A study on the screening of potential biomarkers for rheumatoid arthritis based on nontarget metabolomics

Abstract

To explore changes in blood metabolites with rheumatoid arthritis (RA) using non-targeted metabolomics techniques, to identify potential metabolic biomarkers in RA patients. Twenty patients with rheumatoid arthritis (RA group) who attended the Sichuan Province Orthopaedic Hospital between July and December 2023 were recruited, and twenty healthy controls were included. Serum candidate metabolites were analysed using ultra-performance liquid chromatography-mass spectrometry (UPLC-MS). Variable Importance in Projection (VIP) value > 1 and p < 0.05 as criteria for identifying specific differentially expressed candidate metabolites. Differential candidate metabolites and associated pathways were annotated via the KEGG database, and diagnostic value was detected by receiver operating characteristic (ROC) curves. 579 metabolites were detected, and based on the screening criteria, 16 were identified as important differentially expressed candidate metabolites associated with RA. The 10 up-regulated metabolites included PC(16:1(9Z)/20:0), PC(20:1(11Z)/18:4(6Z,9Z,12Z,15Z)), Palmitoylcarnitine, 3-O-Sulfogalactosylceramide (d18:1/20:0), GPEtn(16:1/22:4), Sphinganine, 3-Methylindole, PC(16:1(9Z)/20:5(5Z,8Z,11Z,14Z,17Z)), Ortho-Hydroxyphenylacetic acid, PE(18:1(11Z)/18:3(9Z,12Z,15Z)). The 6 downregulated metabolites included PE-NMe(18:4(6Z,9Z,12Z,15Z)/18:4(6Z,9Z,12Z,15Z)), PE(22:0/15:0), PS(18:2(9Z,12Z)/22:4(7Z,10Z,13Z,16Z)), CDP-DG(a-13:0/i-12:0), 3a,7a,12a-Trihydroxy-5b-cholestan-26-al, and Serotonin. Using an area under the curve (AUC) of > 0.8 as a criterion, four potential biomarkers for RA were identified, including PE(22:0/15:0) (AUC = 0.8333, 95% CI: 0.6244–1), PE-NMe(18:4(6Z,9Z,12Z,15Z)/18:4(6Z,9Z,12Z,15Z))(AUC = 0.9722, 95% CI: 0.6571–1), Sphinganine (AUC = 0.9167, 95% CI: 0.7143–1), GPEtn(16:1/22:4)(AUC = 0.8889, 95% CI: 0.7429–1). The AUC for the combined diagnosis of RA using the four differentially expressed candidate metabolites was 0.961 (95% CI: 0.9463–0.9757). Glycerophospholipid metabolism was the most significantly altered pathway in RA, according to pathway enrichment analysis. Our research identified four potential biomarkers: PE(22:0/15:0), PE-NMe(18:4(6Z,9Z,12Z,15Z)/18:4(6Z,9Z,12Z,15Z)), Sphinganine, and GPEtn(16:1/22:4). Glycerophospholipid metabolism may be involved in RA. This study was entirely exploratory research and framed as hypothesis-generating only, which will require further validation in subsequent research.