Growth Hormone Induces Cellular Insulin Resistance by Uncoupling Phosphatidylinositol 3-Kinase and Its Downstream Signals in 3T3-L1 Adipocytes

Porcine insulin and human GH were kindly provided by Eli Lilly (Indianapolis, IN) and Novo Nordisk (Bagsvaerd, Denmark), respectively. Dulbecco’s modified Eagle’s medium (DMEM), fetal bovine serum (FBS), and human recombinant platelet-derived growth factor (PDGF)-BB were purchased from Life Technologies/Gibco (Burlington, ON, Canada). 2-Deoxy-d-[³H]glucose was purchased from New England Nuclear (Boston, MA). [γ-³²P]ATP and protein G-Sepharose were purchased from Amersham Pharmacia Biotech. Anti–IR-β subunit (C-19), anti-JAK2, anti-STAT5b, and horseradish-peroxidase–linked anti-rabbit and anti-mouse IgG antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-rat carboxyl-terminal IRS-1 and anti-human IGF-1 antibodies were purchased from Upstate Biotechnology (Lake Placid, NY). Anti-phosphotyrosine (PY20, PY20H) and anti-p85 antibodies were purchased from Transduction Laboratories (Lexington, KY). Anti-Akt, anti–phospho-Akt (Ser473), and anti–phospho-Akt (Thr308) antibodies were purchased from New England Biolabs (Beverly, MA).

3T3-L1 fibroblasts were obtained from the American Type Culture Collection (Rockville, MD) and maintained at 37°C in DMEM/high glucose supplemented with 10% FBS, penicillin (100 units/ml), and streptomycin (100 μg/ml) in a 10% CO₂ environment. Cells were differentiated 2 days postconfluence by the addition of the same media supplemented with 1 μmol/l insulin, 1 μmol/l dexamethasone, and 500 μmol/l isobutyl-methylxanthine for 3 days, and then the medium containing 0.8 μmol/l insulin for another 3 days. The medium was then changed every 3 days until the cells were used for the experimentation (i.e., 13–16 days after the induction of differentiation), when >95% of the cells had the morphorogical and biochemical properties of adipocytes.

The recombinant adenoviruses Ad5-p110^CAAX (containing bovine p110α cDNA with the CAAX motif at the COOH terminus) and Ad5-CT (which has no insert) were gifts from Dr. Olefsky (University of California, San Diego, CA). They were amplified in human embryonic kidney 293 cells, and viral stock solutions with viral titers of >10⁸ plaque-forming units/ml were prepared. 3T3-L1 adipocytes were infected with the adenoviruses by incubating the cells at the indicated multiplicity of infection (MOI) of viral stock solution in DMEM containing 2% heat-inactivated FBS for 16 h. The media were changed to the regular culture media, and the cells were used for experiments after 48 h of infection (48).

3T3-L1 adipocytes were washed with PBS and incubated with Krebs-Ringer phosphate (KRP)-HEPES buffer (10 mmol/l HEPES, pH 7.4, 131.2 mmol/l NaCl, 4.7 mmol/l KCl, 1.2 mmol/l MgSO₄, 2.5 mmol/l CaCl₂, and 2.5 mmol/l NaH₂PO₄) containing 1% bovine serum albumin (BSA) with the indicated concentration of insulin for 15 min at 37°C. The adipocytes were then incubated with 2-deoxy-d-[³H]glucose for 4 min, and the reaction was terminated by the addition of 10 μmol/l cytochalasin B (Sigma, St Louis, MO). Cells were washed three times with ice-cold PBS. Radioactivity taken up by the cells was measured by a liquid scintillation counter.

3T3-L1 adipocytes were rinsed twice with PBS and once with HES buffer (255 mmol/l sucrose, 20 mmol/l HEPES, pH 7.4, 1 mmol/l EDTA, 1 mmol/l phenylmethylsulfonyl fluoride [PMSF], 1 mmol/l Na₃VO₄, 2 μg/ml aprotinin, and 50 nmol/l okadaic acid) (49) and immediately homogenized by 20 strokes with a motor-driven Teflon/glass homogenizer in HES buffer at 4°C. The homogenates (two 10-cm dishes per condition) were subjected to subcellular fractionation as previously described to isolate PM, high-density microsomes (HDMs), low-density microsomes (LDMs), and cytosol (49,50), with some modification. Briefly, the homogenate was centrifuged at 19,000g for 20 min. The resulting supernatant was centrifuged at 41,000g for 20 min, yielding a pellet of HDMs. The supernatant from this spin was centrifuged at 250,000g for 90 min, yielding a pellet of LDMs. The remaining supernatant was concentrated by Centricon-30 (Amicon, Beverly, MA) and used as cytosol. The pellet obtained from the initial spin was resuspended in HES buffer, layered onto a 1.12 mol/l sucrose cushion, and centrifuged at 100,000g in swing rotor for 60 min. A white fluffy band at the interface was collected, resuspended in HES buffer, and centrifuged at 40,000g for 20 min, yielding a pellet of PM. All fractions were adjusted to a final protein concentration of 1–3 mg/ml and stored at −80°C. The protein concentrations of these fractions were measured using the Bradford method.

3T3-L1 adipocytes were solubilized in cell lysis buffer containing 20 mmol/l Tris, pH 7.5, 140 mmol/l NaCl, 1% Nonidet P-40, 1 mmol/l EDTA, 1 mmol/l EGTA, 2.5 mmol/l sodium pyrophosphate, 1 mmol/l β-glycerophosphate, 2 mmol/l Na₃VO₄, 50 mmol/l sodium fluoride, 10 μg/ml aprotinin, and 1 mmol/l PMSF on ice. In some experiments, each subcellular fraction was solubilized in the homogenization buffer supplemented with 1% Nonidet P-40. After centrifugation, the supernatants were mixed with Laemmli buffer and boiled for 5 min. Immunoprecipitation was performed as follows: the supernatants were incubated with the indicated antibody at 4°C overnight, and the immune complexes were collected on protein G–Sepharose for 2 h at 4°C, washed three times with the lysis buffer, and boiled in Laemmli buffer. Samples were run on 7.5% SDS-PAGE gels, transferred onto polyvinylidene difluoride membrane, and immunoblotted with the indicated antibodies.

3T3-L1 adipocytes were grown in 10-cm dishes, washed with washing buffer A (20 mmol/l Tris, pH 7.6, 137 mmol/l NaCl, 1 mmol/l MgCl₂, 1 mmol/l CaCl₂, and 1 mmol/l PMSF), solubilized in lysis buffer (20 mmol/l Tris, pH 7.6, 137 mmol/l NaCl, 1 mmol/l MgCl₂, 1 mmol/l CaCl₂, 100 μmol/l Na₃VO₄, 1% Nonidet P-40, 10% glycerol, 1 mmol/l PMSF, 0.1 mg/ml aprotinin, and 1 mmol/l dithiothreitol), placed on ice for 20 min, then centrifuged at 14,000 rpm for 20 min. In some experiments, each subcellular fraction was solubilized in the homogenization buffer supplemented with 1% Nonidet P-40. After centrifugation, the supernatants were immunoprecipitated with anti–IRS-1 antibody overnight, followed by adsorption of the resulting immune complexes to protein G–Sepharose beads for 2 h at 4°C. The pelleted beads were washed with washing buffer A, washing buffer B (100 mmol/l Tris, 500 mmol/l LiCl, and 100 μmol/l Na₃VO₄), and then washing buffer C (10 mmol/l Tris, 100 mmol/l NaCl, and 1 mmol/l EDTA) twice each. The immunoprecipitates were resuspended in 10 μl of buffer C and then incubated with a PI solution containing 0.5 mg/ml PI, 50 mmol/l HEPES, pH 7.6, 1 mmol/l NaH₂PO₄, and 1 mmol/l EGTA. The PI 3-kinase reaction was started by the addition of reaction mixture containing 250 μmol/l [γ-³²P]-ATP (0.37 MBq/tube), 100 mmol/l HEPES, and 50 mmol/l MgCl₂. After 5 min at 25^oC, the reaction was terminated by the addition of 15 μl of 8 N HCl. The products were extracted by the addition of 130 μl chloroform/methanol (1:1) followed by centrifugation. The mixture was centrifuged to separate the phases, and the lower organic phase was removed, applied to a silica gel thin-layer chromatography plate (Merk, Darmstadt, Germany), developed in CHCl₃:CH₃OH:H₂O:NH₄OH (60:47:11.3:2), and dried. ³²P incorporated into PI was visualized by autoradiography and quantitated using a Fujix Bas 2000.

The quantitation of every immunoblot band was performed by video scanning densitometry, and the results are presented as the means ± SE. Statistical significance between the two groups was evaluated using Student’s t test. The criterion for significance was set at P < 0.05 or P < 0.01.

It is widely accepted that GH causes insulin resistance in humans and animals (30,31,32,33,34). However, it is less clear whether GH impairs insulin signaling at the cellular level. To address this issue, we first determined the effect of chronic GH treatment on basal and insulin-stimulated 2-DOG uptake in fully differentiated 3T3-L1 adipocytes. After incubation with 500 ng/ml GH for 16 h, cells were stimulated with various concentrations of insulin for 15 min, and 2-DOG uptake was then determined (Fig. 1). Insulin stimulation increased 2-DOG uptake in a dose-dependent manner, and the maximal response was observed at 20 nmol/l (Fig. 1). Chronic GH treatment had no effect on the basal 2-DOG uptake but significantly reduced insulin-stimulated 2-DOG uptake by 25, 30, 26, 22, and 21% at insulin concentrations of 0.1, 1, 10, 20, and 100 nmol/l, respectively (Fig. 1).

To investigate the molecular mechanism by which GH impairs insulin-stimulated glucose transport, we next examined which component(s) of the insulin signaling pathway is affected in GH-mediated insulin resistance. First, we determined the effect of GH on the tyrosine phosphorylation of cellular proteins by immunoblotting whole-cell lysates with anti-phosphotyrosine antibody (Fig. 2). Insulin stimulation increased the tyrosine phosphorylation of the 90- to 95-kDa and 180- to 185-kDa proteins (Fig. 2), which were considered as IR and IRS proteins, respectively. As shown in Fig. 2 (lane 2), chronic GH treatment alone also increased tyrosine phosphorylation of the proteins of similar molecular weight ranges. When stimulated with the submaximal concentrations of insulin, GH significantly enhanced the tyrosine phosphorylation of the proteins in both of these molecular weight ranges (Fig. 2, lanes 4 and 6). However, when the cells were stimulated with the maximal concentration of insulin (20 nmol/l), the tyrosine phosphorylation of 90- to 95-kDa proteins was not further affected by GH treatment, whereas the tyrosine phosphorylation of 180- to 185-kDa proteins was increased in GH-treated cells (Fig. 2, lane 8).

To determine whether chronic GH treatment affects tyrosine phosphorylation of the IR-β subunit, the cell lysates were immunoprecipitated with anti–IR-β subunit antibody, followed by immunoblotting with anti-phosphotyrosine antibody (Fig. 3A). GH treatment did not significantly affect tyrosine phosphorylation of the IR-β subunit in either the absence or presence of various concentrations of insulin (Fig. 3A, upper panel). Chronic GH treatment was without effect on the IR protein levels, as determined by immunoblotting with anti–IR-β subunit antibody (Fig. 3A, lower panel). Because GH stimulation is known to result in tyrosine phosphorylation of STATs through the activation of JAK2, and because STAT5b has been reported to be the major tyrosine-phosphorylated STAT induced by GH (39,40), tyrosine phosphorylation of STAT5b was then determined (Fig. 3B). Chronic GH pretreatment increased the tyrosine phosphorylation of STAT5b (92 kDa), which was not affected by the subsequent insulin stimulation (Fig. 3B, upper panel). Chronic GH treatment did not affect the protein level of STAT5b (Fig. 3B, lower panel). To exclude the possibility that the increased tyrosine phosphorylation of 90- to 95-kDa proteins also includes activated IGF-1 receptor, the effect of chronic GH treatment on tyrosine phosphorylation of IGF-1 receptor was then examined (Fig. 3C). Neither chronic GH treatment nor 20 nmol/l insulin detectably increased tyrosine phosphorylation of the IGF-1 receptor β subunit, although 20 nmol/l IGF-1 clearly increased the phosphorylation (Fig. 3C). Thus, the results indicated that the increased tyrosine phosphorylation of the 90- to 95-kDa proteins observed in GH-treated cells was due at least in part to tyrosine-phosphorylated STAT5b, not IGF-1 receptor, and that chronic GH treatment does not affect the basal or insulin-stimulated tyrosine phosphorylation of the IR-β subunit.

We next determined the effect of GH on the tyrosine phosphorylation of IRS-1 and the association of the p85 subunit of PI 3-kinase with IRS-1. The whole-cell lysates were immunoprecipitated with anti–IRS-1 antibody, followed by immunoblotting with either anti-phosphotyrosine or anti-p85 antibody (Fig. 4). As shown in Fig. 4A (upper panel, lane 2) and B, chronic GH treatment alone increased tyrosine phosphorylation of IRS-1. Furthermore, chronic GH treatment significantly enhanced the insulin-stimulated tyrosine phosphorylation of IRS-1 (Fig. 4A [upper panel] and B). Similarly, the association of p85 with IRS-1 was increased by GH treatment alone (Fig. 4A [middle panel, lane 2] and C), and the insulin-stimulated association of p85 with IRS-1 was significantly enhanced in GH-treated cells (Fig. 4A [middle panel] and C). IRS-1 protein levels were not significantly affected by chronic GH treatment, although a slight retardation of the electrophoretic mobility was observed (Fig. 4A, lower panel).

We then determined PI 3-kinase activity in the anti–IRS-1 antibody immunoprecipitates. Chronic GH treatment alone increased IRS-1–associated PI 3-kinase activity 1.7-fold, and GH significantly enhanced insulin-stimulated PI 3-kianse activity by 45, 39, and 25% at insulin concentrations of 0.2, 2.0, and 20 nmol/l, respectively (Fig. 5A).

Because the studies above indicated that chronic GH treatment increases insulin-stimulated PI 3-kinase activity but decreases glucose transport, which is regulated by PI 3-kinase, we next examined the effect of GH on the activation of Akt, a Ser/Thr kinase that is also regulated downstream of PI 3-kinase (18,19). Activation of Akt was assessed by immunoblotting with anti-phosphospecific antibodies directed against Thr308 and Ser473, the two regulatory phosphorylation sites of Akt (27). The results showed that phosphorylation of Akt at both of these sites was increased by chronic GH treatment alone (Fig. 5B, lane 2). However, the insulin-stimulated phosphorylation of these sites was significantly reduced in GH-treated cells (Fig. 5B). Immunoblotting with nonphosphospecific Akt antibody showed that chronic GH treatment did not affect protein levels of Akt (Fig. 5B).

GH stimulates autophosphorylation and tyrosine kinase activity of JAK2, leading to tyrosine phosphorylation of STATs and IRS-1 (35,36,42,43). To evaluate the role of JAK2 in the observed effects of chronic GH treatment, we compared the time course of JAK2, STAT5b, and pp185 in tyrosine phosphorylation, activation of Akt, and uptake of 2-DOG. GH rapidly increased the autophosphorylation of JAK2, tyrosine phosphorylation of STAT5b and pp185, and phosphorylation of Akt at Ser473 at 5–20 min after stimulation (Fig. 6). These acute effects of GH were accompanied by an increase in 2-DOG uptake by GH alone (Fig. 6C), which gradually decreased thereafter (Fig. 6). After chronic GH treatment (16 and 24 h), insulin-stimulated 2-DOG uptake was significantly decreased (P < 0.05, n = 3) (Fig. 6C), and pp185 tyrosine phosphorylation and Akt phosphorylation were significantly increased again (P < 0.05, n = 3) compared with the values before GH treatment; however, JAK2 autophosphorylation declined almost to basal levels, and STAT5b tyrosine phosphorylation decreased to ∼30% of the maximal level (Figs. 6A and B).

To determine whether the inhibitory effects of GH on the insulin-stimulated glucose transport and Akt activation are specific to insulin signaling, we next examined the effects of chronic GH treatment on 2-DOG uptake and Akt activation induced by other stimuli. As described previously (48), adenovirus-mediated expression of p110^CAAX, a membrane-targeted form of the catalytic subunit of PI 3-kinase, stimulated 2-DOG uptake (Fig. 7A) and phosphorylation of Akt at Ser473 (Fig. 7B) in a viral dose–dependent manner. PDGF stimulated phosphorylation of Akt in a dose-dependent manner (Fig. 7C). Chronic GH treatment did not affect the p110^CAAX-induced 2-DOG uptake (Fig. 7A) or the phosphorylation of Akt induced by p110^CAAX (Fig. 7B) or PDGF (Fig. 7C), although chronic GH treatment alone increased phosphorylation of Akt (Figs. 7B and C), as described above (Figs. 5B and 6). Immunoblotting with nonphosphospecific Akt antibody showed that expression of p110^CAAX or stimulation with PDGF did not significantly affect protein levels of Akt (data not shown).

It has been reported that stimulation with growth factors elicits translocation of Akt from the cytosol to the PM, which is suggested to be one of the necessary processes for the activation of Akt (27,28). To examine whether the decrease of the insulin-stimulated activation of Akt in GH-treated cells results from impaired translocation of Akt to the PM, we studied the subcellular localization of Akt. In control cells, most of the Akt was localized in the cytosol (Fig. 8). After insulin stimulation, a marked increase in Akt in the PM was observed (Fig. 8), reflecting insulin-stimulated translocation of Akt. Akt levels in other fractions (including the cytosol) were not significantly changed (Fig. 8), presumably because of a large pool of Akt in the cytosol (28,49). Chronic GH treatment alone increased Akt translocation to the PM to a small extent (Fig. 8), consistent with the small increase in Akt activation by GH alone. The insulin-stimulated translocation of Akt to the PM was significantly reduced (∼40%) in GH-treated cells (Fig. 8). In parallel, activation of Akt in each fraction was reduced in GH-treated cells (Fig. 8).

To test the possibility that altered subcellular localization of the insulin-signaling components may be responsible for the dissociation of the increased insulin-stimulated PI 3-kinase activity and the downstream effects (including glucose transport and Akt activation), we next determined the subcellular distribution of the observed increases in tyrosine phosphorylation of IRS-1, association of p85 with IRS-1, and IRS-1–associated PI 3-kinase activity in GH-treated cells. Insulin stimulation for 10 min resulted in increased tyrosine phosphorylation of IRS-1 in both the LDM and the cytosol (Fig. 9A). Chronic GH treatment alone also increased tyrosine phosphorylation of IRS-1 in both fractions to a smaller extent (Fig. 9A). Chronic GH treatment enhanced insulin-stimulated tyrosine phosphorylation of IRS-1 in the LDM (Fig. 9A, lane 4) and the cytosol (Fig. 9A, lane 8) by 42 and 32%, respectively, compared with the tyrosine phosphorylation of IRS-1 stimulated by insulin alone in each fraction. IRS-1–associated p85 was increased in both fractions in response to insulin (Fig. 9B), and chronic GH treatment alone also increased it in both fractions (Fig. 9B). Insulin-stimulated association of p85 with IRS-1 was enhanced by GH, with higher levels of association in the cytosol (79%) than in the LDM (12%) (Fig. 9B). Similarly, insulin-stimulated IRS-1–associated PI 3-kinase activity was enhanced by GH to a larger extent in the cytosol (57%) than in the LDM (21%) (Fig. 9C). Interestingly, although GH alone increased IRS-1–associated PI 3-kinase activity in both fractions (Fig. 9C), the extent of the increase for PI 3-kinase activity was less than that for the association of p85 with IRS-1 (Fig. 9B).

It is well known that GH produces insulin resistance in humans and animals (30,31,32,33,34). However, the molecular mechanism of GH-induced insulin resistance has not been clarified. Although decreased IR tyrosine kinase activity was noted in animals after chronic GH administration (33,34), such a change could be secondary to the altered metabolic state associated with insulin resistance. In this study, we attempted to clarify the primary cause of GH-induced insulin resistance. To the best of our knowledge, this is the first report to address the mechanism of GH-induced insulin resistance by focusing on the direct effects of chronic GH treatment on the insulin signaling pathway in insulin target cells.

The present study demonstrated that chronic GH treatment of 3T3-L1 adipocytes reduces insulin-stimulated glucose transport and Akt activation, both of which are downstream effects of PI 3-kinase (12,13,18,19), despite enhanced IRS-1–associated PI 3-kinase activity. Both basal and insulin-stimulated tyrosine phosphorylation of IRS-1 were increased by GH, which appears to be responsible for the increased association of p85 with IRS-1 as well as IRS-1–associated PI 3-kinase activity. Therefore, the tyrosine residues involved in the increased tyrosine phosphorylation of IRS-1 appear to interact with the SH2 domains of p85, resulting in increased activation of PI 3-kinase. However, because the increase in IRS-1–associated PI 3-kinase activity was relatively small compared with the increase in the association of p85 with IRS-1 (Fig. 9), the increased tyrosine phosphorylation of IRS-1 by GH may involve tyrosine residues, which are different from the sites phosphorylated by insulin, so that the two SH2 domains in p85 may not simultaneously interact with the phosphorylated tyrosine residues, and p110 may not be fully activated. The increased IRS-1 tyrosine phosphorylation was not caused by augmented IR tyrosine kinase activity because the increased tyrosine phosphorylation of 90- to 95-kDa proteins observed in the cells stimulated with or without submaximal concentrations of insulin was at least partly attributable to tyrosine-phosphorylated STAT5b, and thus IR autophosphorylation was not affected by GH. The possibility that tyrosine phosphorylation is generally increased in GH-treated cells can also be excluded because neither tyrosine phosphorylation of the basal or PDGF-stimulated PDGF receptor (data not shown) nor autophosphorylation of the IR-β subunit were increased by GH.

Our time course study indicated that the rapid activation of JAK2 by GH (which mediates tyrosine phosphorylation of IRS proteins and activation of Akt) coincided with the rapid insulin-like effect of GH on 2-DOG uptake, as has been previously reported (35,36,42,43). However, activation of JAK2 does not seem to directly mediate GH-induced insulin resistance because tyrosine phosphorylation of pp185 and phosphorylation of Akt declined and then increased again during the course of chronic GH treatment, and insulin-stimulated 2-DOG uptake decreased only after prolonged GH treatment, despite continuous decreases in JAK2 autophosphorylation and tyrosine phosphorylation of STAT5b. Therefore, the effects of chronic GH treatment on insulin signaling seem to be mediated by alterations in gene expression. A likely candidate for the mediator is IGF-1, which may be produced by the GH-treated cells and may stimulate IRS-1 tyrosine phosphorylation through activation of IGF-1 receptors (51,52,53). However, our data indicated that IGF-1 may not be involved in the increased IRS-1 tyrosine phosphorylation because chronic GH treatment did not stimulate autophosphorylation of the IGF-1 receptor. Furthermore, neutralization of IGF-1 with anti–IGF-1 antibody did not prevent increases in tyrosine phosphorylation of IRS-1 and phosphorylation of Akt during chronic GH treatment (data not shown). We also failed to detect IGF-1 in the culture media after chronic GH treatment (data not shown). Further studies are required to identify the mediator of GH-induced alterations in insulin signaling.

It can be postulated that the uncoupling between PI 3-kinase and its downstream signals observed in GH-treated cells results from inhibition of the downstream elements of the PI 3-kinase pathway. For example, reduced activity or altered localization of PDK-1 would impair activation of Akt and other kinases regulated by PDK-1 (29,54,55,56,57). Depletion of substrates for PI 3-kinase or increased lipid phosphatase activity would reduce 3′-phosphoinositides produced by PI 3-kinase, thereby attenuating downstream effects. In addition, increased Ser/Thr phosphatase activities could result in dephosphorylation and inactivation of Akt and other signaling molecules (58). However, the current data argue against these possibilities because chronic GH treatment did not affect 2-DOG uptake and Akt activation induced by p110^CAAX or Akt activation stimulated by PDGF, which are assumed to use the same downstream elements of PI 3-kinase as those used by insulin (48). The failure of GH to inhibit p110^CAAX-induced 2-DOG uptake also excludes the possibility that chronic GH treatment affects glucose transport machinery. This notion is supported by the observation that chronic GH treatment did not change the GLUT4 protein level (data not shown). These results, therefore, indicate that the insulin-stimulated PI 3-kinase pathway is specifically inhibited in GH-induced insulin resistance.

Unlike other growth factor receptors, IR uses intracellular substrate proteins, such as IRS-1, to recruit and activate SH2 domain–containing signaling molecules that transmit downstream signals (6,7,8,9,10,11,12,13). Therefore, the specific inhibition of the insulin-stimulated PI 3-kinase pathway could result from impairment in the activation process of IRS–associated PI 3-kinase. Growth factors, such as PDGF, activate PI 3-kinase, yet the majority of previous reports indicates that they have almost no effect on glucose transport (59,60,61). Furthermore, PDGF is not as effective as insulin in activating Akt (62,63,64,65). Because insulin-stimulated PI 3-kinase activation occurs largely in the LDM, whereas PDGF stimulates PI 3-kinase in the PM, appropriate localization of IRS-1–associated PI 3-kinase in a particular intracellular membrane compartment may be important in eliciting insulin-specific effects, such as GLUT4 translocation (49,50,66,67). Consistent with this hypothesis, insulin stimulation elicits translocation of IRS-1–associated PI 3-kinase from the LDM to the cytosol, which may play a role in the termination of insulin signals (49,50). In the cells treated with GH alone, tyrosine phosphorylation of IRS-1 was increased in both the LDM and in the cytosol, an effect similar to that of insulin, and it appears to have accounted for the increased Akt activation by GH alone through increased association and activation of PI 3-kinase. However, the finding that chronic GH treatment by itself did not increase 2-DOG uptake suggests that the increased PI 3-kinase activity in GH-treated cells is qualitatively different from that induced by insulin. Such a difference might be caused by the targeting of IRS-1–associated PI 3-kinase to a distinct subcompartment in the intracellular membrane.

The enhancement of insulin-stimulated association of p85 with IRS-1 and IRS-1–associated PI 3-kinase activity by chronic GH treatment was significantly larger in the cytosol than in the LDM. The preferential localization of IRS-1–associated p85 and PI 3-kinase activity in the cytosol in GH-treated cells does not seem to be caused by preferential increases in the tyrosine phosphorylation of cytosolic IRS-1 because insulin-stimulated tyrosine phosphorylation of IRS-1 was increased by GH in both the LDM and the cytosol to a similar extent. In addition, GH treatment alone did not significantly affect IRS-1 distribution, and insulin-induced translocation of IRS-1 from the LDM to the cytosol was not accelerated by chronic GH treatment (data not shown). Therefore, although the precise mechanism is unknown, the preferential distribution of insulin-stimulated PI 3-kinase in the cytosol by chronic GH treatment seems to result from increased binding or decreased dissociation between p85 and the tyrosine-phosphorylated IRS-1 in the cytosol.

Our data indicated that chronic GH treatment impairs insulin-induced translocation of Akt from the cytosol to the PM, suggesting that insulin-stimulated production of 3′-phosphoinositides in the PM may be decreased. Such impaired production of 3′-phosphoinositides at a particular intracellular location may result from altered targeting of activated PI 3-kinase. Although we failed to find a decrease in IRS-1–associated PI 3-kinase in a particular fraction, the preferential increase of insulin-stimulated IRS-1–associated PI 3-kinase in the cytosol may reflect changes in its localization in the subcompartment of the intracellular membrane. Alternatively, the increased PI 3-kinase activity in the cytosol might somehow inhibit the function of PI 3-kinase in other membrane compartments.

In conclusion, these findings indicate that chronic GH treatment induces cellular insulin resistance in 3T3-L1 adipocytes by uncoupling between the insulin-stimulated PI 3-kinase and its downstream effects, including glucose transport and Akt activation through expression of certain gene(s). The GH-induced uncoupling between PI 3-kinase and the downstream signals is specific to the insulin-stimulated PI 3-kinase pathway, and it might be caused by the altered subcellular distribution of IRS-1–associated PI 3-kinase.

This study was supported in part by grant-in-aids for Scientific Research from the Ministry of Education, Science, Sports, and Culture of Japan (10671060 and 12671103 to T.H.).