Identification of Potential Diagnostic Biomarkers for Systemic Juvenile Idiopathic Arthritis by Integrative Transcriptomic Analysis

Abstract

Introduction: Currently the diagnostic criteria for systemic juvenile idiopathic arthritis (sJIA) is lack of specificity. Diagnostic biomarkers are needed to be identified to help with the early diagnosis of sJIA and prevent lethal complications like MAS. The aim of this study was to identify potential diagnostic biomarkers of sJIA. Methods A JIA cohort study from Gene Expression Omnibus (GEO) database was adopted to identify hub genes of sJIA comparing to healthy or non-sJIA JIA group by using integrated bioinformatic analysis which combined differentially expressed gene (DEG) analysis, weighted co-expression network analysis (WGCNA) and protein-protein network interaction (PPI) analysis. Least absolute shrinkage and selection operator (LASSO) regression analysis was further applied to screen out biomarker genes with most diagnostic potential for sJIA. A prediction model based on the selected genes was constructed and validated in three independent GEO cohort to testify their potency as reliable diagnostic markers to distinguish sJIA patients from healthy population as well as other different types of JIA. Also, CIBERSORT was applied to evaluate the immune cells infiltration and the correlation coefficient between three diagnostic genes and each immune cell subgroup was calculated in the correlation analysis. Results Totally 761 DEGs were acquired by comparing the gene expression profiles in peripheral blood mononuclear cell (PBMC) samples between the sJIA patients and the health controls, the up-regulated genes in sJIA group were mostly enriched in innate immunity and erythrocyte related biological process, while the down-regulated genes were mostly enriched in nature killer cells related biological process. Up to 22 hub genes were identified via combining DEGs with WGCNA and PPI network analysis. All the hub genes were processed to LASSO regression analysis and eventually three genes, 5’-Aminolevulinate Synthase 2 (ALAS2), S100 Calcium Binding Protein A9 (S100A9) and S100 Calcium Binding Protein A12 (S100A12) were screened out as the most potential diagnostic genes. The three genes-based prediction nomogram model was verified and presented good diagnostic performance in all three independent validation datasets. Erythrocyte related gene ALAS2 was with the most significance among all three genes, and specifically higher in sJIA patients comparing with the health controls and other JIA categories. Immune related genes S100A9 and S100A12 also showed significant difference in most conditions, but the difference was less dramatic when comparing with polyarthritis. ALAS2 was also highly expressed in familial hemophagocytic lymphohistiocytosis (FHLH) and systemic lupus erythematosus (SLE), which can develop to MAS and lead to hemophagocytosis. While S100A9 and S100A12 were commonly up-regulated in inflammatory disease. Conclusions ALAS2, S100A9 and S100A12 were highly relevant to sJIA and showed better performance in diagnosis of sJIA when applied comprehensively. ALAS2 may be associated with the predisposition to hemophagocytosis in sJIA, while S100A9 and S100A12 were mainly associated with the hyperinflammation.