Exploration of Targets Potentially Linked to IL-17A Inhibitor Response in Psoriasis Using Machine Learning.

Background: This study aimed to elucidate IL-17 inhibitors' mechanisms in psoriasis, offering a theoretical basis for tackling clinical issues like treatment resistance and relapse. Methods: Datasets GSE226244 and GSE31652 served as the training set, and GSE201827 served as the testing set. Differential hub genes post-IL-17 inhibitor treatment identified via Limma and WGCNA. DEGs were defined by a |log2 fold-change (FC)| greater than 0.585 and a stringent FDR threshold of less than 0.05. CIBERSORT evaluated immune cell infiltration. Comprehensive analysis of 113 machine learning methods identified optimal predictive model. qPCR validated CLCNKB and GFRA3 expression in psoriasis cell models post-IL-17 inhibitor treatment. Mendelian randomization analysis explored causal links between CLCNKB, GFRA3 and cytokines. Results: Analysis of gene expression in psoriasis patients treated with IL-17 inhibitors identified 95 differential genes enriched in FoxO signaling, Lysine degradation, and cGMP-PKG pathways. The LASSO-glmBoost (a hybrid machine learning method combining Lasso regularization with gradient boosting) model exhibited superior diagnostic performance (AUC: 0.920 in training, 0.858 in test), highlighting CLCNKB and GFRA3 as key genes in the optimal predictive framework. qPCR confirmed their upregulation in IL-17-inhibitor-treated psoriasis cells, and Mendelian randomization linked both genes causally to cytokine dysregulation. Conclusion: The study reveals new insights into IL-17 inhibitors' mechanisms in psoriasis, suggesting that upregulation of CLCNKB and GFRA3, along with cytokine dysregulation (eg, IL-13, IL-10, IL-12, TGF-β, TNF-α), may underlie potential resistance and relapse in patients. This work demonstrates a novel approach to clinical outcome prediction with potential utility for specific clinical application, warranting further validation in clinical settings.