Enhanced Basal Activation of Mitogen-Activated Protein Kinases in Adipocytes From Type 2 Diabetes: Potential Role of p38 in the Downregulation of GLUT4 Expression

Specimens of human subcutaneous adipose tissue were obtained from the abdominal region of nondiabetic and type 2 diabetic subjects (Table 1). Subjects were recruited from the Unit of Endocrinology and Diabetes at University Hospital San Martin (Corrientes, Argentina) and Northwestern University Medical School (Chicago, IL). Subjects signed an informed consent document indicating that they understood the risks and benefits of participation in this experiment. The experimental procedures were approved by the appropriate scientific review committees at University Hospital Corrientes and Northwestern University Medical School.

Adipose cells were prepared as described previously (7,24). Briefly, the adipocytes were isolated by digesting ∼0.6 g tissue for 50 min at 37°C in medium 199 containing 25 mmol/l HEPES, 4% BSA, and 5.5 mmol/l glucose and collagenase (Sigma) at 0.8 mg/ml in a shaking water bath. Adipocyte cell size and number were determined as described previously (24). Isolated human adipocytes were distributed into plastic vials (12–15% cell suspension) in a final incubation volume of 400 μl. Cells were preincubated with or without 6.9 nmol/l insulin for 15 min, immediately separated by centrifugation through silicon oil, lysed in 0.4 ml lysis buffer containing 25 mmol/l Tris-HCl (pH 7.4), 0.5 mmol/l EGTA, 25 mmol/l NaCl, 1% Nonidet P-40, 1 mmol/l Na₃VO₄, 10 mmol/l NaF, 0.2 mmol/l leupeptin, 1 mmol/l benzamidine, and 0.1 mmol/l 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride, and rocked for 40 min at 4°C. Detergent-insoluble material was sedimented by centrifugation at 12,000g for 10 min at 4°C. Protein concentration of the cell lysates was determined by bicinchoninic acid assay (Pierce Chemical, Rockford, IL). This insulin concentration (6.9 nmol/l) has been previously determined to achieve maximal insulin stimulation of human adipocytes (7).

To minimize any researcher bias, the researchers were blinded to the identity of the subjects from whom samples were taken. Total protein (10 μg) was separated on 10% acrylamide SDS-PAGE gels (BioRad) and transferred to nitrocellulose (Millipore). Membranes were blocked in 4% nonfat dry milk in TBS-T (Tris-buffered saline with 0.5% Tween-20) (Sigma), and primary antibodies were diluted in either 4% nonfat dry milk/TBS-T or 0.1% BSA/TBS-T. The blots were probed with various primary antibodies: phospho-specific antibodies toward Erk 1/2 and Jun kinase (JNK) (Promega, Madison, WI) phospho-specific antibodies to p38 MAP kinase (Cell Signaling Technologies, Beverly, MA); Erk 1/2, JNK, and p38 independent of phosphorylation state (Upstate Biotechnology, Lake Placid, NY; Santa Cruz Biotechnology, Santa Cruz, CA; and Cell Signaling Technologies, respectively); IRS-1, IRS-2, and the p85 regulatory subunit of PI3 kinase (whole antiserum) (Upstate Biotechnology); and GLUT4 (Chemicon, Temecula, CA). Membranes were washed thoroughly and incubated with donkey anti-rabbit antibodies conjugated to horseradish peroxidase (Amersham). After washing, protein bands were visualized using enhanced chemiluminescence (ECL) (Amersham) and exposure to Hyperfilm ECL X-ray film (Amersham). Densitometric analysis of the bands was performed using a laser densitometer (PDSI; Amersham) and analyzed by ImageQuant software (Amersham).

The half-maximal inhibitory concentration (IC₅₀) toward p38 and JNK of the compounds used in this experiment were determined using in vitro kinase assays. The new compounds (A-291077 and A-304000) are aza-azulene–based compounds designed to inhibit p38 MAP kinase. Recombinant p38 (α) or JNK were activated by incubation at 30°C for 1 h with recombinant MKK6 (4 mU) in 25 mmol/l HEPES (pH 7.4), 25 mmol/l β-glycerophosphate, 1 mmol/l Na₃VO₄, 1 mmol/l DTT, 25 mmol/l MgCl₂, 5 mmol/l EDTA, 1 mmol/l NaF, and 50 μmol/l ATP. After activation of the kinase, activated p38 was incubated with 8 μg recombinant TRDH-ATF2 in 25 mmol/l HEPES (pH 7.4), 25 mmol/l β-glycerophosphate, 1 mmol/l Na₃VO₄, 1 mmol/l dithiothreitol (DTT), 25 mmol/l MgCl₂, 5 mmol/l EDTA, and 1 mmol/l NaF. Reactions were started with the addition of [³³P]ATP (10 μmol/l final; Amersham). Kinase reactions were allowed to proceed at 30°C for 1 h, after which the reactions were stopped with 5% phosphoric acid. The samples were transferred to filtration plates prewetted with 1% phosphoric acid and incubated for 15 min at room temperature. Vacuum filter plates were washed three times each with 1% phosphoric acid and counted in Microscint-20 on a Packard Topcount instrument.

3T3-L1 cells were grown and maintained as fibroblasts in Dulbecco’s modified Eagle’s medium (DMEM) (Life Technologies, Gaithersburg, MD) with high glucose containing 10% FBS (HyClone, Logan, UT) in a humidified atmosphere composed of 95% air and 5% CO₂. Cells were differentiated into adipocytes by exposure to DMEM high glucose with 10% FBS, 0.4 μg/ml dexamethasone, 0.5 mmol/l isobutylmethylxanthine, and 5 μg/ml insulin. After 3 days, the medium was changed and the cells were maintained with DMEM containing 10% FBS. Day 10 adipocytes were starved for 2 h in DMEM/0.1% FBS before incubation in DMEM/0.1% FBS with or without insulin (1 μmol/l) for the time specified. In some cases, the cells were pretreated for 30 min before exposure to insulin with various reagents including SB203580, PD95085, A-291077, and A-304000. At the appropriate time, medium was aspirated from the cells and the cells were washed with ice-cold PBS. Subsequently, cells were lysed in lysis buffer [25 mmol/l Tris-HCl (pH 7.4), 0.5 mmol/l EGTA, 25 mmol/l NaCl, 1% Nonidet P-40, 1 mmol/l Na₃VO₄, 10 mmol/l NaF, 0.2 mmol/l leupeptin, 1 mmol/l benzamidine, and 0.1 mmol/l 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride] (Calbiochem, La Jolla, CA) and rocked for 30 min at 4°C. Detergent-insoluble material was sedimented by centrifugation at 12,000g for 10 min at 4°C. Protein concentration of the cell lysates was determined by bicinchoninic acid assay (Pierce Chemical). Protein (20 μg) was separated on 7.5% SDS-PAGE gels (BioRad, Hercules, CA) and transferred to nitrocellulose. Western blotting was performed as described above.

Data are presented as means ± SD. Statistical significance was set at P < 0.05. Statistically significant differences in subject characteristics between healthy control subjects and type 2 diabetic patients were detected with an unpaired Student’s two-tailed t test. Differences in phosphorylation status of the MAP kinases in the basal and insulin-stimulated state, as well as between healthy and type 2 diabetic patients, were determined using repeated measures ANOVA (JMP, version 3.1; SAS Institute, Cary, NC). A least significant difference test was used for posthoc analysis.

The type 2 diabetic patients examined in this study fit many of the previously described characteristics associated with obesity and type 2 diabetes, including BMI, blood glucose, and insulin concentrations (1,6–9). Type 2 diabetic patients had significantly greater blood glucose and blood insulin concentrations compared with healthy control subjects (P < 0.05) (Table 1). Although the difference was not statistically significant (P = 0.06), there was a tendency for adipocyte size to be greater in type 2 diabetic patients. BMI was greater in patients with type 2 diabetes compared with healthy control subjects. Because the control subjects were not matched for BMI, the following experiments cannot differentiate between the effects of obesity and type 2 diabetes on the parameters examined.

Insulin resistance in adipocytes is frequently characterized by diminished protein abundances of IRS-1 and GLUT4 (2–9). In the current report (Fig. 1), IRS-1 and GLUT4 protein abundances are decreased 70% and 50%, respectively, in adipocytes from type 2 diabetic patients compared with healthy control subjects. This decrement is similar in magnitude to previously reported comparisons of IRS-1 and GLUT4 protein abundances in adipocytes from type 2 diabetic patients and healthy control subjects (6–9).

In adipocytes from healthy humans, 15 min of maximal insulin stimulation resulted in 1.8- and 2.0-fold elevated phosphorylation of Erk 1/2 and JNK, respectively (Fig. 2). p38 phosphorylation in response to insulin was observed in only two of seven healthy subjects examined. Interestingly, insulin treatment resulted in the phosphorylation of only a subset of JNK isoforms, specifically the p46 isoforms in both human adipocytes and 3T3-L1 adipocytes (data not shown). Although previous experiments have demonstrated JNK to be activated in response to insulin (17,18,20), this is the first report of an isoform-specific response. Furthermore, this is the first report to describe the basal phosphorylation status and insulin responsiveness of members of the MAP kinase family in human adipocytes.

In contrast to healthy adipocytes, Erk 1/2 and JNK phosphorylation was not increased by insulin in type 2 diabetic adipocytes. The failure of insulin to increase MAP kinase phosphorylation may have been due to an elevated basal level of phosphorylation, which was significantly greater than in untreated healthy adipocytes. The most intriguing result, however, was the profound fivefold increase in p38 phosphorylation observed in the basal state of type 2 diabetic adipocytes. Insulin stimulation did not further increase p38 phosphorylation in type 2 diabetic adipocytes. The protein abundances of Erk 1/2, JNK, and p38 were not significantly different between healthy control subjects and patients with type 2 diabetes.

To gain further insight into the potential mechanisms by which basal activation of MAP kinases in adipocytes may contribute to the development of insulin resistance, we systematically tested the role of Erk 1/2, JNK, and p38 using a panel of specific inhibitors in insulin-resistant 3T3-L1 adipocytes. 3T3-L1 adipocytes can be made insulin resistant through chronic exposure to insulin and elevated glucose concentrations (25–27). This model of insulin resistance mimics the expression pattern of human adipocytes in that there is a decreased protein abundance of IRS-1 and GLUT4 (10,11,25–27).

The panel of inhibitors used in this experiment included the MEK inhibitor PD98059 (to inhibit the Erk 1/2 pathway) as well as the p38 inhibitor SB203580 (28). In addition, we used two new compounds (Table 2) that show inhibitory properties toward p38 and JNK (A-291077) or p38-alone (A-304000) (29). Table 2 presents the chemical structures and IC₅₀ toward p38 and JNK as determined with in vitro kinase assays. The compound A-304000 is a selective inhibitor of p38, with little inhibitory activity toward JNK. Furthermore, these new compounds, specifically A-304000, exhibited minimal inhibitory activity in in vitro kinase reactions using AKT/PKB, PKA, MAPKAP2, casein kinase2, PKCγ, AMP kinase, c-Met, IGF receptor, Polo-like kinase, and LCK kinase. Thus the pattern of inhibition achieved by this panel would test the contribution of Erk 1/2 (PD98059), JNK, and p38 (A-291077) and p38-alone (A-304000) in the downregulation of GLUT4 expression and/or IRS-1 protein abundance in a cell culture model of insulin resistance. The overlapping specificity of A-291077 and A-304000 would allow us to infer the role that either JNK or p38 may play in the insulin-stimulated loss of IRS-1 and/or GLUT4.

Chronic exposure to insulin in a high-glucose environment significantly decreased protein levels of IRS-1 and GLUT4, by 50% and 40%, respectively (Fig. 3). This effect was restricted to IRS-1 and GLUT4, as insulin exposure had no effect on the levels of p85 PI3 kinase or IRS-2 (data not shown) as previously reported (11). Treatment of the 3T3-L1 adipocytes with selective inhibitors of Erk 1/2, JNK, or p38 did not prevent the insulin-stimulated loss of IRS-1.

The insulin-stimulated loss of GLUT4 protein abundance was prevented only by inhibitors of p38 (Fig. 3). Chronic exposure of the 3T3-L1 adipocytes to insulin resulted in a significant decrease in GLUT4 protein levels. Treatment of the 3T3-L1 adipocytes with p38 inhibitors without insulin stimulation did not significantly alter the protein abundance of GLUT4. Treatment with 10 μmol/l of the p38 selective inhibitors SB203580 or A-304000 did not prevent the insulin-stimulated downregulation of GLUT4. However, 50 μmol/l of SB203580 or A-304000 resulted in a GLUT4 protein level that was not statistically different from that of untreated/unstimulated cells or cells treated with 50 μmol/l of compound without insulin. These data indicate that inhibition of p38 is sufficient to prevent the insulin-stimulated downregulation of GLUT4. Pretreatment of the 3T3-L1 adipocytes with A-291077, an inhibitor of both JNK and p38, resulted in attenuation of the insulin-stimulated loss of GLUT4. This effect was observed at 10- and 50-μmol/l concentrations of compound. By contrast, inhibition of the Erk 1/2 pathway with the MEK inhibitor PD98059 had no effect on the insulin-stimulated loss of GLUT4 protein.

A-304000 prevented the insulin-induced loss of GLUT4 protein abundance in a dose-responsive manner (Fig. 4), with a >40% recovery of GLUT4 protein abundance observed with 25 μmol/l and a full recovery at 50 μmol/l. Taken together, these results indicate that inhibition of p38, but not Erk 1/2, can prevent the insulin-induced downregulation of GLUT4.

Insulin stimulation resulted in an increased p38 phosphorylation by approximately twofold at 5 and 10 min, before returning to baseline for the remainder of the experiment (Fig. 5A). Inhibition of p38 with A-304000 (Fig. 5B) or SB203580 (data not shown) prevented the insulin-induced p38 phosphorylation at 5 and 10 min of insulin stimulation. p38 protein levels were unaffected by insulin stimulation or the inhibitors.

This report is the first to describe the phosphorylation status and insulin responsiveness of the MAP kinases in human adipocytes from healthy control subjects and patients with type 2 diabetes. Members of the MAP kinase family have the potential to contribute to the perturbations in insulin signaling associated with type 2 diabetes. Thus we hypothesized that if one or more of the MAP kinases examined were involved with type 2 diabetes, there should be differential phosphorylation status of that particular MAP kinase between healthy adipocytes and type 2 diabetic adipocytes. In this report, we have demonstrated that basal MAP kinase phosphorylation, especially that of p38, is elevated in adipocytes from type 2 diabetic patients, as in previous experiments that have found elevated serine kinase activity in adipocytes from insulin-resistant animals (12), and our results are consistent with MAP kinases contributing to type 2 diabetes. The increased basal phosphorylation was most striking for p38, which exhibited a fivefold increase in adipocytes from type 2 diabetic patients compared with healthy control subjects. In contrast, Erk 1/2 and JNK phosphorylation was increased approximately twofold in adipocytes from type 2 diabetic patients compared with healthy control subjects.

The elevated basal phosphorylation state of members of the MAP kinase family in adipocytes from type 2 diabetic subjects is associated with a reduction in IRS-1 and GLUT4 protein levels. In humans, diminished GLUT4 and IRS-1 protein abundance can predict the future development of type 2 diabetes (8,9), suggesting that dysregulation of the expression of these proteins may be an early step toward the development of type 2 diabetes. To begin to understand the mechanism by which activation of the MAP kinases might contribute to insulin resistance and type 2 diabetes, we systematically tested the role of these pathways in regulating IRS-1 and GLUT4 protein levels in a cell-culture model of insulin resistance. We used a panel of inhibitors that selectively inhibited Erk 1/2 (PD98059), p38 alone (A-304000), and p38 and JNK (A-291077) to test the contribution of all three MAP kinase pathways toward insulin-stimulated downregulation of IRS-1 and GLUT4. The insulin-stimulated loss of IRS-1 was not prevented by inhibition of any of the MAP kinase pathways examined, consistent with a prior publication that demonstrated that PD98059 has no effect on the insulin-stimulated loss of IRS-1 (11). These results expand on previous findings to rule out JNK and p38 as playing a role in IRS-1 degradation. Inhibition of mTOR with rapamycin has been previously shown to prevent serine phosphorylation and subsequent degradation of IRS-1 (11). Although these results indicate that these MAP kinases do not contribute to insulin-induced IRS-1 degradation, we cannot rule out the possibility that they may be involved with serine phosphorylation of IRS-1, potentially leading to the development of insulin resistance.

The insulin-induced loss of GLUT4 protein abundance could be prevented by selective inhibitors of p38. This effect was also observed when both p38 and JNK were inhibited by A-291077, whereas inhibition of the Erk 1/2 pathway had no effect. Because the inhibition of both p38 and JNK simultaneously prevented the insulin-stimulated downregulation of GLUT4 in a manner similar to inhibition of p38 alone, we conclude that p38 is a major contributor to the insulin-stimulated loss of GLUT4. Although we cannot completely rule out the possibility that JNK and/or other kinases may be involved, our results are consistent with the findings of Fujishiro et al. (30), who demonstrated that constitutive activation of the MKK6-p38 pathway, but not the MKK7-JNK pathway, was sufficient for the downregulation of GLUT4 expression.

The role of p38 in the insulin-stimulated downregulation of GLUT4 is further supported by our use of two structurally distinct inhibitors, SB203580 and A-304000. These selective p38 inhibitors prevented the insulin-stimulated downregulation of GLUT4 with similar potencies. Although treatment with 50 μmol/l SB203580 resulted in decreased protein levels of IRS-1, IRS-2, and p85 regardless of insulin stimulation, treatment did not decrease GLUT4 protein levels and prevented the insulin-stimulated loss of GLUT4. This suggests that high doses of SB203580 may be toxic to the cells but selectively preserve GLUT4 expression, consistent with p38 acting as a negative regulator of GLUT4 expression. These cytotoxic effects appear to be limited to SB203580 itself and are not related to inhibition of p38 per se, because the other p38 inhibitors (A-304000 and A-291077) did not affect IRS-1, p85, or IRS-2 levels. Thus, our results expand on the findings of Fujishiro et al. (30) by demonstrating that the p38 pathway is necessary for the insulin-stimulated downregulation of GLUT4. Furthermore, our observation that p38 phosphorylation is increased in adipocytes isolated from type 2 diabetic patients strongly supports a role for p38 to contribute to insulin resistance and the development of type 2 diabetes.

The 3T3-L1 adipocyte model used in this experiment (chronic exposure to high insulin and glucose concentrations) mimics very closely the conditions found in insulin-resistant human adipocytes. For example, chronic exposure of 3T3-L1 adipocytes to insulin and glucose results in 1) diminished insulin-stimulated glucose uptake, 2) failure of insulin to inhibit lipolysis, and 3) the loss of IRS-1 and GLUT4 (present data; 11,31). These similarities support the use of the 3T3-L1 adipocyte to test the role of p38 in the regulation of IRS-1 and GLUT4. Thus we can infer that activation of p38 observed in type 2 diabetic adipocytes may contribute to the low levels of GLUT4 observed in those cells.

The mechanisms by which p38 may regulate GLUT4 are currently unknown. In adipocytes from type 2 diabetic patients, p38 phosphorylation was increased, in association with reduced GLUT4 and IRS-1. However, unlike the human type 2 diabetic adipocytes, insulin stimulation of 3T3-L1 adipocytes resulted in a transient increase in p38 phosphorylation. Inhibition of p38 was able to prevent the loss of GLUT4; thus only a transient activation of p38 may be necessary for downregulation of GLUT4. Downregulation of GLUT4 mRNA transcription in response to insulin occurs within 2 h (25); thus it is possible that early changes in p38 activity may be capable of modulating GLUT4 expression. GLUT4 mRNA and protein are decreased in adipose tissue from type 2 diabetic and animal models of insulin resistance (9,32,33) and are also diminished in 3T3-L1 adipocytes treated with insulin (25–27), tumor necrosis factor (TNF)-α (34), and other factors (35,36). It has not yet been determined if p38 regulates GLUT4 expression through modifications of mRNA stability or mRNA transcription; however, many transcription factors are known to be downstream of the p38 pathway (13), consistent with gene transcription as a potential mechanism.

Many regions within the GLUT4 promoter have been identified as contributing to the downregulation of GLUT4 expression in response to insulin (37), TNF-α (38), and other factors (36,39). Nuclear factor 1 (NF-1) and Olf-1/early B-cell factor (O/E) are capable of binding to an element within the GLUT4 promoter that mediates the insulin and cAMP-induced downregulation of GLUT4 transcription (40,41). Alternatively, CCAAT-enhancer binding proteins (C/EBP) have been shown to play important roles in regulating GLUT4 expression during adipogenesis (42). NF-1 becomes phosphorylated in response to insulin (43); however, there are no data regarding the ability of p38 to phosphorylate NF-1. p38 has been shown to be capable of phosphorylating C/EBP β (44). Thus the potential exists for the p38 pathway to mediate repression of GLUT4 expression at the transcriptional level through directly modifying factors such as NF-1 and C/EBP β. We are currently pursuing further experimentation to test these possibilities.

The current report has not addressed the nature of the stimulus that may be responsible for the basal activation of p38 in human adipocytes. Because the human adipocytes used in the current experiment were isolated from abdominal subcutaneous biopsies, and thus underwent a substantial isolation process before insulin stimulation and examination, it is likely that any systemic factors that may have contributed to the elevated p38 phosphorylation were removed before analysis. Thus the elevated phosphorylation of p38 is likely due to biochemical characteristics intrinsic to type 2 diabetes adipocytes. Factors such as cAMP (35), oxidative stress (36), and numerous cytokines and interleukins (34) have been shown to decrease GLUT4 protein in 3T3-L1 adipocytes and activate p38 in other cell systems (36 and references therein). Future experiments will be necessary to further define the cause of p38 activation in type 2 diabetic adipocytes.

In conclusion, the current report has identified p38 as a potential candidate for mediating the loss of GLUT4 expression observed in adipocytes from type 2 diabetic subjects. This is based on an elevated basal phosphorylation status of p38 MAPK in human adipocytes from type 2 diabetic patients compared to healthy control subjects. Furthermore, we have demonstrated that p38 is necessary for the insulin-induced loss of GLUT4 expression in 3T3-L1 adipocytes. This may be of great importance because diminished GLUT4 expression results in the development of insulin resistance, diabetes, and many related complications (4,5). In addition, the current experiment has ruled out the MAP kinases as contributing to the insulin-induced loss of IRS-1 protein abundance in 3T3-L1 adipocytes. Future work is necessary to define the mechanisms involved with the activation of p38 in insulin-resistant and diabetic adipocytes, as well as the connection between p38 activation and regulation of GLUT4 protein abundance.

C.J.C. was supported by an Abbott Laboratories/Northwestern University Post-Doctoral Fellowship in Diabetes.

The authors thank Dr. A. Montalvo of Northwestern University Medical School and Dr. Silvana Gross and members of the Department of Biochemistry, School of Medicine, National University of Northeast (Corrientes, Argentina) for their contributions to the project. Additionally, the authors thank Dr. E. Morphew and Dr. J. Trevillyan for constructive comments during preparation of the manuscript.