Hepatocyte mARC1 promotes fatty liver disease

Abstract

Background & aims: Non-alcoholic fatty liver disease (NAFLD) has a prevalence of ∼25% worldwide, with significant public health consequences yet few effective treatments. Human genetics can help elucidate novel biology and identify targets for new therapeutics. Genetic variants in mitochondrial amidoxime-reducing component 1 (MTARC1) have been associated with NAFLD and liver-related mortality; however, its pathophysiological role and the cell type(s) mediating these effects remain unclear. We aimed to investigate how MTARC1 exerts its effects on NAFLD by integrating human genetics with in vitro and in vivo studies of mARC1 knockdown.

Methods: Analyses including multi-trait colocalisation and Mendelian randomisation were used to assess the genetic associations of MTARC1. In addition, we established an in vitro long-term primary human hepatocyte model with metabolic readouts and used the Gubra Amylin NASH (GAN)-diet non-alcoholic steatohepatitis mouse model treated with hepatocyte-specific N-acetylgalactosamine (GalNAc)-siRNA to understand the in vivo impacts of MTARC1.

Results: We showed that genetic variants within the MTARC1 locus are associated with liver enzymes, liver fat, plasma lipids, and body composition, and these associations are attributable to the same causal variant (p.A165T, rs2642438 G>A), suggesting a shared mechanism. We demonstrated that increased MTARC1 mRNA had an adverse effect on these traits using Mendelian randomisation, implying therapeutic inhibition of mARC1 could be beneficial. In vitro mARC1 knockdown decreased lipid accumulation and increased triglyceride secretion, and in vivo GalNAc-siRNA-mediated knockdown of mARC1 lowered hepatic but increased plasma triglycerides. We found alterations in pathways regulating lipid metabolism and decreased secretion of 3-hydroxybutyrate upon mARC1 knockdown in vitro and in vivo.

Conclusions: Collectively, our findings from human genetics, and in vitro and in vivo hepatocyte-specific mARC1 knockdown support the potential efficacy of hepatocyte-specific targeting of mARC1 for treatment of NAFLD.

Impact and implications: We report that genetically predicted increases in MTARC1 mRNA associate with poor liver health. Furthermore, knockdown of mARC1 reduces hepatic steatosis in primary human hepatocytes and a murine NASH model. Together, these findings further underscore the therapeutic potential of targeting hepatocyte MTARC1 for NAFLD.

Keywords: Hepatic steatosis; Mendelian randomisation; NASH; Triglycerides.

Graphical abstract

Fig. 1

Statistical colocalisation of cardiometabolic traits at the MTARC1 locus. (A) Regional association plots at the MTARC1 gene locus, showing the shared genetic associations with liver enzymes (ALT and AST), lipids (TC), and liver fat accumulation. Details about the statistical analysis and source of the data are given in the Materials and Methods section. Colour key indicates the correlation, r², with the respective lead variants in the GWAS. (B) Plot showing the colocalisation posterior probabilities explained by each of the genetic variants at the extended MTARC1 locus region in the colocalisation analysis. The table shows the parameters used for HyPrColoc analysis and results from HyPrColoc. ALT, alanine aminotransferase; AST, aspartate transaminase; GWAS, genome-wide association studies; HLX, H2.0 like homeobox; HyPrColoc, Hypothesis Prioritisation in multi-trait Colocalisation; MARK1, Microtubule Affinity Regulating Kinase 1; MTARC1, mitochondrial amidoxime-reducing component 1; TC, total cholesterol.

Fig. 2

mARC1 knockdown decreases lipid accumulation and apo B secretion in PHH. Validation of (A) mRNA and (B) protein levels following MTARC1 siRNA knockdown in PHH over 17 days. (C) Workflow for culturing of PHH, siRNA transfection, and endpoint collections. (D) Lipid accumulation data following siRNA-mediated knockdown of MTARC1 and DGAT2 relative to NT siRNA, n = 25 independent experiments. (E) siRNA knockdown of MTARC1 and DGAT2 reduces lipid accumulation with and without FFA mix loading in PHH. Fluorescence images of nuclei (blue) and lipid droplets (hot) from siRNA knockdown PHH treated with FFA mix. Scale bar: 50 μm. (F) Reduced apo B expression observed following knockdown with MTARC1 siRNA relative to NT siRNA, n = 16 independent experiments. (G) Increased media TG levels observed following knockdown with MTARC1 siRNA relative to NT siRNA, n = 8 independent experiments. Data are presented as mean ± 95% CI, ∗p ≤0.033, ∗∗p ≤0.002, ∗∗∗p ≤0.001 (D and F, two-way ANOVA with Tukey’s post hoc testing within FFA mix conditions; G, paired t test). 5C, five chemicals; apo B, apolipoprotein B; FC, fold change; FFA, free fatty acid; KD, knockdown; MTARC1, mitochondrial amidoxime-reducing component 1; NT, non-targeting; PHH, primary human hepatocytes; RQ, relative quantification; siDGAT2, siRNA targeting DGAT2; siMTARC1, MTARC1 targeting siRNA; TBP, TATA-box binding protein; TG, triglycerides.

Fig. 3

Hepatocyte-specific Mtarc1 knockdown reverses steatosis in a DIO-NASH mouse model. C57BL/6JRj mice fed a high-fat, high-fructose, high-cholesterol diet for 44 weeks were randomised based on biopsy results and treated weekly with a GalNAc-siMtarc1 or PBS for 8 weeks. (A) mRNA expression for Mtarc1. (B) % Liver weight of BW. (C) Liver TG as mg/g of liver. (D) Liver TC as mg/g of liver. (E) Plasma TG (mmol/L). (F) Plasma TC (mmol/L). Data are presented as mean ± 95% CI, n = 10–14, ∗p ≤0.033, ∗∗∗p ≤0.001 (unpaired t test). BW, body weight; GalNAc, N-acetylgalactosamine; GalNAc-siMtarc1, GalNAc-conjugated siRNA targeting murine Mtarc1; Mtarc1, mitochondrial amidoxime-reducing component 1; NASH, non-alcoholic steatohepatitis; TC, total cholesterol; TG, triglycerides.

Fig. 4

Hepatocyte-specific Mtarc1 knockdown decreases markers of fibrosis in a DIO-NASH mouse model. C57BL/6JRj mice fed a high-fat, high-fructose, high-cholesterol diet for 44 weeks were randomised based on biopsy results and treated weekly with a GalNAc-siMtarc1 or PBS for 8 weeks. (A) Fibrosis as PSR, % area fraction. (B) Col1 as % area fraction. (C) α-SMA as % area fraction. (D) mRNA expression for fibrosis-associated genes; Timp1, TgfB, Col1a1, Mmp2, and Mmp9. Data are presented as mean ± 95% CI, n = 10–14, n.s., ∗p ≤0.033, ∗∗p ≤0.002, ∗∗∗p ≤0.001 (unpaired t test). α-SMA, α-smooth muscle actin; GalNAc, N-acetylgalactosamine; GalNAc-siMtarc1, GalNAc-conjugated siRNA targeting murine Mtarc1; Mtarc1, mitochondrial amidoxime-reducing component 1; NASH, non-alcoholic steatohepatitis; PSR, Picro-Sirius Red.

Fig. 5

Effect of MTARC1 siRNA in PHH and mouse samples evaluated by omics approaches. (A) RNA sequencing data showing fold change of significantly dysregulated mRNAs and (B) metabolomic data showing significantly dysregulated metabolites in PHH (green) and mouse tissue from bulk liver prep (purple). FDR ≥0.05 (empty points) or FDR <0.05 (solid points). 3-Hydroxybutyrate (purple star) was significantly reduced in both PHH and mouse samples. (C) Protein-interaction network analysis of significant metabolites revealed significant enrichment in the Kennedy pathway caused primarily by dysregulation of phosphocholine, phosphoethanolomine, and serine (green star). FDR, false discovery rate; MTARC1, mitochondrial amidoxime-reducing component 1; PHH, primary human hepatocytes.