Reduced glucose uptake precedes insulin signaling defects in adipocytes from heterozygous GLUT4 knockout mice

Transcription

1 Reduced glucose uptake precedes insulin signaling defects in adipocytes from heterozygous GLUT4 knockout mice JING LI,* KAREN L. HOUSEKNECHT, ANTINE E. STENBIT,* ELLEN B. KATZ,* AND MAUREEN J. CHARRON*,1 *Department of Biochemistry, Albert Einstein College of Medicine, Bronx, New York 10461; USA; and Department of Animal Sciences, Purdue University, West Lafayette, Indiana 47907, USA ABSTRACT Decreased GLUT4 expression, impaired insulin receptor (IR), IRS-1, and pp60/irs-3 tyrosine phosphorylation are characteristics of adipocytes from insulin-resistant animal models and obese NIDDM humans. However, the sequence of events leading to the development of insulin signaling defects and the significance of decreased GLUT4 expression in causing adipocyte insulin resistance are unknown. The present study used male heterozygous GLUT4 knockout mice (GLUT4( / )) as a novel model of diabetes to study the development of insulin signaling defects in adipocytes with the progression of whole body insulin resistance and diabetes. Male GLUT4( / ) mice with normal fed glycemia and insulinemia (N/N), normal fed glycemia and hyperinsulinemia (N/H), and fed hyperglycemia with hyperinsulinemia (H/H) exist at all ages. The expression of GLUT4 protein and the maximal insulin-stimulated glucose transport was 50% decreased in adipocytes from all three groups. Insulin signaling was normal in N/N adipose cells. From 35 to 70% reductions in insulin-stimulated tyrosine phosphorylation of IR, IRS-1, and pp60/irs-3 were noted with no changes in the cellular content of IR, IRS-1, and p85 in N/H adipocytes. Insulin-stimulated protein tyrosine phosphorylation was further decreased to 12 23% in H/H adipose cells accompanied by 42% decreased IR and 80% increased p85 expression. Insulin-stimulated, IRS-1-associated PI3 kinase activity was decreased by 20% in N/H and 68% reduced in H/H GLUT4( / ) adipocytes. However, total insulinstimulated PI3 kinase activity was normal in H/H GLUT4( / ) adipocytes. Taken together, these results strongly suggest that hyperinsulinemia triggers a reduction of IR tyrosine kinase activity that is further exacerbated by the appearance of hyperglycemia. However, the insulin signaling cascade has sufficient plasticity to accommodate significant changes in specific components without further reducing glucose uptake. Furthermore, the data indicate that the cellular content of GLUT4 is the rate-limiting factor in mediating maximal insulinstimulated glucose uptake in GLUT4( / ) adipocytes. Li, J., Houseknecht, K. L., Stenbit, A. E., Katz, E. B., Charron, M. J. Reduced glucose uptake precedes insulin signaling defects in adipocytes from heterozygous GLUT4 knockout mice. FASEB J. 14, (2000) Key Words: adipocyte glucose transport NIDDM hyperglycemia PI3 kinase Insulin resistance is a progressive metabolic disorder characterized by reduced glucose uptake in response to normal concentrations of insulin and is most likely the result of a combination of polygenic defects and environmental factors (1 3). Diseases such as non-insulin-dependent diabetes mellitus (NIDDM) and obesity are characterized by insulin resistance (3, 4). Insulin signaling pathways involving the activation of PI3-kinase are responsible for the metabolic regulation of carbohydrate, lipid, and protein utilization (5). These pathways are logical targets for defects associated with insulin resistance and diabetes. However, there is a paucity of data concerning the progression of insulin signaling defects with the onset of insulin resistance and diabetes (2). This is especially true for the signaling events in adipocytes. The GLUT4 heterozygous knockout mouse, GLUT4( / ), is a novel mouse model for studying the development of insulin resistance and diabetes (6). In an outbred genetic background, male GLUT4( / ) mice display the full spectrum of phenotypes leading to insulin resistance and NIDDM (6). At all ages there are GLUT4( / ) mice with normal fed glycemia and insulinemia (N/N), normal fed glycemia and hyperinsulinemia (N/H), and fed hyperglycemia with hyperinsulinemia (H/H). Whereas 53% of males were able to maintain normal 1 Correspondence: Department of Biochemistry, Albert Einstein College of Medicine, 1300 Morris Park Ave., Bronx, NY 10461, USA. charron@aecom.yu.edu /00/ /$02.25 FASEB 1117

2 glucose and insulin levels at ages of around 2 to 4 months, more than 85% of male GLUT4( / ) mice develop fed hyperinsulinemia, and 60% of them develop fed hyperglycemia at 5 to 7 months of age (6). The diabetic H/H GLUT4( / ) mice have hypertension, diabetic cardiomyopathy, and reduced insulin-stimulated skeletal muscle glucose uptake but not dyslipidemia or obesity (6). In addition, prediabetic N/H GLUT4( / ) mice have severely reduced whole body glucose turnover, glycolysis, glycogen synthesis, and insulin action but normal hepatic glucose fluxes (7). It is the successive development of fed hyperinsulinemia, followed by hyperglycemia, that makes the male GLUT4( / ) mice an exceptional model for studying progressive changes in insulin signaling pathways associated with diabetic pathogenesis. Furthermore, the absence of obesity in these mice makes it possible to dissociate the effect of obesity on the diabetic phenotype, which complicates analysis in existing animal models of diabetes (8). Insulin has diverse functions in different tissues and cell types reflected by the complicated signaling cascade (5, 9, 10). The binding of insulin to its receptor activates the tyrosine kinase activity of the insulin receptor (IR), which leads to tyrosine phosphorylation of IR itself and the IR substrates IRS-1, IRS-2, Src homology collagen (Shc), and pp60/irs-3 (2, 9, 11). Tyrosine-phosphorylated substrates then recruit other proteins such as the p85 subunit of phosphatidylinositol 3-kinase (PI3-kinase) and Grb protein via SH2 domains (9, 10). It has been shown that the activation of PI3 kinase leads to translocation of GLUT4-containing vesicles and subsequently increased glucose uptake in skeletal muscle and adipocytes (1, 12). Studies examining insulin signaling defects in skeletal muscle, liver, and adipose tissue of insulinresistant animal models and obese or diabetic human subjects have shown reduced insulin-stimulated IR and IRS-1 tyrosine phosphorylation and decreased insulin-stimulated, IRS-1-associated PI3 kinase activity (13 19). Studies with IRS-1-deficient mice suggested that insulin-stimulated, IRS- 1-associated PI3 kinase activity is responsible for insulin-stimulated glucose uptake in skeletal muscle (20). However, in adipose tissue, the role of insulin-stimulated, IRS-1-associated PI3 kinase activity in determining insulin-stimulated glucose uptake is not clear (21). Recently, Rondinone and colleagues (22) demonstrated that whereas IRS-1 protein expression is reduced, IRS-2 becomes the main docking protein for PI3 kinase in adipocytes from NIDDM patients. Reduced GLUT4 protein expression is a common characteristic of adipose tissue from animal models of diabetes and insulin resistant (23). However, down-regulation of GLUT4 protein expression in adipocyte seems to be the consequence of the development of whole body insulin resistance in obese insulin-resistant mice (24). The contribution of decreased GLUT4 content per se to adipocyte insulin resistance is unknown. To date, no studies have described alternations in the insulin signaling cascade in adipocytes during the development of diabetes. GLUT4( / ) mice present a unique opportunity to study the onset and progression of insulin signaling defects concomitant with the progression of whole body insulin resistance and diabetes (6). As GLUT4 content is decreased first in adipocytes of GLUT4( / ) mice, when most of the animals have normal serum glucose and insulin concentrations (6), it is possible to answer the question of whether reduced GLUT4 protein content alone can cause insulin resistance and insulin signaling defects. In the present study, we describe the progression of defects in insulin signaling in adipocytes from GLUT4( / ) mice that accompany the pre- and overt diabetic stages. MATERIALS AND METHODS Animals Heterozygous GLUT4 knockout mice were generated by mating GLUT4 null mice with the wild-type CD1 strain and subsequent heterozygous mating. They were genotyped as described previously (25). The mice were housed in rooms with a regular day/night cycle and were fed standard rodent chow and water ad libitum. Blood glucose and plasma insulin determinations Ad libitum fed mice were bled at 1:30 a.m. for fed serum levels of glucose and insulin. Whole blood was drawn from the orbital sinus and centrifuged immediately. Serum was immediately frozen on dry ice and kept at 70 C until further use. Serum glucose levels were measured by the glucose oxidase method (Sigma, Inc., St. Louis, Mo.). Serum insulin levels were measured by radioimmunoassay (Linco Research, Inc., St. Louis, Mo.) using rat insulin standards. Isolation of epididymal adipocyte Mice were killed by cervical dislocation at around 11:00 a.m. and epididymal fat pads were dissected and weighed. Adipocytes were isolated by collagenase digestion (Worthington, Inc., Freehold, N.J.; 1 mg/g fat pad) at 37 C as described previously (26). Digested fat pads were washed and resuspended in Krebs-Ringer-bicarbonate-HEPES (KRPH) buffer (ph 7.4) supplemented with 2.5% bovine serum albumin (BSA; Cohn fraction V, Intergen, Inc., Purchase, N.Y.) and 200 nm adenosine (26) Vol. 14 June 2000 The FASEB Journal LI ET AL.

3 Immunoblotting Isolated adipocytes were lysed in a buffer containing 50 mm HEPES, 150 mm NaCl, 2.0 mm NaO3V, 20 mm NaF, 4 mm EDTA, 20% glycerol, 10% Nonidet P-40, 4 mm PMSF, 20 nm leupeptin, and 20 nm aprotinin. Crude adipocyte lysates were used for immunoblot analysis. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), electrophoretically transferred to Hybond ECL membrane (Amersham, Inc., Arlington, Ill.), and reacted with specific antibodies. For GLUT4 protein detection, 30 g of total lysates were loaded and a polyclonal antibody against the carboxyl terminus of GLUT4 was used (25). For detection of p85 protein, 40 g of total lysates were loaded and an antibody against the SH2 domain of p85 protein was used (a kind gift from Dr. J. Backer). For detection of IR protein, 100 g of total lysates were loaded and an antibody against the carboxyl terminus of human IR subunit was used (a kind gift from Dr. B. Cheatham). For detection of IRS-1 protein, 2 mg of total lysates were used in immunoprecipitations with an antibody specific against the PH domain of IRS-1 (a kind gift from Dr. M. White) and then blotted using the same antibody. Determination of adipose cell size and number Resuspended adipose cells were aliquoted in two triplicate sets. One set was fixed in 2% osmic acid, washed in distilled water, and resuspended in 1 ml of counting solution that contained 55% (v/v) glycerol, 1% (v/v) Triton X-100 and 3.85% NaCl. Triplicates (6 ml) of the resuspended cells were counted using a hemocytometer under a microscope. The other set of triplicates was used to determine cell sizes. Cells were extracted using heptane and lipid content was weighed. Adipose cell size was determined as microgram lipid/cell as described previously (27, 28). Adipocyte glucose transport Isolated adipocytes were incubated for 30 min in KRPH buffer supplemented with 2.5% BSA at 37 C in the presence or absence of various concentrations of porcine insulin (a kind gift from Dr. R. Chance). [U- 14 C]glucose (Amersham; 0.3 mm final concentration) uptake was performed as described previously (26). Briefly, [U- 14 C]glucose was added to the cell suspension for 20 min and the reaction was terminated by spinning the cell suspension through dinonylphthalate oil (ICN Biomedicals, Inc., Aurora, Ohio). The upper phase, which contains adipose cells, was collected and subjected to liquid scintillation counting. Under the assay conditions, uptake of [ 14 C]glucose has been shown to directly reflect glucose transport (26). Determination of PI3 kinase activity Isolated adipocytes were divided into two groups: basal and insulin stimulated. In the basal group, cells were immediately lysed in lysis buffer. In the insulin-stimulated group, cells were incubated with 100 nm insulin for 5 min prior to lysis. The lysate was centrifuged at 2500 g for 5 min; 1 mg of solubilized lysate was used for immunoprecipitations with 8 l of anti-p85 or 10 l of anti-irs-1 antibodies. Immunoprecipitation was carried out at 4 C overnight with constant rotating. Protein A-Sepharose beads were added and incubation was continued for 3 more hours. Beads-associated PI3 kinase activity was assayed based on phosphorylation of phosphatidylinositol in the presence of [ - 32 P]ATP (Amersham) (14). Reaction products were separated by thin layer chromatography (TLC), exposed to X-ray film, and quantified by densitometry. Quantification was also performed on representative plates by Cerenkov counting of [ 32 P] phosphate incorporated into PI(3)P. In all cases, the fold difference was comparable to that measured by densitometry. Activity of PI3 kinase was expressed as the percentage of control activity in the basal state. Immunoblot analysis of tyrosine-phosphorylated proteins on insulin stimulation One milligram of total adipose cell lysate protein from basal and insulin-stimulated cells was immunoprecipitated with 10 ml of the antiphosphotyrosine antibody PY20 (Transduction Labs, Inc., Lexington, Ky.) for 4 h at 4 C. The immunocomplex was brought down by protein A-Sepharose beads, loaded onto 8% SDS-PAGE, and immunoblotted onto ECL membrane (Amersham). The filter was then reacted with the antiphosphotyrosine antibody RC20 (Transduction Labs, Inc.) according to the manufacturer s instructions. Statistical analysis Data are presented as mean se. Significance between groups was determined by using the Student s t test. P 0.05 was accepted as statistically significant. Pooled samples were treated as one for analysis. RESULTS Characterization of heterozygous GLUT4( / ) knockout mice Three groups of male GLUT4( / ) mice were used in this study based on their serum glucose and insulin profiles. They represented the three stages of development of diabetes in GLUT4( / ) mice (Table 1). Glucose between 200 and 220 mg/dl was categorized as normal glucose while glycemia higher than 350 mg/dl was grouped as hyperglycemia. Insulin concentrations lower than 10 ng/ml were considered normal insulinemia whereas concentrations higher than 25 ng/ml were grouped as hyperinsulinemia. The 3-month-old GLUT4( / ) mice with normal fed glucose and fed insulin levels were referred to as normal/normal (N/N). The 3- to 4-month-old prediabetic GLUT4( / ) mice with normal fed glucose but high fed insulin were referred to as normal/high (N/H). Since it was difficult to get enough animals with both hyperglycemia and hyperinsulinemia at 3 to 4 months of age, 6- to 9-month-old diabetic GLUT4( / ) mice with both fed hyperglycemia and hyperinsulinemia were used and referred to as high/high (H/H). Control mice in each group were age and sex matched. All GLUT4( / ) mice had body and epididymal fat pad weights similar to their age-matched controls (Table 2). Epididymal fat cell size was similar in both normal N/N and prediabetic N/H groups, but was INSULIN SIGNALING IN GLUT4( / ) ADIPOCYTES 1119

4 TABLE 1. Fed glucose and insulin levels of 3-month-old N/N, 3 to 4-month-old N/H, 6- to 9-month-old H/H GLUT4( / ), and ageand sex-matched control mice N/N group N/H group H/H group GLUT4( / ) Control GLUT4( / ) Control GLUT4( / ) Control Glucose (mg/dl) * (n 17) (n 11) (n 9) (n 11) (n 26) (n 12) Insulin (ng/ml) * * (n 17) (n 11) (n 9) (n 11) (n 26) (n 12) * P % 0.02 larger in diabetic H/H group compared to controls (Table 2). Expression of GLUT4 protein in adipocytes of heterozygous GLUT4( / ) knockout mice The expression of GLUT4 protein was assessed by Western blot analysis using a specific antibody against GLUT4. GLUT4 protein content in adipocytes was reduced by 50% 0.01 in all three groups (Fig. 1). In the diabetic H/H group, the reduction in GLUT4 was slightly greater than the other two groups (Fig. 1C). GLUT1 protein content in the fat was measured in a previous study in the diabetic H/H and control groups, and no significant difference in expression was detected (6). Glucose uptake in adipocytes of heterozygous GLUT4( / ) knockout mice To delineate the effect of reduced GLUT4 expression and whole body metabolic milieu on glucose uptake, [U- 14 C]glucose accumulation in the absence or the presence of 1 nm and 100 nm insulin concentrations was studied. Basal glucose uptake was normal in adipocytes from all three groups of GLUT4( / ) mice. In the presence of 1 nm insulin, control cells for both N/N and N/H adipocytes showed a significant increase in glucose uptake. Adipocytes from N/N GLUT4( / ) mice also showed a comparable increase in glucose uptake at 1 nm of insulin. However, adipocytes from N/H animals exhibited a 50% decrease in glucose uptake at 1 nm insulin compared to controls. No significant increase in glucose uptake was detected in the 6- to 9-month-old control and H/H groups in the presence of 1 nm insulin ( in the H/H adipocytes; in the controls). Consequently, the insulin concentration was increased to 5 nm for these groups. At this insulin concentration, a significant increase in glucose uptake in the control adipocytes was achieved, whereas an attenuated increase in glucose uptake in the H/H adipocytes was seen (50% of control at the same insulin concentration) (Fig. 2). The reduction of glucose uptake in N/H and H/H GLUT4( / ) adipocytes under such concentrations of insulin stimulation suggested that insulin sensitivity was reduced in these cells. The maximal fold stimulation of glucose uptake by insulin was similar in all three groups of GLUT4( / ) adipocytes (2.7-fold in N/N, 2.8-fold in N/H, and 2.6-fold in H/H). On the other hand, the maximal insulin-stimulated glucose uptake was 6.0- to 7.4-fold in controls. Thus, the relative fold decrease in maximal insulin-stimulated glucose uptake in all groups of TABLE 2. Body weight, epididymal fat pad weight, and fat cell size of 3-month-old N/N, 3- to 4-month-old N/H, 6- to 9-month-old H/H GLUT4( / ), and age- and sex-matched control mice. a N/N group N/H group H/H group GLUT4( / ) Control GLUT4( / ) Control GLUT4( / ) Control Body weight (g) (n 13) (n 10) (n 8) (n 11) (n 16) (n 13) Fat pad weight (g) (n 13) (n 10) (n 8) (n 11) (n 23) (n 21) Cell size * (n 9) (n 8) (n 12) (n 6) (n 9) (n 8) a The fat cell size is determined as g lipids/cell. * p Vol. 14 June 2000 The FASEB Journal LI ET AL.

5 Figure 1. Expression of GLUT4 protein in N/N (A). N/H (B), and H/H (C) adipocytes. Expression of GLUT 4 protein was determined by immunoblot analysis and quantified by densitometry. Experiments were repeated using 6 fat cell lysates per group. Expression of GLUT4 protein in controls was set to 100% in each group. A representative gel is shown on top of each bar graph. *P GLUT4( / ) adipocytes was 50%, which was the same as the reduction of GLUT4 protein content. Insulin-stimulated protein tyrosine phosphorylation in adipocytes of heterozygous GLUT4( / ) knockout mice The difference in insulin sensitivity among N/N, N/H, and H/H GLUT4( / ) adipocytes suggested modulation of insulin signaling in these cells. Insulin-stimulated protein tyrosine phosphorylation profiles were examined using isolated adipocytes from all three groups. Tyrosine phosphorylation of three major bands that migrate at the positions of the IR subunit, IRS-1, and IRS-3 (11, 29) were uniformly identified in controls. No changes in the intensity of tyrosine phosphorylation of these three bands were detected in adipocytes from N/N GLUT4( / ) mice (Fig. 3A). However, reductions (64% 5% of IR, 71% 4% of IRS-1, and 35% 1% of IRS-3) were observed in adipocytes from N/H GLUT4( / ) mice (Fig. 3B). Insulin-stimulated tyrosine phosphorylation of IR subunit, IRS-1, and pp60/irs-3 was further reduced in the diabetic H/H GLUT4( / ) adipocytes (Fig. 3C). IRS-1 tyrosine phosphorylation was barely detectable on immunoblot analysis of H/H adipocytes. Tyrosine phosphorylation of IR subunit and pp60/irs-3 in H/H GLUT4( / ) adipocytes was reduced to 12% 1% and 23% 3% of control levels, respectively. Expression of IR, IRS-1, and p85 in adipocytes of heterozygous GLUT4( / ) knockout mice The total protein content of IR, IRS-1, and p85 was measured in adipocytes from all three groups (Fig. 4). No difference in the content of these proteins was noted in adipocytes of either N/N or N/H groups compared to controls. However, a 42% 3% reduction of IR protein was noted in adipocytes from H/H GLUT4( / ) adipocytes (Fig. 4A). Expression of p85 was increased 80% 3% over controls in H/H GLUT4( / ) adipocytes (Fig. 4B). Insulin-stimulated PI3 kinase activity in adipocytes of heterozygous GLUT4( / ) knockout mice Figure 2. [ 14 C]U-Glucose uptake in N/N (A), N/H (B), H/H (C), and control adipocytes. In each group, sex- and agematched controls were used. Results shown are mean se of 4 6 mice in each group. *P 0.5 GLUT4( / ) vs. controls. PI3 kinase is involved in mediating insulin-stimulated GLUT4 translocation and insulin-stimulated glucose uptake in adipocytes. Insulin-stimulated, INSULIN SIGNALING IN GLUT4( / ) ADIPOCYTES 1121

6 Figure 3. Insulin-stimulated protein tyrosine phosphorylation of N/N (A), N/H (B), and H/H (C) GLUT4( / ), and control adipocytes. In each group, experiments were repeated 3 5 times using adipocytes pooled from 9 GLUT4( / ) and 6 controls. Results were quantified by densitometry. IRS-1-associated PI3 kinase activity was normal in adipocytes from N/N GLUT4( / ) mice compared to controls (Fig. 5A). A modest decrease in IRS-1- associated PI3 kinase activity of 20% 1% was noted in adipocytes from N/H GLUT4( / ) mice (data not shown). This was further decreased to 32% 1% of controls in adipocytes from H/H GLUT4( / ) mice (Fig. 5B). To understand the significance of increased p85 protein expression in adipocytes of H/H GLUT4( / ) mice, p85-associated PI3 kinase activity was studied. Basal levels of p85-associated PI3 kinase activity in H/H GLUT4( / ) adipocytes were 50% 3% higher than controls (Fig. 6). Similar p85-associated PI3 kinase activity was measured in H/H GLUT4( / ) compared to control adipocytes under maximal insulin stimulation (Fig. 6). Figure 4. Expression of IR (A) and p85 (B) proteins in N/N, N/H, and H/H adipocytes. In each group, experiments were repeated more than 3 times with multiple lanes of samples from different adipocyte lysates at each time. Results were quantified by densitometry. DISCUSSION Previous studies have identified insulin signaling defects in tissues of diabetic humans and animals (13, 15, 16, 30). Decreased insulin-stimulated IRS-1 phosphorylation and PI3 kinase activity have been observed in adipocytes from human subjects with NIDDM and gold thioglucose-induced obese hyperinsulinemic mice (17 19). However, the earliest alterations in cellular and molecular aspects of insulin action are poorly defined as very few studies of prediabetic populations have been performed. In contrast, individual or additive effects of hyperinsulinemia and hyperglycemia on insulin signaling pathways have been suggested (4). However, data are lacking defining the modulation of the insulin signaling cascade in adipocytes after the onset of insulin resistance and overt diabetes. On the other hand, decreased GLUT4 expression has been uniformly observed in adipocytes from animal models of insulin resistance and in obese and NIDDM patients (23). Down-regulation of GLUT4 protein expression in adipocytes seems to be a consequence of development of whole body insulin resistance in obese insulin-resistant mice (24). However, the significance of reduced GLUT4 content relative to adipocyte insulin sensitivity is not well understood. It is unclear whether reduced GLUT4 protein or impaired insulin signaling is the rate-limiting step in determining insulin-stimulated glucose uptake in adipocytes Vol. 14 June 2000 The FASEB Journal LI ET AL.

7 Figure 5. Insulinstimulated, IRS-1-associated PI3 kinase activity in N/N, H/H GLUT4( / ), and control adipocytes. Results shown are mean se (n 3) using adipocytes pooled from 1 or 2 mice in each experiment. Targeted disruption of the GLUT4 gene leads to different phenotypes in homozygous and heterozygous knockout mice (6, 25). The existence of N/N, N/H, and H/H GLUT4( / ) mice make them an ideal system for studying the progressive development of diabetes (6). In GLUT4( / ) mice, the decrease in GLUT4 expression was first detected in white adipose tissue of 2-month-old males, preceding changes in skeletal muscle GLUT4 protein content (6). GLUT4( / ) mice at this age had normal fed glycemia and insulinemia. Insulin signaling in adipocytes from these mice was normal, as indicated by normal insulin-responsive protein tyrosine phosphorylation and IRS-1-associated PI3 kinase activation (Fig. 3A, Fig. 5A). Studies of insulin-stimulated protein phosphorylation in N/H and H/H GLUT4( / ) adipocytes delineate the progression of insulin signaling defects along with the development of hyperinsulinemia and hyperglycemia. Decreased insulin-stimulated protein tyrosine phosphorylation was observed in adipocytes once GLUT4( / ) mice developed hyperinsulinemia, as seen in N/H GLUT4( / ) adipocytes (Fig. 3B). It was further decreased in H/H GLUT4( / ) adipocytes (Fig. 3C). In addition, the development of hyperglycemia was accompanied by decreased IR expression in H/H GLUT4( / ) adipocytes (Fig. 4). Consistent with the reduced IRS-1 phosphorylation, insulin-stimulated, IRS-1-associated PI3 kinase activity was reduced in both N/H and H/H adipocytes (Fig. 5). Since the expression of IR was not changed in N/H adipocytes, the data suggested that IR tyrosine kinase activity is susceptible to changes in circulating insulin concentration. Chronic hyperglycemia, together with hyperinsulinemia, leads to the down-regulation of IR expression and the further reduction of insulin-stimulated protein tyrosine phosphorylation in H/H GLUT4( / ) adipocytes (Figs. 3, 4). Tyrosine-phosphorylated IRS-1 and pp60/irs-3 serve as docking proteins for the binding of other proteins such as the p85 subunit of PI3 kinase (9, 29, 31). Binding of p85 to IRS-1 and/or pp60/irs-3 activates PI3 kinase (9, 29, 31). It has been suggested that PI3 kinase is involved in mediating insulinregulated glucose metabolism (32, 33). Decreased IRS-1 tyrosine phosphorylation was accompanied by reduced insulin-stimulated, IRS-1-associated PI3 kinase activity in N/H and H/H GLUT4( / ) adipocytes (Fig. 5). However, expression of p85 protein was increased by 80% in diabetic H/H GLUT4( / ) adipocytes (Fig. 4). Along with increased p85 expression, p85-associated PI3 kinase activity was increased by 50% in the basal state and insulin-stimulated, p85-associated PI3 kinase activity was not altered in H/H GLUT4( / ) adipocytes compared to controls (Fig. 6). Although the mechanism of p85 protein up-regulation is unknown, the normal insulin-stimulated total p85 activity suggests that other pathways are activated and compensate for the reduction in IRS-1-associated PI3 kinase activity in H/H adipocytes. Such unchanged insulinstimulated total PI3 kinase activity might be responsible for the absence of further reduction in glucose uptake in H/H adipocytes. The reduction of GLUT4 expression was similar Figure 6. p85-associated PI3 kinase activity in H/H GLUT4 ( / ) and control adipocytes. Results shown are mean se (n 9) using adipocytes pooled from 1 or 2 mice in each experiment. A) Representative TLC analysis under basal and insulinstimulated conditions. B) Densitometric quantification of results. The p85-associated PI3 kinase activity of control adipocytes in the basal state was set to 100%. INSULIN SIGNALING IN GLUT4( / ) ADIPOCYTES 1123

8 among N/N, N/H, and H/H adipocytes of GLUT4( / ) mice (Fig. 1). In addition, maximal insulin-stimulated glucose uptake was similar in all three groups of GLUT4( / ) adipocytes (Fig. 2). The reduction in insulin-stimulated glucose uptake was similar to the decrease in GLUT4 protein content. The progressive and severe reduction of protein tyrosine phosphorylation and the decrease in insulin-stimulated, IRS-1-associated PI3 kinase activity in N/H and H/H adipocytes did not cause further reduction in maximal insulin-stimulated glucose uptake. However, glucose uptake at low insulin concentrations (1 nm in N/H and 5 nm in H/H) in N/H and H/H adipocytes of GLUT4( / ) was significantly reduced compared to controls. On the other hand, a normal stimulation of glucose uptake was observed in N/N GLUT4( / ) adipocytes at an insulin concentration of 1 nm (Fig. 2A), which is consistent with the normal insulin-stimulated glu protein tyrosine phosphorylation in N/N adipocytes. Thus, it is clear that cellular GLUT4 content is the rate-limiting factor in determining maximal insulinstimulated glucose uptake in GLUT4( / ) adipocytes. However, the N/N adipocytes have normal glucose uptake at physiological insulin levels, with only half the normal amount of GLUT4. This demonstrates that adipocytes with normal insulin signaling can maintain normal glucose uptake even with a severely reduced GLUT4 content. The lack of effect of reduced IRS-1-associated PI3 kinase activity on maximal insulin-stimulated glucose uptake in GLUT4( / ) adipocytes was also observed in adipocytes of aged gold thioglucose-induced obese mice and 3T3 L1 adipocytes (17, 34 36). One explanation could be that low levels of insulin-stimulated, IRS-1-associated PI3 kinase activity are sufficient to fully stimulate GLUT4 mediated glucose transport in H/H GLUT4( / ) adipocytes. Early studies in human adipocytes suggested that only a small portion of IRS-1 phosphorylation is required for full stimulation of glucose uptake (19). In addition, the extent to which IRS-1-associated PI3 kinase activity was reduced was not proportional to the reduction of IRS-1 tyrosine phosphorylation. It also suggested that in the insulin signaling cascade, protein is phosphorylated in excess to that required for the stimulation of the next step. Another possibility that has been proposed is that IRS-1-independent pathways may lead to the activation of PI3 kinase, GLUT4 translocation and glucose uptake (34 36). This is supported by the observation that total insulin-stimulated PI3 kinase activity was normal in H/H GLUT4( / ) adipocytes. Conflicting results exist for the involvement of IRS-1 in insulin-stimulated GLUT4 translocation and glucose uptake. Insulin-activated, IRS-1-associated PI3 kinase was identified in GLUT4-containing vesicles (37), whereas microinjection of antibodies against IRS-1 did not block insulin-stimulated, GLUT4-containing vesicle translocation (38). Recent studies of 3T3 L1 adipocytes as well as adipocytes from NIDDM patients suggested that other mechanisms that activate PI3 kinase, such as through docking to IRS-2, could play an important role in insulin-stimulated adipocytes glucose uptake (22, 34 36). Thus, increased IRS-2 or pp60/irs-3 expression or docking of PI3 kinase to these proteins could be the potential mechanism for the maintenance of insulin-stimulated PI3 kinase activity in H/H GLUT4( / ) adipocytes. The maintenance of insulin-stimulated PI3 kinase activity in H/H GLUT4( / ) adipocytes appears to protect cells from further reductions in glucose uptake. In summary, these experiments represent the first longitudinal study describing changes in the insulin signaling pathway that occur in adipocytes from normal to an insulin-resistant state. Changes in insulin signaling coincide with the development of circulating hyperinsulinemia. Hyperglycemia and hyperinsulinemia ultimately lead to decreased IR protein expression, decreased IRS-1 and pp60/irs-3 tyrosine phosphorylations, and decreased IRS-1-associated PI3 kinase activation. Expression of p85 protein is up-regulated and total insulin-stimulated PI3 kinase activity remains normal in H/H GLUT4( / ) adipocytes, suggesting a compensatory response to the reduced GLUT4 expression and insulin signaling. Despite the impaired insulin signaling in N/H and H/H adipocytes, maximal insulin-stimulated glucose uptake is not further reduced, demonstrating that the cellular GLUT4 content is the rate-limiting factor in mediating maximal insulin-stimulated glucose uptake in adipocytes. Finally, the experiments described above also show that the insulin signaling cascade has sufficient plasticity to accommodate significant changes in specific components without further reducing glucose uptake. We wish to thank Drs. M. F. White, J. Smith-Hall, J. Backer, and B. Cheatham for the generous provision of antibodies. We also thank Drs. T.-S. Tsao and R. W. Gelling for fruitful discussions. This work is submitted in partial fulfillment of the requirements for the Ph.D. degree for the Albert Einstein College of Medicine (J.L.). This work was supported by grants from the National Institutes of Health (DK 47425, HL 58119), the American Diabetes Association, the American Heart Association, and the Cancer Center of Albert Einstein of Medicine (to M.J.C.), and by a grant from Purdue University Agricultural Research Program (to K.L.H.). M.J.C. is the recipient of an Irma T. Hirschl Career Scientist award. REFERENCES 1. Czech, M. P. (1995) Molecular actions of insulin on glucose transport. Annu. Rev. Nutr. 15, Reaven, G. M. (1995) Pathophysiology of insulin resistance in human disease. Physiol. Rev. 75, Vol. 14 June 2000 The FASEB Journal LI ET AL.

9 3. Kahn, C. R. (1994) Banting Lecture. Insulin action, diabetogenes, and the cause of type II diabetes. Diabetes 43, Olefsky, J. M, and Nolan, J. J. (1995) Insulin resistance and non-insulin-dependent diabetes mellitus: cellular and molecular mechanisms. Am. J. Clin. Nutr. 61 (Suppl.), 980S 986S 5. Cheatham, B, and Kahn, C. R. (1995) Insulin action and the insulin signaling network. Endocr. Rev. 16, Stenbit, A. E., Tsao, T.-S., Li, J., Burcelin, B., Geenen, D. L., Factor, S. M., Houseknecht, K. L., Katz, E. B., and Charron, M. J. (1997) GLUT4 heterozygous knockout mice develop muscle insulin resistance and diabetes. Nature Med. 3, Rossetti, L., Stenbit, A. E., Chen, W., Hu, M., Barzilai, N., Katz, E. B., and Charron, M. J. (1997) Peripheral but not hepatic insulin resistance in mice with one disrupted allele of the glucose transporter type 4 (GLUT4) gene. J. Clin. Invest. 100, Coleman, D. J. (1978) Obese and diabetes: two mutant genes causing diabetes-obesity syndromes in mice. Diabetologia 14, White, M. F., and Kahn, C. R. (1994) The insulin signaling system. J. Biol. Chem. 269, Lee, J., and Pilch, P. F. (1994) The insulin receptor: structure, function, and signaling. Am. J. Physiol. 266, C319 C Lavan, B. E., Lane, W. S., and Lienhard, G. E. (1997) The 60-kDa phosphotyrosine protein in insulin-treated adipocytes is a new member of the insulin receptor substrate family. J. Biol. Chem. 272, Stephens, J. M., and Pilch, P. F. (1995) The metabolic regulation and vesicular transport of GLUT4, the major insulinresponsive glucose transporter. Endocr. Rev. 16, Goodyear, L. J., Giorgino, F., Sherman, L. A., Carey, J., Smith, R. J., and Dohm, G. L. (1995) Insulin receptor phosphorylation, insulin receptor substrate-1 phosphorylation, and phosphatidylinositol 3-kinase activity are decreased in intact skeletal muscle strips from obese subjects. J. Clin. Invest. 95, Folli, F., Saad, M. J., Backer, J. M., and Kahn, C. R. (1993) Regulation of phosphatidylinositol 3-kinase activity in liver and muscle of animal models of insulin-resistant and insulin-deficient diabetes mellitus. J. Clin. Invest. 92, Saad, M. J., Folli, F., Kahn, J. A., and Kahn, C. R. (1993) Modulation of insulin receptor, insulin receptor substrate-1, and phosphatidylinositol 3-kinase in liver and muscle of dexamethasone-treated rats. J. Clin. Invest. 92, Saad, M. J. A., Areki, E., Miralpeix, M., Rothenberg, P. L., White, M. F., and Kahn, C. R. (1992) Regulation of insulin receptor substrate-1 in liver and muscle of animal models of insulin resistance. J. Clin. Invest. 90, Heydrick, S. J., Gautier, N., Olichon-Berthe, C., Van Obberghen, E., and Le Marchand-Brustel, Y. (1995) Early alteration of insulin stimulation of PI 3-kinase in muscle and adipocyte from gold thioglucose obese mice. Am. J. Physiol. 268, E604 E Freidenberg, G. R., Henry, R. R., Klein, H. H., Reichart, D. R., and Olefsky, J. M. (1987) Decreased kinase activity of insulin receptors from adipocytes of non-insulin-dependent diabetes subjects. J. Clin. Invest. 79, Thies, R. S., Molina, J. M., Ciaraldi, T. P., Freidenberg, G. R., and Olefsky, J. M. (1990) Insulin-receptor autophosphorylation and endogenous substrate phosphorylation in human adipocytes from control, obese, and NIDDM subjects. Diabetes 39, Yamauchi, T., Tobe, K., Tamemoto, H., Ueki, K., Kaburagi, Y., Yamamoto-Honda, R., Takahashi, Y., Yoshizawa, F., Aizawa, S., Akanuma, Y., Sonenberg, N., Yazaki, Y., and Kadowaki, T. (1996) Insulin signalling and insulin actions in the muscles and livers of insulin-resistant, insulin receptor substrate 1-deficient mice. Mol. Cell. Biol. 16, Araki, E., Lipes, M. A., Patti, M.-E., Bruning, J. C., Haag, B. I., Johnson, R. S., and Kahn, C. R. (1994) Alternative pathway of insulin signalling in mice with targeted disruption of IRS-1 gene. Nature (London) 372, Rondinone, C. M., Wang, L. M., Lonnroth, P., Wesslau, C., Pierce, J. H., and Smith, U. (1997) Insulin receptor substrate (IRS) 1 is reduced and IRS-2 is the main docking protein for phosphatidylinositol 3-kinase in adipocytes from subjects with non-insulin-dependent diabetes mellitus. Proc. Natl. Acad. Sci. USA 94, Zierath, J. R., Houseknecht, K. L., and Kahn, B. B. (1996) Glucose transporters and diabetes. Semin. Cell Dev. Biol. 7, Koranyi, L., James, D., Mueckler, M., and Permutt, M. A. (1990) Glucose transporter levels in spontaneously obese (db/db) insulin-resistant mice. J. Clin. Invest. 85, Katz, E. B., Stenbit, A. E., Hatton, K., DePinho, R., and Charron, M. J. (1995) Cardiac and adipose tissue abnormalities but not diabetes in mice deficient in GLUT4. Nature (London) 377, Houseknecht, K. L., Zhu, A. X., Gnudi, L., Hamann, A., Zierath, J. R., Tozzo, E., Flier, J. S., and Kahn, B. B. (1996) Overexpression of Ha-ras selectively in adipose tissue of transgenic mice. Evidence for enhanced sensitivity to insulin. J. Biol. Chem. 271, Hirsch, J., and Gallian, E. (1968) Methods for the determination of adipose cell size in man and animals. J. Lipid Res. 9, Cushman, S. W., and Salans, L. B. (1978) Determination of adipose cell size and number in suspensions of isolated rat and human adipose cells. J. Lipid Res. 19, Smith-Hall, J., Pons, S., Patti, M. E., Burks, D. J., Yenush, L., Sun, X. J., Kahn, C. R., and White, M. F. (1997) The 60 kda insulin receptor substrate functions like an IRS protein (pp60irs3) in adipose cells. Biochemistry 36, Kerouz, N. J., Hörsch, D., Pons, S., and Kahn, C. R. (1997) Differential regulation of insulin receptor substrate-1 and -2 (IRS-1 and IRS-2) and phosphatidylinositol 3-kinase isoforms in liver and muscle of the obese diabetic (ob/0b) mouse. J. Clin. Invest. 100, Lavan, B. E., and Lienhard, G. E. (1993) The insulin-elicited 60-kDa phosphotyrosine protein in rat adipocytes is associated with phosphatidylinositol 3-kinase. J. Biol. Chem. 268, Shepherd, P. R., Nave, B. T., and O Rahilly, S. (1996) The role of phosphoinositide 3-kinase in insulin signalling. J. Mol. Endocrinol. 17, Shepherd, P. R., Siddle, K., and Nave, B. T. (1997) Is stimulation of class-1 phosphatidylinositol 3-kinase activity by insulin sufficient to activate pathways involved in glucose metabolism? Biochem. Soc. Trans. 25, Sharma, P., Egawa, K., Gustafson, T., Martin, J., and Olefsky, J. (1997) Adenovirus-mediated overexpression of IRS-1 interacting domains abolishes insulin-stimulated mitogenesis without affecting glucose transport in 3T3 L1 adipocytes. Mol. Cell. Biol. 17, Staubs, P., Nelson, J., Reichart, D., and Olefsky, J. (1998) Platelet-derived growth factor inhibits insulin stimulation of insulin substrate-1-associated phosphatidylinositol 3-kinase in 3T3 L1 adipocytes without affecting glucose transport. J. Biol. Chem. 273, Sharma, P., Egawa, K., Huang, Y., Martin, J., Huvar, I., Boss, G., and Olefsky, J. (1998) Inhibition of phosphatidylinositol 3-kinase activity by adenovirus-mediated gene transfer and its effect on insulin action. J. Biol. Chem. 273, Heller-Harrison, R. A., Morin, M., Guilherme, A., and Czech, M. P. (1996) Insulin-mediated targeting of phosphatidylinositol 3-kinase to GLUT4-containing vesicles. J. Biol. Chem. 271, Morris, A. J., Martin, S. S., Haruta, T., Nelson, J. G., Vollenweider, P., Gustafson, T. A., Mueckler, M., Rose, D. W., and Olefsky, J. M. (1996) Evidence for an insulin receptor substrate 1 independent insulin signaling pathway that mediates insulinresponsive glucose transporter (GLUT4) translocation. Proc. Natl. Acad. Sci. USA 93, Received for publication August 3, Revised for publication January 13, INSULIN SIGNALING IN GLUT4( / ) ADIPOCYTES 1125