Exploration of Diagnostic Deubiquitinating Enzymes in Endometriosis and Its Immune Infiltration.

Abstract

The mechanism involved in the pathogenesis of endometriosis is poorly understood. The purpose of this study is to identify key deubiquitinating enzymes (DUBs) for endometriosis diagnosis and elucidate the possible mechanism, offering novel insights for noninvasive early diagnosis and treatment. Four gene expression datasets were employed from the Gene Expression Omnibus to identify differentially expressed genes (DEGs) between endometriosis and normal controls. GO and KEGG pathways were performed for enrichment analysis. Calibration curves, ROC, DCA, and clinical impact curves verified the clinical usefulness of the nomogram model. In addition, the ssGSEA method was conducted to estimate 23 types of immune cells. A specific DUB gene signature was constructed with Lasso regression, univariate logistic regression, and SVM analysis. RT-qPCR validated the expression of biomarkers. A total of 85 endometriosis-related DUBs were identified in the eutopic endometrium. Among them, 20 DUBs were found to be correlated with the severity of endometriosis. A diagnostic risk model based on five DUB-related genes (USP21, USP48, ZRANB1, COPS5, and EIF3F) was developed using lasso-cox regression analysis. The nomogram model exhibited a strong predictive ability to diagnose endometriosis. KEGG analysis revealed that ubiquitin-mediated proteolysis was activated in patients suffering from severe symptoms. Analysis of immune cell infiltration revealed a positive correlation between USP21 and multiple immune cells in the eutopic endometrium. However, EIF3F showed an opposite relationship. Dysregulation of DUBs was related to the immune microenvironment in endometriosis. Results from RT-qPCR confirmed the expression of DEGs in clinical samples. In summary, the diagnostic model for endometriosis constructed using five differentially expressed DUB genes demonstrates strong diagnostic capability, suggesting that these genes could serve as potential candidate biomarkers and therapeutic targets.