Moving towards endotypes in atopic dermatitis: identification of patient clusters based on serum biomarker analysis

Publication date: Available online 13 April 2017
Source:Journal of Allergy and Clinical Immunology
Author(s): Judith Lydia Thijs, Ian Strickland, Catharina Ansfrieda Francisca Maria Bruijnzeel-Koomen, Stefan Nierkens, Barbara Giovannone, Eszter Csomor, Bret Richard Sellman, Tomas Mustelin, Matthew Alexander Sleeman, Marjolein Saskia de Bruin-Weller, Athula Herath, Julia Drylewicz, Richard David May, DirkJan Hijnen
BackgroundAtopic dermatitis (AD) is a complex, chronic, inflammatory skin disease with a diverse clinical presentation. It is however unclear whether this diversity exists at a biological level.ObjectiveTo test the hypothesis that AD is heterogeneous at the biological level of individual inflammatory mediators.MethodsSerum from 193 moderate to severe adult AD patients (geomean (95%CI) SASSAD of 22.3 (21.3, 23.3) and 39.1 (37.5, 40.9) respectively) and 30 non-AD healthy controls was analysed for 147 serum mediators, total IgE and 130 allergen specific IgEs. Population heterogeneity was assessed by principal component analysis (PCA) followed by unsupervised k-means cluster analysis of the principal components.ResultsAD patients showed pronounced evidence of inflammation compared to healthy controls. PCA of AD serum data revealed the presence of four potential AD patient clusters. Fifty-seven principal components (PCs) described approximately 90% of the variance. Unsupervised k-means cluster analysis of the 57 largest PCs delivered 4 distinct clusters of AD patients. Cluster 1 had high SASSAD and BSA with the highest levels of PARC, TIMP-1 and sCD14. Cluster 2 had low SASSAD with the lowest levels of IFN-α, TIMP-1 and VEGF. Cluster 3 had high SASSAD with the lowest levels of IFN-β, IL-1 and epithelial cytokines. Cluster 4 had low SASSAD but highest levels of inflammatory markers: IL-1, IL-4, IL-13 and TSLP.ConclusionAD is a heterogeneous disease both clinically and biologically. Four distinct AD patient clusters have been identified that could represent endotypes with unique biological mechanisms. Elucidation of these endotypes warrants further investigation and will require future intervention trials with specific agents such as biologics.



http://ift.tt/2nLFyrA