Liver lipid droplet cholesterol content is a key determinant of metabolic dysfunction-associated steatohepatitis

Abstract

Metabolic dysfunction-associated steatohepatitis (MASH) represents a progressive form of steatotic liver disease which increases the risk for fibrosis and advanced liver disease. The accumulation of discrete species of bioactive lipids has been postulated to activate signaling pathways that promote inflammation and fibrosis. However, the key pathogenic lipid species is a matter of debate. We explored candidates using various dietary, molecular, and genetic models. Mice fed a choline-deficient L-amino acid-defined high-fat diet (CDAHFD) developed steatohepatitis and manifested early markers of liver fibrosis associated with increased cholesterol content in liver lipid droplets within 5 d without any changes in total liver cholesterol content. Treating mice with antisense oligonucleotides against Coenzyme A synthase (Coasy) or treatment with bempedoic acid or atorvastatin decreased liver lipid droplet cholesterol content and prevented CDAHFD-induced MASH and the fibrotic response. All these salutary effects were abrogated with dietary cholesterol supplementation. Analysis of human liver samples demonstrated that cholesterol in liver lipid droplets was increased in humans with MASH and liver fibrosis and was higher in PNPLA3 I148M (variants rs738409) than in HSD17B13 variants (rs72613567). Together, these data identify cholesterol in liver lipid droplets as a critical mediator of MASH and demonstrate that Coenzyme A synthase knockdown and bempedoic acid are therapeutic approaches to reduce liver lipid droplet cholesterol content and thereby prevent the development of MASH and liver fibrosis.

Keywords: Coenzyme A synthase; cholesterol; lipid droplet; metabolic dysfunction–associated steatohepatitis (MASH); phosphatidylcholine.

Fig. 1.

Choline-deficient L-amino acid–defined high-fat diet acutely induces hepatic lipid accumulation followed by inflammation and fibrosis. (A) Study design. C57BL/6J mice were divided into RC fed mice or CDAHFD fed mice. CDAHFD was provided for 1, 2, 3, 4, and 5 d, respectively. (B) Plasma ALT, AST, total cholesterol, and triglycerides levels. ALT and AST increased in CDAHFD-fed mice. Plasma total cholesterol and triglycerides decreased in CDAHFD-fed mice. (C) Liver sections were stained with HE, CD68, BODIPY, and Filipin. CDAHFD induced liver lipid droplets time-dependently. CD68 staining revealed crown-like structures on day 5. BODIPY and Filipin staining demonstrated that free cholesterol was included in liver lipid droplets. (D) The number of crown-like structures increased in the liver of CDAHFD-fed mice time-dependently. (E) Total cholesterol, triglycerides, and fatty acids content in whole liver lysate. Total cholesterol did not change. Triglycerides and fatty acids increased in CDAHFD-fed mice time-dependently. (F) The levels of phospholipids in the liver. The concentration of PC in the plasma membrane exhibited a decline starting on day 1. The level of PE in the plasma membrane did not change. In liver lipid droplets, PC and PE levels increased from day 1. (G) The ratio of PC to PE in the plasma membrane and lipid droplet in the liver, respectively. PC/PE ratio in CDAHFD-fed mice was lower than RC-fed mice. (H) Total cholesterol and triglycerides in the liver lipid droplet increased in CDAHFD-fed mice from day 1. Data are presented as mean ± SEM. Groups were compared by Unpaired one-sided Student’s t test.

Fig. 2.

Choline-deficient L-amino acid–defined high-fat diet induces steatosis via enhanced hepatic uptake of chylomicron remnants. (A) Study design. C57BL/6J mice were divided into RC-fed mice or CDAHFD-fed mice for 1 wk. (B) Liver sections were stained with HE, BODIPY, Filipin, Cathepsin D, Terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) and CD68. CDAHFD induced macrovesicular steatosis. BODIPY and Filipin staining demonstrated that free cholesterol was included in liver lipid droplets. Leakage of cathepsin D into cytoplasm indicated disruption of the lysosomal membrane. TUNEL staining revealed dead hepatocytes. CD68 staining revealed crown-like structures. (C) Representative transmission electron microscopy images of 1 wk CDAHFD-fed mice’s liver. Transmission electron microscopy demonstrated lipid-laden lysosome, disruption of lysosomal membrane, disruption of cell membrane of hepatocyte and macrophage engulfing lipid droplets. (D) RT-qPCR analysis of liver tissues. CDAHFD 1 wk feeding increased mRNA expression of MASH associated macrophage markers (Gpnmb, Trem2, and Lgals3), and fibrosis markers (αSMA, and Col1a1). (E) Study design. C57BL/6J mice were divided into a CDAHFD-fed mice, CDAHFD with 2% cholesterol-fed mice, GAN diet-fed mice. (F) Plasma ALT, AST, total cholesterol, and triglycerides levels. CDAHFD with 2% cholesterol-fed mice showed increased ALT and AST levels compared to CDAHFD-fed mice and GAN-diet fed mice. GAN diet-fed mice showed increased plasma total cholesterol and triglyceride levels compared to CDAHFD with 2% cholesterol-fed mice. (G) Liver sections were stained with HE, CD68, BODIPY, and Filipin. CDAHFD and CDAHFD with 2% cholesterol feeding for 1 wk induced steatosis. GAN diet feeding for 1 wk did not cause apparent steatosis. (H) The number of crown-like structures in the liver increased in CDAHFD with 2% cholesterol-fed mice compared to CDAHFD-fed mice and GAN diet-fed mice. (I) The metabolites levels in liver lipid droplet. Total cholesterol and triglycerides increased in CDAHFD-fed mice and CDAHFD with 2% cholesterol-fed mice compared to GAN diet-fed mice. In addition, total cholesterol in CDAHFD with 2% cholesterol-fed mice was higher than that of CDAHFD-fed mice. The level of PC and PE increased in CDAHFD-fed mice and CDAHFD with 2% cholesterol-fed mice compared to GAN diet-fed mice. The ratio of PC to PE decreased in CDAHFD-fed mice and CDAHFD with 2% cholesterol-fed mice compared to GAN diet-fed mice. Data are presented as mean ± SEM. Groups were compared by one-way ANOVA followed by Tukey’s multiple comparisons test.

Fig. 3.

Bempedoic acid alleviates liver inflammation and fibrosis due to choline-deficient L-amino acid–defined high-fat diet via decreasing cholesterol content in liver lipid droplets. (A) Study design. C57BL/6J mice were divided into a control group and bempedoic acid–treated group. Both groups were fed a CDAHFD for 1 wk. (B) ALT and AST in bempedoic acid (BemA) treated mice decreased compared to control mice. (C) Liver sections were stained with HE, CD68, BODIPY, and Filipin. BemA treatment prevented steatosis. CD68 staining revealed decreased crown-like structures. BODIPY and Filipin staining demonstrated decreased free cholesterol in liver lipid droplet. (D) The number of crown-like structures decreased in BemA-treated mice liver. (E) The metabolites levels in liver lipid droplet. Total cholesterol and triglycerides decreased in BemA-treated mice. The level of PC decreased in BemA-treated mice. PE did not show a difference. The ratio of PC to PE did not show a difference. (F) RT-qPCR analysis of liver tissues demonstrated a decreased mRNA expression of MASH-associated macrophage markers (Gpnmb, Trem2, and Lgals3), and fibrosis marker (Col1a1) in BemA-treated mice. Data are presented as mean ± SEM. Groups were compared by Unpaired one-sided Student’s t test.

Fig. 4.

COASY knockdown alleviates liver inflammation and fibrosis due to Choline-deficient L-amino acid–defined high-fat diet via decreasing cholesterol in liver lipid droplets. (A) CoA biosynthesis pathway. COASY converts 4′-Phosphopantetheine into CoA by adenylation and phosphorylation. CoA subsequently catalyzes the synthesis of acetyl CoA. Acetyl CoA functions as a substrate throughout the cholesterol synthesis or fatty acids synthesis pathway. (B) Study design. C57BL/6J mice were divided into GalNAc control ASO-treated mice and GalNAc Coasy ASO-treated mice. Both groups were fed a CDAHFD for 1 wk. (C) RT-qPCR analysis of liver tissues confirmed the knockdown effect of Coasy ASO. Consistently, Coasy ASO treatment decreased hepatic CoA species, including CoA, Acetyl CoA, Malonyl CoA, and Acyl CoA. (D) ALT and AST in Coasy ASO-treated mice decreased compared to control ASO-treated mice. (E) Liver sections were stained with HE, CD68, BODIPY, and Filipin. Coasy ASO treatment alleviated steatosis. Accordingly, CD68 staining revealed decreased crown-like structures. BODIPY and Filipin staining demonstrated decreased free cholesterol in liver lipid droplets. (F) The number of crown-like structures decreased in Coasy ASO-treated mice liver. (G) The metabolites levels in liver lipid droplet. Total cholesterol and triglycerides decreased in Coasy ASO-treated mice. The level of PC and PE decreased in Coasy ASO-treated mice. The ratio of PC to PE remained unchanged. (H) RT-qPCR analysis of liver tissues demonstrated a decreased mRNA expression of MASH-associated macrophage markers (Gpnmb, Trem2, and Lgals3) and fibrosis markers (αSMA, and Col1a1) in Coasy ASO-treated mice. Data are presented as mean ± SEM. Groups were compared by Unpaired one-sided Student’s t test.

Fig. 5.

Cholesterol supplementation negates the protective effect of Coasy ASO on a Choline-deficient L-amino acid–defined high-fat diet 1-wk mice model. (A) Study design. C57BL/6J mice were divided into CDAHFD fed mice and CDAHFD with 2% cholesterol fed mice. Both groups were injected with GalNAc Coasy ASO. (B) Cholesterol (Chol) dietary supplementation increased plasma ALT and AST levels compared to control (Ctr) mice. (C) Liver sections were stained with HE, CD68, BODIPY, and Filipin. Dietary cholesterol supplementation induced steatosis. CD68 staining revealed increased crown-like structures. BODIPY and Filipin staining demonstrated increased free cholesterol in liver lipid droplet. (D) The number of crown-like structures increased in dietary cholesterol-supplemented mice liver. (E) The metabolites levels in liver lipid droplet. Total cholesterol and triglycerides increased in cholesterol-overload mice. The level of PC and PE increased in cholesterol-supplemented mice. The ratio of PC to PE remained unchanged. (F) RT-qPCR analysis of liver tissues showed an elevated mRNA expression of MASH-associated macrophage markers (Gpnmb, Trem2, and Lgals3) as well as fibrosis markers (αSMA and Col1a1) in cholesterol-overload mice. Data are presented as mean ± SEM. Groups were compared by Unpaired one-sided Student’s t test. (G) Summary of four studies of 1-wk CDAHFD studies. (H) Enrichment analysis of comprehensive proteome analysis. Three comparisons were made: (1) CDAHFD-fed mice (n = 6) vs. RC-fed mice (n = 6); (2) Coasy ASO-CDAHFD-treated mice (n = 6) vs. Control ASO-CDAHFD-treated mice (n = 6); and (3) Coasy ASO-CDAHFD with 2% cholesterol-treated mice (n = 6) vs. Coasy ASO-CDAHFD-treated mice (n = 6). (I) Hepatic protein expression of targets of SREBP1c, MLXIPL, and SREBP2.

Fig. 6.

HSD17B13 knockdown alleviates MASH due to CDAHFD via decreasing cholesterol in liver lipid droplets and Patatin-like phospholipase domain-containing protein 3 I148M knock-in aggravates MASH due to CDAHFD via increasing in liver lipid droplets. (A) Study design. C57BL/6J mice were divided into GalNAc control ASO-treated mice and GalNAc Hsd17b13 ASO-treated mice. Both groups were fed a CDAHFD for 1 wk. (B) RT-qPCR analysis of liver tissues confirmed the knockdown effect of Hsd17b13 ASO. (C) Plasma ALT and AST levels in Hsd17b13 ASO-treated mice decreased compared to control mice. (D) Liver sections were stained with HE, and CD68. Hsd17b13 ASO did not significantly affect steatosis. CD68 staining revealed decreased crown-like structures. (E) The number of crown-like structures decreased in Hsd17b13 ASO-treated mice liver. (F) The metabolites levels in liver lipid droplet. Total cholesterol decreased in Hsd17b13 ASO-treated mice. The level of triglycerides, PC and PE did not change. The ratio of PC to PE did not change. (G) RT-qPCR analysis of liver tissues demonstrated a decreased mRNA expression of MASH-associated macrophage markers (Gpnmb, Trem2, and Lgals3) and fibrosis markers (αSMA and Col1a1) in Hsd17b13 ASO-treated mice. (H) Study design. Patatin-like phospholipase domain-containing protein 3 (Pnpla3) I148M knock-in mice and wild-type mice were fed with a CDAHFD for 1 wk. (I) ALT and AST in Pnpla3 I148M knock-in mice increased compared to wild-type mice. (J) Liver sections were stained with HE, and CD68. Pnpla3 I148M showed macrovesicular steatosis compared to wild-type mice. CD68 staining revealed increased crown-like structures. (K) The number of crown-like structures increased in Pnpla3 I148M knock-in mice liver. (L) The metabolite levels in the liver lipid droplet. Total cholesterol and triglycerides increased in Pnpla3 I148M knock-in mice. The level of PC and PE did not change. The ratio of PC to PE did not change. (M) RT-qPCR analysis of liver tissues demonstrated an increased mRNA expression of MASH-associated macrophage markers (Gpnmb, Trem2, and Lgals3), and fibrosis markers (αSMA, and Col1a1) in Pnpla3 I148M knock-in mice. Data are presented as mean ± SEM. Groups were compared by Unpaired one-sided Student’s t test.

Fig. 7.

Lipid droplet cholesterol is a key factor in explaining how common genetic variants both cause and protect against MASH. (A) The metabolite levels in human liver lipid droplets. Total cholesterol, triglycerides, PC and PE in liver lipid droplets were highest in human MASH with fibrosis compared to other groups. There were no differences in the ratio of PC to PE. Data are presented as mean ± SEM. Groups were compared by one-way ANOVA followed by Tukey’s multiple comparisons test. (B) Comparison of the metabolite in human liver lipid droplet levels between PNPLA3 I148M variants without HSD17B13 splice site variants (high genetic risk) and HSD17B13 splice variants (rs72613567) without PNPLA3 I148M variants (low genetic risk). (C) The CDAHFD induces the formation of large lipid droplets, including cholesterol, in the liver. Cholesterol accumulation in lipid droplets leads to lysosome-dependent hepatocyte death due to membrane leakage, which results from a relative PC deficiency on the lipid droplet surface. This process subsequently triggers macrophage crown-like formation and stellate cell activation. Cholesterol-lowering drugs, such as statins and bempedoic acid, exhibit protective effects in the CDAHFD model by reducing cholesterol levels in lipid droplets. Additionally, we identified COASY as a potential therapeutic target for decreasing cholesterol in lipid droplets. Furthermore, cholesterol accumulation in liver lipid droplets contributes to MASH and is influenced by genetic variants, including the PNPLA3 I148M variant and the HSD17B13 splice site variant.