Screening and Identification of Potential Key Biomarkers in Psoriasis : Evidence From Bioinformatic Analysis

Abstract

Background: Psoriasis is a chronic and prevalent skin condition brought on by various genetic and external factors. Till date, there is no cure for this disease. It is, therefore, crucial to examine the underlying mechanisms leading up to psoriasis. The goal of our study was to explore the mechanistic pathways involved in the molecular pathogenesis of psoriasis.Methods: Using Gene Expression Omnibus (GEO), we performed an extensive analysis of the transcript expression profile in psoriasis patients. The datasets GSE13355, GSE30999, and GSE106992, containing 239 pairs of normal and psoriatic skin samples were arbitrarily assigned to two non-overlapping cohorts for cross-validated differential gene expression analysis. The gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment and gene set enrichment analysis (GSEA) were employed for interpretation, visualization, and unified recognition. The STRING database was used to construct the protein–protein interaction (PPI) network and the hub genes were established using Cytoscape. Additionally, both differentially regulated genes and functional likeness of hub genes were further examined in terms of gene activation and functional outcome. The correlation between normal tissue and infiltrating immune cells was analyzed by CIBERSORT. Moreover, ROC analysis was performed to distinguish between skin lesion samples and skin non-lesion samples. In addition, a signaling axis involving lnRNA, miRNA, mRNA, and ceRNA was generated with information from the DEmiRNA, DElncRNA, and DEmiRNA-DEmRNA relationship. Lastly, immunohistochemical evaluations were used to analyze the highest expression of single gene in the whole body within the Human Protein Atlas (HPA) database.Results: The genetic profiles of 239 pairs of normal and lesional skin samples were downloaded from three datasets in the GEO database. PPI network revealed a tight interaction among 197 differentially expressed genes (DEGs). Moreover, gene ontology analyses indicated that psoriasis-related DEGs mostly included viral defense genes, type I interferon axis, and its corresponding cellular responses. The Kyoto encyclopedia of genes and the DEGs enrichment analysis showed involvement of the NOD-like receptor signaling network, cytokine-cytokine receptor binding, and IL-17 signaling axis in psoriasis verses non-psoriatic tissues. GSEA analysis demonstrated that CXCL8 was only enriched in the "complement characteristic" pathway. ROC curves indicated that CXCL8 expression was highly effective in classifying both lesional and non-lesional skin samples (with AUC 0.941, 0,935, and 0.794 for GSE13355, GSE30999, and GSE 106992). Furthermore, CIBERSORT database indicated that CXCL8 was correlated to 22 types of infiltrating immune cells. In addition, 6 miRNAs were predicted to be related to CXCL8, including hsa-miR-1294, hsa-miR-140-3p, hsa-miR-185-5p, hsa-miR-4306, hsa-miR-4644, and hsa-miR-493-5p. Lastly, immunohistochemical analysis showed that CXCL8 was most widely distributed in lymphoid tissues.Conclusions: Based on our analysis, CXCL8 plays a key role in psoriatic development. Our comprehensive bioinformatics analysis of the GEO data provided new insights into the exploration of molecular mechanisms while searching for highly efficient therapeutic targets for treating psoriasis.