Insulin is synthesized as a single polypeptide called proinsulin that is processed into the disulfide-linked A and B chains of mature insulin and the excised C-peptide. Insulin possesses two asymmetric receptor binding surfaces designated S1 and S2 (Fig. 8-1A). Referred to as the “classical site,” S2 is defined by residues from the A and B chain, including Asn^A21, Gly^B23, Phe^B24, Phe^B25, and Tyr^B26 (see Fig. 8-1A).³² Systematic analysis by site-directed mutagenesis reveals a distinct binding surface (S1) that is necessary for high-affinity receptor binding. S1 includes amino acid residues outside the classical site, especially Leu^A13 and Leu^B17 (see Fig. 8-1A). Insulin binding kinetics suggest that S1 binds to the receptor before S2.³² Human IGF-1 and IGF-2 display high sequence similarity to both the A and B chains of insulin but retain the homologous connecting peptide; IGFs also have an extension at the C-terminus known as the D-domain.³³

FIGURE 8-1 Insulin and insulin receptor structure. A, Insulin structure showing the position of critical amino acids that compose the two binding surfaces (S1^ins and S2^ins) that interact with the L1 and L2 regions of the insulin receptor. B, A linear diagram of the insulin receptor (IR) precursor protein showing the position of important modules in the α and β subunits, which are generated by cleavage at the processing site: insulin contact points (L1 and L2), cysteine-rich region (CR), disulfide bonds, the IRA/IRB splice site (CT-IRA/-IRB), transmembrane region (TM), juxtamembrane insulin receptor substrate binding motif (NPXpY), site of receptor auto-tris-phosphorylation (activation loop), and carboxyl-terminal tyrosines (CT). C, Diagram of the mature insulin receptor composed of two extracellular α subunits and two largely intracellular β subunits. Contiguous modules of the two α subunits are indicated by black or white letters. The holoreceptor is stabilized extracellularly by disulfide bonds between cysteine residues (S-S) in the α and β subunits as well as by noncovalent interactions. Regions within the α subunit that contribute to insulin binding include the L1 and L2 regions and the extra 12 amino acids encoded by exon 11, found only in the B form of the insulin receptor. The β subunit contains the tyrosine kinase catalytic domain with an ATP binding site (Lys₁₀₃₀) and a number of tyrosine phosphorylation sites, including those in the juxtamembrane region (pY₉₇₂), activation loop (pY_{1158, 1162, 1163}), and carboxyl-terminal regions.

The mammalian receptors for insulin (IR) or insulin-like growth factors (IGF1R) are composed of an extracellular ligand-binding domain that regulates the activity of an intracellular tyrosine kinase.34–36 Most receptor tyrosine kinases are activated by ligand-induced dimerization that promotes tyrosine phosphorylation of the kinase-activation loop (A-loop) and other sites that recruit cellular substrates.³⁷ However, the homologous IR and IGF1R reside in the plasma membrane as inactive covalent dimers. High-affinity insulin or IGF binding adds a transient cross-link to stabilize the active conformation, which promotes autophosphorylation and exposes the catalytic domain to other cellular proteins.^38–40

The IR is encoded by a 150-kb gene on human chromosome 19p13.3-p13.2 that contains 22 exons. Exon 11 is alternatively spiced, depending upon the tissue and developmental stage, to produce two IR isoforms: IRA lacks the residues encoded by exon 11, and IRB includes the 12-amino-acid residues encoded by exon 11.⁴¹^,⁴² By comparison, the IGF1R gene is assembled without alternative splicing from its 19-exon gene located on human chromosome 15. A third gene encodes the homologous insulin-related receptor (IRR), which can promote male sexual development in mice; however, its high-affinity ligand is unknown, and how it interacts with the IR and IGF1R needs to be investigated.⁴³

The IR and IGF1R proreceptors are synthesized with a classic signal sequence followed by well-defined modules that include two leucine-rich motifs (L1 and L2) flanking a cysteine-rich (CR) region followed by three fibronectin-III motifs (FnIII1, FnIII2, and FnIII3) (see Fig. 8-1B). FnIII2 is interrupted by a 120-amino-acid insert domain containing a cleavage site that generates the α and β subunits of the mature receptor (see Fig. 8-1B). During translation, the proreceptors for insulin and IGF-1 assemble as homodimers to produce the IR or the IGF1R, or as heterodimers to produce receptor hybrids (IR::IGF1R). Since the insulin receptor occurs in two isoforms, a total of five receptors types can be produced from two genes.⁴⁴ All of the proreceptor dimers are linked by disulfide bonds and cleaved into covalently stabilized tetramers composed of two α and two β subunits (see Fig. 8-1C; Fig. 8-2).The α subunits are entirely extracellular and create the ligand-binding site. Each β subunit contains a transmembrane-spanning segment that separates the extracellular FnIII2 and FnIII3 regions from the intracellular tyrosine kinase (see Fig. 8-1C).³⁴^,³⁵ Under nonreducing conditions, the holoreceptor (α₂β₂) has an approximate molecular mass of 350,000—larger than expected, owing to glycosylation of the α and β subunits.⁴⁵^,⁴⁶ In denaturing polyacrylamide gel electrophoresis, the α and β subunits migrate near 135 kD and 95 kD, respectively.^47,48

FIGURE 8-2 The insulin and insulin-like growth factor family. The insulin/IGF family consists of three peptide hormones: insulin, insulin-like growth factor 1 (IGF-1), and insulin-like growth factor 2 (IGF-2). These bind as indicated in the figure to five distinct homo- or heterodimeric receptor isoforms that generate cytoplasmic signals: two insulin receptor isoforms, IRA and IRB; the insulin-like growth factor receptor, IGF1R; and two hybrid receptors, IRA::IGF1R and IRB::IGF1R. The insulin receptor is the primary target for insulin throughout development and life. The IGF-1 receptor is the primary target for IGF-1. IGF-2 binds to IRA during embryonic development, and to the IGF1R throughout life; binding of IGF-2 to the mannose 6-phosphate receptor (M6Pr) results in its cellular uptake and destruction. The activated IR or IGF1R tyrosine kinases phosphorylate the cytoplasmic insulin receptor substrate proteins, IRS1 and IRS2, which mediate somatic growth and metabolism. The approximate binding affinities (K_d, nM) are shown on the figure.

Receptor/Ligand Binding

Insulin binds with high affinity (K_d < 0.5 nM) to the homodimeric IRB, which predominates in the classic insulin target tissues—adult liver, muscle and adipose tissues (see Fig. 8-2). IRB is selective for insulin because its K_d for IGF-1 and IGF-2 is at least 50- to 100-fold higher. Adult liver and adipose are purely insulin-responsive tissues, expressing IRB but lacking IGF1R.⁷ By comparison, IRA predominates in fetal tissues, the adult central nervous system, and hematopoietic cells.^49–52 IRA binds insulin almost as well as IRB but also binds IGF-2 with moderate affinity (see Fig. 8-2). IGF-1 and IGF-2 bind with high affinity (K_d < 1 nM) to the homodimeric IGF1R and to the hybrid receptors (IGF1R::IRA and IGF1R::IRB), whereas insulin barely binds to these proteins (see Fig. 8-2).⁵³^,⁵⁴ Thus, under ordinary conditions, insulin never activates the IGF1R tyrosine kinase, whereas IGF-1 and IGF-2 can activate the IR tyrosine kinase, especially when it forms a hybrid with the IGF1R.

Site-directed mutagenesis reveals the location of two insulin-binding sites in the α subunit of the insulin receptor.⁵⁵^,⁵⁶ Chimeric receptors between the α subunits of the insulin and IGF-1 receptors confirm that one of the sites is located within the first leucine-rich (L1) region of the insulin receptor.⁵⁷ Photo-affinity labeling of mutant insulin receptors reveals the second insulin binding site in the FnIII1 region.⁵⁸ High-affinity insulin binding depends upon α subunit dimerization in the IR tetramer, insofar as monomeric α subunits bind insulin with low affinity. Insulin bound to the IR tetramer at high affinity dissociates slowly unless the concentration of soluble insulin increases, explaining how the binding of a second insulin molecule to a low-affinity site can reduce the affinity of the first.⁵⁹^,⁶⁰

Insulin binding apparently begins through interactions between the S1 site on insulin and the FnIII1 region in the α subunit.³² Sixteen amino acid residues at the COOH-terminus (CT) of FnIII2 interacts with the L1::CR-region to create a composite insulin binding site—L1::CR::CT—that interacts with the classical receptor binding surface on insulin (S2) (see Fig. 8-1C). Although the α subunits are arranged symmetrically in the dimer, there is a sharp bend between the L2 and FnIII1 regions that juxtaposes the L1::CR::CT domain antiparallel to FnIII1 (see Fig. 8-1C).^61–63 Owing to this arrangement of α subunits, insulin binds to the L1::CR::CT domain of one α subunit and to FnIII1 of the adjacent α subunit, creating the cross-link that activates the kinase. Space constraints allow only one insulin molecule to bind with high affinity.^40,61,63 Inclusion of exon 11 in IRB lengthens the CT-region by 12 amino acids, which modifies the L1::CR::CT domain to exclude IGF-2 binding and increase the insulin-binding affinity.

Insulin Receptor Tyrosine Kinase

The tyrosine kinase activity of the insulin receptor was originally found by biochemical experiments using [³²P]-labeled hepatoma cells or partially purified insulin receptors incubated with insulin and [γ³²P]ATP.64–66 However, until the insulin receptor cDNA was isolated and sequenced,³⁴^,³⁵ many other mechanisms for signal transduction continued to be investigated.⁶⁷^,⁶⁸ The discovery that rare cases of severe insulin resistance in humans are associated with insulin receptor mutations that inactivate the tyrosine kinase—without altering insulin binding—supported the central role of tyrosyl phosphorylation in the mechanism of insulin action.⁶⁹

The intracellular portion of the insulin receptor β-subunit is composed of three distinct regions that contain tyrosyl phosphorylation sites: Y₉₆₅ and Y₉₇₂ in the juxtamembrane region (IRB) between the transmembrane helix and the cytoplasmic tyrosine kinase domain; Y₁₁₅₈, Y₁₁₆₂, and Y₁₁₆₃ in the activation loop (A-loop) of the core catalytic domain (IRB); and Y₁₃₂₈ and Y₁₃₃₄ in the COOH-terminus (IRB).41,70,71 The unphosphorylated Tyr₁₁₆₂ folds into the catalytic site to restrict ATP binding, which explains the high apparent K_m for ATP before insulin stimulation.⁷²^,⁷³ However, the closed A-loop is in equilibrium with an alternate conformation that allows occasional access by ATP to mediate basal autophosphorylation.⁷⁴ Infrequent oscillation from the inactive to the active conformation might be coupled to complementary changes in the α subunits, which are definitively stabilized upon insulin binding to accelerate ATP entry and autophosphorylation of Tyr₁₁₆₂ (Fig. 8-3A). Once initiated, the autophosphorylation cascade progresses rapidly to Tyr₁₁₅₈, resulting in a bis-phosphorylated and active kinase.⁷⁰ Although autophosphorylation of Tyr₁₁₆₃ is relatively slow, the resulting tris-phosphorylated A-loop stabilizes the open conformation to allow unrestricted access by Mg-ATP and protein substrates (see Fig. 8-3A). Autophosphorylation probably occurs through the interaction between adjacent β subunits, since the reaction progresses through first-order kinetics.⁶⁶ This model of kinase regulation is validated by activation of the kinase via substitution of Asp₁₁₆₁ in the middle of the A-loop with Ala₁₁₆₁—which shifts the steady-state conformation of the unphosphorylated A-loop toward the open configuration.⁷⁴ Substitution of Tyr₁₁₆₂ with phenylalanine also increases basal autophosphorylation, consistent with its role in stabilizing the closed conformation.⁷⁵

FIGURE 8-3 The role of insulin receptor autophosphorylation. A, The structure of the insulin receptor activation loop is shown as a series of ribbon diagrams of the insulin receptor kinase domain. The activation loop (A-loop) is darkly shaded; the three activation loop tyrosine residues (Y₁₁₅₈, Y₁₁₆₂, and Y₁₁₆₃) are shown with their side chains. In the inactive, unphosphorylated state (left panel), the activation loop blocks access by potential substrates. Following phosphorylation, however (right panels), the activation loop moves, allowing substrates such as YMXM peptides of the IRS proteins to access the active site. B, A structural representation of the binding of dimeric APS SH2 domains to the phosphorylated A-loop of the insulin receptor. These structures were based upon published coordinates.73,77,101

Regulation of the insulin receptor kinase is modulated by heterologous protein interactions with the phosphorylated A-loop. At least two phosphotyrosine residues in the A-loop are completely solvent-exposed, creating sites that bind to various src-homology-2 (SH2)-domain-containing proteins, including APS and Grb10/14.⁷⁶ Structural analysis reveals how the dimerized Src homology-2 (SH2) domains of APS can bind to the A-loop of adjacent β subunits to stabilize the active conformation (see Fig. 8-3B).⁷⁷ By contrast, binding of the dimeric SH2-domains of Grb10 or Grb14 to adjacent A-loops orients a 40-amino-acid motif into in the catalytic site, which inhibits the IR kinase.⁷⁸ These structural predictions are consistent with gene deletion studies that confirm tissue-specific negative regulation of the insulin receptor by Grb14.⁷⁹ Variable expression of these or other regulatory proteins—due to genetic variation or physiologic or environmental challenge—should be expected to modulate insulin signaling and glucose tolerance.⁷⁶

The Insulin Signaling Cascade

Introduction

Following discovery of the insulin receptor tyrosine kinase, many groups searched for insulin receptor substrates that might mediate downstream signals.⁸⁰^,⁸¹ The “substrate” hypothesis was difficult to prove for any receptor tyrosine kinase, because the only known substrates were abundant proteins of unknown physiologic importance. The first evidence for a cellular substrate of a receptor tyrosine kinase came from antiphosphotyrosine antibody immunoprecipitates, which revealed a 185-kD phosphoprotein (pp 185) in insulin-stimulated hepatoma cells.⁸² This substrate seemed to be biologically important because it was phosphorylated immediately after insulin stimulation but not phosphorylated by catalytically and biologically inactive insulin receptors. Importantly, a few catalytically active but biologically inactive insulin receptor mutants fail to phosphorylate pp185.⁸³ Together, these data provided the first clue that substrate phosphorylation following receptor autophosphorylation are the initial steps in signal transduction.

Purification and molecular cloning of pp185 revealed one of the first signaling scaffolds and the first insulin receptor substrate, called IRS1.⁸⁴ Four IRS-protein genes exist in rodents, but only three (IRS1, IRS2, and IRS4) are expressed in humans.⁸⁵ IRS1 and IRS2 are broadly expressed in mammalian tissues, whereas IRS4 is largely restricted to the hypothalamus.⁸⁶ The IRS proteins are adapter molecules that link the IR and IGF1R to common downstream signaling cascades and heterologous regulatory mechanisms (Fig. 8-4A). IRS proteins are targeted to the activated receptors through an NH₂-terminal pleckstrin homology (PH) domain and a phosphotyrosine-binding (PTB) domain (see Fig. 8-4A). During insulin and IGF-1 stimulation, tyrosines in the COOH-terminal portion of the IRS proteins are phosphorylated and bind to the SH2 domains in various signaling proteins (see Fig. 8-4A).⁸⁷ The interaction between IRS1 and the 85-kD regulatory subunit (p85) of the class 1A phosphatidylinositol 3-kinase (PI3K) was the first insulin signaling cascade to be reconstituted successfully in cells and test tubes.⁸⁸ In addition to the IRS proteins, the insulin receptor phosphorylates other proteins in various contexts, including Shc, APS, SH2B, Gab1/2, Dock1/2, and Cbl.^89–96 Quantitative mass spectrometry–based analysis of proteins extracted from insulin-stimulated 3T3-L1 adipocytes reveals 122 tyrosine phosphorylation sites on 89 proteins, including IRS1 and IRS2 (see Fig 8-4A).⁹⁷ A wide variety of proteins are revealed by this approach, including components of the trafficking machinery for the insulin-responsive glucose transporter, GLUT4.⁹⁷ Although the role of each of these substrates merits attention, work with transgenic mice reveals that many insulin responses, especially those associated with somatic growth and carbohydrate and lipid metabolism, are initiated through IRS1 and IRS2.³⁶

FIGURE 8-4 Insulin receptor substrate proteins IRS1 and IRS2, with structure of the IRS2 kinase regulatory loop binding (KRLB) domain. A, Comparison of IRS1 and IRS2 protein sequences, including amino-terminal pleckstrin homology (PH) and phosphotyrosine binding (PTB) domains and numerous known or potential tyrosine phosphorylation sites. The amino acid sequences surrounding tyrosine sites are shown. Phosphorylation sites revealed by tandem mass spectrometry (MS/MS) are indicated. The kinase regulatory loop binding (KRLB) domain in IRS2 is centered about Y624 and is not conserved in IRS1. B, The backbone of the insulin receptor catalytic domain, including the location of the regulatory loop in the inactive kinase, with Tyr1158 indicated for orientation. C, The backbone of the insulin receptor catalytic domain, including the location of the phosphorylated Tyr1158 and Tyr1162 residues within the regulatory loop, oriented away from the catalytic pocket of the active kinase. The interaction of the IRS2 peptide surrounding Tyr628—but lacking Tyr621—is shown. D, The same diagram as C, but Tyr621 is included to show its interaction with the ATP binding pocket of the insulin receptor.

Insulin Receptor Substrate Recruitment and Phosphorylation Site Selection

The structure of the activated insulin receptor β subunit reveals a mechanism by which the activated IR kinase phosphorylates tyrosine residues within specific amino acid motifs, including the YMXM, YVNI, and YIDL motifs.98–101 These motifs are targeted as antiparallel β strands to the COOH-terminal end of the open A-loop, allowing hydrophobic residues in the Y+1 and Y+3 positions to occupy two small hydrophobic pockets on the activated kinase (see Fig. 8-3A). Tyrosine residues lying within amino acid motifs that contain charged or bulky side chains at the Y+1 and Y+3 positions fit poorly into this site, excluding them from phosphorylation.¹⁰¹

In addition to the phosphorylation of specific tyrosine-containing motifs, selective and regulated recruitment of cellular proteins to the activated IR and IGF1R establishes signaling specificity. In this regard, the phosphorylated NPEY₉₇₂ motif in the juxtamembrane region of the insulin receptor (IRB-numbering) is essential.⁸³^,¹⁰² The juxtamembrane region is about 35 residues long and connects the transmembrane helix of the IR β subunit to the kinase domain (see Figs. 8-1B and C). Tyr₉₈₄ in the juxtamembrane region plays an important structural role to maintain the kinase in the basal state.⁷⁸ However, unlike other receptor tyrosine kinases, the insulin receptor kinase is not regulated by autophosphorylation in the juxtamembrane region.¹⁰³^,¹⁰⁴ Phosphorylation of Tyr₉₇₂ creates a docking site for the phosphotyrosine-binding (PTB) domains in the IRS proteins and SHC.⁸³ In intact cells, phosphorylation of the NPXY₉₇₂ motif is among the first sites of insulin-stimulated autophosphorylation.⁸³^,¹⁰⁵ The juxtamembrane region appears to compete with the A-loop for access to the catalytic site, supporting a model in which the juxtamembrane region is phosphorylated before tris-phosphorylation of the A-loop.⁷² In this mechanism, substrate recruitment can be initiated before the catalytic domain is maximally activated. The NPXY motif in the IR juxtamembrane region fills an L-shaped cleft on the PTB domain, while the N-terminal residues of the bound peptide form an additional strand in the β sandwich.¹⁰² These interactions can explain why the IR and IGF1R—and a few cytokine receptors, including the IL-4 receptor—select IRS1 for tyrosine phosphorylation.

The IRS proteins also contain a pleckstrin homology (PH) domain that is structurally similar but functionally distinct from the PTB domain.¹⁰⁶ Although the PH domain promotes the interaction between IRS proteins and the IR, it does not bind phosphotyrosine, and its mechanism remains poorly understood. PH domains are generally thought to bind phospholipids, but the PH domains in IRS1 and IRS2 are poor examples of this specificity.¹⁰⁷^,¹⁰⁸ Regardless, the PH domain in the IRS protein plays an important and specific role because it can be interchanged among the IRS proteins without noticeable loss of bioactivity. However, heterologous PH domains inhibit IRS1 function when substituted for the IRS1 PH domain, confirming a specific functional role.¹⁰⁹

IRS2 utilizes an additional motif located between amino acid residues 591 and 786—especially Tyr₆₂₄ and Tyr₆₂₈—to interact with the activated insulin receptor.¹¹⁰^,¹¹¹ This binding region in IRS2 was originally called the kinase regulatory loop binding (KRLB) domain because tris-phosphorylation of the IR A-loop was required to observe its interaction.¹¹⁰ Autophosphorylation moves the A-loop out of the catalytic site so the functional part of the KRLB domain—residues 620 to 634 in murine IRS2—can fit into the catalytic site (see Fig. 8-4B-D).¹¹² This interaction aligns Tyr₆₂₈ of IRS2 for phosphorylation; however, it also inserts Tyr₆₂₁ into the ATP-binding pocket, which might attenuate signaling by blocking ATP access to the catalytic site. By contrast, this interaction might promote signaling by opening the catalytic site before tris-autophosphorylation. Interestingly, the KRLB motif does not bind to the IGF1R, which might explain signaling differences between IR and IGF1R, as well as the receptor hybrids.¹¹²

The IRS→PI3K→AKT Cascade

One of the best studied and most important signaling cascades activated by insulin involves the production of phosphatidylinositol lipids by the class 1A phosphatidylinositide 3-kinase (PI3K) (Fig. 8-5). The PI3K plays a role in many signaling cascades. It is composed of a regulatory subunit that contains two src-homology-2 (SH2) domains and a catalytic subunit that phosphorylates the D3 position of the inositol ring.¹¹³ The catalytic subunit is unstable and only detected in association with the regulatory subunit. The PI3K is inactive until both SH2 domains in the regulatory subunit are occupied by phosphorylated YXXM motifs, especially those in the IRS proteins.⁸⁸ Inhibition of the PI3K by chemical or genetic means blocks almost all metabolic responses stimulated by insulin—including glucose influx, glycogen and lipid synthesis, and adipocyte differentiation—confirming that the PI3K is a critical node coordinating insulin action.^114–116

FIGURE 8-5 Insulin and insulin-like signaling cascade. A, Two main branches propagate signals generated via the IRS proteins: the phosphatidylinositide 3-kinase (PI3K)→3-phosphoinositide-dependent protein kinase-1(PDK1)→Akt and Grb2/Sos→ras kinase cascades. Activation of the receptors for insulin and IGF-1 results in tyrosine phosphorylation of the IRS proteins, which bind PI3K and Grb2/SOS. The Grb2/SOS complex promotes GDP/GTP exchange on p21^ras, which activates the ras→raf→MEK→ERK1/2 cascade. Activated ERK stimulates transcriptional activity by direct phosphorylation of elk1 and by indirect phosphorylation of fos through p90^rsk. The activation of PI3K by IRS-protein recruitment produces PI 3,4P₂ and PI 3,4,5P₃ (antagonized by the action of PTEN or SHIP2), which recruit PDK1 and AKT to the plasma membrane. AKT is activated via phosphorylation at T308 by PDK1 and at S473 by mTOR in complex with rictor. The mTOR kinase is activated by Rheb^GTP, which accumulates upon inhibition of the GAP activity of the TSC1::TSC2 complex following PKB-mediated phosphorylation of TSC2. The S6K is primed for mTOR-mediated activation through phosphorylation by PDK1. AKT phosphorylates many cellular proteins, inactivating PGC1α, p21^kip, GSK3β, BAD, and AS160 and activating PDE3b and eNOS. AKT-mediated phosphorylation of forkhead proteins, including FOXO1, results in their sequestration in the cytoplasm, which inhibits their influence upon transcriptional activity. Insulin stimulates protein synthesis by altering the intrinsic activity or binding properties of key translation initiation and elongation factors (eIFs and eEFs, respectively) as well as critical ribosomal proteins. Components of the translational machinery that are targets of insulin regulation include eIF2B, eIF4E, eEF1, eEF2, and the S6 ribosomal protein.¹⁵⁸ TNF-α activates JNK, which can phosphorylate IRS1 and inhibit its interaction with the insulin receptor and subsequent tyrosine phosphorylation. IRS2 expression is promoted by nuclear FOXO, which increases IRS2 concentration in the fasted liver.

PI-3,4,5-P₃ produced by the activated PI3K recruits several Ser/Thr-kinases to the plasma membrane, including the 3-phosphoinositide-dependent protein kinase-1 (PDK1) and v-akt murine thymoma viral oncogene (AKT, also known as PKB), where AKT is activated by PDK1-mediated phosphorylation of Thr₃₀₈ in its activation loop. AKT has a central role in cell biology, phosphorylating many proteins that control cell survival, growth, proliferation, angiogenesis, metabolism, and migration (see Fig. 8-5).¹¹⁷ Phosphorylation of several AKT substrates is especially relevant to insulin-like signaling: GSK3α/β (blocks inhibition of glycogen synthesis), AS160 (promotes GLUT4 translocation), the BAD•BCL2 heterodimer (inhibits apoptosis), the FOXO transcription factors (regulates gene expression in liver, β cells and the hypothalamus), p21^CIP1 and p27^KIP1 (blocks cell cycle inhibition), eNOS (stimulates NO synthesis and vasodilatation), and PDE3b (hydrolyzes cyclic adenosine monophosphate [cAMP]) (see Fig. 8-5).

AKT directly stimulates cell growth through the activation of mTORC1 complex—composed of the mammalian target of rapamycin (mTOR), mLST8, and raptor—which phosphorylates the S6-kinase and eIF4E-BP1 (also known as PHAS-1) to stimulate protein synthesis (see Fig. 8-5).^117–119 The regulatory mechanism involves several steps, beginning with AKT-mediated phosphorylation of at least five sites (Ser939, Ser981, Ser1130, Ser1132, and Thr1462) on TSC2 (tuberin), which in complex with TSC1 (hamartin) functions as a GTPase activating protein for the small G protein, RHEB (Ras homolog enriched in brain). RHEB accumulates in its GTP-bound form when TCS1/2 is inhibited by AKT-mediated phosphorylation, which activates mTORC1 (see Fig. 8-5). In one possible regulatory mechanism, FKBP38 (FK506 binding protein 8) binds to mTORC1 and inhibits mTOR until RHEB-GTP binds to FKBP38 to disinhibit the kinase.^120–122 This complex regulatory cascade is augmented by other pathways, including the direct inhibition of the mTOR catalytic site by PRAS40 (the proline-rich AKT substrate of 40 kD) until AKT-mediated phosphorylation of PRAS40 reverses the inhibition.¹¹⁷ The mTORC2 complex is also composed of mTOR and mLST8, but instead of raptor, this rapamycin-insensitive complex contains rictor and mSIN1 and is mainly regulated by nutrients. Regardless of this complexity, the insulin-mediated control of mTORC1 is directly linked to the IRS→PI3K→AKT cascade and plays an important role in cell growth and feedback regulation.

Downstream effectors of mTORC1—p70^S6K and 4E-BP1—control protein synthesis and cell growth. The p70^S6K occurs as two isoforms, p70^S6K1 and p70^S6K2. Disruption of the S6k1 gene in mice causes glucose intolerance due to reduced size of pancreatic islet β cells; however, peripheral insulin action is enhanced, suggesting that p70^S6K1 contributes to feedback inhibition of insulin signaling.¹²³ Activation of p70^S6K1 is accomplished by multisite phosphorylation events from various kinase activities in response to insulin and other mitogens when amino acids are available.¹²⁴^,¹²⁵ Nutrient sensitivity might arise when mTORC1 mediates the phosphorylation of a hydrophobic motif needed for interaction of p70^s6k with PDK1.

Genetic Validation of the IRS-Based Insulin Signaling Cascade

Systemic Inactivation of Insulin or IGF-1 Receptors

Even though the insulin and IGF-1 receptors have some overlapping functions during development and adult life, complete deletion of the IR or the IGF1R has serious physiologic consequences that cause death shortly after birth.²¹ The IGF-1/2→IGF1R signaling is the principle growth regulatory pathway in fetal mice, which is not influenced by growth hormone or augmented by the insulin receptor until after birth.¹²⁶ IGF1R-deficient mice are born 50% smaller than normal littermates and die after a few days, owing to developmental defects.¹²⁷ By contrast, mice lacking the insulin receptor are nearly normal size at birth, except for a reduced adipose tissue mass. Regardless, insulin receptor–deficient mice also die within a few days of birth, owing to severe hyperglycemia.¹²⁶ Mice retaining at least 20% of normal insulin-receptor expression throughout the body can survive, with severe postnatal growth retardation and hyperglycemia that resembles human leprechaunism.²¹ This growth defect might arise, at least in part, from elevated hepatic IGFBP1 that reduces IGF-1 bioavailability. Thus, small mice with severely reduced insulin or IGF-1 signaling have short lifespans due to developmental and metabolic defects. These mouse models reflect many but not all the features of humans with inactivating human insulin-receptor mutations.¹⁹

Insulin Receptor Substrate–Protein Signaling

Irs1 and Irs2 couple the receptors for insulin and IGF-1 to the PI3K→Akt cascade in mice and humans. Deletion of Irs1 produces small, insulin-resistant mice with nearly normal glucose homeostasis, owing to β-cell expansion and compensatory hyperinsulinemia.¹²⁸ Irs1 appears to mediate most of the Igf1 signal for somatic growth. Mice lacking Irs2 display nearly normal growth but develop life-threatening diabetes between 8 and 15 weeks of age as a result of reduced β-cell mass and insufficient compensatory insulin secretion.¹²⁹ Whereas the complete deletion of Irs1 and Irs2 is lethal, littermates retaining one allele of Irs1 (Irs1^+/−::Irs2^−/−) or one allele of Irs2 (Irs1^−/−::Irs2^+/−) can be born alive.¹³⁰ Irs1^+/−::Irs2^−/− mice develop severe fasting hyperglycemia and die by 4 weeks of age because Irs2 is required for pancreatic β-cell survival and growth (Fig. 8-6). By contrast, Irs1^−/−::Irs2^+/− mice reach only 30% of normal size, but they display nearly normal glucose tolerance and circulating insulin concentrations at 6 months of age¹³⁰ (see Fig. 8-6). Regardless, the small Irs1^−/−::Irs2^+/− mice are very fragile and require extraordinary care to live beyond this age. Thus, healthy glucose tolerance and small size are not definitive determinants of lifespan.

FIGURE 8-6 The lifespan of genetically modified mice, indicated in days or as a percentage of the lifespan of C57BL/6 mice. The targeted genes or PDX1 transgene (PDX1^tg) are indicated on the left. Except where indicated, mice are wild-type and transgene-negative. Genotypes: [+/+] = wild-type; [−/−] = homozygous knockout; [+/−] = heterozygous for knockout allele; [Tg⁺] = transgene-positive.

The central role of IRS proteins in the PI3K→AKT signaling cascade is validated by a wide array of cell-based and mouse-based experiments. Although IRS1 was originally purified and cloned from rat hepatocytes, the principle role of Irs1 and Irs2 during insulin signaling in hepatocytes in vivo was verified only recently.131–134 The simplest experiments employ an intraperitoneal injection of insulin into ordinary mice or mice lacking hepatic Irs1, Irs2, or both. In ordinary mice, insulin rapidly stimulates Akt phosphorylation and the phosphorylation of its downstream substrates, Foxo1 and Gsk3α/β. However, both Irs1 and Irs2 must be deleted before insulin receptors are uncoupled from the PI3K→AKT cascade.¹³⁵ These results confirm the shared but absolute requirement for IRS1 or IRS2 for the hepatic insulin response.

Downstream Insulin Signaling Components

Except for PDK1, the IRS→PI3K→AKT cascade is robust as a result of the expression of multiple isoforms of the proximal components. The type 1_A PI3K is a dimer composed of a catalytic subunit—either p110α, p110β, or p110δ—and one of five regulatory subunit isoforms encoded by three different genes—Pik3r1, Pik3r2, and Pik3r3.¹¹⁷ In most cells, including the liver, products of Pik3r1—p85α, p55α, and p50α—and Pik3r2 (p85β) stabilize and inhibit the catalytic subunits.¹³⁶^,¹³⁷ Partial loss of the regulatory subunits increases insulin sensitivity, which appears to be related to diminished negative feedback to the IRS proteins.¹³⁸ By contrast, the complete disruption of hepatic Pik3r1 and Pik3r2 markedly reduces insulin-stimulated PI3K activity and PIP3 accumulation—at least in part by destabilizing the catalytic subunits—which dysregulates glucose and lipid homeostasis, hepatic size, and function.¹¹⁶ Thus, PI3K activity is responsible for nearly all of insulin’s actions in the liver in vivo.

The PI3K catalytic subunits, p110α and p110β, are ubiquitously expressed, but p110δ is predominantly expressed in leukocytes. Whereas deletion of either p110α or p110β causes embryonic lethality,¹³⁹^,¹⁴⁰ deletion of p110δ causes immune-system defects.¹⁴¹ Mice heterozygous for deletion of either p110α or p110β show no phenotypic abnormalities, but combined heterozygosity for p110α and p110β causes mild glucose intolerance and fasting hyperinsulinemia.¹⁴² Mutation of Asp₉₃₃ of p110α to alanine abolishes its lipid kinase activity, and knock-in mice homozygous for this mutation (D933A) die during embryogenesis.¹⁴³ By contrast, heterozygous D933A mice are viable and fertile but develop severe insulin resistance, glucose intolerance, hyperphagia, and adiposity, indicating a function of p110α not shared by p110β. This unique function derives from the highly selective recruitment of p110α to IRS signaling complexes following insulin stimulation.¹⁴³

The mammalian genome contains three genes that separately encode the AKT isoforms, AKT1, AKT2, and AKT3. The analysis of knockout mice confirms that AKT can regulate important biological functions, including cell proliferation, growth, survival, and differentiation and glucose metabolism in vivo; however, the AKT isoforms are not redundant components of the insulin-like signaling cascade.¹⁴⁴ AKT1 has a major role in embryonic development, growth, and survival but minor effects upon metabolism.¹⁴⁵ By comparison, AKT2-deficient mice display metabolic defects, and AKT3-deficient mice display neural defects.¹¹⁷ AKT2 rather than AKT1 is primarily responsible for insulin-stimulated GLUT4 translocation.¹⁴⁶ Targeted disruption of AKT impairs insulin-stimulated glucose uptake in muscle and adipocytes and prevents the suppression of hepatic glucose output by insulin, which causes glucose intolerance and insulin resistance that progresses to diabetes and β-cell failure.¹⁴⁷ Two human subjects with a dominant-negative mutation in AKT2 display many features of type 2 diabetes—hyperglycemia, increased lipogenesis, elevated liver fat content, TG-enriched VLDL, hypertriglyceridemia, and low HDL cholesterol levels.¹⁹^,¹⁴⁸

Although PDK1 is a master upstream kinase controlling the activation of numerous kinases including AKT, it occurs as a single isoform encoded by one gene. Unlike most kinases, PDK1 is constitutively active, so phosphorylation of its substrates is regulated by mechanisms that control the interaction of substrates with PDK1.¹⁴⁹ PDK1 is required for normal development, inasmuch as mouse embryos lacking PDK1 die at day E9.5; however, PDK1 hypomorphic mice with 90% reduction of PDK1 expression in all tissues are viable, fertile, and small.¹⁵⁰ Surprisingly, activation of AKT1 and S6K1 by insulin is largely normal in PDK1 hypomorphic mice. Moreover, liver-specific PDK1^−/− mice display normal blood glucose and insulin concentrations under ordinary fasting and postprandial conditions; however, they are markedly glucose intolerant and paradoxically fail to normalize their blood glucose after injection of insulin.¹⁵¹ Mistargeted PDK1 due to mutation in the pleckstrin homology domain causes insulin resistance.¹⁵² Thus, PDK1 might not be a rate-limiting step in the insulin signaling cascade; however, excess capacity might be important during acute metabolic challenge when definitive activation of AKT is required for tight metabolic control.

Modulation of Insulin Signaling by Protein and Lipid Phosphatases

Many phosphatases can oppose the action of insulin by dephosphorylating key proteins in the signaling cascade. The protein tyrosine phosphatase, nonreceptor type 1 (PTP1b) and the phosphatase and tensin homolog (PTEN) are two phosphatases that attenuate insulin signals at early points in the pathway. PTP1B reduces insulin sensitivity because it dephosphorylates the insulin receptor and the IRS proteins. PTP1B^−/− mice display increased insulin sensitivity, lower circulating insulin concentrations, and decreased pancreatic β-cell mass.¹⁵³ Complete deletion of PTP1B also increases peripheral insulin sensitivity of compound Irs2^−/−::PTP1B^−/− mice, due at least in part to a reduced rate of β-cell death that sustains insulin secretion in these mice.¹⁵⁴ However, Irs2^−/−::PTP1B^−/− mice eventually lose sufficient β-cell mass between 8 and 9 months of age, and life-threatening hyperglycemia develops (see Fig. 8-6). Thus, inhibition of PTP1B can be effective only when some signaling capacity remains available.

PTEN is a potent negative regulator of insulin action and cellular proliferation and one of the most frequently mutated genes in many forms of human cancer.¹⁵⁵^,¹⁵⁶ PTEN attenuates downstream insulin-like signaling by dephosphorylating PI(3,4)P₂ and PI(3,4,5)P₃ at the 3-position, which reduces the recruitment to the plasma membrane and activation of PDK1 and AKT.¹⁵⁶ Pten^+/− mice are glucose tolerant, even as β-cell mass and circulating insulin levels decrease.¹⁵⁷ PTEN heterozygosity also increases peripheral insulin sensitivity in Irs2^−/−::PTEN^+/− mice and normalizes glucose tolerance, because the small islets in these mice produce sufficient insulin until death from lymphoproliferative disease between 10 and 12 months age (see Fig. 8-6). These experiments highlight the complex relation between nutrient homeostasis, insulin sensitivity and secretion, and cancer that emerges in rodents and humans. Despite this complexity, mild inhibition of PTEN, especially if it can be accomplished in a tissue-specific way, might have therapeutic value.

Regulation of Protein Synthesis by Insulin

Insulin stimulates protein synthesis by altering the intrinsic activity or binding properties of key translation initiation and elongation factors (eIFs and eEFs, respectively) as well as critical ribosomal proteins. Components of the translational machinery that are targets of insulin regulation include eIF2B, eIF4E, eEF1, eEF2, and the S6 ribosomal protein.¹⁵⁸ The eIF2B multi-subunit guanine nucleotide exchange factor for eIF2 is kept inactive via phosphorylation of the eIF2Bε subunit at Ser⁵³⁵ by glycogen synthase kinase-3 (GSK3).¹⁵⁹ Inhibition of GSK3 by insulin-stimulated AKT kinase activity results in dephosphorylation eIF2Bε, promoting the formation of eIF2^GTP that recruits the initiator methionyl-tRNA to the ribosome.^159–161 The insulin-stimulated activation of eIF2B leads to an overall increase in translation initiation.¹⁶²

The eIF4F complex—including eIF4A/4G/4E and other proteins—is required for cap-dependent translation initiation. The mRNA cap-binding protein, eIF4E, is inactive during association with 4E-BP1 (see Fig. 8-5). Insulin activates eIF4E by stimulating mTORC1-mediated phosphorylation of 4E-BP1. Phosphorylated 4E-BP1 dissociates to facilitate the interaction between eIF4E and eIF4G, the scaffold protein for the eIF4F complex.¹⁶³ Mnk, an insulin-stimulated kinase activated through the Ras/ERK cascade, also resides in the eIF4F complex where it phosphorylates eIF4E at Ser²⁰⁹ (see Fig. 8-5).¹⁶⁴^,¹⁶⁵ Phosphorylation of eIF4E increases the binding affinity for mRNA caps, enhancing translation initiation. The phosphorylation of 4E-BP1 has an important systemic role in insulin action, insofar as deletion of the 4E-BP1 gene dramatically decreases white adipose tissue depots while increasing insulin sensitivity.¹⁶⁶

The effect of insulin-stimulated phosphorylation of eEF1 on translation elongation remains incompletely clear.¹⁵⁸ An elongation factor critical for ribosomal translocation along the mRNA, eEF2, is inactive when phosphorylated at Thr⁵⁶ by the eEF2 kinase (eEF2K).¹⁶⁷^,¹⁶⁸ Insulin stimulates the dephosphorylation of eEF2 via a rapamycin-sensitive route potentially involving the phosphorylation and inactivation of eEF2K by p70^S6K.167 In vitro, p70^S6K phosphorylates eEF2K at Ser³⁶⁶, a modification that greatly reduces the activity of the kinase.¹⁶⁷ Insulin-stimulated phosphorylation of the ribosomal S6 protein by p70^S6K may promote elongation of specific mRNAs corresponding to components of the translational machinery.¹⁶⁹ The atypical protein kinase C isoforms, PKCζ and PKCλ, also promote insulin-stimulated protein synthesis, although their downstream effectors remain undetermined.¹⁷⁰ The IRS1→PI3K appears to represent a bifurcation in the insulin signaling pathway to protein synthesis: one branch leads through PKCλ/ζ to general protein synthesis and the other through Akt and the target of rapamycin (mTOR) to growth-regulated protein synthesis and cell-cycle progression. Convergence between insulin and nutritional signaling occurs at the level of mTORC1 and mTORC2 regulation.^169,171,172

Insulin-Regulated Glucose Transport

Glucose transport is the prototype insulin response. Insulin stimulates glucose influx into adipose and cardiac muscle and into skeletal muscle, where it is the major mechanism for disposal of exogenous glucose. However, the molecular mechanisms linking the insulin signal to increased glucose influx have been challenging to resolve.¹⁷³ Cells express one or more members of the glucose transporter family, including 12 isoforms organized into three classes. Class I includes the well-characterized GLUT1, GLUT2, GLUT3, and GLUT4 transporters; class II includes GLUT5, GLUT7, GLUT9, and GLUT11; and class III includes GLUT6, GLUT8, GLUT10, and GLUT12 (reviewed in Ref. 174).

GLUT1 is expressed in most cells and tissues and resides permanently on the plasma membrane, where it constitutively transports glucose from the extracellular space into the cell. Although insulin does not stimulate translocation of GLUT1, chronic insulin treatment can increase the expression of cellular GLUT1 via the p21^ras→Erk kinase pathway (see Fig. 8-5).^175,176 GLUT2 is mainly expressed in liver and pancreatic β cells, where its relatively low affinity and high transport capacity provides a constant flux of glucose into these organs at physiologic plasma glucose concentrations. In the β cell, the uptake of glucose through GLUT2 is the first step in the detection of circulating glucose concentrations needed to stimulate insulin secretion. GLUT3 has a relatively high affinity for glucose and is most abundant in the central nervous system, where glucose concentrations are lower than in the bloodstream. GLUT8 might be insulin sensitive, since it contains GLUT4-like intracellular targeting motifs and is widely expressed in many insulin-responsive (as well as insulin-independent) tissues¹⁷⁷^,¹⁷⁸; GLUT12 is expressed in prostate, small intestine, placenta, skeletal muscle, and adipocytes and might also be insulin responsive.¹⁷⁹^,¹⁸⁰

GLUT4 is the principle insulin-responsive glucose transporter. GLUT4 is glycosylated in its first exofacial loop and contains a phosphorylation site (Ser₄₈₈) and a di-leucine motif in its COOH-terminus that plays a role in endocytosis.¹⁸¹ Upon synthesis, GLUT4 is unavailable to transport glucose across the plasma membrane because it enters a continuously recycling pathway that concentrates it in intracellular membranes of unstimulated cells.¹⁷³ During insulin stimulation, GLUT4-containing vesicles move from their intracellular sequestration compartment to the cell surface through targeted exocytosis (Fig. 8-7). Simultaneously, GLUT4 endocytosis is repressed, leading to an overall increase in glucose uptake.¹⁸² It is not clear whether the effect of insulin to stimulate glucose uptake in muscle and fat can be accounted for in its entirety by translocation; other factors that increase transporter activity could be involved.

Buy Membership for Endocrinology, Diabetes and Metabolism Category to continue reading. Learn more here