C3G reduced dyslipidemia and hyperglycemia in HFD-fed mice
First, we performed mouse feeding studies with oral administration of C3G to mice fed a high-fat diet (HFD). The oral administration of C3G for 8 weeks significantly reduced plasma and hepatic triglyceride (TG) concentrations (P < 0.05 and P < 0.01, respectively); C3G reduced the number of intracellular lipid droplets, and F4/80 levels in C3G-HFD livers, suggesting that C3G improves non-alcoholic fatty liver disease (NAFLD) (Fig. 1a, b). Intracellular TG concentrations were also reduced in HepG2 cells stimulated with C3G compared to control HepG2 cells (Fig. 1c). GW7647, a PPARα agonist, reduced intracellular triglyceride levels and fatty acid oxidation rate but increased fatty acid synthesis rate in lipid-loaded HepG2 cells. Total cholesterol, low-density lipoprotein (LDL)-cholesterol, and high-density lipoprotein (HDL)-cholesterol concentrations were unaltered by C3G administration in mice (Supplementary Fig. 1a). By analysing hepatic fatty acid metabolic rates, C3G was shown to stimulate the rate of fatty acid oxidation but suppress the rate of fatty acid synthesis in both HepG2 cells and mouse livers (Fig. 1d, e). The oral administration of C3G had an antiobesogenic effect on mice, as indicated by reduced body weight, visceral fat tissue weight, adipocyte size, the ratio of white-to-brown adipose tissues and the increased ratio of white adipose tissue-to-skeletal muscle weight in HFD mice administered C3G (Fig. 1f, g, Supplementary Fig. 1b, and Supplementary Table 1). Aortic atherosclerotic plaque formation was decreased in mice administered C3G with a HFD (Supplementary Fig. 1c).
The oral administration of C3G improved key parameters indicating insulin sensitivity in mice. C3G decreased fasting plasma glucose and insulin concentrations and increased adiponectin concentrations without affecting liver glycogen and FGF21 levels (Fig. 2a–c). C3G also improved glucose and insulin tolerance, HOMA-IR, and the insulin sensitivity index in mice fed a HFD (Fig. 2d–f). To investigate glucose uptake upon treatment with C3G, we induced insulin resistance in cultured C2C12 myotubes and HepG2 cells by treating them with 400 µM lipids for 24 h and then measured their glucose uptake. Insulin promoted glucose uptake in C2C12 myotubes and HepG2 cells that were not lipid-loaded. C3G improved the impaired glucose uptake in lipid-loaded C2C12 myoblasts and HepG2 cells (Fig. 2g). These results collectively demonstrate that C3G improved NAFLD, adiposity, glucose tolerance, hyperglycaemia, and insulin sensitivity in mice fed a HFD.
C3G alters the hepatic fatty acid oxidation metabolites
We next examined the hepatic metabolomes of fasting mice by CE-MS and GC-TOF-MS and by assay kits. HFD-fed mice were orally administered either vehicle (saline) or C3G for 8 weeks, and the liver metabolome was analysed in mouse livers in a fasting state. CE-MS and GC-TOF-MS analyses identified 92 metabolites (selected key metabolites are shown in Table 1 and Supplementary Table 1). Multivariate statistical analysis was conducted to determine the significance of any differences in the metabolomes of liver tissues from mice fed different diets for 8 weeks. By using partial least squares regression (PLSR) analysis, the metabolite profiles of each group were distinctively clustered. These findings suggest that C3G caused unique and specific alterations in the mouse liver metabolome (Fig. 3a).
C3G reduced fasting plasma glucose concentrations (Fig. 2a) but increased the levels of metabolites in the first phase of glycolysis including glucose 6-phosphate and fructose 6-phosphate (Table 1). These results suggest that C3G elevated hepatic glucose uptake by metabolic trapping. C3G indeed increased glucose uptake in HepG2 cell and C2C12 myotubes (Fig. 2g). Although fructose 1,6-bisphosphate, a product of phosphofructokinase-1, was unaltered, the levels of metabolites in the second phase of glycolysis including pyruvic acid and lactic acid levels were substantially reduced in the C3G group (Table 1), which suggest decreased rate of glycolysis by C3G was independent of phosphofructokinase-1 activity. The 6-phosphoglucuronate levels were increased in the liver in the C3G group (Table 1), and hepatic glycogen concentrations were not different between the control and C3G groups (Fig. 2c). These results suggest that glucose 6-phosphate may be directed to the pentose phosphate pathway or other pathways than to glycolysis and glycogen synthesis in C3G fed livers.
Acetyl CoA levels were unaltered, but the level of citrate was dramatically reduced and nearly undetectable in the C3G group compared with the controls (Table 1). Citrate, the first metabolite in the citric acid cycle, is an activator of the anabolic pathway and an inhibitor of the catabolic pathway of fatty acid and glucose metabolism30. The concentration of malonyl-CoA, an inhibitor of mitochondrial fatty acid oxidation31, was reduced substantially following C3G administration (Fig. 3b). Furthermore, the level of β-hydroxybutyrate was substantially increased by C3G (Fig. 3b). By immunoblotting, pyruvate dehydrogenase kinase 4 (PDK4) expression and PDK4-mediated phosphorylation of pyruvate dehydrogenase complex (PDH, pSer293E1α subunit) were increased by C3G administration (Fig. 3c). Glycerol 3-phosphate was increased in C3G-treated livers (Table 1). Glycerol 3-phosphate can be produced from dihydroxyacetone phosphate by glycerol 3-phosphate dehydrogenase or from the transported lipolyzed glycerol in adipose tissue32. These results suggest that C3G suppressed fatty acid synthesis but increased catabolic pathways including fatty acid oxidation, ketogenesis, and at least in part increased lipolysis in white adipose tissues.
The levels of carnitine and its related metabolites choline and betaine aldehyde were increased in the livers of C3G-treated mice compared with those of control mice (Table 1). Carnitine is critical in mitochondrial fatty acid transport and subsequent fatty acid oxidation; thus, increased levels in the blood after carnitine administration have been shown to reduce body weight and body mass index33. Thus, elevated levels of carnitine metabolites could contribute to increase fatty acid oxidation and body fat reduction. The levels of α-ketoglutarate, fumarate, and malic acid were reduced but succinate levels were increased in the C3G group compared with the control group (Table 1).
In addition, C3G increased the levels of amino acids, particularly, branched amino acids, compared with those in HFD-fed control mice (Supplementary Table 2). These findings indicate that C3G induces cellular protein degradation pathways, such as autophagy. Branched amino acids have diverse roles and it has been reported that branched chain amino acid could protect hepatic steatosis and NAFLD, thus increased branched chain amino acids may reinforce improvement of NAFLD in mice34.
The level of reduced glutathione (GSH) and the ratio of reduced GSH to oxidized GSSH were decreased in the livers of C3G-treated mice (Table 1), which implies that C3G increased antioxidant capacity in the liver. As above mentioned, C3G increased glucose 6-phosphate conversion to 6-phosphogluconate in pentose phosphate pathway. NADPH produced from pentose phosphate pathway did not induce but decreased fatty acid synthesis. These findings suggest that pentose phosphate pathway may be mildly increased, and the produced NADPH may be utilized to increase GSH levels antioxidant capacity in the liver.
The results of the metabolomic analysis collectively indicated that hepatic glucose uptake was increased by C3G and that conversion of glucose 6-phosphate to the pentose phosphate pathway or other pathways was increased. Most importantly, C3G stimulates catabolic energy metabolism including hepatic fatty acid oxidation, ketogenesis, and possibly white adipose lipolysis, which indicates a shifting energy metabolism towards the fat-consuming mode.
C3G binds and activates PPARs
C3G activates PPARs, as it has been reported that C3G induces PPAR gene expression and has hypolipidaemic and antiobesogenic effects35. C3G also ameliorated hypertriglyceridaemia and NAFLD and increased the rate of fatty acid oxidation (Fig. 1). Hepatic metabolomic analysis suggested induction of fatty acid oxidation and ketogenesis by C3G, with decreased malonyl-CoA and citrate levels (Table 1 and Fig. 3b). These phenotypic and metabolomic profiles suggested PPAR activation. Accordingly, we investigated whether C3G is a ligand of PPARs. First, the interaction of C3G with the ligand-binding domain (LBD) of PPARs was quantified with coactivator recruitment assays. GW7647, troglitazone, and GW0742 were PPARα, -γ, and -δ/β agonists and the agonist interacted with the PPAR isoforms, respectively. C3G was found to interact with the LBD of three PPAR isoforms and exhibited the highest binding affinity for PPARα. The EC50 values of C3G were 1.1, 10.8, and 31.1 µM for the PPARα, -γ, and -δ/β subtypes, respectively (Fig. 4a and Supplementary Table 3). Second, surface plasmon resonance experiments demonstrated that C3G directly interacted with three PPAR subtypes and exhibited the highest affinity for PPARα (Supplementary Fig. 2a and Supplementary Table 3). By qPCR analysis, the mRNA expression of PPARα and its target genes, Acox and Ucp2, was shown to be induced in the mouse livers of C3G group (Fig. 4b). Two additional dietary anthocyanins-pelargonidin—3-O-glucoside and delphinidin-3-O-glucoside—were also shown to interact directly with PPAR LBDs by surface plasmon resonance experiments (Supplementary Fig. 2b). These results demonstrate that agonistic PPARα and -γ activity is not limited to C3G but extends to other anthocyanins; thus, pelargonidin-3-O-glucoside and delphinidin-3-O-glucoside may exert hypolipidaemic and antiobesogenic effects similar to those of C3G.
In addition, the gene expression of PPARγ and its target genes are also induced in HepG2 and C2C12 myotubes (Supplementary Fig. 3). Previous pharmacokinetic studies reported that the Cmax of C3G is 0.14–14 µM11,36,37,38, suggesting that dietary C3G activates the PPARα and PPARγ subtypes in vivo. These results demonstrate that PPARα may be a major and direct target protein of C3G during its regulation of hyperlipidaemia and insulin resistance and it is also possible that C3G also activates PPARγ, which further contributes to the improvement of lipid and glucose metabolism.
Next, we performed feeding studies in PPARα-deficient mice orally administered C3G for 8 weeks. Reductions in plasma TG and fasting glucose concentrations were completely abrogated in PPARα-deficient mice (Fig. 5a). Changes in the rates of fatty acid synthesis and oxidation, body weight, fat mass, white-to-brown adipose tissue weight, mRNA expression levels of the PPARα responsive genes, Acox1 and Ucp2, malonyl-CoA, and ketone bodies in the liver were also nullified in PPARα-deficient mice (Fig. 5b–f and Supplementary Table 4). These results demonstrated that the hypolipidaemic, hypoglycaemic, and antiobesogenic effects of C3G are primarily dependent on PPARα activation.
C3G reduces adiposity with increases energy expenditure
Metabolome analysis revealed increased oxidative metabolism induced by C3G. C3G also reduced adiposity in HFD-fed mice, and the antiobesogenic mechanism of C3G was further investigated in mice. C3G administration for 8 weeks reduced visceral adipose tissue and adipocyte size, while the reduction in white adipose tissues induced by C3G was abolished in PPARα-deficient mice. Cells in the brown adipose tissue of HFD-fed mice administered C3G were smaller (Fig. 6a), and the mRNA expression levels of PPARα, PGC-1α and UCP1 in brown adipose tissue were higher; however, these C3G-induced changes were not observed in PPARα-deficient mice (Fig. 6b). Activation of the PPARα-PGC-1α-UCP1 signalling axis by C3G in brown adipose tissue increased oxygen consumption and energy expenditure and caused adiposity in vivo.
To investigate the effects of C3G on respiratory metabolism, indirect calorimetry was performed in wild-type mice orally administered C3G for 2 weeks. Compared with the vehicle control treatment, C3G increased oxygen consumption and energy expenditure, especially in the dark cycle (Fig. 6c). These results demonstrated that C3G reduces body fat accumulation via increased mitochondrial oxidative metabolism and thermogenesis in brown adipose tissue and energy expenditure via the activation of PPARs. The concentrations of malonyl-CoA and β-hydroxybutyrate in the liver were not different between PPARα-deficient mice and control mice (Fig. 5f). These results collectively demonstrate that C3G reduces adiposity in mice by inducing hepatic fatty acid oxidation and brown adipocyte thermogenesis in a PPARα-dependent manner. These results demonstrate that PPARα is a major target protein for C3G in the regulation of energy metabolism.