DGAT2 inhibition blocks SREBP-1 cleavage and improves hepatic steatosis by increasing phosphatidylethanolamine in the ER

Abstract

Diacylglycerol acyltransferase 2 (DGAT2) catalyzes the final step of triglyceride (TG) synthesis. DGAT2 deletion in mice lowers liver TGs, and DGAT2 inhibitors are under investigation for the treatment of fatty liver disease. Here, we show that DGAT2 inhibition also suppressed SREBP-1 cleavage, reduced fatty acid synthesis, and lowered TG accumulation and secretion from liver. DGAT2 inhibition increased phosphatidylethanolamine (PE) levels in the endoplasmic reticulum (ER) and inhibited SREBP-1 cleavage, while DGAT2 overexpression lowered ER PE concentrations and increased SREBP-1 cleavage in vivo. ER enrichment with PE blocked SREBP-1 cleavage independent of Insigs, which are ER proteins that normally retain SREBPs in the ER. Thus, inhibition of DGAT2 shunted diacylglycerol into phospholipid synthesis, increasing the PE content of the ER, resulting in reduced SREBP-1 cleavage and less hepatic steatosis. This study reveals a new mechanism that regulates SREBP-1 activation and lipogenesis that is independent of sterols and SREBP-2 in liver.

Keywords: DGAT2; PE; SREBPs; diacylglycerol acyltransferase 2; hepatic steatosis; lipogenesis; phosphatidylethanolamine.

Figure 1.. DGAT2 inhibition suppresses nSREBP-1 and reduces SREBP-1-regulated lipid synthesis and secretion

(A) Relative mRNA levels of genes involved in lipogenesis in livers of C57BL/6J mice that were fed chow or chow diet supplemented with iDgat2 (n = 6 per group). Total RNA was extracted from livers of control and iDgat2-treated C57BL/6J mice described in Table 1 and subjected to quantitative real-time PCR analysis. Expressions of the genes were normalized to cyclophilin. (B) Immunoblot analysis of SREBP protein levels in C57BL/6J mice that were fed chow or chow diet supplemented with iDgat2. Membrane and nuclear proteins were prepared from individual livers, and immunoblot analysis was performed as described in the STAR Methods. Precursor SREBPs (P) were evaluated using membrane protein and activated nuclear forms of SREBPs (N) were measured using nuclear protein. Calnexin and LSD1 were used as loading controls for membrane and nuclear proteins, respectively. (C) Protein intensities of immunoblots from (B) were quantified, and the intensities of precursor (P) and nuclear (N) SREBPs were normalized to calnexin and LSD1, respectively. (D) Liver lipid secretion in C57BL/6J mice that were fed chow or chow diet supplemented with iDgat2 (n = 5 per group, 9 weeks of age). Male C57BL/6J mice were fed chow or chow diet supplemented with iDgat2 (0.004%) for 7 days. Mice were fasted for 4 h prior to the study. (E) Plasma TG secretion rates were calculated for each mouse from the linear regression analysis of the time vs. TG concentrations. Data are the mean ± SD. Statistical significance was assessed by two-tailed Student’s t test, *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 2.. DGAT2 inhibition suppresses hepatic nSREBP-1 in TghSREBP-1c rats

(A) Relative mRNA levels of genes involved in lipogenesis in livers from TghSREBP-1c rats that were fed chow or chow diet supplemented with iDgat2 (n = 5–6 per group). Total RNA was extracted from livers of control and iDgat2-treated TghSREBP-1c rats described in Table 1 and subjected to quantitative real-time PCR analysis. Expressions of the genes were normalized to cyclophilin. (B) Immunoblot analysis of SREBP-1 protein in TghSREBP-1c rats that were fed chow or chow diet supplemented with iDgat2. Membrane and nuclear proteins were prepared and subjected to immunoblot analysis as described in the STAR Methods. Precursor SREBPs (P) were measured in membrane protein and activated nuclear forms of SREBPs (N) were measured in the nuclear fractions. SREBP-1 denotes the total SREBP-1 (rat endogenous and human transgene). HA denotes the HA-tagged transgenic human SREBP-1 protein. Calnexin and LSD1 were used as loading controls for membrane and nuclear proteins, respectively. (C) Protein intensities of immunoblots from (B) were quantified, and the intensities of precursor (P) and nuclear (N) SREBPs were normalized to calnexin and LSD1, respectively. Data are presented as mean ± SD. Statistical significance was assessed by two-tailed Student’s t test, *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 3.. Lipidomic analysis of liver ER fractions from animals treated with iDgat2

(A–C) Lipid species of ER fractions prepared from iDgat2-treated TghSREBP-1c rats (n = 5 for control and n = 6 for iDgat2), C57BL/6J mice (n = 6 for control and n = 5 for iDgat2), and ob/ob mice (n = 6 per group). Lipids in the liver ER fractions of animals treated with iDgat2 were measured as described. Each individual value was normalized to the average value of the control group and shown as arbitrary units (a.u.)/μg protein. (D and E) Fatty acid compositions of PC and PE in liver ER fractions of C57BL/6J mice. Data are presented as mean ± SD. Statistical significance was assessed by two-tailed Student’s t test, *p < 0.05, **p < 0.01.

Figure 4.. In vitro enrichment of PE in the ER suppresses SREBP-1 activation

Primary hepatocytes from TghSREBP-1c rats or Hepa-1c1c7 mouse hepatoma cells were treated with control, low-PE-containing liposomes, or high-PE-containing liposomes for 3 h as described in the STAR Methods. (A) ER fractions and whole-cell membrane fractions were prepared from TghSREBP-1c rat primary hepatocytes, and PE levels in each fraction were measured. (B) Immunoblot analysis of precursor and nuclear forms of hSREBP-1c in primary hepatocytes treated with liposomes containing different concentrations of PE. Membrane and nuclear proteins were prepared from the primary hepatocytes and subjected to immunoblot analysis. (C) Immunoblot analysis of precursor and nuclear forms of SREBP-1 in Hepa-1c1c7 cells treated with liposomes containing different concentrations of PE. Whole-cell lysates were prepared from the Hepa-1c1c7 cells, and SREBP-1 levels were evaluated by immunoblot analysis. P, SREBP-1 precursors; N, SREBP-1-activated nuclear form.

Figure 5.. Insig is not required for the suppression of SREBP-1 cleavage by PE

Immunoblot analysis of precursor and nuclear forms of SREBP-1 in different cellular conditions. (A) SV589 cells were cultured in DMEM with 5% lipoprotein-deficient serum supplemented with 0, 10, 20, or 50 μM ethanolamine (EthA) overnight. ALLN was added to the cells 2 h before the cells were harvested for immunoblot analysis. (B) Control SV589 cells or SV589 cells that lack Insig-1 and Insig-2 were cultured in DMEM with 5% lipoprotein-deficient serum supplemented with vehicle (control, lanes 1 and 5), 100 μM palmitate (Palm, lanes 2 and 6), 100 μM EthA (lanes 3 and 7), or 100 μM of both Palm and EthA (lanes 4 and 8) overnight. ALLN was added to the cells 2 h before the cells were harvested. Protein was prepared from whole-cell lysates and applied to SDS-PAGE for immunoblot analysis. P, SREBP precursors; N, SREBP-activated nuclear forms. Nuclear forms of SREBP-1 in (B) high-exposure mode (upper) and low exposure mode (lower).

Figure 6.. Liver ER PE concentrations in DGAT2-deficient mice in the presence or absence of Insigs

DGAT2 and/or Insigs were knocked out using sgRNAs packaged in AAV-DJ and injected into Cas9-expressing mice. (A and B) Relative mRNA levels of Dgat2, Insig-1, Insig-2a, Insig-2b, and lipogenesis genes in Dgat2 and/or Insig-1 and −2 (Insig) hepatocyte-specific knockout mice (n = 5 per group). Total RNA was extracted from livers of the mice 6 weeks after the AAV injection and subjected to quantitative real-time PCR analysis. (C) Immunoblot analysis of precursor and nuclear forms of SREBP-1 in Dgat2 and/or Insig knockout mice. Membrane and nuclear proteins from individual livers were prepared and loaded to 8% SDS-PAGE (3 samples from each group were loaded for the blot on the left panel, and the other 2 samples from each group were loaded for the blot on the right panel) and subjected to immunoblot analysis. Precursor SREBP-1 (P) and Insig proteins were detected in membrane fractions, and nSREBP-1 (N) was measured in the nuclear protein fractions. Calnexin and LSD1 were used as loading controls for membrane and nuclear proteins, respectively. (D) Protein intensities of immunoblots from (C) were quantified, and intensities of the precursor (P) and nuclear (N) SREBP-1 were normalized to calnexin and LSD1, respectively. # denotes a nonspecific bond detected by Insig antibody. (E) PE and PC content in the ER fractions were measured using LC-MS/MS. Data are presented as mean ± SD. Statistical significance was assessed by two-tailed Student’s t test. *p < 0.05, **p < 0.01, ***p < 0.001.