Different pathways of postreceptor desensitization following chronic insulin treatment and in cells overexpressing constitutively active insulin receptors.

We have reported previously that substitution of the transmembrane domain of the insulin receptor with that of the erbB-2 oncogene (IRerbV→E) results in constitutive activation of the insulin receptor kinase. Compared to NIH3T3 cells overexpressing wild-type insulin receptors (IRwt), cells overexpressing IRerbV→E displayed a decrease in IRS-1 protein content by 55%, but basal tyrosine phosphorylation of IRS-1 was increased. This resulted in an increased association of IRS-1 with the p85 subunit of phosphatidylinositol 3-kinase, increased phosphatidylinositol 3-kinase activity in anti-IRS-1 immunoprecipitates, constitutive activation of p70 S6 protein kinase, and an increased association of Grb2 with Shc in the absence of ligand. However, Grb2 association with IRS-1 could not be detected in the basal or insulin-stimulated states, and mitogen-activated protein kinase (MAPK) activity could not be stimulated by insulin, epidermal growth factor, or platelet-derived growth factor. In contrast to IRerbV→E, the insulin receptor content and its tyrosine phosphorylation were significantly decreased in IRwt cells chronically stimulated (>24 h) with insulin. With decreased IRS-1 content, tyrosine phosphorylation of IRS-1 was decreased by over 75%, leading to decreased IRS-1-associated PI 3-kinase and Grb2. In addition, Grb2 association with Shc and activation of MAPK and the p70 S6 kinase were insensitive to insulin stimulation. By contrast, association of Grb2 with Shc and activation of MAPK, but not the p70 S6 kinase, could be stimulated by epidermal growth factor or platelet-derived growth factor. These data suggest that there are multiple levels of postreceptor desensitization to insulin action. These are used somewhat differently in these two different models, probably due in part to the difference in receptor down-regulation.

We have reported previously that substitution of the transmembrane domain of the insulin receptor with that of the erbB-2 oncogene (IR erbV3 E ) results in constitutive activation of the insulin receptor kinase. Compared to NIH3T3 cells overexpressing wild-type insulin receptors (IR wt ), cells overexpressing IR erbV3 E displayed a decrease in IRS-1 protein content by 55%, but basal tyrosine phosphorylation of IRS-1 was increased. This resulted in an increased association of IRS-1 with the p85 subunit of phosphatidylinositol 3-kinase, increased phosphatidylinositol 3-kinase activity in anti-IRS-1 immunoprecipitates, constitutive activation of p70 S6 protein kinase, and an increased association of Grb2 with Shc in the absence of ligand. However, Grb2 association with IRS-1 could not be detected in the basal or insulin-stimulated states, and mitogen-activated protein kinase (MAPK) activity could not be stimulated by insulin, epidermal growth factor, or platelet-derived growth factor.
In contrast to IR erbV3 E , the insulin receptor content and its tyrosine phosphorylation were significantly decreased in IR wt cells chronically stimulated (>24 h) with insulin. With decreased IRS-1 content, tyrosine phosphorylation of IRS-1 was decreased by over 75%, leading to decreased IRS-1-associated PI 3-kinase and Grb2. In addition, Grb2 association with Shc and activation of MAPK and the p70 S6 kinase were insensitive to insulin stimulation. By contrast, association of Grb2 with Shc and activation of MAPK, but not the p70 S6 kinase, could be stimulated by epidermal growth factor or plateletderived growth factor. These data suggest that there are multiple levels of postreceptor desensitization to insulin action. These are used somewhat differently in these two different models, probably due in part to the difference in receptor down-regulation.
The insulin receptor (IR) 1 is an integral membrane protein comprised of two extracellular ␣-subunits, which bind insulin, and two transmembrane ␤-subunits, which contain a tyrosine kinase activity that is activated following insulin binding (1). Activation of the receptor tyrosine kinase appears to be essential for most, if not all, insulin actions (1). In addition to insulin binding, the receptor-associated tyrosine kinase activity can be activated by tryptic cleavage of the ␣-subunit or removal of the ␣-subunit by in vitro mutagenesis, suggesting that the ␣-subunit acts to repress kinase activity (1). Recently, we have shown that substitution of the insulin receptor transmembrane domain with that from the erbB-2 oncogene (IR erbV3 E ) also results in a constitutively active insulin receptor kinase (2,3).
Both in intact animals and in cultured cells, chronic stimulation by insulin induces a desensitization and down-regulation of insulin signaling. This occurs in part through internalization and degradation of the insulin receptor (4,5), as well as downregulation of IRS-1 (3). In hyperinsulinemic conditions, tyrosine phosphorylation of IRS-1 and association with PI 3-kinase are also decreased, consistent with the insulin resistance that is present (6,7). In acute insulin signaling, there may be other levels of feedback inhibition, including serine phosphorylation of the insulin receptor and IRS-1, feedback inhibition of MAPK at the level of Ras and Raf-1 association (8), and dissociation of SOS from Grb2 (9 -11). In this study, we have examined the mechanism of down-regulation of insulin signaling and the relationship to the changes of insulin actions using two different cell culture models. The first is NIH3T3 cells overexpressing wild-type human IR (IR wt ) and chronically treated by insulin, in which desensitization presumably occurs at the level of the insulin receptor. The second is NIH3T3 cells overexpressing the constitutively active insulin receptor (IR erbV3 E ), in which desensitization may occur at postreceptor levels, but the level of the activated insulin receptors remains high.
The parental mouse NIH3T3 fibroblasts were transfected with the wild-type or mutant insulin receptor expression plasmids as described previously (3). Stable transfectants were selected after selection in the presence of G418 (500 g/ml) for 8 weeks. Resistant clones were selected by limiting dilution, expanded, and screened further by immunoblotting for insulin receptor expression.
Preparation of Total Cell Lysates and Immunoblotting-Cells were serum-starved for 24 h before experiments. For chronic treatment with insulin, IR wt cells were incubated with 10 Ϫ6 M insulin for 24 h. Cells were acutely treated with ligands, washed twice with ice-cold phosphate-buffered saline, and harvested in lysis buffer (50 mM HEPES, 10 mM EDTA, 100 mM NaF, 10 mM sodium pyrophosphate, 1 mM Na 3 VO 4 , 50 M okadaic acid, 1% (v/v) Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, and 0.1 mg/ml aprotinin at pH 7.5). After centrifugation, supernatants were normalized for protein concentration and subjected to immunoprecipitation and immunoblotting.
For immunoblotting, equal amounts of protein (10 -50 g) were subjected to SDS-PAGE and electroblotted to nitrocellulose filters. The filters were blocked with 3% bovine serum albumin and then incubated with a respective antibody, washed, reacted with anti-rabbit or antimouse IgG coupled to peroxidase, and developed with enhanced chemiluminescence reagents as instructed by the manufacturer. For immunoprecipitation, lysates were incubated with primary antibody for 2 h at 4°C and with Sepharose-protein A beads for an additional 1.5 h, followed by three washes with lysis buffer before elution into SDS-PAGE sample buffer. The samples were then loaded onto a SDS-acrylamide gel and subjected to electrophoresis.
Phosphatidylinositol 3-Kinase Assay-Confluent cells were treated with or without insulin (10 Ϫ7 M) for 5 min at 37°C, lysed in buffer containing 1% Nonidet P-40 and immunoprecipitated with anti-phosphotyrosine antibody or anti-IRS-1 antibody and protein A-Sepharose. The immunoprecipitates were washed and subjected to the PI 3-kinase assay as described (13). In parallel, the immunoprecipitates were also subjected to immunoblotting for IRS-1 and the 85-kDa subunit of PI 3-kinase.
MAPK Assay-Confluent cells were treated with insulin (10 Ϫ7 M), EGF (130 nM), or PDGF-BB (25 ng/ml) for 5 min at 37°C. The cell lysates were incubated with anti-MAPK and protein A-Sepharose at 4°C for 3 h. Immunoprecipitates were washed three times with lysis buffer and twice with kinase reaction buffer. The reactions were allowed to proceed for 10 min at 30°C in a final volume of 40 l containing 40 mM HEPES, pH 8.0, 10 mM MgCl 2 , 0.25 mg/ml myelin basic protein, 2 M protein kinase inhibitor (Sigma P-3294), and 50 M [␥-32 P]ATP. The reaction was stopped by the addition of 20 l of 3 ϫ Laemmli sample buffer containing 300 mM dithiothreitol. The mixtures were boiled for 5 min and subjected to SDS-PAGE. The dried gels were then subjected to PhosphorImager analysis and/or autoradiography.

RESULTS
Basal and Insulin-stimulated Tyrosine Phosphorylation-As reported previously (2,3), substitution of the transmembrane domain of the wild-type human insulin receptor with that from the oncogenic form of erbB-2 (IR erbV3 E ) results in constitutive activation of the insulin receptor kinase. This resulted in increased basal tyrosine phosphorylation of both the ␤-subunit of the insulin receptor and IRS-1 with little or no further stimulation by insulin (Fig. 1, lanes 1 and 2). However, the tyrosine phosphorylation of the ␤-subunit of the insulin receptor was decreased compared with that in acutely insulin-stimulated IR wt cells (Fig. 1, lanes 3 and 4). The tyrosine phosphorylation of IRS-1 was also decreased compared with that following acute insulin stimulation of IR wt cells. This was due, at least in part, to the decrease in IRS-1 protein content (see below). The results with IR wt cells treated with insulin for 24 h were somewhat different, with no detectable tyrosine phosphorylation of the insulin receptor in the basal state and only minimally increased IRS-1 phosphorylation. The acute addition of fresh insulin to these cells resulted in a small stimulation of IR phosphorylation but no further increase in tyrosine phosphorylation of IRS-1 ( Fig. 1, lanes 5 and 6).
In IR wt cells, insulin stimulated the tyrosine phosphorylation of 46-, 52-, and 60-kDa proteins, which include Shc and other low molecular substrates. In IR erbV3 E cells, tyrosine phosphorylation of the 46-, 52-, and 60-kDa proteins was observed independent of insulin stimulation, and insulin further stimulated tyrosine phosphorylation of the 52-and 60-kDa proteins (Figs. 1 and 4 (see below)). In contrast, in chronically insulin-treated IR wt cells, tyrosine phosphorylation of 52-and 60-kDa proteins was not observed, whereas that of the 46-kDa protein was moderately increased. In addition, in IR wt cells, there was constitutive tyrosine phosphorylation of a 120-kDa band as observed previously (14). This band includes FAK (14) and possibly other proteins. In both the IR erbV3 E cells and in chronically insulin-treated IR wt cells, the phosphorylation of pp120 was decreased.
Protein Content of the Insulin Receptor and Downstream Signaling Components-IR wt and IR erbV3 E cells expressed similar amounts of insulin receptors ( Fig. 2A). As noted in other cell types, chronic insulin treatment of IR wt cells produced a time-dependent down-regulation of the insulin receptor (Fig. 2B). By 48 h, there was a 90% reduction in the amount FIG. 1. Tyrosine phosphorylation in NIH3T3 cells overexpressing IR erbV3 E , IR wt , and chronically insulin-treated IR wt cells. NIH3T3 cells overexpressing IR erbV3 E , IR wt cells, or IR wt cells, which were pretreated with 10 Ϫ6 M insulin for 24 h (I-treated), were stimulated with or without 10 Ϫ7 M insulin for 5 min. Cells were harvested as described, and total cell lysates were subjected to immunoblotting with antiphosphotyrosine antibody.

FIG. 2. Cellular content of the insulin signaling constituents.
Monolayers of NIH3T3 cells overexpressing IR erbV3 E and IR wt were treated with or without 10 Ϫ7 M insulin for 5 min. Total cell lysates were separated by SDS-PAGE and transferred to nitrocellulose membrane; the content of the insulin receptor, IRS-1, the 85-kDa subunit of PI 3-kinase, Shc, Grb2, and MAPK was measured by quantitative immunoblotting (A). To examine an effect of chronic insulin treatment on the content of insulin signaling constituents, IR wt cells were incubated with 10 Ϫ6 M insulin for the indicated period of time, and total cell lysates were analyzed by quantitative immunoblotting for determination of the content of the insulin receptor, IRS-1, the 85-kDa subunit of PI 3-kinase, and Shc (B).
of the insulin receptors detected by immunoblotting. In addition, both chronic insulin treatment (Fig. 2B) and overexpression of IR erbV3 E ( Fig. 2A) caused a reduction of the levels of IRS-1. IRS-1 was decreased in IR erbV3 E cells by 55-60% and in chronically insulin-treated IR wt cells by 20 -30% compared to untreated IR wt cells. Interestingly, although chronic insulin treatment of IR wt cells did not change the content of Shc (Fig.  2B), the content of all isoforms of Shc (especially of 66 kDa) was increased in IR erbV3 E cells ( Fig. 2A). The levels of the 85-kDa subunit of PI 3-kinase (p85), Grb2, and MAPK were not significantly affected in IR erbV3 E cells ( Fig. 2A) or in chronically insulin-stimulated cells (Fig. 2B). The content of other constituents of the insulin signaling pathway, including mSOS, SHPTP2, Ras, and MEK1, was also unaffected (data not shown).
Tyrosine Phosphorylation of IRS-1 and Association with PI 3-Kinase and Grb2-As noted above, in a direct blot of cell lysates with anti-phosphotyrosine, tyrosine phosphorylation of IRS-1 appeared to decrease in both IR erbV3 E cells and chronically insulin-treated IR wt cells as compared with that in IR wt cells acutely stimulated with insulin ( Fig. 1). In addition, tyrosine phosphorylation of IRS-1 occurred without any treatment in both IR erbV3 E cells and chronically insulin-treated IR wt cells, and acute addition of insulin had little further effect on IRS-1 phosphorylation. To assess the apparent stoichiometry of IRS-1 tyrosine phosphorylation, the phosphorylation of IRS-1 was normalized for IRS-1 protein content. When adjusted for the IRS-1 content, the tyrosine phosphorylation of IRS-1 was increased in the basal state in both IR erbV3 E and chronically insulin-treated IR wt cells (23.9 and 34.0%, respectively, compared to acutely stimulated IR wt cells versus 6.5% in nonstimulated IR wt cells) but showed no further insulin stimulation.
In IR erbV3 E cells, the constitutively elevated tyrosine phosphorylation of IRS-1 lead to an insulin-independent increase in the formation of the IRS-1⅐p85 complex and constitutive activation of PI 3-kinase activity (Fig. 3). In IR wt cells, insulin stimulated p85 binding and PI 3-kinase activity (by 25-and 13-fold, respectively), whereas in cells expressing IR erbV3 E , insulin had no further effect on PI 3-kinase association or activity. In chronically insulin-treated IR wt cells, the formation of the IRS-1⅐p85 complex and PI 3-kinase activity in IRS-1 immunoprecipitates was also increased in the "basal" state. However, this increase was less than that in IR erbV3 E cells. Further stimulation by insulin was also greatly reduced.
The p85 subunit and PI 3-kinase activity could not be detected in the anti-phosphotyrosine immunoprecipitates in cell extracts from IR erbV3 E cells or chronically insulin-treated IR wt cells (Fig. 3). Compared to IR wt cells, IRS-1 associated with the anti-phosphotyrosine precipitates was undetectable in IR erbV3 E cells and Ͼ90% decreased in chronically insulintreated cells. In contrast to the increased basal association of IRS-1 with PI 3-kinase, there was no detectable IRS-1⅐Grb2 complex in IR erbV3 E cells before or after treatment with insulin, whereas in IR wt cells, the association of IRS-1 with Grb2 was easily detected following stimulation with insulin (Fig. 4A,  left and center panels). In chronically insulin-treated IR wt cells, some association between Grb2 and IRS-1 could be detected, although it was decreased compared with that in IR wt cells (Fig. 4A, right panel).
In summary, IRS-1 association with PI 3-kinase, but not with Grb2, is preserved in IR erbV3 E cells despite the decrease in IRS-1 protein amount and its tyrosine phosphorylation. By contrast, both PI 3-kinase and Grb2 associations with IRS-1 appear to be decreased in chronically insulin-treated IR wt cells.
Tyrosine Phosphorylation of Shc and its Association with Grb2 and Activation of MAPK-As described previously (17) and as shown in Fig. 4B, insulin stimulates the tyrosine phosphorylation of the 46-and 52-kDa isoforms of Shc. Tyrosine phosphorylation of the 66-kDa isoform was absent, consistent with the previous report (17), indicating that the tyrosinephosphorylated 60-kDa protein in Fig. 1 is not Shc but may be the GAP-associated protein (18).
Tyrosine phosphorylation of Shc leads to its association with the SH2 domains of Grb2. In IR erbV3 E cells, the tyrosine phosphorylation of Shc and its association with Grb2 was detected even in the basal state, and insulin had no further affect on this interaction. By contrast, in chronically insulin-treated IR wt cells, Shc phosphorylation and Grb2 association were not detected in the basal state or following acute insulin treatment.
Recent reports indicate that insulin-stimulated activation of MAPK occurs mainly via Shc-mediated, rather than IRS-1mediated, pathways (19,20). In IR wt cells, after a 5-min incubation with insulin, MAPK activity was stimulated by 3.6-fold (Fig. 5). By contrast, in IR erbV3 E cells, the basal activity of MAPK was low and was not stimulated by insulin, despite the fact that the tyrosine phosphorylation of Shc and its association with Grb2 was increased in these cells in the absence of ligand. In chronically insulin-treated IR wt cells, insulin did not activate MAPK, consistent with the absence of the Shc⅐Grb2 complex. These data indicate that desensitization of the insulininduced MAPK pathway occurs in both IR erbV3 E cells and chronically insulin-treated IR wt cells. However, in these two settings, the mechanism of insulin resistance appears to be different.
Insulin Regulation of p70 S6 Kinase-Recently, it has been shown that several different signaling pathways may lead to phosphorylation and activation of the p70 S6 protein kinase (15). The major pathway for stimulation of p70 S6 kinase by insulin requires activation of PI 3-kinase (12,16). In IR wt cells subjected to acute insulin stimulation, the electrophoretic mobility of p70 S6 kinase was decreased, consistent with phosphorylation and activation (Fig. 6). The 85-kDa isoform of p70 S6 kinase was also detected in the blot, and this isoform was also shifted after insulin treatment (Fig. 6). By contrast, in IR erbV3 E cells, both isoforms of p70 S6 kinase were shifted to the slower mobility form in the absence of insulin. The further addition of insulin had no effect. These data coincide with the elevated PI 3-kinase activity recovered in the IRS-1 immunoprecipitates from IR erbV3 E cells. In chronically insulin-treated IR wt cells, the mobility of p70 S6 kinase was not decreased, and acute addition of insulin did not induce a mobility shift, consistent with the lower levels of PI 3-kinase activity recovered in the IRS-1 immunoprecipitates, both in the basal and stimulated states.
EGF or PDGF Actions in IR erbV3 E Cells and Chronically Insulin-treated IR wt Cells-To better understand the mechanisms of desensitization, we examined the EGF-and PDGFstimulated signaling pathways to see whether the desensitization process(es) that affected insulin signaling in IR erbV3 E cells and in chronically insulin-treated IR wt cells would also occur in other pathways. In IR wt cells, EGF and PDGF activated MAPK to the same extent as insulin (Fig. 5). In addition, in these cells both EGF and PDGF stimulated tyrosine phosphorylation of the 46-and 52-kDa Shc proteins and the formation of Shc⅐Grb2 complexes (Fig. 4B). However, in IR erbV3 E cells, neither EGF nor PDGF stimulated MAPK activity (Fig. 5). There was also no further increase of the already elevated levels of tyrosinephosphorylated Shc or in the formation of Shc⅐Grb2 complexes. These data suggest that at least one site of desensitization is common to the insulin, EGF, and PDGF signaling pathways, and that this step is probably downstream of the formation of Shc⅐Grb2 complexes. In addition, although EGF-induced tyrosine phosphorylation of EGF receptors was difficult to observe, PDGF-induced tyrosine phosphorylation of PDGF receptors could be clearly observed (data not shown), indicating that overexpression of IR erbV3 E did not induce desensitization at the level of the PDGF receptor itself. In contrast to cells expressing IR erbV3 E , in chronically insulin-treated IR wt cells, EGF and PDGF stimulated the formation of Shc⅐Grb2 complexes (Fig. 4B) and activated MAPK (2.9-and 2.2-fold, respectively), whereas insulin could not (Fig. 5). The tyrosine phosphorylation of 52-kDa Shc was also induced by EGF or PDGF treatment, although it was reduced compared to that in IR wt cells (Fig. 4B). Likewise, PDGF-induced tyrosine phosphorylation of PDGF receptors was also markedly reduced in chronically insulin-treated cells (data not shown). These data suggest that in chronically insulin-treated cells, EGF or PDGF signaling to Shc is partially desensitized but is sufficient to induce the formation of Shc⅐Grb2 complexes and activate MAPK.
These results are somewhat different from those observed for activation of the p70 S6 kinase. Thus, in IR wt cells, the electrophoretic mobility of both isoforms of p70 S6 kinase was decreased following EGF or PDGF treatment as well as by insulin (Fig. 6). In IR erbV3 E cells, p70 S6 kinase was constitutively in the slower mobility form and was not further altered by EGF or PDGF treatment. In contrast, neither EGF nor PDGF could induce the mobility shift in chronically insulintreated IR wt cells, suggesting that chronic insulin treatment induces desensitization of not only insulin but also EGF and PDGF signaling pathway to p70 S6 kinase.

DISCUSSION
Chronic insulin stimulation induces down-regulation in various steps in insulin signaling. Insulin has been shown to induce the internalization and degradation of the insulin receptor (4,5). Insulin is also reported to decrease IRS-1 content in CHO (3) and 3T3-L1 cells (21) in culture. Prolonged exposure to insulin has also been shown to result in a desensitization of the final effectors of the insulin action pathway, such as stimulation of glucose transport (22). In the present study, we have compared desensitization pathways in NIH3T3 cells overexpressing a constitutively activated insulin receptor to that observed in cells overexpressing wild-type insulin receptors chronically treated with insulin. We find that chronic insulin treatment of IR wt cells decreases the content of insulin receptor protein and produces a significant decrease in insulin-induced tyrosine phosphorylation of the insulin receptor and IRS-1, although the content of IRS-1 is only moderately decreased. This leads to a decrease in the stoichiometry of insulin-induced tyrosine phosphorylation of IRS-1 as compared with that ob-  6. Gel shift assay of p70 S6 kinase. NIH3T3 cells overexpressing IR erbV3 E , IR wt cells, or chronically insulin-treated IR wt cells (Itreated) were stimulated with or without 10 Ϫ7 M insulin for 10 min. Total cell lysates were subjected to SDS-PAGE, and p70 S6 kinase was detected by immunoblotting with anti-p70 S6 kinase. A and B are the same blot; however, the autoradiograph in A was exposed for a longer period to detect the 85-kDa isoform of p70 S6 kinase. served in IR wt cells and a decrease in both IRS-1-associated PI 3-kinase and Grb2. In addition, insulin-induced tyrosine phosphorylation of Shc and its association with Grb2 is also decreased. The decreases in these intermediate pathways lead to a loss in the ability of insulin to stimulate MAPK and p70 S6 kinase. Thus, there is a "broad spectrum" desensitization of insulin action following chronic insulin stimulation both at the level of the insulin receptor and involving several postreceptor mechanisms. Findings in the cells overexpressing the constitutively activated receptor help sort out the role of these various defects.
In IR erbV3 E cells, the basal level of insulin receptor kinase activity is high, and receptors are not down-regulated. Despite a significant decrease in IRS-1 content and in the apparent stoichiometry of IRS-1 tyrosine phosphorylation, IRS-1 is tyrosine-phosphorylated and associates with the 85-kDa subunit of PI 3-kinase in a ligand-independent manner, resulting in constitutively elevated PI 3-kinase activities. In these cells, p70 S6 kinase is also constitutively phosphorylated. These results are consistent with the model in which PI 3-kinase is an upstream regulator of p70 S6 kinase (12,16). Furthermore, these data also suggest that continuous signals generated from PI 3-kinase downstream to p70 S6 kinase do not significantly alter p70 S6 kinase protein levels or appear to affect any intermediate signals between PI 3-kinase and p70 S6 kinase.
Interestingly, PI 3-kinase activity and the 85-kDa subunit of PI 3-kinase could not be recovered in anti-phosphotyrosine antibody immunoprecipitates in IR erbV3 E cells, although these activities could be detected in the anti-phosphotyrosine immunoprecipitates from IR wt cells after insulin treatment. Moreover, in IR erbV3 E cells, no association between IRS-1 and Grb2 could be detected. These results suggest that tyrosine residues other than the PI 3-kinase binding sites on IRS-1 are not phosphorylated or are quickly dephosphorylated in IR erbV3 E cells. Furthermore, although the apparent stoichiometry of IRS-1 tyrosine phosphorylation is decreased to a similar extent in IR erbV3 E cells compared to that in chronically insulintreated IR wt cells, more IRS-1⅐PI 3-kinase complex is formed in IR erbV3 E cells. These findings are consistent with the recent report (23) that a constitutively active insulin-like growth factor 1 receptor (IGF-1R), which is mutated in the transmembrane region (Val 922 3 Glu), turns on IGF-1R signaling in CHO cells to an extent similar to that seen with stimulation of wild-type IGF-1R with submaximal concentration of IGF-1. In addition, they also raised the possibility that the Val 922 3 Glu mutant of IGF-1R may induce a conformational change of IGF-1R to phosphorylate the PI 3-kinase-binding site on IRS-1 but not some of other sites on IRS-1 (23).
Another possibility for the decreased stoichiometry of IRS-1 tyrosine phosphorylation is an increase in protein-tyrosine phosphatase activities induced by constitutive activation of the insulin signaling pathway. Recently, several protein-tyrosine phosphatases have been shown to be induced by insulin and other growth factors (24), and several protein-tyrosine phosphatases, including PTP CD45 (25), PTP-LAR (26), PTP1B (27,28), and PTP␣ and PTP⑀ (29), have been reported to inhibit insulin signaling. Consistent with the latter hypothesis, the tyrosine phosphorylation of several proteins, including pp120, the ␤-subunit of IR, and IRS-1, is decreased in IR erbV3 E cells. It has been reported that in vitro SHPTP2 preferentially dephosphorylates phosphotyrosines of the Grb2 and SHPTP2 binding sites on IRS-1, as compared to the PI 3-kinase binding sites (30). Thus, increased activity of a protein-tyrosine phosphatase possessing substrate specificity similar to SHPTP2 could result in selective desensitization of Grb2-mediated effects in IR erbV3 E cells.
The situation in chronically insulin-treated cells is similar, but not identical, to that in IR erbV3 E cells. For example, in chronically insulin-treated cells, PDGF-induced tyrosine phosphorylation of PDGF receptors is also decreased, in contrast to apparently normal stimulation in IR erbV3 E cells, and there is desensitization of insulin-, EGF-, and PDGF-stimulated p70 S6 kinase. There is also a decrease of tyrosine phosphorylation of some constitutively phosphorylated proteins (for example, pp120) in chronically insulin-treated cells, again raising the possibility that an increase in protein-tyrosine phosphatase activity might play some role in the desensitization. However, in contrast to IR erbV3 E cells, in which Grb2, but not PI 3-kinase, association is markedly decreased, in chronically stimulated cells, Grb2 and PI 3-kinase in anti-IRS-1 immunoprecipitates are decreased to a similar extent. Although this could simply reflect the decrease of insulin receptor kinase activity, the global desensitization of p70 S6 kinase must reflect another mechanism.
In terms of the MAPK pathway, MAPK activity is low and not activated by insulin in either IR erbV3 E or in chronically insulin-treated IR wt cells. However, in IR erbV3 E cells, tyrosine phosphorylation of Shc and its association with Grb2 occurs in the absence of insulin, and neither insulin, EGF, nor PDGF can further increase the Grb2 association of Shc or activate MAPK. These results suggest that some pathway of MAPK activation that is common to insulin, EGF, and PDGF signaling is desensitized downstream of the Shc⅐Grb2 complex. However, PDGF can fully activate MAPK, despite the lower levels of detectable Shc⅐Grb2 complexes in IR wt cells, and PDGF may utilize pathways other than Shc for MAPK activation (31). In addition, several recent reports have suggested that there are several potential MEK (MAPK/extracellular signal-regulated kinase) activators, that may be differentially regulated by various growth factors (32). This indicates the complexity and redundancy of pathways for MAPK activation. Nonetheless, PDGF activation of MAPK is also severely compromised in IR erbV3 E cells, suggesting that desensitization may occur at a common site that appears downstream or independent of Shc/Grb2generated signals. By contrast, in chronically insulin-treated IR wt cells, EGF or PDGF can stimulate tyrosine phosphorylation of Shc and its association with Grb2 and activate MAPK, whereas insulin could not. Although the levels of EGF-and PDGF-stimulated formation of Shc⅐Grb2 complexes were similar to those observed in IR wt cells, the levels of MAPK activity were slightly less. This suggests that the EGF or PDGF signaling from the receptor to Shc and Grb2 is not lost in these cells, but some desensitization may occur at steps thereafter.
One possibility that may explain these findings is a change in factors leading to deactivation of MAPK. Recently, it has been shown that angiotensin 2 can deactivate MAPK by enhancing MAPK phosphatase 1 (MKP1), a protein-tyrosine phosphatase that dephosphorylates and inactivates MAPK (33). MKP1 and other vaccinia H-1 gene product-like proteintyrosine phosphatases have also been shown to be induced by several growth factors (34,35). Therefore, it is possible that an increased MAPK deactivation activity accounts for the failure of MAPK activation by insulin, EGF, and PDGF in IR erbV3 E cells and the decreased activation of MAPK by EGF and PDGF in chronically insulin-treated IR wt cells.
In summary, in chronically insulin-treated IR wt cells, desensitization in insulin signaling involves multiple downstream responses but appears to be due mainly to the decrease of the insulin receptor content and insulin receptor kinase activity. However, since tyrosine phosphorylation of PDGF receptors and p70 S6 kinase activation by EGF or PDGF are also compromised, chronic insulin stimulation must induce significant changes in a postreceptor step(s) common to the insulin, EGF, and PDGF signaling pathways. Overexpression of the constitutively active insulin receptor also induces some profound changes in insulin, EGF, and PDGF signaling pathways that are different from those in chronically insulin-treated cells. In these cells, the activated insulin receptor kinase leads to constitutive activation of IRS-1⅐PI 3-kinase complex formation and p70 S6 kinase. However, IRS-1⅐Grb2 complex formation cannot be detected, and MAPK activation by insulin, EGF, and PDGF treatment does not occur, although Shc⅐Grb2 complex is present in the absence of external stimulation. Taken together, these data indicate that there are multiple levels of desensitization in the insulin action pathway, and that desensitization of some postreceptor sites may also result in alterations in signaling by other growth factors.