Clinical lipidomics analysis reveals biomarkers of lipid peroxidation in serum from patients with rheumatoid arthritis

Abstract

Rheumatoid arthritis (RA) is a common chronic inflammatory disorder characterized by inflammation of the synovium which causes joint damage, chronic pain, and disability. Changes in lipid profile are seen before overt RA disease manifestations suggesting that lipids contribute to the inflammation-driven metabolic changes in tumor necrosis factor. Currently, the pathogenesis of RA is unclear. Moreover, it is important to search for lipids biomarkers for early diagnosis of the disease. Herein, we compared the lipid profile of healthy individuals and RA patients to provide evidence for the diagnosis and management of persons with RA. Serum samples were collected from 511 RA patients and 396 health controls (HC). The samples were analyzed by ultra-performance liquid chromatography Q-Exactive mass spectrometry. Potential lipid biomarkers were screened and validated using the partial least squares discriminant analysis (PLS-DA), orthogonal partial least squares-discriminant analysis (OPLS-DA), random forest, binary logistic regression (BLR), receiver operating characteristic (ROC) and counter propagation artificial neural network (CP-ANN) analysis. A total of 36 differential lipid metabolites were identified, including phosphatidylethanolamine (PE), triglyceride (TG), acylcarnitine (ACar) and phosphatidylcholine (PC). Among them, PE 16:0-18:2, TG 18:0-18:1-18:2 and PE 18:2-18:2 were identified as specific biomarkers for differential diagnosis of RA. ROC analysis showed that the biomarkers had a sensitivity of 68% with a specificity of 63% in discriminating RA from HC individuals. The resulting CP-ANN can be used to identify unknown RA samples with predictive accuracy of 97% as determined through cross validation. The external validation accuracy of CP-ANN was 100% in all evaluated cases. The three lipid serum biomarkers were associated with the disease activity in RA. These lipid biomarkers showed that lipid peroxidation is disrupted in RA.