Insulin and its evolving partnership with leptin in the hypothalamic control of energy homeostasis

Transcription

1 Review TRENDS in Endocrinology and Metabolism Vol.15 No.8 October 2004 Insulin and its evolving partnership with leptin in the hypothalamic control of energy homeostasis Kevin D. Niswender 1, Denis G. Baskin 2 and Michael W. Schwartz 3 1 Diabetes, Endocrinology and Metabolism, 722 Preston Research Building, Vanderbilt University Medical Center, 2220 Pierce Avenue, Nashville, TN , USA 2 Division of Endocrinology/Metabolism, VA Puget Sound Health Care System, 1660 South Columbian Way, Mail Stop 151, Seattle, WA 98108, USA 3 Metabolism, Endocrinology and Nutrition, Harborview Medical Center and University of Washington, Box , 325 9th Avenue, Seattle, WA 98104, USA Despite an alarming increase in the burden of obesity worldwide, body adiposity seems to be a regulated physiological variable. Regulation of adiposity occurs through a classical endocrine feedback loop, in which the pancreatic b-cell-derived hormone insulin and the adipocyte-derived hormone leptin signal the status of body energy stores to the hypothalamus. Recent advances in our understanding of the signal transduction mechanisms used by insulin and leptin in the hypothalamus to modulate neuronal firing suggest that intracellular cross-talk occurs at several levels and is a potentially important determinant of regulated body weight. These pathways are thus an attractive target for pharmacological intervention in the treatment of obesity. Obesity currently qualifies as a worldwide health epidemic [1,2] and is a major impetus driving research on the mechanisms of body weight regulation. Much scientific work has yielded an impressive understanding of the role of specific molecules that are essential for energy homeostasis but, unfortunately, has yet to generate substantial advances in the treatment of obesity [3,4]. The concept of adiposity negative feedback, whereby hormonal signals generated in proportion to fat mass and recent energy balance act in the brain to reduce food intake and body weight, is fundamental to the biology of energy homeostasis. Although leptin is the prototypic adiposity signal, insulin qualifies as a second adiposity signal that has received relatively less attention. In this review, we highlight both historical and recent findings that support a central role for insulin and insulin signal transduction pathways, in relation to the action of leptin, in the central nervous system (CNS)-mediated regulation of energy homeostasis. In so doing, we highlight areas where key issues remain and outline hypotheses for the pathophysiology of obesity. Corresponding author: Kevin D. Niswender (kevin.niswender@vanderbilt.edu). Available online 23 August 2004 Insulin as a central effector of energy homeostasis Although the cloning of leptin [5] and the elucidation of its crucial role in energy homeostasis [6] were seminal events in energy homeostasis research, Woods, Porte and their coworkers [7 9] demonstrated the existence of neural circuits that respond to insulin and regulate food intake and adipose mass many years before leptin was discovered. Work conducted in the late 1970s and early 1980s revealed that chronic infusion of insulin into the cerebral ventricles of baboons reduces food intake and body adiposity [7], providing crucial early evidence in support of the feedback loop model of energy homeostasis on which current models are based [4]. Indeed, insulin shares characteristics with leptin that are requisite for an adiposity signal: it is secreted in proportion to body fat mass [10] and interacts with key neurons found in the hypothalamic arcuate nucleus (ARC) that express both insulin [11,12] and leptin [13] receptors (Figure 1). Although leptin is recognized as a principal regulator of fat mass, the concept that insulin functions as a catabolic adiposity negative feedback signal through the regulation of key hypothalamic neurons is relatively less appreciated, even though the strength of evidence in support of this hypothesis is equally as robust as that in support of the role of leptin. Part of the reason for this might be that the notion of insulin as a centrally acting catabolic hormone runs counter to its well-recognized role as the prototypical anabolic (and glucoregulatory) hormone in peripheral tissues. For example, administration of insulin can lead to hypoglycemia and counter-regulatory increases in food intake and, if repeated, can cause weight gain via this mechanism [14]. In addition, insulin functions in the periphery to promote storage of energy in the forms of carbohydrate, protein and fat. A combination of these factors probably explains the general clinical observation that insulin usage can lead to weight gain in individuals with diabetes [15], rather than the weight loss predicted to occur with the administration of an adiposity signal. As we elaborate in this review, the idea that central insulin action is fundamentally catabolic (i.e. reducing /$ - see front matter Q 2004 Elsevier Ltd. All rights reserved. doi: /j.tem

2 Review TRENDS in Endocrinology and Metabolism Vol.15 No.8 October (a) (b) Second-order and downstream neurons 1 r Neuron Mc4r Food intake Energy expenditure 3V lepr/insr Third ventricle (3v) NP/ AGRP + POMC lepr/insr First-order neurons Arcuate nucleus c (c) Insulin leptin Pancreas 3V Adipose tissue Figure 1. Neurons in the arcuate nucleus (ARC) express insulin and leptin receptors and integrate peripheral signals to maintain energy homeostasis. (a) According to current models, the ARC contains anabolic neurons coexpressing neuropeptide (NP) and agouti-related protein (AgRP) and catabolic neurons expressing pro-opiomelanocortin (POMC). Both types of cell respond to insulin and leptin but do so in the opposite manner: NP/AgRP neurons are inhibited and POMC neurons are activated. It is thought that this coordinate dual regulation of opposing cell types ultimately regulates complex responses, such as changes in feeding and energy expenditure, to maintain body fat stores. In this model, insulin and leptin have similar roles. Reproduced, with permission, from Ref. [70]. (b) Radiolabeled insulin administered to rats binds to neurons found in the ARC (oval outline), as well as various other brain areas. Image is in pseudocolor with highest intensity binding indicated in purple. (c) Radiolabeled leptin administered to rats binds primarily to neurons found in the ARC (oval outline). Image is in pseudocolor with highest intensity binding indicated in purple. food intake and body weight), whereas its peripheral actions are anabolic (i.e. increasing energy storage and potentially increasing body weight if an individual has too much insulin), is ultimately consistent with, and illustrative of, the concept of an endocrine feedback loop in the control of energy stores [4]. Like most physiological systems, the peripheral and central actions of insulin are balanced and function to promote euglycemia, optimal body composition and reproductive fitness. The opposing central and peripheral effects of insulin are evident in the condition known as diabetic hyperphagia, a prominent behavioral manifestation of uncontrolled human type I diabetes or experimental diabetes in animals. When b cells have been destroyed, the absolute lack of insulin results in hyperglycemia and an inability to store energy in peripheral tissues, ultimately causing a marked loss of body weight. The lack of catabolic insulin action in the brain, coupled with the decrease in serum leptin that accompanies weight loss, however, leads to marked increases in food intake that compensates, in part, for energy wasting and prevents a more rapid demise. Stated differently, if the catabolic tone of hypothalamic insulin and leptin action were not relieved by the absence of these hormones, weight loss would be greater and survival impaired. In fact, either insulin or leptin given systemically or directly into the third cerebral ventricle of rats with streptozotocin-induced diabetes blunts or completely blocks this compensatory hyperphagia [16,17] and results in a more rapid loss of weight. The fact that either hormone given alone can block diabetic hyperphagia is evidence of overlap in their functions in the CNS. A brief comment on the access of insulin, as a central effector of energy homeostasis, to the CNS is warranted. The hypothesis that circulating insulin provides a physiological signal to brain areas involved in energy homeostasis implies a mechanism whereby this peptide traverses the blood brain barrier (BBB) to enter brain interstitial fluid. Early studies demonstrated that, during intravenous infusion, increases in plasma insulin are paralleled by proportionate (albeit much smaller in absolute terms) increases in cerebrospinal fluid insulin levels in experimental animals [18] and humans [19]. Subsequent studies showed that this insulin uptake process involves a saturable transport mechanism [20] and, combined with the findings that insulin receptors are expressed by endothelial cells including those present in brain capillaries [21], and that these receptors are capable of transendothelial insulin transport [22], led to the conclusion that insulin delivery across the BBB involves an active, receptor-mediated mechanism. Specifically, plasma insulin is proposed to bind to its receptor on the luminal endothelial surface and, after being internalized into an endocytic vesicle (still bound to receptor), is ultimately released intact from the abluminal surface into interstitial fluid [21,23]. The observation in dogs that insulin transport into the brain is attenuated by the consumption of a high-fat diet [24] raises the possibility that acquired reductions in brain

3 364 Review TRENDS in Endocrinology and Metabolism Vol.15 No.8 October 2004 insulin delivery might respond with an increase in body weight. The extent to which a regulated transport mechanism mediates insulin delivery to specific key neuronal targets involved in energy homeostasis is, however, uncertain. This is an especially important issue with respect to the hypothalamic ARC. This brain area is situated adjacent to the median eminence, which lacks a BBB, and available data have yet to determine whether or not ARC neurons are protected by the BBB [25,26]. Indeed, capillary permeability is reported to vary extensively across the ARC itself, being greater in its mediobasal than its lateral aspect [27]. Thus, additional studies are needed to clarify whether circulating insulin gains access to ARC neurons by simple diffusion or via a more complex, regulated transport mechanism. Current model of energy homeostasis: proposed role of insulin A current and well-accepted model of the regulation of adipose tissue stores has been reviewed in detail elsewhere (Figure 1a) [3,4]. In this model, both leptin, the adipocyte-derived hormone, and insulin, the pancreatic b-cell-derived hormone, are proposed to function as adiposity signals in the feedback regulation of adipose mass [3,4]. Both hormones circulate in the bloodstream in proportion to body fat mass [10,28] and regulate the activity of neurons found in regions of the brain associated with body weight regulation. Neurons found in the hypothalamic ARC express receptors for, and bind to, both insulin (Figure 1b) and leptin (Figure 1c) [12,13], and are increasingly well-characterized targets of both hormones [29 31]. ARC neurons project to other key brain areas and are thought to function as primary neurons in a series of neural circuits that regulate food intake, energy expenditure, hypothalamus pituitary function and sympathetic outflow [32]. Situated in the ARC are at least two distinct neuronal cell types (Figure 1a) that respond to adiposity signals and exert opposing effects on energy balance: one is anabolic in nature, whereas the other is catabolic [4,31]. Anabolic neurons are those that, when activated, promote an increase in food intake and a decrease in energy expenditure, leading to the storage of energy. In the ARC, these neurons coexpress the orexigenic neuropeptides neuropeptide (NP) and agouti-related protein (AgRP). NP/AgRP neurons are inhibited by rising leptin and insulin levels, and this inhibition is thought to mediate some of the anorectic actions of these hormones. For example, insulin infusion directly into the third cerebral ventricle reduces expression of the NP gene (a marker of neuronal activation) in fasted or STZ diabetic rats, which in effect blocks the increase in NP gene expression that would otherwise occur in these rats [33]. Conversely, antisense RNA knockdown of hypothalamic insulin receptors causes an increase in both food intake and expression of the NP and AgRP genes [34]. These findings add physiological relevance to the observation that insulin given intracerebroventricularly (i.c.v.) reduces food intake in baboons [7], sheep [35] and rodents [36,37] by suggesting that tonic action of insulin to inhibit the production and release of a potent orexigenic neuropeptide normally constrains food intake. Furthermore, in an isolated neuronal preparation, the firing rate (and thus release of neuropeptide) of a subset of ARC neurons proposed to be NP/AgRP neurons was found to be reduced by insulin treatment and similar effects were seen after leptin treatment [29,30]. The effect of both insulin and leptin to hyperpolarize this population of hypothalamic neurons is dependent on signaling through phosphatidylinositol 3-kinase (PI3K), a classical insulin signaling target, and on the opening of ATP-sensitive K C channels [29,30]. Acting in opposition to the NP/AgRP neuron is the catabolic neuron, which expresses pro-opiomelanocortin (POMC). According to current models, these cells release the anorexic neuropeptide a-melanocyte-stimulating hormone when stimulated by insulin and leptin, and thereby promote a decrease in food intake and body weight. Insulin, like leptin, increases POMC gene expression when given i.c.v. [38] and activates downstream melanocortin target neurons [38]. Thus, POMC and NP/AgRP neurons comprise two primary but opposing sets of neurons found in the ARC that are reciprocally regulated by both insulin and leptin, establishing a system of dual anabolic and catabolic regulation (Figure 1a) [4]. These ARC neurons are especially interesting because of their proposed role as the primary neurons that respond directly to insulin and leptin and that sense input related to body fuel stores to regulate, in turn, downstream neurons and circuits [4]. Many other brain areas are crucial targets of ARC neurons: the identification of these areas is key, because it is ultimately the net activity of these neural circuits that integrates various inputs pertinent to energy homeostasis [32]. Notably, both insulin and leptin have been implicated in the cognitive aspects of feeding (reward, learning and memory) via receptors in cortico-limbic sites [39], an emerging area of interest that could be particularly relevant to populations that become obese in association with the consumption of highly palatable diets. Experimental models supporting a role for insulin First proposed in the 1970s, the hypothesis that insulin acts in the CNS to control energy homeostasis has enjoyed a resurgence of interest in response to several recent models of insulin action in the CNS. Interestingly, nematodes such as Caenorhabditis elegans express homologs of insulin, insulin receptors and insulin signal transduction molecules [40,41]. Because these organisms regulate fat stores but do not express leptin signaling homologs, these observations suggest that, from an evolutionary perspective, the insulin signaling system might have a more primitive role in energy homeostasis than the leptin pathway. Null mutants of insulin signal transduction molecules such as DAF-2, an insulin receptor homolog, show increased fat storage in the intestine, increased longevity and reduced reproductive function (Figure 2) [40,41]. This outcome seems to indicate that insulin signaling cascades have a role in the nervous system of C. elegans, where insulin action has the catabolic effect of limiting energy storage, because the mutants have larger fat deposits. It is

4 Review TRENDS in Endocrinology and Metabolism Vol.15 No.8 October (a) TRENDS in Endocrinology & Metabolism Figure 2. Insulin-like signaling in the nervous system of Caenorhabditis elegans controls energy storage (images courtesy of G. Ruvkun, Harvard, Boston, MA). Nematodes such as C. elegans express numerous homologs of insulin, insulin receptors and insulin signal transduction molecules. Null mutations in these insulin-like signaling molecules, such as those in daf-2, which encodes an insulin receptor homolog, result in excess fat storage in the gut (b) as compared with wildtype control organisms (a) as detected here by Nile red staining. Notably, nematodes do not possess homologs of leptin or leptin receptors. also noteworthy that these primitive animals do not seem to regulate carbohydrate flux actively, suggesting that insulin had an important role in the control of energy balance, life cycle and reproduction before the evolution of a sophisticated insulin-regulated system of glucose homeostasis [40 42]. Likewise, the fruitfly, Drosophila melanogaster, expresses insulin signaling homologs that might have a similar role in the regulation of energy metabolism and lifespan [43]. Although nematodes and flies provide an intriguing evolutionary perspective, recently developed molecular genetics models in rodents have shed additional light on the role played by insulin in both reproduction and energy balance. Both brain-specific knockouts of the insulin receptor and animals lacking insulin receptor substrate-2 (IRS2), a key molecule linking activated insulin receptors to signal transduction via PI3K (Figure 3), show a phenotype of obesity and reproductive dysfunction [44,45]. Similarly, knockdown of insulin receptor expression locally in the mediobasal hypothalamus by RNA antisense technology [34] was found to result in a cumulative food intake of 152% and fat mass of 186% relative to control animals treated with nonsense RNA [34]. The consensus from various molecular genetics approaches is, therefore, that insulin action in the hypothalamus is essential in the regulation of energy homeostasis. A broader perspective reveals that neurochemical [33,38], electrophysiological [30], behavioral [7,36,37,46] and molecular genetics [34,44,45] approaches each support a role for insulin in the regulation of hypothalamic neurons and therein of energy homeostasis. Given the strength of this evidence and similarities in the action of insulin and leptin in the hypothalamus, an exploration of the intracellular signal transduction mechanisms used by these two hormones in this brain region is warranted. Insulin and leptin signal transduction Although originally associated with divergent signal transduction pathways [47,48], the cellular and physiological responses to insulin and leptin in the hypothalamus are similar (Figure 1a) [49]. In brief, ligand (b) binding to the heterotetrameric insulin receptor activates an intrinsic tyrosine kinase, leading to the autophosphorylation of several tyrosine residues on the intracellular portion of the receptor (Figure 3). Subsequently, a member of a family of proteins known as IRS molecules is activated by phosphorylation mediated by the insulin receptor. After this tyrosine phosphorylation step, IRS proteins activate the kinase PI3K, which consists of a regulatory subunit (typically called p85) and a catalytic domain (p110). By binding to the p85 subunit, phosphorylated IRS molecules activate the p110 subunit of PI3K, which in turn catalyzes the phosphorylation of phosphatidylinositol (4,5)-bisphosphate to phosphatidylinositol (3,4,5)-trisphosphate [PtdIns(3,4,5)P 2 ]. The latter functions as an active signal transduction factor, leading to the activation of several downstream molecules such as 3-phosphoinositide-dependent kinase-1, glycogen synthase kinase-3 and protein kinase B [47,50,51]. Unlike insulin, leptin is a cytokine-like molecule whose receptor is a member of the type 1 cytokine receptor family that includes receptors for ciliary neurotrophic factor, leukemia inhibitory factor and interleukin-6 [48]. Although several splice variants exist, the fully active signaling isoform of the leptin receptor is the long form or B isoform. Binding of leptin leads to the recruitment of Janus-activated kinase-2 (JAK2), a cytosolic tyrosine kinase (Figure 3). JAK2 functions as an extrinsic tyrosine kinase and phosphorylates tyrosine residues on the intracellular portion of the leptin receptor. Signal transducer and activator of transcription-3 (STAT3) molecules are transcription factors that are activated, in turn, by binding to the activated leptin receptor, whereupon they are phosphorylated by JAK2. Activated STAT3 then dimerizes and translocates to the nucleus to regulate gene transcription. Both the leptin and the insulin receptor also couple to additional signal transduction pathways (including extracellular signal-regulated kinase and mitogen-activated protein kinase pathways) [48,50]. Cross-talk between leptin and insulin in the hypothalamus Much of our knowledge about insulin and leptin signaling arises primarily from studies in peripheral tissues and cell culture. In several of these signaling models, leptin has the ability to activate both the JAK/STAT pathway and the classically insulin-like PI3K pathway [52 55], in a phenomenon referred to as cross-talk. Given similarities in the hypothalamic responses to insulin and leptin in energy homeostasis, an attractive hypothesis is that this cross-talk also exists in key brain areas such as the ARC [49]. Addressing this possibility, Spanswick, Ashford and their coworkers [29,30] carried out a series of electrophysiological studies in hypothalamic slice preparations, which showed that both leptin and insulin acutely regulate the membrane potential and firing rates of a specific subset of neurons. For both hormones, this effect was dependent on signaling through PI3K [29,30] a clear demonstration of cross-talk in the CNS. Other studies have addressed the role of hypothalamic PI3K signaling in the regulation of food intake. Our

5 366 Review TRENDS in Endocrinology and Metabolism Vol.15 No.8 October 2004 Insulin receptor Leptin receptor IRS PIP2 p110 IRS p110 p85 PIP3 S PKB T IRS JAK2 PKC PDK1 JAK2 JAK2 Stat3 Stat3 Stat3 Socs3 Figure 3. Insulin and leptin signal transduction and cross-talk at the level of phosphatidylinositol 3-kinase (PI3K). Binding of insulin to its receptor activates an intrinsic tyrosine kinase that phosphorylates key residues on the intracellular domain of the receptor. Insulin receptor substrate (IRS) proteins bind to these phosphorylated residues and become activated by tyrosine phosphorylation. In turn, activated IRS proteins interact with and activate PI3K, which consists predominantly of a regulatory (p85) and a catalytic subunit (p110). PI3K catalyzes the phosphorylation of phosphatidylinositol (4,5)-bisphosphate (PIP2) to phosphatidylinositol (3,4,5)-trisphosphate (PIP3), a cofactor and signal transduction molecule that activates downstream targets such as 3-phosphoinositide-dependent kinase-1 (PDK1), glycogen synthase kinase-3 (GSK3) and protein kinase B (PKB/Akt). Leptin is a cytokine-type molecule that signals through the long form of the leptin receptor, which is homologous to members of the type 1 cytokine receptor family. Ligand binding results in receptor dimerization and recruitment of an extrinsic kinase, Janus-activated kinase-2 (JAK2). This kinase phosphorylates the intracellular domain of the leptin receptor, providing a binding site for signal tranducer and activator of transcription-3 (STAT3) molecules, which are then phosphorylated and activated by JAK2. In turn, activated STAT3 molecules dimerize and translocate to the nucleus, where they participate in leptin target gene regulation. Notably, in various peripheral tissues leptin has been shown to possess the ability to activate PI3K, perhaps via JAK2. Several investigators have now shown that leptin is capable of similar cross-talk in the hypothalamus. Reproduced, with permission, from Ref. [49]. research group and others have demonstrated in vivo that insulin activates the IRS/PI3K signaling cascade in mediobasal hypothalamus and, furthermore, that PI3K signaling is required for the inhibition of food intake by insulin administered i.c.v. [37,56]. Studies using an immunohistochemical method to detect the PI3K-catalyzed reaction product PtdIns(3,4,5)P 3 have shown that PI3K seems to be activated in ARC neurons that coexpress IRS2 [57], as predicted [37]. We have also tested the hypothesis that at least some of the hypothalamic effects of leptin might be mediated by PI3K signaling. Like insulin, leptin activates PI3K in hypothalamic neurons and this activation is required for the ability of leptin administered i.c.v. to reduce food intake [58], an observation that has been confirmed by other groups [59]. These findings support a model in which activation of IRS/PI3K signaling in hypothalamic neurons is a crucial event in the ability of adiposity-related hormones to regulate energy homeostasis. Notably, as discussed above, insulin-related signaling in C. elegans and Drosophila also uses a pathway involving IRS and PI3K homologs [40,41]. Important unanswered questions Several important questions regarding hypothalamic insulin/pi3k signaling remain unanswered (Box 1), of which we mention only a few here. First, what are the relative contributions of the PI3K pathway and JAK/STAT signaling to insulin and leptin action and ultimately to integrated energy homeostasis? For example, leptin (ob/ob) or leptin receptor (db/db) mutant mice have a more severe obesity phenotype than do the insulin signaling pathway mutants described above, suggesting that insulin and leptin signaling have differential participation in the processes underlying energy homeostasis. An illustrative example is provided by a germline altered mutant leptin receptor mouse that is genetically defective in its ability to activate STAT signaling [60]. The observations that these mice are hyperphagic and obese (more obese than insulin signaling mutants, but not as obese as db/db control mice) suggest that JAK/STAT signaling has an important role in leptin-mediated control of body adiposity. Unlike db/db mice that lack all leptin receptor function, however, these mice do not express reproductive defects and have an attenuated diabetes phenotype [60]. Thus, STAT3 activation is necessary for some, but not all, leptin receptor functions. As mentioned above, brain-specific knockouts of the insulin receptor and animals lacking IRS2 show a phenotype of milder obesity with reproductive impairment and glucose intolerance [44,45]. These data support a model in which insulin/pi3k signaling and reproduction might be localized predominantly to the NP/AgRP neurons in the ARC, whereas the control of energy expenditure, food intake and fat mass might rely more heavily on signaling via the POMC neuronal subtype [61]. An alternative, but not mutually exclusive, hypothesis states that the insulin and leptin neuronal signaling pathways are intricately related and, for example, that defects in leptin signal transduction lead to insulin resistance, as is observed in the db/db mouse model [49]. In short, the available information has yet to clarify how both insulin and leptin input to a specific neuron is integrated in the control of either reproduction or energy homeostasis. Given the interrelatedness of these systems, a careful dissection of the individual contributions of physiological insulin and leptin action in the hypothalamus will require novel approaches.

6 Review TRENDS in Endocrinology and Metabolism Vol.15 No.8 October Box 1. Outstanding questions How is neuronal input from both insulin and leptin ultimately integrated to regulate energy homeostasis? Do insulin and leptin regulate the same or distinct populations of hypothalamic neurons via PI3K? Does leptin signal via both PI3K and JAK/STAT in the same neuron? Does insulin use signaling pathways in addition to the PI3K pathway in hypothalamic neurons? How does PI3K activation couple to neuronal firing in POMC neurons and to neuronal inhibition in NP/AgRP neurons? Does insulin signaling in the hypothalamic extra-arcuate nucleus and non-hypothalamic brain areas contribute to energy homeostasis? Because PI3K is a target of insulin resistance in peripheral tissues, does hypothalamic insulin resistance at the level of PI3K activation contribute to the pathophysiology of obesity? A second compelling, and related, fundamental issue is whether PI3K signaling is required for the response of NP/AgRP neurons, POMC neurons, or both sets of neurons to adiposity signals. Because POMC neurons are activated by insulin and leptin, whereas NP/AgRP neurons are inhibited, it seems counterintuitive to suggest that both responses involve an insulin-like PI3K-dependent mechanism. What key difference might allow this regulation to occur in opposite directions in adjacent cell populations? Finally, to what extent are the overlapping actions of leptin and insulin in the CNS attributable to intracellular cross-talk at the neuronal level? A principal challenge for this work is to develop key methods with which to examine cellular signaling events and neuronal responses in the context of the most complex of tissues containing numerous cell types and almost infinite cell cell interactions. Additional unanswered research questions are listed in Box 1. Hypothalamic insulin signaling and obesity As in most biomedical research, the ultimate goal underlying studies of hypothalamic insulin action is to generate insight into human disease and its treatment. Human obesity seems to be characterized by hypothalamic resistance to adiposity signals, such as insulin and leptin, in the sense that obese individuals have markedly increased serum insulin [10] and leptin [28] levels, reflecting an increase in body adipose mass. et, even in the face of this increase in adiposity signals, food intake remains normal or high, analogous to the hyperinsulinemia and normo- or hyperglycemia of type 2 diabetes. Does this phenomenon represent an acquired metabolic defect in hypothalamic signal transduction? Such a scheme might involve mechanisms analogous to those involved in peripheral insulin resistance in obesity and diabetes. In peripheral tissues, activation of PI3K in response to insulin is a central event in the activation and translocation of glucose transporter molecules to the cell membrane, which in turn facilitate glucose clearance from the bloodstream [47,51,62,63]. Intravenous fat infusion [64 66], high-fat feeding [67,68] and obesity [69] per se have each been shown to impair the ability of insulin to activate PI3K and thus glucose transport. Because obesity and type 2 diabetes occur in the same metabolic milieu, it is intriguing to propose that a mechanism similar to that causing type 2 diabetes causes insulin and leptin resistance in the hypothalamus, as has been shown for genetic models of obesity [56]. Thus, demonstration that the classically insulin-like signal transduction pathway involving PI3K is relevant to the CNS control of energy homeostasis might yield new therapeutic opportunities for the treatment of obesity. Conclusions A compelling body of work, both historical and more recent, using various model systems and experimental approaches strongly supports the idea that insulin, together with leptin, has an important role in the hypothalamic control of energy homeostasis. Further study of the role of IRS/PI3K signaling in the hypothalamus could reveal new targets for the treatment of obesity and will address outstanding fundamental research issues concerning the cellular organization of essential signaling systems in a complex tissue such as the brain. It is likely that evermore sophisticated experimental approaches will be necessary to understand how complex inputs from the periphery are integrated in the CNS to maintain homeostasis, but such information is needed to improve our understanding of the pathogenesis of common obesity. Acknowledgements This work supported by grants DK (to K.D.N.) and DK17047 (to the Diabetes and Endocrinology Research Center at the University of Washington), by the Department of Veterans Affairs Career Scientist and Merit Review Research Program (to D.G.B.), and by grants NS32273, DK12829, DK52989 (to M.W.S.). References 1 Kopelman, P.G. (2000) Obesity as a medical problem. Nature 404, Mokdad, A.H. et al. (2001) The continuing epidemics of obesity and diabetes in the United States. J. Am. Med. Assoc. 286, Spiegelman, B.M. and Flier, J.S. (2001) Obesity and the regulation of energy balance. Cell 104, Schwartz, M.W. et al. (2000) Central nervous system control of food intake. Nature 404, Zhang,. et al. (1994) Positional cloning of the mouse obese gene and its human homologue. Nature 372, Halaas, J.L. et al. (1995) Weight-reducing effects of the plasma protein encoded by the obese gene. Science 269, Woods, S.C. et al. (1979) Chronic intracerebroventricular infusion of insulin reduces food intake and body weight of baboons. Nature 282, Woods, S.C. and Seeley, R.J. (2001) Insulin as an adiposity signal. Int. J. Obes. Relat. Metab. Disord. 25(Suppl. 5), S35 S38 9 Woods, S.C. and Porte, D., Jr. (1978) The central nervous system, pancreatic hormones, feeding, and obesity. Adv. Metab. Disord. 9, Bagdade, J.D. et al. (1967) The significance of basal insulin levels in the evaluation of the insulin response to glucose in diabetic and nondiabetic subjects. J. Clin. Invest. 46, Corp, E.S. et al. (1986) Localization of 125 I-insulin binding sites in the rat hypothalamus by quantitative autoradiography. Neurosci. Lett. 70, Marks, J.L. et al. (1990) Localization of insulin receptor mrna in rat brain by in situ hybridization. Endocrinology 127, Schwartz, M.W. et al. (1996) Identification of targets of leptin action in rat hypothalamus. J. Clin. Invest. 98,

7 368 Review TRENDS in Endocrinology and Metabolism Vol.15 No.8 October Dryden, S. et al. (1998) Hyperphagia induced by hypoglycemia in rats is independent of leptin and hypothalamic neuropeptide (NP). Peptides 19, Carlson, M.G. and Campbell, P.J. (1993) Intensive insulin therapy and weight gain in IDDM. Diabetes 42, Sindelar, D.K. et al. (1999) Low plasma leptin levels contribute to diabetic hyperphagia in rats. Diabetes 48, Sipols, A.J. et al. (1995) Effect of intracerebroventricular insulin infusion on diabetic hyperphagia and hypothalamic neuropeptide gene expression. Diabetes 44, Woods, S.C. and Porte, D., Jr. (1977) Relationship between plasma and cerebrospinal fluid insulin levels of dogs. Am. J. Physiol. 233, E331 E Wallum, B.J. et al. (1987) Cerebrospinal fluid insulin levels increase during intravenous insulin infusions in man. J. Clin. Endocrinol. Metab. 64, Baura, G.D. et al. (1993) Saturable transport of insulin from plasma into the central nervous system of dogs in vivo. A mechanism for regulated insulin delivery to the brain. J. Clin. Invest. 92, Pardridge, W.M. (1986) Receptor-mediated peptide transport through the blood brain barrier. Endocr. Rev. 7, Hachiya,H.L. et al. (1988) Intracellular pathways of insulin transport across vascular endothelial cells. Am. J. Physiol. 255, C459 C Woods, S.C. et al. (2003) Insulin and the blood brain barrier. Curr. Pharm. Des. 9, Kaiyala, K.J. et al. (2000) Obesity induced by a high-fat diet is associated with reduced brain insulin transport in dogs. Diabetes 49, Krisch, B. and Leonhardt, H. (1978) The functional and structural border of the neurohemal region of the median eminence. Cell Tissue Res. 192, Merchenthaler, I. (1991) Neurons with access to the general circulation in the central nervous system of the rat: a retrograde tracing study with fluoro-gold. Neuroscience 44, Shaver, S.W. et al. (1992) Morphology and function of capillary networks in subregions of the rat tuber cinereum. Cell Tissue Res. 267, Considine, R.V. et al. (1996) Serum immunoreactive-leptin concentrations in normal-weight and obese humans. N. Engl. J. Med. 334, Spanswick, D. et al. (1997) Leptin inhibits hypothalamic neurons by activation of ATP-sensitive potassium channels. Nature 390, Spanswick, D. et al. (2000) Insulin activates ATP-sensitive K C channels in hypothalamic neurons of lean, but not obese rats. Nat. Neurosci. 3, Cone, R.D. et al. (2001) The arcuate nucleus as a conduit for diverse signals relevant to energy homeostasis. Int. J. Obes. Relat. Metab. Disord. 25(Suppl. 5), S63 S67 32 Berthoud, H.R. (2002) Multiple neural systems controlling food intake and body weight. Neurosci. Biobehav. Rev. 26, Schwartz, M.W. et al. (1991) Central insulin administration reduces neuropeptide mrna expression in the arcuate nucleus of fooddeprived lean (Fa/Fa) but not obese (fa/fa) Zucker rats. Endocrinology 128, Obici, S. et al. (2002) Decreasing hypothalamic insulin receptors causes hyperphagia and insulin resistance in rats. Nat. Neurosci. 5, Foster, L.A. et al. (1991) Food intake and serum insulin responses to intraventricular infusions of insulin and IGF-I. Physiol. Behav. 50, Air, E.L. et al. (2002) Acute third ventricular administration of insulin decreases food intake in two paradigms. Pharmacol. Biochem. Behav. 72, Niswender, K.D. et al. (2003) Insulin activation of phosphatidylinositol 3-kinase in the hypothalamic arcuate nucleus: a key mediator of insulin-induced anorexia. Diabetes 52, Benoit, S.C. et al. (2002) The catabolic action of insulin in the brain is mediated by melanocortins. J. Neurosci. 22, Figlewicz, D.P. (2003) Adiposity signals and food reward: expanding the CNS roles of insulin and leptin. Am. J. Physiol. Regul. Integr. Comp. Physiol. 284, R882 R Wolkow, C.A. et al. (2002) Insulin receptor substrate and p55 orthologous adaptor proteins function in the Caenorhabditis elegans daf-2/insulin-like signaling pathway. J. Biol. Chem. 277, Wolkow, C.A. et al. (2000) Regulation of C. elegans life-span by insulinlike signaling in the nervous system. Science 290, Lee, S.S. et al. (2003) DAF-16 target genes that control C. elegans lifespan and metabolism. Science 300, Garofalo, R.S. (2002) Genetic analysis of insulin signaling in Drosophila. Trends Endocrinol. Metab. 13, Burks, D.J. et al. (2000) IRS-2 pathways integrate female reproduction and energy homeostasis. Nature 407, Bruning, J.C. et al. (2000) Role of brain insulin receptor in control of body weight and reproduction. Science 289, Clegg, D.J. et al. (2003) Differential sensitivity to central leptin and insulin in male and female rats. Diabetes 52, Kido,. et al. (2001) The insulin receptor and its cellular targets. J. Clin. Endocrinol. Metab. 86, Banks, A.S. et al. (2000) Activation of downstream signals by the long form of the leptin receptor. J. Biol. Chem. 275, Niswender, K.D. and Schwartz, M.W. (2003) Insulin and leptin revisited: adiposity signals with overlapping physiological and intracellular signaling capabilities. Front. Neuroendocrinol. 24, Saltiel, A.R. and Pessin, J.E. (2002) Insulin signaling pathways in time and space. Trends Cell Biol. 12, Saltiel, A.R. and Kahn, C.R. (2001) Insulin signalling and the regulation of glucose and lipid metabolism. Nature 414, Anderwald, C. et al. (2002) Short-term leptin-dependent inhibition of hepatic gluconeogenesis is mediated by insulin receptor substrate-2. Mol. Endocrinol. 16, Harvey, J. et al. (2000) Essential role of phosphoinositide 3-kinase in leptin-induced K ATP channel activation in the rat CRI-G1 insulinoma cell line. J. Biol. Chem. 275, Kellerer, M. et al. (1997) Leptin activates PI-3 kinase in C2C12 myotubes via Janus kinase-2 (JAK-2) and insulin receptor substrate-2 (IRS-2) dependent pathways. Diabetologia 40, Zhao, A.Z. et al. (2000) Leptin induces insulin-like signaling that antagonizes camp elevation by glucagon in hepatocytes. J. Biol. Chem. 275, Carvalheira, J.B. et al. (2003) Selective impairment of insulin signalling in the hypothalamus of obese Zucker rats. Diabetologia 46, Niswender, K.D. et al. (2003) Immunocytochemical detection of phosphatidylinositol 3-kinase activation by insulin and leptin. J. Histochem. Cytochem. 51, Niswender, K.D. et al. (2001) Key enzyme in leptin-induced anorexia. Nature 413, Zhao, A.Z. et al. (2002) A phosphatidylinositol 3-kinase phosphodiesterase 3B cyclic AMP pathway in hypothalamic action of leptin on feeding. Nat. Neurosci. 5, Bates, S.H. et al. (2003) STAT3 signalling is required for leptin regulation of energy balance but not reproduction. Nature 421, Myers, M.G., Jr. (2004) Leptin receptor signaling and the regulation of mammalian physiology. Recent Prog. Horm. Res. 59, Kanai, F. et al. (1993) Insulin-stimulated GLUT4 translocation is relevant to the phosphorylation of IRS-1 and the activity of PI3- kinase. Biochem. Biophys. Res. Commun. 195, Cushman, S.W. et al. (1998) Molecular mechanisms involved in GLUT4 translocation in muscle during insulin and contraction stimulation. Adv. Exp. Med. Biol. 441, Dresner, A. et al. (1999) Effects of free fatty acids on glucose transport and IRS-1-associated phosphatidylinositol 3-kinase activity. J. Clin. Invest. 103, u, C. et al. (2002) Mechanism by which fatty acids inhibit insulin activation of IRS-1 associated phosphatidylinositol 3-kinase activity in muscle. J. Biol. Chem. 277, Griffin, M.E. et al. (1999) Free fatty acid-induced insulin resistance is associated with activation of protein kinase Cq and alterations in the insulin signaling cascade. Diabetes 48, Wang, J. et al. (2001) Overfeeding rapidly induces leptin and insulin resistance. Diabetes 50, Buettner, R. et al. (2000) Correction of diet-induced

8 Review TRENDS in Endocrinology and Metabolism Vol.15 No.8 October hyperglycemia, hyperinsulinemia, and skeletal muscle insulin resistance by moderate hyperleptinemia. Am. J. Physiol. Endocrinol. Metab. 278, E563 E Houmard, J.A. et al. (2002) Effect of weight loss on insulin sensitivity and intramuscular long-chain fatty acyl-coas in morbidly obese subjects. Diabetes 51, Barsh, G.S. and Schwartz, M.W. (2002) Genetic approaches to studying energy balance: perception and integration. Nat. Rev. Genet. 3, Endeavour the quarterly magazine for the history and philosophy of science ou can access Endeavour online via ScienceDirect, where you ll find a collection of beautifully illustrated articles on the history of science, book reviews and editorial comment. Featuring Sverre Petterssen and the Contentious (and Momentous) Weather Forecasts for D-Day, 6 June 1944 by J.R. Fleming Food of Paradise: Tahitian breadfruit and the Autocritique of European Consumption by P. White and E.C. Spary Two Approaches to Etiology: The Debate Over Smoking and Lung Cancer in the 1950s by M. Parascandola Sicily, or sea of tranquility? Mapping and naming the moon by J. Vertesi The Prehistory of the Periodic Table by D. Rouvray Two portraits of Edmond Halley by P. Fara and coming soon Fighting the microbe of sporting mania : Australian science and Antarctic exploration in the early twentieth century by P. Roberts Learning from Education to Communicate Science as a Good Story by A. Negrete and C. Lartigue The Traffic and Display of Body Parts in the Early-19th Century by S. Alberti and S. Chaplin The Rise, Fall and Resurrection of Group Selection by M. Borrello Pomet s great Compleat History of Drugs by S. Sherman Sherlock Holmes: scientific detective by L. Snyder The Future of Electricity in 1892 by G.J.N. Gooday The First Personal Computer by J. November Baloonmania: news in the air by M.G. Kim and much, much more... Locate Endeavour on ScienceDirect (