Appetite Regulation and Thermogenesis
The classification of body weight as a regulated physiologic parameter is relatively novel. While obesity, including morbid obesity, has been recognized for thousands of years, the possibility that body weight was determined by a complex interaction between internal regulatory systems and the environment first received scientific attention in the mid-20th century. Studies of brain-lesioned animals indicated that disruption of the ventromedial hypothalamus produced a syndrome of obesity and hyperphagia,1–3 and ablation of the lateral hypothalamus resulted in aphagia, adipsia, and dramatic weight loss,1,4 suggesting that these areas were critical to the maintenance of energy balance. Findings also emerged demonstrating that obesity could be caused by administration of goldthioglucose,5 which damaged the ventromedial nucleus of the hypothalamus (VMH).6 Subsequent studies using monosodium glutamate–induced obesity indicated that the arcuate nucleus also played a role in maintaining energy balance.7 A series of parabiosis experiments performed between genetically obese mice, ob/ob and db/db, suggested that circulating factors might play a role in determining adiposity.8,9 In humans, the potential role of the hypothalamus in obesity came from the observation that patients with craniopharyngioma frequently became obese following resection of the tumor.10
Despite significant efforts along a number of lines of investigation which resulted in increased understanding of obesity, a full appreciation of obesity as an endocrine syndrome did not arrive until the discovery of leptin gene in 199411 and its receptor.12 Mutations in either gene served to explain two mouse obesity syndromes: the ob/ob mouse, which lacks the hormone leptin, and the db/db mouse, which lacks the long form of the leptin receptor. Recognition of the fact that mutations in these molecules led to similar phenotypes in humans rapidly followed and confirmed their importance.13,14 These discoveries led to a significant paradigm shift in the understanding of obesity and the nature of the fat cell. Existence of a hormone-specific obesity syndrome made it clear that body weight was subject to physiologic regulation. Furthermore, identification of the adipocyte as the source of the hormone changed the perception of the fat cell from that of a passive depot of energy stores to a regulator important to overall energy homeostasis.15,16 This overall shift redefined understanding of the processes involved in regulating overall energy balance in mammalian organisms, including humans. As a consequence, a sophisticated understanding of the interconnection of multiple organ systems in the brain and periphery and their interaction with the environment is currently evolving.
Components of Feeding
Intake of calories occurs for multiple reasons (Fig. 1-1). Perhaps the most important is the net caloric deficit that begins after the digestion of a meal. As calories are used, the increasing deficit that accrues eventually leads to hunger, food seeking, and food ingestion. This component of feeding may be called “homeostatic hunger” because it reflects a true metabolic deficit, and consumption is aimed at maintaining energy stores. Since this activity is critical to survival, it appears to be linked to reward pathways so that animals will “work” for food.17 For example, feeding involves flavor and texture and thus engages gustatory pathways that involve taste and smell. The rewards associated with ingestion of palatable flavors lead to eating past the point of metabolic repletion.18 Food variety also appears to engage reward pathways; increased variety prevents malnutrition. However, availability of an increased variety of energy-dense foods is associated with obesity.19 Motivation and reward pathways have been extensively explored with regard to drug addiction. While addiction to drugs of abuse has no homeostatic value, reinforcement of food rewards may occur through similar pathways. For example, agents that alter opioid and dopaminergic signals also act to modulate motivation for palatable food.20 Furthermore, eating is also linked to stress, which may predispose to hyperphagia in an environment where palatable food is readily available.19
Components of Energy Expenditure
Traditional models categorize energy expenditure (EE) into basal (obligate) and adaptive (or facultative) thermogenesis (AT). Obligate EE includes all pathways involved in the maintenance of basic metabolic and physiologic processes and is also referred to as resting metabolic rate (RMR) or more recently resting energy expenditure (REE). RMR includes both sleeping metabolic rate (SMR) and the increase in metabolic rate that is seen with arousal. AT includes cold and diet-induced thermogenesis. The cellular mechanisms that regulate obligatory and adaptive thermogenesis are often similar. Finally, physical activity represents a third category which has at least two components, nonexercise activity thermogenesis (NEAT) and sports-like exercise.21 These distinct categories of EE are in fact only approximate, and regulatory mechanisms are overlapping (Fig. 1-2). For example, thyroid hormone (TH) is required for up to 30% of basal EE, and adaptive increases in TH are required for normal cold-induced thermogenesis.22,23 Furthermore, physical activity (PA) can have long-lasting effects on REE, and PA may be enacted by stimuli that are traditionally considered stimulants of facultative thermogenesis, such as caloric excess.24 Approximate contributions for the components of energy expenditure are REE, 70%; PA, 20%; and facultative, 10%, with PA representing the most variable component.25
FIGURE 1-2 Approximate contribution of components of energy expenditure. Thermic effect of food (TEF) and resting metabolic rate (RMR) can be measured using a ventilatory hood. RMR and nonexercise activity thermogenesis (NEAT) can be measured in a respiratory chamber. Total energy expenditure can be measured using doubly labeled water. SMR, Sleeping metabolic rate.
Integration of Energy Balance
Inputs from a number of neuropeptides and neurotransmitters in the brain as well as peripheral signals integrate information to mediate energy balance.26 The interactions and pathways engaged by these signals are already complex, and understanding of the pathways is still evolving. Furthermore, it is not clear that all possible important signals have been identified. One interesting aspect of the signals involved is that many regulate both energy intake and energy expenditure in a coordinated fashion. Although a correlation of obesity with decreased sympathetic activity has been long recognized,27,28 the potential pathways regulating energy expenditure were thought to be separate from those regulating feeding and satiety. However, it is now recognized that peptides regulating appetite also play a role in regulating energy expenditure in an inverse manner. Thus, peptides that stimulate feeding decrease energy expenditure, promoting energy storage, while those that inhibit feeding increase energy expenditure (Table 1-1).
In the brain, both neurotransmitters and neuropeptides have complex functions.29 The role of neurotransmitters in regulating feeding behavior was recognized before the role of neuropeptides was appreciated. However, mechanisms of these neurotransmitters have been more difficult to define; they may act through multiple receptors, and effects may vary depending on the anatomic area targeted. Monoamine neurotransmitters may be stimulatory or inhibitory. Glutamate and γ-aminobutyric acid (GABA), which are the most abundant neurotransmitters in the hypothalamus, act to increase feeding, and some of the neurons expressing orexigenic neuropeptides appear to be GABAergic.30 One view of the interaction of transmitters and peptides is that peptides act as essential modulators of GABA and glutamate action.31 Serotonin, which acts through multiple receptors, is inhibitory,32 and some of these receptors are being considered as specific pharmacologic targets for the treatment of obesity.33 The roles of epinephrine, norepinephrine, and dopamine are more complex, and these transmitters may act to either stimulate or inhibit feeding.34,35 Although the mechanism of action is poorly understood, these pathways are the targets of the limited pharmacologic therapies that are currently available for the treatment of obesity. Biogenic amines currently in use include phentermine and sibutramine. Phentermine, an analog of amphetamine, acts to increase catecholamine release in the paraventricular nucleus of the hypothalamus. Mazindol has a similar mechanism of action. Sibutramine, which acts through its active metabolites, prevents reuptake of 5-HT but does not cause release. Sibutramine also inhibits noradrenaline reuptake.36
Recently, the importance of interaction of peptidergic systems with neurotransmitters has been recognized as playing a role. For example, leptin37,38 and MCH39,40 appear to regulate dopaminergic tone. In the arcuate, GABAergic AgRP/NPY neurons play a critical role in regulating pro-opiomelanocortin (POMC) neurons. Although ablation of AgRP or NPY individually or both peptides in combination does not result in an identifiable phenotype,41 ablation of the neurons in adult mice leads to rapid starvation.42 Consistent with the role of GABA regulating POMC neurons, mice with impaired GABA release are lean and resistant to diet-induced obesity.43
In addition to leptin, other signals from the periphery add to the complex pathways that are involved in the regulation of body weight. This includes multiple signals from the gut44,45 (Fig. 1-3). Cholecystokinin (CCK) is synthesized in the duodenum and jejunum and was recognized as a peptide capable of inhibiting appetite as early as 1973.46 CCK acts in the hindbrain to reduce meal size and duration.47 Peptide YY (PYY) is secreted by the distal portion of the gastrointestinal tract, and in addition to inhibiting gastric emptying,48 it also crosses the blood-brain barrier to act on arcuate nuclei and inhibit feeding.44 However, the role of PYY remains controversial, because when injected into the lateral ventricles of animals, PYY acts to increase food intake, and not all investigators have been able to reproduce the satiety effects.49 Glucagon-like peptide 1 (GLP-1) and oxyntomodulin are products of the preproglucagon gene and are synthesized in the gut and brain. Both act to inhibit food intake, through different mechanisms.50 Thus far, the only gut peptide known to stimulate appetite is ghrelin, which is secreted by the stomach and also acts on neurons in the arcuate nucleus.51
FIGURE 1-3 Schema of selected gut-to-brain signals that may play a role in mediating energy balance. A number of peptides from the gut play a role in mediating appetite and energy expenditure. Only one, ghrelin from the stomach, is orexigenic. All segments of the gut may also contribute to signals from vagal afferents. Finally, information of oxidation of fatty acids in the liver may also be transmitted by the vagus and play a role in mediating appetite. Within the brain, signaling acquires additional complexity; networks between the hypothalamus and hindbrain and cortex are all involved in regulating both intake and output.
Other factors derived from the gut that may play a role in appetite include fatty acid amides. For example, the cannabinoid receptor agonist, anandamide, is an ethanolamide synthesized in the gut and possibly regulated by nutritional status, since local concentrations increase with fasting.52 Anandamide may act centrally through cannabinoid receptors, which are known to promote appetite and increase energy homeostasis.53 In contrast oleoylethanolamide, which is also synthesized in the gut, acts to inhibit food intake and stimulate lipolysis through the activation of the peroxisome proliferator–activated receptor alpha (PPAR-α).54 Thus fatty acid–derived molecules may serve as an additional pathway of gut-mediated regulation of appetite.
Peripheral signals regarding the state of the gastrointestinal tract may also be integrated by the vagus nerve,55 which sends afferents to multiple brain areas, including the dorsal motor nucleus and the nucleus of the solitary tract. These areas appear to be involved in responses to neuropeptides, including ghrelin56 and α-MSH. The vagus may also play a role in conveying information on fatty acid oxidation in the liver. Fatty acid oxidation appears to play a role in mediating appetite; inhibition of this process is associated with increased appetite.57,58
Insulin may also play a role in inhibiting appetite, and it is known that neurons in the arcuate express insulin receptors59 and respond to insulin. Female mice lacking insulin receptor expression in the brain eat more than normal animals, and both genders develop mild obesity when placed on a high-fat diet.60 Insulin and glucose may play a role in meal initiation and meal termination.61
Thus, body weight is regulated by a complex interaction of signals involving both the gut and the brain. Additional complexity derives from the fact that these signals act through specific receptors. The receptors have anatomic-specific expression. Furthermore, some involve relatively large receptor families, as seen with the melanocortin receptors (see The Melanocortin System: α-MSH, Agouti-Related Peptide, and Central Melanocortin Receptors). For some of these pathways, the finding of spontaneous mutations associated with obesity in human populations has provided proof that body weight is regulated similarly in humans as in rodents (Table 1-2).
Table 1-2
Sample of Characterized Mutations Leading to Obesity in Humans
| Mutation | Reference |
| Leptin | 62, 13 |
| Leptin receptor | 14 |
| Melanocortin 4 receptor | 94, 95, 205 |
| Melanocortin 3 receptor | 206 |
| Prohormone convertase 1 | 207 |
| PPAR-γ | 208 |
| POMC | 96 |
POMC, Pro-opiomelanocortin; PPAR, peroxisome proliferator–activated receptor.
Specific Hormones and Neuropeptides
Leptin, identified through positional cloning of the ob gene,11 is a 167-amino-acid peptide hormone secreted by adipocytes. It signals through a membrane receptor that has six splice variants and belongs to the class I cytokine receptor family.12 Leptin signaling is required, although not sufficient alone, for normal energy balance. Animals, including mice and humans, lacking leptin62 or the leptin receptor14 have a syndrome of severe hyperphagia and obesity. In the case of leptin mutations, leptin administration leads to a marked resolution of the syndrome in both ob/ob mice63,64 and in rare human patients with leptin mutations.65,66 However, in the vast majority of obese mammals, leptin levels are elevated, correlating well with available fat stores,67,68 and administration of peripheral leptin has little effect on appetite. These findings revised the perception of leptin. While the complete absence of leptin has major consequences on appetite, the incremental increases in leptin that are seen with increased adiposity have little effect on the continued ingestion of calories or the storage of calories as fat. In contrast, repletion of leptin with fasting leads to attenuation of many of the neuroendocrine changes seen with fasting.69 In human females, leptin replacement leads to some of the abnormalities seen in hypothalamic amenorrhea secondary to strenuous exercise or low body weight.70 Thus the critical physiologic role of leptin appears to be to signal caloric deficiency and thus mediate the appropriate metabolic changes rather than to signal caloric excess.
Leptin targets specific neurons in the brain, specifically in the hypothalamus, although leptin receptors are also seen in other areas, including the ventral tegmental area (VTA). The best-characterized neurons are the NPY/AgRP and POMC neurons in the arcuate.30,71 NPY and AgRP are both orexigenic (appetite-inducing) peptides synthesized by the same population of neurons. POMC is expressed in a different population of arcuate neurons that process the preprohormone to a number of peptides, including α-MSH, which acts to suppress appetite. To date, the leptin-to-arcuate pathway represents the best-characterized pathway involved in the regulation of body weight, especially insofar as mutations disrupting this pathway have also been shown to be important in human obesity as well as rodent obesity (Fig. 1-4).
FIGURE 1-4 The leptin pathway. Leptin, a hormone secreted by adipocytes, crosses the blood-brain barrier to act on neurons in the arcuate nucleus. One set of target neurons are those synthesizing prepro-opiomelanocortin (POMC). Leptin acts to stimulate these neurons to synthesize POMC and release one of the POMC gene products, α-melanocyte-stimulating hormone (α-MSH). This peptide mainly acts through the melanocortin 4 receptor (MC4R) to decrease feeding, and through this receptor and the melanocortin 3 receptor (MC3R) to increase energy expenditure. Mutations in this pathway lead to disruption of the appropriate signals and to obesity. Known mutations leading to obesity in humans are marked with asterisks. (See text and Table 1-2 for details.)
Adipocytes synthesize many biologically active proteins with potential endocrine function (reviewed in Ref. 58 and Chapter 11). These include cytokines, immune-related proteins, complement and complement-related proteins, enzymes involved in steroid metabolism, and proteins of the rennin-angiotensin system. Furthermore, receptors for traditional endocrine hormones, nuclear hormones, cytokines, and catecholamines are all expressed by adipose tissue. These peptides are likely to form causal links between obesity, insulin resistance, and cardiovascular disease. Two recently discovered peptides, adiponectin and resistin, may play a role in modulating insulin resistance. Adiponectin inversely correlates with insulin resistance, declines with obesity, and increases with weight loss.72 In contrast, resistin impairs glucose tolerance and insulin sensitivity, and secretion increases with increasing adiposity.73 They may also be involved in determining predisposition to obesity and responses to a high-fat diet. Recently, excess adiposity has been associated with finding increased expression of multiple inflammatory markers in fat tissue. Thus, expression of interleukin 1 (IL-1), IL-5, plasminogen activator inhibitor 1 (PAI-1), tumor necrosis factor (TNF), and suppressor of cytokine signaling 3 (SOCS3) are all increased in obesity.74 These factors play a role in the decrease in insulin sensitivity associated with obesity; however, it is unclear (and seems unlikely) that any of these factors have direct effects on either appetite or energy expenditure.
The Hypothalamus
The potential role of neuropeptides in feeding behavior was first suggested by studies indicating that NPY was synthesized by arcuate75 neurons and elicited a robust feeding response when injected intracerebroventricularly (ICV).76,77 Chronic infusion of NPY leads to obesity in rats.78,79 Furthermore, expression of NPY increases with fasting, indicating that neurons making NPY respond to peripheral signals, signaling the state of energy balance.80,81 Interestingly, ablation of the NPY gene was not associated with altered body weight or feeding,82 although when mice lacking NPY were bred to mice lacking leptin, substantial attenuation of the ob/ob obesity syndrome was noted.83 However, animals without NPY show an abnormal response to refeeding after short-term fasting84 and also show an attenuated feeding response to hypoglycemia.85
The Melanocortin System: α-MSH, Agouti-Related Peptide, and Central Melanocortin Receptors
Humans and rodents require an intact melanocortin system in order to maintain normal body weight. The effect of α-MSH to decrease appetite was described in the late 1980s.86 However, the key role of melanocortins in the physiology of energy balance was not appreciated until the molecular mechanism of obesity of the yellow Ay mouse was identified. In this model, obesity is secondary to a mutation in the gene encoding a protein, agouti, which mediates coat color and leads to ectopic expression of the protein in all tissues, including the central nervous system.87 Subsequently it was discovered that agouti protein acted on melanocortin receptors to block melanocyte-stimulating hormone (MSH) action.88,89 These findings led to speculation that another protein normally expressed in the brain might have an action similar to that of agouti and to the discovery of agouti-related peptide, AgRP,90 which is expressed in the hypothalamus and interacts with the central melanocortin receptors, MC3R and MC4R.91 Overexpression of AgRP recapitulated an obesity syndrome similar to that seen in the Ay mouse, as did disruption of the MC4R.92 Mice with targeted disruption of the MC3R demonstrated a small increase in body fat and feeding efficiency, suggesting that at least in rodents, MC4 plays the dominant role in energy homeostasis.
The profound effects caused by disruption of the melanocortin pathway stimulated a search of MC4R mutations in humans, especially in children with early-onset obesity and a strong family history of obesity. Several such mutations were readily identified,93,94 and currently it is estimated that 5% of persons with severe familial early-onset obesity have MC4 mutations.95 In humans, obesity has also been associated with mutations in the POMC gene, which encodes multiple transcripts. Disruption of this gene leads to deficiency in both adrenocorticotropic hormone (ACTH) and MSH, and patients present with adrenal insufficiency and obesity. Since MSH expression outside of the central nervous system mediates hair color, patients with POMC mutations will frequently also have red hair.96
Melanin-Concentrating Hormone
Melanin-concentrating hormone is a 19-amino-acid peptide synthesized in magnocellular neurons of the lateral hypothalamus; it plays a key role in maintaining energy balance in animals.97,98 The peptide structure and anatomic distribution is highly conserved, and the sequence is identical in rodents, sheep, and humans. When injected ICV in rats, MCH induces an acute increase in feeding behavior. Chronic infusions in mice lead to a syndrome of mild obesity associated with decreased energy expenditure.99 Deletion of both the MCH and the MCH receptor genes are associated with leanness.100–102 In the case of the receptor knockouts, leanness is secondary to increased expenditure, because animals without the receptor eat as much or more than wild-type animals. Deletion of MCH from mice lacking leptin leads to a marked attenuation of the obesity phenotype, which is secondary to changes in energy expenditure rather than feeding.103 Pharmacologic blockade of the MCH receptor leads to leanness and reduces meal size.104 Chronic infusions of MCH agonists also lead to obesity similar to that seen with MCH infusion.105 The importance of the MCH system has not been validated in humans, since a phenotype of MCH deficiency would present with a lean phenotype. However, the homology of MCH in all strains of mammals examined strongly suggests that MCH will play a role in humans.
The Gut
Ghrelin, produced in the stomach, was identified as the endogenous ligand for the receptor responsible for growth hormone secretion.106 Subsequently it was found to produce adiposity in rodents, an effect that is independent of its ability to stimulate growth hormone secretion.107 Although infusions of ghrelin induce hunger and increased feeding,108 endogenous levels are low in obese individuals and increase with weight loss.109 This rise is not seen after gastric bypass surgery, which may help to explain the success of this procedure in mediating weight loss in obese humans.110,111 Ghrelin levels are extremely high in the Prader-Willi syndrome of genetic obesity.112 Ghrelin is transported into the brain, where it acts to stimulate NPY/AgRP neurons in the arcuate nucleus and is thus part of a circuit mediating energy homeostasis involving the stomach and the hypothalamus.113
Peptide YY
Peptide YY is synthesized and secreted throughout the intestine, although concentrations are higher in the distal portion, particularly in the colon and rectum, and the 3-36 form crosses the blood-brain barrier. Food intake stimulates PYY release, and higher serum concentrations are seen after fatty meals. As release occurs prior to nutrients reaching the distal parts of the gastrointestinal tract, neural reflexes may act to stimulate release, possibly through the vagus. PYY 1-36 has structural similarity to NPY and binds with high affinity to all five NPY receptors; the 3-36 form binds preferentially to the Y2 receptor. PYY acts on both the intestine and the brain. In the intestine, it increases fluid absorption and delays gastric emptying. In the brain, the 3-36 form has substantial effects on appetite. When given intravenously to human volunteers, it reduces caloric intake and increases the sensation of satiety.114 Similar effects have been reported in rats115; however, this effect is controversial because other investigators have been unable to reproduce the satiating effect.49 ICV injection of PYY clearly increases feeding, presumably through targeting a different receptor subset.116
PYY levels are low in patients with morbid obesity. One report suggests that levels rise after weight loss secondary to gastric bypass surgery.117 This suggests a potential role of PYY in the treatment of obesity.
Glucagon-Like Peptide-1 and Oxyntomodulin
GLP-1 and oxyntomodulin, along with GLP-2, are products of the preproglucagon gene and result from posttranslational processing by prohormone convertases. The preproglucagon gene is expressed in the central nervous system, in intestinal L cells, and in the pancreas. Both GLP-1 and oxyntomodulin act as satiety signals through the GLP-1 receptor.50,118,119 Release from the small intestine is seen after food ingestion; however, the peptides are rapidly cleaved by dipeptidyl peptidase IV and thus have a short half-life. GLP-2 does not affect satiety.120 GLP-1 has effects on insulin secretion and beta cell mass, while GLP-2 affects the growth of intestinal epithelial cells.121
