Modulation of Resistin Expression by Retinoic Acid and ... - Diabetes

Modulation of Resistin Expression by Retinoic Acid and Vitamin A Status Francisco Felipe, M. Luisa Bonet, Joan Ribot, and Andreu Palou

This work identifies retinoic acid (RA), the acid form of vitamin A, as a signal that inhibits the expression of resistin, an adipocyte-secreted protein previously proposed to act as an inhibitor of adipocyte differentiation and as a systemic insulin resistance factor. Both 9-cis and all-trans RA reduced resistin mRNA levels in white and brown adipocyte cell model systems; the effect was time- and dose-dependent, was followed by a reduced secretion of resistin, and was reproduced by selective agonists of both RA receptors and rexinoid receptors. Association of CCAAT/enhancer-binding protein ␣ (a positive regulator of the resistin gene) and its coactivators p300, cAMP response element– binding protein binding protein, and retinoblastoma protein with the resistin gene promoter was reduced in RA-treated adipocytes. RA administration to normal mice resulted in reduced resistin mRNA levels in brown and white adipose tissues, reduced circulating resistin levels, reduced body weight, and improved glucose tolerance. Resistin expression was also downregulated after dietary vitamin A supplementation in mice. The results raise the possibility that vitamin A status may contribute to modulate systemic functions through effects on the production of adipocyte-derived protein signals. Diabetes 53:882– 889, 2004

A

dipose tissue is an important secretory organ, and genes encoding adipocyte secreted proteins (adipokines) potentially involved in the regulation of energy homeostasis, adipose tissue development, and insulin sensitivity may be the targets for nutrient interactions relevant to the likelihood of development of obesity and diabetes. One adipokine that has been claimed to be involved in the modulation of

From the Laboratory of Molecular Biology, Nutrition and Biotechnology, Department of Fundamental Biology and Health Sciences, University of the Balearic Islands, Palma de Mallorca, Spain. Address correspondence and reprint requests to Dr. M. Luisa Bonet, Laboratori de Biologia Molecular, Nutricio´ i Biotecnologia, Departament de Biologia Fonamental i Cie`ncies de la Salut, Universitat de les Illes Balears, Edifici Guillem Colom, Crta. Valldemossa Km 7.5, 07122 Palma de Mallorca, Spain. E-mail: [email protected]. Received for publication 29 July 2003 and accepted in revised form 23 December 2003. ATRA, all-trans retinoic acid; BAT, brown adipose tissue; CBP, cAMP response element– binding protein binding protein; C/EBP, CCAAT/enhancerbinding protein, CREB, cAMP response element– binding protein; DMEM, Dulbecco’s modified Eagle’s medium; eWAT, epididymal white adipose tissue; iBAT, interscapular BAT; iWAT, inguinal white adipose tissue; PPAR-␥, peroxisome proliferator–activated receptor-␥; RA, retinoic acid; RAR, RA receptor; RB, retinoblastoma protein; RXR, rexinoid receptor; TTNPB, p-[(E)2-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)-1-propenyl]benzoic acid; TZD, thiozolidinedione; UCP, uncoupling protein; WAT, white adipose tissue. © 2004 by the American Diabetes Association. 882

systemic insulin sensitivity is resistin, a 12.5-kDa cysteinerich secretory protein that in the mouse is adipocytespecific. Resistin was identified (1) in cultured 3T3-L1 adipocytes as a gene product downregulated by thiozolidinediones (TZDs), a family of antidiabetic drugs that are high-affinity ligands of peroxisome proliferator–activated receptor-␥ (PPAR-␥). A causative role of elevated resistin in obesity-associated insulin resistance remains controversial (reviewed in 2– 4), but recent reports have demonstrated metabolic actions of resistin impairing insulin action in the liver (5) and implicated resistin in the pathophysiology of the human insulin resistance syndrome (6). Resistin is induced during differentiation of 3T3-L1 preadipocytes (1,7) and inhibits adipose conversion of these cells (7), suggesting that it may also serve as a feedback signal to restrict adipose tissue expansion. Vitamin A is a nutrient with important effects on adipose tissue development and metabolism (reviewed in 8). Retinoic acid (RA), its carboxylic acid form, promotes (9) or inhibits (10) adipogenesis of preadipose cells in culture depending on the dose, negatively affects preadipocyte survival (11), and induces the expression of uncoupling proteins (UCPs) in cultured brown adipocytes (12–14). Moreover, both RA treatment and vitamin A status influence body adiposity and the expression of adipogenic/ lipogenic transcription factors and UCPs in white (WAT) and brown (BAT) adipose tissues and of UCP3 in muscle of rodents (13,15–17). Most of the cellular effects of RA rely on changes in gene transcription and are mediated by two types of retinoid receptors that, like the PPARs, belong to the nuclear hormone receptor superfamily: the RA receptors (RARs), which respond to both all-trans RA and 9-cis RA, and the rexinoid receptors (RXRs), which respond specifically to 9-cis RA (reviewed in 18). Liganded retinoid receptors regulate transcription directly, through binding as RAR:RXR or RXR:RXR to specific response elements in promoter/enhancers, and indirectly, by interfering with the activity of other transcription factors. Extragenomic actions of liganded retinoid receptors (19) and effects of retinoids nonmediated by retinoid receptors (20) of potential impact on gene expression have also been described. A number of considerations led us to hypothesize that RA could affect resistin expression. First, PPAR-␥, which is likely to mediate the inhibitory effect of TZDs on the resistin gene, regulates transcription as heterodimer with RXR, and the PPAR:RXR, unlike other RXR-containing heterodimers, can be activated by ligands of both partners (21). Second, synthetic RXR agonists behave as insulin sensitizers in rodent models of obesity and diabetes (21,22). Third, resistin gene expression is induced by CCAAT/ enhancer-binding protein-␣ (C/EBP-␣) (23–25), and RA is DIABETES, VOL. 53, APRIL 2004

F. FELIPE AND ASSOCIATES

known to inhibit the transcriptional activity of C/EBP transcription factors (26). The aim of this work was to test this hypothesis, studying retinoid regulation of resistin expression in adipocyte cell model systems and in intact mice, and gaining first insights into the molecular mechanism(s) involved and the possible functional relevance of this regulation on glucose tolerance. RESEARCH DESIGN AND METHODS (⫾) Arterenol bitartrate salt (norepinephrine), all-trans RA (ATRA); isopropyl-(E,E)-(R,S)-11-methoxy-3,7,11-trimethyldodeca-2,4-dienoate (methoprene); and p-[(E)-2-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)-1propenyl]benzoic acid (TTNPB) were from Sigma (St. Louis, MO). 9-cis RA, Ro 40-6055, and Ro 25-7386 were gifts from Hoffmann-La Roche (Basel, Switzerland). Newborn calf serum and FBS were from Linus (Madrid, Spain), and other cell culture reagents were from Sigma. Cell culture. 3T3-L1 murine fibroblasts (American Type Culture Collection, LGC Deselaers SL, Barcelona, Spain) were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% newborn calf serum under 8% CO2/92% air at 37°C. Differentiation was induced 2 days after confluence by incubating the cells for 48 h in DMEM containing 5 ␮mol/l insulin, 0.5 mmol/l 3-isobutyl-1-methyl-xanthine, 0.25 ␮mol/l dexamethasone, and 10% FBS, followed by 48 h in DMEM containing 10% FBS and 5 ␮mol/l insulin (no insulin thereafter). Adipogenesis was monitored by morphologic examination of the cells for accumulation of lipid droplets. Cells were exposed to ATRA and other compounds of interest on days 6 –7 after induction of differentiation, when ⬎80% expressed the adipocyte phenotype, and RNA was routinely extracted 24 h later. In some experiments, the culture medium was collected to determine resistin protein levels. Cytotoxicity of vehicle (DMSO) and all agonists (including ATRA) of retinoid receptors at the working concentrations used was measured by the lactate dehydrogenase method (27), using a commercial kit (Roche Diagnostics, Mannheim, Germany). BAT precursor cells from 4-week-old NMRI mice were isolated and cultured as described previously (28). Cells were exposed to RA or norepinephrine on day 7, when they had reached confluence and started to differentiate, and RNA was extracted 24 h later. Chromatin immunoprecipitation assays. Day 6 after hormonal onset of adipogenesis, 3T3-L1 adipocytes were treated with 0.1 ␮mol/l ATRA, 10 ␮mol/l ATRA, or vehicle (DMSO) for 24 h. Cells were then washed with PBS, submitted to cross-linking with 1% formaldehyde in PBS at 37°C for 10 min, rinsed and scraped in ice-cold PBS, centrifuged for 4 min at 700g, and resuspended in lysis buffer (1% SDS, 5 mmol/l EDTA, 50 mmol/l Tris-HCl [pH 8.1]). After a 20-min incubation on ice, samples were sonicated at 15-s pulses three times on ice. The lysates were centrifuged at 14,000g for 10 min; a small aliquot of the supernatant served as the input control, and the rest was diluted in buffer (1% Triton X-100, 2 mmol/l EDTA, 150 mmol/l NaCl, 20 mmol/l Tris-HCl [pH 8.1]) containing protease inhibitors. Samples were precleared with 2 ␮g of sheared salmon sperm DNA and 45 ␮l of protein A-Sepharose beads for 2 h. Immunoprecipitation with antibodies against the following proteins was performed overnight: C/EBP-␣, p300, cAMP response element– binding protein (CREB) binding protein (CBP), retinoblastoma protein (RB), and normal rabbit IgG (all from Santa Cruz Biotechnology, Santa Cruz, CA). Samples were then incubated with 45 ␮l of protein A-Sepharose beads for 1 h. Precipitates were washed as described previously (23) and extracted by incubating with elution buffer (1% SDS, 0.1 mol/l NaHCO3) at 65°C for 6 h. DNA fragments were purified using a commercial kit (Qiagen, Hilden, Germany). Two to 5 ␮l of purified sample was used in 31 cycles of PCR. Primers surrounding the C/EBP-␣ site in the resistin promoter had sequences 5⬘-GTCTTGGCTCCTAGCCTTGC-3⬘ and 5⬘-GTTGACTTCTGGCCCATCC-3⬘ (23). PCR products were run on agarose gels, stained with ethidium bromide, recorded using an image recording system (Gelprinter; TDI, Madrid, Spain), and quantified using the Kodak 1D Image Analysis Software 3.5 for Windows (Eastman Kodak, Rochester, NY). Animal studies. Animals were obtained from CRIFFA (Barcelona, Spain) and were treated in accordance with our institutional guidelines. RA treatment experiment. Twelve-week-old NMRI male mice that were fed regular diet (Panlab, Barcelona, Spain) received one daily subcutaneous injection of ATRA at a dose of 10, 50, or 100 mg/kg body wt during the 4 days before they were killed (seven animals/group). Controls received the vehicle (olive oil). Vitamin A supplementation experiments. Four-week-old C57BL/6J (B6) male mice were fed either a normal-fat or a high-fat diet (10 and 45% of total energy as fat, mainly lard, respectively) with either the standard vitamin A DIABETES, VOL. 53, APRIL 2004

content or a 40-fold excess (8 mg and 320 mg retinyl palmitate/kg diet, respectively) for 18 weeks, as previously described (17) (four animals/group). In an independent experiment, 4-week-old NMRI mice were fed the normal-fat diet supplemented or not with vitamin A for 11 weeks (six animals/group). Diets were from Research Diets (New Brunswick, NJ). The vitamin A–supplemented animals did not show any external signs of vitamin A toxicity or significant liver enlargement. The animals were kept at 22°C under 12-h light/dark cycles (lights on at 0800) and were killed with CO2 followed by decapitation at the start of the light cycle. Blood was collected from the neck, and serum was prepared and frozen at ⫺20°C. Interscapular BAT (iBAT), inguinal WAT (iWAT), and epididymal WAT (eWAT) were excised in their entirety, weighed, frozen in liquid nitrogen, and stored at ⫺70°C. Glucose tolerance tests. Tests were conducted in four 12-week-old NMRI male mice and in the same animals 1 week later, after they had received four subcutaneous injections of ATRA (at a dose of 100 mg/kg body wt, in 100 ␮l of olive oil; one daily injection the 4 days preceding the test). Glucose (3.375 g/kg animal) was administered intraperitoneally after a 6-h fast (from 2400 to 0600), and glucose levels were measured in tail blood collected at the indicated times. RNA extraction and Northern blotting. Total RNA was extracted with TriPure reagent (Roche Diagnostics). Ten to 20 ␮g was fractionated by agarose gel electrophoresis, blotted onto a nylon membrane by capillary blotting, fixed with ultraviolet light, and hybridized as described previously (17). An antisense oligonucleotide (5⬘-TCCCACGAGCCACAGGCAGAGCCAC AGGAGCAGC-3⬘) based on the mouse resistin cDNA sequence labeled at both ends with a single digoxigenin ligand was used as a hybridization probe (29). Hybridization signals were detected by chemiluminescence using CDP-Star (Roche Diagnostics) as the substrate, with membranes exposed to Hyperfilm ECL (Amersham Biosciences UK, Buckinghamshire, U.K.) for up to 60 min. Bands in films were analyzed by scanner photodensitometry and quantified as described above. Blots were stripped and reprobed for other mRNAs of interest and 18S rRNA to check equal loading of gels and transfer efficiency. Antisense oligonucleotides used to detect leptin mRNA, UCP1 mRNA, and 18S rRNA were as described previously (17). Western blotting. A total of 50 ␮l of culture medium or 10 ␮l of serum was solubilized and boiled in Laemmli sample buffer containing 20% 2-mercaptoethanol, fractionated in 15% SDS-PAGE gels, and electrophoretically transferred onto nitrocellulose membranes. Blots were stained with Ponceau S solution to visualize the amount of total protein in each lane. After blocking in PBS-Tween containing 5% fat-free dried milk, membranes were incubated with a polyclonal antibody against mouse resistin (1:1,000; Linco Research, St. Charles, MO) for 1 h, washed, and incubated with a horseradish peroxidase– conjugated secondary antibody (1:10000; Research Diagnostics, Flanders, NJ). Blots were developed using an enhanced chemiluminescence Western blotting analysis system (Amersham Biosciences UK). Bands in films were analyzed and quantified as described above. Other parameters determined. Circulating levels of triacylglycerols, glucose, and insulin were determined using commercial kits (Triglyceride [INT]20 from Sigma Diagnostics; D-glucose ultraviolet method from Roche Biopharm, Darmstadt, Germany; and rat-insulin enzyme-linked immunosorbent assay from DRG Instruments, Marburg, Germany, respectively). Statistical analysis. Data are expressed as mean ⫾ SE. Statistical significance was assessed by Student’s t test or by one- or two-way ANOVA followed by least-significant difference and Student’s t test post hoc comparisons, respectively. The minimum significance level was set at P ⬍ 0.05.

RESULTS

RA inhibited resistin expression in brown and white adipocytes differentiated in culture. Resistin mRNA levels were dose-dependently downregulated by ATRA and 9-cis RA both in primary brown adipocytes differentiated in culture and in 3T3-L1 white adipocytes (Fig. 1A and B). In both cell systems, ATRA was more effective than 9-cis RA. As little as 0.01 ␮mol/l ATRA reduced resistin mRNA levels by 80% in primary brown adipocytes and by 60% in 3T3-L1 white adipocytes. In the brown adipocytes, this dose was insufficient to trigger the induction of the UCP1 gene, which was triggered, as expected (12,13), by higher doses of the retinoids (Fig. 1A). The inhibitory effect of RA isomers on resistin expression in brown adipocytes was only partially reproduced by 1 883

RESISTIN AND RETINOIC ACID

FIG. 1. Inhibition of resistin expression in brown and white adipocyte cell model systems by RA. A: Dose-dependence curve of the effect of 9-cis RA and ATRA on resistin and UCP1 mRNA levels in mouse primary brown adipocytes differentiated in culture. Confluent cells were treated with vehicle (C) or the retinoids for 24 h before harvesting and RNA extraction. Representative Northern blots of the effect of ATRA and 1 ␮mol/l norepinephrine (NA) are shown at the bottom. B: Dose-dependence curve of the effect of 9-cis RA and ATRA on resistin mRNA levels in 3T3-L1 adipocytes. Eighty percent differentiated cultures were treated with vehicle (C) or the retinoids for 24 h before harvesting and RNA extraction. Representative Northern blots of the effect of ATRA are shown at the bottom. C: Time course of the downregulatory effect of RA on resistin mRNA levels in 3T3-L1 adipocytes. Differentiated cells were harvested at the indicated time intervals after addition of 1 ␮mol/l ATRA. Representative Northern blots are shown at the bottom. D: Inhibition of resistin secretion by RA in 3T3-L1 adipocytes. At time 0, the culture medium in contact with the differentiated cells was changed, and 1 ␮mol/l ATRA or vehicle (control) was added. The culture medium was collected at the indicated time intervals after retinoid/vehicle addition and used to determine resistin protein levels. Representative Western blots are shown at the bottom. Data are means ⴞ SE of duplicates of at least two independent experiments. Student’s t test significances, P < 0.05: *ATRA vs. control; #9-cis RA vs. control; †time point vs. time 0.

␮mol/l norepinephrine, which, similar to retinoids, induced the expression of the UCP1 gene, as expected (Northern blot in Fig. 1A). The inhibitory effect of retinoids on resistin mRNA levels exhibited a relatively lengthy time course, with a half-maximal reduction observed after ⬃10 h of exposure of 3T3-L1 adipocytes to 1 ␮mol/l ATRA (Fig. 1C), and was followed by reduced resistin accumulation in the culture medium (Fig. 1D). Expression of resistin transcripts in differentiated primary brown adipocytes agrees with the recently reported differentiation-depen884

dent expression of resistin in the brown adipocyte cell line T37i (30). Both RAR- and RXR-dependent pathways seem to mediate the inhibition of resistin expression by retinoids. We assayed whether nonisomerizable synthetic compounds that selectively bind and activate RARs or RXRs could reproduce the inhibitory effect of RA on resistin gene expression in 3T3-L1 adipocytes (Table 1). Both TTNPB, which activates all RAR subtypes, and Ro 40-6055, a selective RAR-␣ agonist, reduced resistin mRNA DIABETES, VOL. 53, APRIL 2004


TABLE 1 Effect of RAR and RXR agonists on resistin mRNA levels in 3T3-L1 adipocytes Treatment Vehicle (DMSO) ATRA 1 ␮mol/l Ro 40-6055 1 ␮mol/l 10 ␮mol/l 100 ␮mol/l TTNPB 1 ␮mol/l 10 ␮mol/l 100 ␮mol/l Ro 25-7386 1 ␮mol/l 10 ␮mol/l 100 ␮mol/l Methoprene 1 ␮mol/l 10 ␮mol/l 100 ␮mol/l

Resistin mRNA (%) 100 ⫾ 7.9 27.3 ⫾ 3.4* 52.5 ⫾ 13.9* 45.1 ⫾ 4.2* 14.4 ⫾ 2* 27.1 ⫾ 4.4* 10.2 ⫾ 2.4* 3.8 ⫾ 0.4* 88.8 ⫾ 13 54.9 ⫾ 10.5* 20.5 ⫾ 2.1* 69.4 ⫾ 14.4 54.9 ⫾ 4.5* 15.2 ⫾ 5.4*

Eighty percent differentiated cultures were treated as indicated 24 h before harvesting and RNA extraction. Data are means ⫾ SE of duplicate plates of two independent experiments. Student’s t test significances, P ⬍ 0.05: *agonist vs. vehicle. Ro 40-6055, RAR-␣ agonist; TTNPB, agonist of all RARs; Ro 25-7386, RXR-␣ agonist; methoprene, agonist of all RXRs.

abundance in a dose-dependent manner. The pan-RAR agonist was more effective than the selective RAR-␣ agonist, suggesting that different RAR subtypes mediate the inhibitory effect of RA on the resistin gene. Although with lower potency than ATRA and TTNPB, methoprene, a pan-RXR agonist, and Ro 25-7386, a selective RXR-␣ agonist, also reduced resistin mRNA abundance in a dose-dependent manner; the two compounds were similarly effective, reproducing the inhibitory effect of 1 ␮mol/l ATRA at 100 ␮mol/l, which for methoprene is the concentration known to activate RXR in mammalian cells (31). Only TTNPB and Ro 25-7386 at the highest concentration tested (100 ␮mol/l) had a significant cytotoxic effect over vehicle (DMSO), leading to a 2.6 –2.8 increment of lactate dehydrogenase activity in the medium (P ⬍ 0.001 for both compounds, Student’s t test). RA treatment reduced the association of C/EBP-␣ and C/EBP-␣ coactivators with the resistin gene promoter in 3T3-L1 adipocytes. We reasoned that the inhibitory effect of RA on resistin expression could result, at least in part, from impaired binding of C/EBP-␣ and/or its coactivators to the resistin gene promoter. To test this, we performed chromatin immunoprecipitation assays measuring C/EBP-␣, p300, CBP, and RB association with a proximal region of the resistin promoter containing the C/EBP-␣ binding site (23) in mature 3T3-L1 adipocytes treated with ATRA (0.1 and 10 ␮mol/l) or vehicle for 24 h. p300 coactivates C/EBP-␣ (32), and both p300 and the closely related coactivator CBP were reported to be recruited to the mouse resistin gene promoter in differentiated 3T3-L1 mature adipocytes through their interaction with DNA-bound C/EBP-␣ (23). Hypophosphorylated RB also coactivates C/EBP-␣, this activity being of importance for adipocyte terminal differentiation (33–35). At 10 DIABETES, VOL. 53, APRIL 2004

FIG. 2. Effect of RA on the association of C/EBP-␣ and C/EBP-␣ coactivators with the resistin gene promoter in vivo. Differentiated 3T3-L1 adipocytes were treated with vehicle or ATRA (0.1 and 10 ␮mol/l) for 24 h before formaldehyde cross-linking and harvesting. Cross-linked fragmented chromatin was incubated with a nonspecific IgG antibody (mock) and with antibodies against C/EBP-␣, p300, CBP, and RB. Relative abundance of the resistin gene in the immunoprecipitated DNA was analyzed by PCR using primers surrounding the C/EBP-␣ binding site in the resistin gene promoter amplifying a 219-bp fragment DNA. Fragmented chromatin not subjected to immunoprecipitation served as input control.

␮mol/l, ATRA treatment resulted in reduced association of C/EBP-␣, p300, CBP, and RB with the resistin gene promoter by 45 ⫾ 9%, 33 ⫾ 10%, 27 ⫾ 6%, and 46 ⫾ 12%, respectively (n ⫽ 4; Fig. 2). The reduced association of C/EBP-␣ and its coactivators to the resistin gene promoter was already evident or even more marked in the cells that were treated with 0.1 ␮mol/l ATRA (Fig. 2). RA administration reduced adipose tissue resistin mRNA levels and circulating resistin levels and improved glucose tolerance in mice. Resistin mRNA levels in eWAT, iWAT, and iBAT (Fig. 3) and serum resistin levels (Fig. 4) were severely reduced in ATRA-treated NMRI mice. In accordance with previous reports (15,36), RA administration also triggered a reduction of adipose tissue leptin expression that was evident in the three depots (Fig. 3) and a reduction of body weight and of the mass of all fat pads analyzed (Table 2). Circulating levels of insulin, glucose, and triacylglycerols in ad libitum feeding conditions trended lower in the ATRA-treated NMRI mice (Table 2), suggesting an improvement of insulin sensitivity. To address this further, we performed glucose tolerance tests in a separate group of NMRI mice before and after ATRA administration. The treatment triggered a 13.3% reduction of body weight. Circulating glucose levels at 0, 30, 90, and 180 min after the glucose load were, respectively, 4.0 ⫾ 0.5, 28.9 ⫾ 1.9, 16.6 ⫾ 2.8, and 9.4 ⫾ 0.6 mmol/l before the ATRA treatment and 4.6 ⫾ 0.2, 21.0 ⫾ 3.7, 9.6 ⫾ 1.9, and 4.4 ⫾ 0.8 mmol/l afterward. The difference was of statistical significance at t ⫽ 180 min (P ⫽ 0.002, Student’s t test). The area under the glucose curve was 39.5 ⫾ 3 cm2 before ATRA treatment and 25.6 ⫾ 3.6 cm2 after it, indicating a statistically significant (P ⫽ 0.025, Student’s t test) enhancement of glucose disposal upon ATRA treatment. Dietary vitamin A supplementation reduced adipose tissue resistin mRNA levels and circulating resistin 885


FIG. 3. Downregulation of resistin and leptin expression in WAT and BAT of NMRI mice by RA treatment. Twelve-week-old NMRI male mice were treated with ATRA (daily subcutaneous injection during the 4 days before they were killed) at a dose of 10, 50, or 100 mg/kg body wt. Control animals received the vehicle (olive oil). Data are means ⴞ SE of seven animals per group. Leptin and resistin mRNA levels in adipose tissues were determined by Northern blotting. Student’s t test significances, P < 0.05: *ATRA injected vs. vehicle injected.

levels in mice. We analyzed resistin expression in mice that were chronically fed diets with a 40-fold excess of vitamin A relative to the usual content (Table 3). A similar dietary manipulation was previously shown to result in a high vitamin A status in rats, as revealed by the increment of liver retinol concentration (37). Our assumption was that supplementation could result in increased availability of RA to tissues and could reproduce, at least in part, the effects of acute RA treatment. In fact, in both rats and mice, vitamin A supplementation was shown to reproduce the effects of acute RA doses downregulating leptin expression and upregulating UCP1 expression (17,37). Resistin mRNA levels in eWAT and circulating resistin levels were reduced in both B6 and NMRI mice that were chronically fed a normal-fat high–vitamin A diet (for 18 and 11 weeks, respectively) compared with controls that were fed the same diet with the normal vitamin A content. Vitamin A supplementation did not significantly affect body weight or adiposity of either B6 (17) or NMRI mice, although in the latter, the two parameters trended lower after supplementation (by 9.7 and 30%, respectively). In the obesity-prone B6 mice, high-fat diet feeding for 18 weeks caused 26% overweight, as reported previously (17), and a significant reduction of both eWAT resistin mRNA

FIG. 4. Reduction of circulating resistin levels in NMRI mice by RA treatment. Animal groups were as described in the legend to Fig. 3. Resistin levels were determined in 10 ␮l of serum by Western blotting. Data are means ⴞ SE of seven animals per group. Representative blots are shown inside the figure. Student’s t test significances, P < 0.05: *ATRA injected vs. vehicle injected. 886

levels and serum resistin levels. Supplementation of the high-fat diet with vitamin A did not affect the final obesity reached (17) and did not lead to further significant decreases of eWAT resistin mRNA levels or serum resistin levels. DISCUSSION

The present study shows that RA inhibits resistin expression in adipocyte cell model systems and in vivo, after its administration to intact mice, both at mRNA and protein levels. Both RAR and RXR agonist reproduced the inhibitory effect of RA on resistin expression in 3T3-L1 adipocytes, strongly suggesting that both RAR- and RXRdependent signaling pathways mediate the RA effect. Liganded RXR may be acting as a heterodimer with PPAR-␥, which, although results are conflicting (38 – 40), has been previously implicated in the inhibition of the resistin gene (1,23,24,41– 43). Indeed, the finding that both PPAR-␥ agonists and RXR agonists—including the endogenous ligand, as first demonstrated here—inhibit resistin gene transcription reinforces the idea of the involvement of PPAR-␥:RXR, because this entity is known to respond to ligands of both partners (21). In any case, our results are the first to suggest that not only liganded RXR but also liganded RARs can repress resistin gene transcription. Liganded retinoid receptors may act on the resistin gene through different, nonmutually exclusive mechanisms. The fact that the inhibitory effect of both ATRA (Fig. 1C) and the TZD PPAR-␥ ligand rosiglitazone (23) on resistin expression in 3T3-L1 adipocytes follows a relatively lengthy time course suggests that liganded PPAR-␥:RXR may be inducing a protein or proteins that repress resistin gene transcription. Our chromatin immunoprecipitation assay results (Fig. 2) indicate that impairment of C/EBP-␣ activity on the resistin promoter may also contribute to the ATRA effect. We favor the idea that this impairment is mediated by RARs because liganded RARs are known to interfere with the transcriptional activity of C/EBPs (26); in addition, rosiglitazone treatment did not affect the occupancy of the resistin promoter by C/EBP-␣ or p300/ CBP (23), making it unlikely that the interference detected here is mediated by liganded PPAR-␥:RXR. Liganded retinoid receptors may also repress resistin gene expression through direct binding to negative RA and/or PPAR reDIABETES, VOL. 53, APRIL 2004


TABLE 2 Effects of ATRA administration on body weight, food intake, fat-pad mass, and circulating parameters of NMRI mice

Body weight before treatment (g) Body weight after treatment (g) Food intake during treatment (kcal 䡠 g animal⫺1 䡠 day⫺1) iBAT mass (mg) eWAT mass (mg) iWAT mass (mg) Serum insulin (pmol/l) Serum triacylglycerol (mmol/l) Serum glucose (mmol/l)

Control

10 mg 䡠 kg⫺1 䡠 day⫺1



ANOVA

45.2 ⫾ 1.7 46.2 ⫾ 1.5a

44.4 ⫾ 0.9 41.5 ⫾ 1.1b

45.7 ⫾ 1.5 38.8 ⫾ 1.3bc

43.9 ⫾ 1.4 37.5 ⫾ 1.2c

RA

398 ⫾ 16 206 ⫾ 17a 1,151 ⫾ 133 592 ⫾ 54a 15.7 ⫾ 3.3 3.73 ⫾ 0.38a 7.69 ⫾ 0.67

329 ⫾ 16 121 ⫾ 12b 926 ⫾ 138 524 ⫾ 52ab 13.4 ⫾ 3.1 1.91 ⫾ 0.13b 7.37 ⫾ 0.58

404 ⫾ 70 90 ⫾ 8b 885 ⫾ 78 431 ⫾ 47bc 11.0 ⫾ 1.9 1.72 ⫾ 0.25b 6.88 ⫾ 0.41

350 ⫾ 27 85 ⫾ 8b 821 ⫾ 148 370 ⫾ 36c 12.2 ⫾ 1.8 2.22 ⫾ 0.23b 7.07 ⫾ 0.71

RA RA RA

Twelve-week-old NMRI male mice received a daily subcutaneous injection of ATRA at a dose of 10, 50, or 100 mg/kg body wt during the 4 days before they were killed. Control animals received vehicle (olive oil). Data are means ⫾ SE of seven animals per group. ANOVA significances P ⬍ 0.05: RA, effect of ATRA treatment. Values within a row not sharing a common letter are statistically different by least significant difference post hoc comparison (P ⬍ 0.05).

sponse elements in its promoter or in an enhancer. No conventional RA response elements were reported in the promoter region of either the mouse or the human resistin gene analyzed to date (23,25), and no functional PPAR-␥ binding site was found within 6.2 kb upstream of the transcriptional start site of the mouse resistin gene (23), but it remains possible that negative response elements are present at other locations. Both resistin and leptin mRNA levels were reduced in adipose tissues of ATRA-treated NMRI mice (Fig. 3). These results add to the striking similarity between the regulation of these two adipokines, both of which seem to be negatively controlled by PPAR-␥ (23,24,44) and positively controlled by C/EBP-␣ (23,24,45) at the transcriptional level. Acute RA treatment causes in rodents a reduction of body weight and adiposity, not explained solely by the observed changes in food intake, that has been reproduced here (Table 2) and was previously shown to correlate with a depressed adipogenic/lipogenic potential in adipose tissues (16) and an increased expression of UCPs in BAT and muscle (13,15,17). Teleologically, reduction of both a signal that inhibits food intake and enhances energy expenditure (leptin) and a signal that may function as an inhibitor of adipocyte differentiation and as a resistance factor to the anabolic action of insulin (resistin) may be part of a security mechanism to avoid fat depletion in the

context of an overall effect of high RA doses promoting fat mobilization. An improvement of glucose tolerance paralleled the reduction of adipose tissue resistin mRNA and circulating resistin levels in ATRA-treated NMRI mice, which is compatible with resistin’s acting as an insulin-resistance factor in normal mice. However, it is likely that the improvement of glucose tolerance is, at least in part, a secondary response to the reduction of body weight elicited by ATRA treatment. In addition, other effects of ATRA that may contribute to changes in glucose tolerance—including effects on the expression and circulating levels of other adipokines potentially related to insulin sensitivity, such as tumor necrosis factor-␣ or adiponectin— cannot be discarded. Further studies are needed to evaluate precisely the relationship among vitamin A status, insulin sensitivity, and adipokine expression. Not only acute ATRA treatment but also chronic dietary vitamin A supplementation reduced resistin expression, as evidenced in two different strains of mice (Table 3). Remarkably, although resistin expression was reduced in high-fat diet–fed B6 mice as compared with their lean controls, the inhibitory effects of high-fat diet and vitamin A supplementation on the resistin gene were not additive (Table 3), suggesting that common mechanisms are involved. Reduced resistin mRNA levels in fat depots of

TABLE 3 Effect of chronic dietary vitamin A supplementation on eWAT resistin mRNA levels and circulating resistin levels in B6 and NMRI mice

B6 mice eWAT resistin mRNA levels (%) Serum resistin levels (%) NMRI mice eWAT resistin mRNA levels (%) Serum resistin levels (%)

NF

NF ⫹ A

HF

HF ⫹ A

ANOVA

100 ⫾ 10.6 100 ⫾ 7.4

40.5 ⫾ 14.1* 45.9 ⫾ 12.4*

41.5 ⫾ 8.4† 44.8 ⫾ 12.1†

27.9 ⫾ 5.5 45.0 ⫾ 11.3

F, A, FxA F, A, FxA

100 ⫾ 6.8 100 ⫾ 11.1

65.0 ⫾ 5.4* 61.3 ⫾ 10.0*

Four-week-old B6 mice were fed one of the following diets for 18 weeks: a normal-fat diet with the standard vitamin A content (NF group), a normal-fat diet with a 40-fold excess vitamin A (NF ⫹ A group), a high-fat diet with the standard vitamin A content (HF group), or a high-fat diet with a 40-fold excess vitamin A (HF ⫹ A group). In an independent experiment, 4-week-old NMRI mice were fed the normal-fat diet with the standard vitamin A content or a 40-fold excess vitamin A for 11 weeks. Data are means ⫾ SE of four to six animals per group and are expressed relative to the mean value of the corresponding NF group, which was set at 100%. ANOVA significances, P ⬍ 0.05: F, effect of high-fat diet; A, effect of dietary vitamin A supplementation; FxA, interaction of high-fat diet and vitamin A supplementation. Student’s t test significances, P ⬍ 0.05: *vitamin A supplemented vs. nonsupplemented; †high fat vs. normal fat. DIABETES, VOL. 53, APRIL 2004

887


obese animals is well established (reviewed in 2– 4). Concerning circulating resistin, our finding of reduced levels in obese B6 mice is in line with results of Rajala et al. (46) and Maebuchi et al. (47) and in contrast with the increased serum levels reported by Steppan et al. (1) in rodents with genetic and diet-induced obesity and diabetes. Changes of circulating resistin in human obesity are also controversial (48,49). The explanation for these differences is unclear: for diet-induced obesity models, it may be related to differences in the composition of the high-fat diets used. Reduced circulating resistin in obese animals apparently conflicts with resistin’s being a critical link between obesity and insulin resistance. However, as noted previously (4), the correlation between circulating levels of a bioactive peptide and its effects may not be linear, because receptor and postreceptor mechanisms (which for resistin are largely unknown) may be critical. Gene-nutrient interactions are likely to contribute to the current epidemic of obesity and type 2 diabetes in industrialized societies. The inhibition of resistin expression by vitamin A characterized in this work may be a specific gene-nutrient interaction of interest in this context, in view of the potential relationship of resistin with insulin resistance, which is currently the subject of an intense debate (2– 4). ACKNOWLEDGMENTS

This work was supported by the Spanish Government (Direccio´ n General de Investigacio´ n, BFI2000-0988-C0601; Ministerio de Sanidad y Consumo, FIS 01/1379 and G03/028), and the European Union (COST Action 918; DLARFID project, QLRT-2001-00183). F.F. was the recipient of a doctoral fellowship from the Spanish Government. We thank Dr. Francisca Serra for defining the resistin probe used in the Northern blot analysis. REFERENCES 1. Steppan CM, Bailey ST, Bhat S, Brown EJ, Banerjee RR, Wright CM, Patel HR, Ahima RS, Lazar MA: The hormone resistin links obesity to diabetes. Nature 409:307–312, 2001 2. Ukkola O: Resistin—a mediator of obesity-associated insulin resistance or an innocent bystander? Eur J Endocrinol 147:571–574, 2002 3. Steppan CM, Lazar MA: Resistin and obesity-associated insulin resistance. Trends Endocrinol Metab 13:18 –23, 2002 4. Hotamisligil GS: The irresistible biology of resistin. J Clin Invest 111:173– 174, 2003 5. Rajala MW, Obici S, Scherer PE, Rossetti L: Adipose-derived resistin and gut-derived resistin-like molecule-beta selectively impair insulin action on glucose production. J Clin Invest 111:225–230, 2003 6. Smith SR, Bai F, Charbonneau C, Janderova L, Argyropoulos G: A promoter genotype and oxidative stress potentially link resistin to human insulin resistance. Diabetes 52:1611–1618, 2003 7. Kim KH, Lee K, Moon YS, Sul HS: A cysteine-rich adipose tissue-specific secretory factor inhibits adipocyte differentiation. J Biol Chem 276:11252– 11256, 2001 8. Bonet ML, Ribot J, Felipe F, Palou A: Vitamin A and the regulation of fat reserves. Cell Mol Life Sci 60:1311–1321, 2003 9. Safonova I, Darimont C, Amri EZ, Grimaldi P, Ailhaud G, Reichert U, Shroot B: Retinoids are positive effectors of adipose cell differentiation. Mol Cell Endocrinol 104:201–211, 1994 10. Kuri-Harcuch W: Differentiation of 3T3 F442A cells into adipocytes is inhibited by retinoic acid. Differentiation 23:164 –169, 1982 11. Chawla A, Lazar MA: Peroxisome proliferator and retinoid signaling pathways co-regulate preadipocyte phenotype and survival. Proc Natl Acad Sci U S A 91:1786 –1790, 1994 12. Alvarez R, De Andre´ s J, Yubero P, Vin˜ as O, Mampel T, Iglesias R, Giralt M, Villarroya F: A novel regulatory pathway of brown fat thermogenesis. 888

Retinoic acid is a transcriptional activator of mitochondrial uncoupling protein gene. J Biol Chem 270:5666 –5673, 1995 13. Puigserver P, Va´ zquez F, Bonet ML, Pico´ C, Palou A: In vitro and in vivo induction of brown adipocyte uncoupling protein (thermogenin) by retinoic acid. Biochem J 317:827– 833, 1996 14. Carmona MC, Valmaseda A, Iglesias R, Mampel T, Vinas O, Giralt M, Villarroya F: 9-cis retinoic acid induces the expression of the uncoupling protein-2 gene in brown adipocytes. FEBS Lett 441:447– 450, 1998 15. Bonet ML, Oliver J, Pico´ C, Felipe F, Ribot J, Cinti S, Palou A: Opposite effects of vitamin A deficient diet-feeding and retinoic acid treatment on brown adipose tissue UCP1, UCP2 and leptin expression. J Endocrinol 166:511–517, 2000 16. Ribot J, Felipe F, Bonet ML, Palou A: Changes of adiposity in response to vitamin A status correlate with changes of PPAR gamma 2 expression. Obes Res 9:500 –509, 2001 17. Felipe F, Bonet ML, Ribot J, Palou A: Up-regulation of muscle uncoupling protein 3 gene expression in mice following high fat diet, dietary vitamin A supplementation and acute retinoic acid-treatment. Int J Obes Relat Metab Disord 27:60 – 69, 2003 18. Aranda A, Pascual A: Nuclear hormone receptors and gene expression. Physiol Rev 81:1269 –1304, 2001 19. Teruel T, Hernandez R, Benito M, Lorenzo M: Rosiglitazone and retinoic acid induce uncoupling protein-1 (UCP-1) in a p38 mitogen-activated protein kinase-dependent manner in fetal primary brown adipocytes. J Biol Chem 278:263–269, 2003 20. Radominska-Pandya A, Chen G, Czernik PJ, Little JM, Samokyszyn VM, Carter CA, Nowak G: Direct interaction of all-trans-retinoic acid with protein kinase C (PKC): implications for PKC signaling and cancer therapy. J Biol Chem 275:22324 –22330, 2000 21. Mukherjee R, Davies PJ, Crombie DL, Bischoff ED, Cesario RM, Jow L, Hamann LG, Boehm MF, Mondon CE, Nadzan AM, Paterniti JR Jr, Heyman RA: Sensitization of diabetic and obese mice to insulin by retinoid X receptor agonists. Nature 386:407– 410, 1997 22. Liu YL, Sennitt MV, Hislop DC, Crombie DL, Heyman RA, Cawthorne MA: Retinoid X receptor agonists have anti-obesity effects and improve insulin sensitivity in Zucker fa/fa rats. Int J Obes Relat Metab Disord 24:997–1004, 2000 23. Hartman HB, Hu X, Tyler KX, Dalal CK, Lazar MA: Mechanisms regulating adipocyte expression of resistin. J Biol Chem 277:19754 –19761, 2002 24. Song H, Shojima N, Sakoda H, Ogihara T, Fujishiro M, Katagiri H, Anai M, Onishi Y, Ono H, Inukai K, Fukushima Y, Kikuchi M, Shimano H, Yamada N, Oka Y, Asano T: Resistin is regulated by C/EBPs, PPARs, and signaltransducing molecules. Biochem Biophys Res Commun 299:291–298, 2002 25. Seo JB, Noh MJ, Yoo EJ, Park SY, Park J, Lee IK, Park SD, Kim JB: Functional characterization of the human resistin promoter with adipocyte determination- and differentiation-dependent factor 1/sterol regulatory element binding protein 1c and CCAAT enhancer binding protein-alpha. Mol Endocrinol 17:1522–1533, 2003 26. Schwarz EJ, Reginato MJ, Shao D, Krakow SL, Lazar MA: Retinoic acid blocks adipogenesis by inhibiting C/EBP␤-mediated transcription. Mol Cell Biol 17:1552–1561, 1997 27. Decker T, Lohmann-Matthes ML: A quick and simple method for the quantitation of lactate dehydrogenase release in measurements of cellular cytotoxicity and tumor necrosis factor (TNF) activity. J Immunol Methods 115:61– 69, 1988 28. Puigserver P, Herron D, Gianotti M, Palou A, Cannon B, Nedergaard J: Induction and degradation of the uncoupling protein thermogenin in brown adipocytes in vitro and in vivo. Biochem J 284:393–395, 1992 29. Oliver J, Pico´ C, Serra F, Palou A: Resistin expression in different adipose tissue depots during rat development. Mol Cell Biochem 252:397– 499, 2003 30. Viengchareun S, Zennaro MC, Pascual-Le Tallec L, Lombes M: Brown adipocytes are novel sites of expression and regulation of adiponectin and resistin. FEBS Lett 532:345–350, 2002 31. Harmon MA, Boehm MF, Heyman RA, Mangelsdorf DJ: Activation of mammalian retinoid X receptors by the insect growth regulator methoprene. Proc Natl Acad Sci U S A 92:6157– 6160, 1995 32. Erickson RL, Hemati N, Ross SE, MacDougald OA: p300 coactivates the adipogenic transcription factor CCAAT/enhancer-binding protein alpha. J Biol Chem 276:16348 –16355, 2001 33. Puigserver P, Nadal-Ginard B, Palou A: Expression and interaction of C/EBP␣ adipogenic transcription factor and retinoblastoma protein in adipocytes during differentiation (Abstract). Int J Obes Relat Metab Disord 18 (Suppl. 2):113, 1994 34. Chen P, Riley DJ, Chen Y, Lee W: Retinoblastoma protein positively regulates terminal adipocyte differentiation through direct interaction with C/EBPs. Genes Dev 10:2794 –2804, 1996 DIABETES, VOL. 53, APRIL 2004


35. Puigserver P, Ribot J, Serra F, Gianotti M, Bonet ML, Nadal-Ginard B, Palou A: Involvement of the retinoblastoma protein in brown and white adipocyte cell differentiation: functional and physical association with the adipogenic transcription factor C/EBP␣. Eur J Cell Biol 77:117–123, 1998 36. Kumar MV, Scarpace PJ: Differential effects of retinoic acid on uncoupling protein-1 and leptin gene expression. J Endocrinol 157:237–243, 1998 37. Kumar MV, Sunvold GD, Scarpace PJ: Dietary vitamin A supplementation in rats: suppression of leptin and induction of UCP1 mRNA. J Lipid Res 40:824 – 829, 1999 38. Way JM, Gorgun CZ, Tong Q, Uysal KT, Brown KK, Harrington WW, Oliver WR Jr, Willson TM, Kliewer SA, Hotamisligil GS: Adipose tissue resistin expression is severely suppressed in obesity and stimulated by peroxisome proliferator-activated receptor gamma agonists. J Biol Chem 276:25651– 25653, 2001 39. Fukui Y, Motojima K: Expression of resistin in the adipose tissue is modulated by various factors including peroxisome proliferator-activated receptor alpha. Diabetes Obes Metab 4:342–345, 2002 40. Li Y, Lazar MA: Differential gene regulation by PPAR␥ agonist and constitutively active PPAR␥2. Mol Endocrinol 16:1040 –1048, 2002 41. Moore GB, Chapman H, Holder JC, Lister CA, Piercy V, Smith SA, Clapham JC: Differential regulation of adipocytokine mRNAs by rosiglitazone in db/db mice. Biochem Biophys Res Commun 286:735–741, 2001 42. Haugen F, Jorgensen A, Drevon CA, Trayhurn P: Inhibition by insulin of resistin gene expression in 3T3–L1 adipocytes. FEBS Lett 507:105–108, 2001 43. Shojima N, Sakoda H, Ogihara T, Fujishiro M, Katagiri H, Anai M, Onishi Y, Ono H, Inukai K, Abe M, Fukushima Y, Kikuchi M, Oka Y, Asano T:

DIABETES, VOL. 53, APRIL 2004

Humoral regulation of resistin expression in 3T3-L1 and mouse adipose cells. Diabetes 51:1737–1744, 2002 44. Hollenberg AN, Susulic VS, Madura JP, Zhang B, Moller DE, Tontonoz P, Sarraf P, Spiegelman BM, Lowell BB: Functional antagonism between CCAAT/enhancer binding protein-alpha and peroxisome proliferator-activated receptor-gamma on the leptin promoter. J Biol Chem 272:5283–5290, 1997 45. Hwang CS, Mandrup S, MacDougald OA, Geiman DE, Lane MD: Transcriptional activation of the mouse obese (ob) gene by CCAAT/enhancer binding protein alpha. Proc Natl Acad Sci U S A 93:873– 877, 1996 46. Rajala MW, Lin Y, Ranalletta M, Yang XM, Qian H, Gingerich R, Barzilai N, Scherer PE: Cell type-specific expression and coregulation of murine resistin and resistin-like molecule-alpha in adipose tissue. Mol Endocrinol 16:1920 –1930, 2002 47. Maebuchi M, Machidori M, Urade R, Ogawa T, Moriyama T: Low resistin levels in adipose tissues and serum in high-fat fed mice and genetically obese mice: development of an ELISA system for quantification of resistin. Arch Biochem Biophys 416:164 –170, 2003 48. Azuma K, Katsukawa F, Oguchi S, Murata M, Yamazaki H, Shimada A, Saruta T: Correlation between serum resistin level and adiposity in obese individuals. Obes Res 11:997–1001, 2003 49. Lee JH, Chan JL, Yiannakouris N, Kontogianni M, Estrada E, Seip R, Orlova C, Mantzoros CS: Circulating resistin levels are not associated with obesity or insulin resistance in humans and are not regulated by fasting or leptin administration: cross-sectional and interventional studies in normal, insulin-resistant, and diabetic subjects. J Endocrinol Metab 88:4848 – 4856, 2003

889