Blood transcriptome profiling identifies 2 candidate endotypes of atopic dermatitis

Abstract

Background: Few studies have analyzed the blood transcriptome in atopic dermatitis (AD).

Objective: We explored blood transcriptomic features of moderate to severe AD.

Methods: Blood messenger RNA sequencing on 60 adults from the TREATgermany registry including 49 patients before and after dupilumab treatment, as well as from an independent cohort of 31 patients and 43 controls was performed. Patient clustering, differential expression, correlation and coexpression network analysis, and unsupervised learning were conducted.

Results: AD patients showed pronounced inflammatory expression signatures with increased myeloid and IL-5-related patterns, and clearly segregated into 2 distinct clusters, with striking differences in particular for transcripts involved in eosinophil signaling. The eosinophil-high endotype showed a more pronounced global dysregulation, a positive correlation between disease activity and signatures related to IL-5 signaling, and strong correlations with several target proteins of antibodies or small molecules under development for AD. In contrast, the eosinophil-low endotype showed little transcriptomic dysregulation and no association between disease activity and gene expression. Clinical improvement with receipt of dupilumab was accompanied by a decrease of innate immune responses and an increase of lymphocyte signatures including B-cell activation and natural killer cell composition and/or function. The proportion of super responders was higher in the eosinophil-low endotype (32% vs 11%). Continued downregulation of IL18RAP, IFNG, and granzyme A in the eosinophil-high endotype suggests a residual disturbance of natural killer cell function despite clinical improvement.

Conclusion: AD can be stratified into eosinophilic and noneosinophilic endotypes; such stratification may be useful when assessing stratified trial designs and treatment strategies.

Keywords: Atopic dermatitis; RNA-Seq; blood transcriptome; dupilumab; endotypes; eosinophil signatures; gene expression.