Hierarchical cluster analysis of cytokine profiles reveals a cutaneous vasculitis-associated subgroup in dermatomyositis

• Published in: Clinical rheumatology Vol. 40; no. 3; pp. 999 - 1008
• Main Authors: Bai, Jingjing, Wu, Chanyuan, Zhong, Danli, Xu, Dong, Wang, Qian, Zeng, Xiaofeng
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.03.2021; Springer Nature B.V
• Subjects: Autoimmune diseases; Cluster Analysis; Correlation analysis; Cytokines; Dermatomyositis; Dermatomyositis - complications; Humans; Interleukin 1 receptor antagonist; Interleukin 1 receptors; Interleukin 22; Interleukin 6; Interleukin 7; Interleukin 8; IP-10 protein; Medicine; Medicine & Public Health; Monocyte chemoattractant protein 1; Myositis; Original Article; Phenotypes; RANTES; Regression analysis; Rheumatology; Serum levels; Tumor Necrosis Factor-alpha; Tumor necrosis factor-α; Vasculitis; α-Interferon; γ-Interferon; Cluster analysis; Cutaneous vasculitis; Cytokine; Dermatomyositis
• DOI: 10.1007/s10067-020-05339-2
• ISSN: 0770-3198; 1434-9949

Objectives Dermatomyositis (DM) is a chronic inflammatory autoimmune disease with notable heterogeneity. The intent of this study was to explore the difference in cytokine profiles of different subsets in DM based on the disease activity and myositis-specific antibodies, and to identify the clinical phenotypes associated with different cytokine profiles. Methods Serum levels of 34 cytokines were prospectively measured in 47 consecutive DM patients and healthy controls. Concentrations of the cytokines were compared between the active and stable groups. Univariate and multivariate logistic regression models were used to identify the cytokines associated with DM disease activity. The cytokine profiles of anti-MDA5 and anti-TIF1γ subsets were compared, and the correlation analysis was performed between the elevated cytokines and clinical parameters in the two subsets. Hierarchical cluster analysis was used to establish clinical-cytokine subgroups in DM. Results Serum levels of MIP-1α, IP-10, IL-8, IL-1RA, MCP-1, GRO-α, and IL-22 were significantly higher in DM patients compared with healthy controls. IP-10, IL-6, IL-1RA, IFN-α, and MCP-1 were significantly elevated in the DM-active subset than the DM-stable subset. The combination of three cytokines (IP-10, IL-1RA, and MCP-1) had a better performance in differentiating between the active subset and the stable subset than the conventional inflammatory markers. SDF-1α, IP-10, IL-7, IL-17A, RANTES, IFN-γ, TNF-α, MIP-1β, IFN-α, MCP-1, GRO-α, and IL-1α were significantly higher in the anti-MDA5 subset than in the TIF1γ subset. Cluster analysis revealed a hypercytokinemic-vasculitis subgroup in patients with DM. Conclusions Multiple cytokine signatures were depicted in different subsets of DM. A vasculitis-associated subgroup was firstly identified in DM with regards of cytokinome and deserves further mechanistic study. Key Points • The multivariate regression model of three cytokines (IP-10 , IL-1RA , and MCP-1) could be a promising tool for distinguishing between the active and stable subset in DM. • Cytokine profiles of anti-MDA5-DM and anti-TIF1γ-DM were compared to identify the immunopathological differences between the two subsets. • Cluster analysis revealed a hypercytokinemic-vasculitis subgroup in patients with DM.