Ethanol induction of FGF21 in the liver is dependent on histone acetylation and ligand activation of ChREBP by glycerol-3-phosphate

Abstract

Ethanol rapidly stimulates the liver to synthesize the hormone fibroblast growth factor 21 (FGF21), which then acts on the brain to elicit a multifaceted protective response. We show that in mice, this induction of FGF21 occurs at the level of gene transcription and is regulated by two byproducts of ethanol metabolism, glycerol-3-phosphate (G3P) and acetyl-CoA. Using cell-based reporter and thermal shift binding assays, we show that G3P binds to a conserved domain and activates the transcription factor carbohydrate-responsive element-binding protein (ChREBP), which regulates the Fgf21 gene promoter. The stimulation of Fgf21 gene transcription by ethanol also requires its metabolism to acetyl-CoA and correlates with histone acetylation. Accordingly, a p300/CBP histone acetyltransferase inhibitor blocks histone acetylation, ChREBP recruitment, and transcriptional activation at the Fgf21 promoter. Together, these findings reveal a dual regulatory mechanism driven by both G3P and acetyl-CoA that explains ethanol's robust stimulatory effect on Fgf21 and possibly other ChREBP target genes in the liver.

Keywords: ChREBP; FGF21; alcohol; liver; transcription.

Fig. 1.

Ethanol stimulates hepatic FGF21 synthesis. (A) Hepatic Fgf21 mRNA and (B) plasma FGF21 levels in C57BL/6 mice 1 h after giving a single bolus of water or increasing concentrations of ethanol by gavage. Values are means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 for water versus ethanol gavage groups (n = 6 to 8) by one-way ANOVA with Tukey’s multiple comparisons test. (C) Genome browser tracks of PRO-seq data for the Fgf21 gene in the livers of C57BL/6 mice 1 h after giving water or ethanol (1 g/kg) by oral gavage. Browser tracks show merged data from 3 mice per group. Note that the Fgf21 gene is transcribed from the bottom (–) strand.

Fig. 2.

Ethanol induction of FGF21 synthesis requires metabolism to acetyl-CoA. (A) Schematic showing ethanol metabolism by alcohol dehydrogenase 1 (ADH1), aldehyde dehydrogenase 2 (ALDH2), and acyl-CoA synthetase short-chain family member 2 (ACSS2). (B–D) Hepatic Fgf21 mRNA and plasma FGF21 levels in global Adh1-KO (n = 10 to 13), Aldh2-KO (n = 6 to 12), or Acss2-KO (n = 7 to 12) mice and their corresponding wild-type (WT) (n = 12) mice 1 h after giving water (–) or 1 g/kg ethanol (EtOH) (+) by oral gavage. Values are means ± SEM. **P < 0.01, ***P < 0.001, ****P < 0.0001 for water versus ethanol gavage groups by two-way ANOVA with Tukey’s multiple comparisons test.

Fig. 3.

ChREBP-dependent induction of Fgf21 and other genes by ethanol does not require ChREBP acetylation. (A) Hepatic Fgf21 mRNA and (B) plasma FGF21 levels in control Chrebp^fl/fl (n = 7 to 9) and Chrebp-HepKO (n = 7 to 9) mice 1 h after giving water (–) or 1 g/kg ethanol (+) by oral gavage. (C) ChIP analysis showing the effect of ethanol on ChREBP binding to the Fgf21 promoter in mouse livers (n = 3). (D) Hepatic Fgf21 mRNA in WT mice 1 h after giving either water (–) or 1 g/kg ethanol (+) by oral gavage plus vehicle or p300/CBP inhibitor (C646, 20 mg/kg) by i.p. injection n = 6 to 7). (E) Hepatic Fgf21 mRNA levels in Chrebp-HepKO, Chrebp-HepKO + Chrebp, and Chrebp-HepKO + Chrebp^KR mice 1 h after giving water (–) or 1 g/kg ethanol (+) by oral gavage (n = 5). (F) Hepatic mRNA levels for Chrebpβ, glucose-6-phosphatase (G6pc), thioredoxin-interacting protein (Txnip), and Krüppel-like factor 10 (Klf10) in Chrebp-HepKO, Chrebp-HepKO + Chrebp, and Chrebp-HepKO + Chrebp^KR mice under the same conditions as in (E) (n = 4 to 5). To create hepatocyte knockouts of ChREBP, Chrebp^fl/fl mice were infected with hepatocyte-selective AAVs expressing GFP (Chrebp^fl/fl) or CRE (Chrebp-HepKO). To test the effects of acetylation, Chrebp^fl/fl mice were infected with an AAV expressing CRE together with AAVs expressing wild-type ChREBP (Chrebp-HepKO+Chrebp) or the acetylation sites mutant ChREBP (Chrebp-HepKO+Chrebp^KR). Values are means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 for water versus ethanol gavage groups by two-way ANOVA with Tukey’s multiple comparisons test.

Fig. 4.

Ethanol-mediated induction of FGF21 requires histone acetylation of the Fgf21 promoter. (A) Western blot images (Left panel) and quantification (Right panel) of H3K9 acetylation of the Fgf21 promoter 1 h after giving water (–) or 1 g/kg ethanol (+) by oral gavage to WT mice. Values are means ± SEM. *P < 0.05 by unpaired t test. (B) ChIP analysis showing H3K9 acetylation of the Fgf21 promoter in mouse liver 1 h after giving water (–) or 1 g/kg ethanol (+) by oral gavage plus vehicle or p300/CBP inhibitor (C646, 20 mg/kg) by i.p. injection. Immunoprecipitation was performed with IgG or anti-H3K9ac antibodies. (C) ChIP analysis of ChREBP binding to the hepatic Fgf21 promoter under the same conditions as in (B). (D) ChIP analysis demonstrating p300 binding to the hepatic Fgf21 promoter 1 h after giving water (–) or 1 g/kg ethanol (+) by oral gavage to control Chrebp^fl/fl or Chrebp-HepKO mice (n = 3). Data are expressed as means ± SEM from n = 3 individual mice per group. In (B–D) *P < 0.05, **P < 0.01, and ****P < 0.0001 by two-way ANOVA with Tukey’s multiple comparisons test comparing water to EtOH treatment in each group unless otherwise noted.

Fig. 5.

Glycerol-3-phosphate (G3P) activates ChREBP. (A) G3P and glyceraldehyde-3-phosphate (Ga3P) concentrations in mouse liver 1 h after giving water (–) or 1 g/kg ethanol (+) by oral gavage. (B) G3P concentrations in HEK-293T cells in either the absence or presence of GPD1 coexpression. (C) The GRACE domain of ChREBP is required for G3P induction of the Fgf21 promoter. Luciferase reporter assay of the Fgf21 promoter in HEK-293T cells expressing either WT ChREBP or ChREBP–ΔGRACE in the absence (–) or presence (+) of GPD1 coexpression. (D) Deletion of the ChREBP response element (ΔChoRE) in the Fgf21 promoter abolishes G3P-mediated transcription. In (C and D) HEK-293T cells were cotransfected with empty vector or vectors expressing WT ChREBP, GRACE-deleted (ΔGRACE) ChREBP, or GPD1, as indicated, and either a WT Fgf21-luciferase reporter plasmid (in panel C and D) or an Fgf21-luciferase reporter plasmid lacking the ChoRE (ΔChoRE, in panel D). All values are means ± SEM. ***P < 0.001, ****P < 0.0001. In (A, C, and E), p-values refer to comparisons versus no ethanol or GPD1 by two-way ANOVA with Tukey’s multiple comparisons test. Significance in (B) was determined by an unpaired t test.

Fig. 6.

G3P binds to ChREBP. (A and B) Thermal shift analysis was performed using lysates from HEK-293T cells expressing WT or GRACE-domain deleted (ΔGRACE) ChREBP that were treated with vehicle or 1 mM G3P for 30 min. Insets show representative thermal shift stabilities of ChREBP proteins. (C) Model of the dual mechanism for activation of Fgf21 gene expression by ethanol.