Psoriasis, an immune-mediated chronic inflammatory skin disease, is primarily diagnosed through clinical assessment. Currently, specific markers for the accurate diagnosis and prediction of psoriatic disease are lacking. Here, we employed a three-step designed study to perform untargeted metabolomics, with the aim of identifying candidate biomarkers for psoriasis. Through comprehensive multivariate and univariate statistical analyses, we screened eight potential biomarkers specific to psoriasis, with five structurally identified. Two dipeptide biomarkers, γ-GluSer and ThrGly, along with a lysine glycation metabolite, Nα-fructosyl-lysine (Fruc-Lys), were found to be psoriasis biomarkers for the first time. Receiver operating characteristic curve analysis revealed that the area under the curve (AUC) values of these eight metabolites/features ranged from 0.68 to 0.94. A biomarker panel comprising ThrGly and feature m/z 120.0656 (C4H9NO3) demonstrated high diagnostic accuracy (AUC = 0.97) in distinguishing psoriasis patients from healthy controls. Overall, our study identified and validated a panel of plasma metabolites/features that could serve as potential biomarkers for the diagnosis of psoriasis, providing new insights into the diagnosis and pathogenesis of this disease.
Read full abstract