0013-7227/91/1285-2415$03.00/0 Endocrinology Copyright © 1991 by The Endocrine Society
Vol. 128, No. 5 Printed in U.S.A.
Glucose and Insulin Regulate Insulin Sensitivity in Primary Cultured Adipocytes without Affecting Insulin Receptor Kinase Activity* FABIO B. LIMAf, R. SCOTT THIES, AND W. TIMOTHY GARVEY Departmente Fisiologia e Biofisica, Institute Ciencias Biomedicas, Universidade de Sao Paulo (F.B.L.), Sao Paolo SP 05508, Brazil; the Genetics Institute (R.S.T.), Cambridge, Massachusetts 02140; and the Department of Medicine, Division of Endocrinology/Metabolism, Indiana University, and the Indianapolis Veterans Administration Medical Center (W.T.G.), Indianapolis, Indiana 46202
ABSTRACT. We have previously shown in primary cultured adipocytes that chronic insulin exposure decreases insulin's subsequent ability to maximally restimulate the glucose transport system, and that extracellular glucose potentiates this ligandinduced defect in maximal insulin responsiveness. To examine whether glucose could also modulate insulin sensitivity (i.e. acute insulin effects at submaximal concentrations), adipocytes were cultured for 5 and 24 h in the absence and presence of various glucose and insulin concentrations. Then, after washing cells to remove any insulin and allow for full deactivation of transport, we assessed the dose response of insulin's acute ability to stimulate 2-deoxyglucose transport, bind to cell surface receptors, and activate insulin receptor tyrosine kinase activity. After 5 h, glucose and insulin alone had no chronic regulatory effects; however, in combination, these agents were able to decrease insulin sensitivity. In cells preincubated with 50 ng/ml insulin, the insulin ED60 for acute stimulation of glucose transport was increased by 65% and 116% as medium glucose was raised to 5 and 20 mM, respectively, relative to that at 0 mM glucose. After 24 h, chronic exposure to either glucose (20 mM) or insulin (50 ng/ml) alone increased the ED50 value by 52%, and, together they acted synergistically to increase the ED50 by 183%. While glucose and insulin independently and synergistically impaired insulin sensitivity, both agents were necessary for coregulation of maximal insulin responsiveness (confirming our previous observation). Insulin receptor down-regulation (18%) was observed after 24 h (but not 5 h) in insulin-treated cells; however,
I
N Noninsulin-Dependent Diabetes Mellitus (NIDDM) and obesity, insulin resistance in adipocytes is characterized by a decrease in both maximal insulin responsiveness and insulin sensitivity of the glucose transport system (1-6). Classically, a decrement in
Received October 15,1990. Address all correspondence and requests for reprints to: W. Timothy Garvey, M.D., Section of Endocrinology, Indiana University School of Medicine, Veterans Administration Hospital 111-E, 1481 West 10th Street, Indianapolis, Indiana 46202. * This work was supported by NIH Grant DK-38765 and a grant from the V.A. Medical Service. t Recipient of a Fogarty International Research Fellowship Award from the NIH and a Postdoctoral Research Fellowship from the Sao Paulo State Research Foundation.
the major portion of the decrease in insulin sensitivity was due to uncoupling of occupied insulin receptors from stimulation of the glucose transport system. To further determine the mechanism for postbinding desensitization, we tested for concordant regulation of insulin receptor kinase activity. Insulin's ability to stimulate the receptor tyrosine kinase was assessed by multiple methods, including 1) receptor autophosphorylation and phosphorylation of Gh^-Tyr1 by solubilized insulin receptors activated in vitro, 2) histone-2B phosphorylation by receptors that were stimulated in intact cells and then solubilized under conditions that preserve the in celluto phosphorylation state, and 3) receptor autophosphorylation and phosphorylation of ppl80 in intact cells. Long term treatment (24 h) with glucose (10 mM) and insulin (50 ng/ml) markedly decreased insulin sensitivity (and receptor coupling), but did not affect insulin receptor kinase activity in any of these studies. We conclude that 1) glucose and insulin can independently and synergistically regulate insulin sensitivity by uncoupling ligand-bound insulin receptors from stimulation of glucose transport, and 2) glucose/insulin-induced desensitization occurred without any effect on insulin receptor tyrosine kinase activity. These observations suggest that decreased insulin sensitivity in target tissue may be induced by hyperglycemia and hyperinsulinemia in noninsulin-dependent diabetes mellitus via a receptor kinase-independent mechanism. (Endocrinology 128: 2415-2426, 1991)
maximal insulin responsiveness has been taken as evidence of a postreceptor defect, probably involving the hormone effector system (7), and we have, in fact, demonstrated cellular depletion of glucose transporter proteins in adipocytes isolated from these patients (8). On the other hand, a decrease in insulin sensitivity, defined as reduced effects at submaximal insulin concentrations, leading to a rightward shift in the insulin biological action dose-response curve, can be caused by insulin receptor down-regulation and/or an abnormality in the insulin signal coupling ligand binding to stimulation of the hormone effector system (7). In adipocytes from patients with NIDDM (1-3) and obesity (4, 5), receptor loss accounts for some, but not all, of the rightward shift
2415
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 July 2015. at 03:23 For personal use only. No other uses without permission. . All rights reserved.
2416
GLUCOSE AND INSULIN REGULATE INSULIN SENSITIVITY
in the insulin-glucose transport dose-response curve. The postbinding defect accounting for the remaining loss of insulin sensitivity has not been identified, but could involve a decrease in insulin receptor tyrosine kinase activity, since evidence suggests that insulin receptor kinase is involved in insulin signal transduction and couples insulin binding to stimulation of glucose transport (9-13). Furthermore, tyrosine kinase activity of solubilized insulin receptors from various tissues (14-17) and intact adipocytes (18-20) is decreased in NIDDM. We have previously suggested in NIDDM and obesity that hyperglycemia and/or hyperinsulinemia may directly induce insulin action defects in target tissue, based on our studies assessing the regulatory effects of glucose and insulin in primary cultured adipocytes (21-23). In these cells, chronic exposure to insulin per se decreased the ability of insulin to acutely stimulate glucose transport by sequentially acting at receptor and multiple postreceptor sites in the insulin action pathway (21). Furthermore, glucose was shown to potentiate insulin's ability to decrease maximal insulin responsiveness by impairing the translocation efficiency of intracellular glucose transporters to the cell surface (22). In the current study we assessed whether glucose could also modulate insulin sensitivity in cultured adipocytes. Secondly, we examined whether insulin's ability to decrease insulin sensitivity was due to effects on insulin receptor kinase activity. Thus, we hoped to gain some insight into potential factors that could impair insulin sensitivity in insulin-resistant disease states such as NIDDM and obesity.
Materials and Methods Materials Porcine monocomponent insulin and A-14-mono-[125I]insulin were generously supplied by Dr. Ronald Chance of Eli Lilly Co. (Indianapolis, IN). Antiphosphotyrosine antibody was kindly supplied by M. Kamps and B. Sefton (Salk Institute, San Diego, CA), and antiinsulin receptor antibodies by L. Mandarino (University of Pittsburg, Pittsburg, PA). L-[l-3H]Glucose was purchased form New England Nuclear (Boston, MA); 2-deoxy-D-[3H]glucose, [7-32P]ATP (3000 Ci/mmol), and [125I] protein-A from Amersham Corp. (Arlington Heights, IL); BSA (fraction V) from Armour Pharmaceutical Co. (Tarrytown, NY); collagenase from Worthington Biochemical Corp. (Freehold, NJ); Dulbecco's Modified Eagle's Medium (DMEM) from Gibco (Grand Island, NY); protein-A (formalin-fixed Staphylococcus aureus cells) from Bethesda Research Laboratories (Gaithersburg, MD); and CnBr-activated Sepharose 4B from Pharmacia (Piscataway, NJ). Histone-2B, poly (Glu4Tyr1), and all other reagents were obtained from Sigma (St. Louis, MO). Materials for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) were purchased from Bio-Rad (Richmond, CA).
Endo • 1991 Voll28»No5
Primary culture of adipocytes Isolated adipocytes were obtained from the epididymal fat pads of male Sprague-Dawley rats under sterile conditions and maintained in primary culture, as previously described (21-23). Adipocytes (5 X 104 cells/ml) were then incubated for 5 or 24 h at 37 C in the absence (control) and presence of various Dglucose and insulin concentrations. The presence of 2% fetal calf serum led to a final D-glucose concentration of 0.12 mM without further addition of D-glucose. Glucose transport assays After incubation, cells were washed three times in equal volumes of glucose- and insulin-free buffer, pH 7.4, containing 20 mM HEPES, 120 mM NaCl, 1.2 mM MgSO4) 2.0 mM CaCl2, 2.5 mM KC1, 1.0 mM NaH2PO4, 1.0 mM sodium pyruvate, and 1% BSA. The cells were then incubated in this buffer for an additional 40 min at 37 C to dissociate any remaining receptorbound insulin. These washing procedures effectively removed all extracellular and receptor-bound insulin and allowed the glucose transport system of insulin-stimulated cells to deactivate to basal levels. After washing, adipocytes were resuspended in the same buffer and incubated (1 ml) in the absence and presence of various insulin concentrations in a shaking water bath for 45 min at 37 C. Initial rates of 2-deoxyglucose uptake (substrate concentration, 0.2 mM) were measured over 3 min (2-3 X 105 cells/ml), as previously described (21), and the distribution space of radiolabeled L-glucose was used to correct for nonspecific carryover of radioactivity with the cells and the uptake of hexose by simple diffusion. 3-O-Methylglucose transport was assayed using a modification of the method of Whitesell and Gliemann (24), as previously described by Foley et al. (25). Measurement of insulin receptors Cells were washed to remove extracellular and receptorbound insulin, as described above, and then incubated (~2 X 105 cells/ml) in a shaking water bath at 16 C with 0.2 ng [125I] insulin and various concentrations of nonradioactive insulin. After 2 h the binding reaction was terminated, and specific [125I]insulin binding was determined, as previously described (21). Insulin binding to intact adipocytes was performed at 16 C, because insulin internalization is negligible at this low temperature (26); thus, cell-associated [125I]insulin essentially reflects binding only to cell surface receptors. Aliquots of wheat germ eluates (partially purified preparation of solubilized insulin receptors) were incubated with [I25I]insulin (final concentration, 0.5-1 ng/ml) at 4 C for 18 h in the presence or absence of increasing concentrations of unlabeled insulin, as previously described (18, 27). The amount of receptor-bound hormone was determined by previously described methods (28), and insulin-binding capacities of the receptor preparations were estimated by Scatchard analysis. Partial purification of insulin receptors and measurement of tyrosine kinase activity in vitro For the determination of receptor tyrosine kinase activity in vitro, cells were solubilized, and insulin receptors were partially purified over wheatgerm-agarose (27). Insulin-induced auto-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 July 2015. at 03:23 For personal use only. No other uses without permission. . All rights reserved.
GLUCOSE AND INSULIN REGULATE INSULIN SENSITIVITY phosphorylation of partially purified receptors and phosphorylation of the artificial kinase substrate Glu4-Tyr1 were performed as previously described (18, 27). Briefly, cultured adipocytes were disrupted and solubilized by agitating the cells in an ice-cold solubilization mixture containing 1.5% Triton X100, 7.5 mg/ml bacitracin, 15 mM benzamidine, 1500 kallikrein inhibitor units/ml aprotinin, and 2.5 mM phenylmethylsulfonylfluoride (PMSF), pH 7.4. The cellular extract was clarified by centrifugation, diluted 1:4 with buffer containing 25 mM HEPES (pH 7.4), 10% (wt/vol) glycerol, 0.05% Triton, 120 mM NaCl, 2.5 mM KC1, 1 mM CaCl2, 1 mM MgCl2, 15 mM benzamidine, 1500 kallikrein inhibitor units/ml aprotinin, and 2.5 mM PMSF, and then recycled twice over wheatgerm-agarose at 4 C. The glycoproteins were desorbed with 0.3 M iV-acetyl-Dglucosamine in the recycling buffer. Fractions containing insulin receptors were combined in each treatment group, and frozen at —70 C for use in subsequent phosphorylation and insulin binding studies. The protein content of the frozen aliquots was determined by the Bradford dye method (29), using crystalline BSA as the standard (Bio-Rad protein assay). Phosphorylation assays were performed as previously described (18, 27), using equal numbers of insulin receptors from control and treated cells. Receptors were preincubated in the absence and presence of various insulin concentrations for 18 h at 4 C, and then the phosphorylation reaction was initiated in a reaction mixture containing 5 mM MnCl2, 4 mM MgCl2, 500 MM CTP, and 50 MM [ T - 3 2 P ] A T P (11 MCi/nmol). The reaction was terminated in a solution containing 25 mM HEPES (pH 7.6), 0.05% Triton, 10% glycerol, 2 mM sodium orthovanadate, 100 mM NaF, 1 mM PMSF, 5 mM EDTA, 50 mM ATP, and 10 mM sodium pyrophosphate. The phosphorylated insulin receptors were immunoprecipitated with a 1:50 dilution of serum containing human antiinsulin receptor antibodies, as previously described (18), and then analyzed by SDSPAGE on 10% resolving gels according to the method of Laemmli (30). Under these conditions, 90-95% of the phosphorylated receptors were immunoprecipitated. The /3-subunit of the insulin receptor was visualized by autoradiography, and the degree of phosphorylation was quantitated by excising this band from the dried gel and counting in a scintillation counter. Equal areas of dried gel, judged free of radioactivity by autoradiography, were cut, counted, and subtracted as background. To assess phosphorylation of exogenous substrate, lectinpurified insulin receptors were preincubated with and without insulin as described above, and then reacted with the synthetic substrate Glu4-Tyr' (final concentration, 2 mg/ml), 5 mM MnCl2, 12 mM MgCl2, 500 MM CTP, and 50 fiM [7-32P]ATP (5 Ci/nmol). After incubation at 4 C for different intervals of time, the reaction was terminated by the addition of 50 mM ATP, and the 32P incorporated into Glu^Tyr1 was determined by the filter paper method, as previously described (18). Measurement of insulin receptor tyrosine kinase activity in intact cells The tyrosine kinase activity of receptors stimulated with insulin in intact adipocytes was estimated in two ways. By one approach, cells were incubated with and without insulin, insulin receptors were isolated under conditions designed to preserve the phosphorylation state found in the intact cells, and then
2417
receptor tyrosine kinase activity was assayed by its ability to phosphorylate histone-2B in vitro, as previously described (31). Adipocytes were preincubated with and without glucose/insulin, washed, and acutely stimulated with various insulin concentrations, as described above. Insulin receptors were then partially purified, as described above, but with the addition of phosphatase and kinase inhibitors to all buffers. Wheatgerm eluates were passed over insulin-Sepharose beads in order to purify and immobilize the insulin receptors. The phosphorylation reaction was carried out in a solution containing 5 mM MnCl2, 12 mM MgCl2, 500 nM cytidine triphosphate, 1 mg/ml histone, and 4 ^M [7-32P]ATP (80 AtCi/nmol). Previous detailed experiments have shown that at this low concentration of ATP, there is no additional activation of insulin receptor kinase activity as a result of in vitro manipulations subsequent to cell solubilization (31). Phosphorylated proteins were analyzed by SDS-PAGE on 15% gels according to the method of Laemmli (30), as previously described (31). In the second technique to measure receptor kinase activity in intact cells (20), adipocytes stimulated with and without insulin were solubilized in 2% SDS containing 2 mM Na3VO4 and 100 mM NaF to inhibit phosphatases. Cell extracts were electrophoresed on SDS-PAGE gels (10%), and the resolved protein bands were transferred to nitrocellulose for Western blotting. The nitrocellulose blot was incubated with polyclonal antiphosphotyrosine antibodies, followed by 125I-labeled protein-A, as described by Kamps and Sefton (32), to selectively label proteins containing phosphotyrosine. The 125I-labeled protein bands were visualized by autoradiography and quantitated by scanning densitometry. Titers of cell extract blotted with antiphosphotyrosine antibodies confirmed linearity between the amount of phosphoprotein and the density of autoradiographic images. Statistics Values are given as the mean ± SEM, and statistical significance was determined using Student's t test for paired and unpaired data where appropriate.
Results Glucose transport studies To assess the ability of glucose and insulin to chronically regulate the glucose transport system, isolated adipocytes were cultured with various glucose concentrations (0, 5, and 20 mM) in the absence and presence of 50 ng/ml insulin. After 5-24 h, the cells were washed to remove any insulin and allow full deactivation of the glucose transport system. The cells were then acutely restimulated with insulin, and 2-deoxyglucose uptake rates were determined. Our previous data (22), showing that glucose acted in combination with insulin to decrease basal and maximally stimulated glucose transport rates, were confirmed. The purpose of the current study was to test whether glucose and insulin were also able to regulate insulin sensitivity (i.e. acute stimulatory effects at submaximal insulin concentrations). Therefore, we
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 July 2015. at 03:23 For personal use only. No other uses without permission. . All rights reserved.
2418
GLUCOSE AND INSULIN REGULATE INSULIN SENSITIVITY
assessed the full dose-response curve for insulin-stimulated glucose transport and replotted transport rates as a percentage of the maximal insulin effect, as previously described (21). This allows optimal determination of the insulin ED50 as a quantitative measure of the degree to which the dose-response curve is right or left shifted. As is evident in Fig. 1, cells pretreated with either glucose (0, 5, or 20 mM) or insulin (50 ng/ml) alone for 5 h showed no change in insulin sensitivity (Fig. 1A). HowPreincubation Insulin Concentration 50 ng/ml 0 ng/ml
0.7
A: 5 hours preincubation i
0.6 0.5 0.4 c
0.3
O Q
0.2
UJ
I
0.1 0.7 w c 0) CO
B: 24 hours preincubation
0.6 0.5 0.4
ever, glucose and insulin in combination induced a rightward shift in the dose-response curve; glucose in the presence of insulin led to 61% and 116% increases in the ED50 values (i.e. a decrease in insulin sensitivity) as medium glucose was raised to 5 and 20 mM, respectively (P < 0.05). Additional regulatory effects were observed with more prolonged exposure to glucose and/or insulin for 24 h (Fig. IB). With longer term treatment, both glucose (20 mM) alone and insulin alone increased the insulin ED50 value by 52% (P < 0.05), and together interacted synergistically to cause a marked 183% increment in the ED50 compared to that in control cells (P < 0.01). Thus, glucose and insulin independently and synergistically regulated insulin sensitivity. We have previously shown (22) that D-glucosamine and D-mannose in addition to D-glucose could interact with insulin to diminish maximal insulin responsiveness, but that other compounds, such as D-fructose, sodium pyruvate, L-glucose, sucrose, and alanine, were inactive. In a similar manner, we examined substrate specificity for regulation of insulin sensitivity. Cells were cultured with and without the substrates listed in Table 1 (at 5 mM) in the presence of 50 ng/ml insulin. Then, after 24 h, the cells were washed, and the dose-response of insulin's acute ability to restimulate 2-deoxyglucose transport was assessed. Table 1 shows effects on the insulin ED50 as an index of insulin sensitivity. In this series of experiments, the ED50 in cells preincubated with insulin alone was further increased by coincubation with either D-mannose or D-glucose (P < 0.05). Thus, these agents greatly potentiated insulin's chronic ability to decrease insulin sensitivity and also led to reduced basal and maximally insulin-stimulated transport rates (data not TABLE 1. Ability of substrates to decrease insulin sensitivity of the glucose transport system
0.3 0.2 0.1
Endo • 1991 Voll28«No5
0
20
0
20
Preincubation D-glucose Concentration (mM) Fir,. 1. Glucose and insulin regulate insulin sensitivity in cultured adipocytes. Cells were cultured for 5 h (A) and 24 h (B) in medium containing 0, 5, and 20 mM glucose in the absence {left) and presence (rif>ht) of 50 ng/ml insulin, and then washed to remove any extracellular and receptor-bound insulin. After glucose transport was allowed to fully deactivate, dose-response curves for insulin's (0-50 ng/ml) acute ability to stimulate 2-deoxyglucose transport were assessed. The insulin ED50 was calculated as follows. The basal 2-deoxyglucose transport rate was subtracted from all insulin-stimulated transport rates, and the value at each insulin concentration was expressed as a percentage of the maximal insulin increment. The insulin ED5o, a measure of insulin sensitivity, was then determined from individual dose-response curves as the concentration yielding 50% of the maximal effect. Please note that the ordinate in this figure begins at 0.1. Data represent the mean ± SE of three to six experiments.
Substrate present during preincubation (5 mM)
Insulin sensitivity, ED50 (ng/ml)
Effect on maximum insulin responsiveness
None D-Glucose D-Mannose D-Fructose Sodium pyruvate L-Glucose
0.31 ± 0.01 1.05 ± 0.12° 1.23 ± 0.15" 0.75 ± 0.08° 0.55 ± 0.06° 0.28 ± 0.09
Yes Yes No No
No
Adipocytes were cultured for 24 h at 37 C in the absence (none) and presence of the indicated substrates (5 mM) and in the presence of 50 ng/ml insulin. All cells were then washed in insulin- and substrate-free buffer, and glucose transport rates were allowed to fully deactivate. We then assessed the dose response of insulin's ability (0-50 ng/ml) to acutely restimulate 2-deoxyglucose transport. The insulin ED60 value was calculated from individual dose-response curves, as described for Fig. 1. Insulin responsiveness denotes effects on the maximal glucose transport rate at 50 ng/ml insulin. Data represent the mean ± SE of three experiments. 0 P < 0.05 vs. controls with no added substrate.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 July 2015. at 03:23 For personal use only. No other uses without permission. . All rights reserved.
GLUCOSE AND INSULIN REGULATE INSULIN SENSITIVITY
shown), as previously reported (22). A different pattern was observed with D-fructose and sodium pyruvate; these compounds did not affect basal and maximal glucose transport rates (data not shown) (22), but were still able to potentiate insulin-induced decrements in insulin sensitivity. Specifically, the ED50 value was increased 2.4fold by D-fructose and 1.8-fold by sodium pyrvvate over that in cells treated with insulin alone (P < 0.05). In contrast, L-glucose, which is not metabolized and does not enter the cell via facilitative transporters, did not affect basal and maximal glucose transport rates and did not alter the insulin ED50 value (P = NS). In these studies, quantitatively similar effects on the insulin sensitivity (ED50) and insulin responsiveness of 2-deoxyglucose transport were observed using 3-O-methylglucose as the glucose analog in the transport assay (data not shown). Insulin binding studies Changes in insulin sensitivity could result from differences in cell surface insulin binding, a change in the coupling efficiency between insulin receptors and the biological effector system, or a combination of both. Figure 2 shows insulin competition binding curves to insulin receptors in cells cultured for 24 h in the presence
2419
of glucose (5 mM) plus insulin (50 ng/ml) and in controls. Glucose plus insulin led to an 18% decrease in insulin binding at trace concentrations of [125I]insulin without significantly affecting the IC50 value. This is generally consistent with a decrease in the number of insulin receptors and agrees with our previously reported studies of insulin-induced receptor down-regulation in primary cultured adipocytes (21). However, while receptor loss could be expected to impair insulin sensitivity in these cells, the relatively small decrease in insulin binding (~18%) could not by itself explain a 3-fold increase in the insulin ED50 value (see Fig. IB). The results indicate that the major mechanism accounting for the decrease in insulin sensitivity involves steps distal to ligand binding in the insulin action sequence. To examine whether the decrease in insulin sensitivity involved a decrease in insulin receptor-glucose transport system coupling efficiency, we plotted glucose transport rates as a percentage of the maximal insulin effect us. the actual amount of insulin bound to cell surface insulin receptors at each of the insulin concentrations used to construct the dose-response curve (Fig. 3). In other words, these plots assess the coupling efficiency between T
100 3.0
1
i
r
0 mM glucose No Insulin
80
2.5
i! £ I 60
2.0
Vol. 128, No. 5 Printed in U.S.A.
Glucose and Insulin Regulate Insulin Sensitivity in Primary Cultured Adipocytes without Affecting Insulin Receptor Kinase Activity* FABIO B. LIMAf, R. SCOTT THIES, AND W. TIMOTHY GARVEY Departmente Fisiologia e Biofisica, Institute Ciencias Biomedicas, Universidade de Sao Paulo (F.B.L.), Sao Paolo SP 05508, Brazil; the Genetics Institute (R.S.T.), Cambridge, Massachusetts 02140; and the Department of Medicine, Division of Endocrinology/Metabolism, Indiana University, and the Indianapolis Veterans Administration Medical Center (W.T.G.), Indianapolis, Indiana 46202
ABSTRACT. We have previously shown in primary cultured adipocytes that chronic insulin exposure decreases insulin's subsequent ability to maximally restimulate the glucose transport system, and that extracellular glucose potentiates this ligandinduced defect in maximal insulin responsiveness. To examine whether glucose could also modulate insulin sensitivity (i.e. acute insulin effects at submaximal concentrations), adipocytes were cultured for 5 and 24 h in the absence and presence of various glucose and insulin concentrations. Then, after washing cells to remove any insulin and allow for full deactivation of transport, we assessed the dose response of insulin's acute ability to stimulate 2-deoxyglucose transport, bind to cell surface receptors, and activate insulin receptor tyrosine kinase activity. After 5 h, glucose and insulin alone had no chronic regulatory effects; however, in combination, these agents were able to decrease insulin sensitivity. In cells preincubated with 50 ng/ml insulin, the insulin ED60 for acute stimulation of glucose transport was increased by 65% and 116% as medium glucose was raised to 5 and 20 mM, respectively, relative to that at 0 mM glucose. After 24 h, chronic exposure to either glucose (20 mM) or insulin (50 ng/ml) alone increased the ED50 value by 52%, and, together they acted synergistically to increase the ED50 by 183%. While glucose and insulin independently and synergistically impaired insulin sensitivity, both agents were necessary for coregulation of maximal insulin responsiveness (confirming our previous observation). Insulin receptor down-regulation (18%) was observed after 24 h (but not 5 h) in insulin-treated cells; however,
I
N Noninsulin-Dependent Diabetes Mellitus (NIDDM) and obesity, insulin resistance in adipocytes is characterized by a decrease in both maximal insulin responsiveness and insulin sensitivity of the glucose transport system (1-6). Classically, a decrement in
Received October 15,1990. Address all correspondence and requests for reprints to: W. Timothy Garvey, M.D., Section of Endocrinology, Indiana University School of Medicine, Veterans Administration Hospital 111-E, 1481 West 10th Street, Indianapolis, Indiana 46202. * This work was supported by NIH Grant DK-38765 and a grant from the V.A. Medical Service. t Recipient of a Fogarty International Research Fellowship Award from the NIH and a Postdoctoral Research Fellowship from the Sao Paulo State Research Foundation.
the major portion of the decrease in insulin sensitivity was due to uncoupling of occupied insulin receptors from stimulation of the glucose transport system. To further determine the mechanism for postbinding desensitization, we tested for concordant regulation of insulin receptor kinase activity. Insulin's ability to stimulate the receptor tyrosine kinase was assessed by multiple methods, including 1) receptor autophosphorylation and phosphorylation of Gh^-Tyr1 by solubilized insulin receptors activated in vitro, 2) histone-2B phosphorylation by receptors that were stimulated in intact cells and then solubilized under conditions that preserve the in celluto phosphorylation state, and 3) receptor autophosphorylation and phosphorylation of ppl80 in intact cells. Long term treatment (24 h) with glucose (10 mM) and insulin (50 ng/ml) markedly decreased insulin sensitivity (and receptor coupling), but did not affect insulin receptor kinase activity in any of these studies. We conclude that 1) glucose and insulin can independently and synergistically regulate insulin sensitivity by uncoupling ligand-bound insulin receptors from stimulation of glucose transport, and 2) glucose/insulin-induced desensitization occurred without any effect on insulin receptor tyrosine kinase activity. These observations suggest that decreased insulin sensitivity in target tissue may be induced by hyperglycemia and hyperinsulinemia in noninsulin-dependent diabetes mellitus via a receptor kinase-independent mechanism. (Endocrinology 128: 2415-2426, 1991)
maximal insulin responsiveness has been taken as evidence of a postreceptor defect, probably involving the hormone effector system (7), and we have, in fact, demonstrated cellular depletion of glucose transporter proteins in adipocytes isolated from these patients (8). On the other hand, a decrease in insulin sensitivity, defined as reduced effects at submaximal insulin concentrations, leading to a rightward shift in the insulin biological action dose-response curve, can be caused by insulin receptor down-regulation and/or an abnormality in the insulin signal coupling ligand binding to stimulation of the hormone effector system (7). In adipocytes from patients with NIDDM (1-3) and obesity (4, 5), receptor loss accounts for some, but not all, of the rightward shift
2415
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 July 2015. at 03:23 For personal use only. No other uses without permission. . All rights reserved.
2416
GLUCOSE AND INSULIN REGULATE INSULIN SENSITIVITY
in the insulin-glucose transport dose-response curve. The postbinding defect accounting for the remaining loss of insulin sensitivity has not been identified, but could involve a decrease in insulin receptor tyrosine kinase activity, since evidence suggests that insulin receptor kinase is involved in insulin signal transduction and couples insulin binding to stimulation of glucose transport (9-13). Furthermore, tyrosine kinase activity of solubilized insulin receptors from various tissues (14-17) and intact adipocytes (18-20) is decreased in NIDDM. We have previously suggested in NIDDM and obesity that hyperglycemia and/or hyperinsulinemia may directly induce insulin action defects in target tissue, based on our studies assessing the regulatory effects of glucose and insulin in primary cultured adipocytes (21-23). In these cells, chronic exposure to insulin per se decreased the ability of insulin to acutely stimulate glucose transport by sequentially acting at receptor and multiple postreceptor sites in the insulin action pathway (21). Furthermore, glucose was shown to potentiate insulin's ability to decrease maximal insulin responsiveness by impairing the translocation efficiency of intracellular glucose transporters to the cell surface (22). In the current study we assessed whether glucose could also modulate insulin sensitivity in cultured adipocytes. Secondly, we examined whether insulin's ability to decrease insulin sensitivity was due to effects on insulin receptor kinase activity. Thus, we hoped to gain some insight into potential factors that could impair insulin sensitivity in insulin-resistant disease states such as NIDDM and obesity.
Materials and Methods Materials Porcine monocomponent insulin and A-14-mono-[125I]insulin were generously supplied by Dr. Ronald Chance of Eli Lilly Co. (Indianapolis, IN). Antiphosphotyrosine antibody was kindly supplied by M. Kamps and B. Sefton (Salk Institute, San Diego, CA), and antiinsulin receptor antibodies by L. Mandarino (University of Pittsburg, Pittsburg, PA). L-[l-3H]Glucose was purchased form New England Nuclear (Boston, MA); 2-deoxy-D-[3H]glucose, [7-32P]ATP (3000 Ci/mmol), and [125I] protein-A from Amersham Corp. (Arlington Heights, IL); BSA (fraction V) from Armour Pharmaceutical Co. (Tarrytown, NY); collagenase from Worthington Biochemical Corp. (Freehold, NJ); Dulbecco's Modified Eagle's Medium (DMEM) from Gibco (Grand Island, NY); protein-A (formalin-fixed Staphylococcus aureus cells) from Bethesda Research Laboratories (Gaithersburg, MD); and CnBr-activated Sepharose 4B from Pharmacia (Piscataway, NJ). Histone-2B, poly (Glu4Tyr1), and all other reagents were obtained from Sigma (St. Louis, MO). Materials for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) were purchased from Bio-Rad (Richmond, CA).
Endo • 1991 Voll28»No5
Primary culture of adipocytes Isolated adipocytes were obtained from the epididymal fat pads of male Sprague-Dawley rats under sterile conditions and maintained in primary culture, as previously described (21-23). Adipocytes (5 X 104 cells/ml) were then incubated for 5 or 24 h at 37 C in the absence (control) and presence of various Dglucose and insulin concentrations. The presence of 2% fetal calf serum led to a final D-glucose concentration of 0.12 mM without further addition of D-glucose. Glucose transport assays After incubation, cells were washed three times in equal volumes of glucose- and insulin-free buffer, pH 7.4, containing 20 mM HEPES, 120 mM NaCl, 1.2 mM MgSO4) 2.0 mM CaCl2, 2.5 mM KC1, 1.0 mM NaH2PO4, 1.0 mM sodium pyruvate, and 1% BSA. The cells were then incubated in this buffer for an additional 40 min at 37 C to dissociate any remaining receptorbound insulin. These washing procedures effectively removed all extracellular and receptor-bound insulin and allowed the glucose transport system of insulin-stimulated cells to deactivate to basal levels. After washing, adipocytes were resuspended in the same buffer and incubated (1 ml) in the absence and presence of various insulin concentrations in a shaking water bath for 45 min at 37 C. Initial rates of 2-deoxyglucose uptake (substrate concentration, 0.2 mM) were measured over 3 min (2-3 X 105 cells/ml), as previously described (21), and the distribution space of radiolabeled L-glucose was used to correct for nonspecific carryover of radioactivity with the cells and the uptake of hexose by simple diffusion. 3-O-Methylglucose transport was assayed using a modification of the method of Whitesell and Gliemann (24), as previously described by Foley et al. (25). Measurement of insulin receptors Cells were washed to remove extracellular and receptorbound insulin, as described above, and then incubated (~2 X 105 cells/ml) in a shaking water bath at 16 C with 0.2 ng [125I] insulin and various concentrations of nonradioactive insulin. After 2 h the binding reaction was terminated, and specific [125I]insulin binding was determined, as previously described (21). Insulin binding to intact adipocytes was performed at 16 C, because insulin internalization is negligible at this low temperature (26); thus, cell-associated [125I]insulin essentially reflects binding only to cell surface receptors. Aliquots of wheat germ eluates (partially purified preparation of solubilized insulin receptors) were incubated with [I25I]insulin (final concentration, 0.5-1 ng/ml) at 4 C for 18 h in the presence or absence of increasing concentrations of unlabeled insulin, as previously described (18, 27). The amount of receptor-bound hormone was determined by previously described methods (28), and insulin-binding capacities of the receptor preparations were estimated by Scatchard analysis. Partial purification of insulin receptors and measurement of tyrosine kinase activity in vitro For the determination of receptor tyrosine kinase activity in vitro, cells were solubilized, and insulin receptors were partially purified over wheatgerm-agarose (27). Insulin-induced auto-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 July 2015. at 03:23 For personal use only. No other uses without permission. . All rights reserved.
GLUCOSE AND INSULIN REGULATE INSULIN SENSITIVITY phosphorylation of partially purified receptors and phosphorylation of the artificial kinase substrate Glu4-Tyr1 were performed as previously described (18, 27). Briefly, cultured adipocytes were disrupted and solubilized by agitating the cells in an ice-cold solubilization mixture containing 1.5% Triton X100, 7.5 mg/ml bacitracin, 15 mM benzamidine, 1500 kallikrein inhibitor units/ml aprotinin, and 2.5 mM phenylmethylsulfonylfluoride (PMSF), pH 7.4. The cellular extract was clarified by centrifugation, diluted 1:4 with buffer containing 25 mM HEPES (pH 7.4), 10% (wt/vol) glycerol, 0.05% Triton, 120 mM NaCl, 2.5 mM KC1, 1 mM CaCl2, 1 mM MgCl2, 15 mM benzamidine, 1500 kallikrein inhibitor units/ml aprotinin, and 2.5 mM PMSF, and then recycled twice over wheatgerm-agarose at 4 C. The glycoproteins were desorbed with 0.3 M iV-acetyl-Dglucosamine in the recycling buffer. Fractions containing insulin receptors were combined in each treatment group, and frozen at —70 C for use in subsequent phosphorylation and insulin binding studies. The protein content of the frozen aliquots was determined by the Bradford dye method (29), using crystalline BSA as the standard (Bio-Rad protein assay). Phosphorylation assays were performed as previously described (18, 27), using equal numbers of insulin receptors from control and treated cells. Receptors were preincubated in the absence and presence of various insulin concentrations for 18 h at 4 C, and then the phosphorylation reaction was initiated in a reaction mixture containing 5 mM MnCl2, 4 mM MgCl2, 500 MM CTP, and 50 MM [ T - 3 2 P ] A T P (11 MCi/nmol). The reaction was terminated in a solution containing 25 mM HEPES (pH 7.6), 0.05% Triton, 10% glycerol, 2 mM sodium orthovanadate, 100 mM NaF, 1 mM PMSF, 5 mM EDTA, 50 mM ATP, and 10 mM sodium pyrophosphate. The phosphorylated insulin receptors were immunoprecipitated with a 1:50 dilution of serum containing human antiinsulin receptor antibodies, as previously described (18), and then analyzed by SDSPAGE on 10% resolving gels according to the method of Laemmli (30). Under these conditions, 90-95% of the phosphorylated receptors were immunoprecipitated. The /3-subunit of the insulin receptor was visualized by autoradiography, and the degree of phosphorylation was quantitated by excising this band from the dried gel and counting in a scintillation counter. Equal areas of dried gel, judged free of radioactivity by autoradiography, were cut, counted, and subtracted as background. To assess phosphorylation of exogenous substrate, lectinpurified insulin receptors were preincubated with and without insulin as described above, and then reacted with the synthetic substrate Glu4-Tyr' (final concentration, 2 mg/ml), 5 mM MnCl2, 12 mM MgCl2, 500 MM CTP, and 50 fiM [7-32P]ATP (5 Ci/nmol). After incubation at 4 C for different intervals of time, the reaction was terminated by the addition of 50 mM ATP, and the 32P incorporated into Glu^Tyr1 was determined by the filter paper method, as previously described (18). Measurement of insulin receptor tyrosine kinase activity in intact cells The tyrosine kinase activity of receptors stimulated with insulin in intact adipocytes was estimated in two ways. By one approach, cells were incubated with and without insulin, insulin receptors were isolated under conditions designed to preserve the phosphorylation state found in the intact cells, and then
2417
receptor tyrosine kinase activity was assayed by its ability to phosphorylate histone-2B in vitro, as previously described (31). Adipocytes were preincubated with and without glucose/insulin, washed, and acutely stimulated with various insulin concentrations, as described above. Insulin receptors were then partially purified, as described above, but with the addition of phosphatase and kinase inhibitors to all buffers. Wheatgerm eluates were passed over insulin-Sepharose beads in order to purify and immobilize the insulin receptors. The phosphorylation reaction was carried out in a solution containing 5 mM MnCl2, 12 mM MgCl2, 500 nM cytidine triphosphate, 1 mg/ml histone, and 4 ^M [7-32P]ATP (80 AtCi/nmol). Previous detailed experiments have shown that at this low concentration of ATP, there is no additional activation of insulin receptor kinase activity as a result of in vitro manipulations subsequent to cell solubilization (31). Phosphorylated proteins were analyzed by SDS-PAGE on 15% gels according to the method of Laemmli (30), as previously described (31). In the second technique to measure receptor kinase activity in intact cells (20), adipocytes stimulated with and without insulin were solubilized in 2% SDS containing 2 mM Na3VO4 and 100 mM NaF to inhibit phosphatases. Cell extracts were electrophoresed on SDS-PAGE gels (10%), and the resolved protein bands were transferred to nitrocellulose for Western blotting. The nitrocellulose blot was incubated with polyclonal antiphosphotyrosine antibodies, followed by 125I-labeled protein-A, as described by Kamps and Sefton (32), to selectively label proteins containing phosphotyrosine. The 125I-labeled protein bands were visualized by autoradiography and quantitated by scanning densitometry. Titers of cell extract blotted with antiphosphotyrosine antibodies confirmed linearity between the amount of phosphoprotein and the density of autoradiographic images. Statistics Values are given as the mean ± SEM, and statistical significance was determined using Student's t test for paired and unpaired data where appropriate.
Results Glucose transport studies To assess the ability of glucose and insulin to chronically regulate the glucose transport system, isolated adipocytes were cultured with various glucose concentrations (0, 5, and 20 mM) in the absence and presence of 50 ng/ml insulin. After 5-24 h, the cells were washed to remove any insulin and allow full deactivation of the glucose transport system. The cells were then acutely restimulated with insulin, and 2-deoxyglucose uptake rates were determined. Our previous data (22), showing that glucose acted in combination with insulin to decrease basal and maximally stimulated glucose transport rates, were confirmed. The purpose of the current study was to test whether glucose and insulin were also able to regulate insulin sensitivity (i.e. acute stimulatory effects at submaximal insulin concentrations). Therefore, we
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 July 2015. at 03:23 For personal use only. No other uses without permission. . All rights reserved.
2418
GLUCOSE AND INSULIN REGULATE INSULIN SENSITIVITY
assessed the full dose-response curve for insulin-stimulated glucose transport and replotted transport rates as a percentage of the maximal insulin effect, as previously described (21). This allows optimal determination of the insulin ED50 as a quantitative measure of the degree to which the dose-response curve is right or left shifted. As is evident in Fig. 1, cells pretreated with either glucose (0, 5, or 20 mM) or insulin (50 ng/ml) alone for 5 h showed no change in insulin sensitivity (Fig. 1A). HowPreincubation Insulin Concentration 50 ng/ml 0 ng/ml
0.7
A: 5 hours preincubation i
0.6 0.5 0.4 c
0.3
O Q
0.2
UJ
I
0.1 0.7 w c 0) CO
B: 24 hours preincubation
0.6 0.5 0.4
ever, glucose and insulin in combination induced a rightward shift in the dose-response curve; glucose in the presence of insulin led to 61% and 116% increases in the ED50 values (i.e. a decrease in insulin sensitivity) as medium glucose was raised to 5 and 20 mM, respectively (P < 0.05). Additional regulatory effects were observed with more prolonged exposure to glucose and/or insulin for 24 h (Fig. IB). With longer term treatment, both glucose (20 mM) alone and insulin alone increased the insulin ED50 value by 52% (P < 0.05), and together interacted synergistically to cause a marked 183% increment in the ED50 compared to that in control cells (P < 0.01). Thus, glucose and insulin independently and synergistically regulated insulin sensitivity. We have previously shown (22) that D-glucosamine and D-mannose in addition to D-glucose could interact with insulin to diminish maximal insulin responsiveness, but that other compounds, such as D-fructose, sodium pyruvate, L-glucose, sucrose, and alanine, were inactive. In a similar manner, we examined substrate specificity for regulation of insulin sensitivity. Cells were cultured with and without the substrates listed in Table 1 (at 5 mM) in the presence of 50 ng/ml insulin. Then, after 24 h, the cells were washed, and the dose-response of insulin's acute ability to restimulate 2-deoxyglucose transport was assessed. Table 1 shows effects on the insulin ED50 as an index of insulin sensitivity. In this series of experiments, the ED50 in cells preincubated with insulin alone was further increased by coincubation with either D-mannose or D-glucose (P < 0.05). Thus, these agents greatly potentiated insulin's chronic ability to decrease insulin sensitivity and also led to reduced basal and maximally insulin-stimulated transport rates (data not TABLE 1. Ability of substrates to decrease insulin sensitivity of the glucose transport system
0.3 0.2 0.1
Endo • 1991 Voll28«No5
0
20
0
20
Preincubation D-glucose Concentration (mM) Fir,. 1. Glucose and insulin regulate insulin sensitivity in cultured adipocytes. Cells were cultured for 5 h (A) and 24 h (B) in medium containing 0, 5, and 20 mM glucose in the absence {left) and presence (rif>ht) of 50 ng/ml insulin, and then washed to remove any extracellular and receptor-bound insulin. After glucose transport was allowed to fully deactivate, dose-response curves for insulin's (0-50 ng/ml) acute ability to stimulate 2-deoxyglucose transport were assessed. The insulin ED50 was calculated as follows. The basal 2-deoxyglucose transport rate was subtracted from all insulin-stimulated transport rates, and the value at each insulin concentration was expressed as a percentage of the maximal insulin increment. The insulin ED5o, a measure of insulin sensitivity, was then determined from individual dose-response curves as the concentration yielding 50% of the maximal effect. Please note that the ordinate in this figure begins at 0.1. Data represent the mean ± SE of three to six experiments.
Substrate present during preincubation (5 mM)
Insulin sensitivity, ED50 (ng/ml)
Effect on maximum insulin responsiveness
None D-Glucose D-Mannose D-Fructose Sodium pyruvate L-Glucose
0.31 ± 0.01 1.05 ± 0.12° 1.23 ± 0.15" 0.75 ± 0.08° 0.55 ± 0.06° 0.28 ± 0.09
Yes Yes No No
No
Adipocytes were cultured for 24 h at 37 C in the absence (none) and presence of the indicated substrates (5 mM) and in the presence of 50 ng/ml insulin. All cells were then washed in insulin- and substrate-free buffer, and glucose transport rates were allowed to fully deactivate. We then assessed the dose response of insulin's ability (0-50 ng/ml) to acutely restimulate 2-deoxyglucose transport. The insulin ED60 value was calculated from individual dose-response curves, as described for Fig. 1. Insulin responsiveness denotes effects on the maximal glucose transport rate at 50 ng/ml insulin. Data represent the mean ± SE of three experiments. 0 P < 0.05 vs. controls with no added substrate.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 12 July 2015. at 03:23 For personal use only. No other uses without permission. . All rights reserved.
GLUCOSE AND INSULIN REGULATE INSULIN SENSITIVITY
shown), as previously reported (22). A different pattern was observed with D-fructose and sodium pyruvate; these compounds did not affect basal and maximal glucose transport rates (data not shown) (22), but were still able to potentiate insulin-induced decrements in insulin sensitivity. Specifically, the ED50 value was increased 2.4fold by D-fructose and 1.8-fold by sodium pyrvvate over that in cells treated with insulin alone (P < 0.05). In contrast, L-glucose, which is not metabolized and does not enter the cell via facilitative transporters, did not affect basal and maximal glucose transport rates and did not alter the insulin ED50 value (P = NS). In these studies, quantitatively similar effects on the insulin sensitivity (ED50) and insulin responsiveness of 2-deoxyglucose transport were observed using 3-O-methylglucose as the glucose analog in the transport assay (data not shown). Insulin binding studies Changes in insulin sensitivity could result from differences in cell surface insulin binding, a change in the coupling efficiency between insulin receptors and the biological effector system, or a combination of both. Figure 2 shows insulin competition binding curves to insulin receptors in cells cultured for 24 h in the presence
2419
of glucose (5 mM) plus insulin (50 ng/ml) and in controls. Glucose plus insulin led to an 18% decrease in insulin binding at trace concentrations of [125I]insulin without significantly affecting the IC50 value. This is generally consistent with a decrease in the number of insulin receptors and agrees with our previously reported studies of insulin-induced receptor down-regulation in primary cultured adipocytes (21). However, while receptor loss could be expected to impair insulin sensitivity in these cells, the relatively small decrease in insulin binding (~18%) could not by itself explain a 3-fold increase in the insulin ED50 value (see Fig. IB). The results indicate that the major mechanism accounting for the decrease in insulin sensitivity involves steps distal to ligand binding in the insulin action sequence. To examine whether the decrease in insulin sensitivity involved a decrease in insulin receptor-glucose transport system coupling efficiency, we plotted glucose transport rates as a percentage of the maximal insulin effect us. the actual amount of insulin bound to cell surface insulin receptors at each of the insulin concentrations used to construct the dose-response curve (Fig. 3). In other words, these plots assess the coupling efficiency between T
100 3.0
1
i
r
0 mM glucose No Insulin
80
2.5
i! £ I 60
2.0
