PHD3 regulates glucose metabolism by suppressing stress-induced signalling and optimising gluconeogenesis and insulin signalling in hepatocytes

Abstract

Glucagon-mediated gene transcription in the liver is critical for maintaining glucose homeostasis. Promoting the induction of gluconeogenic genes and blocking that of insulin receptor substrate (Irs)2 in hepatocytes contributes to the pathogenesis of type 2 diabetes. However, the molecular mechanism by which glucagon signalling regulates hepatocyte metabolism is not fully understood. We previously showed that a fasting-inducible signalling module consisting of general control non-repressed protein 5, co-regulator cAMP response element-binding protein binding protein/p300-interacting transactivator with Glu/Asp-rich carboxy-terminal domain 2, and protein kinase A is required for glucagon-induced transcription of gluconeogenic genes. The present study aimed to identify the downstream effectors of this module in hepatocytes by examining glucagon-induced potential target genes. One of these genes was prolyl hydroxylase domain (PHD)3, which suppressed stress signalling through inhibition of the IκB kinase-nuclear factor-κB pathway in a proline hydroxylase-independent manner to maintain insulin signalling. PHD3 was also required for peroxisome proliferator-activated receptor γ coactivator 1α-induced gluconeogenesis, which was dependent on proline hydroxylase activity, suggesting that PHD3 regulates metabolism in response to glucagon as well as insulin. These findings demonstrate that glucagon-inducible PHD3 regulates glucose metabolism by suppressing stress signalling and optimising gluconeogenesis and insulin signalling in hepatocytes.

Figure 1

Phd3 expression is induced by glucagon–cAMP–PKA signalling in mouse hepatocytes. (A) qRT-PCR detection of mRNA levels of three PHD isoforms and Vegf in primary hepatocytes with or without exposure to pCPT-cAMP (100 μM, 6 h) or hypoxia (1% O₂, 6 h). (B,C) qRT-PCR (B) and immunoblot (C) analyses of PHD3 in primary mouse hepatocytes incubated in the absence or presence of 100 μM pCPT-cAMP for indicated times. Cell lysates were subjected to immunoblot analysis of PHD3, Ser157-phosphorylated VASP, total VASP, or α-tubulin. (D) qRT-PCR analysis of Phd3 mRNA level in primary hepatocytes with or without exposure to pCPT-cAMP (100 μM, 6 h) and with or without pre-treatment with H89 (20 μM, 30 min). (E) qRT-PCR analysis of Phd3 mRNA level in various tissues of C57BL/6 J mice fed normal chow (NC) or a high-fat diet (HFD) for 20 weeks in the fed state. BAT, brown adipose tissue; EWAT, epididymal white adipose tissue; Gastro, gastrocnemius muscle; SWAT, subcutaneous white adipose tissue. (F,G) qRT-PCR analysis (F) and immunoblot analysis (G) of PHD3 in the liver of db/db and db/m (control) mice fed NC at 8 weeks of age after food deprivation for 16 h. α-Tubulin served as the loading control for immunoblotting. *Non-specific. Complete immunoblots are presented in Supplementary Fig. S7. Quantitative data are shown as mean ± SEM (n = 3 (A,B,D), 7 (E), or 6 (F)); results in (A–D) are representative of at least two independent experiments. Differences between groups were evaluated by ANOVA followed by Bonferroni’s post hoc test (A,D) or with the unpaired Student’s t test (E,F). *P < 0.05, **P < 0.01 vs. control or as indicated.

Figure 2

GCN5–CITED2–PKA signalling module mediates cAMP induction of PHD3 gene expression. (A) Effects of shRNA-mediated CITED2 knockdown on mRNA expression of three PHD isoforms in primary mouse hepatocytes with or without exposure to pCPT-cAMP (100 μM, 6 h). (B) Effects of ectopic CITED2 expression on Phd3 gene expression in primary mouse hepatocytes with or without exposure to pCPT-cAMP (100 μM, 6 h). (C,D) Effects of shRNA-mediated GCN5 (C) or PGC-1α (D) knockdown on Phd3 gene expression in primary mouse hepatocytes with or without exposure to pCPT-cAMP (100 μM, 6 h). Data are shown as mean ± SEM (n = 3) and are representative of at least two independent experiments. Differences between groups were evaluated by ANOVA followed by Bonferroni’s post hoc test. **P < 0.01 vs. indicated groups. Adenoviral vectors encoding CITED2 and shRNAs targeting CITED2, GCN5, and PGC-1α were used for experiments.

Figure 3

PHD3 is required for cAMP-dependent induction of gluconeogenesis. (A,B) Effects of shRNA-mediated PHD3 knockdown on gluconeogenic gene expression (A) and glucose production (B) in primary mouse hepatocytes with or without exposure to pCPT-cAMP for 6 and 22 h, respectively. (C,D) Effects of ectopic expression of shRNA-resistant PHD3(WT) or PHD3(ΔPH) on gluconeogenic gene expression (C) and glucose production (D) in primary mouse hepatocytes with or without shRNA-mediated knockdown of PHD3 in the presence of pCPT-cAMP (100 μM, 6 h). (E) Effects of shRNA-mediated PHD3 knockdown on HIF-1α and -2α protein levels in hepatocytes with or without exposure to DMOG (1 mM, 4 h). α-Tubulin served as the loading control for immunoblotting. Complete immunoblots are presented in Supplementary Fig. S7. Quantitative data are shown as mean ± SEM (n = 3) and are representative of at least two independent experiments. Differences between groups were evaluated by ANOVA followed by Bonferroni’s post hoc test. **P < 0.01 vs. indicated groups. Adenoviral vectors encoding PHD3 shRNA, sh-R PHD3(WT), or sh-R PHD3(ΔPH) were used for experiments.

Figure 4

PHD3 interacts with CITED2 and GCN5 and regulates PGC-1α-induced gluconeogenesis. (A) Effects of PHD3 depletion on phosphorylation of various PKA substrates induced by pCPT-cAMP (100 μM for 0, 10, and 30 min) in primary hepatocytes. (B) Mouse hepatocytes with or without FLAG-tagged GCN5 expression and with or without PHD3 knockdown were exposed to 100 μM pCPT-cAMP for 30 min or left untreated, and then subjected to immunoprecipitation with antibodies against phosphorylated PKA substrates followed by immunoblot analysis with anti-DYKDDDDK antibody. (C,D) Immunoprecipitation and immunoblot analyses of the interaction between FLAG-PHD3 and HA-CITED2 (C), FLAG-PHD3 and Myc-GCN5, and PHD3 and FLAG-GCN5 (D) in AD-293 cells. (E) Effect of PHD3 depletion on the interaction between FLAG-GCN5 and HA-CITED2 in AML12 cells. (F,G) Effects of PHD3 depletion on PGC-1α-induced gluconeogenic gene expression (F) and glucose production (G) in primary mouse hepatocytes with or without FLAG-PGC-1α expression in the absence of pCPT-cAMP. (H) Immunoprecipitation and immunoblot analyses of PGC-1α acetylation in primary hepatocytes expressing FLAG–PGC-1α with or without PHD3 depletion. α-Tubulin served as the loading control for immunoblotting. *Non-specific. Complete immunoblots are presented in Supplementary Fig. S7. All quantitative data are shown as mean ± SEM (n = 3 (F,G)) and are representative of at least two independent experiments. Differences between groups were evaluated by ANOVA followed by Bonferroni’s post hoc test (F,G). **P < 0.01 vs. indicated groups. Adenoviral vectors encoding PHD3 shRNA, control shRNA, FLAG-GCN5, HA-CITED2, and FLAG-PGC-1α were used for experiments.

Figure 5

PHD3 depletion impairs insulin signalling associated with NF-κB and JNK activation. (A) Effects of PHD3 depletion on insulin-induced (10 or 100 nM, 10 min) phosphorylation of Akt at Thr³⁰⁸ and Ser⁴⁷³, GSK-3α/β at Ser^21/9, and extracellular signal-regulated kinase (ERK)1/2 at Thr²⁰²/Tyr²⁰⁴, and total protein levels of Akt, GSK-3β and ERK1/2 in mouse hepatocytes, as determined by immunoblotting. (B) Effects of shRNA-mediated PHD3 knockdown on Srebf1c and Dgat1 mRNA expression in primary mouse hepatocytes with or without exposure to 10 nM insulin for 6 h. (C) Mouse hepatocytes with or without PHD3 depletion were exposed to 10 nM insulin for 1 min or left untreated, and then subjected to immunoblot analysis with an antibody specific to Tyr1146-phosphorylated IR β subunit (IRβ). Cells were also subjected to immunoprecipitation with antibodies against IRS-1 or -2, followed by immunoblot analysis with antibodies against phosphorylated tyrosine (αPY), PI3K p85 subunit, or IRS-1 or -2. (D) Effects of PHD3 depletion on Irs1 and Irs2 mRNA levels in primary mouse hepatocytes with or without exposure to 100 μM pCPT-cAMP for 6 h. (E) Immunoblot analysis of the effects of enforced expression of shRNA-resistant PHD3(WT) or PHD3(ΔPH) on insulin-induced (10 nM, 10 min) phosphorylation of Akt at Thr³⁰⁸ and Ser⁴⁷³ in primary mouse hepatocytes with or without PHD3 depletion. (F,G) Effects of PHD3 depletion on phosphorylation of NF-κB p65 subunit at Ser⁵³⁶ (F) and JNK at Thr¹⁸³/Tyr¹⁸⁵ (G) in mouse hepatocytes with or without exposure to LPS (100 ng/ml) (F) or TNF-α (20 ng/ml) (G) for indicated times. α-Tubulin served as the loading control for immunoblotting. Complete immunoblots are presented in Supplementary Fig. S7. Quantitative data are shown as mean ± SEM (n = 3 (B,D)) and are representative of at least two independent experiments. Differences between groups were evaluated by ANOVA followed by Bonferroni’s post hoc test (B,D). **P < 0.01 vs. indicated groups. Adenoviral vectors encoding PHD3 shRNA, sh-R PHD3(WT), or sh-R PHD3(ΔPH) were used for experiments.

Figure 6

PHD3 represses NF-κB-mediated transcription of proinflammatory genes independent of prolyl hydroxylase activity. (A,B) Effects of shRNA-mediated PHD3 knockdown on proinflammatory gene expression in primary mouse hepatocytes with or without exposure to 100 ng/ml LPS (A) or 20 ng/ml TNF-α (B) for 2 h, as detected by qRT-PCR. (C) Effects of ectopic expression of shRNA-resistant PHD3(WT) or PHD3(ΔPH) on IL-6 and iNOS gene expression in PHD3-depleted primary mouse hepatocytes in the presence of LPS (100 ng/ml, 2 h), as detected by qRT-PCR. Quantitative data are shown as mean ± SEM (n = 3) and are representative of at least two independent experiments. Differences between groups were evaluated by ANOVA followed by Bonferroni’s post hoc test. **P < 0.01 vs. indicated groups. Adenoviral vectors encoding PHD3 shRNA, sh-R PHD3(WT), or sh-R PHD3(ΔPH) were used for experiments.

Figure 7

PHD3 depletion potentiates IL-6–STAT3 signalling and the PERK–ATF4 branch of the UPR^ER pathway independent of prolyl hydroxylase activity. (A,B) Effects of shRNA-mediated PHD3 knockdown on Tyr⁷⁰⁵ phosphorylation and STAT3 protein level in mouse hepatocytes without (A) or with (B) exposure to 100 ng/ml LPS for indicated times, as detected by immunoblotting. The lines in (B) indicate the deletion of non-relevant bands from the blots. (C) Time course analysis of IL-6 mRNA expression in mouse hepatocytes with or without PHD3 depletion exposed to 100 ng/ml LPS for indicated times. (D) Effects of ectopic expression of shRNA-resistant PHD3(WT) or PHD3(ΔPH) on Atf4 and Chop gene expression induced by PHD3 depletion in primary mouse hepatocytes, as detected by qRT-PCR. (E) Effect of PHD3 depletion on Thr⁹⁸⁰ phosphorylation of PERK and total PERK, XBP1s, XBP1u, PHD3, and α-tubulin levels in whole cell lysates, and total amount of nuclear ATF4 and histone H3 in AML12 cells with or without exposure to 5 μg/ml tunicamycin for 24 h, as determined by immunoblotting. Histone H3 and α-tubulin served as loading controls for immunoblot analyses of the nuclear fraction and whole cell lysates, respectively. *Non-specific. Complete immunoblots are presented in Supplementary Fig. S7. Quantitative data are shown as mean ± SEM (n = 3 (C,D)) and are representative of at least two independent experiments. Differences between groups were evaluated by ANOVA followed by Bonferroni’s post hoc test. **P < 0.01 vs. indicated groups. Adenoviral vectors encoding PHD3 shRNA, sh-R PHD3(WT), or sh-R PHD3(ΔPH) were used for experiments.

Figure 8

Proposed mechanism by which PHD3 depletion leads to decreased gluconeogenesis, impaired insulin signalling, and enhanced stress signalling in hepatocytes. PY: phosphorylation of tyrosine residues.