Transcriptomics-driven metabolic pathway analysis reveals similar alterations in lipid metabolism in mouse MASH model and human

Abstract

Background: Metabolic dysfunction-associated steatotic liver disease (MASLD) is a prevalent chronic liver disease worldwide, and can rapidly progress to metabolic dysfunction-associated steatohepatitis (MASH). Accurate preclinical models and methodologies are needed to understand underlying metabolic mechanisms and develop treatment strategies. Through meta-analysis of currently proposed mouse models, we hypothesized that a diet- and chemical-induced MASH model closely resembles the observed lipid metabolism alterations in humans.

Methods: We developed transcriptomics-driven metabolic pathway analysis (TDMPA), a method to aid in the evaluation of metabolic resemblance. TDMPA uses genome-scale metabolic models to calculate enzymatic reaction perturbations from gene expression data. We performed TDMPA to score and compare metabolic pathway alterations in MASH mouse models to human MASH signatures. We used an already-established WD+CCl4-induced MASH model and performed functional assays and lipidomics to confirm TDMPA findings.

Results: Both human MASH and mouse models exhibit numerous altered metabolic pathways, including triglyceride biosynthesis, fatty acid beta-oxidation, bile acid biosynthesis, cholesterol metabolism, and oxidative phosphorylation. We confirm a significant reduction in mitochondrial functions and bioenergetics, as well as in acylcarnitines for the mouse model. We identify a wide range of lipid species within the most perturbed pathways predicted by TDMPA. Triglycerides, phospholipids, and bile acids are increased significantly in mouse MASH liver, confirming our initial observations.

Conclusions: We introduce TDMPA, a methodology for evaluating metabolic pathway alterations in metabolic disorders. By comparing metabolic signatures that typify human MASH, we show a good metabolic resemblance of the WD+CCl4 mouse model. Our presented approach provides a valuable tool for defining metabolic space to aid experimental design for assessing metabolism.

Fig. 1. Effect of WD + CCl4 on mouse liver.

A Microscopy of hematoxylin and eosin (H&E)-stained liver sections showing diffuse macro-vesicular steatosis, lobular inflammation, and the presence of ballooned hepatocytes in the WD+CCl4 group. B Histological scoring of the WD+CCl4 group livers, given as median values. Values in parentheses denote the range of values across all mice. C Oil Red O staining comparing neutral lipid content of control and WD+CCl4 group (unpaired t-test; ****p < 0.0001). D Microscopy of Sirius Red-stained liver sections showed increased fibrosis in WD+CCl4 group (unpaired t-test; **p < 0.005). E Liver transaminases (ALT and AST), and total cholesterol were significantly increased in WD+CCl4 group (unpaired t-test; ****p < 0.0001). All box plot error bars represent standard deviation. WD+CCl4: Western diet supplemented by carbon tetrachloride. (WD+CCl4 n = 7, control n = 5).

Fig. 2. TDMPA results for the compared cases.

a WD+CCl4 mouse model vs. Control, b human MASH vs. Control, c human MASH F3 vs. MASLD (MASH progression). d Corresponding Venn diagram of the statistically significant (FDR < 0.05) altered pathways across the three datasets. Number labels on the left panels (a, b, c) correspond to the pathways as indicated in the table on the right. WD+CCl4: Western diet supplemented by carbon tetrachloride, MASLD: Metabolic dysfunction-associated steatotic liver, MASH: Metabolic dysfunction-associated steatohepatitis.

Fig. 3. Metabolic network of the most affected pathways in mouse MASH model and corresponding reaction change ratios.

Green and red colors denote a calculated increase and decrease of reaction flux, respectively. Blue triangles denote a good agreement with corresponding human data (MASH vs. CTRL (2)). Reaction change ratios are reported as log2-fold-changes. MASH Metabolic dysfunction-associated steatohepatitis, CTRL control, HMG-CoA 3-hydroxy-3-methylglutaryl coenzyme A, MEV mevalonate, IPP isopentenyl pyrophosphate, FPP farnesyl pyrophosphate, SQL squalene, LANST lanosterol, ZYMST zymosterol, DESMST desmosterol, CHOL cholesterol, CHOL-EST cholesterol ester, G3P glycerol-3-phosphate, LPA lysophosphatidic acid, PA phosphatidic acid, CDP-DG cytidine diphosphate diacylglyceride, PI phosphatidylinositol, DG diacylglyceride, TG triacylglyceride, MG monoacylglyceride, PC phosphatidylcholine, LPC lysophosphatidylcholine, PE phosphatidylethanolamine, LPE lysophosphatidylethanolamine, PS phosphatidylserine, FA fatty acid, aCRN: acylcarnitine, aCoA acyl-coenzyme A, 2eCoA 2-enoyl coenzyme A, 3haCoA 3-hydroxyacyl coenzyme A, 3kaCoA 3-ketoacyl coenzyme A, acCoA acetyl coenzyme A, TCA tricarboxylic acid, ATP: adenosine triphosphate.

Fig. 4. Effect of WD + CCl4 on (I) mitochondrial bioenergetics and (II) fatty acid oxidation.

a High-resolution respirometry of oxygen consumption (O2 flux) in mitochondria in liver homogenates of control vs. WD+CCl4 group. b Immunoblot showing expression of mitochondrial pyruvate carrier MPC1 and MPC2 in liver homogenates. Vinculin served as the loading control. Both MPC1 and MPC2 were lower in WD+CCl4 vs. the control group. (unpaired t-test; *p < 0.05; **p < 0.005; ***p < 0.001). c fatty acid oxidation was measured by high-resolution respirometry of oxygen consumption (O2 flux) in mitochondria in liver homogenates of the control vs. WD+CCl4 group. d Immunoblot showing expression of carnitine palmitoyltransferase CPT1α and CPT2 in liver homogenates. Vinculin served as the loading control. CPT-2 was lower in WD+CCl4 vs. the control group. (unpaired t-test; *p < 0.05). All box plot error bars represent standard deviation. WD+CCl4: Western diet supplemented by carbon tetrachloride. (a: WD+CCl4 n = 7, control n = 5, b, d: WD+CCl4 n = 5, control n = 5, c: WD+CCl4 n = 6, control n = 4).

Fig. 5. Metabolite changes in MASH vs. healthy mouse liver tissue.

a PCA plot, b volcano plot, c heatmap of the most changed lipids. MASH metabolic dysfunction-associated steatohepatitis, PCA principal component analysis. (MASH n = 17, control n = 9).

Fig. 6. Fatty acyl chain lengths of significantly altered lipids in MASH vs. healthy mouse liver tissue.

a Distribution per the direction of change and lipid class, b Box plots of log2-fold-changes per lipid class. The solid line denotes the median of the distribution and the upper and lower hinges correspond to the first and third quartiles, respectively (Tukey representation). MASH metabolic dysfunction-associated steatohepatitis, TG triacylglyceride, DG diacylglyceride, MG monoacylgyceride, PL phospholipid, LPL lysophospholipid, CE cholesterol ester. (MASH n = 17, control n = 9).