Transcriptional Regulation of the Angptl8 Gene by Hepatocyte Nuclear Factor-1 in the Murine Liver

Figure 1. The expression of hepatic Angptl8 mRNA and protein were increased by refeeding after fasting. (a) After a 12 h fast, 20-week-old male mice were re-fed ad libitum and euthanised after 60 and 240 min. Total RNA of the liver was isolated and subjected to qPCR to quantify Angptl8 mRNA levels. Data represent mean ± SEM from 5 mice at each time point from triplicate PCR samples. The experiment was repeated once with similar results. Asterisks indicate significant differences from fasting Angptl8 expression levels (***p < 0.001). (b) Whole cell lysates of the liver were prepared from three different mice at each time point and subjected to immunoblotting using anti-Angptl8 and anti-cyclophilin antibodies. The Angptl8 and the cyclophilin blots were conducted by re-probing the same parts of the same membrane.

Figure 2. Promoter deletion analyses of mouse Angptl8 gene in liver- or non-liver-derived cell lines and mouse primary hepatocytes. (a) Schematic representation of mouse Angptl8 gene promoter constructs. The serially deleted promoter regions (−2291/+101, −1508/+101, −966/+101, −309/+101, and −59/+101) were ligated to the pGL4.10 luciferase reporter vector (Luc). The position of the consensus-binding motif for hepatocyte nuclear factor-1 (HNF-1) at −84/−68 (GGTTAACCATTGACCAG) is indicated as oblique boxes. These promoter constructs were transfected to mouse hepatoma-derived Hepa1–6 cells (b), human hepatoma-derived HepG2 cells (c), mouse primary hepatocytes (d) or HeLa cells originated from human uterine cervical cancer (e). The luciferase activity of the −59/+101 construct was set as 1 and data represent mean ± SEM from three independent experiments with triplicate determinations. Asterisks indicate significant differences from the activity of −59/+101 construct (*p < 0.05, **p < 0.01 or ***p < 0.001).

Figure 3. (a–d) Mutations of the HNF-1 binding site at −84/−68 abrogated the basal Angptl8 gene promoter activities in hepatocytes. (a) Point mutations (indicated as lower case letters) or 12-base pair deletion in the HNF-1 binding site at −84/−68 were introduced into the −2291/+101 construct (see Materials and Methods) and transfected into Hepa1–6 cells (b), HepG2 cells (c) or mouse primary hepatocytes (d). Promoter activities are expressed as luciferase activities relative to those of the wildtype −2291/+101 construct, which was assigned a value of 1. Data represent mean ± SEM from three independent transfections with triplicate determinations. Asterisks indicate significant differences from the wildtype −2291/+101 activities (**p < 0.01 or ***p < 0.001). (e) Basal promoter activity of the mouse Angptl8 gene requires co-transfected HNF-1 in non-hepatocytes. The −309/+101 reporter construct with the wildtype HNF-1 binding motif at −84/−68 or −59/+101 construct lacking the HNF-1 binding site was co-transfected with HNF-1α and/or β expression vector into HeLa cells. A CMV10 expression vector was co-transfected as a control. Promoter activities are expressed as luciferase activities relative…

Figure 4. Knockdown of endogenous Hnf-1 reduced Angptl8 protein expression levels in Hepa1–6 cells (a) and mouse primary hepatocytes (b). At 24 h after transfection of both siHnf-1α and siHnf-1β or control siRNA (siCtr), whole cell lysates were prepared and subjected to SDS-PAGE. Immunoblotting was performed using anti-Angptl8, anti-cyclophilin and anti-β-actin antibodies. Increasing amounts of transfected siHnf-1α and siHnf-1β reduced Angptl8 protein levels in a dose-dependent manner in both Hepa1–6 cells and primary hepatocytes. The blot of Hnf-1 and the blot of Angptl8 were cropped from different parts of the same membrane. The blot of cyclophilin and β-actin was cropped from a different membrane but using the same whole cell lysates used for anti-Hnf-1 and anti-Angptl8 blotting.

Figure 5. Binding of Hnf-1 to the −84/−68 region of the Angptl8 gene. (a) Electrophoretic mobility-shift assays (EMSA) showed that Hnf-1 in whole cell extracts from Hepa1–6 cells significantly bound to the wildtype probe of the Angptl8 promoter lesion –94/−60 that included the Hnf-1 binding consensus motif. This complex was supershifted with a specific antibody against Hnf-1. By contrast, no protein binding was observed using mutant probes (M1 or M2) that contain nucleotide mutations in the Hnf-1 binding consensus motif. (b) The amount of Hnf-1 bound to the region –94/−60 was reduced in a dose-dependent manner by adding 20- or 40-fold molar excess cold (non-32P labelled) probes. Adding non-labelled probes including mutated nucleotides in the Hnf-1 binding consensus motif did not alter the amount of Hnf-1 binding. W.C.E indicates the whole cell extract and N.S indicates the non-specific binding.

Figure 6. (a,b) The mRNA levels of Hnf-1α, but not Hnf-1β, in the liver were significantly increased after refeeding. After a 12 h fast, mice were refed ad libitum and sacrificed after 60 and 240 min. Total RNA of the liver was isolated and subjected to qPCR to quantify mRNA levels of Hnf-1α or β. Data represent mean ± SEM from 5 mice at each time point from triplicate PCR samples and the experiment was repeated once with similar result. Asterisks indicate the significant difference from mRNA levels in the liver of fasted mice (***p < 0.001). (c) Hepatic Hnf-1 protein levels were increased by refeeding. After a 12 h fast, mice were refed ad libitum and euthanised after 60 and 240 min. Whole cell lysates of the liver were prepared from three different mice at each time point and subjected to immunoblotting using anti-Hnf-1 and anti-cyclophilin antibodies. The blot of Hnf-1 and the blot of cyclophilin were cropped from different parts of the same membrane.

Figure 7. Insulin increased the expression of Hnf-1 mRNA and protein in Hepa1–6 cells (a,b). After 12 h from cell splits, Hepa1–6 cells were incubated with serum-free medium for 10 h and then transferred to a medium including 10 nM insulin. The cells were then collected at the indicated time points. (a) Total RNA isolated at each time point was subjected to qPCR to quantify Hnf-1α mRNA levels. Data represent mean ± SEM from triplicate PCR samples and the experiment was repeated once with similar results. Asterisks indicate a significant difference in the mRNA levels in Hepa1–6 cells at 0 min (***p < 0.001). (b) Whole cell lysates were prepared and subjected to immunoblotting using anti-Hnf-1, anti-p-Akt and anti-cyclophilin antibodies. The p-AKT was phosphorylated Akt. The blot of Hnf-1, the blot of p-AKT and the blot of cyclophilin were cropped from different parts of the same membrane. The blot of Angptl8 and cyclophilin was cropped from different membrane using the same whole cell lysates used for anti-Hnf-1, p-AKT and cyclophilin blotting. (c,d) Incubation with 10 nM insulin stimulated the promoter activity of Angptl8, including consensus binding motif for Hnf-1, in Hepa1–6 cells (c)…

Figure 8. Insulin increased the protein expression level of Angptl8 in primary hepatocytes and required Hnf-1 (a,b). At 12 h after cell splits, mouse primary hepatocyte cells were incubated with serum-free medium for 2 h and then switched to a medium with 10 nM of insulin and cells were collected at the indicated time points. Whole cell lysates were prepared and subjected to immunoblotting using anti-Hnf-1, anti-Angptl8 and anti-cyclophilin antibodies. The blot of Hnf-1 and cyclophilin were cropped from different parts of the same membrane. The blot of Angptl8 was cropped from different membrane using the same whole cell lysates used for anti-Hnf-1and cyclophilin blotting (a). The same experiment was performed after Hnf-1 knockdown in mouse primary hepatocytes (b) using anti-β actin antibodies instead of anti-cyclophilin antibodies. At 24 h after transfection of both 40 nM siHnf-1α and siHnf-1β or control siRNA (siCtr), the cells were stimulated by incubation in medium containing 10 nM insulin. Whole cell lysates were prepared and subjected to SDS-PAGE. SiCtr samples were collected after 60 min from insulin administration. In the absence of Hnf-1, insulin did not increase the expression…