Search

Insulin stimulates the activity of mitogen-activated protein kinase (MAPK) via its upstream activator, MAPK kinase (MEK), a dual specificity kinase that phosphorylates MAPK on threonine and tyrosine. The potential role of MAPK activation in insulin action was investigated with the specific MEK inhibitor PD98059. Insulin stimulation of MAPK activity in 3T3-L1 adipocytes (2.7-fold) and L6 myotubes (1.4-fold) was completely abolished by pretreatment of cells with the MEK inhibitor, as was the phosphorylation of MAPK and pp90Rsk, and the transcriptional activation of c-fos. Insulin receptor autophosphorylation on tyrosine residues and activation of phosphatidylinositol 3'-kinase were unaffected. Pretreatment of cells with PD98059 had no effect on basal and insulin-stimulated glucose uptake, lipogenesis, and glycogen synthesis. Glycogen synthase activity in extracts from 3T3-L1 adipocytes and L6 myotubes was increased 3-fold and 1.7-fold, respectively, by insulin. Pretreatment with 10 microM PD98059 was without effect. Similarly, the 2-fold activation of protein phosphatase 1 by insulin was insensitive to PD98059. These results indicate that stimulation of the MAPK pathway by insulin is not required for many of the metabolic activities of the hormone in cultured fat and muscle cells.

Activation of the tyrosine kinase activity of the insulin receptor by autophosphorylation leads to phosphorylation of cellular substrates on tyrosine. Thus far, the best characterized is the insulin receptor substrate (IRS) 1, which has been proposed to serve as a docking protein for other molecules involved in signal transduction. A number of other proteins that become phosphorylated in response to insulin have been identified, some of which are reported to be tissue-specific. A 60 kDa phosphoprotein has been detected in adipocytes after insulin stimulation [Lavan and Lienhard (1993) J. Biol. Chem. 268, 5921-5928]. We have identified a protein of similar molecular mass in rat hepatoma cells transfected with the human insulin receptor. The 60 kDa protein in hepatoma cells is tyrosine-phosphorylated in response to insulin in a dose-dependent manner, with maximal phosphorylation occurring at 50 nM insulin. Although the dose-response of p60 phosphorylation mirrors that of IRS-1, the time course is slightly slower, with maximal phosphorylation observed 5 min after addition of insulin. Like the adipocyte protein, the 60 kDa protein detected in liver cells binds to the SH2 domain of the p85 regulatory subunit of phosphatidylinositol 3-kinase, but not to other SH2 domains. Binding of p60 to p85 is similar to the interaction between p85 and IRS-1 in that a tyrosine-phosphorylated peptide containing the YVXM motif can inhibit the association. The presence of this 60 kDa tyrosine-phosphorylated protein in adipocytes and hepatoma cells suggests that it represents another important intermediate in the insulin-receptor signal-transduction pathway.

The precise mechanism by which insulin regulates glucose metabolism is not fully understood. However, it is known that insulin activates two enzymes, phosphatidylinositol 3'-kinase (PI 3'-K) and mitogen-activated protein kinase (MAPK), which may be involved in stimulating the metabolic effects of insulin. The role of these enzymes in glucose metabolism was examined by comparing the effects of insulin, platelet-derived growth factor (PDGF) and epidermal growth factor (EGF) in 3T3-L1 adipocytes. Treatment of the cells with PDGF or EGF for 5 min increased the MAPK activity 3-5-fold, while insulin treatment produced a 2.5-fold increase. The MAPK activity remained elevated for 1 h after either PDGF or insulin treatment. PDGF and insulin, but not EGF, caused a transient increase in the amount PI 3'-K activity coprecipitated with tyrosine phosphorylated proteins. Although PDGF and insulin caused a similar increase in the activities of these two enzymes, only insulin caused substantial increases in glucose utilization. Insulin increased the transport of glucose and the synthesis of lipid 4- and 17-fold, respectively, while PDGF did not affect these processes significantly. Glycogen synthesis was increased 15-fold in response to insulin and only 3-fold in response to PDGF. Thus, the activation of MAPK and PI 3'-K are not sufficient for the complete stimulation of glucose transport, lipid synthesis, or glycogen synthesis by hormones in 3T3-L1 adipocytes, suggesting a requirement for other signaling mechanisms that may be uniquely responsive to insulin.

Although the insulin-dependent hydrolysis of glycosyl-phosphatidylinositol (GPI) may play an important role in insulin action, an absolute requirement for this glycolipid has not been demonstrated. Human K562 cells were mutated to produce a cell line (IA) incapable of the earliest step in PI glycosylation, the formation of PI-GlcNAc. Another cell line (IVD) was deficient in the deacetylation of PI-GlcNAc to form PI-GlcN and subsequent mannosylated species. Each line was transfected with wild-type human insulin receptors. Similar insulin-stimulated receptor autophosphorylation was observed in all three lines, along with a nearly identical increase in the association of phosphorylated insulin receptor substrate 1 with endogenous PI 3-kinase. Both normal and GPI-defective lines also displayed a similar 2- to 3-fold increase in phosphorylation of the Shc protein and its association with growth factor receptor-bound protein 2 in response to insulin. In contrast to these results, striking differences were noted in insulin-stimulated glycogen synthesis. In normal cells, glycogen synthesis was significantly increased by insulin, whereas no insulin stimulation was observed in GPI-deficient IA cells, and only a trace of stimulation was detected in IVD cells. These results indicate that tyrosine phosphorylation produced by insulin is not dependent on GPI synthesis, and this effect is not sufficient to elicit at least some of the metabolic effects of the hormone. In contrast, GPI synthesis is required for the stimulation of glycogen synthesis by insulin in these cells. These findings support the existence of divergent pathways in the action of insulin.

Naturally occurring human insulin receptor mutants Ser1200 and Thr1134, and a site-directed mutant Arg1030 overexpressed in Chinese hamster ovary (CHO) cells, bind insulin with affinities identical to wildtype receptors but are apparently kinase deficient. Cells expressing the Ser1200 receptor exhibit insulin stimulation of glycogen synthesis similar to these bearing the wildtype receptor, but fail to mediate insulin-responsive DNA synthesis. In contrast, the Thr1134 and Arg1030 mutants exhibit no response to insulin. The activity of Mitogen Activated Protein (MAP) kinase in cells transfected with wildtype receptor is more responsive to insulin than that detected in untransfected parental cells, while cells bearing any of the mutant receptors are less responsive than the parental cells. These differences in the stimulation of MAP kinase activity are paralleled by differences in insulin-dependent phosphorylation of the enzyme. These results suggest that the p42 MAP kinase is not universally required for the metabolic effects of insulin.