Identification and validation of shared biomarkers and drug repurposing in psoriasis and Crohn's disease: Integrating bioinformatics, machine learning, and experimental approaches

Psoriasis and Crohn's disease (CD) are chronic inflammatory diseases that involve complex immune-mediated mechanisms. Despite clinical overlap and shared genetic predispositions, the molecular pathways connecting these diseases remain incompletely understood. The present study seeks to identify shared biomarkers and therapeutic targets for psoriasis and CD. Initially, differentially expressed genes (DEGs) were identified from publicly available transcriptomic datasets related to psoriasis and CD. Simultaneously, weighted gene co-expression network analysis (WGCNA) was performed to identify gene modules associated with the clinical traits of psoriasis and CD. Subsequently, biomarkers were prioritized from shared key genes by integrating protein-protein interaction (PPI) networks with machine learning models. Gene Set Enrichment Analysis (GSEA), along with Gene Ontology (GO) and KEGG pathway analysis, highlighted the significance of cell cycle regulation and immune response pathways in comorbidities. Immune infiltration analysis underscored the involvement of hub genes in immune regulation, while single-cell transcriptomic analysis revealed the cellular localization of these hub genes.Additional targeted molecular biology experiments validated the shared biomarkers.Finally, KIF4A, DLGAP5, NCAPG, CCNB1, and CEP55 were identified as key regulatory molecules and shared biomarkers. DSigDB predictions and molecular docking simulations indicated strong therapeutic potential for Etoposide, Lucanthone, and Piroxicam, with Etoposide showing the highest affinity for key targets. In cellular models, Etoposide demonstrated promising therapeutic effects by significantly downregulating the expression of psoriasis-related keratinocytes marker genes (KRT6, KRT16) and CD-related inflammatory cytokines (IL6, IL8, TNF-α), highlighting its potential in treating psoriasis and CD. This study integrates bioinformatics, machine learning, and molecular validation to identify the shared molecular mechanisms of psoriasis and CD, uncovering novel biomarkers and potential combined therapeutic candidates. These findings provide valuable insights into potential treatment strategies for these diseases.