Bromodomain Inhibition Reveals FGF15/19 As a Target of Epigenetic Regulation and Metabolic Control

Abstract

Epigenetic regulation is an important factor in glucose metabolism, but underlying mechanisms remain largely unknown. Here we investigated epigenetic control of systemic metabolism by bromodomain-containing proteins (Brds), which are transcriptional regulators binding to acetylated histone, in both intestinal cells and mice treated with the bromodomain inhibitor JQ-1. In vivo treatment with JQ-1 resulted in hyperglycemia and severe glucose intolerance. Whole-body or tissue-specific insulin sensitivity was not altered by JQ-1; however, JQ-1 treatment reduced insulin secretion during both in vivo glucose tolerance testing and ex vivo incubation of isolated islets. JQ-1 also inhibited expression of fibroblast growth factor (FGF) 15 in the ileum and decreased FGF receptor 4-related signaling in the liver. These adverse metabolic effects of Brd4 inhibition were fully reversed by in vivo overexpression of FGF19, with normalization of hyperglycemia. At a cellular level, we demonstrate Brd4 binds to the promoter region of FGF19 in human intestinal cells; Brd inhibition by JQ-1 reduces FGF19 promoter binding and downregulates FGF19 expression. Thus, we identify Brd4 as a novel transcriptional regulator of intestinal FGF15/19 in ileum and FGF signaling in the liver and a contributor to the gut-liver axis and systemic glucose metabolism.

Figure 1

JQ-1 induces hyperglycemia in vivo. A: Experimental protocol. GTT, glucose tolerance test; ITT, insulin tolerance test; ip, intraperitoneal. B: Body weight changes during JQ-1 treatment (n = 5–6). C: Fasting blood glucose levels (4-h fasting, day 10; n = 5–6). Veh, vehicle. Blood glucose levels (D) and AUC (E) during ipGTT (day 6). Plasma insulin (F) and glucagon (G) levels during ipGTT. H: Hematoxylin and eosin (H&E) and insulin staining of pancreatic sections. Scale bar, 200 and 100 μm. I: Islet size (n = 5–6; >100 islets/mouse). ipITT (day10) (J) and glycerol tolerance test (day 4) (K). Data are expressed as means ± SEM. *P < 0.05, **P < 0.01 vs. vehicle-treated mice (Veh).

Figure 2

JQ-1 reduces hepatic Fgfr signaling. A–C: Microarray analysis in the liver was performed in JQ-1 vs. mice vehicle (Veh)-treated mice (n = 3 per group). Minus log₁₀ (P value) (A) and upregulated or downregulated gene numbers (B) in pathways related to FGFR signaling. C: Heat map representing log₂ fold change (JQ-1 vs. vehicle) in gene expression in pathways related to FGFR ligand binding and activation. D and E: Western blot analysis of total and phosphorylated (p) FRS2 in liver lysates. β-actin is shown as a loading control. **P < 0.01 vs. vehicle (Veh).

Figure 3

JQ-1 decreases ileal Fgf15 expression without the change in overall villus structure. A: Western blot analysis of Brd4 in ileum and liver lysates. B: Schematic of bile acid signaling in ileal enterocytes. Expression levels of genes related to bile acid signaling in the ileum (Fgf15) (C), Nr0b2 (SHP) (D), Slc51a (OSTα) (E), and Slc51b (OSTβ) (F) (n = 5–6). Expression levels were normalized by Rpl13 expression. G: Plasma Fgf15 level in JQ-1 vs. vehicle (Veh)-treated mice (n = 5–6). *P < 0.05, **P < 0.01 vs. vehicle (Veh)-treated mice. Data are expressed as means ± SEM. H: Hematoxylin and eosin staining of intestinal sections. Scale bar, 100 μm.

Figure 4

BRD4 binds promoter regions of FGF19 and SHP to modulate expression in human intestinal cells. A–E: HT-29 cells were treated with 250 or 500 nmol/L JQ-1 or DMSO (Veh) for 12–24 h. Gene expression levels of BRD4 (A), MYC (B), FGF19 (C), and NR0B2 (SHP) (D) (n = 4). The levels were normalized by those of RN18S. E: FGF19 level in conditioned media (n = 4). F–H: HT-29 cells were treated with siRNA for BRD4 and JQ-1. Gene expression levels of BRD4 (F), and FGF19 (G) (n = 3). The levels were normalized by those of 36B4. H: Primer design for ChIP assays for FGF19 and NR0B2 (SHP) promoter regions. I and J: Cells were incubated with JQ-1 or DMSO (Veh) for 24 h and then harvested for the ChIP assay. The precipitated DNA was analyzed by PCR (I) and real-time PCR (J) (n = 4). Data are expressed as mean ± SEM. *P < 0.05, **P < 0.01 vs. vehicle (Veh)-treated mice. N.D., not detected.

Figure 5

Overexpression of FGF19 reverses hyperglycemia induced by JQ-1. A: Experimental protocol. GTT, glucose tolerance test; ip, intraperitoneal. B: Plasma FGF19 levels in AAV-GFP– and AAV-FGF19–treated mice. C: Body weight in AAV-GFP– and AAV-FGF19–treated mice during JQ-1 treatment. Blood glucose levels (D) and AUC (E) during ipGTT (day10). F: Plasma insulin levels during ipGTT. Data are expressed as means ± SEM. *P < 0.05, **P < 0.01 vs. AAV-GFP–treated mice (n = 9–10). N.D., not detected.

Figure 6

Hypothetical model illustrating how inhibition of Brd4 induced hyperglycemia in vivo.