Diabetes Publish Ahead of Print, published online December 27, 2007

Forkhead FoxO1 in adipose tissues regulates energy storage and expenditure

Jun Nakae1, Yongheng Cao1, Miyo Oki1, Yasuko Orba2, Hirofumi Sawa3, Hiroshi Kiyonari4, Kristy Iskandar5, Koji Suga1, Marc Lombes6, Yoshitake Hayashi5

1: 21st Century COE Program for Signal Transduction Disease: Diabetes Mellitus as Model, Department of Clinical Molecular Medicine, Division of Diabetes, Digestive and Kidney Disease, Kobe University Graduate school of Medicine, Kobe 650-0017, Japan 2: Laboratory of Molecular & Cellular Pathology, Hokkaido University Graduate School of Medicine、N15, W7, Kita-ku, Sapporo 060-8638, Japan 3: 21st Century COE Program for Zoonosis Control, Department of Molecular Pathobiology, Hokkaido University Research Center for Zoonosis Control, N18, W9, Kita-ku, Sapporo 060-0818, Japan 4: Laboratory for Animal Resources and Genetics Engineering Team, RIKEN Center for Developmental Biology (CDB), 2-2-3 Minatojima-minami-cho, Chuou-ku Kobe, Hyogo 650-0047, Japan 5: Division of Molecular Medicine and Medical Genetics, International Center for Medical Research and Treatment (ICMRT), Kobe University Graduate School of Medicine, Kobe 650-0017, Japan 6: INSERM U 693, Faculte de Medecine Paris-Sud, 63, rue Gabriel Peri, 94276 Le Kremlin Bicêtre Cedex, France

Running title: FoxO1 in adipose tissues

Correspondence: Jun Nakae, MD Department of Clinical Molecular Medicine, Division of Diabetes, Digestive and Kidney Diseases, Kobe University Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku, Kobe 650-0017, Japan. [email protected]

Received for publication 26 May 2007 and accepted in revised form 12 December 2007.

Additional information for this article can be found in an online appendix at http://diabetes.diabetesjournals.org.

Copyright American Diabetes Association, Inc., 2007 FoxO1 in adipose tissues

ABSTRACT

Objective: Adipose tissues serve as integrators of various physiological pathways, energy balance and glucose homeostasis. Forkhead transcription factor FoxO (Forkhead bOX-containing O subfamily) 1 mediates insulin action at transcriptional level. However, physiological roles of FoxO1 in adipose tissues remain unclear.

Research Design and Methods: In the present study, we generated adipose tissue- specific FoxO1 transgenic mice (aP2-FLAG-∆256) using adipose-specific adipocyte protein 2 (aP2) promoter/enhancer and a mutant FoxO1 (FLAG∆256) in which the carboxyl terminal transactivation domain was deleted. Using these mice, we analyzed effects of overexpression of the FLAG∆256 on glucose metabolism and energy homeostasis.

Results: The aP2-FLAG-∆256 showed improved glucose tolerance and insulin sensitivity accompanied with smaller size of adipocytes and increased Adiponectin (Adipoq) and Glut 4 (Slc2a4) and decreased Tnfα (Tnf) and Ccr2 (Ccr2) expression levels in white adipose tissue (WAT) under high fat diet. Furthermore, the aP2-FLAG-∆256 had increased oxygen consumption accompanied with increased expression of PGC-1α protein and Ucp-1 (Ucp1), Ucp-2 (Ucp2), and β3-AR (Adrb3) in brown adipose tissue (BAT). Overexpression of the FLAG∆256 in T37i cells, which are derived from hibernoma of SV40 large T antigen transgenic mice, increased expression of PGC-1α protein and Ucp1. Furthermore, knockdown of endogenous FoxO1 in T37i cells increased Pgc1α (Ppargc1a), Pgc1β (Ppargc1b), Ucp1 and Adrb3 gene expression.

Conclusions: These data suggest that FoxO1 modulates energy homeostasis in WAT and BAT through regulation of adipocyte size and adipose tissue-specific gene expression in response to excessive calorie intake.

2 FoxO1 in adipose tissues

he incidence of obesity has been Adaptive thermogenesis is defined as increasing in the world wide. heat production in response to cold TO besity is the result of energy exposure or overfeeding and it serves the imbalance and it can develop when purpose of protecting the organism from energy intake exceeds energy cold exposure or regulating energy expenditure (1). Energy intake is almost balance after changes in diet. Diet is a entirely determined by food intake (minus potent regulator of adaptive whatever fails to be absorbed), which is thermogenesis. Starvation or calorie regulated by leptin, insulin, adiponectin, restriction decreases resting metabolic interleukins, cholecystokinin, PYY3-36, rate. In contrast, feeding increases ghrelin, glucagon-like peptide-1, malonyl energy expenditure. Brown adipose CoA and long-chain fatty acid through tissue (BAT) and skeletal muscle are the actions on the mediobasal hypothalamus two major organs involved in adaptive or the nucleus of the solitary tract in the thermogenesis (8). While rodents have brainstem (2)(3)(4; 5). On the other hand, prominent brown fat depots, this isn’t the energy expenditure has more case in larger mammals, including components, including basal metabolism, humans, although there may be brown fat physical activity and adaptive cells dispersed among the adipocytes of thermogenesis (1). white adipose tissue (WAT) (1). The In situations of excess energy intake, adaptive thermogenic program in both excess calories will be stored primarily in BAT and skeletal muscle involves the adipose tissues. However, in a situation stimulation of mitochondria biogenesis, of chronic calorie overload, subcutaneous increased fatty acid oxidation, and the adipose tissue eventually reaches its uncoupling of oxidative phosphorylation upper limit for further triglyceride storage, (9). In response to cold exposure, and this may trigger adipose inflammation peroxisome proliferators-activated as well as lipid ‘spill-over’. This means -γ (PPAR-γ)-coactivator-1α that energy storage will be partitioned (PGC-1α) expression is increased. PGC- towards the visceral fat depot and 1α enhances the expression of BAT- subsequently into ectopic fat depots, specific uncoupling protein 1 (UCP-1), which include intrahepatocellular and which dissipates the proton gradient intramyocellular lipids, both of which will across the inner mitochondrial have direct negative effects on insulin membranes that is produced by the action in these tissues (1; 6). The action of the electron transport chain (9; capacity of fat storage could be 10). The essential role of PGC-1α in determined by genetic programming of adaptive thermogenesis is demonstrated recruitment of new adipocytes or by by the observation that PGC-1α-knockout neuroendocrine interactions as well as mice are unable to withstand a cold adipose tissue inflammation. It has been stress with reduced UCP-1 expression suggested that an increased adipocyte (11). size is associated with insulin resistance Insulin and/or IGF-1 signaling are also (7). Therefore, the prevention of involved in development and functional generation of enlarged adipocytes might maintenance of BAT. BAT-specific insulin lead to inhibition of development of insulin receptor (IR) knockout mice (BATIRKO) resistance in whole body. showed an age-dependent atrophy of

3 FoxO1 in adipose tissues

brown adipose tissue (12). Cells lacking overexpression of the FLAG∆256 in both IRS-1 and IRS-3 failed to adipose tissues ameliorated insulin differentiate with reduced expression of resistance induced by high-fat diet Ppargca1 and Ucp1 (13). These effects of accompanied with smaller sizes of insulin/IGF-1 and IRSs on function of adipocytes in WAT and increased oxygen brown adipocytes might be mediated consumption through increased through PI 3-kinase and Akt (14). expression of PGC-1α protein. These However, very little is known about data indicate that FoxO1 in adipose downstream targets of PI 3-kinase and tissues might regulate energy Akt and transcriptional regulation by homeostasis through regulation of insulin/IGF-1 yet. adipose tissue-specific gene expression. Forkhead transcription factors of the FoxO (Forkhead bOX-containing protein, RESEARCH DESIGN AND METHODS O subfamily) family are phosphorylated Antibodies. We purchased anti-Flag mainly by Akt and regulated by (M2), anti-tubulin and anti-UCP1 (M-17) insulin/IGF-1. FoxO family is conserved from Sigma, anti-FOXO1 (N18 and H128) across many species (15). In mammals, and anti-PGC-1 (H-300) from Santa Cruz InsR/IGF1R-PI3K-Akt signaling inhibits Biotechnology, anti-HA monoclonal transcription by FoxO1, FoxO3a and antibody (12CA5) from Roche, and anti- FoxO4. These possess a phospho FOXO1 (pS256 and pT24) and forkhead DNA binding domain consisting anti-4E-BP1 from Cell Signaling of around 110 amino acids and a Technology. We used anti-FOXO1 transactivation domain in the C-terminus. antibody (H128) only when we tried to FoxOs bind to consensus FoxO binding discriminate endogenous FoxO1 from sites (T(G/A)TTTT(G/T)) in the promoter FLAG∆256 clearly because the epitope of region of their target and activate H128 was the carboxyl terminus of gene expression (16). It has been FOXO1 which FLAG-∆256 didn’t have. reported that FoxOs were expressed in Generation of adipose tissue-specific adipose tissues and that a constitutively FoxO1 transgenic mice. We cloned a nuclear mutant FoxO1 inhibited mutant FoxO1 cDNA, bearing a truncated differentiation of preadipocyte cell line FoxO1 (a.a 1-255) (∆256) that has no 3T3-F442A cells and that transactivation domain and a FLAG haploinsufficiency of FoxO1 reduced cell peptide tag, into the SmaI site of plasmid size and increased cell numbers of white pCMV5-aP2, in which ClaI and SmaI- adipocytes of mice under high fat-diet treated 5.4-kb promoter-enhancer (17). However, there have been no fragment of the mouse aP2 gene was reports about physiological roles of subcloned into the ClaI/SmaI-treated FoxO1 in adipose tissues, which include pCMV5/cMyc vector (19). We excised the both WAT and BAT using a tissue- transgene with ClaI and XhoI, gel-purified specific manner. and injected it into fertilized eggs of FVB In the present study, we generated X Bl6 hybrid mice. The resulting embryos adipose tissue-specific FoxO1 transgenic were implanted into CD-1 foster mothers. mice of a mutant FoxO1 (FLAG∆256) (18) We screened offspring for transgene under the control of mouse adipocyte transmission by PCR. Of five independent protein 2 (aP2) promoter. Using transgenic lines obtained, two founders transgenic mice, we demonstrated that transmitted the transgene through the

4 FoxO1 in adipose tissues

germline. Primers for genotypying were: mouse divided by its body weight. We FLAG-S1: 5’-atggactacaaagacgatgac-3’ performed oxygen consumption of three and SE1AS: 5’-gtcgagttggactggttaaac-3’. to four mice in each genotype in each Animal studies, analytical procedures, condition (normal chow or high fat-dieted glucose tolerance and insulin condition). Representative graphs were tolerance test. We used only male mice drawn from the mean and standard error for following analyses. Animals were fed calculated from data obtained in each standard chow diet and water ad libitum measurement. in sterile cages in a barrier animal facility RNA isolation, real-time PCR and with 12/12-hour light /dark cycle. All Northern blotting. Isolation of total RNA experimental protocols using mice were from tissues and cells was performed approved by the animal ethics committee using an SV Total RNA Isolation System of Kobe University Graduate School of (Promega) according to manufacturer’s Medicine. High fat diet was begun at protocol. Real-time PCR was also weaning (4 weeks of age) and continued performed as described previously (17). for 15 weeks. We used the same high-fat The primers used in this study were diet as described previously (20). We described in Supplemental table 1 measured glucose levels with a Glutest (available at Pro (Sanwa Kagaku Kenkyusho Co.), and http://diabetes.diabetesjournals.org). insulin, leptin and resistin by Northern blotting was performed radioimmunoassay (Linco). We carried according to the standard techniques. For out all assays in duplicate. Each value preparation of the probe for FoxO1, we represents the means of two independent treated pFLAG CMV-2 WT FoxO1 (22) determinations. Analysis was limited to with EcoRI and the probe β-actin used male mice, as they are more susceptible here were described previously (18). to insulin resistance and diabetes. Western blotting. We homogenized Glucose tolerance and insulin tolerance tissues in buffer containing 50mM test were performed as described (21). Tris.HCl (pH 8.0), 250 mM NaCl, 1% Hepatic glycogen measurement was NP40, 0.5% deoxycholate, 0.1% SDS, performed as described previously (21). and protease inhibitors (Roche CT scanning was performed using Diagnostics). After centrifugation to LaTheta LCT-100M (ALOKA CO., LTD). remove insoluble material, each 50 µg of Measurement of oxygen consumption. lysate was electrophoresed in 8% or 14% Six-month-old mice under normal chow SDS-PAGE, and Western blotting was diet and mice under high-fat diet for 15 performed using indicated antibodies. weeks were monitored individually in a Immunohistochemistry and metabolic cage (The O2/CO2 metabolism histological analysis. For measuring system for small animals immunohistochemistry of WAT and BAT, Model MK-5000, MUROMACHI KIKAI we incubated WATs and BATs overnight CO., LTD.) with free access to a normal in 4% paraformaldehyde and embedded chow diet and drinking water for 48 hours. them in paraffin. We then mounted Each cage was monitored for oxygen consecutive 5-µm sections on slides. consumption at 5 minutes interval After rehydration and permeabilization, throughout 48 hours period. Total oxygen we immunostained sections using OctA- consumption was calculated as Probe (D-8) (sc-807, SANTA CRUZ accumulated oxygen uptake for each BIOTECHNOLOGY, INC.). After washing

5 FoxO1 in adipose tissues with PBS, we incubated the sections into mature adipocytes was achieved sequentially with biotinylated anti-rabbit under standard conditions by incubating IgG reagent (Vector Laboratories) and subconfluent undifferentiated T37i cells VECTASTAIN Elite ABC reagent (Vector with 2 nM triiodothyronine (T3; Sigma Laboratories) and visualized with Liquid Chemical, St Louis, MO) and 20 nM DAB Substrate Chromogen System insulin (Invitrogen) for 6 days. (DakoCytomation). For histological Construction of adenoviruses encoding analysis, we removed interscapular tissue FoxO1 mutants was described elsewhere from 15-week high-fat dieted mice, fixed (18). For infection with adenoviruses, the specimens in 10% paraformaldehyde, after differentiation into mature and embedded them in paraffin. We adipocytes, cells were infected with mounted consecutive 10µm sections on adenoviruses at MOIs between 10 and slides and stained them with hematoxylin 20. At 72 hours after infection, cells were and eosin. We calculated adipose cell harvested. size with NIH Image 1.62 software by Knockdown of endogenous FoxO1 in manual tracing of at least 500 adipocytes T37i cells. We used RNAi-Ready for each genotype. Measurement of size pSIREN-Shuttle vector (Clontech) with of adipocytes in WAT was performed GCTGCAGTACTCTCCTTATGG as the using FLVFS-LS software (FLOVEL, target sequence and selected target- Tokyo) by manual tracing of at least 500 sequence of FoxO1 by INVITROGEN adipocytes for each genotype. BLOCK-iTTM RNAi Designer (Invitrogen). Fractionation of cytoplasmic and We generated scrambled control plasmid nuclear proteins of BAT. After treatment using mismatch sequence, with 1 mg/ml of Collagenase I (Roche), GCGGCAGTACTCCCCTTAGGG. After cytoplasmic and nuclear extracts of 150 transfection of cells with the knockdown- µg of BAT tissues from wild type and vector using Lipofectamine (Invitrogen), transgenic mice were fractionated by cells were induced by differentiation differential lysis of cells in two medium. At day 5 after induction, cells consecutive steps using the NE-PER were harvested and real-time PCR or extraction reagents (Pierce). Protein western blotting was performed. concentrations in the cytoplasmic and Statistical analyses. We calculated nuclear extracts were determined using descriptive statistics using ANOVA the Micro BSA protein assay kit (Pierce), followed by Fisher’s test (Statview; SAS and aliquots (30 µg) were resolved on 8 Institute Inc.). P values of less than 0.05 or 14 % SDS-PAGE, followed by western were considered significant. blotting using the indicated antibodies. Culture of a brown adipocyte cell line, RESULTS T37i cells and infection with Generation of adipose tissue-specific adenoviruses encoding FoxO1. T37i mutant FoxO1 transgenic mice. It has cells were cultured in DMEM-Ham’s F12 been already reported that FoxO family medium (Invitrogen) supplemented with members (FoxOs) were expressed in 10% fetal calf serum, 2 mM glutamine, both WAT and BAT and that 100 IU/ml penicillin, 100 µg/ml haploinsufficiency of FoxO1 altered streptomycin, and 20 mM HEPES, and histological appearance and gene were grown at 37oC in a humidified expression in WAT (17). However, it is atmosphere with 5% CO2. Differentiation not known whether FoxO1 has some

6 FoxO1 in adipose tissues

physiological roles in mature adipose with endogenous FoxOs and inhibit their tissues, especially in BAT. In order to transcriptional activity on promoter investigate roles of FoxO1 in adipose regions of their target genes in BAT. For tissues, we generated adipose tissue- subsequent analyses, we used line 2, specific FoxO1 transgenic mice (aP2- which has a higher transgene expression FLAG-∆256) using adipose tissue-specific levels in BAT than line 1. 5.4-kb promoter-enhancer fragment of the Improved insulin sensitivity in adipose mouse aP2, also called Fatty Acid Binding tissue-specific FoxO1 transgenic mice. Protein 4 (FABP4) and a FLAG-tagged In order to investigate the functional mutant FoxO1 (FLAG∆256), which lacks consequences of FLAG∆256 carboxyl terminal transactivation domain overexpression in adipose tissues on (18) (Figure 1a). It has been reported glucose metabolism and energy that aP2 promoter was active in WAT homeostasis, we measured body weight, and/or BAT (23)(24)(25) and that the blood glucose and serum insulin levels FLAG∆256 can compete with and performed intraperitoneal glucose endogenous FoxOs and inhibit and insulin tolerance tests. The aP2- expression of FoxOs’ target genes (26). FLAG-∆256 mice have a tendency to be We generated two transgenic lines in reduced body weight under normal chow which FLAG∆256 was expressed mainly diet. Especially, from 12 to 14 weeks of in BAT under normal chow condition age, they showed statistically significant (Figure 1b). Transgene expression level lower body weight than wild type mice in BAT in line 2 was more pronounced (Figure 2a). At the age of 3 months, aP2- than in line 1 (Figure 1b, lane 2 and 4). FLAG-∆256 had no significant differences The FLAG∆256 mRNA in line 2 was of fed blood glucose and serum insulin expressed in adipose tissue, especially levels compared with wild type mice (data BAT (Figure 1c, lane 2). The FLAG∆256 not shown). Intraperitoneal glucose protein was also expressed to a greater tolerance test, insulin secretion and extent in BAT than in WAT in both lines insulin tolerance test didn’t show any (Figure 1d). Finally, we examined significant differences between wild type intracellular localization of FLAG∆256 in and aP2-FLAG-∆256 mice (data not BAT from line 2 and detected a nuclear shown). At the age of 6 months, although staining of the FLAG∆256 in BAT from body weight and fed blood glucose level line 2 (Figure 1e). In BAT under normal of aP2-FLAG-∆256 were not different from chow diet, endogenous FoxO1 is wild type mice, fed serum insulin level of phosphorylated (Figure 1f, lane 1 and 2) aP2-FLAG-∆256 was significantly lower and localized both in cytosol and in than wild type mice (Figure 2b and 2c). nucleus (Figure 1g, upper panel). Intraperitoneal glucose tolerance test However, on high-fat diet, demonstrated that aP2-FLAG-∆256 phosphorylation of endogenous FoxO1 is tended to be more glucose-tolerant decreased (Figure 1f, lane 3 and 4) and compared with wild type mice but it did FoxO1 is localized mainly in nucleus not reach statistical significance (Figure 1g, upper panel). In contrast, the (Supplemental figure 1a). However, FLAG∆256 is localized constitutively in insulin secretion during intraperitoneal nucleus both in normal chow and high fat glucose tolerance test in aP2-FLAG-∆256 diet (Figure 1g, the middle panel). was significantly lower than in wild type Therefore, the FLAG∆256 may compete mice (Supplemental figure 1b). Insulin

7 FoxO1 in adipose tissues

tolerance test didn’t show any significant (Figure 2i). Insulin tolerance test also difference between aP2-FLAG-∆256 and demonstrated that insulin sensitivity of wild type mice (data not shown). aP2-FLAG-∆256 was improved Next, in order to clarify whether or not significantly compared with wild type mice the FLAG∆256 overexpression in adipose (Figure 2j). In order to assess hepatic tissues may affect glucose metabolism glucose production indirectly, we and energy homeostasis, aP2-FLAG- examined gene expression levels of ∆256 and wild type mice were fed on G6Pase (G6pc) and Pepck (Pck1), which high-fat diet for 15 weeks after weaning. are involved in , and Western blotting of endogenous FoxO1 in measured hepatic glycogen content. In WAT demonstrated that FoxO1 is both normal chow and high-fat diet, phosphorylated under normal chow diet although hepatic G6pc expression in aP2- but, under high-fat dieted condition, FLAG-∆256 was decreased compared FoxO1 is dephosphorylated (Figure 2d). with wild type mice significantly, Pck1 Interestingly, a 15-week high-fat diet expression was not changed (Figure 2k induced transgene expression in WAT as and 2l). However, hepatic glycogen well as in BAT (Figure 2e and 2f). Real- content of aP2-FLAG-∆256 was not time PCR revealed that transgene different from wild type (Figure 2m). expression in WAT under high fat diet These data suggest that overexpression increased by 5-fold compared with under of the FLAG∆256 in adipose tissues normal chow diet (Figure 2f). Consistent ameliorates insulin resistance induced by with gene expression level of transgene, high-fat diet. western blotting demonstrated that Overexpression of the mutant FoxO1 FLAG∆256 protein could be detected in in WAT affects size and gene WAT as well as in BAT (Figure 2g). expression of white adipocytes. In Immunohistochemistry showed that the order to investigate the mechanism how FLAG∆256 was mainly expressed in the the FLAG∆256 improves glucose nucleus of white adipocytes (Figure 2h). tolerance and insulin sensitivity, we Body weight was not different from wild examined adiposity, histological analyses, type mice (Figure 2a). However, fed and gene expression in WAT. CT blood glucose level of transgenic mice scanning revealed that visceral fat mass was significantly lower than wild type was increased significantly and mice (Figure 2b) and serum insulin levels subcutaneous fat mass of transgenic of transgenic mice tended to be lower mice was decreased slightly compared than wild type mice but it was not with wild type mice under high-fat diet statistically significant (Figure 2c). Food condition although muscle and whole intake and serum leptin level in aP2- adipose tissue mass of transgenic mice FLAG-∆256 were not different from wild were not changed (Figure 3a, 3b, 3c, and type mice both under normal chow and 3d). Interestingly, histological analyses of high fat-dieted condition (Supplemental epididymal fat demonstrated that size of figure 2). Fasting blood glucose level of adipocytes from transgenic mice was transgenic mice was lower than wild type smaller than wild type mice significantly mice significantly and intraperitoneal (Figure 3e, 3f, and 3g). We performed glucose tolerance test demonstrated that real-time PCR in order to investigate aP2-FLAG-∆256 were significantly more effects of the FLAG∆256 on gene glucose-tolerant than wild type mice expression of WAT-specific metabolic

8 FoxO1 in adipose tissues and inflammatory genes. Under normal be suppressed by the FLAG∆256. chow diet, Slc2a4 and Resistin (Retn) Interestingly, in WAT from aP2-FLAG- gene expression levels were increased ∆256 under normal chow diet, gene and Fatty acid synthase (Fasn) andTnf expression levels of most FoxO1 target gene expression levels in WAT from genes except p21 (Cdkn1a), Cyclin G2 transgenic mice were significantly (Ccng2), and Fas ligand (Fasl) were not decreased compared with wild type mice affected (Figure 3k). However, under (Figure 3h). Furthermore, in WAT from high-fat diet, gene expression levels of transgenic mice under high-fat diet, Fasn, most FoxO1-target genes except Cdkn1a Adipoq, Scl2a4, Retn and Pparγ (Pparg) and Catalase (Cat) were suppressed gene expression levels were increased (Figure 3l). These data suggest that but Tnf gene expression was suppressed overexpression of the FLAG∆256 under significantly (Figure 3i). Serum resistin high-fat diet affects size of adipocytes concentration of transgenic mice wasn’t and gene expression in WAT. different from wild type mice Increased energy expenditure by (Supplemental figure 3). In addition, overexpression of mutant FoxO1 in expression level ofCcr2, a receptor for BAT. In order to investigate effects of monocyte chemoattractant protein overexpression of the FLAG∆256 on modulating metabolic and inflammatory physiological function of BAT, we effects induced by high-fat diet (27), was measured resting rectal temperature. significantly decreased in WAT from Although rectal temperature in aP2-FLAG- transgenic mice under both normal chow ∆256 was similar to wild type mice under and high-fat diet was decreased normal chow diet, rectal temperature significantly compared with wild type mice under high-fat diet was significantly (Figure 3j). These data suggest that higher than wild type mice (Figure 4a). overexpression of a mutant FoxO1 affects Furthermore, oxygen consumption was adipocyte size and gene expression in measured using indirect calorimetry. WAT. Oxygen consumption of aP2-FLAG-∆256 Effects of a mutant FoxO1 on gene was significantly higher than wild type expression of FoxO target genes in mice under both normal chow and high WAT. It has been reported that FoxOs fat-diet conditions (Figure 4b and 4c). have various classes of target genes, Respiratory quotient, which represents which are involved in apoptosis, utilization of carbohydrate or fat for fuel regulation of cell cycle, stress resistance, source, of aP2-FLAG-∆256 was similar to and DNA repair other than glucose wild type mice under normal chow diet metabolism (28) and a constitutively (Figure 4d and 4f). However, under high- nuclear FoxO1 (ADA) inhibited adipocye fat diet, respiratory quotient of of aP - differentiation. Therefore, we 2 FLAG-∆256 tended to be higher than wild hypothesized that overexpression of type mice especially in dark state (Figure FLAG∆256 in WAT may affect gene 4e) and average of respiratory quotient of expression of FoxO target genes and aP -FLAG-∆256 was also higher than adipocyte differentiation and finally affect 2 wild type mice significantly (Figure 4f). adipocyte size and examined gene These data indicate that overexpression expression of FoxO target genes. It can of the FLAG∆256 in BAT increases be speculated that gene expression oxygen consumption and energy levels of typical FoxO target genes should expenditure and may reflect increased

9 FoxO1 in adipose tissues

carbohydrate utilization, which spared (Figure 5e). In contrast, under high-fat triglyceride oxidation and stimulated the diet, overexpression of the FLAG∆256 in accumulation of adipose tissue. BAT significantly decreased 4ebp-1 Overexpression of mutant FoxO1 in (Eif4ebp1) gene expression (Figure 5e). BAT increases the number of brown The Eif4ebp1gene has been reported to adipocytes. In order to investigate the be a target gene of dFOXO in Drosophila effects of overexpression of the (29) (30) and in mice (31). Therefore, it FLAG∆256 on BAT, we measured can be speculated that the FLAG∆256 weights of BAT and investigated suppresses Eif4ebp1 gene expression in histological appearance of BAT from wild BAT. Ppargca1 gene expression under type and transgenic mice under high-fat high-fat diet was decreased but not diet. Weights of BAT in aP2-FLAG-∆256 significant statistically. These data under normal chow diet were not different suggest that overexpression of the compared with wild type mice (data not FLAG∆256 in BAT affects gene shown). Furthermore, we could not detect expression and may ameliorate BAT any significant difference in weights of function under high fat dieted condition. BAT in aP2-FLAG-∆256 under high fat- It has been reported that targeted diet compared with wild type mice (Figure disruption of Eif4ebp1 caused increased 5a). However, numbers of brown translation of PGC1 protein (32). adipocytes in BAT from aP2-FLAG-∆256 Therefore, it is worth to investigate PGC- were increased significantly compared 1α protein level in BAT from aP2-FLAG- with wild type mice (Figure 5b and 5c). ∆256 mice. In order to examine whether These data suggest that overexpression overexpression of the FLAG∆256 in BAT of the FLAG∆256 apparently affects may affect PGC-1α protein expression morphological appearance of brown level or not, we performed western adipocytes and reverses histological blotting using lysates of BAT from wild phenotype of BAT induced by high-fat type and aP2-FLAG-∆256 mice under diet. normal chow or high fat-dieted condition. Pleiotropic effects of mutant FoxO1 on In normal chow, PGC-1α, 4E-BP1, and gene and protein expression in BAT. In UCP-1 protein levels in BAT from aP2- order to investigate effects of the FLAG-∆256 tended to be higher, but FLAG∆256 on BAT-specific gene these effects were not statistically expression, we performed real-time PCR significant (Figure 5f and 5g). However, in using total RNA of BAT from wild type high-fat diet, PGC-1α protein level was and aP2-FLAG-∆256 mice under normal increased and 4E-BP1protein level was chow or high-fat diet condition. Under decreased compared with wild type mice normal chow diet, overexpression of the significantly (Figure 5f and 5h). UCP-1 FLAG∆256 in BAT increased Ppargc1b, protein level in BAT from aP2-FLAG-∆256 Ucp1 andUcp2 gene expression was also increased but not statistically significantly compared with wild type mice significant (Figure 5f and 5h). These data (Figure 5d). Furthermore, in BAT from indicate that overexpression of the aP2-FLAG-∆256 mice under high-fat diet, FLAG∆256 in BAT increases PGC-1α Ppargc1b, Ucp1, Ucp2, andAdrb3, which protein expression level under high-fat have important roles in BAT function, diet. Consistent with the findings above, gene expression levels were increased overexpression of the FLAG∆256 in BAT compared with wild type mice significantly increased gene expression levels for

10 FoxO1 in adipose tissues

many genes of mitochondrial oxidative protein level and increased PGC-1α phosphorylation that are known to be protein level (Figure 6b). Furthermore, enriched in BAT, which include Cox5b, overexpression of the ∆256 FoxO1 CoxIII, Cox8b, Cox4i1, and cytochrome c increased Ucp1 and Adrb3 gene (Cyc) (33), both in normal chow and high- expression in T37i cells (Figure 6a). fat dieted condition (Figure 5i and 5j). These data indicate that inhibition of These data suggest that overexpression transcriptional activity FoxO1 in T37i cells of the FLAG∆256 increases mitochondrial increases PGC-1α protein and Ucp1 and gene expression and finally may up- Adrb3 gene expression levels. regulate physiological function of BAT. Knockdown of FoxO1 in T37i cells The mutant FoxO1 increased Ucp1and enhances BAT-specific gene Adrb3 gene expression levels in T37i expression. Because the truncated cell line. In order to confirm whether the FoxO1 used in the present study lacks DN FoxO1 in BAT increases PGC-1α the C-terminal transactivation domain but expression level and endow BAT-specific retains the N-terminal and forkhead DNA phenotype on cells, we infected binding domain, it is possible that this differentiated T37i cells with adenovirus mutant FoxO1 may affect functions of encoding LacZ, constitutively nuclear other Fox proteins, including FoxC2 and (CN) or ∆256 FoxO1 and analyzed gene FoxA2 (36; 37), and have effects on expression. T37i cells are derived from some FoxO-binding proteins, such as hibernoma of SV40 large T antigen Sirt1 and PGC-1α(15; 38). The best way transgenic mice under control of human to elucidate physiological roles of FoxO1 mineralocorticoid receptor promoter. T37i in adipose tissues may be the generation cells maintain the ability to undergo of adipose tissue-specific FoxO1 terminal differentiation and remain knockout mice. However, at the present capable of expressing Ucp1 upon retinoic time, we could not generate them. acid and β-adrenergic stimulation (34; Therefore, we tried to investigate 35). After T37i cells were differentiated physiological roles of FoxO1 in T37i cells into mature brown adipocytes, we using short hairpin RNA (shRNA). infected cells with adenovirus encoding Transfection with shRNA was performed FoxO1 (Figure 6b). We used at the same day as induction of adipocyte adenoviruses encoding ADA FoxO1 as differentiation and cells were harvested at CN mutant and ∆256 FoxO1 (18). At day day 5 after induction. Expression of 3 after infection, we harvested cells and shRNA reduced expression of FoxO1 by performed real-time PCR. about 80% (Figure 6c, the left panel and Overexpression of the CN FoxO1 6d). In contrast, other FoxOs, which increased Eif4ebp1 and Ppargc1a gene include FoxO3a and FoxO4, gene expression significantly (Figure 6a). In expression levels were increased contrast, Eif4ebp1and Ppargc1a gene significantly compared with cells expression levels in cells infected with the transfected with a scrambled (SCR) ∆256 FoxO1 were not changed compared control shRNA (Figure 6c, the right with the LacZ-infected cells (Figure 6a). panel). Knockdown of endogenous However, although overexpression of the FoxO1 increased gene expression levels CN FoxO1 didn’t affect 4E-BP1 and PGC- of Ppargc1a, Ppargc1b, Ucp1 and Adrb3 1α protein levels, overexpression of the significantly and increased UCP-1 protein ∆256 FoxO1 in cells decreased 4E-BP1 expression in the presence of

11 FoxO1 in adipose tissues

isoproterenol (Figure 6d and 6e). Pparg endogenous FoxOs and inhibit their gene expression in shRNA-transfected transcriptional activity. Indeed, expression cells increased significantly and Eif4ebp1 levels of several FoxO1-target genes, gene expression level was not changed which include p27 (Cdkn1b) (40), pRb compared with SCR-transfected cells (Rb1) (17), p130 (Rbl2) (41), Ccng2 (42), (Figure 6e). These data demonstrate that MnSod (Sod2) (43) and Bim (Bcl2l11) knockdown of FoxO1 confers the (44), were suppressed in WAT. The molecular phenotype of BAT cells onto effects of overexpression of the T37i cells. FLAG∆256 on glucose tolerance and insulin sensitivity under normal chow diet DISCUSSION were smaller than under high-fat diet. In the present study, we generated These can be explained by the findings two lines of adipose tissue-specific that endogenous FoxO1 in adipose mutant FoxO1 transgenic mice (aP2- tissues under normal chow diet is FLAG-∆256) using adipose tissue-specific phosphorylated and overexpression of 5.4-kb promoter-enhancer fragment of the the FLAG∆256 did not result in significant mouse aP2 gene. In both lines, the alterations in adipocytic gene expression FLAG∆256 is expressed in BAT much in adipose tissues under normal chow more than WAT under normal chow diet. diet. However, under high-fat diet, The aP2-FLAG-∆256 showed improved endogenous FoxO1 is dephosphorylated glucose tolerance and insulin sensitivity and may activate expression of its target under high-fat diet. Overexpression of the genes. In that circumstance, the FLAG∆256 in WAT lead to increased FLAG∆256 can inhibit expression of visceral fat mass, increased small FoxOs’ target genes and normalize adipocytes and altered expression levels glucose tolerance, insulin sensitivity, and of adipose tissue-related genes, such as oxygen consumption. Fasn, Adipoq, Slc2a4, Retn, Pparg, Tnf The aP2-FLAG-∆256 mice showed and Ccr2. Furthermore, overexpression of increased expression of PGC-1α protein, the FLAG∆256 in BAT leads to increased which is a key coactivator for adaptive oxygen consumption accompanied with thermogenesis in BAT (9), increased increased expression of PGC-1α, Ucp1 expression of Ucp1 gene, body and Adrb3. temperature and oxygen consumption. It Interestingly, in the present study, has been known that d4E-BP is one of high-fat diet increased expression level of target genes of dFOXO in Drosophila (29) the FLAG∆256 in WAT as well as BAT. and recently it has been reported that Because it has been reported that high- Eif4ebp1 is a target gene of FoxO1 in fat diet induced aP2 gene expression and C2C12 cells (31). Furthermore, Eif4ebp1- content (39), it can be speculated that /- mice exhibited smaller white adipose high fat diet activated aP2 promoter used tissue and an increased metabolic rate in this study and induced transgene due to increased expression of PGC-1α, expression in WAT. Under high-fat diet, Ucp1 and Adrb3 in WAT although the phosphorylation of endogenous FoxO1 is mechanism whereby reduced expression decreased in adipose tissues. The of 4E-BP1 leads to enhanced expression FLAG∆256 in WAT and BAT from of PGC-1α remains to be elucidated (32). transgenic mice was also present in Several genes that are involved in nucleus. Therefore, it can compete with mitochondrial oxidative phosphorylation

12 FoxO1 in adipose tissues

and are known to be enriched in BAT adipocyte among FoxOs. In order to from transgenic mice were increased. elucidate the mechanism how FoxO1 Therefore, it is speculated that the affects differentiation of brown FLAG∆256 in BAT may inhibit adipocytes, further investigations will be Eif4ebp1expression and increases PGC- needed. 1α protein expression through unknown Several mechanisms how the mechanism and drive the molecular FLAG∆256 in adipose tissues improved phenotype of BAT. However, at cellular glucose tolerance and insulin sensitivity levels, we could not detect any significant can be speculated. We demonstrated differences of expression levels of that, in WAT from transgenic mice, the Eif4ebp1 between LacZ- and ∆256 - ratio of small adipocytes was increased infected cells. This discrepancy has been significantly compared with wild type. shown by another study that the same Increased number of small adipocytes ∆256 FoxO1 increased adipoq expression might contribute to improved glucose although it was shown to be a target gene tolerance and insulin sensitivity despite of FoxO1 (45). One of the possibilities increased visceral fat mass. The size of might be that the FLAG∆256 used in the adipocytes is associated with insulin present study still had the N-terminus and sensitivity (46). Failure in recruitment of forkhead DNA binding domain and these new fat cells due to impaired domain could bind to some co-factors for differentiation may lead to a reduction in transcription of FoxO1-target genes in a the capacity of adipose tissue to tissue- or cell type-specific manner. Using accumulate lipid and be paralleled by knockdown of endogenous FoxO1 by enlargement of the existing adipocytes shRNA, we could demonstrate that about and also a ‘spill-over’ and ectopic 80%-knockdown of endogenous FoxO1 in accumulation of lipids in other tissues, T37i cells enhanced the molecular such as muscle and liver (47). phenotype of brown adipocytes through Suppressed expression of several increased expression of Ppargc1a, FoxO1-target genes, such as Cdkn1b, Ppargc1b, Ucp1 and Adrb3. These data Rb1, Rbl2 and Ccng2, which are involved suggest that FoxO1 may suppress gene in cell cycle arrest, might increase expression of BAT-specific genes and proliferation of newly generated also suggest the possibility that FoxO1 adipocytes. In addition, Bcl2l11, which could affect differentiation of T37i cells induces cell death, expression level was into mature brown adipocytes as well as also suppressed in WAT from transgenic the previous report that CN FoxO1 mice. In addition, adipose tissue is an inhibits differentiation of white endocrine organ, which secretes several preadipocyte cell line 3T3-F442A cells adipokines, such as adiponectin, leptin (17) because Pparg expression in T37i and resistin. Enlargement of adipocytes cells transfected with sh-FoxO1 vector might alter the amount of adipokines was increased significantly. Interestingly, released (1). Expression of the although knockdown of FoxO1 increased FLAG∆256 in WAT affects gene FoxO3a and FoxO4 expression, the expression, which include increased molecular phenotype of T37i cells as expression of Fasn, Adipoq, Slc2a4, and brown adipocyte was enhanced.This Pparg and decreased expression of Tnf finding suggest the hypothesis that and Ccr2 gene. These changes of gene FoxO1 may play a specific role in brown expression may improve insulin sensitivity

13 FoxO1 in adipose tissues

under high-fat diet. It has been already The mechanism how overexpression of reported that FoxO1 repressed Pparg the FLAG∆256 affects visceral and promoter activity and transcriptional subcutaneous fat mass differentially is regulation of Slc2a4 (48). Therefore, the still unknown. However, it has been FLAG∆256 may increase gene suggested that different adipocyte expression of Pparg and Slc2a4. precursors are responsible for a specific Transgenic mice in this study showed adipose depot development and have no significant difference of body weight different pattern of gene expression of under high-fat diet even though energy developmental genes (50). Therefore, expenditure of transgenic mice was these intrinsic differences between increased. This odd finding may be due to subcutaneous and visceral fats might effects of the FLAG∆256 in WAT. In the affect differential phenotypes. present study, smaller adipocytes in WAT from transgenic mice were increased CONCLUSION significantly compared with wild type mice The present study demonstrated that under high-fat diet. These data suggest the mutant FoxO1 increased energy store the possibility that overexpression of the in WAT through increased small FLAG∆256 might affect differentiation and adipocytes and spared triglyceride, and proliferation of newly generated increased energy expenditure in BAT adipocytes. Furthermore, the finding that through enhanced thermogenic capacity carbohydrate utilization of transgenic of brown adipocytes. Therefore, it can be mice under high-fat diet was increased speculated that FoxO1 in adipose tissues suggests that triglyceride oxidation was might decrease energy store and energy spared and triglyceride accumulation was expenditure. FoxO1 in adipose tissues increased in WAT. Indeed, Fasn may have an important role for changing expression in WAT from transgenic mice energy homeostasis in response to was increased. Therefore, autonomous excessive energy intake and identify development of WAT may equalize FoxO1 as a potential target in treatment increased energy expenditure and cause of obesity and its associated disorders, unchanged body weight. such as type 2 diabetes. The analyses using CT scanning revealed that visceral fat mass of ACKNOWLEDGEMENTS transgenic mice is increased significantly We thank Dr. Bruce M. Spiegelman compared with wild type mice under high- (Dana-Farber Cancer Institute, Boston, fat diet although subcutaneous fat mass Massachusetts, USA) for providing of aP2-FLAG-∆256 mice is not changed. promoter-enhancer fragment of the In general, increased mass of visceral mouse aP2 gene kindly. We also thank adipose tissue has been known to Dr. Susumu Seino (Division of Cellular produce a much greater risk of diabetes, and Molecular Medicine, Kobe University dyslipidemia, and accelerated Graduate School of Medicine) for making atherosclerosis than subcutaneous his laboratory available to accomplish this adipose tissue (49) (6). However, work kindly. We also thank Dr. adipocytes of visceral fat from transgenic MutsuoTaiji (Dainippon Sumitomo mice were small and could be considered Pharmaceuticals) for making a metabolic to be insulin sensitive and overcome cage available kindly. This work was st insulin resistance induced by high-fat diet. supported by a grant for the 21 Century

14 FoxO1 in adipose tissues

COE Program “Center of Excellence for to J. N, a grant from Takeda Science Signal Transduction Disease: Diabetes Foundation to J. N and a grant from Mellitus as a Model” from the Ministry of Uehara Memorial Foundation to J. N. Education, Culture, Sports, Science and Yasuko Orba is Research Fellow of the Technology of Japan, a grant from The Japan Society for the Promotion of Akiyama Foundation to J. N, a grant from Science. The Mother and Child Health Foundation

15 FoxO1 in adipose tissues

REFERENCES

1. Rosen ED, Spiegelman BM: Adipocytes as regulators of energy balance and glucose homeostasis. Nature 444:847-853, 2006 2. Barsh GS, Schwartz MW: Genetic approaches to studying energy balance: perception and integration. Nat Rev Genet 3:589-600, 2002 3. Lam TK, Schwartz GJ, Rossetti L: Hypothalamic sensing of fatty acids. Nat Neurosci 8:579- 584, 2005 4. Kubota N, Yano W, Kubota T, Yamauchi T, Itoh S, Kumagai H, Kozono H, Takamoto I, Okamoto S, Shiuchi T, Suzuki R, Satoh H, Tsuchida A, Moroi M, Sugi K, Noda T, Ebinuma H, Ueta Y, Kondo T, Araki E, Ezaki O, Nagai R, Tobe K, Terauchi Y, Ueki K, Minokoshi Y, Kadowaki T: Adiponectin stimulates AMP-activated protein kinase in the hypothalamus and increases food intake. Cell Metab 6:55-68, 2007 5. Coll AP, Farooqi IS, O'Rahilly S: The hormonal control of food intake. Cell 129:251-262, 2007 6. Despres JP, Lemieux I: Abdominal obesity and metabolic syndrome. Nature 444:881-887, 2006 7. Okuno A, Tamemoto H, Tobe K, Ueki K, Mori Y, Iwamoto K, Umesono K, Akanuma Y, Fujiwara T, Horikoshi H, Yazaki Y, Kadowaki T: Troglitazone increases the number of small adipocytes without the change of white adipose tissue mass in obese Zucker rats. J Clin Invest 101:1354-1361, 1998 8. Cannon B, Houstek J, Nedergaard J: Brown adipose tissue. More than an effector of thermogenesis? Ann N Y Acad Sci 856:171-187, 1998 9. Puigserver P, Spiegelman BM: Peroxisome proliferator-activated receptor-gamma coactivator 1 alpha (PGC-1 alpha): transcriptional coactivator and metabolic regulator. Endocr Rev 24:78- 90, 2003 10. Lowell BB, Spiegelman BM: Towards a molecular understanding of adaptive thermogenesis. Nature 404:652-660, 2000 11. Lin J, Wu PH, Tarr PT, Lindenberg KS, St-Pierre J, Zhang CY, Mootha VK, Jager S, Vianna CR, Reznick RM, Cui L, Manieri M, Donovan MX, Wu Z, Cooper MP, Fan MC, Rohas LM, Zavacki AM, Cinti S, Shulman GI, Lowell BB, Krainc D, Spiegelman BM: Defects in adaptive energy metabolism with CNS-linked hyperactivity in PGC-1alpha null mice. Cell 119:121-135, 2004 12. Guerra C, Navarro P, Valverde AM, Arribas M, Bruning J, Kozak LP, Kahn CR, Benito M: Brown adipose tissue-specific insulin receptor knockout shows diabetic phenotype without insulin resistance. J Clin Invest 108:1205-1213, 2001 13. Tseng YH, Kriauciunas KM, Kokkotou E, Kahn CR: Differential roles of insulin receptor substrates in brown adipocyte differentiation. Mol Cell Biol 24:1918-1929, 2004 14. Valverde AM, Benito M, Lorenzo M: The brown adipose cell: a model for understanding the molecular mechanisms of insulin resistance. Acta Physiol Scand 183:59-73, 2005 15. Accili D, Arden KC: FoxOs at the crossroads of cellular metabolism, differentiation, and transformation. Cell 117:421-426, 2004 16. Onuma H, Vander Kooi BT, Boustead JN, Oeser JK, O'Brien RM: Correlation between FOXO1a (FKHR) and FOXO3a (FKHRL1) binding and the inhibition of basal glucose-6- phosphatase catalytic subunit gene transcription by insulin. Mol Endocrinol 20:2831-2847, 2006

16 FoxO1 in adipose tissues

17. Nakae J, Kitamura T, Kitamura Y, Biggs WH, 3rd, Arden KC, Accili D: The forkhead transcription factor Foxo1 regulates adipocyte differentiation. Dev Cell 4:119-129, 2003 18. Nakae J, Kitamura T, Silver DL, Accili D: The forkhead transcription factor Foxo1 (Fkhr) confers insulin sensitivity onto glucose-6-phosphatase expression. J Clin Invest 108:1359-1367, 2001 19. Nakae J, Park BC, Accili D: Insulin stimulates phosphorylation of the forkhead transcription factor FKHR on serine 253 through a Wortmannin-sensitive pathway. J Biol Chem 274:15982- 15985, 1999 20. Kanda H, Tamori Y, Shinoda H, Yoshikawa M, Sakaue M, Udagawa J, Otani H, Tashiro F, Miyazaki J, Kasuga M: Adipocytes from Munc18c-null mice show increased sensitivity to insulin-stimulated GLUT4 externalization. J Clin Invest 115:291-301, 2005 21. Nakae J, Biggs WH, Kitamura T, Cavenee WK, Wright CV, Arden KC, Accili D: Regulation of insulin action and pancreatic beta-cell function by mutated alleles of the gene encoding forkhead transcription factor Foxo1. Nat Genet 32:245-253, 2002 22. Nakae J, Cao Y, Daitoku H, Fukamizu A, Ogawa W, Yano Y, Hayashi Y: The LXXLL motif of murine forkhead transcription factor FoxO1 mediates Sirt1-dependent transcriptional activity. J Clin Invest 116:2473-2483, 2006 23. Imai T, Jiang M, Chambon P, Metzger D: Impaired adipogenesis and lipolysis in the mouse upon selective ablation of the alpha mediated by a tamoxifen-inducible chimeric Cre recombinase (Cre-ERT2) in adipocytes. Proc Natl Acad Sci U S A 98:224-228, 2001 24. Abel ED, Peroni O, Kim JK, Kim YB, Boss O, Hadro E, Minnemann T, Shulman GI, Kahn BB: Adipose-selective targeting of the GLUT4 gene impairs insulin action in muscle and liver. Nature 409:729-733, 2001 25. He W, Barak Y, Hevener A, Olson P, Liao D, Le J, Nelson M, Ong E, Olefsky JM, Evans RM: Adipose-specific peroxisome proliferator-activated receptor gamma knockout causes insulin resistance in fat and liver but not in muscle. Proc Natl Acad Sci U S A 100:15712-15717, 2003 26. Nakae J, Barr V, Accili D: Differential regulation of gene expression by insulin and IGF-1 receptors correlates with phosphorylation of a single residue in the forkhead transcription factor FKHR. Embo J 19:989-996, 2000 27. Weisberg SP, Hunter D, Huber R, Lemieux J, Slaymaker S, Vaddi K, Charo I, Leibel RL, Ferrante AW, Jr.: CCR2 modulates inflammatory and metabolic effects of high-fat feeding. J Clin Invest 116:115-124, 2006 28. Greer EL, Brunet A: FOXO transcription factors at the interface between longevity and tumor suppression. Oncogene 24:7410-7425, 2005 29. Puig O, Marr MT, Ruhf ML, Tjian R: Control of cell number by Drosophila FOXO: downstream and feedback regulation of the insulin receptor pathway. Genes Dev 17:2006-2020, 2003 30. Junger MA, Rintelen F, Stocker H, Wasserman JD, Vegh M, Radimerski T, Greenberg ME, Hafen E: The Drosophila forkhead transcription factor FOXO mediates the reduction in cell number associated with reduced insulin signaling. J Biol 2:20, 2003 31. Southgate RJ, Neill B, Prelovsek O, El-Osta A, Kamei Y, Miura S, Ezaki O, McLoughlin TJ, Zhang W, Unterman TG, Febbraio MA: FOXO1 regulates the expression of 4E-BP1 and inhibits mTOR signaling in mammalian skeletal muscle. J Biol Chem 282:21176-21186, 2007

17 FoxO1 in adipose tissues

32. Tsukiyama-Kohara K, Poulin F, Kohara M, DeMaria CT, Cheng A, Wu Z, Gingras AC, Katsume A, Elchebly M, Spiegelman BM, Harper ME, Tremblay ML, Sonenberg N: Adipose tissue reduction in mice lacking the translational inhibitor 4E-BP1. Nat Med 7:1128-1132, 2001 33. Seale P, Kajimura S, Yang W, Chin S, Rohas LM, Uldry M, Tavernier G, Langin D, Spiegelman BM: Transcriptional control of brown fat determination by PRDM16. Cell Metab 6:38-54, 2007 34. Zennaro MC, Le Menuet D, Viengchareun S, Walker F, Ricquier D, Lombes M: Hibernoma development in transgenic mice identifies brown adipose tissue as a novel target of aldosterone action. J Clin Invest 101:1254-1260, 1998 35. Penfornis P, Viengchareun S, Le Menuet D, Cluzeaud F, Zennaro MC, Lombes M: The mineralocorticoid receptor mediates aldosterone-induced differentiation of T37i cells into brown adipocytes. Am J Physiol Endocrinol Metab 279:E386-394, 2000 36. Cederberg A, Gronning LM, Ahren B, Tasken K, Carlsson P, Enerback S: FOXC2 is a winged helix gene that counteracts obesity, hypertriglyceridemia, and diet-induced insulin resistance. Cell 106:563-573, 2001 37. Wolfrum C, Shih DQ, Kuwajima S, Norris AW, Kahn CR, Stoffel M: Role of Foxa-2 in adipocyte metabolism and differentiation. J Clin Invest 112:345-356, 2003 38. Puigserver P, Rhee J, Donovan J, Walkey CJ, Yoon JC, Oriente F, Kitamura Y, Altomonte J, Dong H, Accili D, Spiegelman BM: Insulin-regulated hepatic gluconeogenesis through FOXO1- PGC-1alpha interaction. Nature 423:550-555, 2003 39. Park YS, Yoon Y, Ahn HS: Platycodon grandiflorum extract represses up-regulated adipocyte fatty acid binding protein triggered by a high fat feeding in obese rats. World J Gastroenterol 13:3493-3499, 2007 40. Medema RH, Kops GJ, Bos JL, Burgering BM: AFX-like Forkhead transcription factors mediate cell-cycle regulation by Ras and PKB through p27kip1. Nature 404:782-787, 2000 41. Kops GJ, Medema RH, Glassford J, Essers MA, Dijkers PF, Coffer PJ, Lam EW, Burgering BM: Control of cell cycle exit and entry by -regulated forkhead transcription factors. Mol Cell Biol 22:2025-2036, 2002 42. Martinez-Gac L, Marques M, Garcia Z, Campanero MR, Carrera AC: Control of cyclin G2 mRNA expression by forkhead transcription factors: novel mechanism for cell cycle control by phosphoinositide 3-kinase and forkhead. Mol Cell Biol 24:2181-2189, 2004 43. Kops GJ, Dansen TB, Polderman PE, Saarloos I, Wirtz KW, Coffer PJ, Huang TT, Bos JL, Medema RH, Burgering BM: Forkhead transcription factor FOXO3a protects quiescent cells from oxidative stress. Nature 419:316-321, 2002 44. Dijkers PF, Medema RH, Lammers JW, Koenderman L, Coffer PJ: Expression of the pro- apoptotic Bcl-2 family member Bim is regulated by the forkhead transcription factor FKHR-L1. Curr Biol 10:1201-1204, 2000 45. Qiao L, Shao J: SIRT1 regulates adiponectin gene expression through Foxo1-C/enhancer- binding protein alpha transcriptional complex. J Biol Chem 281:39915-39924, 2006 46. Eriksson JW, Smith U, Waagstein F, Wysocki M, Jansson PA: Glucose turnover and adipose tissue lipolysis are insulin-resistant in healthy relatives of type 2 diabetes patients: is cellular insulin resistance a secondary phenomenon? Diabetes 48:1572-1578, 1999 47. Danforth E, Jr.: Failure of adipocyte differentiation causes type II diabetes mellitus? Nat Genet 26:13, 2000 48. Armoni M, Harel C, Karnieli E: Transcriptional regulation of the GLUT4 gene: from PPAR- gamma and FOXO1 to FFA and inflammation. Trends Endocrinol Metab 18:100-107, 2007

18 FoxO1 in adipose tissues

49. Rosen ED, MacDougald OA: Adipocyte differentiation from the inside out. Nat Rev Mol Cell Biol 7:885-896, 2006 50. Gesta S, Bluher M, Yamamoto Y, Norris AW, Berndt J, Kralisch S, Boucher J, Lewis C, Kahn CR: Evidence for a role of developmental genes in the origin of obesity and body fat distribution. Proc Natl Acad Sci U S A 103:6676-6681, 2006

19 FoxO1 in adipose tissues

FIGURE LEGENDS

Figure 1. Generation of transgenic mice expressing a mutant FoxO1 driven by the aP2 promoter. (a) Diagram of the transgenic construct. In addition to 5.4kb of the aP2 promoter, the construct contains the Flag-∆256 FoxO1 cDNA and the human GH polyadenylation sequences. (b) Northern blotting of transgene expression in WAT (lane 1 and 3) and BAT (lane 2 and 4) from line 1 (lane 1 and 2) and line 2(lane 3 and 4).(c) Tissue survey of transgene expression in line 2 at the age of 6 months under normal chow diet. Total RNA was isolated from the tissues indicated at the bottom of the autoradiogram and analyzed by Northern blotting using a FoxO1 cDNA probe. (d) Western blotting of transgene expression in BAT (lane 1 to 3) and WAT (lane 4 to 6) from wild type (lane 1 and 4), line 1 (lane 2 and 5), line 2 transgenic male mice (lane 3 and 6) at the age of 6 months under normal chow diet. The top panel shows immunoblotting of tissue lysates (50µg) with anti-FLAG mouse monoclonal antibody (M2) (SIGMA). The middle panel shows immunoblotting with anti-FOXO1 (N18) antibody. The bottom panel shows immunoblotting with anti-tubulin mouse monoclonal antibody (SIGMA). (e) Immunohistochemistry of BAT from transgenic (the left panel) and wild type mice (the right panel) using OctA-Probe (D-8) (sc-807, SANTA CRUZ BIOTECHNOLOGY, INC.). Scale bars indicate 20µm. (f) Western blotting of phosphorylated FoxO1 in BAT from wild type mice under normal chow diet (lane 1 and 2) at the age of 6 months or under 15-week high-fat diet (lane 3 and 4) at fed state. The top panel shows western blotting with anti-phospho-FOXO1 (pS256) and the middle panel shows western blotting with anti-FOXO1 (N18) and the bottom panel shows western blotting with anti-tubulin antibody. (g) Fractionation of cytoplasmic (C) and nuclear extracts (N) of BAT from transgenic mice under normal chow at the age of 6 months and 15 week high-fat diet at fed state. The top panel shows western blotting of endogenous FoxO1 with anti-FOXO1 (H128) and the middle panel shows western blotting of DN FoxO1 with anti-FLAG (M2) and the bottom panel shows western blotting of UCP-1 using anti-UCP-1 (M17).

Figure 2. Metabolic characterization of FoxO1 transgenic mice. (a) Body weight of wild type and transgenic mice under normal chow and high-fat diet. Data represent mean + s.e.m of twenty male mice for each genotype. The open circle indicates wild type under normal chow diet, the closed circle indicates transgenic mice under normal chow, the open square indicates wild type mice under high-fat diet, and the closed square indicates transgenic mice under high-fat diet. An asterisk indicates statistically significant difference between wild type and transgenic mice under normal chow diet from the age of 12 to 14 weeks (P<0.05 by one-factor ANOVA). (b) Fed blood glucose levels of wild type (the closed bar) and transgenic mice (the hatched bar) under normal chow at the age of 6 months (NC) and 15-week-high-fat diet (HF). Data represent mean + s.e.m of twenty male mice for each genotype. An asterisk indicates statistically significant difference between wild type and transgenic mice under HF condition (P<0.05 by one-factor ANOVA). (c) Fed serum insulin levels of wild type (closed bar) and transgenic mice (hatched bar) under NC or HF. Data represent mean + s.e.m of fifteen male mice for each genotype. Asterisk indicates a statistically significant difference between wild type and transgenic mice (P<0.005 by one-factor ANOVA). (d)

20 FoxO1 in adipose tissues

Western blotting of endogenous phosphorylated FoxO1 in WAT from wild type mice under NC (lane 1 and 2) and HF (lane 3 and 4). The top panel shows western blotting of phosphorylated FoxO1 with anti-pT24 and the middle panel shows western blotting of total FoxO1 with anti-FOXO1 (N18) and the bottom panel shows western blotting of tubulin. (e) Tissue survey of transgene expression in transgenic mice under 15-week high-fat diet. Total RNA was isolated from the tissues indicated at the bottom of the autoradiogram and analyzed by Northern blotting using a FoxO1 and Actin cDNA probe. (f) Real-time PCR of transgene expression in WAT (closed bar) and BAT (hatched bar) under NC and HF diet. The relative mRNA abundance of transgene is shown as folds of expression level of transgene in WAT under NC. Data represent mean + s.e.m of RNA samples from five transgenic mice for each tissue. An asterisk indicates a statistically significant difference between WAT under NC and HF (P<0.005 by one-factor ANOVA). The primers used in genotyping, FLAG-S1 and SE1AS, were used for this real-time PCR. (g) The transgene is expressed in WAT as well as BAT under a 15-week high-fat dieted condition. Lysates from WAT (lane 1 to 6) and BAT (lane 7 to 12) of wild type (lane 1 to 3 and lane 6 to 9) and transgenic mice (lane 4 to 6 and lane 10 to 12) were electrophoresed and immunoblotted with the indicated antibodies. An arrow (the upper band) indicates FLAG-∆256 and an asterisk (the lower band) indicates non-specific band. (h) Immunohistochemistry of WAT from wild type (left panel) and transgenic mice (right panel). Scale bars indicate 20µm. (i) Intraperitoneal glucose tolerance test in transgenic mice under HF diet (n=10 for each genotype). The open circle indicates wild type and the closed circle indicates transgenic mice. Asterisks indicate statistically significant differences between wild type and transgenic mice (P<0.05 by one-factor ANOVA). (j) Intraperitoneal insulin tolerance test in transgenic mice under HF diet (n=10 for each genotype). An asterisk indicates a statistically significant difference between wild type and transgenic mice (P<0.05 by one-factor ANOVA). The open circle indicates wild type and the closed circle indicates transgenic mice. (k) (l) Real-time PCR of G6pc and Pck1 using liver from wild type (closed bar) and transgenic mice (hatched bar) under normal chow (k) or high-fat diet (l). Data represent mean + s.e.m of RNA samples from five transgenic mice at fasting state. A single or double asterisk indicates a statistically significant difference of G6pc expression between wild type and transgenic mice (*P<0.01 and **P<0.05 by one-factor ANOVA). Primers for real-time PCR of G6pc and Pck1 were described elsewhere (22). (m) Hepatic glycogen content of wild type (closed bar) and transgenic mice (hatched bar) under NC and HF at fasting state.

Figure 3. Adiposity, histological analyses and gene expression of WAT from transgenic mice. (a)-(d) Muscle mass (a), adipose tissue mass (sum of visceral and subcutaneous fats) (b), visceral fat mass (c), and subcutaneous fat mass (d) were calculated as described in “RESEARCH DESIGN AND METHODS”. The data are demonstrated as % of body weight and represent as mean + s.e.m of calculation from five mice of each genotype. An asterisk indicates a statistically significant difference of visceral fat mass between wild type and transgenic mice (*P<0.01 by one-factor ANOVA). (e) Histological appearance of WAT from wild type (left panel) and transgenic mice (right panel). Sections were stained with hematoxylin and eosin. Representative photomicrographs are presented. Scale bars indicate 100µm. (f) Average of size of adipocytes from epididymal fat of wild type (closed bar) and transgenic mice (hatched

21 FoxO1 in adipose tissues

bar) under NC and HF. An asterisk indicates a statistically significant difference between wild type and transgenic mice under HF (*P<0.001 by one-factor ANOVA). (g) Distribution of sizes of adipocytes. Sizes of each adipocyte from wild type (closed bar) or transgenic mice (hatched bar) under HF were measured as described in “RESEARCH DESIGN AND METHODS”. The data represent % of number of adipocytes with each size among all measured cell number. (h) (i) Real-time PCR of WAT-specific genes using WAT from wild type (closed bar) and transgenic mice (hatched bar) under normal chow (NC) (h) or high-fat diet (HF) (i). Data represent mean + s.e.m of RNA samples from each five genotype. An asterisk indicates a statistically significant difference between wild type and transgenic mice (*P<0.01 by one-factor ANOVA). (j) Real-time PCR of Ccl2 and Ccr2 using WAT from wild type (closed bar) and transgenic mice (hatched bar) under normal chow (left panel) or high-fat diet (right panel). Data represent mean + s.e.m of RNA samples from each five genotype. A single or double asterisk indicates a statistically significant difference between wild type and transgenic mice (*P<0.01 and **P<0.05 by one-factor ANOVA). (k) (l) Real-time PCR of FoxO1-target genes using WAT from wild type (closed bar) and transgenic mice (hatched bar) under normal chow (k) or high-fat diet (l). Data represent mean + s.e.m of RNA samples from each five genotype. A single, double, or triple asterisk indicates a statistically significant difference between wild type and transgenic mice (*P<0.001, **P<0.005, and ***P<0.01 by one-factor ANOVA).

Figure 4. Body temperature and metabolic rates of transgenic mice. (a) Rectal temperature of six-month-old male mice (five for each genotype) under NC and fifteen- week-high-fat dieted male mice (five for each genotype) were measured at room temperature. The closed bar indicates wild type mice and the hatched bar indicates transgenic mice. An asterisk indicates a statistically significant difference between wild type and transgenic mice under high-fat dieted condition (P<0.01 by one-factor ANOVA). (b)(c) The oxygen consumption of wild type and transgenic mice under normal chow (NC) at the age of 6 months or under 15-week-high-fat diet (HF). The data represented were mean + s.e.m of 5 male mice for each genotype and each condition. Measurements of oxygen consumption were performed for 72 hours, with the first day allowing the mice to acclimate to the cage environment. The open circle indicates wild type and the closed circle indicates transgenic mice. The dark shadow indicates the dark phase. Asterisks indicate statistically significant differences between wild type and transgenic mice (*P<0.02 by one-factor ANOVA). (d)(e) Respiratory quotient of wild type and transgenic mice. The data represented were mean + s.e.m of 5 male mice for each genotype and each condition as described in Figure 4b and 4c. The open circle indicates wild type and the closed circle indicates transgenic mice. The dark shadow indicates the dark phase. (f) Average of respiratory quotient from Figure 4d and 4e. Closed bar indicates wild type and hatched bar indicates transgenic mice under NC or HF. An asterisk indicates a statistically significant difference between wild type and transgenic mice under HF (*P<0.05 by one-factor ANOVA).

Figure 5. Effects of the mutant FoxO1 on histological appearance, gene and protein expression in BAT. (a) Weights of BAT from wild type and transgenic mice under high-fat dieted condition for 15 weeks. The data presented are mean + s.e.m of

22 FoxO1 in adipose tissues

weight of BAT from each genotype (n=8 for each genotype). (b) Histological appearance of BAT from wild type (left panel) and transgenic mice (right panel) under 15-week-high- fat diet. Interscapular tissue was excised and fixed. Sections were stained with hematoxylin and eosin. Representative photomicrographs are presented. Scale bars indicate 20µm. (c) The number of fat cells within a random microscopic field was determined by counting cells in at least six different random fields per mouse and six mice for each genotype. The data are mean + s.e.m and represented as folds of data of wild type mice. An asterisk indicates a statistically significant difference (P<0.005 by one-factor ANOVA). (d)(e) Real-time PCR of genes that are expressed in BAT of wild type and transgenic mice at the age of 6 months under NC (d) or 15-week high-fat diet (e). Isolation of total RNA and real-time PCR were performed as described in “RESEARCH DESIGN AND METHODS”. The closed bar indicates wild type and the hatched bar indicates transgenic mice. The data presented are mean + s.e.m of three independent experiments (n=4 for each genotype and each condition). Asterisks indicate statistically significant differences between wild type and transgenic mice (*P<0.001 by one-factor ANOVA). (f) Western blotting of PGC-1α, 4E-BP1 and UCP-1 proteins in BAT from wild type and transgenic mice at the age of 6 months (NC) or 15- week high-fat diet (HF). Tissue lysates (50µg) of BAT from wild type (lane 1, 2, 5 and 6) and transgenic mice (lane 3, 4, 7 and 8) under normal chow (left panel) or high-fat dieted condition (right panel) were electrophoresed in SDS-PAGE and immunoblotted with the indicated antibodies. (g) (h) Quantitative analyses of PGC-1α, 4E-BP1 and UCP-1 protein expression levels in BAT from wild type (closed bar) and transgenic mice (hatched bar) at the age of 6 months under normal chow (g) or high fat-diet (h). Intensity of bands blotted with anti-PGC-1α, -4E-BP1 and -UCP-1were measured using NIH Image 1.62. Data represent mean+ s.e.m from three independent experiments. Asterisks indicate statistically significant differences between wild type and transgenic mice (*P<0.01 and **P<0.02 by one-factor ANOVA). (i)(j) Real-time PCR of genes of mitochondrial components in BAT of wild type and transgenic mice at the age of 6 months under NC (i) or 15-week high-fat diet (j). The closed bar indicates wild type and the hatched bar indicates transgenic mice. The data presented are mean + s.e.m of three independent experiments (n=4 for each genotype and each condition). Asterisks indicate statistically significant differences between wild type and transgenic mice (*P<0.001 and **P<0.02 by one-factor ANOVA).

Figure 6. Gene and protein expression in T37i cells infected with adenoviruses encoding the CN or ∆256 FoxO1 or transfected with shRNA of FoxO1. (a) Real-time PCR of Eif4ebp1, Ppargc1a, Ucp1 and Adrb3 using total RNA from T37i cells infected with adenoviruses encoding the indicated protein. The data presented are mean + s.e.m of three independent experiments. Asterisks indicate statistically significant differences (*P<0.05 by one-factor ANOVA). The CN indicates ADA FoxO1 in which Threonine24, Serine 253 and Serine 316 are substituted to alanine or aspartic acid and the ∆256 indicates ∆256 FoxO1 (18). An arrow in lane 2 or 3 in the bottom panel indicates the CN or ∆256 FoxO1, respectively. (b) PGC-1α and 4E-BP1 protein expression in T37i cells infected with adenoviruses encoding indicated protein. (c) Real-time PCR of FoxOs using total RNA from T37i cells transfected with shRNA of FoxO1 (sh-FoxO1) or a scrambled control shRNA (SCR). The data presented are mean + s.e.m of three

23 FoxO1 in adipose tissues independent experiments. Asterisks indicate statistically significant differences (*P<0.005 and **P<0.01 by one-factor ANOVA). (d) Western blotting of lysates of T37i cells transfected SCR (lane 1 and 2) or sh-FoxO1 (lane 3 and 4) with the indicated antibodies. At day 5 after transfection and induction of differentiation, cells were incubated in the absence (lane 1 and 3) or presence (lane 2 and 4) of 1µM isoproterenol for 6 hours and harvested. (e) Real-time PCR of Ppargc1a, Ppargc1b, Ucp1, Ucp2, Adrb3, Pparg and Eif4ebp1 using total RNA from T37i cells transfected with SCR or sh- FoxO1 in the absence (closed bar) or presence (hatched bar) of 1µM of isoproterenol for 6 hours. The data presented are mean + s.e.m of three independent experiments. Asterisks indicate statistically significant differences between SCR and sh-FoxO1 in each condition (*P<0.001 and **P<0.05 by one-factor ANOVA).

24 FoxO1 in adipose tissues

25 FoxO1 in adipose tissues

26 FoxO1 in adipose tissues

27 FoxO1 in adipose tissues

28 FoxO1 in adipose tissues

29 FoxO1 in adipose tissues

30