Metabolomic and microarray analyses of adipose tissue of dapagliflozin-treated mice, and effects of 3-hydroxybutyrate on induction of adiponectin in adipocytes

Sodium/glucose cotransporter 2 (SGLT2) inhibitor improves systemic glucose metabolism. To clarify the effect of dapagliflozin, we performed gene expression microarray and metabolomic analyses of murine adipose tissue. Three groups of mice were used; non-diabetic control KK mice (KK), diabetic KKAy mice (KKAy), and KKAy mice treated with dapagliflozin (KKAy + Dapa). Plasma glucose levels were significantly reduced in KKAy + Dapa compared with KKAy. Food consumption was larger in KKAy + Dapa than KKAy, and there were no significant differences in body and adipose tissue weight among the groups. Metabolomic analysis showed higher levels of many intermediate metabolites of the glycolytic pathway and TCA cycle in KKAy than KK, albeit insignificantly. Dapagliflozin partially improved accumulation of glycolytic intermediate metabolites, but not intermediate metabolites of the TCA cycle, compared with KKAy. Interestingly, dapagliflozin increased plasma and adipose 3-hydroxybutyric acid (3-HBA) levels. Microarray analysis showed that adipocytokines were downregulated in KKAy compared with KK mice, and upregulated by dapagliflozin. In vitro, 3-HBA induced β-hydroxybutyrylation of histone H3 at lysine 9 and upregulation of adiponectin in 3T3-L1 adipocytes independent of their acetylation or methylation. Our results suggest that 3-HBA seems to provide protection through epigenetic modifications of adiponectin gene in adipocytes.

SCientiFiC RepoRts | (2018) 8:8805 | DOI: 10.1038/s41598-018-27181-y compromised hypothalamic control of appetite 19 . As expected, food and water intakes were significantly higher in both KKAy and KKAy + Dapa than KK (Fig. 1b,c), and this effect was observed on the first week of treatment with dapagliflozin (Fig. 1b,c). During the study period, plasma glucose levels were significantly higher in KKAy and KKAy + Dapa than KK (Fig. 1d). Treatment with dapagliflozin significantly reduced plasma glucose levels in KKAy + Dapa compared with KKAy (Fig. 1d). The weight of WAT (subcutaneous, periovarian, and mesenteric fat), was larger in KKAy and KKAy + Dapa than KK, but no such difference was noted between KKAy and KKAy + Dapa (Fig. 1e). Brown adipose tissue (BAT) weight was not significantly different among the three groups (Fig. 1e). Liver weight was significantly larger in KKAy and KKAy + Dapa than KK, and was significantly lower in KKAy + Dapa than KKAy (Fig. 1e). Kidney weight was significantly larger in KKAy and KKAy + Dapa than KK, and tended to be larger in KKAy + Dapa than KKAy (Fig. 1e). Skeletal muscle weight was similar in the three groups (Fig. 1e).
Effects of dapagliflozin on blood parameters. Glycated hemoglobin levels were significantly higher in KKAy than KK, and significantly lower in KKAy + Dapa than KKAy (Fig. 2a), indicating the strong glucose-lowering effect of dapagliflozin. Plasma insulin levels were significantly higher in KKAy than KK, and significantly lower in KKAy + Dapa than KKAy (Fig. 2b), indicating improvement in insulin resistance. Plasma triglyceride levels were also higher in both KKAy and KKAy + Dapa relative to KK. The levels of plasma non-esterified fatty acids (NEFA) was 0.38 ± 0.06 mEq/L in KKAy before treatment, and 0.37 ± 0.04 mEq/L in KKAy and 0.50 ± 0.03 mEq/L in KKAy + Dapa at the end of treatment. NEFA levels tended to be higher in KKAy + Dapa than KK and KKAy (Fig. 2c,d). Plasma adiponectin levels also tended to be lower in KKAy than KK, and higher in KKAy + Dapa than KKAy, albeit insignificantly (Fig. 2e). Plasma levels of 3-HBA, a ketone body, were 215.3 ± 92.7 μmol/L in KKAy before treatment, and 249.7 ± 21.7 μmol/L in KKAy and 325.8 ± 47.4 μmol/L in KKAy + Dapa at the end of treatment. 3-HBA levels tended to be higher in both KKAy and KKAy + Dapa than KK (Fig. 2f). Plasma levels of corticosterone showed no significant differences among groups ( Supplementary  Fig. S1a). Plasma levels of leptin were significantly higher in KKAy than KK, and lower in KKAy + Dapa than KKAy ( Supplementary Fig. S1b).

Metabolomic analysis of adipose tissue.
To investigate the effects of dapagliflozin in adipose tissue, we analyzed the metabolites in periovarian WAT using CE-MS analysis. The levels of metabolites of the glycolytic pathway and tricarboxylic acid (TCA) cycle tended to be higher in KKAy than KK. The levels of metabolites of the glycolytic pathway tended to be lower in KKAy + Dapa than KKAy while those of metabolites of the TCA cycle were not different between these two groups (Fig. 3a). 3-HBA level was significantly higher in KKAy + Dapa than KK and KKAy (Fig. 3a). Carnitine plays an essential role in the transfer of fatty acids across the inner mitochondrial membrane. In the first step of transport, fatty acyl CoA is converted to acylcarnitine, which can enter the mitochondria for energy production via β-oxidation 20 . Palmitoylcarnitine, and acylcarnitine (16:2), (18:0), (18:1), (18:2), and (20:1) were elevated in KKAy + Dapa compared with KK or KKAy ( Supplementary Fig. S2), suggesting enhanced mitochondrial oxidation of fatty acids by dapagliflozin treatment.
Microarray analysis of adipose tissue. We also performed microarray analysis of gene expression in periovarian WAT. The expression levels of pdha1 and pdhb, catalyzing the conversion of pyruvate into acetyl-CoA, showed a significant 58% and 56% reduction in KKAy than KK (Fig. 3a). In addition, expression levels of gapdh, pgam1, and eno1 were significantly reduced in KKAy compared with KK (Fig. 3a). Hk2 expression levels were also significantly reduced in KKAy compared with KK (Fig. 3a). Next, we measured pyruvate carboxylase (PCX) activity in periovarian WAT of KK, KKAy and KKAy + Dapa. PCX catalyzes the physiologically irreversible carboxylation of pyruvate to form oxaloacetate (OAA), and plays an anaplerotic role to supply intermediate metabolites of the TCA cycle for biosynthesis of amino acids and fatty acids 21 . PCX activity was significantly higher in periovarian WAT of KKAy + Dapa than that of KK or KKAy ( Supplementary Fig. S3).
MCT1 and MCT2 are monocarboxylate transporters that import 3-HBA into peripheral tissues, while MTC4 exports it 15 . Succinyl-CoA-3-oxaloacid CoA transferase (SCOT) is the major enzyme in extrahepatic ketone body catabolism 22 . The mRNA expression levels of MCT1, MCT2, and SCOT, tended to be lower in KKAy than KK, and higher in KKAy + Dapa than KKAy (Fig. 3c). Furthermore, the mRNA expression level of MCT4 tended to be higher in KKAy than KK, and lower in KKAy + Dapa than KKAy (Fig. 3c). Next, we measured SCOT activity in periovarian WAT of KK, KKAy and KKAy + Dapa. SCOT activity was significantly higher in KKAy + Dapa than KKAy ( Supplementary Fig. S7).

Metabolomic analysis and microarray analysis of liver.
To investigate the effects of dapagliflozin in liver, we analyzed the metabolites in liver using CE-MS analysis. The levels of phosphoenolpyruvic acid (PEP) and 3-HBA were significantly higher in KKAy + Dapa than KKAy ( Supplementary Fig. S8). Ketogenesis is limited to hepatocytes due to restricted expression of the ketogenic enzyme mitochondrial 3-hydroxymethylglutaryl-CoA synthase (HMGCS2) 23 . In liver, AcCoA levels tended to be lower in KKAy + Dapa compared with KK or KKAy, suggesting the synthesis of 3-HBA from AcCoA ( Supplementary Fig. S8). Hepatic substrates for gluconeogenesis, such as alanine, lactate, glycerol-3-phosphate (G3P), and dihydroxyacetone phosphate (DHAP) tended to be lower in KKAy + Dapa than KK and KKAy ( Supplementary Fig. S9a).
We also performed microarray analysis of gene expression in liver. The expression levels of genes of the reaction enzymes of the glycolytic pathway, such as glucose phosphate isomerase 1 (Gpi1), glyceraldehyde 3-phosphate dehydrogenase (Gapdh), and enolase 1 (Eno1) were significantly lower in KKAy + Dapa than KKAy ( Supplementary Fig. S8).

Effects of 3-hydroxybutyric acid on mRNA expression levels of adiponectin and intracellular adiponectin protein levels in 3T3-L1 adipocytes.
3-HBA provides protection against apoptosis of cardiomyocytes 25 . Based on this observation, we postulated that 3-HBA also provides protection in adipocytes. To test the latter, we incubated 3T3-L1 adipocytes in the presence or absence of 3-HBA. The presence of 3-HBA was associated with overexpression of adiponectin mRNA and downregulation of inflammation-related genes, such as MCP-1, PAI-1, and IL-6 ( Fig. 4a-f, n = 6). Western blot analysis confirmed that the intracellular protein levels of adiponectin were significantly higher in 3T3-L1 adipocytes incubated with 3-HBA than without 3-HBA (Fig. 4g, n = 3). Treatment with 3-HBA did not affect glycerol release from 3T3-L1 adipocytes ( Supplementary Fig. S10), suggesting no major effect of 3-HBA on lipolysis.

Mechanism of 3-hydroxybutyric acid-induced increase in adiponectin.
To determine the mechanism of 3-HBA-induced increase in adiponectin expression in 3T3-L1 adipocytes, promoter analysis was performed as described in detail previously 26,27 . Luciferase activity was induced with pioglitazone in a10390Luc or a908Luc containing 10390 bp or 908 bp of human adiponectin promoter, respectively 27 , but not with 3-HBA (Fig. 5a, n = 3). Based on these results, we speculated that 3-HBA-related induction of adiponectin gene is mediated through epigenetic regulation. As reported by Kim et al. 28 , obesity-induced pro-inflammatory cytokines promote DNA methylation of adiponectin promoter, resulting in reduction of adiponectin expression. We conducted bisulfite sequencing analysis of 3T3-L1 adipocytes treated with or without 3-HBA. 3-HBA had no effect on the DNA methylation levels of the adiponectin gene (Fig. 5b). With regard to other epigenetic modifications, previous studies reported that adiponectin gene expression is associated with methylation or acetylation of histone H3 at lysine 9 (H3K9) in 3T3-L1 adipocytes 29 while another study described lysine β-hydroxybutyrylation (Kbhb), a new type of histone modification 18 . To assess modification of the histone H3K9, we performed chromatin immunoprecipitation assay (ChIP assay) followed by qPCR analysis for adiponectin gene in 3T3-L1 adipocytes (Fig. 5c). 3-HBA induced significant upregulation of Kbhb modification of H3K9 upstream (−700 bp to −100 bp) and downstreatm ( +2300 bp and +7300 bp to +9700 bp) of the translation initiation site of adiponectin gene, whereas acetylation or di-methylation did not (Fig. 5c). These results indicate that induction of adiponectin expression by 3-HBA is associated with enhanced modification of H3K9 by Kbhb in adiponectin gene.

Discussion
Increased food intake was observed in SGLT2-deficient mice 30 , and in diet-induced obese rats treated with dapagliflozin 31 . KKAy mouse is a hyperphagic obese diabetic model due to the antagonism of hypothalamic melanocortin receptor-4 by ectopic expression of the agouti protein 32,33 . In the present study, food intake was significantly higher in KKAy + Dapa than KKAy. Leptin is a satiety hormone that reduces appetite 34 , and we found that plasma leptin levels were significantly lower in KKAy + Dapa than KKAy in the current study. Regulation of appetite involves a balance between excitatory and inhibitory processes. Agouti gene mutation stimulates, whereas,  Fig. S11). Reduced plasma leptin levels by dapagliflozin is supposed to fail in suppression of appetite resulting in further enhancement of hyperphagia. Therefore, reduced leptin levels might be partially responsible for the hyperphagia at least in KKAy + Dapa.
There was no significant difference in the weight of WAT between KKAy and KKAy + Dapa. In periovarian WAT, expressions of genes involved in lipolysis, such as lipe, pnpla2, and mgll showed tendency to elevate, and acylcarnitines tended to be higher in KKAy + Dapa compared with KK or KKAy, suggesting enhancement of both lipolysis and mitochondrial oxidation of fatty acids by dapagliflozin treatment. On the other hand, expressions of genes associated with lipogenesis tended to be higher in periovarian WAT of KKAy + Dapa than KKAy. Collectively, dapagliflozin-induced lipolysis and fatty acid oxidation should be partially compensated by modest increase of lipogenesis in periovarian WAT, resulting in no significant changes in fat weight in the current study. In another condition of enhanced lipolysis by chronic β3-adrenergic receptor stimulation, Mottillo reported the coupling of lipolysis, fatty acid oxidation, and lipogenesis in adipocytes 35 . When lipolysis is activated, greater flux of fatty acids into mitochondria activates fatty acid oxidation. In addition, lipolysis-dependent generation of ligands for PPARs upregulate transcription of lipogenic enzymes 35 . In adipocyte-specific ATGL-deficient mice, a model of reduced adipocyte lipolysis, fatty acid oxidation and lipogenesis were also impaired 36 . Treatment with dapagliflozin exhibited similar metabolic changes in adipose tissues with these lipolysis-modified models.
Previous report showed that expressions of lipogenic genes in liver of HFD-induced obese diabetic models were reduced by the treatment with tofogliflozin or empagliflozin 37,38 . In these reports, both body weight and liver weight were reduced in pair-feeding conditions against control mice. In another report, expressions of lipogenic genes in liver of amylin NASH models were reduced by the treatment with ipragliflozin without changes in body weight 39 . In the current study, dapagliflozin did not reduce body weight probably because of excess calorie intake by hyperphagia. In this condition, liver weight was significantly decreased in KKAy + Dapa compared with KKAy. Moreover, we revealed for the first time that expressions of lipogenic genes were significantly reduced in KKAy mice treated with dapagliflozin without pair-feeding. There are two explanations for the reduced lipogenesis in liver by dapagliflozin. The first is the direct effect of dapagliflozin against hepatocytes. In palmitic acid-induced HepG2 cells, treatment with dapagliflozin significantly attenuates the mRNA expression of fasn and acaca 40 . The second is the effect of lipid mobilizing factor, Azgp1 (zinc alpha 2-glycoprotein). Injection of Azgp1 expression plasmid reduces hepatic TG content and expressions of lipogenic genes 24 . In the current study, expressions of Azgp1 were significantly elevated in liver of KKAy + Dapa compared with KKAy.
It was reported that tofogliflozin, another SGLT2 inhibitor, increases plasma adiponectin levels and reduces body weight in KKAy mice 41 . In our models, dapagliflozin tended to increase plasma adiponectin levels and upregulate adiponectin expression in adipose tissues, compared with KKAy albeit insignificantly, and did not induce significant reduction in body weight compared with KKAy, probably due to hyperphagia. Overfeeding enhances total intake of fat and amino acids (components of normal chow), and excess lipids or glutamate results in reduced adiponectin secretion 42,43 . Collectively, in our setup, overfeeding might have masked the effects of dapagliflozin on the enhancement of adiponectin expression.  after differentiation, each promoter/reporter construct was transfected into 3T3-L1 adipocytes, and next day, the media of 3T3-L1 cells were replaced with KRBB supplemented with or without 10 mM 3-HBAs or 1 μM pioglitazone. Luciferase activity was measured after 24-hr incubation. Luciferase values were normalized by an internal CMV-Renilla control and expressed as relative luciferase activity. Data are mean ± SEM (n = 3). (b) Adiponectin promoter region bisulfite sequencing analysis of 3T3-L1 adipocytes. On day 7 after differentiation, the media of 3T3-L1 cells were replaced with KRBB supplemented with 0 or 10 mM 3-HBA and incubated for 24 hr. Top panel; Each circle represents sequencing results of independent clones. Open circles: unmethylated CpGs, solid circles: methylated CpG. The CpG position relative to upstream transcription start site of mouse adiponectin gene is shown below each column. Bottom panel; Percentage of 5-methylcytosine. Data are mean ± SEM of three independent samples (n = 3). (c) ChIP-qPCR analysis of histone H3 tail at lysine 9 modifications on the adiponectin gene in 3T3-L1 adipocytes. On day 7 after differentiation, the media of 3T3-L1 cells were replaced with KRBB supplemented with 0 or 3 mM 3-HBA and incubated for 24 hr. The genomic DNA was precipitated by antibodies against β-hydroxybutyrylated histone H3 at lysine 9 (H3K9bhb), acetylated histone H3 at lysine 9 (H3K9ac), di-methylated histone H3 at lysine 9 (H3K9me2). ChIP signals of each region of adiponectin gene were detected by quantitative real-time PCR and normalized to input signal as relative to input (%). Data are mean ± SEM (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001. The present results are similar to those of Nagao et al. 43 who conducted metabolomic analysis of epididymal adipose tissue of C57BL/6 J and ob/ob mice, and showed that many metabolites of the TCA cycle, such as 2-oxoglutarate, succinate, fumarate and malate, were significantly higher in ob/ob mice than in C57BL/6J mice. With respect to the glycolytic pathway in periovarian WAT, amount of intermediate metabolites, such as G6P, F6P, and F1,6P tended to be higher in KKAy than KK. Microarray analysis indicated that expressions of insulin signaling molecules, including insr, irs1, and slc2a4 (Glut4), were significantly lower in KKAy than KK, and expression levels of gapdh, pgam1, Eno1, pdha1 and pdhb were significantly lower in KKAy than KK. Considered together, reduced expression of these distal glycolytic enzymes might be responsible for the accumulation of intermediate metabolites of glycolytic pathway in periovarian WAT of KKAy. In KKAy + Dapa, amount of these intermediate metabolites were normalized to the levels of KK. There were no differences in expressions of gapdh, pgam1, Eno1, pdha1 and pdhb between KKAy and KKAy + Dapa, however, PCX activity was significantly higher in periovarian WAT of KKAy + Dapa than that of KK, suggesting the increase of outflow from pyruvate to oxaloacetate. Upregulation of PCX activity might be partially responsible for the improvement of glycolytic pathway by dapagliflozin.
In this study, we first found that 3-HBA upregulated the mRNA expression level of adiponectin and intracellular adiponectin protein level in 3T3-L1 adipocytes. Adiponectin expression is regulated by both epigenetic modifications 28,29 and promoter activity 26,27 . We identified previously a functional PPAR-responsive element and a distal enhancer in adiponectin gene 26,27 . Other studies showed that obesity-induced pro-inflammatory cytokines promoted DNA hypermethylation of a particular region of the adiponectin promoter, and suppressed adiponectin expression through epigenetic regulation in 3T3-L1 adipocytes 28 . Furthermore, acetylation and methylation of histone H3K9 in the adiponectin gene were reported to be associated with adiponectin expression in 3T3-L1 adipocytes 29 . In the present study, while 3-HBA induced adiponectin expression in 3T3-L1 adipocyte, changes were neither noted in adiponectin promoter activity, DNA methylation in adiponectin promoter, nor acetylation or methylation of histone H3K9 on adiponectin gene. We focused on a new type of histone modification, lysine β-hydroxybutyrylation 18 . During prolonged fasting, histone is subject to modification by 3-HBA, which is an active gene mark 18 . In the present study, treatment with 3-HBA induced both adiponectin expression and H3K9bhb in adiponectin gene, independent of their acetylation or methylation. Considering that adiponectin has anti-inflammatory and anti-atherogenic properties 44 , 3-HBA might have protective roles through epigenetic modifications of adiponectin gene. Further study is required to reveal the existence of H3K9bhb in adipose tissues of diabetic patients treated with SGLT2 inhibitor.
In the present study, we identified the effect of dapagliflozin in KKAy mice. We suppose working hypothesis as follows (Supplementary Fig. S11). Ectopic expression of agouti protein in KKAy mice stimulates appetite 32,33 . Excess calorie intake due to hyperphagia results in reduction of adiponectin expression, induction of inflammatory cytokines, and inhibition of lipogenesis, lipolysis, and fatty acid oxidations in periovarian WAT. Treatment with dapagliflozin reduces plasma leptin levels, enhances further hyperphagia, whereas, it improves dysregulation of adipocytokines and lipid metabolism in periovarian WAT.
In summary, dapagliflozin improved glucose levels and induced 3-HBA in obese diabetic mice. 3-HBA can enhance adiponectin expression in adipocytes through direct modification of histone H3K9. Identification of the target genes of 3-HBA is important to understand the molecular mechanisms of SGLT2 inhibitors in protection against cardiovascular events.

Methods
Animal studies. 6-week-old female KK mice and KKAy mice were obtained from CLEA Japan (Tokyo, Japan) and acclimated for 2 week before the experiment. Mice were housed individually in sterile cages, maintained in a room under controlled temperature (23 ± 1.5 °C) and humidity (45 ± 15%) on a 12-h dark/12-h light cycle, and had free access to water and chow (MF; Oriental Yeast, Tokyo). Dapagliflozin (Cay-11574) was purchased from Cayman Chemical (Ann Arbor, MI). Dapagliflozin was dissolved in 100% ethanol at 125 mg /ml, followed by dilution with distilled water (160 μl of dapagliflozin mixture/1000 ml water, with final concentration of 0.02 mg/ml dapagliflozin). This concentration of ethanol (0.016% ethanol) did not affect amount of water intake, body weight and plasma 3-HBA concentrations (data not shown). Drinking water bottles were changed twice a week. Dapagliflozin treatment was conducted by adding 0.02 mg/ml dapagliflozin in drinking water. Three groups of mice were used; non-diabetic control female KK mice (KK), diabetic female KKAy mice (KKAy), and female KKAy mice treated with dapagliflozin (KKAy + Dapa). Food consumption was determined by weighing the metal cage top, including the food. Water intake was determined by measuring the fluid in drinking water bottles. Five weeks after the start of dapagliflozin treatment, the animal was fasted for 4 hours then anesthetized with isoflurane and blood samples were collected from the inferior vena cava, before sacrifice. Samples from seven tissues were collected, including subcutaneous fat, periovarian fat, mesenteric fat, BAT, liver, kidney, and skeletal muscle (gastrocnemius muscle). Gene expression microarray and metabolomic analyses were performed to compare the effects of dapagliflozin in periovarian adipose tissues. All mice studies were approved by the Ethics Review Committee for Animal Experimentation of Osaka University, Graduate School of Medicine, and carried out in accordance with the Institutional Animal Care and Use Committee Guidelines of Osaka University.

Microarray analysis.
Gene expression was profiled with Mouse Gene Expression 8 × 60 K Microarray (Agilent, Santa Clara, CA) by Takara Bio Inc (Tokyo). Cyanine-3 (labeled cRNA was prepared from 0.1 μg total RNA using the Low Input Quick Amp Labeling Kit (Agilent)). Fragmentation, hybridization, and washing steps were also carried out as recommended by the manufacturer (Agilent). Slides were scanned on the Agilent SureScan Microarray Scanner G2600D using one color scan setting for 8 × 60 K array slides. The scanned images were analyzed with Feature Extraction Software 11.5.1.1 (Agilent) using default parameters to obtain the intensity of the processed signal.

Metabolomic analysis.
Approximately 50 mg of frozen periovarian adipose tissue and liver was plunged into 1,500 µL of 50% acetonitrile/Milli-Q water containing internal standards (solution ID: 304-1002, Human Metabolome Technologies, Inc., Tsuruoka, Japan) at 0 °C in order to inactivate the tissue enzymes. The tissue was homogenized thrice at 1,500 rpm for 120 sec using a tissue homogenizer (Micro Smash MS100R, Tomy Digital Biology Co., Tokyo) and then the homogenate was centrifuged at 2,300 × g and 4 °C for 5 min. Subsequently, 800 µL of the upper aqueous layer was centrifugally filtered through a Millipore 5-kDa cutoff filter at 9,100 × g and 4 °C for 120 min to remove proteins. The filtrate was centrifugally concentrated and re-suspended in 50 µL of Milli-Q water for Capillary Electrophoresis-Mass Spectrometry (CE-MS) analysis. Metabolome measurements were carried out through a facility service at Human Metabolome Technologies.
Cell cultures and differentiation. 3T3-L1 adipocytes were cultured in DMEM (Nacalai Tesque, Kyoto, Japan) supplemented with 10% FBS until two days post-confluence. Then (day 0), the cells were treated with adipogenic mixture containing 0.5 mM 3-isobutyl-1-methylxanthine (Nacalai), 1 μM dexamethasone (Sigma-Aldrich, St. Louis, MO), and 1 μM insulin (Nacalai) for 48 hr. After 48 hr (day 2), medium was replaced with DMEM supplemented with 10% FBS. Until day 7, the cells were maintained in DMEM with 10% FBS. RNA extraction, cDNA synthesis, and quantitative real-time PCR. Total RNA was isolated using TRI-Reagent (Sigma) based on the method recommended by the manufacturer. The total amount of RNA was quantified by measurement of optical density (OD) at 260 nm. All RNA purity was determined by measuring the 260/280 and 260/230 ratios with Nanodrop instrument. The cDNA was synthesized using the PrimeScript RT Master Mix (Takara) according to the instructions supplied by the manufacturer. Real-time PCR was performed using THUNDERBIRD SYBR qPCR Mix (Toyobo, Osaka, Japan) on the LightCycler System (Roche, Basel, Switzerland) according to the protocol provided by the manufacturer. The sequences of primers used for real-time PCR are described in Supplementary Table 1.

Effects of 3-hydroxybutyric acid on
Western blot analysis. Cell lysates of 3T3-L1 cells were collected. Then, 4.3 μg of cell lysate protein was subjected to SDS-PAGE under reducing conditions, transferred, and blotted with HRP-conjugated anti-rat Adiponectin antibody (R&D Systems, Minneapolis, MN) followed by HRP-conjugated rat IgG antibody (Amersham Biosciences, Little Chalfont, UK). The membrane was stripped and reproved with anti-mouse β-actin Clone AC-15 (Sigma), followed by HRP-conjugated mouse IgG antibody (Amersham Biosciences).
Plasmids. pGL3 basic were purchased from Promega (Madison, WI). a10390Luc and a908Luc were reported previously or constructed in our laboratory, as described in detail elsewhere 26,27 . Luciferase reporter assay. On day 7 after the induction of differentiation as stated above, the medium of 3T3-L1 cells in 12-well plates was changed to Opti-MEM (Invitrogen, Carlsbad, CA), and the cells were transfected with luciferase reporter plasmids using Lipofectamine 3000 Transfection Reagent (Invitrogen) according to the protocol provided by the manufacturer. Transfection was performed using 2 ng CMV-Renilla (internal standard) and 100 ng reporter plasmids. The next day (day 8), the medium was changed to Krebs-Ringer Bicarbonate buffer, composed as stated above, with or without 10 mM 3-Hydroxybutyric acid (Sigma) or 1 μM pioglitazone (Takeda, Osaka, Japan). After 24-hr incubation, luciferase reporter assays were performed using Promega Dual-Luciferase Reporter Assay System (Promega). Luciferase values were normalized by an internal CMV-Renilla control and expressed as relative luciferase activity.
Statistical analysis. All data were expressed as mean ± SEM values. Differences between two groups were examined for statistical significance by the Student's t-test, and among three groups by one-way analysis of variance (ANOVA) followed by Tukey-Kramer test. A P value < 0.05 denoted the presence of a statistically significant difference. The JMP Pro 13.1.0 software (SAS Institute. Inc., Cary, NC, USA) was used in all statistical analyses.