The transcription factor COUP-TFII is negatively regulated by insulin and glucose via Foxo1- and ChREBP-controlled pathways

Abstract

COUP-TFII has an important role in regulating metabolism in vivo. We showed this previously by deleting COUP-TFII from pancreatic beta cells in heterozygous mutant mice, which led to abnormal insulin secretion. Here, we report that COUP-TFII expression is reduced in the pancreas and liver of mice refed with a carbohydrate-rich diet and in the pancreas and liver of hyperinsulinemic and hyperglycemic mice. In pancreatic beta cells, COUP-TFII gene expression is repressed by secreted insulin in response to glucose through Foxo1 signaling. Ex vivo COUP-TFII reduces insulin production and secretion. Our results suggest that beta cell insulin secretion is under the control of an autocrine positive feedback loop by alleviating COUP-TFII repression. In hepatocytes, both insulin, through Foxo1, and high glucose concentrations repress COUP-TFII expression. We demonstrate that this negative glucose effect involves ChREBP expression. We propose that COUP-TFII acts in a coordinate fashion to control insulin secretion and glucose metabolism.

FIG. 1.

In vivo COUP-TFII expression in pancreas and liver. C57BL/6J mice were either fasted for 24 h or fed an HCHO diet for 18 h. (A and B) RTQ-PCR analysis of COUP-TFII in pancreas (A) and liver (B, left) and quantification from triplicate Western blots of COUP-TFII protein from liver (B, right). (C) Blood glucose, plasma insulin, and regulation of COUP-TFII in liver and pancreas of HGHI and eGHI clamped mice. Results are means ± SEM (where n = 6 mice/group).

FIG. 2.

Ex vivo COUP-TFII expression in pancreatic beta cells. After a 24-h preincubation at low glucose concentrations, 832/13 INS-1 cells were incubated for 4 h or C57BL/6J mouse islets were incubated for 12 h in the presence of glucose or insulin. (A) RTQ-PCR analysis of islet COUP-TFII mRNA after glucose stimulation. (B and C) RTQ-PCR analysis of the effects of different glucose (B) and insulin (C) concentrations on COUP-TFII mRNA levels in 832/13 INS-1 cells. (D) Insulin does not affect the stability of COUP-TFII transcripts in 832/13 INS-1 cells. After a 24-h preincubation in 5 mM glucose, cells were incubated for different periods in the presence of 5 mM glucose with 2.5 μM actinomycin D (Act D) (banded bars) or 5 mM glucose plus 100 nM insulin with 2.5 μM actinomycin D (gray bars). (E) Quantification from triplicate Western blots of COUP-TFII protein from islets and 832/13 INS-1 cells cultured as described above (A to C). Data are the means ± SEM for at least three independent experiments.

FIG. 3.

Glucose-stimulated insulin secretion decreases COUP-TFII mRNA expression levels in 832/13 INS-1 cells. Following a 24-h preincubation in 5 mM glucose, cells were incubated for 4 h with 5 mM glucose, 5 mM glucose plus 100 nM insulin, 5 mM glucose plus 100 μM HNMPA-(AM)₃, 5 mM glucose plus 15 mM KCl, 20 mM glucose, 20 mM glucose plus 20 μM verapamil, or 20 mM glucose plus 100 μM HNMPA-(AM)₃. The resulting COUP-TFII (A), PPARα (B), L-PK (C), and proinsulin I and II (D) mRNA levels were measured by RTQ-PCR. Results are means ± SEM for six independent experiments.

FIG. 4.

COUP-TFII mRNA expression is negatively controlled by insulin via the Foxo1 pathway in 832/13 INS-1 cells. (A and B) Cells were cultured in 5 mM glucose and infected overnight with Ad-GFP or Ad-Foxo1-ADA as indicated (2 to 5 PFU/cell) in the presence of 5 mM glucose plus 100 nM insulin. (A) Representative Western blot of total protein (25 μg/lane) showing HA-Foxo1-ADA (top) as detected with antibodies against the HA tag and α-tubulin as a loading control (bottom). (B) RTQ-PCR of steady-state COUP-TFII mRNA and control PPARα mRNA levels. (C) 832/13 INS-1 beta cells were electroporated with control or with rat Foxo1 siRNA and incubated with 5 mM glucose for 24 h. Shown are data for RTQ-PCR measuring Foxo1, Pdx1, PPARα, and COUP-TFII mRNA levels. Results are means ± SEM for three independent experiments.

FIG. 5.

COUP-TFII represses insulin mRNA expression and insulin release and lowers insulin content and intracellular TG content. 832/13 INS-1 cells were electroporated with control or specific COUP-TFII siRNA and cultured for 48 h. (A) COUP-TFII mRNA levels measured by RTQ-PCR. Results are means ± SEM for eight independent experiments. (B) Representative Western blot with 20 μg/lane of nuclear extracts showing COUP-TFII protein (top) and the GAPDH loading control (bottom). (C) RTQ-PCR analysis of proinsulin I and II and control NeuroD1 and Pdx1 mRNA levels in control and COUP-TFII knockdown 832/13 INS-1 cells. Results are means ± SEM for eight independent experiments. (D and E) 832/13 INS-1 cells (D) and islets (E) were cultured with 11 mM glucose and infected with Ad-GFP or Ad-hCOUP-TFII as indicated (2 to 5 and 200 PFU/cell, respectively). Shown are representative Western blots of 20 μg/lane of nuclear extracts showing COUP-TFII (top) and GAPDH (bottom) as the loading control and RTQ-PCR analysis of proinsulin I and II mRNA levels. (F and G) Insulin release and insulin content in control (white bars) and COUP-TFII siRNA-electroporated (black bars) 832/13 INS-1 cells. (F) Induction of glucose-stimulated insulin secretion in 832/13 INS-1 cells. (G) Increase in insulin content normalized to the DNA content/well. Data are the means ± SEM for three independent experiments. (H) [1-¹⁴C]palmitate esterification into cellular TGs was measured in control and COUP-TFII-suppressed 832/13 INS-1 cells at the end of a 2-h incubation period in the presence of either 2.5 mM or 20 mM glucose. Results are means ± SEM of data from four independent experiments.

FIG. 6.

In hepatocytes, COUP-TFII expression is negatively controlled by insulin via the Foxo1 pathway and by glucose via the ChREBP pathway. (A) C57BL/6J mouse hepatocytes were cultured for 24 h in the presence of 5 mM glucose. Cells were then incubated with either 5 mM or 25 mM glucose with or without 100 nM insulin. After 24 h, steady-state COUP-TFII mRNA levels were determined by RTQ-PCR. Data are the means ± SEM for five independent experiments. (B) C57BL/6J mouse hepatocytes were cultured for 24 h in the presence of 5 mM glucose and infected for 4 h with Ad-GFP or Ad-Foxo1-ADA as indicated (0.5 and 2 PFU/cell) and then cultured for 24 h in the presence of 25 mM glucose plus 100 nM insulin. Shown are data for RTQ-PCR of steady-state COUP-TFII mRNA levels. Results are means ± SEM for three independent experiments. (C) C57BL/6J mouse hepatocytes were cultured for 12 h in the presence of 5 mM glucose. Hepatocytes were then transfected with control or Foxo1 siRNA and incubated with 5 mM glucose for 48 h. RTQ-PCR measured Foxo1, G6Pase, ChREBP, and COUP-TFII mRNA expression levels. (D) Effect of fasting and HCHO refeeding in control and hGK-KO mice. Total RNA was extracted from liver, and COUP-TFII mRNA levels were analyzed by RTQ-PCR. Results are means ± SEM from six mice/group. (E and F) ob+/− and ob/ob mice treated for 7 days with Ad-GFP (−) or Ad-shChREBP (+). (E) RTQ-PCR measuring COUP-TFII mRNA levels in liver. (F) Quantification from triplicate Western blots of COUP-TFII protein from total liver extracts. Cyclophilin expression was used as a loading control. Results are means ± SEM from six mice/group. (G and H) C57BL/6J mouse hepatocytes were cultured for 24 h in the presence of 5 mM glucose and transfected with control or ChREBP siRNA in the presence of 5 mM glucose and incubated with 25 mM glucose plus 100 nM insulin for 24 h the next day. RTQ-PCR measured ChREBP and L-PK (G) and COUP-TFII (H) mRNA expression levels.

FIG. 7.

Proposed scheme of events underlying the regulation of insulin secretion in pancreatic beta cells and glucose metabolism in hepatocytes by COUP-TFII. Shown are positive (pointed arrows) and negative (flat-headed arrows) effects of COUP-TFII in hepatocytes and pancreatic beta cells in fasted (A) and HCHO-fed (B) states.