35 Satiety signals Any endogenous factor that reduces the size of an ongoing meal is considered to be a satiety signal, and it is thought that several different gut peptides normally contribute.39–42 This enables species such as general omnivores that can derive energy from eating plants, other animals or any nutrient source, species that include laboratory rats and humans, to eat whatever is available and to secrete a blend of gut peptides appropriate for digesting the food while at the same time informing the brain precisely what has been consumed. CCK is secreted in response to some carbohydrates and fats, and other peptides are secreted in response to other mixes of macronutrients. Besides CCK, gastrin releasing peptide43, neuromedin B44, enterostatin45,46, somatostatin47, glucagon-like peptide-148,49, apolipoprotein A-IV50 and peptide YY(3-36)51 are all peptides secreted from the gastrointestinal system that have been reported to reduce meal size when administered systemically. In addition, amylin52,53 and glucagon54,55 secreted from the pancreatic islets during meals also have this property. Although the exact way in which each of these peptides signals the central nervous system and contributes to the phenomenon of satiety may differ, most are thought either to activate receptors on vagal afferent fibers passing to the hindbrain (e.g. CCK56–58, glucagon59,60) or else to stimulate the hindbrain directly at sites with a relaxed blood-brain barrier (e.g. amylin61,62). The signals from different peptides, as well as signals related to stomach distension, are thought to be integrated either within the vagal fibers themselves or else in the hindbrain as they generate an overall signal that ultimately causes the individual to stop eating.63–66 Ghrelin Two generalizations can be made regarding gastrointestinal peptides that contribute to satiety. The first is that most if not all of them are also synthesized in areas of the brain involved in overall caloric homeostasis. The second is that, with one exception, the net result of an increased level of any of these signals is a reduction in meal size. The lone exception is the recently described gastric peptide ghrelin, an agonist at the growth hormone secretagogue receptor. Ghrelin levels are increased after fasting and prior to meals67, and the administration of exogenous ghrelin causes large meals to be eaten.68–71 Like the satiety peptides, ghrelin is made in the brain as well as in the stomach72, and there is evidence that the systemic ghrelin signal is carried in vagal afferent nerves to the brain.73 It is worth pondering why it is that out of the numerous peptides secreted from the stomach and intestines that influence food intake, only one, ghrelin, increases it. Because meals generally end well before the physical limit of the stomach is reached, there must be a teleological reason to put a curb on eating when more intake is possible. It has been argued that a major function of gastrointestinal signals is therefore to restrain the consumption of excess calories at one time to minimize the increase of postprandial blood glucose and other nutrients.25,74 Considered in this light, the number and diversity of gastrointestinal signals that reduce meal size are appropriate for signaling the complexity of macronutrients in different meals. The fact that only one

502 S. C. Woods et al

orexigenic gastrointestinal peptide has been discovered speaks of the importance of limiting meal size. Summary In summary, when food is eaten, it interacts with receptors lining the stomach and intestine, causing the release of peptides and other factors that coordinate the process of digestion with the particular food being consumed. Some of the peptides provide a signal to the nervous system, and as the integrated signal accumulates, it ultimately creates the sensation of fullness and contributes to cessation of eating. Administering the same peptides exogenously elicits a dose-dependent reduction of meal size, whereas administering antagonists (e.g. for CCK) or antisera (e.g. for apolipoprotein A-IV) causes increased meal size. An important and generally unanswered question concerns whether agonists for satiety signals, and/or antagonists for ghrelin, have therapeutic value in treating obesity. At a first-order level of analysis, the answer seems clear. That is, if the size of individual meals is reduced (e.g. by administering CCK prior to each meal), individuals adjust by increasing how often they eat and maintaining their total daily intake essentially constant.75,76 However, CCK and the other gut-derived satiety signals have very short half-lives, of the order of magnitude of one or a few minutes. Hence, longacting analogs of the satiety signals may have efficacy in causing weight loss, especially since rats with a genetic absence of functional CCK-1 receptors gradually become obese over their lifetime.77 This is an area of considerable research activity at present.

ADIPOSITY SIGNALS Body fat regulation The other important information needed to regulate body weight pertains to the amount of fat stored in the body (see Figure 1 above). That is, it is well known that if an individual loses body weight (e.g. by dieting), there is a tendency to eat more food and restore body weight when conditions allow. Analogously, if an individual voluntarily overeats, or is forced to overeat, and gain weight, there is a tendency to reduce food intake and lose the gained weight when conditions allow. This is the fundamental tenet of the body weight regulatory system—that as the amount of body fat changes, signals are generated that ultimately interact with the controllers of food intake. A decrease of fat causes increased food intake, whereas an increase of fat causes decreased food intake in a strict negative feedback—type manner.6–8,78 Monitoring stored fat presents a challenge to the brain because the fat is located in multiple depots throughout the body, and when total fat becomes greatly increased, fat becomes stored in other cell types such as liver and skeletal muscle as well.79 While it is possible that sensory nerves could pass from each individual fat depot to the brain, become integrated with one another and provide a ‘total fat signal’, no such neural network has been described. Instead, a different signaling system has evolved in which the secretion of specific hormones into the blood is directly correlated with stored fat. Hence, the brain need only have receptors for those hormonal signals to obtain a reliable estimate of body fat, a hypothesis initially proposed by Kennedy half a century ago.80

Regulation of energy homeostasis by peripheral signals 503

Signals for adiposity At least two hormones meet the criteria to be adiposity signals to the brain: insulin and leptin. Insulin is secreted from pancreatic b cells in response to circulating glucose, fatty acids and amino acids. The amount of insulin secreted both during basal (quiescent and unstimulated) conditions and in response to increased glucose is directly proportional to body fat, such that obese individuals have elevated basal and stimulated insulin compared with lean individuals.81,82 In contrast, leptin is secreted directly from white adipose cells in response to ongoing metabolic activity of the cell, the level of secretion being directly proportional to the amount of stored fat.83,84 Hence, any tissue with receptors for insulin and/or leptin will receive information on the total amount of fat in the body. Insulin and leptin share several features in common as adiposity signals (see reviews in Refs. 6–8,85–87). Both are large peptides secreted into the general circulation that cannot easily diffuse through the blood–brain barrier to gain access to their central receptors. Both are, however, passed from the blood into brain interstitial fluid by receptor-mediated transport through the brain capillary endothelial cells.88–95 Receptors for both insulin and leptin are located on neurons in the hypothalamus and elsewhere in the brain96–102, including the key neuronal populations in the hypothalamic arcuate nucleus that synthesize proopiomelanocorticotropin (POMC neurons) and neuropeptide Y and agouti-related protein (NPY/AgRP neurons).101,103–105 As discussed below, the POMC and NPY/AgRP neurons in the arcuate nucleus are considered to be highly important controllers of food intake and body weight. Finally, changes in the level of either insulin or leptin within the vicinity of the arcuate nucleus cause predictable changes of food intake and body weight. Increased insulin or leptin activity (signaling an increase of body fat) results in the stimulation of POMC and inhibition of NPY/AgRP, decreased food intake, increased energy expenditure and weight loss. Conversely, decreased insulin or leptin activity results in inhibition of POMC and stimulation of NPY/AgRP, increased food intake, decreased energy expenditure and weight gain (see reviews in Refs. 6–8,85,86). The normal reliance of the control system for energy homeostasis on endogenous insulin and leptin is best demonstrated when their influence is removed. Decades before leptin was discovered, its existence was inferred from the overeating and obesity characteristic of animals unable to synthesize leptin or its receptor106,107; in addition, a hallmark symptom of insulin deficiency is extreme hyperphagia. Insulindeficient individuals are lean rather than obese because a major action of systemic insulin is to allow lipid to be stored in fat cells. Insulin deficiency hyperphagia is attenuated by the administration of insulin108 or leptin109 locally in the vicinity of the arcuate nucleus, just as the hyperphagia of leptin-deficient animals is attenuated by the central administration of leptin.110 Consistent with these observations, the administration of antibodies to insulin directly into the hypothalamus causes increased food intake and body weight111,112, and the selective knockout of either the leptin receptor113 or the insulin receptor114,115 uniquely in the brain results in hyperphagia and obesity. While insulin and leptin share many properties as adiposity signals to the brain, there are important differences between them. For one thing, their levels do not reflect the same adipose depots. Most leptin comes from subcutaneous fat, its levels in the blood therefore signaling subcutaneous fat more reliably than total body fat.116,117 In contrast, insulin is secreted in proportion to visceral fat, plasma insulin therefore being a better

504 S. C. Woods et al

correlate of visceral than total body fat.117,118 Hence, the two signals combined provide a more accurate picture of total fat than either alone, and they also provide information on where fat is distributed. Women have relatively more fat in subcutaneous depots, whereas men have more fat in visceral depots.117,119,120 Leptin is therefore a more accurate signal of total body fat in females and insulin a more accurate signal in males. Consistent with this, the brains of female rats are relatively more sensitive to low doses of leptin, whereas the brains of male rats are relatively more sensitive to low doses of insulin with regard to their catabolic actions.121 Leptin has a half-life in the plasma of around 45 minutes, whereas insulin has a half-life of only 2–3 minutes. Thus, leptin represents a relatively stable indicator of activity in subcutaneous fat, whereas insulin is a rapidly changing signal reflecting changes in recent carbohydrate metabolism but nonetheless related to body fat. Central adiposity pathways Both insulin and leptin stimulate POMC neurons and inhibit NPY/AgRP neurons in the arcuate nucleus.85,97,103,122,123 As each interacts with its respective receptor, it initiates a unique chain of intracellular events that ultimately mediates its actions. These intracellular pathways are the subject of intense investigation as potential therapeutic applications of insulin and leptin signaling are considered.86,124–126 An important recent observation is that the intracellular signaling cascades of leptin and insulin overlap, such that drugs that block the activity of phosphatidylinositol–3 kinase prevent either leptin or insulin from exerting its catabolic actions126,127, and both leptin and insulin have been found to stimulate neuronal ATP-sensitive potassium channels.128,129 Hence, there is a convergence of the two adipose signals within the arcuate. Consistent with this, when insulin is administered into the brain of rats with deficient leptin receptors (fatty Zucker rats), they do not reduce their food intake and lose weight.130 Adiponectin A third hormone, adiponectin, is also secreted and circulates in proportion to adiposity. Although it is secreted from adipose cells131, adiponectin levels are, however, inversely proportional to body fat132, especially visceral fat.133,134 When adiponectin is administered locally into the brain, it increases energy expenditure and causes weight loss with no change of food intake135, adding it to the roster of circulating adiposity signals.136 How adiponectin interacts with other adiposity signals has not yet been determined.

INTEGRATION OF SATIETY AND ADIPOSITY SIGNALS The information about total body fat derived from insulin and leptin must be integrated with satiety signals as well as with other signals related to myriad other factors, including learning, the social situation, stress and so on, for the control system to be maximally efficient. Although the nature of these interactions is not well understood, several generalizations or conclusions can be made. For one, the negative feedback circuits related to body fat and meal ingestion can easily be overridden by situational events. As an example, even though satiety signals might indicate that no more food should be eaten during an ongoing meal, the sight, smell and perceived palatability of an offered dessert can stimulate further intake. Likewise, even though an individual is

Regulation of energy homeostasis by peripheral signals 505

severely underweight and food is available, the influence of stressors can preclude significant ingestion. Because of these kinds of interaction, trying to relate food intake within an individual meal to recent energy expenditure or to fat stores is futile, at least in the short term. Rather, the influence of homeostatic signals becomes apparent only when intake is considered over longer intervals. That is, if homeostatic signals predominated, a relatively large intake in one meal should be compensated by a reduced intake in the subsequent meal. Detailed analyses have, however, revealed that such compensation, if it occurs at all, is only apparent when intervals of one or more days are considered in humans.137,138 This phenomenon was initially demonstrated in a rigorous experiment using rabbits, in which weekly intake correlated better with recent energy expenditure than did intake after 1 or 3 days.139 A second conclusion, as discussed above, is that homeostatic regulation is most apparent on meal size rather than on meal initiation or patterning. A large literature has documented that the specific meal pattern adopted by individuals is idiosyncratic and related more to environmental constraints than to metabolism or energy stores24,25,140–142, enabling individuals to be flexible, timing their eating bouts to times that food is available and obtainable while simultaneously coordinating ingestion with other behaviors. Regulation occurs in the determination of how much food is eaten once a meal is underway. A final conclusion is that the homeostatic controls over food intake act by changing the sensitivity to satiety signals. The adiposity signals insulin and leptin have been found to change the sensitivity to CCK. Hence, when an individual has gained excess weight, more insulin and leptin stimulate the brain, and this in turn renders CCK more efficacious at reducing meal size.143–147 An increased insulin signal in the brain also renders individuals more sensitive to the meal size-reducing action of amylin148 and the neuropeptide corticotropin-releasing hormone.149 As discussed above, the integration of satiety signals with other signals that influence meal size occurs in vagal afferent fibers and continues into the hindbrain64,150,151, where meal size is ultimately determined.152 At the same time, the hypothalamic arcuate nucleus receives signals related to adiposity as well as information on ongoing meals from the hindbrain, and it is able to monitor ongoing metabolism directly, providing it with the capacity to integrate multiple signals that determine ingestion.5,7,153–155 CENTRAL SIGNALS RELATED TO ENERGY HOMEOSTASIS As an oversimplification, the neural circuits in the brain that control energy homeostasis can be subdivided into those which receive relevant sensory information (afferent circuits), those which integrate the information and those which control motor, autonomic and endocrine responses (efferent circuits). The discussion above has focused upon some of the relevant afferent satiety and adiposity signals and how they influence food intake. Peptides such as insulin, leptin and CCK can be considered as surrogate signals whose levels reflect key metabolic events such as the number of calories eaten or the number of calories stored in the body. Recent evidence suggests that more direct metabolic signals arise within the brain itself and also influence food intake. Cellular metabolism Most cells in the body are able to oxidize the substrates glucose and/or fatty acids to release and consequently capture the chemical energy they contain. As oxygen

506 S. C. Woods et al

combines with these substrates in the mitochondria of the cell, water and carbon dioxide are produced and the substrate’s energy is transferred into molecules such as ATP that can be used as needed to power cellular processes. Any treatment that compromises the formation of ATP disables cells, and if this were to occur in the brain, it might be expected to generate a signal that led to increased eating, and this in fact occurs.23,156–158 Most cells in the body have complex means of maintaining adequate ATP generation because they are able to oxidize either glucose or fatty acids. Hence, if one or other substrate becomes low, enzymatic changes occur to increase the ability of the cell to take up and oxidize the alternate fuel more rapidly. Energy sensing Historically, several hypotheses have posited that cells in the brain function as sensors of their own oxidation of substrates and thereby generate a signal that interacts with other controllers of energy homeostasis.23,158 Related to this, it has generally been assumed that the brain is sensitive mainly to changes of carbohydrate oxidation because neurons have an obligate need to use glucose for energy and because, unlike most other cells in the body, they do not require insulin to take in and oxidize the glucose. Several recent observations have, however, challenged this long-held belief and suggest that some hypothalamic cells are sensitive to local lipid metabolism as well as to carbohydrate metabolism, and that they can generate powerful signals that integrate the two and influence food intake. When energy is abundant, such as during and immediately after meals, most cells throughout the body have the ability to synthesize fatty acids via the cellular enzyme fatty acid synthase. Although this was not thought to occur in the brain, recent evidence indicates that the entire lipid-metabolizing pathway also exists in the hypothalamus.159 Importantly, when fatty acid synthase activity is inhibited locally in the brain by the drug C75, animals eat less food and, over the course of a few days, selectively lose body fat.160–162 A similar phenomenon occurs if C75 is administered systemically.162 However, interpreting the reduction of food intake following systemic C75 is confounded by a generalized malaise that is not apparent when the drug is given locally into the brain.160 One important implication of these findings is that some hypothalamic cells are able to oxidize fatty acids163, and a second implication is that a signal related to this metabolic process is linked to the controls over energy homeostasis. Because recent evidence suggests that C75 elicits compensatory changes in brain carbohydrate metabolism as well164, the clear implication is that generalized energy sensors must exist in the hypothalamus. Consistent with this, increases of either carbohydrate or long-chain fatty acid availability locally in the arcuate nucleus reduce food intake and send signals to the liver to reduce the secretion of energy-rich fuels into the blood.165 The concept that some brain neurons can utilize either glucose or lipids for energy and hence function as overall energy sensors is not new (see, for example Refs. 22,23, 166). What is new, however, is the application of highly selective molecular probes to assess the role of individual enzymatic steps in the metabolic cascade of brain cells that influence physiology and behavior, and the concept that a population of cells in the brain is sensitive to local manipulations of fatty acids and/or glucose. As originally reported by Oomura167–169, and consistent with glucose-sensing cells in other parts of the body, the electrical activity of these cells can be changed by local fluctuations of glucose, leptin or insulin.170–172 These same cells are also thought to contain receptors and enzymes that are characteristic of glucose-sensing pancreatic b cells. Hence, like b cells, these cells

Regulation of energy homeostasis by peripheral signals 507

can detect changes of glucose and generate signals that influence metabolism and behavior.171,172 There also is evidence that the same or proximally close neurons contain receptors for leptin and insulin. The picture that is emerging is therefore of a population of neurons that collectively sample different classes of energy-rich molecules (i.e. glucose and fatty acids) as well as hormones whose levels reflect metabolism throughout the body (i.e. insulin and leptin). These same neurons appear also to be sensitive to the myriad neuropeptides known to be important regulators of energy homeostasis.23

CENTRAL INTEGRATIVE CIRCUITS Melanocortins This is a rapidly expanding area of research and beyond the scope of this review. Because of this, only a few generalizations will be made. As discussed above, the arcuate nucleus of the hypothalamus comprises a hub, where many kinds of signals related to energy homeostasis converge. Adiposity signals from the circulation interact with specific insulin and leptin receptors in the arcuate nucleus. Satiety signals entering the hindbrain are relayed to the arcuate and other hypothalamic areas. The arcuate itself (or nearby hypothalamic areas) contains neurons that generate signals in proportion to their oxidation of glucose and fatty acids. Although the arcuate nucleus contains numerous kinds of neuron, the POMC and NPY/AgRP neurons are best known. While several possible neuropeptides can be synthesized from POMC, those whose function in arcuate POMC neurons is best understood use a-melanocyte stimulating hormone (aMSH) as the neurotransmitter. POMC neurons project to many other areas of the hypothalamus and brain in general where aMSH is released and causes a net catabolic response, i.e. reduced food intake, increased energy expenditure and loss of body weight. POMC neurons contain specific receptors for leptin and insulin, and insulin and leptin each stimulate the synthesis of POMC and the release of aMSH. Blocking this chain of events at any point, from the insulin or leptin receptors to aMSH to melanocortin receptors (specifically MC3 and MC4 receptors) prevents insulin or leptin from decreasing food intake and body weight. There are many reviews of this system.5,7,8,26,155,173–175 Hence, considerable activity is being focused on the possibility that agonists for these MC3 and/or MC4 receptors would be useful therapeutic agents to reduce the incidence of obesity.175 Anabolic systems NPY/AgRP neurons have a somewhat opposite profile. They also contain receptors for leptin and insulin, and the local administration of leptin or insulin inhibits their activity. NPY acts in other hypothalamic areas to stimulate food intake, reduce energy expenditure and increase body weight. AgRP on the other hand is an antagonist at MC3 and MC4 receptors. Hence, increased activity of NPY/AgRP neurons stimulates anabolic circuits while simultaneously inhibiting melanocortin-driven catabolic circuits.5,7,8,26,155

508 S. C. Woods et al

SUMMARY Several types of signal are involved in the regulation of energy homeostasis by the brain. Because appetite and meal initiation are generally not under homeostatic control, body weight is regulated by how much food is eaten once a meal begins. Satiety signals generated as ingested food interacts with the stomach and intestine contribute to satiation. Adiposity signals are detected directly in the brain and work by changing the sensitivity of satiety signals such that when individuals lose weight, they are less sensitive to satiety signals and eat larger meals until the weight is restored. The converse occurs when one gains weight: smaller meals are eaten until the excess weight is lost. These signals in turn interact with environmental and social factors to determine body weight.

ACKNOWLEDGEMENTS The preparation of this review was supported by several grants from the United States National Institutes of Health.

Practice points † the incidence of obesity has been increasing at a rate termed epidemic; nonetheless, the body’s regulatory system remains robust † hunger and meal initiation are elicited by conditioned and social cues, whereas satiation is caused by gastrointestinal signals indicating how many calories have been consumed † combined pharmaceutical approaches that target more than the area of regulation (such as satiety and adiposity) are more efficacious than those targeting single processes † weight loss strategies differ for men and women † treatments that reduce body weight will reduce symptoms of type 2 diabetes † pharmaceuticals that target the brain are likely to have undesirable side effects

Research agenda † ways to prolong the action of short-acting satiety signals are needed † developing pharmaceuticals that mimic the action of adiposity signals locally in the brain is a high priority † identifying the controls over visceral as opposed to subcutaneous fat should is needed † understanding how energy use by the brain is translated into energy intake and expenditure will lead to new therapeutic strategies

Regulation of energy homeostasis by peripheral signals 509

REFERENCES 1. Berridge KC. Motivation concepts in behavioral neuroscience. Physiology and Behavior 2004; 81: 179–209. 2. Pi-Sunyer FX, Laferrere B, Aronne LJ & Bray GA. Therapeutic controversy: Obesity—a modern-day epidemic. Journal of Clinical Endocrinology and Metabolism 1999; 84: 3–12. 3. Popkin BM & Doak C. The obesity epidemic is a worldwide phenomenon. Nutrition Reviews 1998; 56: 106–114. 4. Stein CJ & Colditz GA. The epidemic of obesity. Journal of Clinical Endocrinology and Metabolism 2004; 89: 2522–2525. *5. Flier JS. Obesity wars: molecular progress confronts an expanding epidemic. Cell 2004; 116: 337–350. *6. Woods SC, Seeley RJ, Porte DJ & Schwartz MW. Signals that regulate food intake and energy homeostasis. Science 1998; 280: 1378–1383. *7. Schwartz MW, Woods SC, Porte DJ et al. Central nervous system control of food intake. Nature 2000; 404: 661–671. 8. Seeley RJ & Woods SC. Monitoring of stored and available fuel by the CNS: implications for obesity. Nature Reviews Neuroscience 2003; 4: 901–909. 9. Schwartz MW, Woods SC, Seeley RJ et al. Is the energy homeostasis system inherently biased toward weight gain? Diabetes 2003; 52: 232–238. 10. Keesey RE & Hirvonen MD. Body weight set-points: determination and adjustment. Journal of Nutrition 1975; 127: 1875S–1883S. 11. Mrosovsky N & Powley TL. Set points for body weight and fat. Behavioral Biology 1977; 20: 205–223. 12. Woods SC, Decke E & Vasselli JR. Metabolic hormones and regulation of body weight. Psychology Review 1974; 81(1): 26–43. 13. Pinel JPJ, Assanand S & Lehman DR. Hunger, eating and ill health. American Psychologist 2000; 55: 1105–1116. 14. Wirtshafter D & Davis JD. Set points, settling points, and the control of body weight. Physiology and Behavior 1977; 19: 75–78. 15. de Castro JM & Stroebele N. Food intake in the real world: implications for nutrition and aging. Clinical and Geriatric Medicine 2002; 18: 685–697. 16. de Castro JM, The control of eating behavior in free living humans. In Stricker EM & Woods SC (eds.) Handbook of Neurobiology. Neurobiology of Food and Fluid Intake, vol.14, no.2. New York: Kluwer Academic, 2004, pp. 467–502. 17. de Graaf C, Blom WAM & Smeets PAM. Biomarkers of satiation and satiety. American Journal of Clinical Nutrition 2004; 79: 946–961. 18. Mayer J. Regulation of energy intake and the body weight: the glucostatic and lipostatic hypothesis. Annals of the New York Academy of Sciences 1955; 63: 14–42. 19. Mayer J & Thomas DW. Regulation of food intake and obesity. Science 1967; 156: 328–337. 20. Friedman MI. An energy sensor for control of energy intake. Proceedings of the Nutrition Society 1997; 56(1A): 41–50. 21. Friedman MI. Fuel partitioning and food intake. American Journal of Clinical Nutrition 1998; 67(supplement 3): 513S–518S. 22. Langhans W. Metabolic and glucostatic control of feeding. Proceedings of the Nutrition Society 1996; 55: 497–515. 23. Peters A, Schweiger U, Pellerin L et al. The selfish brain: competition for energy resources. Neuroscience and Biobehavioral Reviews 2004; 28: 143–180. 24. Strubbe JH & Woods SC. The timing of meals. Psychological Review 2004; 111: 128–141. 25. Woods SC & Strubbe JH. The psychobiology of meals. Psychonomic Bulletin and Review 1994; 1: 141–155. 26. Woods SC, Schwartz MW, Baskin DG & Seeley RJ. Food intake and the regulation of body weight. Annual Review of Psychology 2000; 51: 255–277. 27. Davis JD & Campbell CS. Peripheral control of meal size in the rat. Effect of sham feeding on meal size and drinking rate. Journal of Comparative and Physiological Psychology 1973; 83(3): 379–387.

510 S. C. Woods et al 28. Davis JD & Smith GP. Learning to sham feed: behavioral adjustments to loss of physiological postingestional stimuli. American Journal of Physiology 1990; 259(6 Pt 2): R1228–R1235. *29. Gibbs J, Young RC & Smith GP. Cholecystokinin elicits satiety in rats with open gastric fistulas. Nature 1973; 245: 323–325. 30. Gibbs J, Young RC & Smith GP. Cholecystokinin decreases food intake in rats. Journal of Comparative and Physiological Psychology 1973; 84: 488–495. 31. Kissileff HR, Pi-Sunyer FX, Thornton J & Smith GP. Cholecystokinin decreases food intake in man. American Journal of Clinical Nutrition 1981; 34: 154–160. 32. Muurahainenn N, Kissileff HR, Derogatis AJ & Pi-Sunyer FX. Effects of cholecystokinin-octapeptide (CCK-8) on food intake and gastric emptying in man. Physiology and Behavior 1988; 44: 644–649. 33. Moran TH & Schwartz GJ. Neurobiology of cholecystokinin. Critical Reviews of Neurobiology 1994; 9: 1–28. 34. Smith GP & Gibbs J. The development and proof of the cholecystokinin hypothesis of satiety. In Dourish CT et al (ed.) Multiple Cholecystokinin Receptors in the CNS. Oxford: Oxford University Press, 1992, pp. 166–182. 35. Beglinger C, Degen L, Matzinger D et al. Loxiglumide, a CCK-A receptor antagonist, stimulates calorie intake and hunger feelings in humans. American Journal of Physiology 2001; 280: R1149–R1154. 36. Hewson G, Leighton GE, Hill RG & Hughes J. The cholecystokinin receptor antagonist L364,718 increases food intake in the rat by attenuation of endogenous cholecystokinin. British Journal of Pharmacology 1988; 93: 79–84. 37. Moran TH, Ameglio PJ & Peyton HJ. Blockade of type A, but not type B, CCK receptors postpones satiety in rhesus monkeys. American Journal of Physiology 1993; 265: R620–R624. 38. Reidelberger RD & O’Rourke MF. Potent cholecystokinin antagonist L-364,718 stimulates food intake in rats. American Journal of Physiology 1989; 257: R1512–R1518. 39. Kaplan JM & Moran TH, Gastrointestinal signaling in the control of food intake Handbook of Behavioral Neurobiology. Neurobiology of Food and Fluid Intake, vol. 4, no. 2. New York: Kluwer Academic/Plenum Publishing, 2004, pp. 273–303. 40. Smith GP (ed.) Satiation: from Gut to Brain. New York: Oxford University Press, 1998. 41. Strader AD & Woods SC. and SC Gastrointestinal hormones and food intake. Gastroenterology 2004; in press. *42. Woods SC. Gastrointestinal satiety signals I. An overview of gastrointestinal signals that influence food intake. American Journal of Physiology 2004; 286: G7–G13. 43. Stein LJ & Woods SC. Gastrin releasing peptide reduces meal size in rats. Peptides 1982; 3: 833–835. 44. Ladenheim EE, Wirth KE & Moran TH. Receptor subtype mediation of feeding suppression by bombesin-like peptides. Pharmacology and Biochemistry of Behavior 1996; 54: 705–711. 45. Okada S, York DA, Bray GA & Erlanson-Albertsson C. Enterostatin (Val-Pro-Asp-Pro-Arg), the activation peptide of procolipase, selectively reduces fat intake. Physiology and Behavior 1991; 49: 1185–1189. 46. Shargill NS, Tsuji S, Bray GA & Eralson-Albertsson C. Enterostatin suppresses food intake following injection into the third ventricle of rats. Brain Research 1991; 544: 137–140. 47. Lotter EC, Krinsky R, McKay JM et al. Somatostatin decreases food intake of rats and baboons. Journal of Comparative Physiology and Psychology 1981; 95: 278–287. 48. Larsen PJ, Fledelius C, Knudsen LB & Tang-Christensen M. Systemic administration of the long-acting GLP-1 derivative NN2211 induces lasting and reversible weight loss in both normal and obese rats. Diabetes 2001; 50: 2530–2539. 49. Naslund E, Barkeling B, King N et al. Energy intake and appetite are suppressed by glucagon-like peptide1 (GLP-1) in obese men. International Journal of Obesity and Related Metabolic Disorders 1999; 23: 304–311. 50. Fujimoto K, Machidori H, Iwakari R et al. Effect of intravenous administration of apolipoprotein A-IV on patterns of feeding, drinking and ambulatory activity in rats. Brain Research 1993; 608: 233–237. 51. Batterham RL, Cowley MA, Small CJ et al. Gut hormone PYY(3-36) physiologically inhibits food intake. Nature 2002; 418: 650–654. 52. Chance WT, Balasubramaniam A, Zhang FS et al. Anorexia following the intrahypothalamic administration of amylin. Brain Research 1991; 539: 352–354. 53. Lutz TA, Del Prete E & Scharrer E. Reduction of food intake in rats by intraperitoneal injection of low doses of amylin. Physiology and Behavior 1994; 55: 891–895.

Regulation of energy homeostasis by peripheral signals 511 54. Geary N. Glucagon and the control of meal size. In Smith GP (ed.) Satiation. From Gut to Brain. New York: Oxford University Press, 1998, pp. 164–197. 55. Salter JM. Metabolic effects of glucagon in the Wistar rat. American Journal of Clinical Nutrition 1960; 8: 535–539. 56. Davison JS & Clarke GD. Mechanical properties and sensitivity to CCK of vagal gastric slowly adapting mechanoreceptors. American Journal of Physiology 1988; 255(1 Pt 1): G55–G61. 57. Lorenz DN & Goldman SA. Vagal mediation of the cholecystokinin satiety effect in rats. Physiology and Behaviour 1982; 29: 599–604. 58. Moran TH, Baldessarini AR, Salorio CF et al. Vagal afferent and efferent contributions to the inhibition of food intake by cholecystokinin. American Journal of Physiology 1997; 272(4 Pt 2): R1245–R1251. 59. Geary N, Sauter JL & Noh U. Glucagon acts in the liver to control spontaneous meal size in rats. American Journal of Physiology 1993; 264: R116–R122. 60. Langhans W. Role of the liver in the metabolic control of eating: what we know—and what we do not know. Neuroscience and Biobehavioral Reviews 1996; 20: 145–153. *61. Lutz TA, Del Prete E & Scharrer E. Subdiaphragmatic vagotomy does not influence the anorectic effect of amylin. Peptides 1995; 16: 457–462. 62. Lutz TA, Senn M, Althaus J et al. Lesion of the area postrema/nucleus of the solitary tract (AP/NTS) attenuates the anorectic effects of amylin and calcitonin gene-related peptide (CGRP) in rats. Peptides 1998; 19: 309–317. 63. Edwards GL, Ladenheim EE & Ritter RC. Dorsomedial hindbrain participation in cholecystokinininduced satiety. American Journal of Physiology 1986; 251: R971–R977. 64. Moran TH, Ladenheim EE & Schwartz GJ. Within-meal gut feedback signaling. International Journal of Obesity and Related Metabolic Disorders 2001; 25(supplement 5): S39–S41. 65. Moran TH & Kinzig KP. Gastrointestinal satiety signals. II. Cholecystokinin. American Journal of Physiology Gastrointestinal and Liver Physiology 2004; 286: G183–G188. 66. Rinaman L, Hoffman GE, Dohanics J et al. Cholecystokinin activates catecholaminergic neurons in the caudal medulla that innervate the paraventricular nucleus of the hypothalamus in rats. Journal of Comparative Neurology 1995; 360: 246–256. 67. Cummings DE, Purnell JQ, Frayo RS et al. A preprandial rise in plasma ghrelin levels suggests a role in meal initiation in humans. Diabetes 2001; 50: 1714–1719. 68. Asakawa A, Inui A, Kaga T et al. Ghrelin is an appetite-stimulatory signal from stomach with structural resemblance to motilin. Gastroenterology 2001; 120: 337–345. 69. Tscho¨p M, Smiley DL & Heiman ML. Ghrelin induces adiposity in rodents. Nature 2000; 407: 908–913. 70. Wren AM, Small CJ, Ward HL et al. The novel hypothalamic peptide ghrelin stimulates food intake and growth hormone secretion. Endocrinology 2000; 141: 4325–4328. 71. Wren AM, Seal LJ, Cohen MA et al. Ghrelin enhances appetite and increases food intake in humans. Journal of Clinical Endocrinology and Metabolism 2001; 86: 5992–5995. 72. Cowley MA, Smith RG, Diano S et al. The distribution and mechanism of action of ghrelin in the CNS demonstrates a novel hypothalamic circuit regulating energy homeostasis. Neuron 2003; 37: 649–661. 73. Date Y, Murakami N, Toshinai K et al. The role of the gastric afferent vagal nerve in ghrelin-induced feeding and growth hormone secretion in rats. Gastroenterology 2002; 123: 1120–1128. 74. Woods SC. The eating paradox: how we tolerate food. Psychological Reviews 1991; 98: 488–505. 75. West DB, Fey D & Woods SC. Cholecystokinin persistently suppresses meal size but not food intake in free-feeding rats. American Journal of Physiology 1984; 246: R776–R787. 76. West DB, Greenwood MRC, Marshall KA & Woods SC. Lithium chloride, cholecystokinin and meal patterns: evidence that cholecystokinin suppresses meal size in rats without causing malaise. Appetite 1987; 8: 221–227. 77. Moran TH, Katz LF, Plata-Salaman CR & Schwartz GJ. Disordered food intake and obesity in rats lacking cholecystokinin A receptors. American Journal of Physiology 1998; 274(3 Pt 2): R618–R625. *78. Woods SC & Seeley RJ. Adiposity signals and the control of energy homeostasis. Nutrition 2000; 16: 894–902. 79. Shulman GI. Cellular mechanisms of insulin resistance in humans. American Journal of Cardiology 1999; 84: 3J–10J. 80. Kennedy GC. The role of depot fat in the hypothalamic control of food intake in the rat. Proceedings of the Royal Society of London (Biology) 1953; 140: 579–592.

512 S. C. Woods et al 81. Bagdade JD, Bierman EL & Porte Jr. D. The significance of basal insulin levels in the evaluation of the insulin response to glucose in diabetic and nondiabetic subjects. Journal of Clinical Investigation 1967; 46: 1549–1557. 82. Polonsky KS, Given E & Carter V. Twenty-four-hour profiles and pulsatile patterns of insulin secretion in normal and obese subjects. Journal of Clinical Investigation 1988; 81: 442–448. 83. Havel PJ. Mechanisms regulating leptin production: implications for control of energy balance. American Journal of Clinical Nutrition 1999; 70: 305–306. 84. Havel PJ. Peripheral signals conveying metabolic information to the brain: short-term and long-term regulation of food intake and energy homeostasis. Experimental Biology and Medicine (Maywood) 2001; 226: 963–977. 85. Baskin DG, Figlewicz-Lattemann D, Seeley RJ et al. Insulin and leptin: dual adiposity signals to the brain for the regulation of food intake and body weight. Brain Research 1999; 848: 114–123. 86. Niswender KD & Schwartz MW. Insulin and leptin revisited: adiposity signals with overlapping physiological and intracellular signaling capabilities. Frontiers in Neuroendocrinology 2003; 24: 1–10. *87. Porte DJ, Baskin DG & Schwartz MW. Leptin and insulin action in the central nervous system. Nutrition Reviews 2002; 60: S20–S29. 88. Banks WA, Kastin AJ & Huang W. Leptin enters the brain by a saturable system independent of insulin. Peptides 1996; 17: 305–311. 89. Banks WA & Kastin AJ. Differential permeability of the blood–brain barrier to two pancreatic peptides: insulin and amylin. Peptides 1998; 19: 883–889. 90. Banks WA. The source of cerebral insulin. European Journal of Pharmacology 2004; 490: 5–12. 91. Baura G, Foster DM, Kaiyala K et al. Insulin transport from plasma into the central nervous system is inhibited by dexamethasone in dogs. Diabetes 1996; 45: 86–90. 92. Baura G, Foster D, Porte Jr. D et al. Saturable transport of insulin from plasma into the central nervous system of dogs in vivo: a mechanism for regulated insulin delivery to the brain. Journal of Clinical Investigation 1993; 92: 1824–1830. 93. Schwartz MW, Peskind E, Raskind M et al. Cerebrospinal fluid leptin levels: relationship to plasma levels and to adiposity in humans. Nature Medicine 1996; 2: 589–593. 94. Schwartz MW, Bergman RN, Kahn SE et al. Evidence for uptake of plasma insulin into cerebrospinal fluid through an intermediate compartment in dogs. Journal of Clinical Investigation 1991; 88: 1272–1281. 95. Woods SC, Seeley RJ, Baskin DG & Schwartz MW. Insulin and the blood–brain barrier. Current Pharmaceutical Design 2003; 9: 795–800. 96. Baskin DG, Schwartz MW, Seeley RJ et al. Leptin receptor long form splice variant protein expression in neuron cell bodies of the brain and colocalization with neuropeptide Y mRNA in the arcuate nucleus. Journal of Histochemistry and Cytochemistry 1999; 47: 353–362. 97. Baskin DG, Hahn TM & Schwartz MW. Leptin sensitive neurons in the hypothalamus. Hormone and Metabolic Research 1999; 31: 345–350. 98. Baskin DG, Marks JL, Schwartz MW et al. Insulin and insulin receptors in the brain in relation to food intake and body weight. In Lehnert H et al (ed.) Endocrine and Nutritional Control of Basic Biological Functions. Stuttgart: Hogrefe & Huber, 1990, pp. 202–222. 99. Havrankova J, Roth J & Browstein M. Insulin receptors are widely distributed in the central nervous system of the rat. Nature 1978; 272: 827–829. 100. Corp ES, Woods SC, Porte Jr. D et al. Localization of 125I-insulin binding sites in the rat hypothalamus by quantitative autoradiography. Neuroscience Letters 1986; 70: 17–22. 101. Schwartz MW, Seeley RJ, Campfield LA et al. Identification of hypothalmic targets of leptin action. Journal of Clinical Investigation 1996; 98: 1101–1106. 102. Tartaglia LA. The leptin receptor. Journal of Biological Chemistry 1997; 272: 6093–6096. 103. Benoit SC, Air EL, Coolen LM et al. The catabolic action of insulin in the brain is mediated by melanocortins. Journal of Neuroscience 2002; 22: 9048–9052. 104. Baskin DG, Breininger JF & Schwartz MW. Leptin receptor mRNA identifies a subpopulation of neuropeptide Y neurons activated by fasting in rat hypothalamus. Diabetes 1999; 48: 828–833. 105. Baskin DG, Sipols AJ, Schwartz MW & White MF. Insulin receptor substrate-1 (IRS-1) expression in rat brain. Endocrinology 1994; 134: 1952–1955. 106. Coleman DL. Effects of parabiosis of obese with diabetes and normal mice. Diabetologia 1973; 9: 294–298.

Regulation of energy homeostasis by peripheral signals 513 107. Coleman DL. Obese and diabetes: two mutant genes causing diabetes–obesity syndromes in mice. Diabetologia 1978; 14: 141–148. 108. Sipols AJ, Baskin DG & Schwartz MW. Effect of intracerebroventricular insulin infusion on diabetic hyperphagia and hypothalamic neuropeptide gene expression. Diabetes 1995; 44: 147–151. 109. Sindelar DK, Havel PJ, Seeley RJ et al. Low plasma leptin levels contribute to diabetic hyperphagia in rats. Diabetes 1999; 48: 1275–1280. 110. Campfield LA, Smith FJ, Gulsez Y et al. Mouse OB protein: evidence for a peripheral signal linking adiposity and central neural networks. Science 1995; 269: 546–549. 111. McGowan MK, Andrews KM & Grossman SP. Chronic intrahyphothalamic infusions of insulin or insulin antibodies alter body weight and food intake in the rat. Physiology and Behaviour 1992; 51: 753–766. 112. Strubbe JH & Mein CG. Increased feeding in response to bilateral injection of insulin antibodies in the VMH. Physiology and Behaviour 1977; 19: 309–313. 113. Cohen P, Zhao C, Cai X et al. Selective deletion of leptin receptor in neurons leads to obesity. Journal of Clinical Investigation 2001; 108: 1113–1121. *114. Bru¨ning JC, Gautam D, Burks DJ et al. Role of brain insulin receptor in control of body weight and reproduction. Science 2000; 289: 2122–2125. *115. Obici S, Feng Z, Karkanias G et al. Decreasing hypothalamic insulin receptors causes hyperphagia and insulin resistance in rats. Nature Neuroscience 2002; 5: 566–572. 116. Dua A, Hennes MI, Hoffman RG et al. Leptin: a significant indicator of total body fat but not of visceral fat and insulin insensitivity in African-American women. Diabetes 1996; 45: 1635–1637. 117. Wajchenberg BL. Subcutaneous and visceral adipose tissue: their relation to the metabolic syndrome. Endocrine Reviews 2000; 21(6): 697–738. 118. Cigolini M, Seidell JC, Targher G et al. Fasting serum insulin in relation to components of the metabolic syndrome in European healthy men: the European fat distribution study. Metabolism 1995; 44: 35–40. 119. Bjorntorp P. Body fat distribution, insulin reistance, and metabolic diseases. Nutrition 1997; 13: 795–803. 120. Despres JP. The insulin-resistance-dyslipidemic syndrome of visceral obesity: effect on patients’ risk. Obesity Research 1998; 6(supplement 1): 8S–17S. 121. Clegg DJ, Riedy CA, Smith KA et al. Differential sensitivity to central leptin and insulin in male and female rats. Diabetes 2003; 52: 682–687. 122. Schwartz MW, Sipols AJ, Marks JL et al. Inhibition of hypothalamic neuropeptide Y gene expression by insulin. Endocrinology 1992; 130: 3608–3616. 123. Seeley R, Yagaloff K, Fisher S et al. Melanocortin receptors in leptin effects. Nature 1997; 390: 349. 124. Bates SH, Stearns WH, Dundon TA et al. STAT3 signalling is required for leptin regulation of energy balance but not reproduction. Nature 2003; 421: 856–859. 125. Kloek C, Haq AK, Dunn SL et al. Regulation of Jak kinases by intracellular leptin receptor sequences. Journal of Biological Chemistry 2002; 277: 41547–41555. 126. Niswender KD, Morton GJ, Stearns WH et al. Intracellular signalling. Key enzyme in leptin-induced anorexia. Nature 2001; 413: 794–795. 127. Niswender KD, Morrison CD, Clegg DJ et al. Insulin activation of phosphatidylinositol 3-kinase in the hypothalamic arcuate nucleus: a key mediator of insulin-induced anorexia. Diabetes 2003; 52: 227–231. 128. Spanswick D, Smith MA, Groppi VE et al. Leptin inhibits hypothalamic neurons by activation of ATPsensitive potassium channels. Nature 1997; 390: 521–525. 129. Spanswick D, Smith MA, Groppi VE et al. Insulin activates ATP-sensitive KCchannels in hypothalamic neurons of lean, but not obese rats. Nature Neuroscience 2000; 3: 757–758. 130. Ikeda H, West DB, Pustek JJ et al. Intraventricular insulin reduces food intake and body weight of lean but not obese Zucker rats. Appetite 1986; 7: 381–386. 131. Scherer PE, Williams S, Fogliano M et al. A novel serum protein similar to C1q, produced exclusively in adipocytes. Journal of Biological Chemistry 1995; 26746–26749. 132. Arita Y, Kihara S, Ouchi N et al. Paradoxical decrease of an adipose-specific protein, adiponectin, in obesity. Biochemistry and Biophysics Research Communications 1999; 257: 79–83. 133. Atzmon G, Yang XM, Muzumdar R et al. Differential gene expression between visceral and subcutaneous fat depots. Hormones and Metabolic Research 2002; 34: 622–628.

514 S. C. Woods et al 134. Wajchenberg BL, Giannella-Neto D, da Silva ME & Santos RF. Depot-specific hormonal characteristics of subcutaneous and visceral adipose tissue and their relation to the metabolic syndrome. Hormone and Metabolic Research 2002; 34: 616–621. *135. Qi Y, Takahashi N, Hileman SM et al. Adiponectin acts in the brain to decrease body weight. Nature Medicine 2004; 10: 524–529. 136. Seeley RJ, D’Alessio DA & Woods SC. Fat hormones pull their weight in the CNS. Nature Medicine 2004; 10: 454–455. 137. Birch LL, Johnson SL, Andresen G et al. The variability of young children’s energy intake. New England Journal of Medicine 1991; 324: 232–235. 138. de Castro JM. Prior day’s intake has macronutrient-specific delayed negative feedback effects on the spontaneous food intake of free-living humans. Journal of Nutrition 1998; 128: 61–67. 139. Gasnier A & Mayer A. Recherche sur la re´gulation de la nutrition. II. Me´canismes re´gulateurs de la nutrition chez le lapin domestique. Annals Physiologie Physicoichemie et Biologie 1939; 15: 157–185. 140. Collier G. The dialogue between the house economist and the resident physiologist. Nutrition and Behavior 1986; 3: 9–26. 141. Collier GH, Johnson DF & Mitchell C. The relation between meal size and the time between meals: effects of cage complexity and food cost. Physiology and Behavior 1999; 67: 339–346. 142. Woods SC. The house economist and the eating paradox. Appetite 2002; 38: 161–165. 143. Barrachina MD, Martinez V, Wang L et al. Synergistic interaction between leptin and cholecystokinin to reduce short-term food intake in lean mice. Proceedings of the National Academy of Science of the USA 1997; 94: 10455–10460. 144. Figlewicz DP, Sipols AJ, Seeley RJ et al. Intraventricular insulin enhances the meal-suppressive efficacy of intraventricular cholecystokinin octapeptide in the baboon. Behavioral Neuroscience 1995; 109: 567–569. 145. Matson CA, Wiater MF, Kuijper JL & Weigle DS. Synergy between leptin and cholecystokinin (CCK) to control daily caloric intake. Peptides 1997; 18: 1275–1278. 146. Matson CA, Reid DF, Cannon TA & Ritter RC. Cholecystokinin and leptin act synergistically to reduce body weight. American Journal of Physiology 2000; 278: R882–R890. 147. Riedy CA, Chavez M, Figlewicz DP & Woods SC. Central insulin enhances sensitivity to cholecystokinin. Physiology and Behavior 1995; 58: 755–760. 148. Rushing PA, Lutz TA, Seeley RJ & Woods SC. Amylin and insulin interact to reduce food intake in rats. Hormone and Metabolic Research 2000; 32: 62–65. 149. Richardson RD, Omachi K, Kermani R & Woods SC. Intraventricular insulin potentiates the anorexic effect of corticotropin releasing hormone in rats. American Journal of Physiology 2002; 283: R1321– R1326. 150. Schwartz GJ & Moran TH. Sub-diaphragmatic vagal afferent integration of meal-related gastrointestinal signals. Neuroscience and Biobehavioral Reviews 1996; 20: 47–56. 151. Schwartz GJ, Moran TH, White WO & Ladenheim EE. Relationships between gastric motility and gastric vagal afferent responses to CCK and GRP in rats differ. American Journal of Physiology 1997; 272(6 Pt 2): R1726–R1733. 152. Grill HJ & Kaplan JM. The neuroanatomical axis for control of energy balance. Frontiers in Neuroendocrinology 2002; 23: 2–40. 153. Ahima RS, Saper CB, Flier JS & Elmquist JK. Leptin regulation of neuroendocrine systems. Frontiers of Neuroendocrinology 2000; 21: 263–307. 154. Cone RD, Cowley MA, Butler AA et al. The arcuate nucleus as a conduit for diverse signals relevant to energy homeostasis. International Journal of Obesity and Related Metabolic Disorders 2001; 25(supplement 5): S63–S67. 155. Elmquist JK, Elias CF & Saper CB. From lesions to leptin: hypothalamic control of food intake and body weight. Neuron 1999; 22: 221–232. 156. Ainscow EK, Mirshamsi S, Tang T et al. Dynamic imaging of free cytosolic ATP concentration during fuel sensing by rat hypothalamic neurones: evidence for ATP-independent control of ATP-sensitive K(C) channels. Journal of Physiology 2002; 544: 429–445. 157. Even P & Nicolaidis S. Spontaneous and 2DG-induced metabolic changes and feeding: the ischymetric hypothesis. Brain Research Bulletin 1985; 15: 429–435.

Regulation of energy homeostasis by peripheral signals 515 158. Nicolaidis S & Even P. Mesure du me´tabolisme de fond en relation avec la prise alimentaire: hypothese iscyme´trique. Comptes Rendus Academie de Sciences, Paris 1984; 298: 295–300. 159. Kim EK, Miller I, Landree LE et al. Expression of FAS within hypothalamic neurons: a model for decreased food intake after C75 treatment. American Journal of Physiology 2002; 283: E867–E879. 160. Clegg DJ, Wortman MD, Benoit SC, McOsker CC & Seeley RJ. Comparison of central and peripheral administration of C75 on food intake, body weight, and conditioned taste aversion. Diabetes 2002; 51: 3196–3201. 161. Kumar MV, Shimokawa T, Nagy TR & Lane MD. Differential effects of a centrally acting fatty acid synthase inhibitor in lean and obese mice. Proceedings of the National Academy of Sciences of the USA 2002; 99: 1921–1925. *162. Loftus TM, Jaworsky DE, Frehywot GL et al. Reduced food intake and body weight in mice treated with fatty acid synthase inhibitors. Science 2000; 288: 2299–2300. 163. Obici S, Feng Z, Arduini A et al. Inhibition of hypothalamic carnitine palmitoyltransferase-1 decreases food intake and glucose production. Nature Medicine 2003; 9: 756–761. 164. Wortman MD, Clegg DJ, D’Alessio D et al. C75 inhibits food intake by increasing CNS glucose metabolism. Nature Medicine 2003; 9: 483–485. 165. Obici S, Feng Z, Morgan K et al. Central administration of oleic acid inhibits glucose production and food intake. Diabetes 2002; 51: 271–275. 166. Nicolaidis S. Mecanisme nerveux de l’equilibre energetique. Journe´es Annuelles de Diabetologie de l’HotelDieu 1978; 1: 152–156. 167. Oomura Y, Ono T, Ooyama H & Wayner MJ. Glucose and osmosensitive neurones of the rat hypothalamus. Nature 1969; 222: 282–284. 168. Oomura Y, Ooyama H, Sugimori M et al. Glucose inhibition of the glucose-sensitive neurone in the rat lateral hypothalamus. Nature 1974; 247: 284–286. 169. Oomura Y. Glucose as a regulator of neuronal activity. Advances in Metabolic Disorders 1983; 10: 31–65. 170. Harvey J & Ashford ML. Leptin in the CNS: much more than a satiety signal. Neuropharmacology 2003; 44: 845–854. 171. Levin BE, Dunn-Meynell AA & Routh VH. Brain glucose sensing and body energy homeostasis: role in obesity and diabetes. American Journal of Physiology 1999; 276: R1223–R1231. 172. Levin BE. Glucosensing neurons: the metabolic sensors of the brain? Diabetes Nutrition and Metabolism 2002; 15: 274–280. 173. Cone RD. The central melanocortin system and energy homeostasis. Trends in Endocrinology and Metabolism 1999; 10: 211–216. 174. Cone RD (ed.) The Melanocortin Receptors. New Jersey: Humana Press, 2000. 175. Seeley RJ, Drazen DL & Clegg DJ. The critical role of the melanocortin system in the control of energy balance. Annual Review of Nutrition 2004; 24: 133–149.