β-catenin links hepatic metabolic zonation with lipid metabolism and diet-induced obesity in mice

Abstract

β-catenin regulates the establishment of hepatic metabolic zonation. To elucidate the functional significance of liver metabolic zonation in the chronically overfed state in vivo, we fed a high-fat diet (HFD) to hepatocyte-specific β-catenin transgenic (TG) and knockout (KO) mice. Chow-fed TG and KO mice had normal liver histologic findings and body weight. However, HFD-fed TG mice developed prominent perivenous steatosis with periportal sparing. In contrast, HFD-fed KO mice had increased lobular inflammation and hepatocyte apoptosis. HFD-fed TG mice rapidly developed diet-induced obesity and systemic insulin resistance, but KO mice were resistant to diet-induced obesity. However, β-catenin did not directly affect hepatic insulin signaling, suggesting that the metabolic effects of β-catenin occurred via a parallel pathway. Hepatic expression of key glycolytic and lipogenic genes was higher in HFD-fed TG and lower in KO mice compared with wild-type mice. KO mice also exhibited defective hepatic fatty acid oxidation and fasting ketogenesis. Hepatic levels of hypoxia inducible factor-1α, an oxygen-sensitive transcriptional regulator of glycolysis and a known β-catenin binding partner, were higher in HFD-fed TG and lower in KO mice. KO mice had attenuated perivenous hypoxia, suggesting disruption of the normal sinusoidal oxygen gradient, a major determinant of liver carbohydrate and liver metabolism. Canonical Wnt signaling in hepatocytes is essential for the development of diet-induced fatty liver and obesity.

Figure 1

High-fat diet (HFD)–fed transgenic (TG) mice have a predominantly perivenous pattern of liver steatosis. Mice were fed the HFD or chow diets for 8 weeks. For tissue collection, mice were fasted for 4 hours from 7 AM and sacrificed at 11 AM. A: Gross appearance of the liver from chow and HFD-fed TG and wild-type (WT) mice. B: Liver/body weight ratio. C: Liver sections stained with hematoxylin-eosin and Oil Red O. Note the strongly perivenous pattern of steatosis with periportal sparing in TG/HFD livers. D: Magnified view of hepatocytes from HFD-fed WT and TG liver. E: Liver triglyceride and total cholesterol levels. Data are expressed as means ± SEM of five mice per group (chow) and seven mice per group (HFD). ^∗P < 0.05, ^∗∗P < 0.01. Scale bar = 100 μm (C). Original magnification: ×100 (C, top row); ×200 (C, remaining rows, and D). C, central vein; P, portal area.

Figure 2

High-fat diet (HFD)–fed knockout (KO) mice exhibit periportal steatosis, lobular inflammation, and hepatocyte apoptosis. Mice were fed the HFD or chow diets for 4 weeks. Mice were fasted for 4 hours from 7 AM and sacrificed at 11 AM for tissue collection. A: Hematoxylin and eosin (H&E)–stained liver sections from chow- or HFD-fed wild-type (WT) and KO mice. B: Low power view of H&E-stained liver sections. Note the foci of inflammatory cells in HFD-fed KO liver. C: Gross liver morphologic findings in mice on chow and HFD diets and quantitation of liver/body ratio. D: Liver triglyceride and total cholesterol levels at 4 weeks. Data are presented as means ± SEM of five mice per group. E: Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling staining reveals apoptotic cells in livers from HFD-fed KO mice (black arrowheads). F: Western blot analysis for activated caspase 3 (longer exposure emphasizes differences in C3 expression). Glyceraldehyde-3-phosphate dehydrogenase is the internal loading control (shorter exposure). Scale bar = 100 μm (A and B). Original magnification: ×200 (A); ×100 (B); ×400 (E). ^∗P < 0.05. Neg, negative control; Pos, positive control (spleen sections).

Figure 3

Zonal expression patterns of β-catenin and its target genes on mice fed the high-fat diet (HFD). Mice were fed the HFD or chow diets for 4 weeks. For tissue collection, mice were fasted for 4 hours from 7 AM and sacrificed at 11 AM. A: Liver sections from chow- or HFD-fed knockout (KO), wild-type (WT), and transgenic (TG) mice stained with β-catenin, Cyp2E1, or GS as indicated. Hepatocytes with cytoplasmic and nuclear β-catenin staining on the HFD are marked with white arrowheads. Note the absence of Cyp2E1 and GS staining in KO mice and the perivenous zonal pattern of expression of both genes in WT and TG mice. Also note the correlation between Cyp2E1 staining and hepatocyte steatosis in HFD-fed TG mice.

Figure 4

Hepatic β-catenin regulates systemic energy homeostasis. A: Body weight change, morphologic findings, and perigonadal fat pads from chow- and high-fat diet (HFD)–fed transgenic (TG) and wild-type (WT) mice. Mice were fed chow or HFD for 8 weeks and fasted for 4 hours before sacrifice for tissue collection at 11 AM. B: Body weight change, morphologic findings, and perigonadal fat pads from chow and HFD-fed WT and knockout (KO) mice. Mice were fed chow or HFD for 4 weeks and fasted for 4 hours before sacrifice for tissue collection at 11 AM. C: Respiratory exchange ratio on the HFD. D: Activity level on the HFD expressed as the total of X (horizontal) and Z (vertical) movements measured in the metabolic cage. E: Food intake in HFD-fed WT, TG, and KO mice measured during the metabolic cage experiment. F: Feces dry weight on the HFD during 24 hours. G: Protein immunoblot analysis for β-catenin using lysates prepared from various organs from WT and KO mice. Glyceraldehyde-3-phosphate dehydrogenase was used as a loading control. n = 5 to 8 per group (A); n = 5 to 7 per group (B); n = 5 per group (C–F). ^∗P < 0.05 (A–D and F). Pancr, pancreas; VAT, visceral adipose tissue; Vastus, vastus muscle.

Figure 5

β-catenin indirectly regulates hepatic insulin signaling. A: Fasting or random serum glucose from high-fat diet (HFD)–fed wild-type (WT) and transgenic (TG) mice and serum insulin levels from chow or HFD-fed WT and TG mice collected after 4 hours of fasting. B: I.P. glucose tolerance tests (GTTs) in chow and HFD-fed WT and TG mice. C: Hyperinsulinemic-euglycemic clamp studies with chow-fed WT and TG mice revealing glucose infusion rate, blood glucose levels during the clamp, and hepatic glucose production. D: Hepatic glucose output measured during the insulin clamp study. E: Western blot analysis for proteins in the insulin-signaling pathway. Chow-fed, 16-hour overnight fasted mice were injected i.p. with saline (minus sign) or insulin (plus sign) and livers were harvested after 20 minutes. n = 4 to 5 per group (A), n = 5 per group (B–D). ^∗P < 0.05, ^∗∗∗P < 0.001.

Figure 6

β-catenin regulates hepatic fatty acid oxidation, ketogenesis, and mitochondrial function. A: Palmitate oxidation rate in liver sections from overnight fasted, chow-fed mice. B: Serum β-hydroxybutyrate (BHB) levels from fasted (white bars) and refed (black bars) chow-fed mice. C: Citrate synthase activity in isolated mitochondria from freshly harvested livers of chow-fed wild-type (WT) and knockout (KO) mice. D: Relative hepatic mitochondrial DNA (mtDNA) content of chow- (white bars) and high-fat diet (HFD)–fed mice at 4 weeks (gray bars) and 8 weeks (black bars). mtDNA content was determined as the ratio of the mitochondrial gene COXI and the nuclear gene 18S rRNA measured by real-time PCR. E: Western blot analysis for cytochrome c oxidase (COX) IV and peroxisome proliferator-activated receptor gamma coactivator (PGC)-1α in livers of WT, KO, and transgenic (TG) mice on chow or HFD. Tubulin is the loading control. Mice were fed the chow or HFD diets for 4 weeks and fasted for 4 hours before sacrifice for tissue collection at 11 AM. F: Quantification of Western blot results from E. Results are expressed as means ± SEM of four samples per group normalized to the internal loading control and expressed as relative value compared with the WT/chow group. n = 5 per group (A and B), n = 4 to 5 per group (D). ^∗P < 0.05.

Figure 7

β-catenin regulates expression of glycolytic and lipogenic genes in mice on the high-fat diet (HFD). A: Immunoblots for hepatic fatty acid synthase (FAS), glucokinase (GCK), and peroxisome proliferator-activated receptor (PPAR)-γ from chow or HFD-fed wild-type (WT) and knockout (KO) mice (at 4 weeks) or WT and transgenic (TG) mice (at 8 weeks). B: Real-time PCR assays for steady-state mRNA levels of important metabolic genes as indicated: glycolysis (Gck and Pklr), lipogenesis (Fasn and Scd1), lipid droplet (Plin2), gluconeogenesis (Pck1 and G6pc), ketogenesis (Hmgcs2), fatty acid oxidation (Acadm and Acox1), and regulators (Hnf4a and Srebf1). Results are from chow-fed overnight fasted (white bars) and 6-hour refed (black bars) mice. C: Immunoblot for hepatic GCK from chow-fed fasted and refed mice. α-Tubulin was the internal loading control in A and C. D: Chromatin immunoprecipitation analysis in Hep3B cells for binding of transcription factor (TCF)-4 to the 2-kb liver-specific Gck promoter. SP5 is shown as positive and actin as negative controls. For fasting-refeeding experiments, mice were fasted overnight for 16 hours and refed for 6 hours starting at 7 AM. Fasted mice were sacrificed at 7 AM, and refed mice were sacrificed for tissue collection at 1 PM. n = 5 per group (B). ^∗P < 0.05.

Figure 8

β-catenin regulates zonal oxygen gradients and affects hypoxia-inducible factor (HIF)-1α protein levels. A: Immunoblot for hepatic HIF-1α from chow- and high-fat diet (HFD)–fed mice and its quantification. α-Tubulin was the internal loading control. Results are expressed as means ± SEM of four samples per group normalized to the internal loading control and expressed as relative value compared with the WT/chow group. B: Pimonidazole staining (brown) for tissue hypoxia in livers of chow-fed wild-type (WT) and knockout (KO) mice exposed to room air (Air; 21% oxygen), intermittent hypoxia (IH; 60 cycles hour⁻¹ of 21% oxygen alternating with 6% oxygen for 30 seconds each), and continuous hypoxia (CH; 10% oxygen). Mice were not fasted, and pimonidazole injection was performed at 11 AM followed by sacrifice for tissue collection 1 hour later. C: Hypothetical model reveals the link between β-catenin and hepatocyte bioenergetics, metabolic zonation, and lipid homeostasis. Metabolic functions, such as oxidative phosphorylation and gluconeogenesis, predominate in the oxygen-rich periportal zone, whereas glycolysis and lipogenesis predominate in the relatively hypoxic perivenous zone. Fasting expands the periportal functions, and feeding expands the perivenous functions, conferring metabolic plasticity to the normal liver. β-catenin regulates mitochondrial and microsomal function in both periportal and perivenous hepatocytes, helps to establish the sinusoidal oxygen gradient, and promotes glycolysis and linked lipogenesis in the hypoxic perivenous zone. β-catenin disruption in KO mice causes defects in mitochondrial functions (ketogenesis, gluconeogenesis, and fatty acid oxidation) and perivenous microsomal functions (eg, Cyp2E1) with decreased oxygen use, blunting of the sinusoidal oxygen gradient, and down-regulation of HIF-1α–mediated glycolysis and lipogenesis in the perivenous zone. Additional factors, such as direct or indirect regulation of sinusoidal blood flow (eg, via Hmox1), may also play a role. Decreased hepatic lipogenesis contributes to resistance to diet-induced obesity and systemic insulin sensitivity in KO mice but increases susceptibility of the liver to lipotoxic effects. Transgenic (TG) mice have the opposite phenotype with increased hepatic glycolysis and lipogenesis, leading to diet-induced obesity and systemic insulin resistance. n = 4 mice per group (B). ^∗P < 0.05.