Abstract
Nonalcoholic fatty liver (NAFL) is a precursor of nonalcoholic steatohepatitis (NASH), a condition that may progress to cirrhosis and hepatocellular carcinoma. Markers for diagnosis of NASH are still lacking. We have investigated lipid markers using mouse models that developed NAFL when fed with high fat diet (HFD) or NASH when fed using methionine choline deficient diet (MCDD). We have performed a comprehensive lipidomic analysis on liver tissues as well as on sera from mice fed HFD (n = 5), MCDD (n = 5) or normal diet as controls (n = 10). Machine learning approach based on prediction analysis of microarrays followed by random forests allowed identifying 21 lipids out of 149 in the liver and 14 lipids out of 155 in the serum discriminating mice fed MCDD from HFD or controls. In conclusion, the global approach implemented allowed characterizing lipid signatures specific to NASH in both liver and serum from animal models. This opens new avenue for investigating early and non-invasive lipid markers for diagnosis of NASH in human.