Discovery and isolation of insulin from pancreatic extract in 1922 (Banting et al. 1922) offered life-saving opportunities to diabetes mellitus patients and sparked active research on its mechanism of action. It was not until 1949 that Levine et al. (1949) identified significant reduction of blood sugar following intravenous administration of insulin to dogs, suggesting that insulin facilitates transfer of hexoses across cell membranes. This raised “the question of mechanism by which a molecule of the insulin type may affect the transfer of substances” (Levine et al. 1949). The idea of insulin receptor (IR) was hovering in the air for over two decades. In 1970, binding of radiolabeled insulin to rat liver plasma membranes was demonstrated (House and Weidemann 1970). Soon after, saturable, reversible, and high-affinity (KD = 7 × 10−11 M to 3 × 10−9 M) binding of insulin to membranes of hepatocytes and adipocytes was documented (Freychet et al. 1971; Cuatrecasas et al. 1971; Gammeltoft and Gliemann 1973), pointing to the existence of IR and triggering an avalanche of studies aimed at characterization of this intriguing molecule. IR was shown to be a receptor tyrosine kinase (RTK) that undergoes insulin-induced autophosphorylation at multiple sites (Kasuga et al. 1982). Human INSR gene (22 exons and 21 introns) was shown to be located on chromosome 19, and its cDNA was cloned and sequenced in 1985 (Ebina et al. 1985; Ullrich et al. 1985). IR preproreceptor was predicted to comprise 1370 amino acids (Ullrich et al. 1985) or 1382 amino acids (Ebina et al. 1985). This difference reflected the existence of two IR isoforms resulting from alternative splicing of exon 11: Isoform Long, or IR-B, and Isoform Short, or IR-A (UniProt accession numbers P06213-1 and P06213-2, respectively). Cleavage of the proreceptor by furin and glycosylation during maturation were documented subsequently. Further studies characterized the IR molecule in great detail, identifying the 27-residue signal sequence, the N-terminal extracellular α-subunit, and the C-terminal membrane-spanning β-subunit with the kinase domain and phosphorylation sites, the RKRR furin cleavage site, the α-β and the intermolecular α-α disulfide bonds, as well as similarities with other RTKs such as the epidermal growth factor receptor. The first IR substrate protein (IRS1) to be tyrosine phosphorylated by activated IR was discovered (White et al. 1985) and was later cloned and shown to mediate phosphatidylinositol 3-kinase pathway of cell signaling, thus providing critical insight into the molecular machinery by which insulin regulates cellular metabolism and growth.
IR belongs to the protein kinase superfamily, RTK family, and IR subfamily. The other two members of the latter are the insulin-like growth factor 1 receptor (IGF1R) and the insulin receptor-related receptor (IRR). All three are expressed in vertebrates, except that IRR has not been found in teleost fish, whereas invertebrates have just one IR homolog, Daf-2. IR is produced as a single-chain preproreceptor that sheds the signal sequence co-translationally and undergoes calnexin- and calreticulin-assisted folding in the endoplasmic reticulum (ER) and posttranslational modifications as the protein travels from ER to the Golgi complex and to the plasma membrane as a mature (αβ)2 homodimer. What makes the IR family members special is that they form covalent (disulfide-bridged) homodimers that stay in the basal, inactive state until hormone binding and activation, as opposed to other RTK’s that undergo noncovalent dimerization and activation upon hormone binding. (The author apologizes for skipping important references (available upon request); here the number of references is strictly limited.)
IR is N-glycosylated at 18 or 19 asparagines and O-glycosylated at six threonines and serines of the ectodomain, i.e., the α-subunit and the extracellular part of the β-subunit. Most N-glycans are mannose rich, and some are complex oligosaccharides containing galactose, fucose, N-acetylglucosamine, and sialic acid. Glycosylation facilitates protein processing, folding, and trafficking.
The structures of various fragments of the cytoplasmic TK domain of IR have been determined by X-ray crystallography and suggested important conformational changes in the β-chain upon receptor phosphorylation (Hubbard et al. 1994; Hubbard 1997). Studies on a construct that involved the juxtamembrane region in addition to the tris-phosphorylated TK domain uncovered noncovalent dimerization between the JM domain of one protomer and the N-lobe of the TK domain of the other (Cabail et al. 2015). The dimeric structure was speculated to play a key role in receptor activation. According to the proposed mechanism, in the basal, unphosphorylated state, the JM-N-lobe interaction occurs within same hemireceptor protomers, thus keeping the receptor in an inactive conformation. Receptor phosphorylation and accompanying conformational changes lead to cross dimerization and structural rearrangements, facilitating ATP binding and phosphorylation of substrate proteins (Cabail et al. 2015).
Many key questions regarding the structural basis for IR activation still remain unanswered. It is not clear what prevents the kinase activity of the receptor in the basal state, spatial dissociation of the TK domains, or formation of a nonproductive complex (Cabail et al. 2015). The fundamental questions of what conformational changes occur upon insulin binding and how these changes trigger receptor autophosphorylation are awaiting clarification (De Meyts 2015).
Insulin Binding and Receptor Activation
Both A and B isoforms of IR bind insulin as well as insulin-like growth factors 1 and 2, IGF1 and IGF2. The difference between the two isoforms in terms of ligand-binding properties is that the IR-A (exon 11–) isoform binds IGF2 with much higher affinity than IR-B (exon 11+) isoform and can be involved in cancer-related mitogenic activity (Belfiore et al. 2009; Vigneri et al. 2016). Both isoforms bind IGF1 with relatively low affinity (Hale and Coward 2013). In addition to IR and IGF1R, hybrid receptors, composed of one αβ hemireceptor of IR and one of IGF1R, are produced in all tissues that express both receptors. They can be activated by all three hormones, i.e., insulin, IGF1, and IGF2, but their physiological importance remains to be characterized.
Structural data indicate that IR in the basal state does not provide an open hormone binding pocket (Fig. 2), suggesting that the receptor needs to undergo conformational rearrangements to accommodate high-affinity hormone binding. Crystallographic studies identified that insulin interacts with the αCT segment of IR, which is repositioned against the L1 domain of same protomer, and opens up in a way to allow the C-terminal segment of its B chain to interact with L1 domain of the other IR protomer (Menting et al. 2013, 2014). Thus, both the hormone and the receptor undergo significant conformational changes upon interaction, leading to an “induced fit” and “detachment” model for hormone binding to IR (De Meyts 2015).
Based on the crystal structures of IR and IGF1R ectodomains with and without the bound hormone, Kavran et al. (2014) proposed that the ligand-free, inactive state of these receptors is stabilized by tight interaction between the L1 domain of one protomer and the FnIII-2 and FnIII-3 domains of the other, holding the ∧-shaped ectodomain in a wide open conformation. Engagement of the L1 domain by the hormone frees the FnIII-2/FnIII-3 module, allowing it to swing toward the central axis of the receptor, thereby moving the downstream TM, JM, and TK domains closer to each other and facilitating trans-autophosphorylation (Kavran et al. 2014), as schematically shown in Fig. 3c. This receptor activation mechanism has not been unequivocally supported by structural data, however, because the hormone-bound structure was determined for a construct lacking the FnIII-2/FnIII-3 segments (Menting et al. 2013). Therefore it should be considered hypothetical rather than an established mechanism.
Physiological Function of IR
In mammals, the IR-A isoform is expressed ubiquitously, including nearly all organs, and is upregulated during prenatal development (Belfiore et al. 2009), and IR-B is mainly expressed in the skeletal muscle, liver, adipose tissue, placenta, and kidneys (Hale and Coward 2013). In humans, IR-B expression (at mRNA level) is 85–90% of total IR in the liver, 78% in skeletal muscle, up to 70% in adipose tissue, and 45–50% in placenta and is undetectable in lymphocytes (Benecke et al. 1992). The level of expression of IR in different tissues and organs varies in a wide range. For example, erythrocytes contain only around 40 IR molecules per cell, whereas more than 200,000 IR molecules are present on adipocytes and hepatocytes (White and Kahn 1993).
The proteins that immediately interact with IR following its autophosphorylation are insulin receptor substrate (IRS) molecules. They act as adapter proteins that provide communications between cell signaling components and mediate pathways of normal physiology as well as pathology. IRS1 binds to phosphorylated Tyr972 (IR-B numbering) in the JM domain in a sequence (NPXY)-specific manner (X is any residue). Tyrosine phosphorylation at multiple sites of IRS1makes it an interaction partner for more than a dozen signaling molecules and thereby triggers a downstream polykinase cascade that includes processes ranging from glucose intake and glycogen synthesis to transcription regulation and mitogenesis. Phosphorylated Tyr972 recruits other proteins as well, such as the mitogenic protein SHC1, the transcription factor STAT5B, and the suppressor of cytokine-3 (SOCS3) that is involved in leptin and insulin resistance (http://www.genecards.org/cgi-bin/carddisp.pl?gene=INSR). Phosphorylated tyrosines 1158, 1162, and 1163 in the TK domain bind cytokine suppressor SOCS1 protein and MAPK3 inhibitor adapter protein GRB14. Tyrosine phosphorylation in the CT domain recruits IRS2. Thus, the events downstream to IR autophosphorylation can be divided into two major branches, metabolic, or Akt pathway, and mitogenic, or MAPK pathway.
IR primarily performs a metabolic function and thereby controls nutrient homeostasis, cellular metabolism and growth. High affinity of IR-A for both insulin and IGF2 is important for prenatal growth through the IGF2 pathway and postnatal metabolism through the insulin pathway. These patterns can vary among mammals, however.
The metabolic, or Akt pathway itself, has several components, e.g., glucose import, regulation of glycogen synthesis, gluconeogenesis, and protein synthesis. In this pathway, phosphorylated IRS1 binds to the regulatory subunit of the phosphatidylinositol-3 kinase (PI-3K), which then phosphorylates phosphatidylinositol 4,5-bisphosphate (PIP2) to phosphatidylinositol 3,4,5-trisphosphate (PIP3) in the cytoplasmic face of the cell plasma membrane. PIP3 recruits the PH domain-containing proteins, pyruvate dehydrogenase lipoamide kinase isoform 1 (PDK1) and Akt2. PDK1 phosphorylates and activates PIP3-anchored Akt2, which mediates plasma membrane translocation of the glucose transporter GLUT4 by phosphorylation of C2 domain-containing protein CDP138, followed by glucose influx (Xie et al. 2011). Another function of Akt2 is phosphorylation of glycogen synthase kinase 3 (GSK3), leading to regulation of glycogen synthesis in the liver (Fig. 4). Akt2 also mediates adipogenesis, lipolysis, and triglyceride synthesis in adipocytes (Shearin et al. 2016). Phosphorylation of the forkhead box O (FoxO) family transcription factors by Akt regulates glucose intake, glucose synthesis in hepatocytes, and protein synthesis in skeletal muscle (Martins et al. 2016). Thus, Akt plays a central role in the metabolic branch of IR signaling.
The mitogenic, MAPK pathway, can be triggered through both IRS1 and SHC1 mechanisms of IR signaling. In either case, phosphorylated IRS1 or SHC1 binds to and phosphorylates growth factor receptor-bound protein 2 (GRB2). GRB2 engages the son of sevenless (SOS) protein, which in turn activates the Ras protein and thereby sets in motion the Ras-Raf-MAPK cascade. This results in translocation into the nucleus of MAPK and regulation of various transcription factors, leading to gene expression, mitogenesis, cell survival, and proliferation (Fig. 4) (Vigneri et al. 2016). The mitogenic pathway is usually mobilized by growth factors and their receptors but can be hijacked by IR, depending on many factors, including hormone availability and concentration, receptor expression levels and functional state, and duration of hormone exposure.
Role of IR in Disease
Defects in insulin signaling pathways, coupled with impaired insulin secretion by the β cells, lead to hyperglycemia and insulin resistance, i.e., type 2 diabetes mellitus (T2DM). T2DM constitutes approximately 90% of all diabetes cases and affects around 400 million people worldwide. Diabetes is associated with many pathologies including, but not limited to, metabolic syndrome, obesity, chronic inflammation, dyslipidemia, atherosclerosis, hypertension, nephropathy, retinopathy and blindness, coronary heart disease, stroke, Alzheimer’s disease, and cancer. A fraction of T2DM is caused by mutations or dysfunction of proteins in the insulin signaling cascade, including missense mutations in the hormone binding or the TK domains of IR (Hubbard et al. 1994; Sesti et al. 2001; LeRoith and Accili 2008; Menting et al. 2013). UniProt reports 56 naturally occurring mutations in human IR associated with insulin resistance through defects in receptor processing and trafficking, hormone binding, or the catalytic kinase function of IR.
Adipocyte-specific IR knockout mice showed adipose and hepatic abnormalities and had significantly shortened lifespan (Friesen et al. 2016), whereas induction of hepatic expression of IR, especially the A isoform, in diabetic mice ameliorated glucose intolerance (Diaz-Castroverde et al. 2016). Total prevention of IR expression in humans resulted in severe growth retardation and diabetes, while homozygous IR-null mice or those expressing kinase-deficient IR rapidly developed hyperglycemia and did not survive beyond a few days of postnatal life (Accili 1995; Accili et al. 1996; Lauro et al. 1999). These data underscore the importance of IR in preventing diabetes and improving survival and life expectancy (LeRoith and Accili 2008).
Apart from genetic factors, T2DM can arise from sedentary lifestyle coupled with high-calorie diet. These factors cause obesity, which may be associated with chronic inflammation and T2DM. The latter in turn can cause endothelial dysfunction and defective vascular homeostasis through altered gene expression and cell signaling in vascular endothelium, leading to atherosclerosis, the major life-shortening factor for diabetic patients (Rask-Madsen and King 2013). High affinity of IR-A for IGF2 contributes to atherogenic processes and vascular damage. IGF2 and TNF-α induce overexpression of IR-A and IGF1R and increase the level of IR-A/IGF1R hybrid receptors in vascular smooth muscle cells, which augment atherosclerosis (Gómez-Hernández et al. 2013).
Risk of cardiovascular diseases such as myocardial infarction, heart failure, and ischemic stroke is significantly higher among patients with metabolic syndrome and/or T2DM. Diabetes-related cardiomyopathy has been documented in numerous experimental models using diabetic rodents (Fuentes-Antrás et al. 2015). These adverse effects are related to both proximal defects in the insulin signaling pathways and distal events such as transcription regulation by FoxO1 in the liver resulting in reduced levels of plasma adiponectin and high-density lipoprotein as well as overproduction of very low-density lipoprotein.
Insulin and IR are present in the brain where they support glucose metabolism and downregulate neuronal apoptosis and thereby contribute to neuronal functions such as cognition, memory, and learning. Insulin resistance and hyperinsulinemia are risk factors for Alzheimer’s disease (AD) and dementia possibly through promoting amyloid β (Aβ) peptide deposition (Willette et al. 2015), as well as reduced Aβ degradation and elevated tau protein phosphorylation (Morales-Corraliza et al. 2016). Insulin therapy could improve the cognitive function, but systemic (e.g., intravenous) insulin administration may cause adverse effects such as hypoglycemia. Intranasal administration of insulin to AD patients, which ends up mostly in the brain, indeed significantly ameliorates the memory and cognition of the patients and is associated with increased cerebrospinal Aβ levels and decreased tau-to-Aβ ratios (Craft et al. 2012). (Higher cerebrospinal Aβ means lower brain deposits.) Aβ oligomers were more neurotoxic in the absence than in the presence of insulin, suggesting that direct effect of insulin on the Aβ peptide may play a role in the correlation between T2DM and AD (Luo et al. 2016).
Strong link has been established between aberrant IR signaling and cancer. T2DM and hyperinsulinemia lead to activation of IR-A in tumor cells and promote various cancers (Vigneri et al. 2016). Owing to both its abundance in neoplasm and its high affinity for IGF2, IR-A favors mitogenesis and thus contributes to tumor growth and malignancy. In addition, overexpression of IR-A increases the level of IR-A/IGF1R hybrid receptors, which exert mitogenic effect and support cancer growth through both insulin and IGF pathways (Belfiore et al. 2009; Lundby et al. 2015). Anticancer drugs targeting growth hormone receptors, including IGF1R, have failed or proved ineffective because of a bypass route for the mitogenic activity of IGF2 through the IR-A pathway (Singh et al. 2014; Chan et al. 2016). Anti-IGF1R immunotherapy in transgenic mice with pancreatic carcinogenesis did not significantly affect tumor growth, but the same therapy in IR knockout mice resulted in strong reduction in tumor burden, increased apoptosis, and recovered sensitivity to anti-IGF1R therapy (Ulanet et al. 2010). Likewise, antibodies or small-molecule drugs aimed at IGF1R alone or IR-A alone showed little benefit in breast cancer in humans and mice, whereas combined inhibition of both IGF1R and IR-A was more effective (Rostoker et al. 2013; Chan et al. 2016). Dual IGF1R/IR therapy has also been more successful in treatment of prostate, lung, and colorectal cancers compared to targeted inhibition of IGF1R (Dayyani et al. 2012; Vincent et al. 2013; Vidal et al. 2015; Puzanov et al. 2015; Bendell et al. 2015).
IR is a special RTK that forms a disulfide-linked (αβ)2 type homodimer of two αβ hemireceptors. Insulin is the primary ligand of IR, and insulin binding to IR triggers a signaling cascade necessary for glucose uptake and metabolism, regulation of glycogen synthesis, gluconeogenesis, protein synthesis, lipolysis, and other metabolic functions. IR-A also binds IGF2 with high affinity, which contributes to mobilization of mitogenic pathways of IR signaling, often leading to various pathologies.
Unlike other RTKs, IR and the other two members of the IR family receptors (IGF1R and IRR) are covalent dimers. Receptor activation takes place through hormone-induced conformational changes in the ectodomain that are transmitted to the cytoplasmic TK domains, resulting in receptor trans-autophosphorylation and activation (IRR is less extensively studied in this respect). Much has been elucidated about the structural basis of insulin-induced transmembrane signaling of IR, primarily by X-ray crystal structures of the ectodomain and the TK domain of the receptor. The structure of the full-length IR has not been achieved yet, leaving important questions about the IR signaling mechanism unanswered.
There is strong evidence that IR, especially its A isoform, is involved in progression of a variety of diseases, ranging from T2DM to cardiovascular disease, atherosclerosis, heart failure, stroke, Alzheimer’s disease, and cancer. Similarities between the growth hormone receptor IGF1R and IR pose serious problems in cancer therapy targeting the former because IR-A hijacks the IGF2 pathway of cell proliferation. Drugs designed for dual inhibition of both receptors open promising prospects for treatment of certain cases of cancer.
- Bendell JC, Jones SF, Hart L, Spigel DR, Lane CM, Earwood C, Infante JR, Barton J, Burris HA. A phase Ib study of linsitinib (OSI-906), a dual inhibitor of IGF-1R and IR tyrosine kinase, in combination with everolimus as treatment for patients with refractory metastatic colorectal cancer. Investig New Drugs. 2015;33(1):187–93.CrossRefGoogle Scholar
- Craft S, Baker LD, Montine TJ, Minoshima S, Watson GS, Claxton A, Arbuckle M, Callaghan M, Tsai E, Plymate SR, Green PS, Leverenz J, Cross D, Gerton B. Intranasal insulin therapy for Alzheimer disease and amnestic mild cognitive impairment: a pilot clinical trial. Arch Neurol. 2012;69(1):29–38.PubMedCrossRefGoogle Scholar
- Dayyani F, Parikh NU, Varkaris AS, Song JH, Moorthy S, Chatterji T, Maity SN, Wolfe AR, Carboni JM, Gottardis MM, Logothetis CJ, Gallick GE. Combined inhibition of IGF-1R/IR and Src family kinases enhances antitumor effects in prostate cancer by decreasing activated survival pathways. PLoS ONE. 2012;7(12):e51189.PubMedCrossRefPubMedCentralGoogle Scholar
- Diaz-Castroverde S, Gómez-Hernández A, Fernández S, García-Gómez G, Di Scala M, González-Aseguinolaza G, Fernández-Millán E, González-Rodríguez Á, García-Bravo M, Chambon P, Álvarez C, Perdomo L, Beneit N, Escribano O, Benito M. Insulin receptor isoform A ameliorates long term glucose intolerance in diabetic mice. Dis Model Mech 2016;9(11):1271–81.Google Scholar
- Gómez-Hernández A, Escribano Ó, Perdomo L, Otero YF, García-Gómez G, Fernández S, Beneit N, Benito M. Implication of insulin receptor A isoform and IRA/IGF-IR hybrid receptors in the aortic vascular smooth muscle cell proliferation: role of TNF-α and IGF-II. Endocrinology. 2013;154(7):2352–64.PubMedCrossRefGoogle Scholar
- Lundby A, Bolvig P, Hegelund AC, Hansen BF, Worm J, Lützen A, Billestrup N, Bonnesen C, Oleksiewicz MB. Surface-expressed insulin receptors as well as IGF-I receptors both contribute to the mitogenic effects of human insulin and its analogues. J Appl Toxicol. 2015;35(7):842–50.PubMedCrossRefGoogle Scholar
- Menting JG, Whittaker J, Margetts MB, Whittaker LJ, Kong GK, Smith BJ, Watson CJ, Záková L, Kletvíková E, Jiráček J, Chan SJ, Steiner DF, Dodson GG, Brzozowski AM, Weiss MA, Ward CW, Lawrence MC. How insulin engages its primary binding site on the insulin receptor. Nature. 2013;493(7431):241–5.PubMedCrossRefPubMedCentralGoogle Scholar
- Menting JG, Yang Y, Chan SJ, Phillips NB, Smith BJ, Whittaker J, Wickramasinghe NP, Whittaker LJ, Pandyarajan V, Wan ZL, Yadav SP, Carroll JM, Strokes N, Roberts Jr CT, Ismail-Beigi F, Milewski W, Steiner DF, Chauhan VS, Ward CW, Weiss MA, Lawrence MC. Protective hinge in insulin opens to enable its receptor engagement. Proc Natl Acad Sci U S A. 2014;111(33):E3395–404.PubMedCrossRefPubMedCentralGoogle Scholar
- Morales-Corraliza J, Wong H, Mazzella MJ, Che S, Lee SH, Petkova E, Wagner JD, Hemby SE, Ginsberg SD, Mathews PM. Brain-wide insulin resistance, tau phosphorylation changes, and hippocampal neprilysin and amyloid-β alterations in a monkey model of type 1 diabetes. J Neurosci. 2016;36(15):4248–58.PubMedCrossRefPubMedCentralGoogle Scholar
- Puzanov I, Lindsay CR, Goff L, Sosman J, Gilbert J, Berlin J, Poondru S, Simantov R, Gedrich R, Stephens A, Chan E, Evans TR. A phase I study of continuous oral dosing of OSI-906, a dual inhibitor of insulin-like growth factor-1 and insulin receptors, in patients with advanced solid tumors. Clin Cancer Res. 2015;21(4):701–11.PubMedCrossRefGoogle Scholar
- Ullrich A, Bell JR, Chen EY, Herrera R, Petruzzelli LM, Dull TJ, Gray A, Coussens L, Liao YC, Tsubokawa M, Mason A, Seeburg PH, Grunfeld C, Rosen OM, Ramachandran J. Human insulin receptor and its relationship to the tyrosine kinase family of oncogenes. Nature. 1985;313(6005):756–61.PubMedCrossRefGoogle Scholar
- Vidal SJ, Rodriguez-Bravo V, Quinn SA, Rodriguez-Barrueco R, Lujambio A, Williams E, Sun X, de la Iglesia-Vicente J, Lee A, Readhead B, Chen X, Galsky M, Esteve B, Petrylak DP, Dudley JT, Rabadan R, Silva JM, Hoshida Y, Lowe SW, Cordon-Cardo C, Domingo-Domenech J. A targetable GATA2-IGF2 axis confers aggressiveness in lethal prostate cancer. Cancer Cell. 2015;27(2):223–39.PubMedCrossRefPubMedCentralGoogle Scholar
- Vigneri R, Goldfine ID, Frittitta L. Insulin, insulin receptors, and cancer. J Endocrinol Investig 2016;39(12):1365–76.Google Scholar
- White MF, Kahn CR. Ch. 2, Mechanism of insulin action. In: Moller DE, editor. Insulin resistance. Chichester: Wiley; 1993. p. 9–48.Google Scholar
- Willette AA, Johnson SC, Birdsill AC, Sager MA, Christian B, Baker LD, Craft S, Oh J, Statz E, Hermann BP, Jonaitis EM, Koscik RL, La Rue A, Asthana S, Bendlin BB. Insulin resistance predicts brain amyloid deposition in late middle-aged adults. Alzheimers Dement. 2015;11(5):504–10.PubMedCrossRefGoogle Scholar
- Xie X, Gong Z, Mansuy-Aubert V, Zhou QL, Tatulian SA, Sehrt D, Gnad F, Brill LM, Motamedchaboki K, Chen Y, Czech MP, Mann M, Krüger M, Jiang ZY. C2 domain-containing phosphoprotein CDP138 regulates GLUT4 insertion into the plasma membrane. Cell Metab. 2011;14(3):378–89.PubMedCrossRefPubMedCentralGoogle Scholar