Potentiation of Glucose Uptake in 3T3-L1 Adipocytes by ... - CiteSeerX

0888-8809/01/$03.00/0 Printed in U.S.A.

Molecular Endocrinology 15(10):1729–1738 Copyright © 2001 by The Endocrine Society

Potentiation of Glucose Uptake in 3T3-L1 Adipocytes by PPAR␥ Agonists Is Maintained in Cells Expressing a PPAR␥ Dominant-Negative Mutant: Evidence for Selectivity in the Downstream Responses to PPAR␥ Activation CLAIRE NUGENT, JOHANNES B. PRINS, JONATHAN P. WHITEHEAD, DAVID SAVAGE, JOHN M. WENTWORTH, V. KRISHNA CHATTERJEE, AND STEPHEN O’RAHILLY Departments of Clinical Biochemistry and Medicine, University of Cambridge, Addenbrooke’s Hospital, Cambridge, United Kingdom, CB2 2QR Pharmacological agonists for the nuclear receptor PPAR␥ enhance glucose disposal in a variety of insulin-resistant states in humans and animals. The precise mechanisms whereby activation of PPAR␥ leads to increased glucose uptake in metabolically active cells remain to be determined. Notably, certain novel, synthetic PPAR␥ ligands appear to antagonize thiazolidinedione-induced adipogenesis yet stimulate cellular glucose uptake. We have explored the molecular mechanisms underlying the enhancement of glucose uptake produced by PPAR␥ agonists in 3T3-L1 adipocytes. Rosiglitazone treatment for 48 h significantly increased basal and insulin-stimulated glucose uptake and markedly increased the cellular expression of GLUT1 but not GLUT4. Rosiglitazone increased plasma membrane levels of GLUT1, but not GLUT4, both basally and after insulin stimulation. Surprisingly, adenoviral expression of a

dominant-negative mutant PPAR␥, which was demonstrated to strongly inhibit adipogenesis, completely failed to inhibit rosiglitazone-stimulated glucose uptake. Similar findings were obtained with the non-thiazolidinedione PPAR␥ agonists, GW1929 and GW7845. The insensitivity of PPAR␥ agonist-stimulated glucose uptake to expression of a dominant-negative mutant, compared with the latter’s marked inhibitory effects on preadipocyte differentiation, suggests that, as is the case for other nuclear receptors, the precise molecular mechanisms linking PPAR␥ activation to downstream events may differ depending on the nature of the biological response. The growing evidence that the effects of PPAR␥ on adipogenesis and glucose uptake can be dissociated may have important implications for the development of improved antidiabetic drug treatments. (Molecular Endocrinology 15: 1729–1738, 2001)

T

PPAR␥ agonists, optimized for potency on hPPAR␥, have been described, the in vitro activity of which closely matches their ability to reduce glucose levels in rodent models of type 2 diabetes (7, 8). In addition to its role in the control of insulin sensitivity, PPAR␥ has also been shown to play a crucial role in adipocyte differentiation. Retroviral expression of PPAR␥ in fibroblasts in the presence of weak PPAR␥ activators resulted in their efficient differentiation to adipocytes, as measured by lipid accumulation, changes in cell morphology, and the expression of an adipocyte-specific pattern of genes (9). Adipogenesis has since been shown to involve a complex interplay between PPAR␥ and other families of transcription factors, notably the CAAT/ enhancer binding proteins (C/EBPs) and adipocyte determination and differentiation factor 1 (ADD1)/sterol regulatory element-binding protein 1 (SREBP1) (10). Furthermore, recent gene knockout studies demonstrated conclusively that PPAR␥ is essential for adipocyte differentiation in vivo (11, 12). It is unclear how activation of a transcription factor that is predominantly expressed in adipose tissue can

HE THIAZOLIDINEDIONE (TZD) class of drugs was developed through empirical compound screening in rodent models of type 2 diabetes (1). These compounds were shown to ameliorate insulin resistance and to lower blood glucose levels, while having no effect on insulin secretion. The mechanism of action of the TZDs (1) was initially unknown until strong evidence emerged in the mid-1990s to suggest that PPAR␥ was the molecular target for their antidiabetic activity. The TZDs were shown to function as selective PPAR␥ ligands (2, 3) whose rank order of potency in activating PPAR␥ in vitro correlated closely with their glucose lowering activity in rodents (4, 5). More recently, further evidence has emerged to strengthen the link between PPAR␥ and systemic insulin action. First, loss of function mutations in human PPAR␥ (hPPAR␥) have been identified which result in extreme insulin resistance and type 2 diabetes (6). Second, a series of non-TZD, tyrosine-derived Abbreviations: Ad-GFP, An adenovirus expressing GFP alone; Ad␥m, adenovirus expressing the PPAR␥1 mutant receptor and GFP; aP2, adipocyte P2; GFP, green fluorescent protein; NBCS, newborn calf serum; TZD, thiazolidinedione.

1729

1730 Mol Endocrinol, October 2001, 15(10):1729–1738

improve insulin sensitization and glucose utilization in muscle, the primary site of glucose disposal. One explanation for this apparent paradox is that PPAR␥ agonists may regulate the storage or secretion of adipocyte-derived signaling factors that influence glucose metabolism in muscle. Candidate molecules include FFA, TNF␣, and leptin (13). Alternatively, despite low levels of PPAR␥ expression in muscle, agonists may directly induce expression of genes involved in glucose homeostasis in this tissue. Several studies have demonstrated an enhancement of glucose uptake and glucose transporter expression in adipocytes treated with TZDs, and there is evidence to suggest that a similar effect may occur in muscle (14). Mukherjee et al. recently reported a novel PPAR␥specific ligand that blocks TZD-induced adipogenesis but stimulates insulin-mediated glucose uptake in 3T3-L1 adipocytes (15). This finding raises the possibility that, as with other nuclear receptors, the biological response to receptor occupation may depend on the precise nature of the ligand involved (16, 17). Herein we report further evidence for the dissociation of downstream biological responses to PPAR␥ activation, on this occasion from the use of genetic, rather than solely pharmacological, tools.

RESULTS Effect of Rosiglitazone on Glucose Uptake in 3T3-L1 Adipocytes Rosiglitazone-treated cells showed a trend toward increased insulin-stimulated glucose uptake at all time points (1–48 h), but the effect was only statistically significant after 48 h supplementation (Fig. 1). Rosiglitazone significantly enhanced basal glucose uptake at both 24 h and 48 h with the effect being maximal at 48 h supplementation. This enhancement of basal glucose uptake could be attributed to a prevention of the time-dependent decrease in basal glucose uptake observed in untreated cells rather than a potentiation of glucose uptake per se. Nevertheless, at 48 h, rosiglitazone-treated cells exhibited a 2.6-fold higher level of basal glucose uptake compared with untreated cells, and insulin-stimulated glucose uptake was enhanced by 1.7-fold (both P ⬍ 0.001). All further experiments were performed with 48 h of rosiglitazone treatment. Effect of Rosiglitazone on the Expression and Cellular Localization of Glucose Transporters Rosiglitazone increased total cellular levels of GLUT1 by 1.5-fold (P ⬍ 0.001) but had no discernible effect on levels of GLUT4 (Fig. 2A). Plasma membrane lawns from disrupted cells were probed with anti-GLUT1 or GLUT4 antibodies followed by immunofluorescence detection (Fig. 2B). Consistent with previously published data (18), insulin treatment of 3T3-L1 adipocytes produced a 2-fold increase in GLUT1 and a

Nugent et al. • PPAR␥ Agonists Stimulate Glucose Uptake

Fig. 1. Effect of Rosiglitazone on Glucose Uptake 3T3-L1 adipocytes (day 7–9 post differentiation) were cultured in six-well plates and supplemented with 10⫺7 M rosiglitazone in serum-free DMEM (1 h, 2 h, 4 h, and 6 h) or in DMEM/FBS followed by 2 h in serum-free DMEM (24 h and 48 h). 2-Deoxyglucose uptake was measured over 5 min following stimulation ⫾ insulin for 30 min. Data are mean uptakes ⫾ SE from four or more independent experiments performed in triplicate, normalized to vehicle insulin-stimulated uptake at each time point (mean insulin responses, 16, 800 dpm/well). Open columns, Vehicle; filled columns, rosiglitazone.

4.5-fold increase in GLUT4 at the plasma membrane. Rosiglitazone increased levels of GLUT1 at the plasma membrane in the absence and presence of insulin by 1.8-fold and 1.3-fold, respectively (both P ⬍ 0.001). Rosiglitazone had no effect on levels of GLUT4 at the plasma membrane. Transduction of 3T3-L1 Preadipocytes with a Dominant-Negative PPAR␥ Mutant A compound hPPAR␥1 mutant has previously been described in which the highly conserved hydrophobic and charged residues, Leu (468) and Glu (471), in helix 12 of the ligand-binding domain were mutated to alanine (19). This mutant receptor retains ligand and DNA binding, but exhibits reduced transactivation due to impaired coactivator recruitment. In addition, it silences basal transcription, recruits corepressors more avidly than wild-type PPAR␥, and exhibits delayed ligand-dependent corepressor release. The L468A/ E471A hPPAR␥1 mutant is also a potent dominantnegative inhibitor of wild-type PPAR␥ action, markedly inhibiting the effects of TZDs on human preadipocyte differentiation and aP2 expression (19). We first established that Ad␥m [an adenovirus expressing the PPAR␥1 mutant receptor and green fluorescent protein (GFP)] could infect the 3T3-L1 cell line, resulting in a similar marked inhibition of differentiation induced by the standard differentiation cocktail (Fig. 3A). Infection with Ad-GFP (an adenovirus expressing GFP alone)


Mol Endocrinol, October 2001, 15(10):1729–1738

1731

Fig. 2. Effect of Rosiglitazone on Glucose Transporters 3T3-L1 adipocytes (day 7 post differentiation) were supplemented with 10⫺7 M rosiglitazone in DMEM/FBS for 48 h. A, Western blotting. Cells were lysed as described in Materials and Methods. Forty micrograms were resolved by SDS-PAGE and immunoblotted with anti-GLUT1 or GLUT4 antibody. A representative gel is shown for each data set. Numerical data are percentage means ⫾ SE obtained by quantitation of gels from eight independent experiments, normalized to transporter levels in vehicle-supplemented cells. Open columns, Vehicle; filled columns, rosiglitazone. B, Plasma membrane lawn assay. Cells were serum starved in DMEM for 2 h and stimulated with and without insulin for 30 min before preparation of plasma membrane lawns and assay of glucose transporter translocation. Representative images from a typical experiment are shown. Data from each experiment, utilizing 16 fields for each condition, were quantified as described in Materials and Methods, and overall results are shown as means ⫾ SE from five independent experiments. Open columns, Unstimulated; filled columns, 10 nM insulin.

had no effect on 3T3-L1 differentiation. Ad␥m infection of the 3T3-L1 cell line resulted in a marked increase in levels of PPAR␥1 expression (Fig. 3B). Infection with Ad-GFP had no effect on the endogenous levels of PPAR␥ (data not shown). Expression of the dominantnegative PPAR␥ in 3T3-L1 preadipocytes significantly inhibited the induction of the differentiation marker, aP2, by rosiglitazone (P ⬍ 0.05 compared with both

Ad-GFP and nil virus controls) (Fig. 3C). Furthermore, Ad␥m significantly inhibited adipocyte differentiation in response to the conventional differentiation cocktail (P ⬍ 0.05 compared with both Ad-GFP and nil virus controls) (Fig. 3D). Thus, these studies provide evidence that the human L468A/E471A PPAR␥ mutant receptor can block endogenous PPAR␥ activity in a murine preadipocyte cell line.



Fig. 3. Transduction of 3T3-L1 Preadipocytes with Dominant-Negative PPAR␥ Confluent preadipocytes in 24-well plates were infected for 12 h with 1 ⫻ 109 plaque forming units (pfu)/well Ad-GFP or Ad␥m. A, Microscopy. Cells were differentiated as described in Materials and Methods for 7 d, fixed with 0.5% glutaraldehyde, and stained with Oil Red O. Representative images from a typical experiment are shown. Left panel, Macroscopic view of individual wells of 24-well plate; right panel, microscopic view of same cells. B–D, Medium was changed to DMEM/FBS with differentiation mix or DMEM/NBCS ⫾ 10⫺7 M rosiglitazone for 48 h. B, Western blotting. Infected preadipocytes (minus rosiglitazone) and day 2 adipocytes were scraped and lyzed as described in Materials and Methods. Ten micrograms of total protein were resolved by SDS-PAGE and immunoblotted with anti-PPAR␥ antibody. C, aP2 induction by rosiglitazone in preadipocytes. Infected preadipocytes ⫾ 10⫺7 M rosiglitazone were lyzed and RNA extracted and reverse transcribed. aP2 gene expression was quantified using real time quantitative PCR as described in Materials and Methods. Data are mean percentage rosiglitazone induction of aP2 ⫾ SE D, aP2 induction during differentiation. Infected preadipocytes (minus rosiglitazone) and day 2 adipocytes were lysed and RNA extracted and reverse transcribed. aP2 gene expression was quantified using real time quantitative PCR. Data are mean percentage induction of aP2 during differentiation ⫾ SE. For experiments C and D, data are from four independent experiments performed in triplicate, normalized to the endogenous control, glyceraldehyde-3-phosphate dehydrogenase.

Effect of Dominant-Negative PPAR␥ on the Potentiation of Glucose Uptake by Rosiglitazone To determine whether fully differentiated 3T3-L1 adipocytes could also be successfully infected, cells were incubated overnight with Ad-GFP or Ad␥m. Infectivity was assessed 48 h later by fluorescence microscopy and Western blotting with anti-PPAR␥ antibody (Fig. 4A). Comparable infection efficiencies of 70–90% were observed for both viruses, and Ad␥m-infected cells exhibited an increased PPAR␥1 protein expression compared with Ad-GFP-infected and uninfected cells. Again, adenoviral infection had no effect on endogenous levels of PPAR␥. To assess the effect of the Ad␥m on the potentiation of glucose uptake by rosiglitazone, cells were infected overnight and the 48 h

incubation with rosiglitazone was subsequently commenced. Rosiglitazone was observed to enhance basal glucose uptake in the Ad-GFP-infected cells and there was also a trend for it to increase insulin-stimulated glucose uptake. Contrary to expectations, Ad␥m did not inhibit this rosiglitazone-mediated potentiation of either basal or insulin-stimulated glucose uptake (Fig. 4B). It is noteworthy that there was a trend for Ad␥m to reduce basal and insulin-stimulated glucose uptake in the absence of rosiglitazone, while maintaining a similar fold insulin stimulation over basal, and to reduce insulin-stimulated glucose uptake in the presence of rosiglitazone. Nevertheless, Ad␥m-infected cells treated with rosiglitazone still exhibited an enhanced basal and insulin-stimulated glucose uptake compared with untreated cells.



1733

ligands, optimized for potency on human (h) PPAR␥, are GW1929 and GW7845 (8, 20). GW1929 has been shown to be 2 orders of magnitude more potent than the TZD, troglitazone, in both in vitro and in vivo assays (8). It both lowered plasma glucose levels in Zucker diabetic fatty rats and transactivated hPPAR␥ and mouse PPAR␥ in reporter gene assays more potently than troglitazone. Similarly, GW7845 was shown to be significantly more potent than either rosiglitazone or troglitazone when assayed for induction of adipogenesis in the 3T3-L1 cell line (20). 3T3-L1 adipocytes were supplemented for 48 h with 10⫺7 M rosiglitazone, GW1929, or GW7845 and glucose uptake was measured (Fig. 5). All three compounds had comparable effects on glucose uptake, increasing basal uptake by between 3- and 4-fold (P ⬍ 0.001) and insulin-stimulated uptake by approximately 1.6-fold (P ⬍ 0.01). Having established that the non-TZD PPAR␥ agonists act to increase glucose uptake to a similar extent to the TZD rosiglitazone, we investigated the effect of the dominant-negative PPAR␥ mutant in this system. As before, adipocytes were infected with Ad-GFP or Ad␥m overnight and subsequently supplemented with the GW compounds for 48 h. As with rosiglitazone, the tendency of the GW compounds to increase basal and insulin-stimulated glucose uptake was similar in AdGFP and Ad␥m-infected cells (Fig. 6). In contrast to the lack of effect of Ad␥m on ligand-stimulated glucose uptake, the mutant PPAR␥ inhibited the effects of these ligands on GLUT1 expression (Fig. 7). Fig. 4. Effect of Dominant-Negative PPAR␥ on the Potentiation of Glucose Uptake in 3T3-L1 Adipocytes by Rosiglitazone 3T3-L1 adipocytes (day 7 post differentiation) cultured in 24-well plates were infected for 12 h with 1 ⫻ 109 pfu/well Ad-GFP or Ad␥m. A, Microscopy and Western blotting. Infection efficiency was estimated using fluorescence microscopy. Images from a typical experiment are shown. Images represent an infection efficiency of approximately 90% for both viruses, with individual infected cells demonstrating variable fluorescence intensities relating to variable levels of GFP expression. Infected adipocytes were scraped and lysed as described in Materials and Methods. Ten micrograms of total protein were resolved by SDS-PAGE and immunoblotted with anti-PPAR␥ antibody. A representative gel is shown. B, Glucose uptake. Infected adipocytes were supplemented with 10⫺7 M rosiglitazone in DMEM/FBS for 48 h, serum starved in DMEM for 2 h, and stimulated with and without 10 nM insulin for 30 min. 2-Deoxyglucose uptake was measured over 5 min. Data are mean 2-deoxyglucose uptake ⫾ SE from four independent experiments performed in triplicate, normalized to nil virus, vehicle insulin-stimulated uptake (mean insulin responses, 8,200 dpm/ well). Open columns, Vehicle; filled columns, rosiglitazone.

Effect of Non-TZD PPAR␥ Agonists on Glucose Uptake A novel class of tyrosine-derived, non-TZD, PPAR␥ ligands has recently been developed (7). Two such

Fig. 5. Effect of non-TZD PPAR␥ Agonists on Glucose Uptake 3T3-L1adipocytes (day 7 post differentiation) cultured in six-well plates were supplemented with 10⫺7 M agonist in DMEM/FBS for 48 h and serum-starved in DMEM for 2 h. 2-Deoxyglucose uptake was measured over 5 min following stimulation ⫾ 10 nM insulin for 30 min. Data are mean uptakes ⫾ SE from six or more independent experiments performed in triplicate, normalized to vehicle insulin-stimulated uptake (mean insulin responses 31, 500 dpm/well). Open columns, Unstimulated; filled columns, 10 nM insulin.

1734


Fig. 6. Effect of Dominant-Negative PPAR␥ on the Potentiation Of Glucose Uptake in 3T3-L1 Adipocytes by GW1929 and GW7845 3T3-L1adipocytes (day 7 post differentiation) cultured in 24-well plates were infected for 12 h with 1 ⫻ 109 pfu/well Ad-GFP or Ad␥m. Cells were supplemented with 10⫺7 M agonist in DMEM/FBS for 48 h, serum-starved in DMEM for 2 h, and stimulated ⫾ 10 nM insulin for 30 min. 2-Deoxyglucose uptake was measured over 5 min. Data are mean 2deoxyglucose uptake ⫾ SE from two independent experiments performed in triplicate, normalized to nil virus, vehicle insulin-stimulated uptake (mean insulin responses, 11,360 dpm/well). Open columns, Vehicle; filled columns, agonist.

DISCUSSION TZDs are now widely used for the treatment of type 2 diabetes. Despite this, the precise mechanisms whereby TZDs act to improve insulin sensitivity and glucose disposal in vivo are unclear. In the studies reported above we demonstrate that 48 h of rosiglitazone treatment increases basal and insulin-stimulated glucose uptake and increases cellular expression of GLUT1 in 3T3-L1 adipocytes. Furthermore, we present new information regarding the effects of this compound on the translocation of glucose transporters to the plasma membrane. After recent pharmacological data that the effects of PPAR␥ activation on adipogenesis and glucose uptake can be dissociated (15), we now present evidence that a dominant-negative PPAR␥ mutant has inhibitory effects on ligandstimulated adipogenesis while having no effects on ligand-stimulated glucose uptake. Thus, a body of data is accumulating to suggest that, as is well established with other well studied nuclear receptors (17, 16), selective modulation of the downstream biological responses to PPAR␥ activation may be possible. Several in vivo studies have demonstrated an increase in adipose tissue glucose transport activity and GLUT4 expression after treatment with TZDs (14). Studies in the 3T3-L1 and 3T3-F442A cell lines have similarly reported TZD-induced increases in glucose transport, although effects on glucose transporters appear to be dependent on the differentiation state (21–26). Hence, increases in both GLUT1 and GLUT4 expression have been reported in cells supplemented


Fig. 7. Effect of Dominant-Negative PPAR␥ on the PPAR␥ Agonist-Induced Increase in GLUT1 in 3T3-L1 Adipocytes 3T3-L1 adipocytes (day 7 post differentiation) cultured in 24-well plates were infected for 12 h with 1 ⫻ 109 pfu/well Ad-GFP or Ad␥m. Cells were supplemented with 10⫺7 M agonist in DMEM/FBS for 48 h and scraped and lysed as described in Materials and Methods. Five micrograms of total protein were resolved by SDS-PAGE and immunoblotted with anti-GLUT1 antibody. A representative gel is shown for each data set. V, Vehicle; R, rosiglitazone; 1929, GW1929; 7845, GW7845. Numerical data are mean percentage levels of total cellular GLUT1 ⫾ SE obtained by quantitation of gels from three independent experiments, normalized to GLUT1 levels in vehicle-supplemented cells. Open columns, Vehicle; light gray columns, rosiglitazone; mid gray columns, GW1929; dark gray columns, GW7845.

with TZDs during the differentiation process (21–23), whereas only GLUT1 expression has been shown to be increased in fully differentiated cells (26). In our study, differentiated cells were used to eliminate confounding effects of rosiglitazone on the adipogenic process itself. In agreement with previous TZD studies (24–26), we demonstrated an increase in basal glucose uptake and cellular GLUT1 expression, with no effect on GLUT4 expression, following rosiglitazone treatment of 3T3-L1 adipocytes. In addition to the effects of rosiglitazone on the total cellular expression of GLUT1, we show, for the first time, that this agent increases plasma membrane levels of GLUT1 in both the basal and insulin-stimulated state while having no effects on GLUT4 cellular localization. The marked effects of rosiglitazone on glucose uptake, GLUT1 expression, and GLUT1 translocation occurring in the absence of insulin suggests that the frequently used term “insulin sensitizer” may not fully reflect the molecular mechanisms underlying the effects of this compound on glucose disposal. The in vivo importance of PPAR␥ in the regulation of glucose disposal and insulin sensitivity in the whole


organism is becoming increasingly apparent. In addition to the large body of evidence derived from the efficacy of TZD and non-TZD PPAR␥ agonists in insulin-resistant and diabetic states in rodents, primates, and humans (1, 7, 8), genetic studies have recently provided compelling evidence for a direct link between PPAR␥ and systemic insulin action. Thus, two independent naturally occurring dominant-negative mutations in hPPAR␥ have been reported in subjects with severe insulin resistance and type 2 diabetes (6). PPAR␥ gene knockout studies in mice, however, were unable to substantiate this result due to embryonic lethality (11, 12). Surprisingly, mice heterozygous for PPAR␥ deficiency exhibited an improved insulin sensitivity and were protected from high-fat diet-induced insulin resistance (27, 28). While direct effects of TZDs on glucose uptake have been demonstrated in cell lines, it is still unclear whether their in vivo effects on glucose disposal can be attributed to direct or indirect mechanisms. Thus, TZDs have effects on suppression of the adipocytederived signaling molecules, TNF␣ and leptin, and on FFA partitioning between adipose and muscle tissues (29, 30), all of which may secondarily improve whole organism insulin sensitivity. To investigate whether the potentiating effects of rosiglitazone on glucose uptake in 3T3-L1 adipocytes were mediated via PPAR␥, the experiment was repeated in cells expressing a dominant-negative hPPAR␥ mutant receptor. This mutant form of hPPAR␥ has previously been shown to powerfully inhibit cotransfected wild-type receptor action and to inhibit the TZD-induced differentiation of human preadipocytes (19). We demonstrated that the transduced hPPAR␥ mutant receptor was highly expressed in murine 3T3-L1 preadipocytes and adipocytes and that it inhibited 3T3-L1 differentiation induced by both the standard differentiation cocktail and rosiglitazone. However, there was no discernible effect on rosiglitazone-induced enhancement of glucose uptake, either basal or insulin-stimulated, in cells transduced with the mutant receptor. There have been previous suggestions that TZDs, particularly troglitazone (31), may have PPAR␥-independent metabolic effects. However, the observation that the potentiating effects of two structurally distinct, non-TZD PPAR␥ agonists on glucose uptake were also completely unaffected by the mutant receptor, would imply that a PPAR␥-independent effect on glucose uptake by all three compounds is unlikely. The ability of Ad␥m to block the PPAR␥ agonist-induced increases in total cellular levels of GLUT1 implies that the potentiation of glucose uptake is mediated via a mechanism independent of the increase in total cellular GLUT1 protein. Such mechanisms might include increasing the translocation of GLUT1 to the plasma membrane, as described above, and/or an increase in the intrinsic activity of the glucose transporters. There are several potential explanations to account for the insensitivity of the PPAR␥ agonist-induced po-


1735

tentiation of glucose uptake to the presence of the mutant receptor. A failure to express the mutant adequately is possible, but we demonstrated that the adenoviral expression system consistently led to more than 70% infectivity in preadipocytes and adipocytes, while Western blotting indicated a more than 30-fold increase of PPAR␥ expression in the transduced cells. Another possibility is that PPAR␥ ligands may exert their effects on glucose uptake via a nongenomic mechanism, something that has been established to occur with ligands for other nuclear hormone receptors (32). However, our time course studies indicated that rosiglitazone had its maximal effects after 48 h of exposure, a finding much more consistent with a transcriptional mechanism than with the expected rapid nature of the nongenomic responses to steroids and other lipophilic ligands (32). While we did not undertake formal tests of the dependence of the enhancement of glucose uptake on new protein synthesis, Kreutter et al. (25) reported previously that the enhancement of basal glucose uptake by CP 68722 (racemic englitazone) in 3T3-L1 adipocytes can be inhibited by cycloheximide. Perhaps the most compelling model to explain our data relates to the fact that the occupancy of nuclear receptors by ligands may lead to the activation of selective sets of downstream biological responses that depend on the precise nature of the activating ligand. Thus, tamoxifen and raloxifene, both highaffinity agonists for the ER, mimic the natural ligand in preventing postmenopausal osteoporosis but inhibit the estrogen-dependent proliferation of breast carcinomas (17). The possibility that such selective receptor modulation might occur with PPAR␥ was recently reported by Mukherjee et al. (15). A synthetic highaffinity PPAR␥ agonist, LG100641, was shown to block both TZD-induced differentiation and target gene activation and repression in 3T3-L1 cells, yet it also enhanced basal and insulin-stimulated glucose uptake in fully mature cells. Selective receptor modulation is thought to relate to the fact that the pattern of receptor interaction with the complex of coactivator and corepressor proteins is dependent on the precise conformation of the particular ligand-receptor complex (16). Compared with other nuclear receptors, PPAR␥ has a particularly large ligand-binding pocket, and there is already strong evidence that different ligands make distinct structural contacts with receptor (33, 34). Thus the a priori possibility of selective receptor modulation is particularly strong for PPAR␥. The growing evidence for the dissociability of biological responses downstream from nuclear hormone receptors has come largely from pharmacological rather than genetic studies. Our studies suggest that a mutant form of PPAR␥, which acts as powerful dominant-negative repressor of adipogenesis, may have no effect on the stimulation of glucose uptake, providing further support for the idea that the repertoire of coactivators and corepressors involved in the promotion of adipogenesis and

1736


glucose uptake may be distinct. The existing structural model of PPAR␥ would predict that mutating both L468 and E471 would grossly interfere with the interaction of the AF2 helix with both ligand and coactivator (19, 35). This suggests that other areas of the PPAR␥ molecule, e.g. the N-terminal activation domain, may be more intimately involved with transcriptional responses relevant to glucose uptake. Indeed, the coactivators PGC1 and PGC2 are not thought to require the AF2 helix for binding to PPAR␥ (36, 37). In summary, we have demonstrated that PPAR␥ agonists potentiate basal and insulin-stimulated glucose uptake in 3T3-L1 adipocytes and increase total cellular and plasma membrane expression of GLUT1. The potentiation of glucose uptake is maintained in cells expressing a PPAR␥ mutant that inhibits adipogenesis, thus providing further evidence for selectivity in the downstream responses to PPAR␥ activation. If such selective modulation of receptor function could be achieved with PPAR␥, drugs might be developed that could improve insulin sensitivity without promoting fat accumulation. In this regard, a PPAR␥ agonist, GW0072, which lowers plasma insulin and triglycerides in insulin-resistant Zucker rats without causing weight gain, has recently been described (T. M. Willson, personal communication).


PPAR␥ Agonist Solutions Solutions (10⫺4 M) of rosiglitazone, GW1929, and GW7845 were prepared in dimethylsulfoxide. These were diluted 1:1,000 into serum-free DMEM for studies requiring an incubation time ⱕ6 h, or into DMEM/FBS for studies with an incubation time ⱖ6 h, to give a final concentration of 10⫺7 M. Glucose Uptake Adipocytes (at day 7–9 after initiation of differentiation) in 6 (or 24)-well plates were incubated in medium containing 10⫺7 M PPAR␥ agonist for the specified length of time. In the time course experiment, rosiglitazone incubations were terminated simultaneously on day 9. Cells that had been supplemented for 24 h and 48 h were serum starved in DMEM for 2 h before the assay. Glucose uptake assays were performed as described previously (18). Western Blotting Treated cells (adipocytes/preadipocytes) were solubilized by scraping and passing 10 times through a 25G needle in lysis buffer as described previously (18). The lysate was clarified by centrifugation at 13,500 ⫻ g for 10 min at 4 C. Crude cell extracts were resolved by SDS-PAGE before electroblotting to polyvinylidene difluoride membranes (Millipore Corp., Bedford, MA). Membranes were blocked in 1% BSA, and specific proteins were detected by incubation with appropriate primary and secondary (horseradish peroxidase-conjugated) antibodies in 150 mM NaCl, 50 mM Tris, 0.1% Tween 20. Proteins were then visualized using an enhanced chemiluminescence kit.

MATERIALS AND METHODS Plasma Membrane Lawn Assay Materials 3

-2-Deoxy-D-[2,6- H]glucose and the enhanced chemiluminescence kit were purchased from Amersham Pharmacia Biotech (Piscataway, NJ). Rosiglitazone was provided by Dr. S. Smith (SmithKline Beecham Pharmaceuticals). GW-1929 and GW-7845 were a gift from Dr. T. Willson (GlaxoSmithKline). Rabbit anti-GLUT4 antibody was a gift from Professor G. Gould (University of Glasgow, Glasgow, UK), rabbit anti-GLUT1 was a gift from Professor S. Baldwin (University of Leeds, Leeds, UK), and rabbit anti-PPAR␥ was a gift from Dr. M. Lazar (University of Pennsylvania School of Medicine, Philadelphia, PA). The RNeasy total RNA kit was from QIAGEN (Valencia, CA). Reverse transcription reagents were obtained from Promega Corp. (Madison, WI) and TaqMan reagents were from PE Applied Biosystems (Foster City, CA). All other reagents were from Sigma (St. Louis, MO). Tissue Culture 3T3-L1 fibroblasts (ATCC, Manassas, VA) were maintained at no higher than 70% confluence in DMEM containing 10% newborn calf serum (NBCS), 25 mM glucose, 2 mM glutamine, and antibiotics (DMEM/NBCS). For differentiation they were grown 2 d post confluence in DMEM/NBCS and then for 2 d in medium containing FBS (DMEM/FBS) supplemented with 0.83 ␮M insulin, 0.25 ␮M dexamethasone, and 0.5 mM isobutylmethylxanthine. The medium was then changed to DMEM/ FBS supplemented only with 0.83 ␮M insulin for 2 d and then to DMEM/FBS alone for an additional 3–5 d. Differentiated cells were only used when at least 95% of the cells showed an adipocyte phenotype by accumulation of lipid droplets.

3T3-L1 adipocytes (day 7), grown on collagen-coated glass coverslips, were treated for 48 h with 10⫺7 M rosiglitazone in DMEM/FBS. Cells were serum starved for 2 h and incubated with and without 10 nM insulin for 30 min, and a modified version of the plasma membrane lawn assay (18) was performed. Cells were washed twice in ice-cold buffer A (50 mM HEPES, 10 mM NaCl, pH 7.2), twice in ice-cold buffer B (20 mM HEPES, 10 mM KCl, 2 mM CaCl2, 1 mM MgCl2, pH 7.2), and sonicated using a probe sonicator (Kontes, Vineland, NJ) to generate a lawn of plasma membrane fragments attached to the coverslip. The membranes were washed twice again in ice-cold buffer B and fixed to the coverslips for 15 min using freshly prepared 3% paraformaldehyde. Membranes were then serially washed: ⫻3 PBS, ⫻3 50 mM NH4Cl in PBS over 10 min, ⫻3 PBS, ⫻3 PBS-gelatin (PBS containing 0.2% gelatin and 1 ␮l/ml goat serum) over 5 min, and finally ⫻3 PBS. Membranes were incubated in either anti-GLUT4 or anti-GLUT1 antibody (1:100 dilution in PBS-gelatin) for 1 h at room temperature. After washing ⫻3 PBS-gelatin and ⫻3 PBS, the coverslips were incubated with the secondary antibody, fluorescein isothiocyanate-conjugated donkey antirabbit IgG, for 1 h at room temperature, washed ⫻3 PBSgelatin and ⫻3 PBS and mounted on glass slides. Coverslips were viewed using a 60⫻ objective lens on a Optiphot-2/ Biorad MRC-1000 microscope (Nikon, Melville, NY) operated in laser scanning confocal mode. Samples were illuminated at 488 nm, and images were collected at 510 nm. Duplicate coverslips were prepared at each experimental condition, and eight random images of plasma membrane lawn were collected from each. The images were quantified using MRC1000 confocal microscope operating software [CoMOS, version 6.05.8 (Bio-Rad Laboratories, Inc., Hercules, CA)], on an AST premmia SE P/60 personal computer.



Adenovirus Expression

erator-activated receptor-␥ (PPAR␥). J Biol Chem 270: 12953–12956 Forman BM, Tontonoz P, Chen J, Brun RP, Spiegelman BM, Evans RM 1995 15-Deoxy-⌬12,14-prostaglandin J2 is a ligand for the adipocyte determination factor PPAR␥. Cell 83:803–812 Willson TM, Cobb JE, Cowan DJ, et al. 1996 The structure-activity relationship between peroxisome proliferator-activated receptor-␥ agonism and the antihyperglycemic activity of thiazolidinediones. J Med Chem 39: 665–668 Berger J, Bailey P, Biswas C, et al. 1996 Thiazolidinediones produce a conformational change in peroxisomal proliferator-activated receptor-␥: binding and activation correlate with antidiabetic actions in db/db mice. Endocrinology 137:4189–4195 Barroso I, Gurnell M, Crowley VE, et al. 1999 Dominant negative mutations in human PPAR␥ associated with severe insulin resistance, diabetes mellitus and hypertension. Nature 402:880–883 Henke BR, Blanchard SG, Brackeen MF, et al. 1998 N-(2-Benzoylphenyl)-L-tyrosine PPAR␥ agonists. 1. Discovery of a novel series of potent antihyperglycemic and antihyperlipidemic agents. J Med Chem 41:5020–5036 Brown KK, Henke BR, Blanchard SG, et al. 1999 A novel N-aryl tyrosine activator of peroxisome proliferator-activated receptor-␥ reverses the diabetic phenotype of the Zucker diabetic fatty rat. Diabetes 48:1415–1424 Tontonoz P, Hu E, Spiegelman BM 1994 Stimulation of adipogenesis in fibroblasts by PPAR␥2, a lipid-activated transcription factor. Cell 79:1147–1156 Lowell BB 1999 PPAR␥: an essential regulator of adipogenesis and modulator of fat cell function. Cell 99: 239–242 Barak Y, Nelson MC, Ong ES, et al. 1999 PPAR␥ is required for placental, cardiac, and adipose tissue development. Mol Cell 4:585–595 Rosen ED, Sarraf P, Troy AE, et al. 1999 PPAR␥ is required for the differentiation of adipose tissue in vivo and in vitro. Mol Cell 4:611–617 Spiegelman BM 1998 PPAR␥: adipogenic regulator and thiazolidinedione receptor. Diabetes 47:507–514 Ciaraldi T, Henry RR 1997 Thiazolidinediones and their effects on glucose transporters. Eur J Endocrinol 137: 610–612 Mukherjee R, Hoener PA, Jow L, et al. 2000 A selective peroxisome proliferator-activated receptor-␥ (PPAR␥) modulator blocks adipocyte differentiation but stimulates glucose uptake in 3T3–L1 adipocytes. Mol Endocrinol 14:1425–1433 McKenna NJ, O’Malley BW 2000 An issue of tissues: divining the split personalities of selective estrogen receptor modulators. Nat Med 6:960–962 Levenson AS, Jordan VC 1999 Selective oestrogen receptor modulation: molecular pharmacology for the millennium. Eur J Cancer 35:1974–1985 Urso B, Cope DL, Kalloo-Hosein HE, et al. 1999 Differences in signaling properties of the cytoplasmic domains of the insulin receptor and insulin-like growth factor receptor in 3T3–L1 adipocytes. J Biol Chem 274: 30864–30873 Gurnell M, Wentworth JM, Agostini M, et al. 2000 A dominant-negative peroxisome proliferator-activated receptor-␥ (PPAR␥) mutant is a constitutive repressor and inhibits PPAR␥-mediated adipogenesis. J Biol Chem 275:5754–5759 Suh N, Wang Y, Williams CR, et al. 1999 A new ligand for the peroxisome proliferator-activated receptor-␥ (PPAR␥), GW7845, inhibits rat mammary carcinogenesis. Cancer Res 59:5671–5673 Kletzien RF, Clarke SD, Ulrich RG 1992 Enhancement of adipocyte differentiation by an insulin-sensitizing agent. Mol Pharmacol 41:393–398

Recombinant adenoviruses were generated as described previously (19), expressing GFP (Ad-GFP) or GFP and fulllength L468A/E471A hPPAR␥1 (Ad␥m). 3T3-L1 preadipocyte (2 d post confluence) or day 7 adipocyte cultures in 24-well plates were infected with recombinant virus by addition of 1 ⫻ 109 plaque-forming units/well. Twelve hours later medium containing free virus was removed, and appropriate experimental medium was added. Comparable viral infection efficiency was verified by microscopy using an axiovert 135 inverted fluorescence microscope (Carl Zeiss, Thornwood, NY). Only cells with more than 70% infectivity were used in experiments.

3.

Oil Red O Staining

6.

3T3-L1 adipocytes were fixed with 0.5% glutaraldehyde and stained with Oil Red O for visualization of lipid droplets, according to conventional methods (38). Differentiation efficiency was assessed macroscopically and microscopically using a Nikon Eclipse TE300 inverted microscope.

7.

RNA Extraction/Quantitative RT-PCR

8.

Virally infected preadipocytes/day 2 adipocytes were scraped and total RNA was extracted using the RNeasy mini kit from QIAGEN. Adipocyte P2 (aP2) gene expression was then quantified using real time quantitative RT-PCR. Briefly, cDNA was prepared from 100 ng of RNA using 200 U Moloney-murine leukaemia virus reverse transcriptase (Promega Corp.). Real time quantitative PCR was performed using an ABI-PRISM 7700 Sequence Detection System instrument and software (PE Applied Biosystems, Inc., Foster City, CA) as described previously (39). The primers and probes for aP2 were as follows: Forward, CACCGCAGACGACAGGAAG; Reverse, GCACCTGCACCAGGG; Probe, TGAAGAGCATCAAACCCTAGATGGCGG (all 5⬘-3⬘). Results were normalized to the endogenous control, glyceraldehyde-3-phosphate dehydrogenase.

4.

5.

9. 10. 11. 12. 13. 14.

Statistical Analysis Data are presented as mean ⫾ SE. Statistical significance of treatments was determined using the paired t test (*, P ⬍ 0.05; **, P ⬍ 0.01; ***, P ⬍ 0.001).

15.

Acknowledgments

16.

We would like to thank Dr. Tim Wilson (Glaxo SmithKline) for his kind gift of GW1929 and GW7845. Received January 8, 2001. Accepted June 28, 2001. Address all correspondence and requests for reprints to: Stephen O’Rahilly, Department of Medicine, University of Cambridge, Addenbrooke’s Hospital, Cambridge, United Kingdom, CB2 2QR. E-mail: [email protected]. C.N. was the recipient of a studentship from Diabetes UK. S.O.R., J.B.P., J.P.W., J.M.W., D.S., and V.K.K.C. are supported by the Wellcome Trust.

REFERENCES 1. Day C 1999 Thiazolidinediones: a new class of antidiabetic drugs. Diabet Med 16:179–192 2. Lehmann JM, Moore LB, Smith-Oliver TA, Wilkison WO, Willson TM, Kliewer SA 1995 An antidiabetic thiazolidinedione is a high affinity ligand for peroxisome prolif-

17. 18.

19.

20.

21.

1737


22. Sandouk T, Reda D, Hofmann C 1993 The antidiabetic agent pioglitazone increases expression of glucose transporters in 3T3–F442A cells by increasing messenger ribonucleic acid transcript stability. Endocrinology 133:352–359 23. Shimaya A, Kurosaki E, Shioduka K, Nakano R, Shibasaki M, Shikama H 1998 YM268 increases the glucose uptake, cell differentiation, and mRNA expression of glucose transporter in 3T3–L1 adipocytes. Horm Metab Res 30:543–548 24. Stevenson RW, Hutson NJ, Krupp MN, et al. 1990 Actions of novel antidiabetic agent englitazone in hyperglycemic hyperinsulinemic ob/ob mice. Diabetes 39: 1218–1227 25. Kreutter DK, Andrews KM, Gibbs EM, Hutson NJ, Stevenson RW 1990 Insulin-like activity of new antidiabetic agent CP 68722 in 3T3–L1 adipocytes. Diabetes 39:1414–1419 26. Tafuri SR 1996 Troglitazone enhances differentiation, basal glucose uptake, and Glut1 protein levels in 3T3–L1 adipocytes. Endocrinology 137:4706–4712 27. Miles PD, Barak Y, He W, Evans RM, Olefsky JM 2000 Improved insulin-sensitivity in mice heterozygous for PPAR␥ deficiency. J Clin Invest 105:287–292 28. Kubota N, Terauchi Y, Miki H, et al. 1999 PPAR␥ mediates high-fat diet-induced adipocyte hypertrophy and insulin resistance. Mol Cell 4:597–609 29. Willson TM, Brown PJ, Sternbach DD, Henke BR 2000 The PPARs: from orphan receptors to drug discovery. J Med Chem 43:527–550 30. Martin G, Schoonjans K, Staels B, Auwerx J 1998 PPAR␥ activators improve glucose homeostasis by stimulating


31.

32.

33.

34.

35.

36.

37.

38. 39.

fatty acid uptake in the adipocytes. Atherosclerosis 137: S75–80 Jones SA, Moore LB, Shenk JL, et al. 2000 The pregnane X receptor: a promiscuous xenobiotic receptor that has diverged during evolution. Mol Endocrinol 14:27–39 Schmidt BM, Gerdes D, Feuring M, Falkenstein E, Christ M, Wehling M 2000 Rapid, nongenomic steroid actions: a new age? Front Neuroendocrinol 21:57–94 Gampe RT, Montana VG, Lambert MH, et al. 2000 Asymmetry in the PPAR␥/RXR␣ crystal structure reveals the molecular basis of heterodimerization among nuclear receptors. Mol Cell 5:545–555 Oberfield JL, Collins JL, Holmes CP, et al. 1999 A peroxisome proliferator-activated receptor-␥ ligand inhibits adipocyte differentiation. Proc Natl Acad Sci USA 96: 6102–6106 Nolte RT, Wisely GB, Westin S, et al. 1998 Ligand binding and co-activator assembly of the peroxisome proliferator-activated receptor-␥. Nature 395:137–143 Puigserver P, Adelmant G, Wu Z, et al. 1999 Activation of PPAR␥ coactivator-1 through transcription factor docking. Science 286:1368–1371 Castillo G, Brun RP, Rosenfield JK, et al. 1999 An adipogenic cofactor bound by the differentiation domain of PPAR␥. EMBO J 18:3676–3687 Green H, Kehinde O 1974 Sublines of mouse 3T3 cells that accumulate lipid. Cell 1:113–116 Wang T, Brown MJ 1999 mRNA quantification by real time TaqMan polymerase chain reaction: validation and comparison with RNase protection. Anal Biochem 269:198–201