Editorial: Toward a Systems Biology of Insulin Secretion and Type 2 Diabetes

Although it is clear that β-cell pathology, in particular disordered insulin secretion, is a key underlying pathogenetic feature of most forms of diabetes, complex interactions of the islets with other organs, such as brain, liver, gut, and several peripheral tissues, are also essential for the normal integration of metabolism. It is this area that has been explored in the Seventh Annual Servier-IGIS Meeting, which was held last Spring in St. Jean Cap Ferrat in Southern France. The focus of these meetings, since their inception in 2000, has been the β-cell and the mechanisms underlying its development and function as the source of insulin, the most essential regulator of the blood glucose level.

The familiar pathways of glucose, lipid, and amino acid metabolism in humans and other mammals are, of course, fundamental to almost all organisms, except perhaps for the most highly specialized bacteria, so it is not surprising that insulin-like molecules and the insulin/IGF receptor signaling system are well conserved features of all metazoans that have been studied. With the rise of multicellular organisms in evolution came the need to regulate and coordinate metabolism and growth in order to maintain both the constancy of the internal environment (homeostasis) and also to respond to the external environment. One of the most prominent environmental stimuli had to be the availability of nutrients and fuels for survival and growth. The insulin-like hormones, insulin and IGF, appear to have evolved, along with a panoply of other regulatory substances, to fine-tune the efficient uptake, storage, and utilization of nutrients for either energy production or growth.

Appropriately, the symposium opened with an introductory lecture by Leopold, reviewing our current knowledge of the control of metabolism and growth in the fruit fly, Drosophila melanogaster, by insulin-like peptides and an insulin signaling pathway that is remarkably similar in many of its components to that of man. No less than seven Drosophila insulin-like peptides have been found. Three of these are expressed in specialized neurones in the insect brain, while others are expressed in larval tissues during development. Insulin-like peptides released from the brain cells downregulate hemolymph sugar levels, while a glucagon-like peptide, the adipokinetic hormone, from other neuroendocrine structures opposes it by raising sugar levels. Interestingly, these adipokinetic hormone-producing cells, unlike the insulin-producing cells, express an ATP-sensitive K⁺ channel (K_ATP channel) that responds to tolbutamide to increase their secretory activity, resulting in hyperglycemia. Moreover, mutations that influence components of the insulin receptor pathway result in reduced growth, reflected in smaller cells in smaller adults. Impaired insulin signaling also results in increased longevity, decreased reproduction, and increased stress resistance. Complex interactions with other endocrine (ecdysone, juvenile hormone) and metabolic (TOR, etc.) pathways integrate nutrition and metabolism and eventual organismal size. The powerful genetic tools available for Drosophila make it a rich experimental source for identification of new components of these important conserved pathways.

Much recent work has focused on the role of lipids as modulators of insulin action and as important factors in the pathogenesis of type 2 diabetes via induction of insulin resistance. Shulman and coworkers have used magnetic resonance spectroscopy to explore the molecular mechanisms underlying defective glucose transport and glycogen metabolism in muscle. Increased lipid metabolites such as fatty acyl-CoAs and diacylglycerol activate kinase cascades that impair insulin signaling due to Ser/Thr phosphorylation of IRS-1. Similar mechanisms may operate to impair hepatic insulin signaling due to increased hepatic lipids in insulin-resistant subjects. These changes then lead to relative increases in gluconeogenesis and reduced hepatic glucose uptake via lowered AKT2 and increased FOXO transcriptional effects on several key gluconeogenic rate-controlling enzymes. In related magnetic resonance spectroscopy studies, significant decreases in mitochondrial oxidative phosphorylation activity in muscles and liver have been found in elderly, lean, insulin-resistant volunteers, in association with increased muscle and hepatic lipid content. Similar changes were found in young, lean offspring of parents with type 2 diabetes associated with reduced mitochondrial density, consistent with other reported studies. The mechanisms underlying the reduction in mitochondrial biogenesis in these individuals is an important area for further study that may lead to new targets for therapeutic intervention.

While lipid overload can lead to insulin resistance and impaired β-cell function, fatty acids and other lipids are also important for normal β-cell function. It is well established that fatty acids can augment glucose-stimulated insulin secretion, an effect that may be especially important in meeting the demands for increased insulin in compensated insulin resistance. Lipids can act both through their metabolism as well as via free fatty acid (FFA) receptor (FFAR) activation. Prentki and associates have studied these mechanisms in detail and find that increased cytosolic malonyl-CoA arising from glucose and lipid metabolism acts via AMPK/malonyl-CoA pathways to limit fatty acid oxidation, thus increasing long-chain acyl-CoA signaling molecules. Glucose can then enhance esterification and subsequent lipolysis of long-chain acyl-CoA to renew the FFA pool, which can then interact with FFAR/GPR40, enhancing cytosolic Ca²⁺ and insulin secretion. Glucose may also enhance release of arachidonic acid from phospholipids to activate yet other lipid-signaling pathways in the β-cell.

Since many of the foregoing effects of lipids on insulin secretion depend on glucose-stimulated lipolytic activity, efforts are currently underway to identify β-cell lipases. Indeed, orlistat, a lipase inhibitor, abolishes lipolysis of tri- and diglycerides in islets, inhibiting insulin secretion without perturbing glucose metabolism. Mulder and colleagues have examined hormone-sensitive lipase–null mice but find no evidence of a β-cell secretory effect. Thus, further studies to identify the role of other lipases involved in β-cell stimulus-secretion coupling are needed.

In addition to substrates such as FFA, various adipokines such as tumor necrosis factor-α and resistin are associated with obesity and insulin resistance, whereas others such as leptin and adiponectin sensitize the body to insulin. Adiponectin has been shown to be upregulated by thiazolidinediones acting through peroxisome proliferator–activated receptor-γ. As reviewed by Kadowaki and associates, adiponectin circulates in several multimeric forms, of which the high–molecular weight forms are most active in ameliorating insulin resistance through negative effects on hepatic gluconeogenesis and lowering of FFA through stimulation of skeletal muscle FFA oxidation. He and his associates have carefully dissected the effects of pioglitazone dosage on ob/ob and adipo^−/− ob/ob mice to demonstrate both adiponectin-dependent and -independent pathways of thiazolidinedione action on such parameters as adipocyte size and adiponectin production, as well as target organ effects on hepatic AMPK activation and decreased gluconeogenesis, leading to improved glucose tolerance and diabetes control.

In a session on muscle and liver, Newsholme and colleagues discussed the effects of amino acids on key β-cell processes leading to enhanced insulin secretion, with special emphasis on generation of ATP and the mechanisms coupling amino acid metabolic pathways with the putative generation of messengers of mitochondrial origin.

Turning to factors arising from muscle, interleukin (IL)-6, a cytokine having both pro- and anti-inflammatory actions, is released in large amounts during exercise. Increases in IL-6 production and secretion are associated with increases in AMPK activity in tissues such as muscle and adipose tissue. AMPK enhances ATP generation while inhibiting nonessential energy consuming processes via phosphorylation of selected metabolic enzymes. Ruderman and colleagues have demonstrated decreased AMPK activity in muscle and adipose tissue in young IL-6–null mice and a diminished enzyme response to exercise in these tissues. These animals later develop manifestations of the metabolic syndrome with obesity, dyslipidemia, and impaired glucose tolerance. Key questions are whether these effects of IL-6 contribute to the reported benefits of exercise in reducing the prevalence of type 2 diabetes, coronary atherosclerosis, and other concomitants of the metabolic syndrome in humans.

In liver, as in the β-cell, glucokinase plays a key role as a glucose sensor. However, the complex conformational states and regulatory networks that control glucokinase function differ significantly in these two tissues. As discussed by Baltrusch and Tiedge, these range from the use of alternate promoters to regulate enzyme expression and the shuttling of a high-affinity regulatory protein between cytosol and nucleus in the liver to altered compartmentalization of glucokinase in β-cells and its activation by binding of the bifunctional enzyme 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase (PFK-2/FBPase-2) to increase its V_max. Some of these effects, and those of chemical glucokinase activators, are related to various conformational states with altered catalytic activity, as revealed in recent crystallographic studies.

This session ended with consideration of the time-honored conundrum as to whether the inhibitory effect of insulin on hepatic glucose production (HGP) is direct or indirect. Girard carefully reviews the various known indirect influences on HGP, which include suppression of glucagon levels, plasma nonesterified fatty acid or gluconeogenic precursors from peripheral tissues, along with more recent studies on effects of various adipokines, as well as novel central mechanisms. Infusion of insulin into the third ventricle has been shown to inhibit HGP, an effect reversed by inhibition of insulin receptor signaling. Surprisingly, central infusion of activators of K_ATP channel lowered blood glucose levels by inhibiting HGP, while K_ATP inhibitors reduced the effects of systemic insulin. These effects are mediated via the hepatic branch of the vagus nerve. However, recent studies with mice lacking expression of insulin receptor only in liver, as well as clamp studies on dogs, support direct effects of insulin as being of more paramount importance in suppressing HGP. Girard proposes that the relative importance of glycogenolysis versus gluconeogenesis in various experimental protocols may account in part for various reported species differences in sensitivity to direct versus indirect effects. Moreover, as gluconeogenesis is less sensitive to suppression by insulin than glycogenolysis in type 2 diabetes, therapeutic agents that suppress glucagon hypersecretion, such as glucagon-like peptide (GLP)-1 and/or others, should likely be clinically beneficial for lowering HGP.

The gastrointestinal tract has recently proven to be a rich source of promising endocrine substances with antidiabetic effects. GLP-1, a product of the glucagon gene, is produced in the intestinal L-cells through the action on proglucagon of prohormone convertase (PC)1/3, in contrast to the islet α-cells, where PC2 acts to process mainly glucagon from the same precursor. GLP-1 has emerged as an important incretin hormone, augmenting the effects of oral nutrient stimuli on insulin release. The factors that regulate GLP-1 secretion are clearly of considerable interest, in view of its recently demonstrated efficacy in treating type 2 diabetes. Its actions include potentiation of insulin secretion in response to glucose, enhancement of β-cell growth and survival, and inhibition of glucagon secretion, gastric emptying, and food intake. Both nutrients and various non-nutrient peptides and neuromodulators have been implicated in GLP-1 release, as discussed in reviews by Brubaker and Reimann and coworkers in this session. Clearly, elucidation of these pathways could lead to the development of therapeutic GLP-1 secretogogues.

Gastric inhibitory polypeptide or glucose-dependent insulinotropic peptide (GIP), another member of the glucagon family produced in the intestinal K-cells, exerts incretin effects, both directly on the β-cell and via augmentation of GLP-1 secretion and/or action. Seino and collaborators have studied effects of GIP receptor knockout in mice and found reduced incretin effects on glucose-induced insulin secretion, a defect that is additive with that induced by GLP-1 receptor knockout. These investigators have also documented extra-pancreatic effects of GIP on the accumulation of fat in adipose tissue and of calcium into bone, indicating a broader role for this gut-derived peptide in regulating nutrient uptake. These effects suggest that GIP is the product of a “thrifty gene” and thus may contribute to the incidence of obesity and diabetes. Crossing GIP receptor–null mice with leptin-deficient ob/ob mice resulted in significant amelioration of obesity and dyslipidemia, accompanied by increased insulin sensitivity and glucose tolerance. The authors conclude that effects of both GIP receptor agonists and/or antagonists may provide beneficial effects in certain forms of diabetes.

Another approach to therapy of obesity and/or type 2 diabetes is gastric bypass surgery (GBS) and related surgical procedures. Naslund and Kral review the effects of GBS with special emphasis on its effects on gastrointestinal peptide levels, including ghrelin and three incretin hormones—GLP-1, GLP-2, and peptide YY. This approach to therapy in a majority of cases is curative and results generally in reduced ghrelin levels and enhanced incretin effects, which likely contribute to the favorably altered physiologic state induced by GBS. Alterations in adipokines and neuroregulatory circuits may also contribute to its positive effects.

The islets are the focus of a larger number of peripheral and central inputs in addition to those discussed above. The islets are richly innervated by both parasympathetic and sympathetic nerve fibers, which act through the classical neurotransmitters acetylcholine and norepinephrine, respectively, which exert either stimulatory or inhibitory effect on the β-cells. A variety of neuroregulatory peptides also modulate both insulin and glucagon secretion. The latter are the major focus of a comprehensive review by Ahren and coworkers. The possibility of altered islet innervation in various models of insulin resistance and type 2 diabetes has been investigated by these authors in several animal models of diabetes (GK rats and db/db mice) with findings indicative of increased islet innervation. They propose that “augmented expression of neurotransmitters in the islets is a sign of islet adaptation for normalization of glucose tolerance” and conclude that further exploration of this area may yield new insights into neural mechanisms that contribute to the regulation of both islet cell mass and function in normal and pathologic states.

Another aspect of extrinsic inputs into glucose sensing in the regulation of insulin secretion has been explored recently by Thorens and associates who have developed ingenious methods to identify and study extrapancreatic glucose sensors in both the hepatoportal vein and the central nervous system (CNS). The hepatoportal sensor is activated by glucose gradients between the portal vein and peripheral veins and transmits signals via afferent branches of the hepatic vagal nerve to the CNS. It affects both first-phase insulin secretion and peripheral insulin sensitivity via a variety of pathways. Both extrapancreatic sensors are dependent on GLUT2 expression for their normalizing effects, including those on feeding behavior and glucagon secretion. Identification of their precise localization may provide new insights into the physiology of energy intake and metabolism.

In addition to peripheral glucose sensors, the brain has emerged as a central switchboard, integrating signals from many regions of the body (liver, gut, adipose tissue, and islets) conveyed by afferent nerves and neurotransmitters, as well as by circulating hormones and the major nutrients (glucose, amino acids, and free fatty acids) to regulate energy homeostasis, food intake, reproduction, and even learning and memory. The reports in this section are focused on various aspects of this expanding area of great current interest.

Needless to say, insulin and leptin are key players in this informational game, and they act primarily via receptors localized in hypothalamic centers such as the arcuate nucleus that control feeding behavior, satiety, and energy expenditure. Elevated insulin and leptin signal the availability of excess nutrients and thus tend to reduce food intake and enhance energy expenditure, as discussed by Wood and his coworkers. They point out that insulin with its rapid fluctuations in secretion and short half-life is a monitor of blood glucose and ongoing metabolism, as well as of body adiposity, while leptin’s longer half-life and secretion from adipocytes conveys information on both fat stores and adipocyte metabolic activity. Centers in the arcuate nucleus regulate satiety and respond to these anorexigenic signals by downregulating secretion of orexigenic factors such as neuropeptide Y and agouti-related peptide and upregulating the release of other factors promoting anorexia, such as α-MSH (melanocortin) and CART (cocaine- and amphetamine-related transcript). Other peripheral orexigenic hormones, especially ghrelin, are released from the stomach and upper gut before mealtimes and tend to oppose waning insulin/leptin actions in the hypothalamus, stimulating new cycles of feeding. Woods et al. also discuss possible consequences of insulin signaling in the hippocampus in relation to cognitive function, pointing out the possibility that enhanced learning may facilitate foraging for food. Olfactory insulin signaling may likewise enhance associations between certain foods and specific odors. Clearly, the development of obesity and insulin resistance may also occur centrally, resulting in impairment of normal regulatory processes, as well as cognitive functions.

Levin and coworkers go on to consider in greater detail the properties of glucose-sensing neurons, both inhibitory and excitatory. In early classical studies of the hypothalamus using lesions or electrical stimulation, certain areas were associated with food intake and energy expenditure. Lesions of the ventromedial hypothalamus led to hyperphagia and obesity, while lesions in the lateral hypothalamic area reduced food intake and increased autonomic activity, leading to lower body weight. Glucose-sensing neurons are widely distributed in forebrain and brainstem nuclei, where they integrate both “hard-wired” inputs from the periphery with hormonal, neuropeptide, and substrate signals. These populations appear to express glucose-sensing systems, such as glucokinase and K_ATP channels, similar to those of normal β- and α-cells. They respond to other stimuli such as lactate from glial cells, fatty acids, and ketone bodies, as well as to both insulin and leptin. Their output, in combination with other glucose-sensing neurones, is via a variety of efferent neural pathways that impact all aspects of energy homeostasis—intake, storage, and expenditure.

An interesting related issue, reviewed by Pénicaud and colleagues, is the discovery of the expression of the translocatable glucose transporter GLUT4 and more recently the related transporter, GLUT8, in the brain. These appear to function by translocation from intracellular pools to the plasma membrane, and their expression has been reported to be influenced by glucose and insulin, but whether they respond to insulin in a classical manner is still controversial. GLUT8 has also been noted to translocate to the endoplasmic reticulum, suggesting a possible role in glycoprotein biosynthesis and/or degradation. Gene disruption strategies are needed now to shed new light on the potential physiological roles of these transporters in regulating the brain’s metabolism and/or signaling functions.

In addition to glucose, FFAs are well known to influence carbohydrate metabolism and energy homeostasis via central mechanisms. This area is nicely reviewed by Magnan and coworkers. These authors have previously shown that infusion of lipids such as oleic acid centrally in rats leads to inhibition of food intake and glucose production, while infusion of triglycerides for 24 h leads to hepatic insulin resistance associated with increased glucose-stimulated insulin secretion and accompanied by decreased splanchnic sympathetic nerve activity, the latter effects being dependent on β-oxidation of the substrate. Transient increases in plasma insulin could be induced by a single intracarotid injection of oleic acid without changes in plasma glucose, indicating a direct effect of FFAs on neural control of insulin secretion. These authors have also provided evidence for the existence of hypothalamic subpopulations of neurones that are either excited or inhibited by FFAs in vitro, consistent with their role in central fuel sensing. Thus, central dysregulation of fatty acid signaling could be a factor leading to impaired glucoregulation of insulin secretion.

Ahima and coworkers review the central effects of adipocytokines on metabolism and energy homeostasis, focusing mainly on leptin and adiponectin. A great deal is now known about their mechanisms of production, their various circulating plasma forms, and their putative receptors and signaling pathways. Leptin normally acts to prevent the negative changes associated with starvation and weight reduction, including reduced energy expenditure, insulin resistance, hyperlipidemia, and fertility. Central leptin administration in ob/ob mice suppresses hepatic glucose production followed by food intake and weight loss, while restoring insulin sensitivity. Both leptin and adiponectin act peripherally to increase AMP-activated protein kinase and other enzymes involved in lipid metabolism in liver and muscle, as well as centrally. The mode of entry of adiponectin into the CNS remains unsettled, but adiponectin receptors are present in cerebral microvessels. These issues and a variety of other adipocytokines and putative hormones produced by fat cells remain to be studied in order to clarify their roles in human metabolic pathophysiology.

The conference concluded with an excellent personal overview by Porte of the history and development of our current concepts regarding the central regulation of energy homeostasis and the key role of insulin in this process as they have developed during the past 30 years of his research. He and his colleagues, as well as many others drawn into this field, have made great strides in elucidating the mechanisms by which insulin and other more recently discovered hormones/cytokines enter and act in the brain to regulate and integrate the various organ systems of the body that must all work harmoniously together to maintain an optimal normal state. Porte concludes that “the overwhelming evidence that insulin plays a key dual role in the regulation of carbohydrate metabolism and body weight suggests that further analysis of its CNS effects will continue to be a fruitful area for study and potentially therapeutic intervention.”

We are indebted to the Secretary of the IGIS group, Dr. Alain Ktorza, and to Laurence Alliot’s team at Servier for their outstanding assistance with the organization of the Symposium, as well as to Catriona Donagh for the editorial management of this supplement.