Role
of set-point
body
weight
RUTH
B. S. HARRIS
Kraft
General Foods, Inc.,
theory
Glenview,
Illinois
60025,
Abstract In adult individuals body weight is maintained at a relatively stable level for long periods. The set-point theory suggests that body weight is regulated at a predetermined, or preferred, level by a feedback control mechanism. Information from the periphery is carried by an affector to a central controller located in the hypothalamus. The controller integrates and transduces the information into an effector signal that modulates food intake or energy expenditure to correct any deviations in body weight from set-point. Evidence for involvement of various factors and physiological systems in the control of food intake and regulation of body weight and fat are reviewed within the context of a control model. Current working hypotheses include roles for nutrients, dietary composition and organoleptic properties, hormones, neural pathways, various brain nuclei, and many neurotransmitters in the regulation of food intake. It is concluded that regulation of body weight in relation to one specific parameter related to energy balance is unrealistic. It seems appropriate to assume that the level at which body weight and body fat content are maintained represents the equilibria achieved by regulation of many parameters. - HARRIs, R. B. S. Role of set-point theory in regulation of body weight. FASEB]. 4: 3310-3318; 1990. Key
Words:
body weight
set-point
body fat
CONCEPT OF A SET-POINT FOR body weight has been the subject of discussion and varied interpretation for several decades. Regulation of body weight was once considered a simple feedback control system in which the hypothalamus modulated food intake to compensate for fluctuations in body weight. It is now apparent that maintenance of body weight is achieved through complex interactions between nutrient selection, organoleptic influences, and metabolic responses to diet, and is influenced by hormonal, environmental, and genetic factors. As more becomes known about factors that can influence body size and composition it is apparent that a simple feedback model for regulation is inappropriate. However, the concept of a set-point is still used as a tool to put observations into perspective. THE
3310
in regulation
of
USA
At some point the effort to make complex information fit an oversimplified model may inhibit true understanding of how a variety of systems interact to determine body weight. The objective of this review is to examine current working hypotheses for the regulation of body weight and composition to determine whether setpoint remains a useful concept. As recent reviews have addressed regulation of body weight (1), set-point (2), and control of energy expenditure (3), the reader will be referred elsewhere for details of the classic experiments that stimulated current research in the regulation of body weight. The hypothesis that individuals regulate their weight at a preferred level developed from observations that many species, including rats (4) and humans (5), forced to gain or lose weight by over- or underfeeding will return to their control weight once the stimulus to change is removed. Most studies investigating weight regulation have examined adult animals; however, it appears that growing animals also have a preferred body weight and body composition that is determined by chronological age. If growth is inhibited by a temporary restriction of food intake, there is a subsequent period of compensatory growth during which animals reach control composition either by increasing food intake or by improving efficiency of energy utilization (6). Young normal animals are rarely hyperphagic, and studies designed to increase weight gain indicate that rats are resistant to gaining excess weight until fast growth is complete. However, examination of body composition in cafeteria-fed (7) and hypothalamically lesioned rats (8) revealed that there had been a change in proportional body composition with an increase in carcass fat content at the expense of lean tissue. Drewry et al. (9), feeding by stomach tube, demonstrated in rats as young as 3 wk of age that overfeeding caused a large increase in carcass fat and a smaller but significant
increase
in
lean
mass.
Once
overfeeding
stopped, the rats were hypophagic for a short period when body weight continued to increase slowly, but body fat was depleted until body composition was the same as that of control animals. These data indicate that mechanisms for regulating body weight, body composition,
and
energy
balance
are
intact
growth and can be perturbed by similar used to change body composition in adult
0892-6638/90/0004-3310/$01.50.
during
techniques rats.
© FASEB
om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
Observations on body weight of adult humans over several years indicate that there are fluctuations of several pounds around a mean weight. Hypothesized models for the feedback regulation of body weight around a reference point allow for small deviations from preferred weight as a load error, or deviation from the mean, that is required to initiate a corrective action (2). However, this variability has led to skepticism about the existence of a mechanism that regulates body weight at a fixed level. Garrow (10) has suggested that body weight is maintained by cognitive control, and Payne and Dugdale (11) developed a computer model to demonstrate that body weight could be maintained by reaching a dynamic equilibrium in which trends to change body weight were reversed by associated changes in obligatory energy expenditure. A similar model that also allows for sensory input to modulate food intake has been proposed by Wirtshafter and Davis (12). Although set-point is often used in reference to regulation of total body weight, confusion arises from the fact that stability of body weight may be secondary to regulation of a component of body composition, such as body fat. Bernardis (13) notes that care should be taken to differentiate between body fat set-point and organismic set-point in which relative body composition is maintained at a different body weight. To maintain a constant body composition, a number of independent metabolic processes may be regulated, with body weight expressing the homeostasis of these equilibria. Studies (14) in which body composition was tracked during recovery from underor overfeeding in mature rats suggest that each tissue compartment is regulated independently, resulting in a fixed body weight. Rats were either overfed, food restricted, or starved to cause significant changes in body weight. On returning to ad libitum feeding, carcass fat was recovered faster than either protein or total weight in restricted rats, whereas starved rats regained protein faster than fat. In overfed rats carcass protein was normalized within days, but fat was still elevated weeks after overfeeding ended. A subsequent study (15) with rats recovering from overfeeding demonstrated that even when body weight had returned to control levels, body fat content could still be increased. However, in vitro measurements of adipose and hepatic lipid metabolism indicated that these rats were insulin resistant, suggesting that changes in metabolism were acting to return body composition to normal. Studies in which rats are allowed to recover from a chronic elevation of body weight indicate that the weight gain may not be fully reversible, possibly because of an irreversible increase in adipocyte number (16). Observations that changes in carcass protein are usually smaller than those of carcass fat and often faster may be interpreted as evidence for a more precise regulation of protein than of fat. However, as fat is more easily perturbed and has the potential to cause large changes in body weight, regulation of body energy stores has received considerably more attention than regulation of lean body mass. In addition to having thermoregulatory and structural functions, fat is a large, compact source of energy.
SET-POINT
THEORY
Although body fat may be regulated, Mrosovsky and Powley (2) emphasize that there must be some allowance for change for fat to function as an endogenous source of energy in time of need. However, during times of plenty there appears to be an upper limit for replenishment, indicating activity of a regulatory system. If the size of fat stores is to remain relatively stable, then energy balance must be maintained with intake equalling expenditure over the long term. This regulation appears to be extremely precise, as it has been calculated that the 11 kg gained by an average woman between ages of 25 and 65 years corresponds to an average daily error of 350 mg of food a day despite a total food intake of 20 tons during the same period (17). Passmore (18) calculated that eating an extra half slice of bread a day, which provides 50 kcal, could result in a weight gain of 20 kg over 10 years. Therefore, although there does not appear to be a correlation between energy intake and expenditure of humans on a daily basis (see ref 17), energy balance must be maintained over longer intervals of time. This can be achieved either by changing food intake to compensate for expenditure or by adjusting expenditure to compensate for energy intake. Food intake will increase in response to dilution of dietary caloric concentration (see ref 17) and to compensate for large changes in expenditure, such as during cold exposure (19). This compensation may not always be precise, as exercise training can result in a lowering of body fat content, especially in males (20). Control of energy expenditure also contributes to the maintenance of energy balance, as it is well established that basal metabolic rate decreases during periods of food restriction or starvation, and expenditure may be increased during periods of overconsumption of palatable diets (see ref 3). Keesey and Powley (3) have recently provided a detailed review of maintenance of energy balance by control of energy expenditure. Despite the absence of direct evidence for a set-point for either body weight or body fat content, various models have been hypothesized for the regulation of energy balance assuming that body fat is regulated at a fixed level. The models include closed ioop systems, open loop systems, and open loop systems that affect closed loop systems (2, 11, 12, 17). All models for regulation have four essential components: the quantity that is being regulated, an affector system, a central controller, and an effector system. In a closed ioop system, information relating to the size of body fat stores feeds back to a central control which adjusts food intake to compensate for changes in fat content. In an open loop system, agents such as hormones and environmental factors act through the central control mechanism to influence body fat content (see Fig. 1). Central
controller
Early studies with brain-lesioned rats demonstrated that damage to the hypothalamus caused dramatic changes in body weight. Large bilateral lesions in the ventromedial hypothalamus (VMH) resulted in the
11
om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
LOOP
CLOSED
TJ Eff’erent
Af’f’e rent
‘4
]
Fat Depots OPEN
LOOP *
I
III
*
I
‘Aff’erent I
iv’ u!
CC Ef’ferent
r COMBINED
I
Fat Depots
OPEN AND CLOSED
/
It’
LOOP
Af’fe rent
I,ccI Affer’ent
\Efferent II
Fat
Depots
I
Figure
1. Hypothesized control models for the regulation of body fat content. In a closed-loop system an afferent signal comes directly from the controlled quantity (fat depots) to the central control (CC) system. The control adjusts efferent output to correct deviations from set-point. In an open loop system the afferent signals are independent of the controlled quantity. In the combined open and closed loop system, the central control integrates afferent input from both the controlled quantity and independent variables.
development of obesity whereas lesions in the lateral hypothalamus (LH) led to weight loss. In both situations the rats defended their new body weight when challenged by food restriction or overfeeding (213 22). Therefore it was proposed that these hypothalamic lesions reset body weight set-point. As the changes in weight were almost entirely accounted for by changes in carcass fat content, it would be more accurate to describe the responses as resetting the body fat set-point (8). Initially it was proposed that the VMH contained a satiety center that inhibited activity of a hunger center in the LH. This hypothesis was supported by observations that discrete knife cuts between the VMH and LH could increase food intake, and that electrical stimulation of the LH stimulated feeding whereas stimulation of the VMH inhibited food intake (see refs 1, 23).
A more exacting investigation demonstrated that obesity after VMH lesions was not simply secondary to hyperphagia, but that the lesions produce a syndrome of behavioral, metabolic, and physiological changes that include aggressive behavior, hyperphagia if food is easily available (but decreased motivation to work for food), decreased gastric motility, and increased insulin response to food (see refs 1, 23). Subdiaphragmatic lesions of the vagus nerve in VMH-lesioned rats arrest or reverse many of these changes. Gastric acid and insulin secretion are normalized, and hyperphagia and obesity are arrested if the rats are offered a commercial pelleted diet (24). The sparing effect of vagotomy in VMH-lesioned rats led to the development of an autonomic hypothesis which proposed that obesity caused by hypothalamic damage was secondary to abnormal control by the autonomic system of peripheral metabolism, and that the resulting hyperinsulinemia stimulated food intake and promoted fat deposition (25). However, as VMH-lesioned vagotomized rats will still become hyperphagic and obese when offered palatable diets (26, 27), the dependence of VMH-lesioned obesity on hyperinsulinemia has been questioned. Whether or not the autonomic nervous system metabolism plays a dominant role in the development of VMH obesity, it is apparent that the syndrome cannot be accurately described as a disruption of specific hunger or feeding centers acting as the central control in the feedback regulation of energy balance. Results from studies using more defined lesions indicated that the dual center hypothesis was too simplistic, and that many areas of the central nervous system are involved in control of food intake and regulation of energy balance. However, various hypothalamic nuclei play a major role in monitoring and integrating peripheral neural and metabolic input which is transduced into efferent signals that modulate nutrient selection, food intake, body composition, and energy balance (Fig. 2). Selective knife cuts demonstrated that lesions confined to the ventromedial nucleus (VMN) had no effect on food intake. Damage rostral to the VMN that cut the ventral noradrenergic bundle caused hyperphagia and weight gain, and lesions of the paraventricular nucleus (PVN) caused the largest increases in both food intake and weight gain (28) independent of hyperinsulinemia (29). A series of studies by Leibowitz and co-workers (30) have demonstrated that the PVN plays a vital role in control of food intake and nutrient selection, as different neuropeptides have selective effects on
Abbreviations: VMH, ventromedial hypothalamus; LH, lateral hypothalamus; VMN, ventromedial nucleus; PVN, paraventricular nucleus; DMN, dorsomedial hypothalamus; 2-DTA, 2-deoxytetronic
acid;
3-DPA,
3-deoxypentanoic
acid;
CSF,
fluid; CCK, cholecystokinin; AP, area postrema; medial nucleus of the solitary tract; GAPDAH,
cerebrospinal NTS, caudal
glyceraldehyde
3-phosphate
dehydrogenase; CRF, corticosterone releasing factor; CC, central control; ArcN, arcuate nucleus; ME, median eminence; F, fornix; 3rd V, third ventricle; Opt, optical tract.
3312 Vol. 4 December 1990 The FASEBlournal HARRIS om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
the set-point concept. It seems more likely that there are multiple areas of the brain, including many in the hypothalamus, that monitor and integrate information from the periphery. Morley and Levine (31) suggest that this information is then transduced in a “cascade system involving multiple neurotransmittors.” Regulated
Figure
2. A schematic representation of a coronal section through the hypothalamus to depict locations of various structures: yentromedial hypothalamus (VMH), lateral hypothalamus (LH), fornix (F),
dorsomedial
nucleus
(DMN),
paraventricular
nucleus
(PVN), third ventricle (3rd V), arcuate nucleus (ArcN), median eminence (ME), and optical tract (Opt). The PVN lies in a plane approximately 1 mm anterior to the other nuclei.
feeding behavior. Of those that stimulate feeding, norepinephrine and neuropeptide Y produce a preference for carbohydrate consumption whereas opiates produce a preference for protein and fat. The feeding elicited by norepinephrine appears to depend on corticosteroids. In contrast, dopamine and serotonin inhibit carbohydrate intake. Leibowitz (30) suggests that the neuropeptide/PVN interaction is responsible for replenishing body carbohydrate stores at the beginning of the diurnal feeding cycle or in response to stress. The effects of PVN activity on energy balance appear to be secondary to regulation of short-term carbohydrate metabolism. Lesions of the LH cause aphagia and loss of body fat content to a new defended level (22). Vagotomy of LHlesioned rats reverses the inhibition of gastric acid secretion but causes a further lowering of body weight (24). Although there may be an autonomic component in the LH syndrome, lesions of dopaminergic fibers originating in the substantia nigra and ascending through the LH can replicate many of the responses seen with LH lesions, which implies an important role for dopaminergic transmitters in maintenance of food intake and body fat content (see ref 1). In contrast, lesions of the dorsomedial hypothalamus (DMN) result in smaller animals that have a normal body composition and a normal food intake per unit body weight. Bernardis (13) has described these animals as having reset the organismic set-point, which suggests that this area of the hypothalamus may be involved in determining growth and body size parameters rather than body composition or food intake. The brief summary provided here is not intended to be an exhaustive review of current work investigating the central control of food intake. It is intended to demonstrate that mechanisms known to influence food intake are too numerous and complex in function to fit
SET-POINT
THEORY
quantity
Aside from the indirect observations of stabilitiy of body weight there is little direct evidence to support the concept that total body fat is subject to regulation. Lipectomy studies in which part or all of defined fat depots are removed have provided equivocal results. Early studies by Leibelt et al. (32) demonstrated compensation in total body fat content of lipectomized gold thioglucose-treated obese mice. Similar results were also reported for lipectomized castrated male rats (33). When one epididymal fat pad was removed from normal Sprague Dawley (34) or lean Zucker rats (35), there did not appear to be any compensation for the loss of fat. Faust et al. (35) suggested that variability in the results could be explained if fat cell size rather than total body fat content was regulated. He proposed that individual adipocytes are resistant to accumulation of excess amounts of lipid, as there is a maximum diameter that can be achieved without compromising physical and metabolic functions. When portions of fat pads are removed, the fat will be replaced if it can be accommodated without significantly enlarging existing fat cells. This may be more easily achieved in animals that tend toward obesity and have a greater capacity for adipocyte proliferation. The hypothesis was supported by observations that feeding Sprague Dawley rats a high-fat diet that induced obesity facilitated compensation for lipectomy whereas a commercial pelleted diet did not (36). When the results could not be repeated with adult lipectomized Osbourne Mendel rats fed a high-fat diet (37), the authors suggested that the difference was attributable to genetic and environmental factors. As regulation of body weight has been demonstrated in a variety of species, it is anticipated that the regulatory mechanism would be less susceptible to strain difference. Mrosovsky (38) suggests that the differences in results from lipectomy studies may depend on which fat pad is removed, which is consistent with the site-specific differences in fat cell size, lipid metabolism, and blood supply. To reconcile this hypothesis with a system that regulates total body fat content, Mrosovsky (38) proposes that there is a common afferent system from all fat depots, but that site-specific responses to central signals are determined by differences in efferent neural supply. Secondary to observations that changes in diet composition can affect body weight, Flatt (39) proposed that body fat content may be determined by nutrient intake. He proposed that not only must energy balance be maintained but that nutrient balance must also be achieved by oxidizing the same mixture of nutrients as
3313
om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
is ingested. Carefully performed calorimetry studies with mice have shown that carbohydrate balance is achieved within a day, presumably because there are very small endogenous carbohydrate stores in the body. In the short-term, carbohydrate balance will be maintained even at the expense of fat balance, as there are large stores of fat in the body that can buffer short-term imbalance. The nutrient balance hypothesis can also be used to explain why high-fat diets induce obesity. When dietary fat is diverted into storage, adipocytes enlarge, turnover of fatty acids increases, and serum free fatty acid concentration rises. The rate of hepatic fatty acid oxidation is proportional to the concentration of fatty acids in the circulation; therefore, equilibrium and nutrient balance is achieved when fat cell size is large enough to stimulate fatty acid oxidation at a rate equal to ingestion. This hypothesis for the effect of nutrient balance on body weight is supported by observations that body fat content is proportional to dietary fat concentration. Recent studies examining the responses of individual animals to changes in dietary composition suggest that there is a genetic component involved in the response to high-fat diets. Some animals increase hepatic fatty acid oxidation more effectively than others and thus gain less weight on a high-fat diet (40, 41). Also, studies with fat substitutes (42) suggest that organoleptic properties of diet may play an important role in determining food intake and body composition of rats consuming greasy diets. Bray (43) has elaborated on Flatt’s hypothesis for nutrient balance to develop a model for the regulation of food intake in which the balance of each major nutrient is regulated separately with a time-constant proportional to the ratio of endogenous stores to the amount eaten each day. Carbohydrate has a short timeconstant and fat has a long time-constant for balance. Circulating concentrations of hormones and nutrients, and vagal afferents from the liver would inform the hypothalamus of peripheral nutrient status. The limited ability to store energy as carbohydrate would also make it more difficult to gain weight on a highcarbohydrate diet than on a high-fat diet, which is consistent with observations from overfeeding studies of a decreased efficiency of energy utilization with a highcarbohydrate compared with a high-fat diet (see ref 43). The concept that body weight or body fat content is regulated at a reference level is inadequate to explain observations that factors such as organoleptic properties of food and nutrient composition of the diet can influence body composition. Afferent
system
Kennedy (44) was the first to propose a lipostatic theory in which total body weight is maintained by regulating body fat content. He hypothesized that the hypothalamus sensed the concentration of a circulating metabolite to obtain information on the size of stores of body fat. The hypothalamus would then modulate the information to effect changes in food intake to compensate for changes in body fat content. This hypothesis was
supported by studies with parabiosed rats (45) in which two animals were united surgically to share a common blood supply. When one member of a pair was made hyperphagic and obese by VMH lesions, the partner appeared hypophagic and lost both lean and fat body mass. In a second study (46), using electrical stimulation of the LH to induce obesity, measurement of food intake of the individual members of the pairs indicated that the nonstimulated partner ate progressively less as obesity developed in the stimulated partner. These observations were interpreted as evidence for the hypothalamus responding to a circulating satiety factor that carried information on total body fat stores. The obese rat was unable to respond to the signal due to hypothalamic manipulation, but the partner responded by reducing food intake in an effort to correct energy balance. Although the parabiosis experiments provided evidence for feedback regulation of energy balance, it was impossible to determine whether food intake, body weight, total body fat content, or fat cell size was being monitored by the hypothalamus. Hervey (17) proposed that the circulating satiety-factor could be a fat-soluble agent, such as a steroid, so that as body fat content increased less of the factor would be in circulation; as fat depots decreased concentrations would rise, and the hypothalamus would stimulate feeding. Glucose is an obvious candidate as an afferent signal to the hypothalamus. Mayer (47) proposed that insulinsensitive cells in the brain monitored arterial-venous differences in glucose concentration as a measure of glucose utilization. The presence of glucoreceptors in the hypothalamus has been demonstrated by selective uptake of gold thioglucose. Also, vagal afferents from hepatic glucose receptors ascend via the NTS and into the LH where they may modulate neuronal activity (see ref 1). Oomura and colleagues (48) have demonstrated that activity of neurons within the VMH and LH respond selectively to physiological changes associated with hunger and satiety. Glucose or glucose plus insulin stimulate firing of glucoreceptors located in the VMH, whereas glucose depresses activity of glucosensitive cells in the LH. In contrast, fatty acids depress activity of VMH glucoreceptors and facilitate activity of LH glucose-sensitive neurons. As free fatty acids are at higher circulating concentrations during periods of hunger, the hypothalamic neurons respond appropriately for units involved in the control of food intake. Subsequent work demonstrated that the glucoreceptor and glucose-sensitive cells were responsive to most metabolites and hormones known to be involved in the control of feeding, which suggests that these cell bodies could be more appropriately called hunger-related and satiety-related neurons (49). More recently Oomura and colleagues (49) analyzed the blood of food-deprived rats by mass spectrometry to identify novel compounds that could be rela.ed to hunger and satiety. Two sugars, 2-deoxytetronic acid (2-DTA) and 3-deoxypentanoic acid (3-DPA), were identified as satiety and hunger factors. Infusion of micromolar concentrations of 2-DTA into the third ventricle of food-deprived rats inhibited feeding and
3314 Vol. 4 December 1990 The FASEBIou ma! HARRIS om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
neuronal activity in the LH whereas infusion of 3-DPA stimulated both feeding and nerve activity. The specific effect of these compounds on glucose-sensitive neurons in the LH suggest involvement in the control of feeding behavior. In addition to the sensitivity of hypothalamic neuronal function to circulating concentrations of nutrients and metabolites, there is some evidence that hypothalamic metabolism mimics peripheral energy utilization. Kasser et al. (50) have demonstrated in rats that during conditions of positive energy balance, fatty acid synthesis is increased in the VMH, and fatty acid oxidation is decreased in the LH. During periods of food restriction and negative energy balance, LH fatty acid oxidation is increased. These changes in glucose and fatty acid utilization were specifically localized in hypothalamic tissue and mimicked changes in hepatic nutrient utilization. In a subsequent study (51) hypothalamic metabolism was tracked in rats recovering from over- or underfeeding. LH fatty acid oxidation and body fat content returned to control levels simultaneously. These observations imply a correlation between peripheral energy balance and central nutrient utilization, and may also represent transduction of afferent neuronal and humoral information on peripheral energy status. McHugh and Moran (52) have demonstrated that there is negative feedback on food intake by vagal afferents from gastric stretch receptors. The primary feedback is on rate of gastric emptying; satiety appears to be a secondary response. In addition to neural afferent signals, there is evidence that circulating concentrations of nutrients and hormones may influence feeding behavior. Baskin et al. (53) propose that plasma insulin is a peripheral satiety signal in determining energy balance. It has been shown that insulin can cross the blood-brain barrier, and insulin receptors have been identified in various brain tissues. Chronic intraventricular infusions of insulin cause a dose-related reduction of food intake and body weight in baboons, and when there are chronic elevations of serum insulin, cerebrospinal fluid (C SF) concentrations of insulin also increase. Baskin et al. (53) suggest that insulin either modifies or acts in concert with other neurotransmittors known to influence food intake. Walls and Koopmans (54) have shown that intravenous infusion of glucose plus amino acids or free fatty acids depress food intake of fed rats. The rats compensated more accurately for the calories infused as glucose and amino acids than for those infused as lipid. Chronic central infusions of glucose or glycerol reduce body weight of rats by inhibiting food intake, whereas infusions of bhydroxybutyrate lowered body weight independently of a reduction in food intake (55). A more direct effect of gastric signals in the development of satiety has been suggested from observations that peptides released from the gastrointestinal tract may also play a role as central neurotransmittors. Cholecystokinin (CCK), released from the duodenum in response to distension and nutrient composition, has been shown to induce satiety whether administered
centrally or peripherally. The peripheral effect may be mediated by the vagus nerve (56). The sensory input from the subdiaphragmatic vagus terminates in the area postrema (AP) and caudal medial nucleus of the solitary tract (NTS), areas of the brain that are also accessible to circulating metabolites as there is neither a blood/brain nor blood/C SF barrier in the area. The observations on reversal of VMH obesity by vagotomy have been criticized for the possibility that food intake was depressed by inhibition of gastrointestinal motility. By using AP/NTS lesions it was possible to separate the importance of peripheral sensory vagal afferents in the control of food intake from vagal motor function, which was not affected by the lesions. AP/NTS lesions caused transient hypophagia and reduction of body weight to a new defended level (57). There was no exaggerated preference to a palatable diet in AP/NTS lesioned rats, which suggests that the diet-induced obesity in VMHlesioned vagotomized rats was secondary to loss of motor control of the gastrointestinal tract. Observations that food intake can be influenced by various nutrients, hormones, and peripheral neural stimuli indicate that there can be no single afferent input to the hypothalamus to determine body weight. However, all the different afferent signals could be considered to be factors impinging on the control in an open loop system. Efferent
system
Once the brain has integrated and transformed information on the size of body fat stores, an effector mechanism has to control food intake or energy expenditure to maintain energy balance. Efferent output may be neural or hormonal, and is probably a combination of the two. Genetically obese rodents have been popular models for examining the regulation of energy balance because obesity is the result of a single recessive gene. If one abnormal gene means one abnormal protein, theoretically it should be possible to identify the single abnormality that results in obesity and to gain insight into mechanisms that normally regulate energy balance. However, syndromes of genetic obesity tend to be complex, and it is impossible to determine which abnormalities are primary and which develop as secondary effects of obesity. For example, the obese Zucker rat is hyperphagic, hyperinsulinemic, hypertriglyceridemic, hypothyroid, hypothermic, hypersensitive to glucocorticoids, sterile, and has low rates of sympathetic activity in some organs (58). Therefore, genetic obesity may be more accurately described as one gene, one syndrome. Bray (43, 59) has suggested that some kinds of genetic obesity result from an autonomic/endocrine abnormality. It is proposed that obesity in genetic models such as Zucker rats and ob/ob mice, and in animals made obese by hypothalamic lesions, can be explained by hypersensitivity toward adrenocorticoid hormones that results in abnormal sympathetic tone. Food intake is controlled to maintain nutrient balance. The afferent humoral and neural signals are integrated and trans-
3315 SET-POINT THEORY om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
formed to neural efferent signals. One type of efferent output, modulated by the LH, controls activity that locates and identifies food. The second efferent output controls nutrient balance to maintain homeostasis. This output is mediated by changes in hypothalamic PVN concentrations of corticosterone releasing factor (CRF) such that increased concentrations of CRF increase sympathetic activity and inhibit food intake. In this model, circulating concentrations of corticosterone would have a negative feedback effect on PVN concentrations of CRF. Also, the corticosterone may modify glucose utilization by neurons in other hypothalamic nuclei. Support for this hypothesis comes from a variety of observations. Central administration of CRF depresses food intake. Young VMH-lesioned rats and genetically obese rodents have suppressed sympathetic tone. Conversely, norepinephrine turnover is increased in rats offered sucrose solution or cafeteria diets (see ref 59) and can be affected by isocaloric changes in nutrient intake (60). Obese Zucker rats appear to have exaggerated sensitivity toward the hyperphagic effects of corticosterone (see ref 23). Adrenalectomy of VMHlesioned or genetically obese rodents arrests hyperphagia and the development of obesity. In Zucker rats, adrenalectomy increases sympathetic activity in brown adipose tissue. All of these responses to adrenalectomy are reversed by replacement of glucocorticoids (see ref 59). Preliminary, nonquantitative evidence has suggested that hepatic mRNA for the lipogenic enzyme glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was depressed by adrenalectomy and recovered in a timedependent manner with glucocorticoid treatment in obese Zucker rats. There was no effect of adrenalectomy or glucocorticoid replacement on hepatic GAPDH mRNA isolated from the lean Zucker rat. Unpublished observations are reported for a similar response in adipose malic enzyme mRNA. Bray suggests that the defective gene in the Zucker rat normally produces a peptide that modifies corticoid stimulation of mRNA for lipogenic enzymes. In obese Zucker rats the protein is expressed at low levels, and thus there is an exaggeration of the anabolic action of glucocorticoids (59). The model proposed by Bray (59) indicates that abnormal glucocorticoid/autonomic interaction is the primary modulator of energy balance. However, some evidence suggests that this mechanism does not fully account for VMH-lesion or genetic obesity in rats. Adrenalectomy of obese rats does not correct either body fat content or food intake to the levels measured for lean littermates (see refs 1, 61, 62). When obesity of VMH-lesioned rats is arrested by vagotomy or Zucker rats are adrenalectomized, the animals will still become hyperphagic and obese if they are offered a palatable diet (26, 61). Adrenalectomized Zucker rats remain hypothermic and sterile (see ref 1), indicating that only some aspects of genetic obesity have been corrected. Also Fletcher and McKenzie (61) have reported that corticosteroid replacement in adrenalectomized Zucker rats not only reinstates hyperinsulinemia in obese rats but enhances, to a smaller degree, insulin release in
response to glucose administration in lean littermates, which suggests that the corticosteroid influences on in sulin release are intact in both phenotypes. Considerabi more work is required before it can be determined whether the time-dependent reversal in mRNA levels o lipogenic hormones is a primary response to the replacement of corticosteroid or whether it is secondary to the reinstatement of hyperphagia and hyperinsulinemia. In addition to sympathetic and adrenal system dysfunctions in models of obesity, there is some evidence that humoral factors play a role in determining body composition. Satietin is a glycoprotein that has been isolated from the blood of normal humans and rats. When satietin is infused centrally into rats, there is a transient
depression
of
food
intake
and
a sustained
depression of body weight, which suggests a decrease in efficiency of energy utilization (63). These observations indicate that blood-borne factors may play a role in regulating energy balance not only by controlling food intake but by adjusting energy expenditure. Parabiosis studies
in which
one
member
of a pair
is made
obese
by overfeeding leads to distinctly different results from those observed with genetically obese or VMH-lesioned obese models. It has been repeatedly demonstrated in a series of studies with Sprague Dawley rats (64) that when one rat is made obese by overfeeding by stomach tube, its partner makes a statistically nonsignificant reduction in food intake, retains a normal lean body mass, and loses a large portion of its body fat content. The loss of fat appears to be secondary to a specific inhibition of adipocyte fatty acid synthesis. It has been hypothesized that development of obesity in the overfed rat leads to the release of a blood-borne factor that is carried into the partner fed ad libitum. This rat responds to the increased body fat content of its partner by making ing further
a small reduction in food intake and inhibitdeposition of fat. Although the change in
food intake is not statistically significant, even when summated over 60 days, the energy deficit is large enough to account for the change in body fat stores. There are no measurable changes in hepatic fatty acid synthesis, but measurements of glucose flux through the pentose phosphate shunt and of lipogenic enzyme activity suggest that the liver is also sensitive to the antilipogenic factor. It is possible that the liver uses alternate pathways for maintaining fatty acid synthesis. The changes in body composition of members of parabiosed pairs is not dependent either on the diet used to induce obesity or on the composition of diet eaten ad libitum by the normal partner (unpublished observations). These parabiosis studies do not indicate whether the primary effect of the circulating factor is to reduce food intake or inhibit fatty acid metabolism. A bioassay has demonstrated that serum from obese overfed rats is capable of inhibiting fatty acid synthesis in isolated adipocytes in vitro (65). When fat cells are incubated with serum from overfed obese rats for 12 h, there is a significant
inhibition
of fatty
acid
synthesis
compared
with cells preincubated with serum from control rats. This assay indicates that the serum factor is capable of reducing fatty acid synthesis by acting directly on adi-
HARRIS 3316 Vol. 4 December 1990 The FASEBJournal om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
pocytes and does not necessarily require central mediation for its effect. In addition, further studies have demonstrated that the factor is produced by overfed hypophysectomized rats and is unlikely to be of hypothalamic or pituitary origin. The identity of the factor is unknown, but it has been demonstrated that a bloodborne dye takes 2 h to equilibrate in parabiosed rats (66), which indicates that any factor that is produced by one animal and is active in the partner must have a long half-life in the circulation. The anti-lipogenic factor appears to be released in response to a substantial increase in body fat content rather than to increased food intake. Whether it is involved in the regulation of total body fat content or in regulation of adipocyte size and number remains to be elucidated. CONCLUSION Although long-term can
body weight or body regulation of energy
be achieved
only
through
fat set-point is related to balance, this regulation modulation
of short-term
responses and sensitivities to factors that influence food intake and energy expenditure. There is evidence for involvement of different areas of the brain in the control of various aspects of feeding behavior, ranging from acute organoleptic preferences to long-term caloric compensation. In this context the set-point theory for regulation of body weight in which one specific measure related to energy balance is regulated appears
unrealistic.
It
seems
appropriate
to
assume
that the level at which body fat content is maintained represents the equilibria achieved by regulation of many parameters including nutrient balance, energy balance, and physical aspects of adipose tissue, such as fat cell number and size. Current evidence for the role of many different factors in the control of food intake and regulation of energy balance indicates that mechanisms responsible for determining body weight are far more complex than was once thought. In this context, having an inaccurate and oversimplified preconception of how the various mechanisms may be integrated could hinder progress by limiting the boundaries of acceptable concepts.
8. Bernardis, L. L., Luboshitsky, R., Bellinger, L. L., McEwen, G. (1982) Nutritional studies in the weanling rat normophagic hypothalamic obesity. j Nutr. 112, 1441-1455 9. Drewry, M. M., Harris, R. B. S., and Martin, R. J. (1989) effect of increased adiposity on food intake of juvenile
Physiol. Behav. 45, 381-386
16.
17. 18. 19.
20.
sis and lipolysis associated with recovery from reversible obesity in mature female rats. Proc. Soc. Exp. BioL Med. 191, 82-89 Hill, J. 0., Dorton, J., Sykes, M. N., and DiGirolamo, M. (1989) Reversal of dietary obesity is influenced by its duration and severity. mt. J. Obes. 13, 711-722 Hervey, G. R. (1969) Regulation of energy balance. Nature (London) 222, 629-631 Passmore, R. (1971) The regulation of body weight in man. Proc. Nutr Soc. 30, 122-127 Armitage, G., Harris, R. B. S., Hervey, G. R., and Tobin, G. (1984) The relationship between energy expenditure and environmental temperature in congenitally obese and non-obese Zucker rats. j PhysioL (London) 350, 197-207 Tremblay, A., Desprey, J-P., and Bouchard, C. (1988) Alteration in body fat and fat distribution with exercise. In Fat Distribution during Growth and Later Health Outcomes. (Bouchard, C. and Johnston, F. E., eds) pp. 297-312, Alan Liss Inc., New York
21. Hoebel, normal
B. E. (1986) Neurological regulation of body weight. Grit. Rev. Gun. Neurobiol. 2, 1-60 2. Mrosovsky, N., and Powley, T L. (1977) Set points for body weight and fat. Behav. Biol. 20, 205-223 3. Keesey, R. E., and Powley, T. L. (1986) The regulation of body weight. Annu. Rev. Psychol. 37, 109-133 4. Cohn, C., and Joseph, D. (1962) Influence of body weight and body fat on appetite of normal lean and obese rats. Yalej Biol. Med. 34, 598-601 5. Sims, E. A. H., and Horton, E. 5. (1968) Endocrine and metabolic adaptation to obesity and starvation. Am. j Gun. Nutr. 21, 1455-1470 6. Wilson, P. N., and Osbourn, D. R. (1960) Compensatory growth after undernutrition in animals and birds. Biol. R#{128}v.
Cambridge Philos. Soc. 35, 324-363 N. J., and Stock, M. J. (1982) Effects of feeding a palatable cafeteria diet on energy balance in young and adult lean (+1?) Zucker rats. Br. J. Nutr. 47, 461-471
7. Rothwell,
SET-POINT
THEORY
B. G., and
Teitlebaum,
P. (1966)
Weight
regulation
in
and hypothalamic hyperphagic rats. j Comp. PhysioL Psychol. 61, 189-193 22. Keesey, R. E., Boyle, P. C., Kemnitz, J. W., and Mitchel, J. S. (1976) The role of the lateral hypothalamus in determining the body weight set-point. In Hunger: Basic Mechanisms and Clinical Implications (Novin, D., Wywicka, W., and Bray, G., eds) pp. 243-255, Raven, New York 23. Bray, G. A., and York, D. A. (1979) Hypothalamic and genetic obesity in experimental animals. An autonomic and endocrine hypothesis. Physiol. Rtv. 59, 719-809 24. Powley, T L., and Opshal, C. A. (1976) In Hunger: Basic Mechanisms and Clinical Implications (Novin, D., Wywicka, W., and
Bray, G., eds) pp. 313-326, 25. Berthoud, tromedial
1. Levin,
The rats.
10. Garrow, J. S., and Stalley, S. F. (1975) Is there a set-point to human body weight? Proc. Nutr. Soc. 34, 84A 11. Payne, P. R., and Dugdale, A. E. (1977) Mechanisms for the control of body-weight. Lancet I, 583-586 12. Wirtshafter, D., and Davis, J. D. (1977) Set points, settling points and the control of body weight. PhysioL Be/wv. 19, 75-78 13. Bernardis, L. L., and Bellinger, L. L. (1987) The dorsomedial hypothalamic nucleus revisited: 1986 update. Brain Ret. Rev. 12, 321-381 14. Harris, R. B. S., Kasser, T. R., and Martin, R. J. (1986) Dynamics of recovery of body composition after overfeeding, food restriction or starvation of mature female rats. j NutE 116, 2536-2546 15. Harris, R. B. S., and Martin, R. J. (1989) Changes in lipogene-
sulinemia REFERENCES
and with
H.
R.,
and
Raven,
New York
Jeanrenoud,
and its reversal hypothalamus
B. (1979)
and
Frohman,
mediated hyperinsulinemia Be/wv. Rev. 6, 205-214
hyperin-
after lesions of the yenin anaesthetized rats. Endocrinology
150, 146-151 26. Sclafani, A. (1981) The role of hyperinsulinemia nerve in hypothalamic hyperphagia reexamined. 402-410 27. King, B. M.,
Acute
by vagotomy
L. A. (1982)
in hypothalamic
The
and the vagus Diabetologia 20, role
obesity.
of vagally Neurosci.
28. Gold, R. M., Jones, A. P., Sawchenko, P. E., and Kapatos, G. (1977) Paraventricular area: critical focus of a longitudinal neurocircuitry mediating food intake. Physiol. Be/wv. 18, 1111-1116 29. Sims, J. S., and Lorden, J. F. (1986) Effect of paraventricular nucleus lesions on body weight, food intake and insulin levels. Be/wv. Brain Rev. 22, 265-281 30. Leibowitz, S. F. (1986) Brain monoamines and peptides: role in the control of eating behavior. Federation Proc. 45, 1396-1403 31. Morley, J. E., and Levine, A. S. (1983) The central control of appetite. Lancet I, 398-401
32. Leibelt, R. A., Ichinoue, S., and Nicholson, N. (1965) Regulatory influences of adipose tissue on food intake and body
3317
om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
weight. Ann. NY Acad. &i. 131, 559-582 33. Bailey, J. W., and Anderson, D. B. (1980) Rate of fat compensation and growth efficiency of lipectomized Sprague Dawley rats. Nutr. 110, 1785-1792 34. Kral, J. G. (1976) Surgical reduction of adipose tissue in the male Sprague Dawley rat. Am. j Physiol. 231, 1090-1096 35. Faust, I. M., Johnson, P. R., and Hirsch,J. (1976) Noncompensation of adipose mass in partially lipectomized mice and rats.
51. Kasser, T R., Harris, R. B. of satiety: in vitro energy hypophagic and hyperphagic
j
Am. j 36. Faust,
Physiol. 231, 538-544 I. M., Johnson, P. R., and Hirsch, J. (1977) Adipose regeneration following lipectomy. Science 197, 391-393 37. Faust, I. M., Johnson, P. R., and Hirsch, J. (1979) Adipose tissue regeneration in adult rats. Proc. Soc. Exp. Biol. Med. 161, 111-114 38. Mrosovsky,
N. (1986)
Body
38, 407-414 39. Flatt, J. P. (1987) Dietary
41.
42.
43.
is regulated?
fat, carbohydrate
PhysioL Be/wv.
balance
and weight
effects of exercise. Am. j C/in. Nutr. 45, 296-306 Triscari, J., Nauss-Karol, C., Levin, B. E., and Sullivan, A. C. (1985) Changes in lipid metabolism in diet induced obesity. Metabolism 34, 580-587 Chang, S., Graham, B., Yakubu, F., Lin, D., Peters, J. C., and Hill, J. 0. (1990) Metabolic differences between obesity prone and obesity resistant rats. Am. J. PhysioL In press Jones, W. K., and Harris, R. B. 5. (1990) Physiological responses of mature rats fed low and high fat diets of similar texture. FASEBJ 4, A379 (abstr.) Bray, G. A. (1987) Obesity-a disease of nutrient or energy
balance?
46. Parameswaran,
S. V., Steffens,
A. B., Hervey,
G. R.,
and
de Ruiter, L. (1977) Involvement of a humoral factor in regulation of body weight in parabiotic rats. Am. j Physiol. 232,
R150.-R157 47. Mayer,
J. (1955) Regulation
the glucostatic theory Sci. 63, 15-43 48. Oomura, fatty acid
of energy
and body weight: and lipostatic hypothesis. Ann. NY Acad.
Y. (1976) Significance on the hypothalamic
intake
of glucose, insulin, and feeding and satiety neurons.
free In
Hunger: Basic Mechanisms and Clinical Implications (Novin, D., Wywicka, W., and Bray, G., eds) pp. 145-157, Raven, New York 49. Oomura, Y. (1987) Regulation of feeding by neural responses to endogenous factors. News PhymioL &i. 2, 199-203 50. Kasser, T. R., Harris, R. B. S., and Martin, R. J. (1985) Level of satiety: fatty acid and glucose metabolism in three brain sites associated with feeding. Am. J. Physiol. 248, R447-R452
P. R.,
and
Moran,
regulation of caloric ingestion ioL 236, R254-R259 53. Baskin, Dorsa,
R. J. (1989) Leve in brain dunn recovery. Am. j
D. G., Figlewicz, D. (1987) Insulin
T. H. (1978)
The
accuracy
in the rhesus monkey. Am. j
D. P., Woods, S., Porte, D., in the brain. Annu. Rev. PhysioL
of the
Physand
49,
335-347 E. K., and Koopmans, H. S. (1989) Effect of intravenous nutrient infusions on food intake in rats. Phytiol. Be/wv. 45, 1223-1226 55. Davis, J. D., Wirtshafter, D., Asin, K. E., and Brief, D. (1981) 54. Walls,
56. 57.
58.
59.
60.
Nutr Rev. 45, 33-43
44. Kennedy, G. C. (1953) The role of depot fat in the hypothalamic control of food intake in the rat. Proc. Royal Soc. B BioL Sci. 140, 578-592 45. Hervey, G. R. (1959) The effects of lesions in the hypothalamus in parabiotic rats. J. PhysioL (London) 145, 336-352
metabolism body weight
PhysioL 257, R1322-R1327 52. McHugh,
-
maintenance
40.
fat: what
S., and Martin,
61.
Sustained intracerebroventricular infusion of brain fuels reduces body weight and food intake in rats. Science 212, 81-83 Dockray, G. J. (1982) The physiology of cholecystokinin in brain and gut. Br. Med. BulL 38, 253-258 Hyde, T M., and Miselis, R. R. (1983) Effects of area postremalcaudal medial nucleus of solitary tract lesions on food intake and body weight. Am. j PhysioL 244, R577-R587 Martin, R. J., Harris, R. B. S., and Jones, D. D. (1986) Evidence for a central mechanism of obesity in the Zucker fatty rat (fa/fa). Proc. Soc. Exp. Biol. Med. 183, 1-10 Bray, G. A. (1989) 1989 McCollum award lecture. Genetic and hypothalamic mechanisms for obesity-finding the needle in the haystack. Am. j Glin. Nutr 50, 891-902 Vander Tuig, J. G., and Romsos, D. R. (1984) Effects of dietary carbohydrate, fat and protein on norepinephrine turnover in rats. Metabolism 33, 26-33 Fletcher, J. M., and McKenzie, N. (1988) The effect of dietary fat content on the growth and body composition of lean -and genetically obese Zucker rats adrenalectomized before weaning.
Br j
Nutr. 60, 563-569
62. Fletcher, J. M., and McKenzie, N. (1988) The parasympathetic nervous system and glucocorticoid-mediated hyperinsulinemia in the genetically obese (faJfa) Zucker rat.]. Endocr. 118, 87-92
63. Bellinger, L. L., and Mendel, V. E. (1988) Ingestion, body weight and activity of rats receiving repeated intracerebroventricular infusions of rat satietin. PhycioL Be/wv. 44, 445-452 64. Harris, R. B. S., and Martin, R. J. (1989) Physiological and metabolic changes in parabiotic partners of obese rats. In Hormones, Thermogenesis and Obesity (Lardy, H., and Straatman, F., eds) pp. 233-243, Elsevier Science Publishing Co., New York 65. Harris, R. B. S., Bruch, R. C., and Martin, R. J. (1989) In vitro evidence for an inhibitor of lipogenesis in serum from overfed obese rats. Am. j PhysioL 257, R326-R336 66. Nishizawa, Y., and Bray, G. A. (1980) Evidence for a circulating ergostatic factor: studies on parabiotic rats. Am. j PhysioL 239, R344-R351
3318 Vol. 4 December 1990 The FASEBJournal HARRIS om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
of set-point
body
weight
RUTH
B. S. HARRIS
Kraft
General Foods, Inc.,
theory
Glenview,
Illinois
60025,
Abstract In adult individuals body weight is maintained at a relatively stable level for long periods. The set-point theory suggests that body weight is regulated at a predetermined, or preferred, level by a feedback control mechanism. Information from the periphery is carried by an affector to a central controller located in the hypothalamus. The controller integrates and transduces the information into an effector signal that modulates food intake or energy expenditure to correct any deviations in body weight from set-point. Evidence for involvement of various factors and physiological systems in the control of food intake and regulation of body weight and fat are reviewed within the context of a control model. Current working hypotheses include roles for nutrients, dietary composition and organoleptic properties, hormones, neural pathways, various brain nuclei, and many neurotransmitters in the regulation of food intake. It is concluded that regulation of body weight in relation to one specific parameter related to energy balance is unrealistic. It seems appropriate to assume that the level at which body weight and body fat content are maintained represents the equilibria achieved by regulation of many parameters. - HARRIs, R. B. S. Role of set-point theory in regulation of body weight. FASEB]. 4: 3310-3318; 1990. Key
Words:
body weight
set-point
body fat
CONCEPT OF A SET-POINT FOR body weight has been the subject of discussion and varied interpretation for several decades. Regulation of body weight was once considered a simple feedback control system in which the hypothalamus modulated food intake to compensate for fluctuations in body weight. It is now apparent that maintenance of body weight is achieved through complex interactions between nutrient selection, organoleptic influences, and metabolic responses to diet, and is influenced by hormonal, environmental, and genetic factors. As more becomes known about factors that can influence body size and composition it is apparent that a simple feedback model for regulation is inappropriate. However, the concept of a set-point is still used as a tool to put observations into perspective. THE
3310
in regulation
of
USA
At some point the effort to make complex information fit an oversimplified model may inhibit true understanding of how a variety of systems interact to determine body weight. The objective of this review is to examine current working hypotheses for the regulation of body weight and composition to determine whether setpoint remains a useful concept. As recent reviews have addressed regulation of body weight (1), set-point (2), and control of energy expenditure (3), the reader will be referred elsewhere for details of the classic experiments that stimulated current research in the regulation of body weight. The hypothesis that individuals regulate their weight at a preferred level developed from observations that many species, including rats (4) and humans (5), forced to gain or lose weight by over- or underfeeding will return to their control weight once the stimulus to change is removed. Most studies investigating weight regulation have examined adult animals; however, it appears that growing animals also have a preferred body weight and body composition that is determined by chronological age. If growth is inhibited by a temporary restriction of food intake, there is a subsequent period of compensatory growth during which animals reach control composition either by increasing food intake or by improving efficiency of energy utilization (6). Young normal animals are rarely hyperphagic, and studies designed to increase weight gain indicate that rats are resistant to gaining excess weight until fast growth is complete. However, examination of body composition in cafeteria-fed (7) and hypothalamically lesioned rats (8) revealed that there had been a change in proportional body composition with an increase in carcass fat content at the expense of lean tissue. Drewry et al. (9), feeding by stomach tube, demonstrated in rats as young as 3 wk of age that overfeeding caused a large increase in carcass fat and a smaller but significant
increase
in
lean
mass.
Once
overfeeding
stopped, the rats were hypophagic for a short period when body weight continued to increase slowly, but body fat was depleted until body composition was the same as that of control animals. These data indicate that mechanisms for regulating body weight, body composition,
and
energy
balance
are
intact
growth and can be perturbed by similar used to change body composition in adult
0892-6638/90/0004-3310/$01.50.
during
techniques rats.
© FASEB
om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
Observations on body weight of adult humans over several years indicate that there are fluctuations of several pounds around a mean weight. Hypothesized models for the feedback regulation of body weight around a reference point allow for small deviations from preferred weight as a load error, or deviation from the mean, that is required to initiate a corrective action (2). However, this variability has led to skepticism about the existence of a mechanism that regulates body weight at a fixed level. Garrow (10) has suggested that body weight is maintained by cognitive control, and Payne and Dugdale (11) developed a computer model to demonstrate that body weight could be maintained by reaching a dynamic equilibrium in which trends to change body weight were reversed by associated changes in obligatory energy expenditure. A similar model that also allows for sensory input to modulate food intake has been proposed by Wirtshafter and Davis (12). Although set-point is often used in reference to regulation of total body weight, confusion arises from the fact that stability of body weight may be secondary to regulation of a component of body composition, such as body fat. Bernardis (13) notes that care should be taken to differentiate between body fat set-point and organismic set-point in which relative body composition is maintained at a different body weight. To maintain a constant body composition, a number of independent metabolic processes may be regulated, with body weight expressing the homeostasis of these equilibria. Studies (14) in which body composition was tracked during recovery from underor overfeeding in mature rats suggest that each tissue compartment is regulated independently, resulting in a fixed body weight. Rats were either overfed, food restricted, or starved to cause significant changes in body weight. On returning to ad libitum feeding, carcass fat was recovered faster than either protein or total weight in restricted rats, whereas starved rats regained protein faster than fat. In overfed rats carcass protein was normalized within days, but fat was still elevated weeks after overfeeding ended. A subsequent study (15) with rats recovering from overfeeding demonstrated that even when body weight had returned to control levels, body fat content could still be increased. However, in vitro measurements of adipose and hepatic lipid metabolism indicated that these rats were insulin resistant, suggesting that changes in metabolism were acting to return body composition to normal. Studies in which rats are allowed to recover from a chronic elevation of body weight indicate that the weight gain may not be fully reversible, possibly because of an irreversible increase in adipocyte number (16). Observations that changes in carcass protein are usually smaller than those of carcass fat and often faster may be interpreted as evidence for a more precise regulation of protein than of fat. However, as fat is more easily perturbed and has the potential to cause large changes in body weight, regulation of body energy stores has received considerably more attention than regulation of lean body mass. In addition to having thermoregulatory and structural functions, fat is a large, compact source of energy.
SET-POINT
THEORY
Although body fat may be regulated, Mrosovsky and Powley (2) emphasize that there must be some allowance for change for fat to function as an endogenous source of energy in time of need. However, during times of plenty there appears to be an upper limit for replenishment, indicating activity of a regulatory system. If the size of fat stores is to remain relatively stable, then energy balance must be maintained with intake equalling expenditure over the long term. This regulation appears to be extremely precise, as it has been calculated that the 11 kg gained by an average woman between ages of 25 and 65 years corresponds to an average daily error of 350 mg of food a day despite a total food intake of 20 tons during the same period (17). Passmore (18) calculated that eating an extra half slice of bread a day, which provides 50 kcal, could result in a weight gain of 20 kg over 10 years. Therefore, although there does not appear to be a correlation between energy intake and expenditure of humans on a daily basis (see ref 17), energy balance must be maintained over longer intervals of time. This can be achieved either by changing food intake to compensate for expenditure or by adjusting expenditure to compensate for energy intake. Food intake will increase in response to dilution of dietary caloric concentration (see ref 17) and to compensate for large changes in expenditure, such as during cold exposure (19). This compensation may not always be precise, as exercise training can result in a lowering of body fat content, especially in males (20). Control of energy expenditure also contributes to the maintenance of energy balance, as it is well established that basal metabolic rate decreases during periods of food restriction or starvation, and expenditure may be increased during periods of overconsumption of palatable diets (see ref 3). Keesey and Powley (3) have recently provided a detailed review of maintenance of energy balance by control of energy expenditure. Despite the absence of direct evidence for a set-point for either body weight or body fat content, various models have been hypothesized for the regulation of energy balance assuming that body fat is regulated at a fixed level. The models include closed ioop systems, open loop systems, and open loop systems that affect closed loop systems (2, 11, 12, 17). All models for regulation have four essential components: the quantity that is being regulated, an affector system, a central controller, and an effector system. In a closed ioop system, information relating to the size of body fat stores feeds back to a central control which adjusts food intake to compensate for changes in fat content. In an open loop system, agents such as hormones and environmental factors act through the central control mechanism to influence body fat content (see Fig. 1). Central
controller
Early studies with brain-lesioned rats demonstrated that damage to the hypothalamus caused dramatic changes in body weight. Large bilateral lesions in the ventromedial hypothalamus (VMH) resulted in the
11
om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
LOOP
CLOSED
TJ Eff’erent
Af’f’e rent
‘4
]
Fat Depots OPEN
LOOP *
I
III
*
I
‘Aff’erent I
iv’ u!
CC Ef’ferent
r COMBINED
I
Fat Depots
OPEN AND CLOSED
/
It’
LOOP
Af’fe rent
I,ccI Affer’ent
\Efferent II
Fat
Depots
I
Figure
1. Hypothesized control models for the regulation of body fat content. In a closed-loop system an afferent signal comes directly from the controlled quantity (fat depots) to the central control (CC) system. The control adjusts efferent output to correct deviations from set-point. In an open loop system the afferent signals are independent of the controlled quantity. In the combined open and closed loop system, the central control integrates afferent input from both the controlled quantity and independent variables.
development of obesity whereas lesions in the lateral hypothalamus (LH) led to weight loss. In both situations the rats defended their new body weight when challenged by food restriction or overfeeding (213 22). Therefore it was proposed that these hypothalamic lesions reset body weight set-point. As the changes in weight were almost entirely accounted for by changes in carcass fat content, it would be more accurate to describe the responses as resetting the body fat set-point (8). Initially it was proposed that the VMH contained a satiety center that inhibited activity of a hunger center in the LH. This hypothesis was supported by observations that discrete knife cuts between the VMH and LH could increase food intake, and that electrical stimulation of the LH stimulated feeding whereas stimulation of the VMH inhibited food intake (see refs 1, 23).
A more exacting investigation demonstrated that obesity after VMH lesions was not simply secondary to hyperphagia, but that the lesions produce a syndrome of behavioral, metabolic, and physiological changes that include aggressive behavior, hyperphagia if food is easily available (but decreased motivation to work for food), decreased gastric motility, and increased insulin response to food (see refs 1, 23). Subdiaphragmatic lesions of the vagus nerve in VMH-lesioned rats arrest or reverse many of these changes. Gastric acid and insulin secretion are normalized, and hyperphagia and obesity are arrested if the rats are offered a commercial pelleted diet (24). The sparing effect of vagotomy in VMH-lesioned rats led to the development of an autonomic hypothesis which proposed that obesity caused by hypothalamic damage was secondary to abnormal control by the autonomic system of peripheral metabolism, and that the resulting hyperinsulinemia stimulated food intake and promoted fat deposition (25). However, as VMH-lesioned vagotomized rats will still become hyperphagic and obese when offered palatable diets (26, 27), the dependence of VMH-lesioned obesity on hyperinsulinemia has been questioned. Whether or not the autonomic nervous system metabolism plays a dominant role in the development of VMH obesity, it is apparent that the syndrome cannot be accurately described as a disruption of specific hunger or feeding centers acting as the central control in the feedback regulation of energy balance. Results from studies using more defined lesions indicated that the dual center hypothesis was too simplistic, and that many areas of the central nervous system are involved in control of food intake and regulation of energy balance. However, various hypothalamic nuclei play a major role in monitoring and integrating peripheral neural and metabolic input which is transduced into efferent signals that modulate nutrient selection, food intake, body composition, and energy balance (Fig. 2). Selective knife cuts demonstrated that lesions confined to the ventromedial nucleus (VMN) had no effect on food intake. Damage rostral to the VMN that cut the ventral noradrenergic bundle caused hyperphagia and weight gain, and lesions of the paraventricular nucleus (PVN) caused the largest increases in both food intake and weight gain (28) independent of hyperinsulinemia (29). A series of studies by Leibowitz and co-workers (30) have demonstrated that the PVN plays a vital role in control of food intake and nutrient selection, as different neuropeptides have selective effects on
Abbreviations: VMH, ventromedial hypothalamus; LH, lateral hypothalamus; VMN, ventromedial nucleus; PVN, paraventricular nucleus; DMN, dorsomedial hypothalamus; 2-DTA, 2-deoxytetronic
acid;
3-DPA,
3-deoxypentanoic
acid;
CSF,
fluid; CCK, cholecystokinin; AP, area postrema; medial nucleus of the solitary tract; GAPDAH,
cerebrospinal NTS, caudal
glyceraldehyde
3-phosphate
dehydrogenase; CRF, corticosterone releasing factor; CC, central control; ArcN, arcuate nucleus; ME, median eminence; F, fornix; 3rd V, third ventricle; Opt, optical tract.
3312 Vol. 4 December 1990 The FASEBlournal HARRIS om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
the set-point concept. It seems more likely that there are multiple areas of the brain, including many in the hypothalamus, that monitor and integrate information from the periphery. Morley and Levine (31) suggest that this information is then transduced in a “cascade system involving multiple neurotransmittors.” Regulated
Figure
2. A schematic representation of a coronal section through the hypothalamus to depict locations of various structures: yentromedial hypothalamus (VMH), lateral hypothalamus (LH), fornix (F),
dorsomedial
nucleus
(DMN),
paraventricular
nucleus
(PVN), third ventricle (3rd V), arcuate nucleus (ArcN), median eminence (ME), and optical tract (Opt). The PVN lies in a plane approximately 1 mm anterior to the other nuclei.
feeding behavior. Of those that stimulate feeding, norepinephrine and neuropeptide Y produce a preference for carbohydrate consumption whereas opiates produce a preference for protein and fat. The feeding elicited by norepinephrine appears to depend on corticosteroids. In contrast, dopamine and serotonin inhibit carbohydrate intake. Leibowitz (30) suggests that the neuropeptide/PVN interaction is responsible for replenishing body carbohydrate stores at the beginning of the diurnal feeding cycle or in response to stress. The effects of PVN activity on energy balance appear to be secondary to regulation of short-term carbohydrate metabolism. Lesions of the LH cause aphagia and loss of body fat content to a new defended level (22). Vagotomy of LHlesioned rats reverses the inhibition of gastric acid secretion but causes a further lowering of body weight (24). Although there may be an autonomic component in the LH syndrome, lesions of dopaminergic fibers originating in the substantia nigra and ascending through the LH can replicate many of the responses seen with LH lesions, which implies an important role for dopaminergic transmitters in maintenance of food intake and body fat content (see ref 1). In contrast, lesions of the dorsomedial hypothalamus (DMN) result in smaller animals that have a normal body composition and a normal food intake per unit body weight. Bernardis (13) has described these animals as having reset the organismic set-point, which suggests that this area of the hypothalamus may be involved in determining growth and body size parameters rather than body composition or food intake. The brief summary provided here is not intended to be an exhaustive review of current work investigating the central control of food intake. It is intended to demonstrate that mechanisms known to influence food intake are too numerous and complex in function to fit
SET-POINT
THEORY
quantity
Aside from the indirect observations of stabilitiy of body weight there is little direct evidence to support the concept that total body fat is subject to regulation. Lipectomy studies in which part or all of defined fat depots are removed have provided equivocal results. Early studies by Leibelt et al. (32) demonstrated compensation in total body fat content of lipectomized gold thioglucose-treated obese mice. Similar results were also reported for lipectomized castrated male rats (33). When one epididymal fat pad was removed from normal Sprague Dawley (34) or lean Zucker rats (35), there did not appear to be any compensation for the loss of fat. Faust et al. (35) suggested that variability in the results could be explained if fat cell size rather than total body fat content was regulated. He proposed that individual adipocytes are resistant to accumulation of excess amounts of lipid, as there is a maximum diameter that can be achieved without compromising physical and metabolic functions. When portions of fat pads are removed, the fat will be replaced if it can be accommodated without significantly enlarging existing fat cells. This may be more easily achieved in animals that tend toward obesity and have a greater capacity for adipocyte proliferation. The hypothesis was supported by observations that feeding Sprague Dawley rats a high-fat diet that induced obesity facilitated compensation for lipectomy whereas a commercial pelleted diet did not (36). When the results could not be repeated with adult lipectomized Osbourne Mendel rats fed a high-fat diet (37), the authors suggested that the difference was attributable to genetic and environmental factors. As regulation of body weight has been demonstrated in a variety of species, it is anticipated that the regulatory mechanism would be less susceptible to strain difference. Mrosovsky (38) suggests that the differences in results from lipectomy studies may depend on which fat pad is removed, which is consistent with the site-specific differences in fat cell size, lipid metabolism, and blood supply. To reconcile this hypothesis with a system that regulates total body fat content, Mrosovsky (38) proposes that there is a common afferent system from all fat depots, but that site-specific responses to central signals are determined by differences in efferent neural supply. Secondary to observations that changes in diet composition can affect body weight, Flatt (39) proposed that body fat content may be determined by nutrient intake. He proposed that not only must energy balance be maintained but that nutrient balance must also be achieved by oxidizing the same mixture of nutrients as
3313
om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
is ingested. Carefully performed calorimetry studies with mice have shown that carbohydrate balance is achieved within a day, presumably because there are very small endogenous carbohydrate stores in the body. In the short-term, carbohydrate balance will be maintained even at the expense of fat balance, as there are large stores of fat in the body that can buffer short-term imbalance. The nutrient balance hypothesis can also be used to explain why high-fat diets induce obesity. When dietary fat is diverted into storage, adipocytes enlarge, turnover of fatty acids increases, and serum free fatty acid concentration rises. The rate of hepatic fatty acid oxidation is proportional to the concentration of fatty acids in the circulation; therefore, equilibrium and nutrient balance is achieved when fat cell size is large enough to stimulate fatty acid oxidation at a rate equal to ingestion. This hypothesis for the effect of nutrient balance on body weight is supported by observations that body fat content is proportional to dietary fat concentration. Recent studies examining the responses of individual animals to changes in dietary composition suggest that there is a genetic component involved in the response to high-fat diets. Some animals increase hepatic fatty acid oxidation more effectively than others and thus gain less weight on a high-fat diet (40, 41). Also, studies with fat substitutes (42) suggest that organoleptic properties of diet may play an important role in determining food intake and body composition of rats consuming greasy diets. Bray (43) has elaborated on Flatt’s hypothesis for nutrient balance to develop a model for the regulation of food intake in which the balance of each major nutrient is regulated separately with a time-constant proportional to the ratio of endogenous stores to the amount eaten each day. Carbohydrate has a short timeconstant and fat has a long time-constant for balance. Circulating concentrations of hormones and nutrients, and vagal afferents from the liver would inform the hypothalamus of peripheral nutrient status. The limited ability to store energy as carbohydrate would also make it more difficult to gain weight on a highcarbohydrate diet than on a high-fat diet, which is consistent with observations from overfeeding studies of a decreased efficiency of energy utilization with a highcarbohydrate compared with a high-fat diet (see ref 43). The concept that body weight or body fat content is regulated at a reference level is inadequate to explain observations that factors such as organoleptic properties of food and nutrient composition of the diet can influence body composition. Afferent
system
Kennedy (44) was the first to propose a lipostatic theory in which total body weight is maintained by regulating body fat content. He hypothesized that the hypothalamus sensed the concentration of a circulating metabolite to obtain information on the size of stores of body fat. The hypothalamus would then modulate the information to effect changes in food intake to compensate for changes in body fat content. This hypothesis was
supported by studies with parabiosed rats (45) in which two animals were united surgically to share a common blood supply. When one member of a pair was made hyperphagic and obese by VMH lesions, the partner appeared hypophagic and lost both lean and fat body mass. In a second study (46), using electrical stimulation of the LH to induce obesity, measurement of food intake of the individual members of the pairs indicated that the nonstimulated partner ate progressively less as obesity developed in the stimulated partner. These observations were interpreted as evidence for the hypothalamus responding to a circulating satiety factor that carried information on total body fat stores. The obese rat was unable to respond to the signal due to hypothalamic manipulation, but the partner responded by reducing food intake in an effort to correct energy balance. Although the parabiosis experiments provided evidence for feedback regulation of energy balance, it was impossible to determine whether food intake, body weight, total body fat content, or fat cell size was being monitored by the hypothalamus. Hervey (17) proposed that the circulating satiety-factor could be a fat-soluble agent, such as a steroid, so that as body fat content increased less of the factor would be in circulation; as fat depots decreased concentrations would rise, and the hypothalamus would stimulate feeding. Glucose is an obvious candidate as an afferent signal to the hypothalamus. Mayer (47) proposed that insulinsensitive cells in the brain monitored arterial-venous differences in glucose concentration as a measure of glucose utilization. The presence of glucoreceptors in the hypothalamus has been demonstrated by selective uptake of gold thioglucose. Also, vagal afferents from hepatic glucose receptors ascend via the NTS and into the LH where they may modulate neuronal activity (see ref 1). Oomura and colleagues (48) have demonstrated that activity of neurons within the VMH and LH respond selectively to physiological changes associated with hunger and satiety. Glucose or glucose plus insulin stimulate firing of glucoreceptors located in the VMH, whereas glucose depresses activity of glucosensitive cells in the LH. In contrast, fatty acids depress activity of VMH glucoreceptors and facilitate activity of LH glucose-sensitive neurons. As free fatty acids are at higher circulating concentrations during periods of hunger, the hypothalamic neurons respond appropriately for units involved in the control of food intake. Subsequent work demonstrated that the glucoreceptor and glucose-sensitive cells were responsive to most metabolites and hormones known to be involved in the control of feeding, which suggests that these cell bodies could be more appropriately called hunger-related and satiety-related neurons (49). More recently Oomura and colleagues (49) analyzed the blood of food-deprived rats by mass spectrometry to identify novel compounds that could be rela.ed to hunger and satiety. Two sugars, 2-deoxytetronic acid (2-DTA) and 3-deoxypentanoic acid (3-DPA), were identified as satiety and hunger factors. Infusion of micromolar concentrations of 2-DTA into the third ventricle of food-deprived rats inhibited feeding and
3314 Vol. 4 December 1990 The FASEBIou ma! HARRIS om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
neuronal activity in the LH whereas infusion of 3-DPA stimulated both feeding and nerve activity. The specific effect of these compounds on glucose-sensitive neurons in the LH suggest involvement in the control of feeding behavior. In addition to the sensitivity of hypothalamic neuronal function to circulating concentrations of nutrients and metabolites, there is some evidence that hypothalamic metabolism mimics peripheral energy utilization. Kasser et al. (50) have demonstrated in rats that during conditions of positive energy balance, fatty acid synthesis is increased in the VMH, and fatty acid oxidation is decreased in the LH. During periods of food restriction and negative energy balance, LH fatty acid oxidation is increased. These changes in glucose and fatty acid utilization were specifically localized in hypothalamic tissue and mimicked changes in hepatic nutrient utilization. In a subsequent study (51) hypothalamic metabolism was tracked in rats recovering from over- or underfeeding. LH fatty acid oxidation and body fat content returned to control levels simultaneously. These observations imply a correlation between peripheral energy balance and central nutrient utilization, and may also represent transduction of afferent neuronal and humoral information on peripheral energy status. McHugh and Moran (52) have demonstrated that there is negative feedback on food intake by vagal afferents from gastric stretch receptors. The primary feedback is on rate of gastric emptying; satiety appears to be a secondary response. In addition to neural afferent signals, there is evidence that circulating concentrations of nutrients and hormones may influence feeding behavior. Baskin et al. (53) propose that plasma insulin is a peripheral satiety signal in determining energy balance. It has been shown that insulin can cross the blood-brain barrier, and insulin receptors have been identified in various brain tissues. Chronic intraventricular infusions of insulin cause a dose-related reduction of food intake and body weight in baboons, and when there are chronic elevations of serum insulin, cerebrospinal fluid (C SF) concentrations of insulin also increase. Baskin et al. (53) suggest that insulin either modifies or acts in concert with other neurotransmittors known to influence food intake. Walls and Koopmans (54) have shown that intravenous infusion of glucose plus amino acids or free fatty acids depress food intake of fed rats. The rats compensated more accurately for the calories infused as glucose and amino acids than for those infused as lipid. Chronic central infusions of glucose or glycerol reduce body weight of rats by inhibiting food intake, whereas infusions of bhydroxybutyrate lowered body weight independently of a reduction in food intake (55). A more direct effect of gastric signals in the development of satiety has been suggested from observations that peptides released from the gastrointestinal tract may also play a role as central neurotransmittors. Cholecystokinin (CCK), released from the duodenum in response to distension and nutrient composition, has been shown to induce satiety whether administered
centrally or peripherally. The peripheral effect may be mediated by the vagus nerve (56). The sensory input from the subdiaphragmatic vagus terminates in the area postrema (AP) and caudal medial nucleus of the solitary tract (NTS), areas of the brain that are also accessible to circulating metabolites as there is neither a blood/brain nor blood/C SF barrier in the area. The observations on reversal of VMH obesity by vagotomy have been criticized for the possibility that food intake was depressed by inhibition of gastrointestinal motility. By using AP/NTS lesions it was possible to separate the importance of peripheral sensory vagal afferents in the control of food intake from vagal motor function, which was not affected by the lesions. AP/NTS lesions caused transient hypophagia and reduction of body weight to a new defended level (57). There was no exaggerated preference to a palatable diet in AP/NTS lesioned rats, which suggests that the diet-induced obesity in VMHlesioned vagotomized rats was secondary to loss of motor control of the gastrointestinal tract. Observations that food intake can be influenced by various nutrients, hormones, and peripheral neural stimuli indicate that there can be no single afferent input to the hypothalamus to determine body weight. However, all the different afferent signals could be considered to be factors impinging on the control in an open loop system. Efferent
system
Once the brain has integrated and transformed information on the size of body fat stores, an effector mechanism has to control food intake or energy expenditure to maintain energy balance. Efferent output may be neural or hormonal, and is probably a combination of the two. Genetically obese rodents have been popular models for examining the regulation of energy balance because obesity is the result of a single recessive gene. If one abnormal gene means one abnormal protein, theoretically it should be possible to identify the single abnormality that results in obesity and to gain insight into mechanisms that normally regulate energy balance. However, syndromes of genetic obesity tend to be complex, and it is impossible to determine which abnormalities are primary and which develop as secondary effects of obesity. For example, the obese Zucker rat is hyperphagic, hyperinsulinemic, hypertriglyceridemic, hypothyroid, hypothermic, hypersensitive to glucocorticoids, sterile, and has low rates of sympathetic activity in some organs (58). Therefore, genetic obesity may be more accurately described as one gene, one syndrome. Bray (43, 59) has suggested that some kinds of genetic obesity result from an autonomic/endocrine abnormality. It is proposed that obesity in genetic models such as Zucker rats and ob/ob mice, and in animals made obese by hypothalamic lesions, can be explained by hypersensitivity toward adrenocorticoid hormones that results in abnormal sympathetic tone. Food intake is controlled to maintain nutrient balance. The afferent humoral and neural signals are integrated and trans-
3315 SET-POINT THEORY om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
formed to neural efferent signals. One type of efferent output, modulated by the LH, controls activity that locates and identifies food. The second efferent output controls nutrient balance to maintain homeostasis. This output is mediated by changes in hypothalamic PVN concentrations of corticosterone releasing factor (CRF) such that increased concentrations of CRF increase sympathetic activity and inhibit food intake. In this model, circulating concentrations of corticosterone would have a negative feedback effect on PVN concentrations of CRF. Also, the corticosterone may modify glucose utilization by neurons in other hypothalamic nuclei. Support for this hypothesis comes from a variety of observations. Central administration of CRF depresses food intake. Young VMH-lesioned rats and genetically obese rodents have suppressed sympathetic tone. Conversely, norepinephrine turnover is increased in rats offered sucrose solution or cafeteria diets (see ref 59) and can be affected by isocaloric changes in nutrient intake (60). Obese Zucker rats appear to have exaggerated sensitivity toward the hyperphagic effects of corticosterone (see ref 23). Adrenalectomy of VMHlesioned or genetically obese rodents arrests hyperphagia and the development of obesity. In Zucker rats, adrenalectomy increases sympathetic activity in brown adipose tissue. All of these responses to adrenalectomy are reversed by replacement of glucocorticoids (see ref 59). Preliminary, nonquantitative evidence has suggested that hepatic mRNA for the lipogenic enzyme glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was depressed by adrenalectomy and recovered in a timedependent manner with glucocorticoid treatment in obese Zucker rats. There was no effect of adrenalectomy or glucocorticoid replacement on hepatic GAPDH mRNA isolated from the lean Zucker rat. Unpublished observations are reported for a similar response in adipose malic enzyme mRNA. Bray suggests that the defective gene in the Zucker rat normally produces a peptide that modifies corticoid stimulation of mRNA for lipogenic enzymes. In obese Zucker rats the protein is expressed at low levels, and thus there is an exaggeration of the anabolic action of glucocorticoids (59). The model proposed by Bray (59) indicates that abnormal glucocorticoid/autonomic interaction is the primary modulator of energy balance. However, some evidence suggests that this mechanism does not fully account for VMH-lesion or genetic obesity in rats. Adrenalectomy of obese rats does not correct either body fat content or food intake to the levels measured for lean littermates (see refs 1, 61, 62). When obesity of VMH-lesioned rats is arrested by vagotomy or Zucker rats are adrenalectomized, the animals will still become hyperphagic and obese if they are offered a palatable diet (26, 61). Adrenalectomized Zucker rats remain hypothermic and sterile (see ref 1), indicating that only some aspects of genetic obesity have been corrected. Also Fletcher and McKenzie (61) have reported that corticosteroid replacement in adrenalectomized Zucker rats not only reinstates hyperinsulinemia in obese rats but enhances, to a smaller degree, insulin release in
response to glucose administration in lean littermates, which suggests that the corticosteroid influences on in sulin release are intact in both phenotypes. Considerabi more work is required before it can be determined whether the time-dependent reversal in mRNA levels o lipogenic hormones is a primary response to the replacement of corticosteroid or whether it is secondary to the reinstatement of hyperphagia and hyperinsulinemia. In addition to sympathetic and adrenal system dysfunctions in models of obesity, there is some evidence that humoral factors play a role in determining body composition. Satietin is a glycoprotein that has been isolated from the blood of normal humans and rats. When satietin is infused centrally into rats, there is a transient
depression
of
food
intake
and
a sustained
depression of body weight, which suggests a decrease in efficiency of energy utilization (63). These observations indicate that blood-borne factors may play a role in regulating energy balance not only by controlling food intake but by adjusting energy expenditure. Parabiosis studies
in which
one
member
of a pair
is made
obese
by overfeeding leads to distinctly different results from those observed with genetically obese or VMH-lesioned obese models. It has been repeatedly demonstrated in a series of studies with Sprague Dawley rats (64) that when one rat is made obese by overfeeding by stomach tube, its partner makes a statistically nonsignificant reduction in food intake, retains a normal lean body mass, and loses a large portion of its body fat content. The loss of fat appears to be secondary to a specific inhibition of adipocyte fatty acid synthesis. It has been hypothesized that development of obesity in the overfed rat leads to the release of a blood-borne factor that is carried into the partner fed ad libitum. This rat responds to the increased body fat content of its partner by making ing further
a small reduction in food intake and inhibitdeposition of fat. Although the change in
food intake is not statistically significant, even when summated over 60 days, the energy deficit is large enough to account for the change in body fat stores. There are no measurable changes in hepatic fatty acid synthesis, but measurements of glucose flux through the pentose phosphate shunt and of lipogenic enzyme activity suggest that the liver is also sensitive to the antilipogenic factor. It is possible that the liver uses alternate pathways for maintaining fatty acid synthesis. The changes in body composition of members of parabiosed pairs is not dependent either on the diet used to induce obesity or on the composition of diet eaten ad libitum by the normal partner (unpublished observations). These parabiosis studies do not indicate whether the primary effect of the circulating factor is to reduce food intake or inhibit fatty acid metabolism. A bioassay has demonstrated that serum from obese overfed rats is capable of inhibiting fatty acid synthesis in isolated adipocytes in vitro (65). When fat cells are incubated with serum from overfed obese rats for 12 h, there is a significant
inhibition
of fatty
acid
synthesis
compared
with cells preincubated with serum from control rats. This assay indicates that the serum factor is capable of reducing fatty acid synthesis by acting directly on adi-
HARRIS 3316 Vol. 4 December 1990 The FASEBJournal om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
pocytes and does not necessarily require central mediation for its effect. In addition, further studies have demonstrated that the factor is produced by overfed hypophysectomized rats and is unlikely to be of hypothalamic or pituitary origin. The identity of the factor is unknown, but it has been demonstrated that a bloodborne dye takes 2 h to equilibrate in parabiosed rats (66), which indicates that any factor that is produced by one animal and is active in the partner must have a long half-life in the circulation. The anti-lipogenic factor appears to be released in response to a substantial increase in body fat content rather than to increased food intake. Whether it is involved in the regulation of total body fat content or in regulation of adipocyte size and number remains to be elucidated. CONCLUSION Although long-term can
body weight or body regulation of energy
be achieved
only
through
fat set-point is related to balance, this regulation modulation
of short-term
responses and sensitivities to factors that influence food intake and energy expenditure. There is evidence for involvement of different areas of the brain in the control of various aspects of feeding behavior, ranging from acute organoleptic preferences to long-term caloric compensation. In this context the set-point theory for regulation of body weight in which one specific measure related to energy balance is regulated appears
unrealistic.
It
seems
appropriate
to
assume
that the level at which body fat content is maintained represents the equilibria achieved by regulation of many parameters including nutrient balance, energy balance, and physical aspects of adipose tissue, such as fat cell number and size. Current evidence for the role of many different factors in the control of food intake and regulation of energy balance indicates that mechanisms responsible for determining body weight are far more complex than was once thought. In this context, having an inaccurate and oversimplified preconception of how the various mechanisms may be integrated could hinder progress by limiting the boundaries of acceptable concepts.
8. Bernardis, L. L., Luboshitsky, R., Bellinger, L. L., McEwen, G. (1982) Nutritional studies in the weanling rat normophagic hypothalamic obesity. j Nutr. 112, 1441-1455 9. Drewry, M. M., Harris, R. B. S., and Martin, R. J. (1989) effect of increased adiposity on food intake of juvenile
Physiol. Behav. 45, 381-386
16.
17. 18. 19.
20.
sis and lipolysis associated with recovery from reversible obesity in mature female rats. Proc. Soc. Exp. BioL Med. 191, 82-89 Hill, J. 0., Dorton, J., Sykes, M. N., and DiGirolamo, M. (1989) Reversal of dietary obesity is influenced by its duration and severity. mt. J. Obes. 13, 711-722 Hervey, G. R. (1969) Regulation of energy balance. Nature (London) 222, 629-631 Passmore, R. (1971) The regulation of body weight in man. Proc. Nutr Soc. 30, 122-127 Armitage, G., Harris, R. B. S., Hervey, G. R., and Tobin, G. (1984) The relationship between energy expenditure and environmental temperature in congenitally obese and non-obese Zucker rats. j PhysioL (London) 350, 197-207 Tremblay, A., Desprey, J-P., and Bouchard, C. (1988) Alteration in body fat and fat distribution with exercise. In Fat Distribution during Growth and Later Health Outcomes. (Bouchard, C. and Johnston, F. E., eds) pp. 297-312, Alan Liss Inc., New York
21. Hoebel, normal
B. E. (1986) Neurological regulation of body weight. Grit. Rev. Gun. Neurobiol. 2, 1-60 2. Mrosovsky, N., and Powley, T L. (1977) Set points for body weight and fat. Behav. Biol. 20, 205-223 3. Keesey, R. E., and Powley, T. L. (1986) The regulation of body weight. Annu. Rev. Psychol. 37, 109-133 4. Cohn, C., and Joseph, D. (1962) Influence of body weight and body fat on appetite of normal lean and obese rats. Yalej Biol. Med. 34, 598-601 5. Sims, E. A. H., and Horton, E. 5. (1968) Endocrine and metabolic adaptation to obesity and starvation. Am. j Gun. Nutr. 21, 1455-1470 6. Wilson, P. N., and Osbourn, D. R. (1960) Compensatory growth after undernutrition in animals and birds. Biol. R#{128}v.
Cambridge Philos. Soc. 35, 324-363 N. J., and Stock, M. J. (1982) Effects of feeding a palatable cafeteria diet on energy balance in young and adult lean (+1?) Zucker rats. Br. J. Nutr. 47, 461-471
7. Rothwell,
SET-POINT
THEORY
B. G., and
Teitlebaum,
P. (1966)
Weight
regulation
in
and hypothalamic hyperphagic rats. j Comp. PhysioL Psychol. 61, 189-193 22. Keesey, R. E., Boyle, P. C., Kemnitz, J. W., and Mitchel, J. S. (1976) The role of the lateral hypothalamus in determining the body weight set-point. In Hunger: Basic Mechanisms and Clinical Implications (Novin, D., Wywicka, W., and Bray, G., eds) pp. 243-255, Raven, New York 23. Bray, G. A., and York, D. A. (1979) Hypothalamic and genetic obesity in experimental animals. An autonomic and endocrine hypothesis. Physiol. Rtv. 59, 719-809 24. Powley, T L., and Opshal, C. A. (1976) In Hunger: Basic Mechanisms and Clinical Implications (Novin, D., Wywicka, W., and
Bray, G., eds) pp. 313-326, 25. Berthoud, tromedial
1. Levin,
The rats.
10. Garrow, J. S., and Stalley, S. F. (1975) Is there a set-point to human body weight? Proc. Nutr. Soc. 34, 84A 11. Payne, P. R., and Dugdale, A. E. (1977) Mechanisms for the control of body-weight. Lancet I, 583-586 12. Wirtshafter, D., and Davis, J. D. (1977) Set points, settling points and the control of body weight. PhysioL Be/wv. 19, 75-78 13. Bernardis, L. L., and Bellinger, L. L. (1987) The dorsomedial hypothalamic nucleus revisited: 1986 update. Brain Ret. Rev. 12, 321-381 14. Harris, R. B. S., Kasser, T. R., and Martin, R. J. (1986) Dynamics of recovery of body composition after overfeeding, food restriction or starvation of mature female rats. j NutE 116, 2536-2546 15. Harris, R. B. S., and Martin, R. J. (1989) Changes in lipogene-
sulinemia REFERENCES
and with
H.
R.,
and
Raven,
New York
Jeanrenoud,
and its reversal hypothalamus
B. (1979)
and
Frohman,
mediated hyperinsulinemia Be/wv. Rev. 6, 205-214
hyperin-
after lesions of the yenin anaesthetized rats. Endocrinology
150, 146-151 26. Sclafani, A. (1981) The role of hyperinsulinemia nerve in hypothalamic hyperphagia reexamined. 402-410 27. King, B. M.,
Acute
by vagotomy
L. A. (1982)
in hypothalamic
The
and the vagus Diabetologia 20, role
obesity.
of vagally Neurosci.
28. Gold, R. M., Jones, A. P., Sawchenko, P. E., and Kapatos, G. (1977) Paraventricular area: critical focus of a longitudinal neurocircuitry mediating food intake. Physiol. Be/wv. 18, 1111-1116 29. Sims, J. S., and Lorden, J. F. (1986) Effect of paraventricular nucleus lesions on body weight, food intake and insulin levels. Be/wv. Brain Rev. 22, 265-281 30. Leibowitz, S. F. (1986) Brain monoamines and peptides: role in the control of eating behavior. Federation Proc. 45, 1396-1403 31. Morley, J. E., and Levine, A. S. (1983) The central control of appetite. Lancet I, 398-401
32. Leibelt, R. A., Ichinoue, S., and Nicholson, N. (1965) Regulatory influences of adipose tissue on food intake and body
3317
om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
weight. Ann. NY Acad. &i. 131, 559-582 33. Bailey, J. W., and Anderson, D. B. (1980) Rate of fat compensation and growth efficiency of lipectomized Sprague Dawley rats. Nutr. 110, 1785-1792 34. Kral, J. G. (1976) Surgical reduction of adipose tissue in the male Sprague Dawley rat. Am. j Physiol. 231, 1090-1096 35. Faust, I. M., Johnson, P. R., and Hirsch,J. (1976) Noncompensation of adipose mass in partially lipectomized mice and rats.
51. Kasser, T R., Harris, R. B. of satiety: in vitro energy hypophagic and hyperphagic
j
Am. j 36. Faust,
Physiol. 231, 538-544 I. M., Johnson, P. R., and Hirsch, J. (1977) Adipose regeneration following lipectomy. Science 197, 391-393 37. Faust, I. M., Johnson, P. R., and Hirsch, J. (1979) Adipose tissue regeneration in adult rats. Proc. Soc. Exp. Biol. Med. 161, 111-114 38. Mrosovsky,
N. (1986)
Body
38, 407-414 39. Flatt, J. P. (1987) Dietary
41.
42.
43.
is regulated?
fat, carbohydrate
PhysioL Be/wv.
balance
and weight
effects of exercise. Am. j C/in. Nutr. 45, 296-306 Triscari, J., Nauss-Karol, C., Levin, B. E., and Sullivan, A. C. (1985) Changes in lipid metabolism in diet induced obesity. Metabolism 34, 580-587 Chang, S., Graham, B., Yakubu, F., Lin, D., Peters, J. C., and Hill, J. 0. (1990) Metabolic differences between obesity prone and obesity resistant rats. Am. J. PhysioL In press Jones, W. K., and Harris, R. B. 5. (1990) Physiological responses of mature rats fed low and high fat diets of similar texture. FASEBJ 4, A379 (abstr.) Bray, G. A. (1987) Obesity-a disease of nutrient or energy
balance?
46. Parameswaran,
S. V., Steffens,
A. B., Hervey,
G. R.,
and
de Ruiter, L. (1977) Involvement of a humoral factor in regulation of body weight in parabiotic rats. Am. j Physiol. 232,
R150.-R157 47. Mayer,
J. (1955) Regulation
the glucostatic theory Sci. 63, 15-43 48. Oomura, fatty acid
of energy
and body weight: and lipostatic hypothesis. Ann. NY Acad.
Y. (1976) Significance on the hypothalamic
intake
of glucose, insulin, and feeding and satiety neurons.
free In
Hunger: Basic Mechanisms and Clinical Implications (Novin, D., Wywicka, W., and Bray, G., eds) pp. 145-157, Raven, New York 49. Oomura, Y. (1987) Regulation of feeding by neural responses to endogenous factors. News PhymioL &i. 2, 199-203 50. Kasser, T. R., Harris, R. B. S., and Martin, R. J. (1985) Level of satiety: fatty acid and glucose metabolism in three brain sites associated with feeding. Am. J. Physiol. 248, R447-R452
P. R.,
and
Moran,
regulation of caloric ingestion ioL 236, R254-R259 53. Baskin, Dorsa,
R. J. (1989) Leve in brain dunn recovery. Am. j
D. G., Figlewicz, D. (1987) Insulin
T. H. (1978)
The
accuracy
in the rhesus monkey. Am. j
D. P., Woods, S., Porte, D., in the brain. Annu. Rev. PhysioL
of the
Physand
49,
335-347 E. K., and Koopmans, H. S. (1989) Effect of intravenous nutrient infusions on food intake in rats. Phytiol. Be/wv. 45, 1223-1226 55. Davis, J. D., Wirtshafter, D., Asin, K. E., and Brief, D. (1981) 54. Walls,
56. 57.
58.
59.
60.
Nutr Rev. 45, 33-43
44. Kennedy, G. C. (1953) The role of depot fat in the hypothalamic control of food intake in the rat. Proc. Royal Soc. B BioL Sci. 140, 578-592 45. Hervey, G. R. (1959) The effects of lesions in the hypothalamus in parabiotic rats. J. PhysioL (London) 145, 336-352
metabolism body weight
PhysioL 257, R1322-R1327 52. McHugh,
-
maintenance
40.
fat: what
S., and Martin,
61.
Sustained intracerebroventricular infusion of brain fuels reduces body weight and food intake in rats. Science 212, 81-83 Dockray, G. J. (1982) The physiology of cholecystokinin in brain and gut. Br. Med. BulL 38, 253-258 Hyde, T M., and Miselis, R. R. (1983) Effects of area postremalcaudal medial nucleus of solitary tract lesions on food intake and body weight. Am. j PhysioL 244, R577-R587 Martin, R. J., Harris, R. B. S., and Jones, D. D. (1986) Evidence for a central mechanism of obesity in the Zucker fatty rat (fa/fa). Proc. Soc. Exp. Biol. Med. 183, 1-10 Bray, G. A. (1989) 1989 McCollum award lecture. Genetic and hypothalamic mechanisms for obesity-finding the needle in the haystack. Am. j Glin. Nutr 50, 891-902 Vander Tuig, J. G., and Romsos, D. R. (1984) Effects of dietary carbohydrate, fat and protein on norepinephrine turnover in rats. Metabolism 33, 26-33 Fletcher, J. M., and McKenzie, N. (1988) The effect of dietary fat content on the growth and body composition of lean -and genetically obese Zucker rats adrenalectomized before weaning.
Br j
Nutr. 60, 563-569
62. Fletcher, J. M., and McKenzie, N. (1988) The parasympathetic nervous system and glucocorticoid-mediated hyperinsulinemia in the genetically obese (faJfa) Zucker rat.]. Endocr. 118, 87-92
63. Bellinger, L. L., and Mendel, V. E. (1988) Ingestion, body weight and activity of rats receiving repeated intracerebroventricular infusions of rat satietin. PhycioL Be/wv. 44, 445-452 64. Harris, R. B. S., and Martin, R. J. (1989) Physiological and metabolic changes in parabiotic partners of obese rats. In Hormones, Thermogenesis and Obesity (Lardy, H., and Straatman, F., eds) pp. 233-243, Elsevier Science Publishing Co., New York 65. Harris, R. B. S., Bruch, R. C., and Martin, R. J. (1989) In vitro evidence for an inhibitor of lipogenesis in serum from overfed obese rats. Am. j PhysioL 257, R326-R336 66. Nishizawa, Y., and Bray, G. A. (1980) Evidence for a circulating ergostatic factor: studies on parabiotic rats. Am. j PhysioL 239, R344-R351
3318 Vol. 4 December 1990 The FASEBJournal HARRIS om www.fasebj.org by Kaohsiung Medical University Library (163.15.154.53) on July 28, 2018. The FASEB Journal Vol. ${article.issue.getVolume()}, No. ${article.issue.getIssueNumber()},
