GADD45A as a Potential Biomarker Associated with Endoplasmic Reticulum Stress in Focal Segmental Glomerulosclerosis

Objective: This study aimed to identify and validate potential endoplasmic reticulum stress-related biomarkers of FSGS. Methods: Five microarray datasets were downloaded from the GEO database. The endoplasmic reticulum stress-related genes were extracted from GeneCards database. GSE104948, GSE129973, and GSE121233 datasets were used to identified DEGs by limma R package. WGCNA was performed to obtain hub gene modules. We intersected the DEGs, genes from hub module of WGCNA, and ERSRGs. GO and KEGG pathway enrichment analyses were performed. LASSO, SVM-RFE, and RF algorithms were used to screen characteristic genes. Further, GSE108109 and GSE104066 were used as validated datasets. Box plots, ROC curves, and AUC were created to identify potential biomarkers. A novel nomogram model was constructed using potential biomarkers. Immunohistochemistry was used to validate the expression of the potential biomarkers. Results: Intersecting DEGs, the brown module genes, and ERSRGs, we identified 15 hub ERSRGs of FSGS. AGO2, CCND1, GADD45A, TRAM2, and PTPN1 genes were screened using Lasso, SVM-RFE, and RF. Further, CCND1, GADD45A, and TRAM2 were validated as significant in training and validation datasets. A nomogram based on CCND1, GADD45A, and TRAM2 expression was constructed. The calibration curves, decision curve, and clinical impact curve showed that the nomogram had good consistency and clinical practical benefit. The expression of GADD45A is higher in FSGS patients than control patients in IHC results ( P < 0.0001). Conclusions: GADD45A may be a potential biomarker associated with endoplasmic reticulum stress in FSGS.