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

Abstract

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.

Figure 1

Changes in body weight, food consumption, water intake, blood glucose level, and organ weight of the three mice groups. Three groups of mice are used; non-diabetic control female KK mice (KK), diabetic female KKAy mice (KKAy), and female KKAy mice treated with dapagliflozin (KKAy + Dapa). (a) Body weight, (b) food consumption, and (c) water intake were measured weekly for 5 weeks. (d) Fasting blood glucose levels were measured every two weeks. Data are mean ± SEM (n = 6). **p < 0.01, ***p < 0.001, KK versus KKAy. #p < 0.05, ##p < 0.01, ###p < 0.001, KKAy versus KKAy + Dapa, by one-way ANOVA followed by post hoc analysis (Tukey-Kramer test). (e) Weight of subcutaneous WAT (subWAT), periovarian WAT (ovaWAT), mesenteric WAT (mesWAT), BAT, liver, kidney, skeletal muscle after 5 weeks of dapagliflozin treatment. Data are mean ± SEM (n = 6). *p < 0.05, **p < 0.01, ***p < 0.001, by one-way ANOVA followed by post hoc analysis (Tukey-Kramer test).

Figure 2

Plasma levels of glycated HbA1c, insulin, NEFA, TG, adiponectin, and 3-HBA. (a) Glycated HbA1c (HbA1c), (b) plasma insulin, (c) plasma NEFA, (d) plasma TG, (e) plasma adiponectin, and (f) plasma 3-HBA levels were measured after 5 weeks of treatment with dapagliflozin. Data are mean ± SEM (n = 6). *p < 0.05, **p < 0.01, ***p < 0.001, by one-way ANOVA followed by post hoc analysis (Tukey-Kramer test).

Figure 3

Results of metabolomic and microarray analyses of periovarian WAT (ovaWAT). All analysis were performed after 5 weeks of dapagliflozin treatment. (a) Relative levels of metabolites and gene expressions associated with glycolytic and tricarboxylic cycle (TCA cycle) pathways. Relative levels of metabolites and gene expressions were surrounded by solid line and bottled line, respectively. (b,c) Relative expression levels of genes associated with adipocytokines (b) and the metabolism of ketone body (c). Data are normalized to the values of metabolites or gene expression levels of KK mice, and expressed as mean ± SEM (n = 4). *p < 0.05, **p < 0.01, ***p < 0.001, by one-way ANOVA followed by post hoc analysis (Tukey-Kramer test). G6P, glucose 6-phosphate; F6P, fructose 6-phosphate; F1,6P, fructose 1,6-diphosphate; GAP, glyceraldehyde 3-phosphate; 1,3-DPG, 1,3-di phosphoglycerate; 3-PG, 3-phosphoglycerate; 2-PG, 2-phosphoglycerate; PEP, phosphoenolpyruvate; AcCoA, acetyl CoA; OA, oxaloacetate; cis-Aco, cis-aconitate; 2-OG, 2-oxoglutarate; SucCoA, succinyl-CoA; 3-HBA, 3-hydroxybutyrate; Hk2, hexokinase 2; Gpi1, glucose phosphate isomerase 1; Pfkp, phosphofructokinase, platelet; Aldoc, aldolase C; Gapdh, glyceraldehyde-3-phosphate dehydrogenase; Pgk1, phosphoglycerate kinase-1; Pgam1, phosphoglycerate mutase 1; Eno1, enolase 1; Pkm, pyruvate kinase, muscle; Pdha1, pyruvate dehydrogenase E1 alpha 1; Pdhb, pyruvate dehydrogenase (lipoamide) beta; PPAR-γ, peroxisome proliferator activated receptor gamma; MCP-1, Monocyte chemoattractant protein-1; PAI-1, plasminogen activator inhibitor-1; IL-6, interleukin 6; TNF-α, tumor necrosis factor alpha; MCT1, monocarboxylic acid transporters 1; MCT2, monocarboxylic acid transporters 2; MCT4, monocarboxylic acid transporters 4; SCOT, succinyl-CoA-3-oxaloacid CoA transferase.

Figure 4

Effects of 3-HBA on mRNA expression levels and intracellular protein levels in 3T3-L1 adipocytes. On day 7 after differentiation, the medium of 3T3-L1 cells were replaced with KRBB supplemented with 0 or 10 mM 3-HBA and incubated for 24 hour. (a,f) The relative mRNA expression levels of adiponectin (a), PPAR-γ (b), MCP-1 (c), PAI-1 (d), IL-6 (e), and TNF-α (f) in 3T3-L1 adipocytes were measured by quantitative real-time PCR. Data are normalized to the level of 36B4 mRNA, and expressed as mean ± SEM (n = 6). (g) Intracellular protein levels of adiponectin and β-Actin in 3T3-L1 adipocytes using western blot analysis. Left panel; Representative western blot analysis. Right panel; Quantitative analysis of adiponectin contents in the left panel. Data are mean ± SEM (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 5

Effects of 3-HBA on adiponectin gene in 3T3-L1 adipocytes. (a) Promoter analysis of adiponectin gene in 3T3-L1 adipocytes. A series of fragments of the 5′-flanking region of the human adiponectin gene were subcloned upstream of the luciferase reporter gene as described in Materials and Methods. On day 7 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.