Novel Mechanism of Foxo1 Phosphorylation in Glucagon Signaling in Control of Glucose Homeostasis

Abstract

Dysregulation of hepatic glucose production (HGP) serves as a major underlying mechanism for the pathogenesis of type 2 diabetes. The pancreatic hormone glucagon increases and insulin suppresses HGP, controlling blood glucose homeostasis. The forkhead transcription factor Foxo1 promotes HGP through increasing expression of genes encoding the rate-limiting enzymes responsible for gluconeogenesis. We previously established that insulin suppresses Foxo1 by Akt-mediated phosphorylation of Foxo1 at Ser²⁵⁶ in human hepatocytes. In this study, we found a novel Foxo1 regulatory mechanism by glucagon, which promotes Foxo1 nuclear translocation and stability via cAMP- and protein kinase A-dependent phosphorylation of Foxo1 at Ser²⁷⁶ Replacing Foxo1-S276 with alanine (A) or aspartate (D) to block or mimic phosphorylation, respectively, markedly regulates Foxo1 stability and nuclear localization in human hepatocytes. To establish in vivo function of Foxo1-Ser²⁷⁶ phosphorylation in glucose metabolism, we generated Foxo1-S273A and Foxo1-S273D knock-in (KI) mice. The KI mice displayed impaired blood glucose homeostasis, as well as the basal and glucagon-mediated HGP in hepatocytes. Thus, Foxo1-Ser²⁷⁶ is a new target site identified in the control of Foxo1 bioactivity and associated metabolic diseases.

Figure 1

The effect of hepatic Foxo1 deletion on glucagon-induced glucose production and gluconeogenesis. A: Blood glucose level in random-fed mice in 1 h after i.p. injection of 0.25 mg/kg body weight glucagon. Ten-week-old mice per treatment. n = 6 mice/group. *P < 0.05 vs. control mice (CNTR); #P < 0.05 for the comparison of designated groups. B: Glucagon tolerance tests for mice after an 18-h overnight fast. Blood glucose levels (mean ± SEM) were determined at the indicated time points after i.p. injection of 16 µg/kg body weight glucagon. n = 6 mice/group. *P < 0.05 vs. CNTR. C–E: Quantitative RT-PCR analyses of the livers of mice with i.p. injection of 0.25 mg/kg body glucagon for 1 h. n = 3 mice/treatment. *P < 0.05 vs. CNTR; #P < 0.05 for the comparison of designated groups. F: HGP, glycogenolysis, and gluconeogenesis in primary hepatocytes. Cells were isolated from WT or L-F1KO mice, cultured in DMEM with 2% FBS medium for attachment for 4 h, and then switched to HGP buffer with or without pyruvate substrate. HGP was measured 3 h after 100 nmol/L glucagon stimulation and normalized to total protein levels. n = 3/group. *P < 0.05 vs. CNTR; #P < 0.05 vs. L-F1KO.

Figure 2

Regulation of Foxo1 stability and activity by insulin and glucagon in hepatocytes. A: Insulin and glucagon signaling in HepG2 cells. Cells were cultured in DMEM with 10% FBS and then starved in DMEM with 1% BSA medium overnight prior to the treatment of 100 nmol/L insulin or 100 nmol/L glucagon over a 6-h time course. Western blotting was performed with 150 µg protein of cell lysates to detect total Foxo1 (t-Foxo1), Akt, CREB, phosphorylated (p)Foxo1-S256, pAkt-T308, and pCREB-S133. Quantification of t-Foxo1/GAPDH (B), pFoxo1-S256/t-Foxo1 (C), pAKT-T308/Akt (D), and pCREB-S133/CREB (E) from A was performed using ImageJ. Results are presented as mean ± SEM. *P < 0.05 vs. 0-min treatment. n = 3 experiments. F: The mRNA expression of Foxo1 in HepG2 cells. HepG2 cells were cultured in DMEM with 10% FBS and then starved in DMEM with 1% BSA medium overnight prior to the treatment of 100 nmol/L glucagon over a 6-h time course. The Foxo1 mRNA level was detected by quantitative PCR. n = 3 experiments. CNTR, control.

Figure 3

Identification of novel phosphorylation sites of Foxo1 by PKA. A: PKA activates and Akt inhibits Igfbp-1 promoter activity in cells. HepG2 cells were cultured in a six-well plate with DMEM with 10% FBS and then transfected with 0.5 µg Igfbp-1–Luc reporter gene and 0.1 µg RL-luc internal control together with 0.5 µg expression vector for GFP, PKAc, Myr-Akt, hemagglutinin (HA)-Foxo1, or a combination of the two as designated. Cell lysates were prepared for luciferase assays and protein levels of PKA, Akt, Foxo1, or β-actin determined by Western blotting from 20 µg of total protein. n = 3 experiments. Values are fold changes of luciferase activity relative to GFP group and presented as mean ± SEM. *P < 0.05 vs. GFP; #P < 0.05 vs. Foxo1. B: Putative phosphorylation sites of Foxo1 by PKA among different species. C: PKA phosphorylates Foxo1 in vitro. A total of 20 µg of recombinant Foxo1-GST was mixed with 5 units (0.5 µg) recombinant active PKAc with or without 0.1 mmol/L ATP for 30 min at 30°C and then the Foxo1-GST and PKAc were resolved in SDS-PAGE after the Coomassie Brilliant Blue staining. The intact Foxo1-GST (105 kDa) or cleaved Foxo1 (60 kDa) was incised from the gel and subjected to LC-MS/MS. D: MS/MS spectra of peptide containing S153 and S276 of Foxo1. Antibodies against phosphorylated Foxo1 at S153 and S276 were generated. E and F: Phosphorylation of Foxo1-S276 and S153 by glucagon in cells. E: HepG2 cells were transfected with either scramble siRNA or siRNA–PKA catalytic subunit C (PKACB) for 24 h and then treated with or without 100 nmol/L glucagon for 60 min. Quantification of total Foxo1 (t-Foxo1), phosphorylated (p)Foxo1-S276, pFoxo1-S153, pCREB, and CREB was performed using ImageJ and normalized by GAPDH. *P < 0.05 vs. scramble siRNA; #P < 0.05 vs. scramble siRNA + glucagon. n = 3 experiments. F: Primary hepatocytes were cultured and treated with either 100 nmol/L glucagon or 10 μmol/L 8-Br–cAMP for 30 min or 10 μmol/L PKA inhibitor H89 for 30 min prior to 60 min of 100 nmol/L glucagon treatment. Phosphorylation of Foxo1-S150 and S273 were analyzed by Western blotting using antibodies at 1:500 dilution. Quantification of pFoxo1-S273, pFoxo1-S150, and t-Foxo1 was performed using ImageJ and normalized by β-actin. n = 3 experiments. *P < 0.05 vs. vehicle; #P < 0.05 vs. glucagon.

Figure 4

Glucagon promotes phosphorylation of Foxo1-S153 and S276 in vivo. A: Random-fed mice were i.p. injected with glucagon at 0.25 mg/kg body weight, and the liver was collected 15 min after injection to determine hepatic Foxo1 phosphorylation at S150 and S273 by using Western blotting. Quantification of phosphorylated (p)Foxo1-S273 and pFoxo1-S150 was performed using ImageJ and normalized by β-actin. *P < 0.05 vs. vehicle. n = 4–6 mice/group. B–D: Eight-week-old WT mice were i.v. injected with GCGR antagonist ([des-His¹, Glu⁹]-Glucagon amide; 1 mg/kg body weight) after 6 h fasting. B: Blood glucose level was measured at the indicated time points. *P < 0.05 vs. vehicle. n = 4–6 mice/group. C: Area under the curve (AUC) from B was calculated. *P < 0.05 vs. vehicle. n = 4–6 mice/group. D: Liver was collected 45 min after [des-His¹, Glu⁹]-Glucagon amide injection to determine hepatic total Foxo1 (t-Foxo1) and Foxo1 phosphorylation at S150 and S273 by using Western blotting. Quantification of t-Foxo1, pFoxo1-S273, pFoxo1-S150, pCREB, and CREB was performed using ImageJ and normalized by GAPDH. *P < 0.05 vs. vehicle. n = 4–6 mice/group.

Figure 5

Foxo1-S276 phosphorylation controls cAMP-induced Foxo1 nuclear localization in hepatocytes. HepG2 cells were transfected with 5 µg plasmid DNA expressing GFP-Foxo1-WT, S276A, S276D, S153A, or S153D for 18 h and then treated with or without 10 μmol/L cAMP for 30 min. GFP-positive cells were displayed, stained with DAPI for nucleus, and analyzed under confocal microscope. Representative images are shown.

Figure 6

Glucagon or cAMP stimulates Foxo1 nuclear localization and Foxo1-S276 phosphorylation, enhancing Foxo1 stability in hepatocytes. A: HepG2 cells were transfected with 5 µg plasmid DNA expressing hemagglutinin (HA)-Foxo1 for 18 h and then treated with 100 nmol/L glucagon, 10 μmol/L cAMP, or 100 nmol/L insulin for 30 min prior to extraction of nuclear (N) and cytoplasmic (C) proteins. A total of 20 µg nuclear protein or 100 µg cytoplasmic protein was immunoblotted to determine the abundance of total Foxo1 (t-Foxo1) in nucleus or cytoplasm. Signal intensity was quantified by ImageJ software for statistical comparison. Cytoplasmic and nuclear protein levels were normalized by GAPDH and histone H1, respectively. n = 3 experiments. *P < 0.05 vs. vehicle in cytoplasm; #P < 0.05 vs. vehicle in nucleus. B and C: Distribution of Foxo1-S276 mutant proteins in the nucleus and cytoplasm of cells. HepG2 cells were transfected with 5 µg plasmid DNA expressing Foxo1-WT, S276A, or S276D for 18 h and then treated with or without 100 nmol/L glucagon (Glc) for 30 min. Nuclear (N) or cytoplasmic (C) protein was extracted to determine protein abundance of Foxo1-S276 mutants in nucleus and cytoplasm. Representative images are shown in B and C, and signal intensity was quantified by ImageJ software for statistical comparison. n = 3 experiments. *P < 0.05 vs. vehicle in cytoplasm; #P < 0.05 vs. vehicle in nucleus. D: Foxo1-S276 mutations influence Foxo1 stability in a proteasome-dependent manner. Western blotting was performed to detect t-Foxo1 abundance in HepG2 cells that was transfected with the same amount of plasmid DNA expressing Foxo1-WT, S276A, or S276D for 10 h, and cells were then starved for 8 h and treated with 10 μmol/L CHX or 10 µg/mL MG132 (MG) for 6 h prior cellular protein collection. Representative images are shown. Signal intensity was quantified and normalized by α-actinin for statistical analysis. *P < 0.05 vs. Foxo1-WT; #P < 0.05 for the comparison of designated groups. n = 3 different experiments. E: Foxo1-S276 mutations influence insulin-induced Foxo1 ubiquitination. HepG2 cells were cotransfected with eGFP expression vectors encoding WT Foxo1, Foxo1-S276D, or Foxo1-S276A and/or FLAG-ubiquitin for 10 h; cells were starved for 12 h and then treated with 100 nmol/L insulin for 12 h. Quantification of t-Foxo1 normalized by α-actinin was performed using ImageJ. *P < 0.05 vs. Foxo1-WT and FLAG-ubiquitin group; #P < 0.05 for comparison of designated groups. n = 3 different experiments.

Figure 7

Foxo1-S276D impairs insulin-stimulated Foxo1 nuclear export in hepatocytes. A: HepG2 cells were transfected with 5 µg plasmid DNA expressing eGFP-Foxo1-WT or eGFP-Foxo1-S276D for 18 h, and cells were then serum starved for 8 h prior to 100 nmol/L insulin treatment for 0, 30, 90, or 120 min. The GFP-positive cells were displayed, stained with DAPI for nucleus, and analyzed under the confocal microscope. Representative images are shown. B: HepG2 cells were transfected with 10 µg eGFP expression vectors encoding Foxo1-WT or Foxo1-S276D for 18 h and followed by 8-h serum starvation prior to 100 nmol/L insulin (Ins) stimulation for 90 and 120 min. Western blotting was performed for total Foxo1 (t-Foxo1) and phosphorylated (p)Foxo1-S256 in cytoplasm (C) and nucleus (N). C: Quantification of t-Foxo1 and pFoxo1-S256/t-Foxo1 was performed by ImageJ. Cytoplasmic (C) and nuclear (N) protein levels were normalized by GAPDH and histone H1, respectively. Results are presented as mean ± SEM. n = 3 different experiments. *P < 0.05 vs. 0 min of insulin treatment from Foxo1-WT in cytoplasm; #P < 0.05 vs. 0 min of insulin treatment from Foxo1-WT in nucleus.

Figure 8

Foxo1-S273 mutations impair blood glucose and HGP in mice. A: Blood glucose levels of random-fed or 18-h–fasted mice. n = 4–6 mice/genotype. *P < 0.05 vs. WT mice. B: Quantitative PCR analysis of expression of Igfbp-1 and G6pc in the liver of 18-h–fasted mice. n = 4 mice/genotype. *P < 0.05; **P < 0.01 vs. WT mice. C: Foxo1 protein levels in the livers of the mutant mice. Western blotting results from two mice per genotype are shown representatively. *P < 0.05; **P < 0.01 vs. WT. n = 4/group. D: HGP assays in the primary hepatocytes isolated from WT, Foxo1-S273^A/A, and Foxo1-S273^D/D KI mice. Fresh hepatocytes were isolated from these mice, and HGP were measured 3 h after 100 nmol/L glucagon stimulation and normalized to total protein levels. n = 3/treatment. *P < 0.05 vs. WT with vehicle group; #P < 0.05 for the comparison of designated groups. E and F: Glucagon tolerance tests. E: Glucagon tolerance tests for mice after an 18-h overnight fast. Blood glucose levels (mean ± SEM) were determined at the indicated time points after i.p. injection of 16 µg/kg body weight glucagon. n = 6 mice. *P < 0.05 vs. WT. F: Area under the curve (AUC) from E was calculated. *P < 0.05 vs. WT. n = 4–6/group. G: Blood glucagon, insulin, and glucose levels in WT and db/db mice in an 18-h fasting state. **P < 0.01 vs. WT. n = 4 mice/group. H: Phosphorylated (p)Foxo1-S273, pFoxo1-S253, total Foxo1 (t-Foxo1), pPKA-T179, PKA, histone H1, and GAPDH in 100 µg cytoplasmic protein (C) and 20 µg nuclear protein (N) of mouse livers were determined by Western blotting. I: Quantification of pFoxo1-S273, pS253, t-Foxo1, pPKA-T197, or PKA was performed by ImageJ. The cytoplasmic (C) and nuclear (N) protein levels were normalized by GAPDH and histone H1, respectively. *P < 0.05 vs. WT. n = 3 mice/group. Results are shown as mean ± SEM. J: Blood glucagon, insulin, and glucose levels in LFD- and HFD-fed WT mice in an 18-h fasting state. *P < 0.05; **P < 0.01 vs. LFD. n = 4/group. K: pFoxo1-S273, t-Foxo1, pCREB, and CREB levels of mouse livers were determined by Western blotting. L: Quantification of pFoxo1-S273, t-Foxo1, pCREB, and CREB was performed by ImageJ. *P < 0.05 vs. LFD. n = 4 mice/group. Results are shown as mean ± SEM.