Development of Insulin-Responsive Glucose Uptake and GLUT4 Expression in Differentiating Human Adipocyte Precursor Cells

International Journal of Obesity (1998) 22, 448±453 ß 1998 Stockton Press All rights reserved 0307±0565/98 $12.00 http://www.stockton-press.co.uk/ijo Development of insulin-responsive glucose uptake and GLUT4 expression in differentiating human adipocyte precursor cells

H Hauner, K RoÈhrig, M Spelleken, LS Liu and J Eckel

Diabetes Research Institute at the Heinrich-Heine-University DuÈsseldorf, D-40225 DuÈsseldorf, Germany

OBJECTIVE: In differentiating human preadipocytes glucose uptake in the presence of insulin is a prerequisite for lipid accumulation. The aim of this study was to characterize the insulin-regulated glucose transport system during and after differentiation. DESIGN AND METHODS: Human adipocyte precursor cells kept in primary culture were allowed to differentiate into fat cells under serum-free hormone-supplemented conditions. 2-Deoxy-glucose uptake was measured as a functional parameter of the glucose transport system, the amount of GLUT1 and GLUT4 protein was determined by Western blotting. RESULTS: In the undifferentiated state, cells did not increase 2-deoxy-glucose uptake in response to insulin. On day 16, when cells have acquired the adipocyte phenotype, there was a 3±4-fold stimulation of glucose transport by insulin compared to basal rates, whereas basal glucose uptake was dramatically diminished. Measurement of GLUT4 protein in cell extracts, showed a marked increase in the amount of this insulin-regulated transporter isoform during the differentiation period. On average, the amount of GLUT4 was 16.7-fold greater after than before differentiation. In contrast, the amount of GLUT1 protein decreased during differentiation to almost undetectable levels on day 16. When newly developed adipocytes were maintained in culture for another 14 d, the stimulation of glucose uptake and the amount of GLUT4 remained stable. CONCLUSION: Differentiating human fat cells in primary culture develops an insulin-responsive glucose transport system which exhibits a high stability, thereby providing a valuable model for long-term studies of glucose transport and GLUT4 expression in human adipocytes.

Keywords: adipose differentiation; glucose transport; GLUT1; GLUT4; human adipose tissue

Introduction However, it is well known that clonal preadipocyte cell lines of rodent origin differ in many aspects from other probably more physiological models, including Many studies have demonstrated that insulin stimu- human adipocyte precursor cells in primary culture.9 lates glucose transport in adipocytes by rapid trans- For example, while 3T3 L1 cells require critical cell location of glucose transporters from an intracellular divisions to enter the differentiation process,10 human membrane pool to the cell surface.1±4 It is now adipocyte precursor cells kept in primary culture are established that the insulin effect on glucose uptake able to develop into fat cells without in vitro mitosis, in adipose tissue depends on the presence of the as these cells are apparently in a later stage of the glucose transporter isoform GLUT4.5 In addition, differentiation programme.11 At present, a human adipocytes also express GLUT1, another glucose white adipocyte cell line is not available. Therefore, transporter which is responsible for basal glucose the only possibility to perform long-term studies in uptake, but is also partially responsive to insulin at human adipose tissue is to use the model of in vitro least in 3T3 L1 cells. Kinetic studies in this clonal differentiating human adipocyte in primary culture.12 adipocyte cell line have shown that GLUT4 is not The aim of this study was to characterize the insulin- present in the ®broblastic state, but is progressively regulated glucose transport system in cultured human expressed during the course of adipose differentiation, preadipocytes during the course of differentiation whereas GLUT1 is highly expressed in the undiffer- and to study the usefulness of human fat cells entiated state.6±8 Such models of rodent origin have developed in vitro for long-term studies of glucose been widely used in studies to investigate the control transport. of glucose transport in general and the regulation of GLUT4 expression in particular.3,4 Materials and methods Correspondence: Hans Hauner, MD, Diabetes Research Institute, Heinrich-Heine-University DuÈ sseldorf, Auf'm Hennekamp 65, D-40225 DuÈ sseldorf, Germany. Received 11 February 1997; revised 3 November 1997; accepted Triiodothyronine (T3), human transferrin, pantothe- 9 January 1998 nate and bovine serum albumin (BSA) were obtained Insulin responses in adipocyte precursor cells H Hauner et al 449 from Sigma (Munich, Germany); 1-methyl-3-iso- 2±3 d. In some experiments, cells were kept in culture butylxanthine (MIX) was purchased from Serva (Hei- for up to 30 d. delberg, Germany); collagenase CLS type 1 was obtained from Worthington (Freehold, NJ, USA); human insulin and cortisol were kindly donated by Hoechst (Frankfurt, Germany); culture media and Glucose transport assay 2-Deoxy-D-glucose uptake was determined as a func- fetal calf serum were obtained from Gibco (Berlin, tional parameter of the glucose transport system. Germany); and all other chemicals were from Boeh- Assays were performed at different time points as ringer (Mannheim, Germany) or Merck (Darmstadt, indicated in the Results. The glucose concentration Germany). Sterile plasticware for tissue culture was was reduced to 5 mmol=l, 24 h prior to the assay, and purchased from Flow Laboratories (Irvine, Scotland). both insulin and Hepes were completely removed 2-Deoxy-D-(1-3H)glucose (15 Ci=mmol) and 125I- from the medium. For assessment of the stimulatory labelled protein A (30 mCi=mg) were obtained from effect of insulin, some dishes were incubated with Amersham (Braunschweig, Germany); reagent for human insulin at the concentrations indicated for SDS-PAGE was purchased from Pharmacia (Freiburg, 15 min immediately before the assay. (3H)-labelled Germany) and Sigma (Munich, Germany), respec- 2-deoxy-D-glucose (1 mCi=dish, concentration 4±5 tively; and the polyclonal GLUT1 and GLUT4 anti- mmol=l) was added to the medium which already sera were from Calbiochem (Bad Soden, Germany). contained 5 mmol=l unlabelled glucose. Hexose uptake was measured for 20 min at 37C. Glucose uptake was terminated by transferring the dishes to an Cell culture ice-bath. Cells were repeatedly washed with ice-cold Adipose tissue samples (20 ±100 g wet tissue) PBS and incubated for 20 min with 0.1% SDS. The were obtained from mammary adipose tissue of radioactivity of the cell material was counted in a young, normal-weight females (age < 40 y, BMI liquid scintillation counter (Beckman, Munich, Ger- < 27 kg=m2) undergoing elective mammary reduction many). Values were corrected for non-speci®c uptake surgery. All women were otherwise healthy, in parti- as decribed recently.13 cular free of metabolic or endocrine diseases. The procedure for obtaining human adipose tissue has been approved by the Ethical Committee of the Heinrich-Heine-University DuÈsseldorf. Stromal cells Western blotting from human adipose tissue were prepared as described For the preparation of crude membrane fractions, cells previously.12 Brie¯y, after removing all ®brous mate- were washed in TES buffer (225 mmol=l saccharose, rial and visible blood vessels, adipose tissue samples 1 mmol=l EDTA, 20 mmol=l Tris, 2.5 mg=ml leupep- were cut into small pieces (approximately 10 mg) and tin, 2.5 mg=ml pepstatin, 2.5 mg=ml aprotinin and digested in 10 mmol=l phosphate buffered saline 0.2 mmol=l phenylmethylsulfonyl¯uoride, pH 7.4). (PBS) containing 1 mg=ml crude collagenase and Cells were scraped from the dishes using a rubber 20 mg=ml bovine serum albumin, pH 7.4, for approxi- policeman, immediately frozen in liquid nitrogen and mately 45 min in a shaking water bath. After short stored at 770C. The cell material was homogenized centrifugation at 200 g, the ¯oating fat cells and the in 10 ml TES buffer using a te¯on glass homogenizer incubation solution were aspirated and discarded. The at 4C. The ®rst centrifugation was performed for sedimented cells were resuspended in an erythrocyte 10 min at 1000 g, the supernatant was centrifuged lyzing buffer consisting of 154 mmol=lNH4Cl, again for 90 min at 100 000 g. The membrane pellet 5.7 mmol=lK2HPO4 and 0.1 mmol=l EDTA for was resuspended in TES buffer. After a second 10 min to remove contaminating red blood cells. The homogenization step, the membrane fraction was dispersed material was ®ltered through a nylon mesh stored at 770C. with a pore size of 150 mm. After two washing steps, Membrane protein samples of 30 ±50 mg, but within the sedimented cells were resuspended in Dulbecco's one experiment, identical amounts were subjected to Modi®ed Eagle's=Ham's F-12 medium (v=v, 1:1) SDS-PAGE and transferred to nitrocellulose ®lters in supplemented with 10% fetal calf serum and inocu- a semi-dry blotting apparatus. Filters were blocked for lated into culture dishes at a density of approximately 10 min in Tris-buffered saline containing 0.05% 3±56104=cm2. After a 16 h incubation period for cell Tween-20 and 5% BSA. Then, ®lters were incubated attachment, cells were washed with PBS and refed for 16 h at 4C with a 1:500 dilution of a polyclonal with a serum-free DME=Ham's F-12 medium supple- GLUT1 or GLUT4 antiserum. After extensive wash- mented with 15 mmol=l NaHCO3, 15 mmol=1 Hepes, ing with Tris-buffered saline containing 0.05 Tween- 33 mmol=l biotin, 17 mmol=l pantothenate, 10 mg=ml 20 and 5% BSA, ®lters were incubated for 21 h with human transferrin, 100 U=ml penicillin and 0.1 g=l 125I-protein A (0.3 mCi=ml) at room temperature. streptomycin. To induce adipose differentiation, Filters were again extensively washed, air-dried and cells were exposed to 66 nmol=l insulin, 100 nmol=l exposed to Hyper®lm-MP ®lms using intensifying cortisol, 0.2 nmol=lT3 and for the ®rst three days screens. Autoradiograms were quanti®ed by laser 0.25 mmol=l MIX. The medium was changed every scanning densitometry (LKB, GraÈfel®ng, Germany). Insulin responses in adipocyte precursor cells H Hauner et al 450 Other biochemical assays droplets, GLUT4 protein level increased approxi- Glycerol-3-phosphate dehydrogenase (GPDH) was mately 4.4-fold and on day 16, after terminal differ- determined as a marker enzyme of adipose differen- entiation, the amount of GLUT4 protein was 16.7-fold tiation and fat cell function, according to an estab- above the level in the undifferentiated state (Figure 1). lished method.14 A protein precipitation technique to Thereby, this rapid and dramatic increase of GLUT4 avoid lipid interference was used to measure the was apparently responsible for the emergence of protein content of the culture.15 Speci®c GPDH activ- insulin responsiveness. On the other hand, a large ities are presented. amount of GLUT1 protein was detectable in undifferentiated cells. Similar levels of GLUT1 were still found on day 8, whereas GLUT1 protein was Statistics decreased to almost undetectable levels on day 16 Results are expressed as mean Æ s.d. of at least three (Figure 2). experiments in triplicate. For comparison, analysis of variance was used. P-values of <0.05 were considered as statistically signi®cant. Glucose transport in long-term cultured human adipocytes To examine the suitability of this human cell culture model for long-term studies of glucose transport, we Results then determined glucose transport capacity for two consecutive weeks in newly developed fat cells. Day Glucose transport during adipose differentiation 16 was de®ned as the starting point when cells have Stromal cells isolated from human adipose tissue expressed full adipocyte function. Measurement of the samples and kept in culture for 72 h did not respond activity of GPDH, which is an established marker to insulin stimulation of 2-deoxy-glucose uptake enzyme for adipose differentiation, revealed that (Table 1). However, when cells have undergone enzyme activity has reached a plateau and remained adipose differentiation in a serum-free medium unchanged for the subsequent 14 d (Table 2). containing insulin, triiodothyronine and cortisol as Although fat cell number was stable and no delayed adipogenic agents, a clear increase in the rate of differentiation was occurring (data not shown), there glucose uptake after a 15 min preincubation with was still a signi®cant progression in fat cell morphol- 1079 or 1077 mol=l insulin was observed. Compared ogy as demonstrated in Figure 3. Most cells developed to basal uptake rates, insulin induced a 3± 4-fold a more spherical shape and the lipid vacuoles became increase in 2-deoxy-glucose uptake. The stimulation larger probably due to the con¯uence of small dro- of hexose uptake was comparable at the two insulin plets. Finally, many cells have obtained the character- concentrations, indicating that the maximum insulin istic signet-ring morphology. effect was already obtained at 1079 mol=l (Table 1). The results of 2-deoxy-glucose uptake between day In contrast, basal glucose uptake markedly decreased 16 and day 30 are also presented in Table 1. Com- during the differentiation process, by almost 90% pared to day 16 there was a moderate decrease in the from 1993 Æ 465 on day 3 to 200 Æ 65 fmol=mg pro- basal uptake rate that was not statistically signi®cant. tein 620 min on day 16 when cells have acquired Likewise, insulin-stimulated glucose transport also adipocyte morphology (Table 1). slightly decreased. However, when expressed as stimulation above basal rates, insulin responsiveness remained unchanged during the 14 additional days Development of GLUT4 and GLUT1 protein in culture. There was also no signi®cant variation in As insulin stimulation of glucose transport in adipose the stimulation of glucose transport by either 1079 or tissue depends on the presence of GLUT4 protein, it 1077 mol=l insulin (Table 1). Measurement of the was important to measure the cellular amount of total cellular amount of GLUT4 protein by Western GLUT4 during adipose differentiation. In crude mem- blotting did not reveal any signi®cant change over brane fractions of undifferentiated cells, the amount of time between day 16 and day 30, although there was a detectable GLUT4 protein was very low. On day 8, trend to a decrease in transporter number (Table 2). when cells have already started to accumulate lipid Even at near-physiological insulin concentrations

Table 1 Basal and insulin-stimulated 2-deoxy-D-glucose uptake in cultured human adipocytes before and after differentiation. Glucose transport was determined as described under Methods. Mean Æ s.d. of three experiments in triplicate

(3H)-2-deoxy-glucose uptake (fmol/mg protein/20 min)

Day basal 1079 M insulin fold increase above basal 1077 M insulin fold increase above basal

3 1993 Æ 465 1640 Æ 375 ± 1930 Æ 232 ± 16 200 Æ 65 633 Æ 94 3.2 715 Æ 62 3.6 23 120 Æ 49 513 Æ 87 4.3 375 Æ 78 3.1 30 150 Æ 62 477 Æ 90 3.2 320 Æ 65 2.1 Insulin responses in adipocyte precursor cells H Hauner et al 451 (2610711 mol=l), starting from day 16, the relative expression of GLUT4 remained unchanged during the following 14 days (data not shown).

Discussion

The results of this study show many similarities in the pattern of insulin-stimulated glucose uptake between human adipocyte precursor cells in primary culture Figure 1 Development of GLUT4 protein in differentiating and clonal preadipocytes from rodent origin, both human adipocyte precursor cells. GLUT4 protein levels were before and after differentiation. In the undifferentiated determined by Western blotting as described under Methods. state, basal glucose uptake was high and decreased Panel A shoes mean Æ s.d. of three experiments in duplicate, panel B demonstrates one representative autoradiogramm. during the course of differentiation. Whereas human preadipocytes did not respond to insulin stimulation, fully differentiated cells demonstrated a rapid increase in glucose transport upon insulin stimulation. This was apparently due to a marked increase in GLUT4 protein, which represents the main insulin-regulated glucose transporter isoform in adipose tissue.3 5 In our experiments, the amount of GLUT4 protein increased from very low levels by more than 16- fold, which is in the same magnitude as ®ndings Figure 2 Course of GLUT1 protein in differentiating human 6±8 adipocyte precursor cells. GLUT1 was determined by Western reported by other groups in rodent cell lines. blotting as described under Methods. One representative auto- It has been demonstrated in 3T3 L1 cells that basal radiogram out of four experiments is demonstrated. rates of 2-deoxy-glucose uptake are higher in grow- ing, than in con¯uent, cells and newly developed fat

Figure 3 Photomicrographs of in vitro differentiated human fat cells from one representative cultre. Cells on A) day 16 and B) day 30. Insulin responses in adipocyte precursor cells H Hauner et al 452 Table 2 GLUT4 protein levels and GPDH activity in newly developed human adipocytes during a 14 d culture period. Day 16 was de®ned as the time point when cells have obtained full adipocyte function. The amount of GLUT4 protein on day 16 was de®ned as 100% and the levels of GLUT4 at the other time points expressed in relation to day 16. Results are mean Æ s.d. of three experiments

Day GLUT4 (%) GPDH (mU/mg protein)

16 1077 M insulin 100 385 Æ 51 23 1077 M insulin 85.5 Æ 29.8 364 Æ 82 30 1077 M insulin 80.4 Æ 16.8 484 Æ 132

cells exhibited lower transport rates relative to undif- GLUT4 protein and mRNA levels.13 Drawbacks of ferentiated cells.8 The decline in basal glucose uptake this culture system of human fat cells are that the during the course of differentiation was found to occur culture of cells is expensive and time-consuming, and in parallel with the decrease of GLUT1 level,6 that the number of available cells is limited. although in one of these studies the decline was The lack of insulin responsiveness of undifferen- only moderate.7 In our study, the marked decrease tiated cells found in the present study could also be in basal glucose uptake was due to a dramatic reduc- due to the insulin receptor status, as studies in mouse tion in the number of GLUT1 transporters to almost preadipocyte cell lines indicated that the number of undetectable levels. In contrast to the 3T3 L1 cell line, insulin receptors is low in the ®broblastic stage and human adipocytes exhibit a clearly more pronounced increases dramatically up to 35-fold during differen- loss of GLUT1 during the course of differentiation. tiation. These studies also described a clear correla- An interesting ®nding was the observation that glu- tion between the development of increased insulin cose transport and GLUT4 protein levels remained binding capacity and the increased sensitivity of stable once cells have acquired the adipocyte morphol- glucose transport to insulin.18±20 On the other hand, ogy. Similarly, cultured human adipocytes exhibited human adipocyte precursor cells express high levels stable high activities of the lipogenic marker enzyme of IGF-I receptors,21 as do rodent adipocyte precur- GPDH and a high production and release of leptin.16 sors22 and 3T3 L1 cells23 and, therefore, supraphy- This is in contrast to a recent study in isolated rat siologically high concentrations of insulin should also adipocytes in suspension culture, where a rapid decrease activate glucose transport via IGF-I receptors due to a of GLUT4 mRNA was reported that was not prevented high sequence and functional homology with the by treatment of the cells with insulin.17 In our studies, insulin receptor.23 However, at lower near-physiolo- newly developed fat cells retained their ability to gical insulin concentrations, a similar responsiveness respond to insulin and to express GLUT4 expression was observed with regard to glucose uptake and for at least 14 d. However, with longer culture periods, GLUT4 expression, arguing against a possible inter- cells became more and more spherical and got easily ference by IGF-I. detached from the culture dishes. Therefore, studies on In conclusion, the results of this study indicate that day 30 were to some degree critical and, in single human adipocyte precursor cells are not able to experiments, signi®cant cell loss was already observed increase glucose uptake upon insulin stimulation, before day 30 in culture (data not shown). probably due to a lack, or only very low expression, This culture system may provide a number of of GLUT4, whereas basal glucose transport is advantages. Generally, a primary culture model is approximately 10-fold higher than in differentiated probably more physiological than clonal cell lines, cells. After acquisition of the adipocyte phenotype, which may have undergone spontaneous transforma- cells exhibit high GLUT4 levels which may confer the tion and may differ in their proliferation and differ- marked insulin responsiveness at this stage, whereas entiation pattern in many aspects.9 As cells were of GLUT1 levels are substantially decreased after differ- human origin, they may better resemble the conditions entiation. The glucose transport system of newly in human diseases. However, the main advantage is differentiated human fat cells remains stable for at that this primary culture model is also suitable for least 14 days. Therefore, this culture technique may long-term studies of the regulation of glucose meta- provide an excellent model to study the long-term bolism. In our experience, freshly isolated fat cells regulation of GLUT4 expression in human adipose remain viable for only a few hours in suspension tissue. culture and can be used only for short-term studies. During this short life span, freshly prepared cells are Acknowledgements invariably stressed or damaged by the collagenase We wish to thank Prof. R. Olbrisch and his team from digestion procedure. In contrast, using the culture the Department of Plastic Surgery at the Florence system described here, we recently studied the effect Nightingale Hospital DuÈsseldorf-Kaiserswerth for of a 3 d incubation, with tumor necrosis factor alpha their kind support. This study was supported in part on glucose and lipid metabolism in human fat cells by a grant from the Deutsche Forschungsgemeinschaft and found a highly reproducible suppression of (SFB 351, Teilprojekt C2). Insulin responses in adipocyte precursor cells H Hauner et al 453 References 13 Hauner H, Petruschke T, Russ M, RoÈhrig K, Eckel J. Effects of 1 Cushman SW, Wardzala LJ. Potential mechanism of insulin tumour necrosis factor alpha (TNFa) on glucose transport and action on glucose transport in the isolated rat adipose cell. lipid metabolism of newly-differentiated human fat cells in Apparent translocation of intracellular transport systems to the culture. Diabetologia 1995; 38: 764±771. plasma membrane. J Biol Chem 1980; 255: 4758±4762. 14 Pairault J, Green H. A study of the adipose conversion of 2 Suzuki K, Kono T. Evidence that insulin causes translocation suspended 3T3 cells by using glycerolphosphate dehydrogen- of glucose transport activity to the plasma membrane from an ase as differentiation marker. Proc Natl Acad Sci USA 1979; intracellular storage site. Proc Natl Acad Sci USA 1980; 77: 76: 5138±5142. 2542±2545. 15 Peterson GL. A simpli®cation of the protein assay of Lowry et 3 Gould GW, Holman GD. The glucose transporter family: al. Which is more generally applicable? Anal Biochem 1977; structure, function and tissue-speci®c expression. Biochem J 83: 346±356. 1993; 295: 329±341. 16 Wabitsch M, Jensen PB, Blum WF, Christoffersen CT, Eng- 4 Kahn BB. Glucose transport: pivotal step in insulin action. laro P, Heinze E, Rascher E, Teller W, Tornqvist H, Hauner H. Diabetes 1996; 45: 1644±1654. Insulin and cortisol promote leptin production in cultured 5 James DE, Strube M, Mueckler MM. Molecular cloning and human fat cells. Diabetes 1996; 45: 1435±1438. characterization of an insulin-regulatable glucose transporter. 17 Gerrits PM, Olson AL, Pessin JE. Regulation of the Nature 1989; 338: 83±87. GLUT4=muscle-fat glucose transporter mRNA in adipose 6 Tordjman KM, Leingang KA, James DE, Mueckler MM. tissue of insulin-de®cient diabetic rats. J Biol Chem 1993; Differential regulation of two distinct glucose transporter 268: 640±644. species expressed in 3T3-L1 adipocytes: Effect of chronic 18 Rubin CS, Hirsch A, Fung C, Rosen OM. Development of insulin and tolbutamide treatment. Proc Natl Acad Sci USA hormone receptors and hormonal responsiveness in vitro. 1990; 86: 7761±7765. Insulin receptors and insulin sensitivity in the preadipocyte 7 De Herreros AG, Birnbaum MJ. The acquisition of increased and adipocyte forms of 3T3-L1 cells. J Biol Chem 1978; 253: insulin responsive hexose transport in 3T3-L1 adipocytes 7570±7578. correlates with expression of a novel transporter gene. J Biol 19 Karlsson FA, Grunfeld C, Kahn CR, Roth J. Regulation of Chem 1989; 264: 19 994±19 999. insulin receptors and insulin responsiveness in 3T3-L1 fatty 8 Weiland M, SchuÈrmann A, Schmidt WE, Joost HG. Develop- ®broblasts. Endocrinology 1979; 104: 1383±1391. ment of the hormone-sensitive glucose transport activity in 20 Resh MD. Development of insulin responsiveness of the differentiating 3T3-L1 murine ®broblasts. Role of the two glucose transporter and the Na,K-adenosine triphosphatase glucose transporter species and their subcellular localization. during in vitro adipocyte differentiation. J Biol Chem 1982; Biochem J 1990; 270: 331±336. 257: 6978±6986. 9 Ailhaud G, Hauner H. Development of white adipose tissue. 21 Wabitsch M, Christoffersen CT, Blum WF, Heinze E, Tornq- In: Handbook of Obesity. Bray GA, Bouchard C, James WPT vist H, Teller W. Role of insulin-like growth factor-I in human (eds). 1997; Marcel Dekker: New York, in press. adipose tissue growth and metabolism. Eur J Pediatr 1997, in 10 Hauner H. Complete adipose differentiation of 3T3 L1 cells press. in a chemically de®ned medium: comparison to serum- 22 Nougues J, Reyne Y, Barenton B, Chery T, Darandel V, containing culture conditions. Endocrinology 1990; 127: Soriano J. Differentiation of adipocyte precursors in a 865±872. serum-free medium is in¯uenced by glucocorticoids and 11 Entenmann G, Hauner H. Relationship between replication endogenously produced insulin-like growth factor-I. Int J and differentiation in cultured human adipocyte precursor Obes 1993; 17: 159±167. cells. Am J Physiol 1996; 270: C1011±C1016. 23 Weiland M, Bahr F, Hohne M, Schurmann A, Ziehm D, Joost 12 Hauner H, Entenmann G, Wabitsch M, Gaillard D, Ailhaud G, HG. The signaling potential of the receptors for insulin and Negrel R, Pfeiffer EF. Promoting effects of glucocorticoids on insulin-like growth factor I (IGF-I) in 3T3-L1 adipocytes: the differentiation of human adipocyte precursor cells cultured comparison of glucose transport activity, induction of onco- in a chemically de®ned medium. J Clin Invest 1989; 84: gene c-fos, glucose transporter mRNA, and DNA synthesis. 1663±1670. J Cell Physiol 1991; 149: 428±435.