Insulin Receptor Signaling and Sarco/Endoplasmic Reticulum Calcium ATPase in β-Cells

The insulin-secreting β-cell of the endocrine pancreas has a central role in regulating glucose homeostasis (1). Glucose oxidation by the β-cell is essential for insulin secretion. In particular, glucokinase, the first step in glycolysis, has been convincingly shown to be the β-cell glucose sensor (2). β-Cell metabolism of glucose results in an increase in the ATP-to-ADP ratio, leading to closure of the ATP-sensitive K⁺ channel, depolarization of the β-cell, and influx of extracellular Ca²⁺ through voltage-dependent Ca²⁺ channels (3). The subsequent increase in intracellular Ca²⁺ then activates insulin exocytosis (Fig. 1). Most likely, a number of protein kinases, including the calcium and calmodulin-dependent protein kinase as well as protein kinase C, are involved in these distal steps (4).

The importance of the β-cell in the pathogenesis of type 2 diabetes is now clearly recognized (5). Type 2 diabetes is characterized by 1) insulin resistance in target tissues of insulin, such as skeletal muscle and adipocytes for glucose uptake and liver for glucose production, and 2) impaired insulin secretion (6). Peripheral tissue and hepatocyte insulin resistance lead to reduced glucose uptake and decreased inhibition of hepatic glucose production. To compensate for insulin resistance and to maintain euglycemia, pancreatic β-cell insulin secretion increases significantly, resulting in hyperinsulinemia. Eventually, the β-cell fails, and overt disease appears. Type 2 diabetes is believed to be a polygenic disease that is also affected by various environmental factors such diet, physical activity, and age.

Insulin has a central role in regulating cell metabolism, gene expression, growth, and differentiation (7). Insulin binds the cell surface insulin receptor, which consists of two extracellular 135-kDa α-subunits that bind insulin and two intracellular 95-kDa β-subunits that have intrinsic tyrosine kinase activity. After the insulin binds to its receptor, there is autophosphorylation of tyrosyl residues on the β-subunits and phosphorylation of cellular proteins, including insulin receptor substrate (IRS)-1 and -2 (7). The subsequent cascade of signaling events results in translocation of the insulin-responsive glucose transporter GLUT4 in muscle and adipocytes.

In 1995, we first reported the presence of the insulin receptor in insulin-secreting β-cells and showed that glucose-induced insulin secretion activates the β-cell surface insulin receptor tyrosine kinase and its intracellular signaling transduction pathway (8,9). In the last few years, there has been a rapidly growing body of evidence confirming our original findings that the insulin receptor signaling pathway is active in pancreatic β-cells and plays an important role in β-cell regulation (5,8–15) (Table 1). Activation of the β-cell insulin receptor results in rapid tyrosine phosphorylation of the insulin receptor β-subunit and the IRS proteins (8). Deletion of insulin receptor results in neonatal death in mice (16) and leprechaunism in humans (17). Mice with heterozygous null alleles of insulin receptor and IRS-1 (IR/ IRS-1^+/−) exhibit hyperinsulinemia and β-cell hyperplasia, and they develop overt diabetes (18). Knockouts of IRS-1 and -2 produce different effects. Inactivation of IRS-1 (IRS-1^−/−) leads to mild insulin resistance, hyperinsulinemia, and β-cell hyperplasia with no overt diabetes syndrome (5,13). In contrast, inactivation of IRS-2 (IRS-2^−/−) results in β-cell failure and causes type 2 diabetes (13,14). The differential effects of IRS-1 and -2 knockout indicate the two major IRSs mediate differential signals in β-cells. However, the mechanisms accounting for such differential regulation and for IRS-1 function in the β-cell are still unknown.

Our prior studies (Table 2) have shown that in β-cells stably overexpressing twofold IRS-1, Ca²⁺ homeostasis was perturbed, with an increase in cytosolic Ca²⁺ and increased insulin secretion (19). Further investigation revealed that the increase in cytosolic Ca²⁺ was due to inhibition of sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA), a protein responsible for Ca²⁺ uptake into the endoplasmic reticulum (ER) lumen. When the control β-cell line was treated with thapsigargin, a SERCA inhibitor, the cytosolic Ca²⁺ levels increased to the same level as the β-IRS1 cell line. Furthermore, ER calcium uptake in a digitonin-permeabilized cell system was reduced in the β-IRS1 cells compared with controls. This increase in cytosolic Ca²⁺ was also seen in a cell line overexpressing the insulin receptor, but not in a cell line overexpressing a kinase-deficient mutant of the insulin receptor (11). We thus postulated the existence of a novel autocrine loop of insulin on insulin secretion from the β-cell, mediated by activation of the insulin receptor signaling pathway and IRS-1. In isolated mouse islets, insulin treatment of islets caused an increase in intracellular Ca²⁺ within minutes, with concomitant insulin exocytosis (20–22). In addition, studies with a β-cell-specific insulin receptor knockout mouse and IRS-1-deficient β-cell lines show that glucose-induced insulin secretion is lost under those conditions (15,23). Thus, there is a convergence of data obtained in animal and cellular models that demonstrate the existence of a positive feedback loop of insulin on glucose-induced insulin secretion that is mediated by an interaction of the insulin signal transduction pathway with the Ca²⁺-homeostatic mechanisms of the β-cell.

Cellular Ca²⁺ is a critical element in β-cell function (1,24,25). Abnormal intracellular concentration of Ca²⁺ ([Ca²⁺]_i) is a common defect in both type 1 and type 2 diabetes (26). Altered Ca²⁺ metabolism has also been reported to affect β-cell function, including insulin biosynthesis (27). The ER plays an important role in the regulation of [Ca²⁺]_i (24). Typically, basal Ca²⁺ concentrations are in the 80–100 nmol/l range, and after glucose stimulation they increase three- to fivefold. Once the stimulation is removed, the ER sequesters excess cytosolic Ca²⁺, and the Ca²⁺ levels return to baseline. SERCA is the major calcium pump that sequestrates cytosolic Ca²⁺ into ER lumen. Its biochemical characteristics have been extensively described, and its implication in the regulation of intracellular Ca²⁺ homeostasis is recognized. Thapsigargin, a nonphorboid tumor promoter, specifically inhibits ER Ca²⁺-ATPase activity and leads to elevated cytosolic Ca²⁺ in β-cells and enhanced short-term glucose-stimulated insulin secretion (28).

SERCA2 and -3 have been found in β-cells (28–31). Both SERCA2 and -3 gene transcripts have splicing variants resulting in distinct protein isoforms. SERCA2a is the muscle isoform and is expressed in slow-twitch skeletal muscle. SERCA2b differs from SERCA2a only in the COOH-terminal part and is widely expressed in nonmuscle tissues (32). SERCA3 is coexpressed with SERCA2b in nonmuscle tissues and has three splicing isoforms: SERCA3a, -b, and -c (33). SERCA3 has a lower apparent affinity for Ca²⁺ than other members of the SERCA family (34). SERCA3b and -c have even lower apparent affinities for Ca²⁺ than SERCA3a (33). Recent studies show that loss of SERCA activities and reduction of SERCA3 gene expression in β-cells are associated with diabetes in db/db mice (30) and GK rats (35). Finally, Varadi et al. (36) have identified four rare missense mutations of the SERCA3 gene in type 2 diabetic patients recruited for the U.K. Prospective Diabetes Study: Gln108→His, Val648→Met, Arg674→Cys, and Ile753→Leu, suggesting that the SERCA3 gene locus contributes to genetic susceptibility to type 2 diabetes.

IRSs may directly interact with ER Ca²⁺-ATPase (SERCA1 and -2) in a tyrosine phosphorylation-dependent manner in muscle and heart (37). This finding suggests that insulin may regulate ER Ca²⁺-ATPase activity, therefore influencing cellular Ca²⁺ homeostasis. To dissect the role of IRS-1 in β-cell function, we have overexpressed IRS-1 in an insulin-secreting β-cell line (Table 2). We showed that IRS-1 regulates β-cell Ca²⁺ homeostasis, insulin biosynthesis, and β-cell proliferation, and that elevated expression of IRS-1 induces abnormal Ca²⁺ homeostasis and β-cell dysfunction (19). We suggested a novel functional link between the IRS-1 signaling pathway and the stimulus-secretion pathway in β-cells, which we believe to be physiologically significant. Under basal conditions in the β-cell, this pathway is not activated. However, once glucose or other secretagogues stimulate insulin secretion, the released insulin will feedback to the β-cell insulin receptor and activate the associated signal transduction pathway. Increased signaling results in IRS-1 tyrosine phosphorylation and subsequent inhibition of ER Ca²⁺ uptake. Decreased Ca²⁺ fluxes into the ER can then increase cytosolic Ca²⁺ and further facilitate the maintenance of increased Ca²⁺ levels due to secretagogue-induced Ca²⁺ influx from the extracellular space (Fig. 2).

More recently, we have used three complementary techniques to show that IRS-1 colocalizes to intracellular membranes in pancreatic β-cell lines (P.D.B., B.A.W., unpublished observations). Subcellular fractionation of the βTC6-F7 and the β-IRS1 cell lines showed that IRS proteins localized to the intracellular membranes in the high-speed pellet. Furthermore, overexpressed IRS-1 in the β-IRS1 cell line was primarily in the intracellular membrane fraction. Colocalization of IRS-1 with the ER marker protein immunoglobulin binding protein (BiP)/glucose-regulated protein 78 kDa (GRP78) was demonstrated by confocal microscopy. Although the signals for the IRS proteins and ER marker clearly overlapped, there was some additional immunofluorescence staining in other portions of the cell. Finally, immunoelectron microscopy experiments showed that IRS-1 is present in ER-derived microsomal vesicles from β-cell lines, as confirmed by its coimmunostaining with BiP/GRP78 and SERCA3b. Taken together, these data present a strong case for IRS protein localization to intracellular membranes, in particular the ER, in pancreatic β-cell lines. The lack of complete localization of IRS-1 and -2 to the ER by confocal microscopy indicates that IRS proteins may distribute to several pools within the cell. Considering the dynamic nature of insulin receptor signaling, it is possible that IRS proteins translocate to several locations in the cell in response to upstream signals. However, the majority of both phosphorylated and nonphosphorylated IRS proteins are maintained in intracellular membranes. IRS proteins in this pool can interact with other proteins in the same location and potentially form large protein complexes that are insoluble in detergents. An example of this comes from one study reporting that IRS-1 can bind to the ς3A subunit of the AP-3 adaptor protein complex, resulting in intracellular membrane localization (38). The localization of IRS proteins to the ER increases the likelihood of a direct interaction with SERCA isoforms in β-cells. Indeed, IRS proteins have previously been shown to bind to several SERCA isoforms, including the β-cell isoform SERCA2b (37), and the pancreatic β-cell SERCA isoform SERCA3b coimmunoprecipitates with IRS-1 when expressed in a CHO T-cell line (P.D.B. and B.A.W., unpublished observations). Furthermore, this interaction is enhanced when IRS-1 is phosphorylated by insulin stimulation. Overexpression of IRS-1 in this system results in further enhancement of IRS-1 binding to SERCA3b, which could explain why IRS-1 overexpression in β-IRS1 cells inhibits SERCA function (19).

In contrast to our findings, a recent study demonstrates that insulin can cause β-cell hyperpolarization and diminish cytosolic calcium oscillations in mouse islets incubated in 10 mmol/l glucose via an increase in whole-cell K⁺ conductance (39). This effect could be mediated through a phosphatidylinositol 3-kinase-dependent pathway. However, in another study, mouse islets incubated with the insulin-mimetic compound L-783,281 had increased insulin release at 11 mmol/l glucose (40). There was no change in the pulsatile release of insulin under these conditions. Treatment with the insulin-mimetic at 3 mmol/l glucose decreased the pulse frequency of insulin release, but not the amount. Although Aspinwall et al. (21) showed that insulin-induced insulin exocytosis could occur at 3 mmol/l glucose, their experiments measured exocytosis events and not actual insulin release, which may have been minimal at that glucose concentration. Thus, although insulin can effect the electrophysiology of the β-cell in a negative fashion, it can also increase insulin release. Most likely, the effects of insulin depend on the extracellular glucose concentration.

Because of the importance of SERCA as a downstream effector of IRS-1, we have studied the effect of overexpressing SERCA isoforms in β-cells (J.M., Z.G., and B.A.W., unpublished observations). cDNAs encoding rat SERCA2b and mouse SERCA3b were cloned by PCR into a His-tagged destination vector using Gateway cloning technology (Life Technologies). The Gateway cloning system uses phage λ-based site-specific recombination instead of restriction endonucleases and ligases, and it was completed in a two-step reaction, with the initial step generating an entry clone. The template plasmids were obtained from Dr. J. Lytton (pMT2mSERCA3b; University of Calgary, Alberta, Canada) and F. Wuytack (pMT2rSERCA2b; Katholieke Universiteit Leuven, Leuven, Belgium). Amplification of the cDNAs was performed in a two-step PCR. PCR fragments were gel-purified and subcloned into a Gateway entry vector (pDONR201; Life Technologies) using BP Clonase enzyme mix and proteinase K solution. After transformation into library efficiency DH5α and grown on kanamycin plates, positive clones were sequenced. Plasmid DNAs from clones with correct nucleotide sequence were purified by Qiagen column chromatography. The final step in the cloning process was the creation of expression clones by combining the entry clones (pDONR.serca2b/pDONR.serca3b) with the destination vector pDEST26 containing an NH₂-terminal histidine fusion protein (Life Technologies). Correct clones were identified by sequencing and then amplified. The expression constructs were driven by a cytomegalovirus (CMV) promoter. β-TC6 insulinoma cells were transfected with plasmid DNAs using either FUGENE-6 transfection reagent (Roche Diagnostics, Indianapolis, IN) or by electroporation technique (Gene Pulser II RF Module; BioRad) and were selected for growth in the presence of 300 μg/ml Geneticin (Life Technologies, Grand Island, NY). Single colonies of the Geneticin-resistant cells were expanded and were tested for overexpression of SERCA2b and -3b by Western blotting, using anti-His and anti-SERCA2b and -3b peptide antibodies.

Sequencing information showed that we successfully generated the expression clones of pDT-SERCA2b and pDT-SERCA3b transgenes under the control of a CMV promoter. We then used the constructs to create stable expression in β-TC6 insulin-secreting cells and quantify the endogenous as well as the His-tagged SERCA protein levels. Analysis of SERCA2b cell lysates showed that the protein is expressed in the nontransfected wild type, in cells transfected with the empty vector (control), as well as in cells with the pDT-SERCA2b construct (Fig. 3A). However, further analysis showed that the level of expression was not significantly different in control cells and the cells stably transfected with pDT-SERCA2b. Furthermore, immunoprecipitation of the His-tagged SERCA2b protein using anti-His antibody and immunoblotting with either anti-His or anti-SERCA2b revealed lack of expression of the pDT-SERCA2b construct to any sufficient amount that can be detected using the protocol used (data not shown). Western blot analysis of analysis of SERCA3b also did not show any significant difference in expression levels between the neo controls and the stable cell lines (Fig. 3B, lanes 3–7). Furthermore, immunoprecipitation complex data showed that the SERCA3b protein was also not expressed to any significant extent in the stably transfected β-cells compared with the neo controls (Fig. 3C). To demonstrate that the constructs can be expressed at all, we transiently or stably transfected CHO cells (which do not express SERCA3b) with the pDT-SERCA3b plasmid. As shown in Fig. 3B (lanes 1 and 2) and Fig. 3C (lanes 1, 2, 5, and 6), SERCA3b was robustly overexpressed in CHO cells after transfection (Cho3b), whereas the protein was not detected in the control nontransfected cells. The fact that β-TC6 cells transfected with SERCA3b survived antibiotic selection (compared with nontransfected β-TC6 cells that died under the same conditions) suggests that these cells did in fact incorporate the SERCA DNA into their genome. However, the lack expression of the protein in these cells (as compared with CHO) may suggest a teleological tight regulatory control necessary for the maintenance of calcium homeostasis in the insulin-secreting cells.

To determine the effect of overexpression of the SERCA proteins on cytosolic Ca²⁺, CHO cells stably transfected with pDT-SERCA3b were plated on 25-mm glass coverslips. After 48 h, the cells were loaded with Fura 2-AM (acetoxymethylester) in Ca²⁺-free Krebs-Ringer bicarbonate (KRB) (no EGTA) in the presence or absence of 100 nmol/l insulin. After a 40-min incubation, cells were washed with prewarmed KRB and perifused on top for 10 min with KRB (G15) with or without insulin. We observed no difference between the SERCA3b-expressing cells and the vector controls. Thus, it appears that the contribution of SERCA3b to total calcium regulation in this system is minimal.

Synaptotagmins are a family of 11 isoforms of a Ca²⁺-mediated phospholipid binding protein originally identified in neuron secretory granules (41). Synaptotagmin III (Syt3) has a single transmembrane domain and two Ca²⁺ regulatory C2 domains that are thought to regulate membrane fusion and membrane budding reactions involved in exocytosis; Syt3 is sensitive to submicromolar levels of Ca²⁺. This and other evidence strongly suggests that Syt3 is one of, if not the, Ca²⁺ sensor in the exocytosis of insulin secretory granules (42). To further dissect the role of the insulin receptor signal transduction pathway in insulin exocytosis, we wished to develop a cell-biological technique to monitor movement and exocytosis of insulin secretory granules in real time, by creating a fusion protein of Syt3 and EGFP (M.T., Z.G., B.A.W., unpublished observations). The Syt3 cDNA was generously provided by Dr. Thomas Südhof (University of Texas Southwestern Medical Center). This DNA sequence was inserted into two plasmids that would produce fusion proteins with the EGFP tag at either the COOH terminus or the NH₂ terminus of the Syt3 protein. The vectors in which Syt3 DNA was inserted were pEGFP-C3 (Clontech, Palo Alto, CA) cut with Xma-1 restriction enzyme, resulting in the fusion protein referred to as Fsyt3, and pEGFP-N3 (Clontech) cut with EcoR-1 restriction enzyme, resulting in the fusion protein referred to as Syt3F. An pEGFP-C3 empty vector was used as control and referred to as pEGFP.

β-TC6 cells were stably transfected with the constructs using FuGene. Cells transfected with the Fsyt3 plasmid were referred to as Fsyt3 cells, those transfected with the Syt3F plasmid were referred to as Syt3F cells, and those transfected with empty vector were referred to as EGFP cells. Western blot analysis of the EGFP and Fsyt3 cell lines demonstrated overexpression of the constructs. Immunogold electron microscopy of the transfected cells demonstrated colocalization of Syt2 with insulin secretory granules. When viewed under a confocal microscope, there was a clear difference in the distribution of fluorescence in the EGFP cells as opposed to the Fsyt3 and Syt3F cells. The fluorescence within the EGFP cells is diffuse and spreads throughout the entire cell. In the cells expressing the fusion protein, however, the fluorescence is concentrated into dense clusters. Actual vesicular trafficking appeared to be a rather rare event, however. Movies of ∼60 cells were constructed during the course of this study, but only a handful of these movies clearly showed what appears to be vesicular trafficking. Much optimization, however, is needed to make this technique a useful and efficient tool.

In conclusion, we propose the existence of a novel positive-feedback pathway in which insulin can regulate insulin secretion in pancreatic β-cells. A key component of this pathway is interaction between IRS-1 and SERCA, which regulates intracellular Ca²⁺. IRS-1 is present in the ER and can directly bind to the β-cell isoforms of SERCA, specifically SERCA3b. Insulin stimulation results in increased binding of IRS-1 to SERCA3b, which inhibits the Ca²⁺-ATPase, increases cytosolic Ca²⁺, and augments fractional insulin secretion. Importantly, this positive-feedback loop on insulin secretion is dependent on the presence of glucose. It is conceivable that glucose or products of glucose oxidation may regulate the interaction of IRS-1 and SERCA by some means yet to be determined. Although the physiological significance of this positive feedback loop is not completely elucidated, it could serve to prime islets for optimal insulin release.

This work was supported by grants from the American Diabetes Association and the National Institutes of Health (DK49814). The Radioimmunoassay Core and the Biomedical Imaging Core of the Penn Diabetes Center are supported by the National Institutes of Health (DK19525).