Aims
Non-alcoholic fatty liver disease (NAFLD) is a usual chronic liver disease and lacks non-invasive biomarkers for the clinical diagnosis and prognosis. Extracellular vesicles (EVs), a group of heterogeneous small membrane-bound vesicles, carry proteins and nucleic acids as promising biomarkers for clinical applications, but it has not been well explored on their lipid compositions related to NAFLD studies. Here, we investigate the lipid molecular function of urinary EVs and their potential as biomarkers for non-alcoholic steatohepatitis (NASH) detection.
Background and aims
Non-alcoholic fatty liver disease (NAFLD) is a usual chronic liver disease and lacks non-invasive biomarkers for the clinical diagnosis and prognosis. Extracellular vesicles (EVs), a group of heterogeneous small membrane-bound vesicles, carry proteins and nucleic acids as promising biomarkers for clinical applications, but it has not been well explored on their lipid compositions related to NAFLD studies. Here, we investigate the lipid molecular function of urinary EVs and their potential as biomarkers for non-alcoholic steatohepatitis (NASH) detection.
Conclusion
The lack of non-invasive means for diagnosing NASH causes increasing morbidity. We investigate the NAFLD biomarkers from the insights of urinary EVs, and systematically compare the EV lipidomic profiles of NAFL and NASH, which holds the promise to expand the current knowledge of disease pathogenesis and evaluate their role as non-invasive biomarkers for NASH diagnosis and progression.
Methods
This work includes 43 patients with non-alcoholic fatty liver (NAFL) and 40 patients with NASH. The EVs of urine were isolated and purified using the EXODUS method. The EV lipidomics was performed by LC-MS/MS. We then systematically compare the EV lipidomic profiles of NAFL and NASH patients and reveal the lipid signatures of NASH with the assistance of machine learning.
Results
By lipidomic profiling of urinary EVs, we identify 422 lipids mainly including sterol lipids, fatty acyl lipids, glycerides, glycerophospholipids, and sphingolipids. Via the machine learning and random forest modeling, we obtain a biomarker panel composed of 4 lipid molecules including FFA (18:0), LPC (22:6/0:0), FFA (18:1), and PI (16:0/18:1), that can distinguish NASH with an AUC of 92.3%. These lipid molecules are closely associated with the occurrence and development of NASH.