Leptin and Clinical Medicine: A New Piece in the Puzzle of Obesity

THE DISCOVERY of leptin in 1994 has provided a major new piece in the puzzle of obesity. Leptin is a 167-amino acid peptide made exclusively in adipose tissue. Its absence produces the obese (ob/ob) mouse, which is characterized by obesity, hyperphagia, hyperglycemia, hyperinsulinemia and insulin resistance, hypothermia, and infertility. Leptin corrects these defects. The leptin receptor is a member of the cytokine family of receptors with several splice variants. The intracellular signaling system for leptin appears to involve activation of the Jak-Stat system, with Stat-3 being the phosphorylated intermediate. Genetic defects in the leptin receptor produce the diabetes (db/db) mouse, which is phenotypically identical with the ob/ob mouse when the genes are expressed on the same background strain, and the obese fa/fa Zucker rat. The levels of leptin are directly related to the quantity of body fat, suggesting that it is a signal to the brain and other tissues about the adequacy or inadequacy of fat stores. Leptin receptors are distributed widely, including brain and many peripheral tissues, suggesting that this peptide may provide a wide range of tissues with information about fat stores. Although human obesity may not be a direct result of low leptin activity, understanding gained from the animal models may help in the development of new pharmacotherapeutics for human obesity. The year was 1950. The place was the Jackson Laboratory in Bar Harbor, ME, where George W. Snell was a leader in mouse genetics. In that year Dickie et al. (1) reported the appearance in their colony of a recessive mutant associated with massive obesity. The genetic defect in the obese or ob/ob mouse is a recessively inherited disease manifested early in life that is associated with diabetes. Shortly after discovery of the ob/ob mouse, a second recessively inherited form of obesity, called diabetes or db/db, was described by Hummel and Coleman (2). Rounding out the discovery of recessively inherited forms of obesity were the discovery of the fatty rat (3) and, somewhat later, tub and fat mice (4). The fact that ob/ob and db/db mice are phenotypically identical when expressed on the same genetic background yet genetically different was initially interpreted as a genetic defect at two different steps in an enzymatic cascade (5). The search for a biochemical or physiological mechanism for the genetic basis in these obese animal models occupied many scientists (5), but has provided only one major piece to the puzzle of obesity. This piece of the puzzle was the finding that carboxypeptidase E, which cleaves propeptides to biologically active peptides, was defective in the fat mouse (6). The observation that all animal models of obesity have high levels of insulin and insulin resistance lead to extensive studies of insulin-responsive tissues and insulin receptors (5). The possibility of a fat storage disease similar to the glycogen storage diseases was also explored. Whatever the defect that was eventually identified, it had to account not only for the metabolic abnormalities of hyperphagia, hyperlipidemia, hyperglycemia, insulin resistance, and diabetes, but also for the infertility and hypothermia that characterize these animals (2, 5). One of the intriguing questions that followed the discovery of the ob/ob mouse was its differences from the obesity produced by damage to the medial hypothalamus. Hetherington and Ranson (7) added another piece to the puzzle when they showed that damage to the ventromedial hypothalamus (VMH) would routinely produce obesity. The paraventricular nucleus and central nucleus of the amygdala are other key sites. In addition to lesions, local anesthesia of the ventromedial nucleus (VMN) with procaine will produce hyperphagia and obesity. Similarly, injection of colchicine, which inhibits the cell cycle, or injection of monosodium glutamate, goldthioglucose, or ibotenic acid, which damage cells in the VMH, will all produce obesity (5). The next piece of the puzzle came from the observation that the obesity would develop in the VMH-lesioned rat and in the ob/ob mouse even when food intake was precisely matched to that of lean controls (2, 5). If hyperphagia were allowed, the obesity was of greater magnitude, but clearly hyperphagia was not essential for obesity in either model. A fourth piece of the puzzle of obesity was the demonstration that a disturbance of the autonomic nervous system appeared to be essential for the development of obesity in most animal models, including the ob/ob/mouse, the fa/fa rat, the VMH-lesioned animal, and probably humans as well (8). Each model has low levels of sympathetic activity to brown adipose tissue, a thermogenic tissue. They all show an increased activity of the parasympathetic nervous system, which may play a key role in the development of obesity. When pancreatic islets of rats were removed from vagal innervation by transplanting them under the renal capsule, the hyperphagic effect of VMH lesions was almost entirely prevented. Thus was born the autonomic hypothesis (5), which suggested that reduced activity of the sympathetic nervous system and increased activity of the parasympathetic nervous system played central roles in the development of obesity in the fatty (fa/fa) rat, the VMH-lesioned rat, and the ob/ob mouse (8). This hypothesis has received subsequent support from both animal and human studies. The low levels of sympathetic activity supplying brown adipose tissue offered a plausible explanation for the lowered body temperature in ob/ob mouse and fa/fa rats (5). A sixth piece of the puzzle came from studies of the endocrine system and has been called the endocrine hypothesis (5), although the glucocorticoid hypothesis might be more appropriate. Removal of the adrenal cortex is sufficient to impair the development of all experimental obesities. The obesity of the ob/ob mouse, the db/db mouse, and the fatty rat does not progress after adrenalectomy. More striking, food intake returns to normal, muscle mass increases, body growth resumes, insulin resistance is eliminated, and hyperglycemia abates after adrenalectomy. Thus, the metabolic components of the ob/ob mouse, db/db mouse, fa/fa rat, and yellow mouse are dependent on adrenal glucocorticoids. The infertility, however, is not restored by adrenalectomy, suggesting that there is another function of the ob/ob and db/db gene products. Another piece in the puzzle of obesity was the demonstration that estrogens altered body fat. Castration of the female rodent increases food intake and decreases the activity of the sympathetic nervous system (9). Injection of estrogens or their direct application to hypothalamic structures will reduce food intake. Estrogen may thus serve as a modulator of the messages produced by the medial hypothalamus that control feeding signals, as part of the integration of feeding and the reproductive axis. It is noteworthy that adrenalectomy prevents the obesity of castration just as it does in leptin deficiency and VMH-lesioned induced obesity (9). Based on his studies of parabiotic animals, Coleman proposed that the ob/ob mouse might be deficient in a circulating factor and that the db/db mouse might not respond to this factor (2). These prescient experiments were given a plausible biochemical basis by the discovery of leptin and the leptin receptor. The identification of leptin shows the power of molecular cloning techniques in identifying additional pieces to the puzzle of obesity. Large numbers of animals carrying the ob or db gene were mated with the Mus castaneus, a distantly related mouse. With these large numbers of crosses and backcrosses, it was possible to narrow the genomic distance to within centimorgans of the gene. This essential preliminary work allowed Friedman and his colleagues to complete the cloning of the ob gene by using yeast artificial chromosomes (10). This major piece in the puzzle of obesity was published in December 1994. The importance of the initial work by Leibel and his colleagues to the eventual cloning of the ob gene deserves greater recognition than it has generally been given to date. The primary product of the ob gene, now known as leptin, is a 167-amino acid protein. Northern blot or reverse transcription-PCR analysis of the messenger ribonucleic acid (mRNA) for the ob gene showed that it was expressed only in adipose tissue (10, 11). This was an important piece in this puzzle because it provided the basis for a selective message arising in adipose tissue that could signal other tissues and the brain about the state of the body fat, as proposed initially by Coleman (2). Although such a factor may regulate food intake, it may also be particularly important in the modulation of reproduction. Such a hypothesis was suggested initially by Frisch et al. (12) as the critical fat level hypothesis for the onset of puberty and menstruation. Demonstration that the Ob protein or leptin was biologically active was the next step. Three laboratories simultaneously reported the results of administering leptin (13–15), a word derived from the Greek word leptos, meaning thin. Administration of leptin to ob/ob mice reduced food intake and body fat, reduced glucose, and decreased insulin and has recently been shown to increase the activity of the sympathetic nervous system (16). Thus, leptin was acting, as do other peptides, to regulate reciprocally the activity of the sympathetic nervous system and food intake (17). Of particular importance for the viability of leptin as the key to the obesity in the ob/ob mouse was the fact that the db/db mouse was completely resistant to leptin (15), as Coleman had predicted, and that leptin cured the hypothermia and infertility of ob/ob mice. With these data confirming that leptin was the ob gene product, a big piece of the puzzle was in hand as the search for the biological realities of leptin began. Leptin is produced exclusively in fat cells across a wide range of animal species, including humans (10, 18). This finding has added to the growing literature showing that the fat cell is an important secretory organ in addition to its role as a storage organ for fat. Figure 1 shows a cartoon of leptin secretion and action. Diagram showing the release and action of leptin. Leptin is made exclusively in fat cells. Glucocorticoids and insulin enhance this process. Circulating leptin is bound to receptors that may be the extracellular part of the leptin receptor, which is shown here in the brain. Leptin may act on the brain receptors or be transported into the brain, where it alters NPY, food intake, the sympathetic nervous system, and the reproductive system. The leptin that is secreted by the fat cell appears to bind to one or more proteins in the circulation (19) (Fig. 1). The percentage of free leptin is higher in obese than in lean individuals (19). Indeed, plasma leptin levels are closely correlated with body fat in both humans and rodents (18–25), although there is a wide range of individual leptin values at a specific level of body fat. To date, no obesity in humans has been found to result from a defect in the reading frame of this gene (26, 27). However, linkage of obesity to the human leptin gene region or to regions flanking the leptin gene have been reported in massively obese subjects (28, 29). The increase in leptin levels with increasing body fat and obesity suggests that obesity may be associated with leptin resistance. Complementing the discovery of leptin was the new piece to the puzzle of obesity obtained by cloning of the leptin receptor. Initially, the short form (Ra) of the receptor was identified from a complementary DNA library of choroid plexus (30), but subsequently the long form (Rb) and a number of other splice variants were cloned and sequenced by several groups (31–35). The receptor for leptin is a member of the cytokine receptor family (36–38). A mutation in the intracellular signaling domain of the receptor was identified in db/db mice (31, 32) to confirm the original experimental data from parabiosis (2) and identify the molecular basis for a failure of db/db mice to respond to either endogenous or exogenous leptin. The db and fa mutations were thought to be syntenic. However, the fa mutation was shown to be a glutamine→proline substitution in the extracellular domain of the receptor. In contrast, the db mutation is a G to T substitution that produces a new consensus splice donor site. This results in the insertion of a 106-base sequence of the terminal exon of the short form of the receptor into the mRNA of the long form of the receptor and introduces a premature stop codon that results in a receptor protein lacking the intracellular signaling domain (33–35). The fa/fa rat, but not the db/db mouse, will respond to exogenous leptin when it is delivered intracerebroventricularly (39). The fa mutation may either impair transport of leptin or reduce the number of receptors that are localized in the plasma membrane. The biological effects of leptin are thought to result from the activation of a Jak-Stat signaling pathway, with specific involvement of Stat-3 protein (36, 40). However, it is not clear whether this is the only signaling pathway that is activated by leptin. The long form of the receptor (Rb) has been identified in several brain regions and may also be present in a range of peripheral tissues, including the liver and pancreas (31, 32, 34, 41, 42). Within the brain this receptor is expressed in a number of brain regions that have been associated with the regulation of feeding behavior and energy balance, such as the arcuate nucleus and VMH (43, 44). The latter observations may fit two pieces of the puzzle together by explaining the similarities between syndromes of leptin deficiency and the syndrome that results from VMH lesions. However, the predominant form of the receptor in peripheral tissues is the short splice variant (Ra). The high level of expression of this receptor in the choroid plexus (30) led to the suggestion that its major function may be the transport of leptin into the brain. If this is so, it is not clear why this form of the receptor is so widely distributed in peripheral tissues. Leptin secreted from adipocytes may be bound to a number of different proteins in the circulation (19), including a splice variant of the receptor (Re) that has no transmembrane domain and is soluble. However, only approximately 50% of total leptin is bound in lean individuals, and the level of free leptin is higher in obese individuals than in their lean counterparts (19). This discounts the possibility that any leptin resistance might result from excessive binding of leptin to circulating proteins. However, leptin resistance does appear to be peripheral, as dietary obese mice retain their sensitivity to central leptin but lose the response to peripheral leptin (45). The studies of leptin physiology have used two major approaches. The first has been the regulation of leptin production, measured as tissue mRNA levels (mainly in rodents) or circulating leptin levels (mainly in humans). The second approach has examined the responses to exogenous leptin. These studies have focused principally on food intake, body composition, insulin secretion, and the reproductive system, although effects on adipose tissue, the hematopoietic system, and other targets are now being studied. Both human and animal studies have demonstrated the close association among body fat, leptin mRNA, and plasma leptin levels (18–25). The level of expression of the leptin gene varies from one fat depot to another. The circulating leptin levels increase with percent body fat (r = ∼0.9), but not with visceral fat. Plasma leptin levels are higher in men than in women (15, 20, 46, 47) and show a diurnal rhythm in both sexes (48). Leptin production is affected by several hormones. Leptin gene transcription and circulating leptin levels are modulated by adrenal steroids (49, 50). Both in vivo and in vitro studies in rodents and man have shown that glucocorticoids enhance leptin gene transcription and leptin levels (49, 50). The generally higher levels of leptin in human females than males suggest that sex steroids might also affect leptin production. These effects are generally similar in rodents and humans. Similarly, insulin stimulation of leptin production has been demonstrated in rodents (51, 52) in a range of studies that have included cultured adipocytes and streptozotocin-diabetic rats. In human studies, leptin levels have been correlated with basal insulin levels (46, 47, 54). In man, insulin increases plasma leptin after a delay of 2–4 h (Saad, M. F., unpublished observation), although not all investigators have observed this acute response (21, 55, 56). Plasma leptin levels respond slowly to fasting (22, 57) and do not begin to decrease in humans for 12–14 h. Leptin gene transcription and plasma leptin are severely reduced by longer starvation (24, 58). Conversely, the increase in leptin after feeding is delayed. The nighttime increase in leptin that was initially described by Considine and colleagues (18) appears to be a delayed response to the last meal of the and probably insulin It has been that secreted leptin as a signal to the central nervous system to the level of body fat and the responses of food intake and energy However, et al. suggested that leptin may act as a starvation such that low levels the to respond to In a similar the effects of leptin on the reproductive system suggest that it may act to the to reproductive and a state Leptin has a in the of approximately a that would be with it being the parabiotic factor proposed by Coleman (2, This to that of 12–14 h in plasma may differences in renal The of high levels of leptin in obesity as a of leptin resistance one the effects of leptin as with regulation of food intake. If the leptin signal a of the quantity of fat for of the reproductive system, and the hyperphagia of leptin deficiency results from an to increase fat for the high levels of leptin would be to a different system. This system is the reproductive system. It is possible that the high levels of leptin produce their effects on the reproductive system in the feeding system. When leptin is either or into the brain, it will reduce food intake, particularly in the animal One explanation for the effects of leptin on the brain may be the in the levels of The observation that food intake and produced obesity with is another piece in the puzzle of obesity from in the arcuate nucleus and is in the paraventricular nucleus The decreases in and secretion produced by leptin or a in sensitivity to could account for the decrease in food intake, the increased sympathetic activity, and the of in the ob/ob mouse. However, leptin is in animals with a of NPY, suggesting that is not essential for the action of leptin leptin may decrease food intake release of other peptides, of which and are studies of rodents and identified the and targets that be affected by leptin. In addition to food intake, an autonomic that insulin secretion and brown adipose tissue was such that the mutant rodent would obese the onset of hyperphagia or when hyperphagia was prevented. Thus, leptin enhance the sympathetic to a number of tissues, including reduce the parasympathetic vagal to the endocrine and A in insulin levels by leptin, in of that related to the in food intake has been as has an increase in sympathetic activity in of in ob/ob mice has also been demonstrated studies have shown the to puberty with leptin and suggest that leptin may act to the of secretion Figure is an to leptin into a The absence of leptin, or the leptin receptor, VMH lesions, and of have similar They increase food intake, reduce sympathetic activity, increase parasympathetic activity, and reproductive It has long been known that the metabolic defects in the ob/ob mouse and other rodent including that by could be by or the effects of adrenal glucocorticoids The of these experimental models suggests that they all act to reduce the from the ventromedial hypothalamus. of hypothalamic effects of leptin and other experimental that produce obesity. adrenalectomy is in or the of all of these syndromes of obesity, including lesions of the and the effects of glucocorticoids be from the (Fig. This model also suggests that the effects of the sympathetic nervous system, the parasympathetic nervous system, and glucocorticoid levels may the peripheral level of leptin mRNA in and brown adipose tissue. into the role of adrenal steroids suggest that they might either leptin transport into the brain or the signaling associated with leptin the on the it is that leptin may have direct effects on other tissues. These may adipose tissue, as leptin inhibits and in cultured cells and the endocrine pancreas 42). The expression of leptin receptors in all peripheral tissues suggests that it may be important in the regulation of tissue of its central effects to hypothalamic In as with other that there may be a associated with excessive levels of leptin. Indeed, et al. suggested recently that the effects of leptin may to insulin resistance. The of these observations is as leptin in animal studies, suggesting an in insulin The identification of genes that obesity in rodents has added several new pieces to the puzzle of obesity and has provided an to the search for human obesity Although no human have yet been related to mutations in the regions of the rodent obesity in the regulation of these genes is The of human obesity has been to account for of the in body fat, but the number of genes may be understanding of the rodent obesity genes and identification of human obesity genes may well the development of new for the of obesity.