Glucagon-induced PGC-1⍺4/PPARγ promotes hepatic lipid storage and macrosteatosis in fasting and MASLD
Description
Untargeted lipidomics analysis of liver samples from MASLD mice model. Liver samples were harvested from male and female PGC-1α4HepTg+ and WTCre+ mice (n = 10-12 per group) fed a MASLD-promoting diet during 12 weeks. Untargeted lipidomic analyses were performed using validated protocols for lipid extraction and LC-MS analysis.
Files
Steps to reproduce
Hepatocyte-specific PGC-1α4 overexpressing male and female mice (PGC-1α4HepTg+)(Leveille et al., 2020) were compared to age- and sex-matched littermate Alb-Cre+ (WTCre+) control mice. For MASLD modeling, mice were fed a high-fat, high-fructose, and cholesterol-rich diet (D17010103i, Research Diets, 20 kcal% protein, 40 kcal% carbohydrates, and 40 kcal% fat, with 2% w/w cholesterol) for 12 weeks, starting at 5 weeks of age. Liver samples were harvested from male and female PGC-1α4HepTg+ and WTCre+ mice (n = 10-12 per group) fed a MASLD-promoting diet, and untargeted lipidomic analyses were performed using validated protocols for lipid extraction and LC-MS analysis, as previously described (Forest et al., 2018). Lipids were extracted from 30 mg of liver tissue and spiked with six internal standards (PS(12:0/12:0), PC(14:0/14:0), PE(17:0/17:0), PC(19:0/19:0), Cr(d18:1/10:0), and TG(17:0/17:1/17:0)-d5). Lipids were eluted on a Zobrax Eclipse plus C18, 2.1 x 100 mm, 1.8 µm (Agilent Technologies Inc.) heated to 40°C with a constant flow rate at 0.45 mL/min with an 83 min gradient of mobile phase A (0.2% formic acid and 10 mM ammonium formate in water) and mobile phase B (0.2% formic acid and 5mM ammonium formate in methanol/acetonitrile/MTBE 55:35:10 (v/v/v)). Samples were injected (volume=0.5 µL) into a 1290 Infinity HPLC coupled to a 6550 QTOF equipped with a dual ESI source (Agilent Technologies Inc., Santa Clara, USA) and analyzed in positive mode (scan m/z range = 50 to 1700). Data collection and processing were done using a previously validated in-house pipeline(Forest et al., 2018) including (1) application of a frequency filter of 80% based on a feature’s presence within each condition, (2) data normalization using cyclic loess algorithm (from the limma R package from Bioconductor), (3) imputation of missing values to 90% of the lowest value, and (4) batch correction for manipulator using ComBat. The final dataset consisted of a matrix in which each sample is described by 1778 features. For statistical analysis, Principal Component Analysis (PCA) was performed using mixOmics R package (Rohart et al., 2017) for data overview and outlier identification. Group comparisons were made through linear regression analysis, p-values were adjusted using False Discovery Rate (FDR) and corrected for sex (limma (Ritchie et al., 2015) R package). Most discriminant features between the two groups were chosen using an adjusted p-value <1x10e-6 and a fold-change (FC) >1.25 or <0.8 threshold. Annotation was done using in-house database (based on exact mass, retention time, and MS/MS scans) and public LIPID MAPS database with further validation using MS/MS analysis for the most discriminant features.
Institutions
- Institut de recherches cliniques de Montreal