Background Metabolomics is the study of unique chemical fingerprints of specific cellular processes. Metabolite profiling is recently applied in identifying biomarkers in medical research including rheumatologic diseases.
Objectives The aim is to assess the metabolite profiling of synovial fluid in patients with different forms of inflammatory arthritis and to identify putative biomarker for rheumatoid arthritis (RA) compared to the other inflammatory arthritis.
Methods Synovial fluid samples were obtained from patients with RA (n = 13, mean age 44.2 ± 10.7 yr) and ankylosing spondylitis (AS) (n = 7, mean age 35.4 ± 10.7 yr), Behcet’s disease (BD) (n = 5, mean age 41.6 ± 12.5 yr) and gout (n = 13, mean age 45.9 ± 7.9 yr). To identify putative biomarkers for RA, the synovial fluid samples were divided into two groups; RA versus non-RA (NRA) which included AS, BD and gout. The metabolites of synovial fluid were analyzed using gas chromatography/time-of-flight mass spectrometry (GC/TOF MS). The multivariate statistical analyses by orthogonal partial least squares discriminant analysis (OPLS-DA) were conducted for the comparison between two groups. The potential biomarkers in RA patients were identified and evaluated by variable importance for projection (VIP) values, non-parametric Wilcoxon-Mann-Whitney test and one-way ANOVA test. Chemometric model validation was finally carried out using receiver operating characteristic (ROC) curve and area under the ROC curve (AUC).
Results A total of 119 metabolites were identified from 38 samples. The metabolite profiling between RA and NRA were clearly discriminated by OPLS-DA (Figure 1). Candidates of biomarkers in RA were determined by VIP values extracted from OPLS-DA and 41 metabolites were selected by VIP scores of greater than 1.0, of which 29 metabolites were elevated in RA and 12 metabolites in NRA. After eliminating variables with no significant difference using Wilcoxon-Mann-Whitney test and one-way ANOVA test, 23 of 41 metabolites were selected as putative biomarkers for RA compared to NRA. Fifteen metabolites were higher level in RA (succinic acid, octadecanol, asparagines, terephtalic acid, salicylaldehyde, glutamine, citrulline, tyrosine, uracil, lysine, phenylalanine, ribitol, tryptophan, xylose and pyrophosphate) and 8 metabolites in NRA (isopalmitic acid, glycerol, myristic acid, palmitoleic acid, hydroxylamine, ethanolamine, alanine and serine). These metabolites were validated by AUC, all of which had AUC > 0.8. ROC curve analysis for the power of discrimination of RA from NRA showed a sensitivity of 69.2% and a specificity of 92%.
Conclusions Our study suggests that the synovial fluid metabolomic profiling can be a novel approach in differentiating RA from AS, BD and gout. A set of validated metabolites could be a putative biomarker in synovial fluid of RA patients.
Disclosure of Interest None Declared