Contents
Contributors
Search

Discovery of the insulin receptor began with evidence that insulin was bound to a unique protein on the surface of fat and liver cells, and culminated with cloning of the gene. This section highlights many of the key findings contributing to that discovery, and summarizes some of the advances made since cloning of the insulin receptor. For the purposes of this chapter, investigations of the insulin receptor are divided among three categories (Fig. 1). This chapter will conclude by describing how some observations which were important for discovery of the IR have since been modified, and suggest new discoveries will be aided by the study of model organisms.

Figure 1. Chronology of Major Advances in the Understanding of Insulin Receptor Function.

A Specific Receptor for Insulin at the Cell Surface

A receptor for insulin was deduced from the saturable binding of insulin to fat and liver cells (1, 2). In the mid-1960's, the idea of a receptor was itself novel, but the insulin receptor (IR) was understood to identify a cellular component that recognized insulin and not other hormones. Thus, an insulin receptor would be present in limited numbers on the cell's surface and would bind with an affinity commensurate with circulating insulin concentrations. A second key feature was exclusivity, which became a more important criterion following the discovery of the two insulin-like growth factors, IGF-I and IGF-II (3). A third key property of the insulin receptor is its presence on cells that have known physiological responses to insulin (especially metabolic responses such as the acceleration of glucose uptake, and increased lipid and glycogen storage). Thus, the fourth key feature is that the IR be required to mediate cellular responses to insulin, which has been demonstrated ultimately through molecular genetics. These studies have also unmasked important developmental functions of the IR.

Insulin Receptor Purification and Subunit Composition

An early and still ongoing goal in studying the IR has been to determine its structure, specifically with the hope of finding clues to its mechanism of action and thus to possible causes of insulin resistance and diabetes. Substantial efforts were therefore made to purify the IR (4-6). The basic criterion to identify the IR was its ability to bind specifically 125I-insulin (7). The IR could be affinity labeled by photochemically or chemically crosslinking radiolabeled insulin to a cell surface protein from insulin sensitive tissues such as fat and liver (8, 9). In addition, purification was aided greatly by antibodies present in some patients with acute insulin resistance (10). Together with metabolic labeling studies, these studies demonstrates that the native insulin receptor was composed of two types of subunits in a tetramer (11, 12): two alpha-subunits of 135 kDa and two beta-subunits of 95 kDa which are linked together by disulfide bonds. The alpha-subunits are entirely extracellular, and the beta-subunits cross the plasma membrane. Insulin binding occurs entirely through contacts made to the alpha-subunits. The intracellular portion of the beta-subunit is also essential for insulin action.

Hormone Binding to Insulin and IGF Receptors

Insulin is closely related to insulin-like growth factors-1 and -2, as indicated by the names. The receptors for insulin (IR) and IGF1 (IGF1R) are structurally similar, whereas the receptor for IGF2 (IGF2R) is significantly different. Each receptor binds its cognate ligand with the highest affinity, as summarized in Table 1. Although the IR and IGF-1R have two hormone binding alpha-subunits per receptor - and biochemically it is possible to bind two hormone molecules per receptor at superphysiological hormone concentrations, giving the appearance of negative cooperativity for insulin binding (13) - under physiological conditions each receptor binds one hormone molecule (14).

The biological situation is somewhat more complicated than might be anticipated for three different receptors and cognate ligands. There are two forms of the insulin receptor present in mammals, revealed during cloning of the IR gene and cDNA (described below). The IR-A and IR-B isoforms that respectively lack or include 12 amino acids near the carboxyl terminus of the alpha-subunit (15), have nearly equivalent affinities for insulin, differing by no more than two-fold (16). Neither isoform shows significant binding to IGF1. However, compared to the IR-B isoform, the IR-A isoform has a ten-fold higher affinity for IGF2, although the binding is still weaker than for insulin (Table 1; data from (17, 18). The apparent affinities are typically determined by the ability of each hormone to displace radiolabeled cognate hormone - a competition binding assay - from soluble receptors or from receptors expressed by cultured cells. The cognate hormone-receptor interactions have dissociation constants in the 0.3 to 1 nM range.

Besides the two IR isoforms, the close primary structural similarity between the IR and IGF1R can - and does - produce IGF1R/IR hybrid receptors, with one alpha-beta subunit pair from the IR and the other alpha-beta pair from the IGF1R. As shown in Table 1, the binding properties of these IGF1R/IR hybrid receptors are dominated by the IGF1R half, since they tend to bind IGF1 with the highest affinity and show slightly weaker IGF2 binding; about 0.3 nM IGF1, which is comparable to native IGF1R. However, the hybrid receptor interaction with insulin is strongly dependent on an IR-A half since the IGF1R/IR-B hybrid shows little insulin binding, and the hybrid binds insulin with greater affinity than the IGF1R itself (18).

The Clinical Significance of IR Isoforms and Hybrids

Clinically, the relative abundance of either IR isoform may be more important for growth effects than metabolic effects, where growth includes normal development or cancer. Thus, although there is a two-fold difference in affinity for insulin shown by the IR-A versus IR-B isoform, it may be more significant that the IR-A isoform has a ten-fold greater affinity for IGF2, compared to the IR-B isoform. This is suggested by two observations: First, the IR-A isoform is more abundant in fetal tissue, and there is clear-cut genetic evidence that fetal IR interacts with IGF2 during fetal development (see below). Second, it is apparent that IR-A expression is elevated in many cancers, including breast, colon, lung, ovarian, and thyroid cancers (19-21). This may be accompanied by increased expression of IGF-II in these cancer cells, and/or loss-of-function mutations in the IGF2R. The increase in IR-A and IGF-2 expression, together with the greater affinity of IR-A for IGF-2 has led some investigators to suggest that an autocrine or paracrine loop functions in cell proliferation in these diseased states (21). Furthermore, the normal developmental function of IGF2R is IGF2 clearance (22). In tumorous growth, loss of IGF2R wild-type function may lead to loss of its potential tumor-suppressor activity (23, 24). It is a subject of some controversy as to whether expression of the IR-A versus IR-B isoforms is significantly different in lean versus obese or normal versus diabetic subjects. Never-the-less, the relative expression levels of the two isoforms differ among adult tissues, although different laboratories report diverse and sometimes contradictory findings on the relative abundances of the two isoforms in various tissues (15, 25, 26). The physiological importance of these differences is not yet known, but recent evidence suggests that each IR isoform activates a different PI3'-kinase isoform, ultimately leading to differential gene expression (27).

The clinical significance of the IGF1R/IR hybrids is another open question. The hybrids may constitute an increased fraction of the total IR pool in skeletal muscle of obese and diabetic patients, although the pool size will itself decrease (28). It has also been reported that IGF1R/IR hybrid receptors are increased in placenta during gestational diabetes (29). However, there is some controversy whether this increased fraction of hybrids serves a primary role in insulin resistance (30). It is speculative, but the hybrids may serve novel growth functions, since the IGF1R and IR differ in some specific signaling partners and/or in the extent to which they activate shared signaling pathways (31-33). Thus, a hybrid receptor could regulate a unique mixture of these signaling properties when stimulated by one member of the class of insulin hormones. To date, genetic evidence has not identified developmental events or physiological activities that are controlled specifically by IGF1R/IR hybrid receptors.

Biosynthetic Pathways

Extensive work from several laboratories described the cellular life cycle of the IR, shown schematically in Figure 2 in five cellular compartments (colored blocks).Biosynthesis begins with transcription of the insulin receptor gene and splicing of the 22 exons, which yields two transcripts (plus or minus exon 11). Exons 1-12 encode the alpha-subunit, and exons 13-22 encode the beta-subunit. Studies are ongoing to fully identify cis-elements and trans-acting factors that elevate transcription of the IR gene in insulin sensitive tissues (34, 35). It has been long-known that glucocorticoids such as dexamethasone can upregulate IR levels (36, 37), and more recently that dexamethasone can alter abundance of the A- versus B-isoforms of the IR-mRNA (38). In contrast to pharamcologically elevated IR expression, insulin resistance could result from genetic defects in trans-acting factors if those defects lead to reduced IR gene expression (39). Finally, the promoter appears sensitive to factors associated with tumorigenic transformation, which may account for increased IR expression in breast cancer (40).

Figure 2. Life Cycle of the Insulin Receptor. The IR is shown as four parallel lines connected by three transverse lines; these represent the alpha- and beta-subunits, and the disulfide bonds, respectively. The organelles are: N, nucleus; ER, endoplasmic reticulum; G, golgi; PM, plasma membrane; A, acidified endosome.

Translation of the IR mRNA is accompanied by membrane insertion due to a 27 amino acid leader sequence, which is subsequently cleaved. Both alpha and beta subunits are present in a single 1355 amino acid polypeptide chain The nascent protein is also co-translationaly glycosylated Over a period of 1.5-3 hours, the pro-receptor carbohydrate chains are modified, dimerization and disulfide bond formation take place, and cleavage between the alpha and beta subunits occurs (41-43). Acquisition of insulin binding activity precedes cleavage (44, 45), and this feature may underlie some processing defects observed in alpha-subunit mutants (46). Cleavage is essential for normal insulin binding and signaling, although it is not absolutely required for cell surface expression (47). Each mature alpha subunit is 735 amino acids and has an apparent mass of 125 kDa by sodium dodecylsulfate-polyacrylamide gel electrophoresis. Each beta subunit is 620 amino acids and has an apparent mass of 95 kDa. Movement of the IR precursors in the endoplasmic reticulum is mediated by Hsp70 family of molecular chaperones, which can control sorting to the exocytic compartment versus sorting to degradation (48).

Transit of the mature receptor from the golgi network to the plasma membrane is controlled separately from internalization, and probably is limited by earlier steps in the biosynthesis. The insulin receptor, at the plasma membrane, is exposed to circulating insulin, whose binding rapidly triggers insulin receptor autophosphorylation on tyrosine residues (described below). Prolonged exposure to insulin eventually leads to down-regulation of receptor levels in cultured cells. This also occurs in laboratory rodents made chronically hyperinsulinemic (49, 50), suggesting that some insulin resistance may be expected from hyperinsulinemia. There is also insulin-triggered serine and threonine phosphorylation of the insulin receptor, which is generally thought to decrease signal transduction (51, 52) although it is not obviously linked to diabetes (53). The majority of insulin signal transduction takes place at the cell surface or during internalization, but little if any appears to occur during the recycling phase (54, 55). The signaling compartment includes the membrane because of the importance of phospholipids in insulin signaling, especially through phosphatidylinositol 3'-kinase (56). Insulin signaling is reversibly terminated by dephosphorylation, which occurs in the plasma membrane and endosomal compartments. Insulin action can be inhibited by interference with insulin binding, autophosphorylation, or substrate phosphorylation, as described below.

Receptor internalization is also triggered by insulin binding. Insulin is removed from circulation by the IR (57), through endocytosis. The endosome becomes acidified, thereby promoting the release of insulin. Endosomal sorting leads to recycling of the receptor to the plasma membrane (through or bypassing the golgi), or to a lysosomal compartment where insulin and its receptor are degraded (58, 59 ). In addition, recent evidence shows receptor ubiquitination and subsequent degradation by the proteosome (60). However, this pathway appears to be more significant for insulin receptor substrate degradation, occurring in cell-type and IGF-1 versus insulin dependent manners (61-63). The normal half-life of the IR is 8 to 16 h, depending on cell type, and degradation is accelerated by insulin binding. As mentioned, with prolonged insulin exposure the steady-state abundance of cellular IR decreases (64). The clinical implication of this "down regulation" feature of receptor cell biology is that hyperinsulinemia should lead to fewer cell surface receptors and thereby contribute to insulin resistance (49, 65). In some cells, such as vascular endothelium, there is also IR mediated transcytosis of insulin, a mechanism by which insulin can enter the central nervous system (66). Insulin entry to the CNS is adversely affected by obesity (67). Interestngly, the presence of IR in the CNS was demonstrated thirty years ago, where its proposed function was sensing hypoglycemia (68, 69).

ACTIVATION AND MODULATION OF INSULIN RECEPTOR FUNCTION

The Insulin Receptor is a Protein Kinase

Another key breakthrough in our knowledge of insulin action was discovery of the protein-tyrosine kinase activity of the IR, first reported in 1982 to occur in vitro and in whole cells (70, 71). This kinase is specific for tyrosines, in contrast to the serine and threonine specificity of the earliest known protein kinases, which were also involved in glucose homeostasis (the cAMP-dependent protein kinase and glycogen synthase kinase). The enzymatic activity of the IR is probably essential for most - if not all - of insulin signaling in cells. Insulin binding to the extracellular alpha-subunits of the IR stimulates autophosphorylation of the intracellular kinase domains, which are encoded within the beta-subunits. Seven intracellular tyrosines become autophosphorylated in response to insulin (72); the same sites are used in cells and in vitro (73, 74). Functionally, autophosphorylation sites causes a 200-fold increase in catalytic activity (75), and equally important they form recruitment sites for bringing substrate proteins to the IR (76). Site-specific mutations among these autophosphorylation sites demonstrated that both enhanced catalytic activity and the ability to recruit target proteins are essential for efficient insulin action (77).

Autophosphorylation is the first intracellular event following insulin binding, reaching a maximum level within seconds (78). This is followed within seconds to minutes by substrate protein phosphorylation. These in turn mediate the biological effects of insulin, which can be manifest within minutes (increased glucose uptake in fat or muscle cells), or within minutes to hours (regulation of gene expression and/or protein degradation), or days (as in organ development through a program of cell differentiation). Every kinase is expected to have specific substrates, and the IR has many (79). The first reported substrate was a 120 kDa protein from liver, which is an ecto-ATPase whose function in insulin signaling may be related to IR internalization (80, 81). However, the most prominent targets are now identified as insulin receptor substrates (IRS-1 through IRS-4), a set of homologous proteins which serve as scaffolds on which to make multimolecular signaling complexes (82).

Dephosphorylation as a Mechanism to Terminate Insulin Signaling

For most protein phosphorylation reactions, there should be a dephosphorylation reaction to reverse its effect. (The major exception would be phosphorylation that triggers protein degradation.) Two protein tyrosine phosphatases (PTPases) have been implicated in dephosphorylation of the insulin receptor, and therefore with terminating insulin action without degrading the insulin receptor (allowing it to recycle). The leukocyte common antigen-related phosphatase (LAR, a transmembrane receptor-like PTPase) is one candidate . The other is PTP-!B, an intracellular PTPase. Either PTPase can dephosphorylate the IR in vitro and in cells. There is genetic evidence from PTPase knockout mice indicating that either PTPase has functions in glucose homeostasis that correlate, at least, with levels of IR autophosphorylation (83, 84).

Interference with Kinase Activity and Substrate Phosphorylation - Potential Mechanisms of Insulin Resistance

Tyrosine-kinase activity is intrinsic to the IR, and thus to insulin action. Both autophosphorylation and substrate phosphorylation can be affected by serine and threonine phosphorylation of the IR and IRS molecules. Typically the effect is to decrease tyrosine phosphorylation of the IR and IRS proteins, which would be the molecular basis for inhibiting insulin action (85-87). For example, the phosphorylation of IR on specific residues in the unique carboxy-terminus is stimulated by phorbol esters, implicating protein kinase C iso-enzymes (88). It has also been reported that serine phosphorylation of the IR is elevated in tissues from polycystic ovary syndrome, which may be linked to insulin resistance in that disease (89). Serine phosphorylation of IRS molecules also can be elevated in the tissues of insulin resistant or diabetic subjects; this would be coincident with decreased ability of insulin to promote IRS tyrosine-phosphorylation (e.g.,(90-92). There is evidence potentially linking obesity with insulin resistance via the increased serine phosphorylation of IRS-1 due to elevated serum lipids (93, 94), and/or increased tumor necrosis factor-alpha (95, 96). Furthermore, many enzymes that are normally activated by insulin are themselves protein serine/threonine kinases; some of these apparently phosphorylate the IR and IRS, providing negative and/or positive feedback loops on insulin action (e.g., (97, 98). Thus, interference with insulin-stimulated tyrosine phosphorylation is one likely contributor to insulin resistance. However, not all serine phosphorylation of the IR and IRS proteins is inhibitory (99).

In addition to TNF-alpha, which apparently acts through serine kinases, two proteins that affect insulin receptor function and may be linked to insulin resistance have been described. The membrane glycoprotein plasma cell antigen-1 (PC-1) was isolated originally from fibroblasts of a type 2 diabetic patient (100). It is elevated in other tissues of diabetics, and it appears to decrease insulin stimulated IR autophosphorylation (101). In addition, although its biochemical function is unknown, a polymorphism has been found that is more potent in its ability to inhibit insulin signaling (102). There is some controversy concerning its specificity for the insulin receptor and whether it is broadly linked to insulin resistance (103-105). A second candidate is fetuin, which appears to have a range of biological functions and binds several proteins, including transforming growth factor-beta and matrix metalloproteinases (106, 107). It was identified as an inhibitor of insulin-stimulated IR autophosphorylation, with an apparently selective effect on the mitogenic pathways activated by insulin in cultured cells (108, 109). Never-the-less, the fetuin knockout mouse, in addition to calcium homeostasis dysregulation and bone growth plate defects (106, 110) shows improved insulin sensitivity in liver and skeletal muscle and resist weight gain on a high fat diet (111). Thus, while much of the current effort is focused on intracellular effectors of insulin signaling, future studies should continue to establish extracellular mediators of insulin resistance.

MOLECULAR BIOLOGY OF THE INSULIN RECEPTOR

Cloning the Insulin Receptor cDNA and Gene

An important accomplishment resulting from IR purification was having enough amino acid sequence to permit cloning of the human IR cDNA. This was published in two separate reports in 1985, representing the collaborations of several laboratories (112, 113). Cloning of the IGF1R and IGF2R followed shortly (114-116). Surprisingly, the two IR cDNAs differed by the absence or presence of 36 nucleotides encoding 12 amino acids. Subsequent cloning of the human IR gene identified 22 exons, and showed that the cDNAs arose from alternative splicing of exon 11. These yield an IR-A isoform (without exon 11) and an IR-B isoform (including exon 11). The IR-A isoform is more abundant in fetal tissues, and the IR-B isoform is generally more abundant in adult tissues (17), although it is notably absent from brain.

Given the clinical importance of insulin action in glucose homeostasis, the DNA sequence of the IR opened multiple lines of investigation to probe for relationships among IR structure, function, and insulin resistance. First, a search was made for mutations in the IR that may be linked to insulin resistance or diabetes. A limited number of mutations were found which account for only a very small fraction of patients with insulin resistance or Type 2 diabetes (46, 117). There were a few mutations that altered insulin binding, and several lesions were found in the kinase domain of the IR. Some mutations affected mRNA stability, and many of the mutations leading to failures of insulin signaling were from improper biosynthesis or increased degradation that lead to reduced expression of IR at the cell surface, and are thus best understood in the context of the receptor's cell biology. The second application for the DNA sequence was to use the IR cDNA as a tool for molecular genetic and structural studies of IR function. For example, mutation of a conserved lysine in the kinase domain yielded an intact IR lacking kinase activity. When expressed in cells, it has a dominant negative effect on signaling through the endogenous IR, blocking both auto- and substrate phosphorylation (118). This demonstrated that kinase activity is essential for the IR to transmit the insulin signal. Mutagenesis of individual autophosphorylation sites demonstrated their explicit functions for enzymatic activity versus substrate recruitment, as mentioned above. In addition, a third application of the cDNA was to identify proteins that might interact with the intracellular portion of the beta-subunit, including the kinase domain. These "yeast two-hybrid" screens have been very successfully employed with the IR and IGF1R, and can readily identify proteins that interact more favorably with one receptor than the other (119-125). Finally, the gene's structure has been essential for generating knockout mice as models in which to study insulin receptor function, described below

Structures for Domains of the Insulin Receptor

In 1994, the crystal structure of the catalytic core of the IR was determined, without tyrosine phosphorylation, and three years later the structure of the autophosphorylated kinase domain was determined with a peptide substrate bound in the active site (126, 127). These provided essential insights into the molecular basis for kinase activation by autophosphorylation. Also important, these were the first structures of a protein tyrosine kinase to be determined in its basal and activated states, as well as the first tyrosine kinase to be cocrystallized with a substrate peptide bound. Known mutations in this domain of the IR were mapped onto the crystal structure, indicating most would disrupt enzymatic activity, thus explaining their deleterious effects on insulin signaling. More recently, a crystal structure of the first three extracellular domains of the IGF-1R was determined (the L1-cysteine rich-L2 domains; (128). These encompassed three of the four domains essential for hormone binding. Although the protein was monomeric and did not bind IGF, it has been extremely important in refining our understanding of insulin versus IGF binding at the molecular level. In brief, most of the insulin binding determinants are in the L1 domain (the first 120 amino acids) and the last 30 amino acids of the alpha-subunit (including those encoded by exon 11), whereas a second region - the cysteine-rich domain adjacent to L1 - also participates in IGF binding. The larger binding surface of the IGF1R matches the larger volumes of IGFs versus insulin (129-132). Furthermore, this ectodomain structure should be a basis for the design of insulin-mimetic or antagonistic pharmaceuticals. Importantly, although there is no crystal structure of the entire receptor, there are several three dimensional structures deduced from electron microscopy (133-136). Although these are not in uniform agreement, one recent higher resolution structure of the native IR derived from scanning transmission electron microscopy has incorporated the known crystal structures, providing a detailed (but controversial) molecular analysis of how insulin binding might be translated into activation of kinase domain autophosphorylation (137).

SHIFTS IN PARADIGMS FOR INSULIN RECEPTOR PHYSIOLOGY

Some of the early observations or conclusions about the IR have been modified in the genomic era. The shifts in paradigm have come from a range of experimental approaches, including forward and reverse genetics using model organisms including mice, Drosophila melanogaster, and Caenorhabditis elegans. Perhaps the most important basic shift is that insulin receptor functions in other than the "classically" insulin responsive tissues - skeletal muscle, fat, and liver - also contributes to insulin resistance, as well as to glucose sensing (67). Functions of the IR in non-classical tissues include: In the pancreas, the IR affects insulin secretion. In the vascular tissues, IR is required for the transcytosis of insulin, allowing this hormone to reach the central nervous system (138). There are also direct effects of IR signaling in the vasculature, where insulin resistance contributes to diabetic retinopathy. In the brain, IR is one of several regulators feeding behavior, reproduction, and also memory (139). There are also direct effects of IR on female reproduction. Perhaps most striking is the contribution of insulin signaling to lifespan, which appears to be largely of neuronal origin (140).

Insulin Receptor Knockouts as Models for Insulin Resistance

Mice have been engineered that are null for IR expression, either throughout the organism or in selected tissues (reviewed in (141). Through these studies it was shown that IR is absolutely required for postpartum survival, since mice lacking IR (homozygous null) die from severe ketoacidosis shortly after birth (142). Knockout mice also identified physiological differences between the IR and IGF1R (143). For example, the IR has less importance for intra-uterine growth than IGF1R, but the IR is also a receptor for IGF2, as described above. As indicated in Table 2, combined heterozygous mice, and tissue-specific IR knockouts have been useful for the analysis of insulin resistance and diabetes, since peripheral insulin resistance combined with failure of the endocrine pancreas to make and/or secrete sufficient insulin are the clinical features of diabetes. Thus, the polygenic nature of insulin resistance was given greater experimental support when it was shown that various combinations of IR, IRS-1, and/or IRS-2 heterozygous mice developed different clinical features of insulin resistance. The physiology of these compound heterozygous mice varied in a tissue-specific manner: heterozygous IR with heterozygous IRS-1 affected primarily muscle and displayed beta-cell hyperplasia, heterozygous IR with heterozygous IRS-2 affected liver more than muscle, and mice heterozygous for all three of these proteins showed severe insulin resistance in muscle, liver and also developed beta-cell hyperplasia (144, 145). In tissue-specific monogenic studies to determine tissue-specific features of insulin resistance, the IR has been ablated selectively in white or brown fat, skeletal or cardiac muscle, liver, endocrine pancreas (beta-cells), and brain. A very important finding is that insulin resistance in fat is critical to insulin resistance in skeletal muscle, while also showing that IR signaling is important for fat cell development. Also in skeletal muscle, IGF1R can compensate for the failure of IR to regulate glucose uptake, essentially confirming earlier studies (146). It is now understood that the IGF1R is important for pancreatic beta cell development, while the IR is important for beta cell function. In summary, these studies reveal that fat cell insulin resistance is a major effector of muscle insulin resistance, and that compensation for insulin resistance can occur in tissues other than those lacking one or more components of the IR signaling pathway.

The null and tissue-specific IR deletions summarized above also revealed important developmental roles for the IR (and related receptors and substrates; see (141). In broader terms, the IGF1R pathways seem more important for body size - in mammals - although many of the molecules involved in IGF1R signaling are shared with the IR. Still, physiological differences between the IR and the IGF1R are implicit in their primary structures. As cited above, recent studies continue to identify specific proteins that interact differently with one versus the other receptor and thereby contribute to the different biological processes that each receptor controls.

Insulin Receptor Functions in Worms and Flies as Models for Human Physiology and Disease

The human insulin receptor has known homologs as far down the evolutionary ladder as hydra, based on sequence information available in GenBank. Among invertebrates, mutations have been identified in the insulin receptor homologs of D. melanogaster and C. elegans. Each of these model organisms' genomes has one IR homlog, but carries between seven and thirty-seven insulin-like hormones as well (158, 159). Given that there is only one receptor, it is regarded as a homolog of both the IGF1R and IR; the invariant residues across evolution do not distinguish between the IR and IGF1R, and only among vertebrates are there three closely related receptors (including the insulin receptor-related receptor, of unknown function and unknown ligand).

To date, several remarkably consistent observations have been made about conserved IR function, based on mutations in flies and worms. First, decreased signaling through the invertebrate IR increases adult lifespan. Second, there is decreased female fertility. Third, these effects are due at least partially to neuronal/neuroendocrine control (160-162). There are also defects in lipid metabolism that accompany loss-of-function IR mutations. Equally striking, most of the signaling molecules regulated by IR kinase activity in mammals are also present in essentially the same hierarchies (140). It is expected that the genetic tools available in these model organisms will allow rapid discovery of novel IR/IGF1R functions; for example, it has been found recently that the Drosophila IR is important for axon guidance in development of the visual system (Song, Kohanski and Pick, in press). Another intriguing discovery is that ablation of insulin secreting neurons in Drosophila provides a good model for pancreatic development (163). Finally, at least one molecular component of insulin-sensitive glucose transport in adipocytes has a Drosophila homolog (164). These organisms provide another avenue for analysis of insulin signaling, and may identify new components of these systems. There is always the expectation and hope that these may be good targets for development of pharmacological agents to relieve or cure insulin resistance and diabetes.