PGC-1α in hepatic UPR during high-fat high-fructose diet and exercise training in mice

Abstract

Diet-induced obesity is associated with hepatic steatosis, which has been linked with activation of the unfolded protein response (UPR). PGC-1α is a transcriptional coactivator involved in exercise training-induced adaptations in muscle and liver. Therefore, the aim of this study was to test the hypothesis that PGC-1α is required for exercise training-mediated prevention of diet-induced steatosis and UPR activation in liver. Male liver-specific PGC-1α knockout (LKO) and littermate floxed (lox/lox) mice were divided into two groups receiving either control diet (CON) or high-fat high-fructose diet (HFF). After 9 weeks, half of the HFF mice were treadmill exercise trained for 4 weeks (HFF+ExT), while the rest were kept sedentary. HFF resulted in increased body and liver weight, adiposity, hepatic steatosis and whole body glucose intolerance as well as decreased hepatic IRE1α phosphorylation. Exercise training prevented the HFF-induced weight gain and partially prevented increased liver weight, adiposity and glucose intolerance, but with no effect on liver triglycerides. In addition, BiP protein and CHOP mRNA content increased with exercise training compared with CON and HFF, respectively. Lack of PGC-1α in the liver only resulted in minor changes in the PERK pathway. In conclusion, this study provides evidence for dissociation between diet-induced hepatic triglyceride accumulation and hepatic UPR activation. In addition, PGC-1α was not required for maintenance of basal UPR in the liver and due to only minor exercise training effects on UPR further studies are needed to conclude on the potential role of PGC-1α in exercise training-induced adaptations in hepatic UPR.

Keywords: UPR; Exercise training; PGC-1α; high-fat diet; liver.

Figure 1

Body weight (A), percent body fat (B), blood glucose area under the curve (AUC) from GTT (C) and PTT (D) from liver‐specific PGC‐1α knockout (LKO) and littermate control (lox/lox) mice fed either a control diet (CON) or a high‐fat high‐fructose diet (HFF) for 13 weeks. Half of the HFF fed mice performed treadmill exercise training the last 4 weeks (HFF+ExT). GTT and PTT were performed 1 week before euthanization. AUC is calculated using the trapezoid method. Values are presented as means ± SE, n = 8–10. *Significantly different from CON within given genotype, P < 0.05. ^¤Significantly different from HFF within given genotype, P < 0.05. Parenthesis indicates a tendency for a difference, 0.05 ≤ P < 0.1.

Figure 2

Liver weight (A), triglyceride (B) and glycogen content (C) from liver‐specific PGC‐1α knockout (LKO) and littermate control (lox/lox) mice fed either a control diet (CON) or a high‐fat high‐fructose diet (HFF) for 13 weeks. Half of the HFF mice performed treadmill exercise training the last 4 weeks (HFF+ExT). Liver weight is normalized to total body weight. Values are presented as means ± SE, n = 9–10. *Significantly different from CON within given genotype, P < 0.05. ^¤Significantly different from HFF within given genotype, P < 0.05. Parenthesis indicates a tendency for a difference, 0.05 ≤ P < 0.1.

Figure 3

Hepatic PEPCK (A), FAS (B) and OXPHOS (C) protein content from liver‐specific PGC‐1α knockout (LKO) and littermate control (lox/lox) mice fed either a control diet (CON) or a high‐fat high‐fructose diet (HFF) for 13 weeks. Half of the HFF mice performed treadmill exercise training the last 4 weeks (HFF+ExT). Protein content is given in arbitrary units (AU). Values are presented as means ± SE, n = 8–10. *Significantly different from CON within given genotype, P < 0.05.

Figure 4

Hepatic BiP protein (A), cleaved ATF6 protein (B), IRE1α ^Ser724 phosphorylation (C), IRE1α protein (D), PERK^{T hr980} phosphorylation (E), PERK protein (F), eIF2α ^Ser51 phosphorylation (G) and eIF2α protein (H) content from liver‐specific PGC‐1α knockout (LKO) and littermate control (lox/lox) mice fed either a control diet (CON) or a high‐fat high‐fructose diet (HFF) for 13 weeks. Half of the HFF mice performed treadmill exercise training the last 4 weeks (HFF+ExT). Protein content is given in arbitrary units (AU). Values are presented as means ± SE, n = 7–10. *Significantly different from CON within given genotype, P < 0.05. ^¤Significantly different from HFF within given genotype, P < 0.05. ^#Significantly different from lox/lox within given group, P < 0.05. Parenthesis indicates a tendency for a difference, 0.05 ≤ P < 0.1.

Figure 5

Hepatic spliced (s)XBP1 (A) and CHOP mRNA (B) content in liver from liver‐specific PGC‐1α knockout (LKO) and littermate control (lox/lox) mice fed either a control diet (CON) or a high‐fat high‐fructose diet (HFF) for 13 weeks. Half of the HFF fed mice performed treadmill exercise training the last 4 weeks (HFF+ExT). mRNA is normalized to single‐stranded (ss) DNA. Values are presented as means ± SE, n = 9–10. *Significantly different from CON within given genotype, P < 0.05. ^¤Significantly different from HFF within given genotype, P < 0.05. Parenthesis indicates a tendency for a difference, 0.05 ≤ P < 0.1.

Figure 6

Representative western blots. Phosphoenolpyruvate carboxykinase (PEPCK) protein, fatty acid synthase (FAS) protein, OXPHOS protein, binding immunoglobulin protein (BiP) protein, cleaved activating factor 6 (ATF6) protein, inositol requiring enzyme‐1α (IRE1α) phosphorylation, IRE1α protein, protein kinase RNA‐like ER kinase (PERK) phosphorylation, PERK protein, eukaryotic translation initiation factor 2α (eIF2α) phosphorylation, eIF2α protein, c‐Jun N‐terminal kinase (JNK) phosphorylation, JNK protein and GAPDH protein content in liver from liver‐specific PGC‐1α knockout (LKO) and littermate control (lox/lox) mice fed either a control diet (CON) or a high‐fat high‐fructose diet (HFF) for 13 weeks. Half of the HFF fed mice performed treadmill exercise training the last 4 weeks (HFF+ExT). One sample from each experimental group (CON, HFF and HFF+ExT of both genotypes) is represented on each gel. Because samples from mice run acutely were run together with the samples used in the present manuscript, the relevant bands have been cut out of the gel pictures for the representative blots. However, the representative blots show samples run on the same gel.