Retinol saturase coordinates liver metabolism by regulating ChREBP activity

Abstract

The liver integrates multiple metabolic pathways to warrant systemic energy homeostasis. An excessive lipogenic flux due to chronic dietary stimulation contributes to the development of hepatic steatosis, dyslipidemia and hyperglycemia. Here we show that the oxidoreductase retinol saturase (RetSat) is involved in the development of fatty liver. Hepatic RetSat expression correlates with steatosis and serum triglycerides (TGs) in humans. Liver-specific depletion of RetSat in dietary obese mice lowers hepatic and circulating TGs and normalizes hyperglycemia. Mechanistically, RetSat depletion reduces the activity of carbohydrate response element binding protein (ChREBP), a cellular hexose-phosphate sensor and inducer of lipogenesis. Defects upon RetSat depletion are rescued by ectopic expression of ChREBP but not by its putative enzymatic product 13,14-dihydroretinol, suggesting that RetSat affects hepatic glucose sensing independent of retinol conversion. Thus, RetSat is a critical regulator of liver metabolism functioning upstream of ChREBP. Pharmacological inhibition of liver RetSat may represent a therapeutic approach for steatosis.Fatty liver is one of the major features of metabolic syndrome and its development is associated with deregulation of systemic lipid and glucose homeostasis. Here Heidenreich et al. show that retinol saturase is implicated in hepatic lipid metabolism by regulating the activity of the transcription factor ChREBP.

Fig. 1

RetSat controls glycolytic and lipogenic pathways in primary hepatocytes. a RetSat protein expression in metabolically relevant mouse tissues was analyzed by immunoblotting. b RetSat protein expression, 48 h after transfecting Control or RetSat siRNA, was determined by immunoblotting. c Top five enriched pathways of genes that are regulated by RetSat depletion in hepatocytes were identified by DAVID functional annotation of Affymetrix gene expression profiles. d mRNA expression of selected genes in siControl or siRetSat hepatocytes was determined by qPCR. Data are shown as mean ± s.d., n = 6 independent transfections of hepatocyte cultures from two mice; *P < 0.05 by two-tailed t-test. An independent experiments yielded similar results. e Hepatocytes depleted of RetSat were incubated with ¹³C-glucose for 5 min and ¹³C-pyruvate labelling analyzed by mass spectroscopy. Data are shown as mean ± s.d., n = 4 independent transfections of hepatocyte cultures from the same mouse; *P < 0.05 by two-tailed t-test. An independent experiment yielded similar results. f Incorporation of ¹⁴C-acetate into extractable lipids was assessed in hepatocytes treated as indicated. Data are shown as mean ± s.d., n = 6 independent transfections of hepatocyte cultures from the same mouse; significance between siControl and siRetSat was tested by two-tailed t-test and *P < 0.05. An independent experiment yielded similar results. g Palmitate/palmitoleate ratio in siControl- or siRetSat-treated hepatocytes was assessed by mass spectroscopy. Data are shown as mean ± s.d., n = 4 independent transfections of hepatocyte cultures from the same mouse; *P < 0.05 by two-tailed t-test. An independent experiment yielded similar results

Fig. 2

Hepatic RetSat expression is increased in obese mice and regulates lipid metabolism. a RetSat mRNA expression in livers of lean, NC-fed (n = 5) and obese, HS/HFD-fed (n = 13) mice was determined by qPCR. *P < 0.05 between groups by two-tailed t-test. b Mice were injected with adenoviruses expressing shRNA targeting βGal or RetSat. Six days later, liver protein was analyzed for RetSat protein by immunoblotting. c–j HS/HFD-fed mice were treated as described in b, c liver stained for TGs by Oil Red-O, scale bars = 200 µM. d Liver TGs in mice were determined biochemically. Data are shown as mean ± s.e.m., n = 12 (shβGal), 13 (shRetSat); *P < 0.05 by two-tailed t-test. An independent experiment yielded similar results. e Hepatic mRNA expression of genes involved in lipid metabolism was determined by qPCR. Data are shown as mean ± s.e.m., n = 7 (shβGal), 6 (shRetSat); *P < 0.05 by two-tailed t-test. An independent experiment yielded similar results. f Newly synthesized palmitate in liver was assessed by determining the incorporation of deuterated water. Data are shown as mean ± s.e.m., n = 5 (shβGal), 5 (shRetSat); P value was determined by two-tailed t-test. g Food intake was measured for two 24 h periods. Data are shown as mean ± s.e.m., n = 6 (shβGal), 5 (shRetSat); *P < 0.05 by two-tailed t test. h–j: serum TGs, NEFAs, and blood glucose in ad libitum-fed or 24 h-fasted mice, determined 6 days after virus injection. Data are shown as mean ± s.e.m., n = 13 (shβGal), 12 (shRetSat); *P < 0.05 between both groups by two-tailed t-test. An independent experiment yielded similar results. k Serum insulin in ad libitum-fed mice was determined by ELISA. Data are shown as mean ± s.e.m., n = 9 (shβGal), 10 (shRetSat); *P < 0.05 by two-tailed t-test

Fig. 3

Hepatic RetSat expression in humans correlates with obesity and liver steatosis. RETSAT mRNA expression in human liver samples was determined by qPCR (n = 29) and correlated with a patient body mass index, b degree of hepatic steatosis, c HOMA-IR and d serum TG. e In a subset of 14 liver samples, MUFA content was determined and correlated with the expression of RETSAT. Normality of data was assessed by the Kolmogorov–Smirnov test and, depending on data distribution, significance was determined by Pearson a, c–e or Spearman b correlation coefficient

Fig. 4

RetSat depletion in mouse liver reduces protein levels and target gene expression of ChREBP. Mice fed HS/HFD were injected with adenoviruses expressing shRNA targeting βGal or RetSat. Six days later, livers were analyzed for a mRNA expression of the indicated transcription factors. b RetSat and ChREBP protein expression by immunoblotting, and c known ChREBP target genes by qPCR. In a, c, data are shown as mean ± s.e.m., n = 7 (shβGal), 6 (shRetSat); *P < 0.05 by two-tailed t-test. d RETSAT, PKLR and ACC1 mRNA expression in human liver samples was determined by qPCR (n = 29) and correlated. Normality of data were tested by the Kolmogorov–Smirnov test and significance was determined by the Pearson correlation coefficient

Fig. 5

RetSat controls ChREBP activity and glucose sensing in primary hepatocytes. Primary mouse hepatocytes were treated with Control or RetSat siRNA for 48 h, (a, left) ChREBP mRNA expression determined by qPCR, and RetSat and ChREBP protein levels determined by immunoblotting (a, right). left, Data are shown as mean ± 11s.d., n = 3 independent transfections of hepatocyte cultures from the same mouse. Two independent experiments yielded similar results. b Hepatocytes were treated as described in a and mRNA expression of a selection of known ChREBP target genes visualized in a heatmap. c Primary hepatocytes were depleted of RetSat using two siRNA’s targeting different sites of the RetSat transcript for 48 h, and expression of the indicated genes analyzed by qPCR. Data are shown as mean ± s.d., n = 6 independent transfections of hepatocyte cultures from two different mice; *P < 0.05 between siControl und siRetSat by one-way ANOVA with Bonferroni post test. An independent experiment yielded similar results. d Hepatocytes treated with Control or RetSat siRNA were transfected with a ChoRE-Luc reporter, exposed to low and high glucose concentrations as indicated, and analyzed for luciferase activity. e Hepatocytes treated with Control or RetSat siRNA were exposed to low and high glucose concentrations as indicated, and mRNA expression determined by qPCR. In d, e, data are shown as mean ± s.d., n = 6 independent transfections of hepatocyte cultures from two mice; two-way ANOVA with Bonferroni post test revealed significances between low and high glucose concentrations (^# P < 0.05) and between siControl and siRetSat (*P < 0.05). An independent experiment yielded similar results

Fig. 6

RetSat depletion prevents the glucose-induced nuclear accumulation of ChREBP independent of 13,14-dihydroretinol generation. a Primary hepatocytes were seeded on cover slips in 2.5 mM glucose. 24 h later, hepatocytes were treated with Control or RetSat siRNA overnight. The next day hepatocytes were exposed to 2.5 or 25 mM glucose and insulin as indicated for 24 h. After fixation, endogenous ChREBP was stained by immunocytochemistry and its localization determined by confocal microscopy. Nuclei were stained using DAPI, scale bars = 20 µm. Representative images of three independent experiments are shown. b Quantification of nuclear staining. Data are shown as mean ± s.d., n = 6 random optical fields of averaged nuclei intensities (total of >15 nuclei for each condition) from hepatocyte cultures from the same mouse; two-way ANOVA with Bonferroni post test revealed significances between low and high glucose (^# P < 0.05) and between siControl and siRetSat (*P < 0.05). c Hepatocytes were treated with Control or RetSat siRNA overnight. The next morning, cells were incubated with vehicle (DMSO) or 1 µM 13,14-dhretinol for 24 h at the indicated glucose concentrations and mRNA expression of Txnip determined by qPCR. Data are shown as mean ± s.d., n = 4. Two-way ANOVA with Bonferroni post test revealed significances between low and high glucose (^# P < 0.05) and between siControl and siRetSat (*P < 0.05), treatment with 13,14-dhretinol had no effect. d Incorporation of ¹⁴C-acetate into extractable lipids was assessed in hepatocytes depleted of RetSat for 48 h and supplemented with 13,14-dhretinol for the final 24 h. Data are shown as mean ± s.d., n = 4 independent transfections of hepatocyte cultures from the same mouse; *P < 0.05 between siControl und siRetSat by one-way ANOVA with Bonferroni post test. e, f Primary hepatocytes were treated with Control or RetSat siRNA and adenoviruses expressing GFP or a GFP-ChREBP fusion protein. Forty-eight hours after transfection/infection, e mRNA expression of RetSat and ChREBP and f ChREBP target genes were analyzed by qPCR. Data are shown as mean ± s.d., n = 6 independent transfections/infections of hepatocyte cultures from two mice; two-way ANOVA with Bonferroni post test showed significances between GFP and ChREBP (^# P < 0.05) and between siControl and siRetSat (*P < 0.05). n.s., not significant