Computational recognition of regulator genes and signature for ferroptosis with implications on immunological properties and clinical management of atopic dermatitis

Abstract

BackgroundAtopic dermatitis (AD) is a common chronic dermatitis of autoimmune origin that considerably affects the quality of life of patients. Ferroptosis, a newly regulated form of cell death, is essential for inflammation-related damage-associated molecular patterns (DAMPs). In this study, we aimed to identify ferroptosis regulators relevant to AD pathogenesis and reveal the mechanisms by which ferroptosis regulates the pathogenesis of AD.MethodsWe analyzed the GEO AD cohorts (GSE16161, GSE32924, GSE107361, and GSE120721), identifying AD-related differentially expressed genes (DEGs) using edgeR. Co-expression and STRING database analyses were used to elucidate the interactions between DEGs and ferroptosis markers. Through functional enrichment analysis, we defined potential biological functions within the protein-protein interaction (PPI) network and developed FerrSig using LASSO regression. The utility of FerrSig in guiding the clinical management of AD was evaluated using the GSE32473 cohort. Subsequently, our in silico findings were confirmed, and mechanistic insights were expanded through both in vitro and in vivo studies, validating the relevance of FerrSig.ResultsIn the GEO AD cohort, 278 DEGs were identified, including seven ferroptosis signature genes. Co-expression analysis and STRING database review revealed a 63-node PPI network linked to cell cycle and pro-inflammatory pathways. Four ferroptosis genes (ALOXE3, FABP4, MAP3K14, and EGR1) were selected to create FerrSig, which was significantly downregulated in samples collected from patients with AD. In addition, immune-related signaling pathways were significantly differentially enriched between the stratifications of samples collected from patients with AD with high and low ferritin levels, whereas in the GSE32473 cohort, FerrSig was significantly increased in cohorts effectively treated with pimecrolimus or betamethasone. Finally, in vitro and in vivo models showed a notable FerrSig decrease in patients with AD versus healthy control. Treatment with betamethasone and tacrolimus restored FerrSig, and the magnitude of the increase in FerrSig was higher in samples collected from patients with AD with better efficacy assessments. In addition, FerrSig was significantly positively correlated with the ferroptosis inhibitors GPX4 and SLC7A11 and negatively correlated with reactive oxygen species (ROS) levels and p-STAT3/STAT3. This implies that the FerrSig signature genes may regulate ferroptosis through the JAK/STAT3 signaling pathway.ConclusionOur study further explored the pathogenesis of AD, and FerrSig could serve as a potential biomarker for identifying AD morbidity risks and determining treatment efficacy.