Search

Triglyceride (TG) accumulation in pancreatic beta-cells is thought to be associated with impaired insulin secretory response to glucose (lipotoxicity). To better understand the mechanism of the impaired insulin secretory response to glucose in beta-cell lipotoxicity, we overexpressed a constitutively active form of the sterol regulatory element-binding protein- 1c (SREBP-1c), a master transcriptional factor of lipogenesis, in INS-1 cells with an adenoviral vector. This treatment was associated with strong activation of transcription of the genes involved in fatty acid biosynthesis, increased cellular TG content, severely blunted glucose-stimulated insulin secretion, and enhanced expression of the uncoupling protein-2 (UCP-2), which supposedly dissipates the mitochondrial electrochemical potential. To decrease the up-regulated UCP-2 expression, small interfering RNA for UCP-2 was used. Introduction of the small interfering RNA increased the ATP/ADP ratio and partially rescued the glucose-stimulated insulin secretion in the cells overexpressing SREBP-1c, but did not affect the cellular TG content. Next, the effect of the AMP-activated protein kinase (AMPK) agonist, 5-amino-4-imidazolecarboxamide riboside, was examined in the lipotoxicity model. Exposure of the cells with lipotoxicity to 5-amino-4-imidazolecarboxamide riboside increased free fatty acid oxidation, partially reversed the TG accumulation, phosphorylated AMPK and acetyl-coenzyme A carboxylase, and improved the impaired glucose-stimulated insulin secretion. These results suggest that UCP-2 down-regulation and AMPK activation could be candidate targets for releasing beta-cells from lipotoxicity.

To study the regulation of growth and differentiated function of insulin-secreting cells, the rat insulinoma cell line INS-1 was cultured in a defined serum-free medium containing prolactin, IGF-I, and triiodothyronine, which was originally reported to maintain insulin secretion of islet cells. Growth and viability, as well as cellular insulin content of INS-1 cells in the defined medium, were comparable to the control cells cultured in the complete medium containing 10% fetal calf serum. However, after a 3-day culture in this medium, insulin secretion in response to glucose, pyruvate, and leucine was markedly blunted compared with the control cells (-78, -68, and -56%, respectively), whereas the response to 30 mmol/l K+ was only slightly decreased. In these cells: 1) nutrient metabolism assessed by tetrazolium salt reduction was reduced in response to pyruvate and leucine, which are mainly metabolized in the mitochondria; 2) oxidation of both [3,4-(14)C]glucose and [1-(14)C]pyruvate was decreased (-22 and -32%, respectively); 3) glucose failed to depolarize the membrane potential, whereas tolbutamide was fully active; 4) video imaging analysis of cytosolic Ca2+ showed a decrease in the population of glucose-responsive cells, while the response to 30 mmol/l K+ was preserved; 5) serum replenishment for 3 days restored glucose-induced insulin secretion. Interestingly, conditioned serum-free medium from rat islets maintained the insulin secretory function of INS-1 cells, although glucagon, somatostatin, and some other factors failed to restore the function. In contrast, conditioned media from HepG2, PC12, and human umbilical vein endothelial cells did not substitute for serum. Thus, the impaired insulin secretion of the cells cultured in the defined medium is best explained by defective mitochondrial metabolism. Islet cells, but not INS-1 cells, produce factors required for normal signal generation by nutrient secretagogues.

The signaling pathways whereby glucose and hormonal secretagogues regulate insulin-secretory function, gene transcription, and proliferation of pancreatic beta-cells are not well defined. We show that in the glucose-responsive beta-cell line INS-1, major secretagogue-stimulated signaling pathways converge to activate 44-kDa mitogen-activated protein (MAP) kinase. Thus, glucose-induced insulin secretion was found to be associated with a small stimulatory effect on 44-kDa MAP kinase, which was synergistically enhanced by increased levels of intracellular cAMP and by the hormonal secretagogues glucagon-like peptide-1 and pituitary adenylate cyclase-activating polypeptide. Activation of 44-kDa MAP kinase by glucose was dependent on Ca2+ influx and may in part be mediated by MEK-1, a MAP kinase kinase. Stimulation of Ca2+ influx by KCl was in itself sufficient to activate 44-kDa MAP kinase and MEK-1. Phorbol ester, an activator of protein kinase C, stimulated 44-kDa MAP kinase by both Ca(2+)-dependent and -independent pathways. Nerve growth factor, independently of changes in cytosolic Ca2+, efficiently stimulated 44-kDa MAP kinase without causing insulin release, indicating that activation of this kinase is not sufficient for secretion. In the presence of glucose, however, nerve growth factor potentiated insulin secretion. In INS-1 cells, activation of 44-kDa MAP kinase was partially correlated with the induction of early response genes junB, nur77, and zif268 but not with stimulation of DNA synthesis. Our findings suggest a role of 44-kDa MAP kinase in mediating some of the pleiotropic actions of secretagogues on the pancreatic beta-cell.

Nutrient metabolism was examined with regard to insulin secretion in purified rat islet beta- and non-beta-cells, beta-cell lines, and hepatocytes. Lactate dehydrogenase (LDH) activity (nanomoles.min-1.mg protein-1) was remarkably low in the glucose-sensitive INS-1 cell line (15.7) and in beta-cells (22.3). Thus, beta-cell LDH was respectively 8-, 122-, 17-, and 136-fold lower than in islet non-beta, liver, HIT-T15, and RINm5F cells. Plasma membrane lactate transport activity was 3-10-fold lower in beta- or INS-1 cells than in the other cell types. Conversely, mitochondrial glycerol phosphate dehydrogenase was strongly expressed only in beta- and INS-1 cells. The significance of these findings to nutrient recognition was explored using INS-1 cells as a model of native beta-cells. Glucose-stimulated lactate output and glucose utilization were, respectively, 12- and 5-fold lower in INS-1 than in RINm5F cells. Each process was entirely blocked by respiratory chain inhibitors in INS-1 cells, whereas glucose utilization was barely affected and lactate output stimulated in RINm5F cells. Glucose oxidation represented 73% of total utilization in INS-1 cells, but only 9% in RINm5F cells. Absolute rates of glucose oxidation, and the extent of mitochondrial NAD(P) reduction, were similar in the two cell types, and glucose stimulated insulin secretion 1.9-fold in INS-1 and 1.4-fold in RINm5F cells. The mitochondrial substrates, monomethyl succinate, pyruvate, and leucine, each triggered secretion in INS-1 cells. The balance of LDH, plasma membrane lactate transport, and mitochondrial glycerol phosphate dehydrogenase activities therefore appear to be important in beta- and INS-1 cell glucose recognition to ensure that mitochondrial oxidation is the principle fate of pyruvate and NADH produced by glycolysis. The resultant close coupling of glycolysis with mitochondrial oxidation explains the absence in beta-cells of Crabtree and Pasteur effects.