Cytoplasmic Endonuclease G promotes nonalcoholic fatty liver disease via mTORC2-AKT-ACLY and endoplasmic reticulum stress

Abstract

Endonuclease G (ENDOG), a nuclear-encoded mitochondrial intermembrane space protein, is well known to be translocated into the nucleus during apoptosis. Recent studies have shown that ENDOG might enter the mitochondrial matrix to regulate mitochondrial genome cleavage and replication. However, little is known about the role of ENDOG in the cytosol. Our previous work showed that cytoplasmic ENDOG competitively binds with 14-3-3γ, which released TSC2 to repress mTORC1 signaling and induce autophagy. Here, we demonstrate that cytoplasmic ENDOG could also release Rictor from 14-3-3γ to activate the mTORC2-AKT-ACLY axis, resulting in acetyl-CoA production. Importantly, we observe that ENDOG could translocate to the ER, bind with Bip, and release IRE1a/PERK to activate the endoplasmic reticulum stress response, promoting lipid synthesis. Taken together, we demonstrate that loss of ENDOG suppresses acetyl-CoA production and lipid synthesis, along with reducing endoplasmic reticulum stress, which eventually alleviates high-fat diet-induced nonalcoholic fatty liver disease in female mice.

Fig. 1. Loss of ENDOG represses lipid accumulation in mouse livers and hepatocytes.

a–c Representative electron microscopic images and quantification of lipid size and number in Endog^+/− or Endog^−/− mouse livers after 24 hours of starvation. N: Nuclear; LD: lipid droplet; n = 8mice. d Measurement of total triglycerides in mouse livers after 24 hours of starvation. n = 8 mice. e, f Representative images of BODIPY staining and the quantitative results of lipid area in liver tissue of mice starved for 24 hours. n = 8 mice. g–i Representative images of Nile red staining, the quantitative results of lipid area, and measurement of triglycerides in WT and ENDOG KO HepG2 cells after treatment of 200 μM oleic acid for 24 h. n = 10 independent samples for Nile red staining and 6 for triglyceride measurement. j–l Representative images of Nile red staining, the quantitative results of lipid area, and measurement of triglycerides in control and ENDOG overexpressing HepG2 cells following the treatment of 200 μM oleic acid for 24 h. pK-Myc and ENDOG overexpressing plasmids were transfected for 48 h. n = 4 independent samples for Nile red staining and 5 for triglyceride measurement. Statistical significance was determined by unpaired Student’s t-test (two-tailed) in (b, c, d, f, h, i, k, l); error bars are mean ± SD. Source data and exact P values are provided in a Source data file. *P < 0.05; **P < 0.01; ***P < 0.001; n.s.: no significance.

Fig. 2. ENDOG promotes lipid synthesis via the AKT-ACLY axis.

a, b Western blot analyses of activation of AKT-ACLY in ENDOG overexpressing and knockout HepG2. pK-Myc and ENDOG overexpressing plasmids were transfected for 48 h. c Measurement of acetyl-CoA in ENDOG overexpressing / knockout HepG2. n = 6 biologically independent samples each group. d Measurement of acetyl-CoA in Endog^+/- and Endog^-/- mice livers. n = 6 mice in Endog^+/- group and 7 mice in Endog^-/- group. e–g Representative images of Nile red, the quantitative results of lipid area, and measurement of triglycerides in wild-type and ENDOG knockout HepG2 after the supplementation of 10 mM citrate for 24 h. n = 5 biologically independent samples for Nile red staining and 4 for triglyceride measurement. h–j Representative images of Nile red, the quantitative results of lipid area, and measurement of triglycerides in wild-type and ENDOG knockout HepG2 after supplementing 50 μM acetyl-CoA for 24 h. n = 6 biologically independent samples for Nile red staining and 4 for triglyceride measurement. k Western blot analyses of AKT-ACLY signaling in ENDOG overexpressing HepG2 following 50 or 100 μM LY294002 treatment for 24 h. pK-Myc and ENDOG overexpressed plasmids were transfected for 48 h. l–n Representative images of Nile red, the quantitative results of lipid area and measurement of ENDOG overexpressing HepG2 following the treatment of 200 μM oleic acid and 50 μM LY294002 for 24 h. n = 5 biologically independent samples for Nile red staining and 4 for triglyceride measurement. o Western blot analyses of AKT-ACLY signaling in wild-type and ENDOG knockout HepG2 after transfection with myr-AKT plasmid for 48 h. p–r Representative images of Nile red, the quantitative results of lipid area, and measurement of triglycerides in the indicated groups. Cells were transfected with plasmids for 24 h and then treated with 200 μM oleic acid for another 24 h. n = 5 biologically independent samples for Nile red staining and 4-6 for triglyceride measurement. Statistical significance was determined by unpaired Student’s t-test (two-tailed) in (c, d, f, g, i, j, m, n, q, r); error bars are mean ± SD. Source data and exact P values are provided in a Source data file. *P < 0.05; **P < 0.01; ***P < 0.001; n.s.: no significance.

Fig. 3. Loss of ENDOG represses HFD-induced lipid accumulation by suppressing the AKT-ACLY axis.

a Body weight of HFD-fed female mice at different ages. n = 6 mice in Endog^+/- group and 7 mice in Endog^-/- group. b, c Gross images of inguinal white adipose (iWAT), epididymal white adipose tissue (eWAT), brown adipose tissue (BAT) and liver of the HFD-fed female mice. d Nonfasting blood glucose (NFBG), the weight of WAT (iWAT and eWAT), BAT, and liver in HFD-fed female mice. n = 5 mice in Endog^+/- group and 7 mice in Endog^-/- group. e Liver H&E staining in HFD female mice. f Steatosis score of HFD female mice liver. n = 6 mice in Endog^+/- group and 7 mice in Endog^-/- group. g, h Measurement of triglycerides in the liver and free fatty acid in serum. n = 6 mice in Endog^+/- group and 7 mice in Endog^-/- group. i–k Body weight, NFBG, the weight of WAT, BAT, and liver in ENDOG liver-specific knockout female mice. n = 5 mice in Endog^+/- group and 6 mice in Endog^-/- group. l, m Liver H&E staining and steatosis score. n = 8 mice each group. n Measurement of triglycerides in the liver. n = 6 mice each group. o Measurement of free fatty acid in serum. n = 12 mice in Endog^+/- group and 9 mice in Endog^-/- group. p, q Western blot analyses of AKT-ACLY signaling in Endog^+/- and Endog^-/- HFD mice livers. n = 6 mice in Endog^+/- group and 7 mice in Endog^-/- group. r, s Western blot analyses of AKT-ACLY signaling in Endog^flox and Endog^LKO HFD mice livers. n = 6 mice in Endog^+/- group and 7 mice in Endog^-/- group. Statistical significance was determined by unpaired Student’s t-test (two-tailed) in (a, d, f, g, h, i, j, k, m, n, o, q, s); error bars are mean ± SD. Source data and exact P values are provided in a Source data file. *P < 0.05; **P < 0.01; ***P < 0.001; n.s.: no significance.

Fig. 4. ENDOG activates the mTORC2-AKT-ACLY axis by competitively binding with 14-3-3γ.

a Co-IP analyses. HepG2 cells were cotransfected with 14-3-3γ-Myc and ENDOG-Flag/ pLVX3-Flag plasmids for 48 h. b Co-IP analyses. HepG2 cells were transfected with ENDOG-Flag or pLVX3-Flag plasmids for 48 h. c Endogenous Co-IP analyses in wild-type and ENDOG knockout HepG2 cells. d, e ENDOG releases to the cytoplasm in HepG2 after being treated with 200 μM oleic acid for 24 h. Cyto, cytoplasm; Mito, mitochondria. n = 4 biologically independent samples. f Representative images of costaining of Tim23 (mitochondrial inner membrane protein) and ENDOG in HepG2 after being treated with 200 μM oleic acid for 24 h. g Endogenous Co-IP analyses in wild-type HepG2 cells after being treated with 200 μM oleic acid for 24 h. h, i Western blot and quantitative results of AKT-ACLY signaling after transfection with ENDOG and 14-3-3γ / ENDOG plasmids together for 48 h. n = 4 biologically independent samples. j–l Representative images of Nile red, the quantitative results of lipid area and measurement of triglycerides after transfection with ENDOG and 14-3-3γ / ENDOG plasmids together for 24 h and then treated with 200 μM oleic acid for another 24 h. n = 10 biologically independent samples for Nile red staining and 4 for triglyceride measurement. Three experiments were repeated independently with similar results in (a, b, c, f, g). Statistical significance was determined by unpaired Student’s t-test (two-tailed) in (i, k, l); error bars are mean ± SD. Source data and exact P values are provided in a Source data file. *P < 0.05; **P < 0.01; ***P < 0.001.

Fig. 5. ENDOG promotes lipid synthesis through the activation of ER stress.

a qPCR results of lipid metabolism genes in wild-type and ENDOG knockout HepG2 cells. n = 4 biologically independent samples. b, c Western blot analyses of lipid synthesis proteins in HepG2. WT, wild-type, KO: ENDOG knockout; pK-Myc or ENDOG plasmids were transfected for 48 h. d GSEA shows endoplasmic reticulum-related gene sets are rich in ENDOG overexpressed HepG2 cells. e Western blot analyses of ER stress-related proteins and lipid synthesis proteins in wild-type and ENDOG knockout HepG2 cells following treatment with 2 μg/ml tunicamycin (TG) and 1 μM thapsigargin (TG) for 24 h. f, g Representative images of Nile red and measurement of triglycerides. Wild-type and ENDOG knockout HepG2 cells following treatment with 2 μg/ml tunicamycin (TM) and 1 μM thapsigargin (TG) for 24 h. n = 4 biologically independent samples. h, i Western blots analyses of ER stress-related proteins in systemic and liver-specific ENDOG knockout female mice livers after the HFD. n = 5 mice each group. j, k Western blot analyses of ER stress-related proteins and lipid synthesis proteins in liver-specific ENDOG knockout female mice livers. Mice were fasting for 12 h and then intraperitoneally injected with tunicamycin (TM) 2 mg/kg body weight for 24 h. n = 6 mice each group. l Measurement of triglycerides in liver. Female mice were fasting for 12 h and then intraperitoneally injected with 2 mg/kg body weight for 24 h. n = 9 mice each group. Three experiments were repeated independently with similar results in (b, c, e). Statistical significance was determined by unpaired Student’s t-test (two-tailed) in (a, i, k, l); error bars are mean ± SD. Source data and exact P values are provided in a Source data file. *P < 0.05; **P < 0.01; ***P < 0.001.

Fig. 6. ENDOG activates ER stress by translocating to the ER and binding with Bip.

a ENDOG binding proteins detected by IP-MASS. Protein Acce.: Protein accession. EmPAI: exponentially modified protein abundance index. b Co-IP analyses. HepG2 cells were transfected with ENDOG-Flag plasmid for 48 h. L: long exposure; S: short exposure. c–e Representative images of costaining of ENDOG with Grp75/Calnexin/Bip after 200 μM oleic acid treatment for 24 h. f, g Representative images of in situ proximity ligation assay (PLA) and quantitative results of PLA puncta per cell. HepG2 Cells were treated with 200 μM oleic acid for 24 h. n = 12 biologically independent samples. h Subcellular fraction isolation by density gradient centrifugation. HepG2 cells were treated with 200 μM oleic acid for 24 h. ER, endoplasmic reticulum fraction; Mito, mitochondrial fraction; p, precursor ENDOG; m, mature ENDOG. i The ER location of ENDOG is analyzed by ER enrichment Kit. HepG2 cells were treated with 200 μM oleic acid for 24 h. Total, total cell lysis; ER, endoplasmic reticulum enriched fraction. j Co-IP analyses. HepG2 cells were cotransfected with Bip-Myc and pK-Myc/ENDOG plasmids for 48 h. k Endogenous Co-IP analyses in wild-type and ENDOG knockout HepG2 cells. l Endogenous Co-IP analyses. HepG2 cells were treated with 200 μM oleic acid for 24 h. Three experiments were repeated independently with similar results in (b, c, d, e, h, i, j, k, l). Statistical significance was determined by unpaired Student’s t-test (two-tailed) in (g); error bars are mean ± SD. Source data and exact P value were provided in a Source data file. ***P < 0.001.

Fig. 7. Inhibiting the release of ENDOG represses oleic acid-induced lipid accumulation.

a, b Representative images of costaining of HSP60/ENDOG and subcellular fraction isolation. HepG2 cells were pretreated with 20 μM VBIT-12 (VDAC inhibitor) for 4 h and then treated with 200 μM oleic acid for 24 h. The red puncta in the dotted circle indicate that ENDOG is released from the mitochondria. S, short time exposure; L, long time exposure. c, d Western blot and quantitative results of ER stress-related proteins, FAS/ACC, and AKT-ACLY activation. HepG2 cells were pretreated with 20 μM VBIT-12 (VDAC inhibitor) for 4 h and then treated with 200 μM oleic acid for 24 h. n = 4 biologically independent samples. e–g Measurement of triglycerides, Nile red staining, and the quantitative results of lipid area. HepG2 cells were pretreated with 20 μM VBIT-12 (VDAC inhibitor) for 4 h and then treated with 200 μM oleic acid for 24 h. n = 4 biologically independent samples for triglycerides measurement and 10 for Nile red staining. h Schematic diagram of ENDOG mediated NAFLD under the HFD chow feeding. HFD promotes the release of ENDOG from the mitochondria. Cytoplasm located ENDOG competitively binds with 14-3-3γ, which makes 14-3-3γ disassociate from Rictor and activates the mTORC2-AKT-ACLY axis, resulting in the production of acetyl-CoA. In addition, cytoplasmic ENDOG also translocates into the ER and interacts with Bip, which releases IRE1a and PERK to activate ER stress and ACC/FAS expression. Increased acetyl-CoA and lipid synthesis enzymes (ACC/FAS) enhanced lipid accumulation. Finally, increased triglyceride and ER stress promote the development of HFD-induced NAFLD. Three experiments were repeated independently with similar results in (a, b). Statistical significance was determined by unpaired Student’s t-test (two-tailed) in (d, e, g); error bars are mean ± SD. Source data and exact P value were provided in a Source data file. *P < 0.05; **P < 0.01; ***P < 0.001.