Identification of molecular markers of immune cell infiltration in diabetic nephropathy by weighted gene co-expression network analysis (WGCNA)

Abstract

Abstract Background Increasing evidence has indicated that infiltrating immune cells play an important role in the pathogenesis of diabetic nephropathy (DN). However, there are relatively few systematic studies on the immunity in DN. Methods The Cell type Identification By Estimating Relative Subsets Of RNA Transcripts (CIBERSORT) algorithm was used to calculate the proportion of immune cells in the GSE96804 and GSE30528 datasets, and to find the differential immune cells between DN and normal samples. The immune cell-related genes were searched by weighted gene co-expression network analysis (WGCNA), and the differentially expressed immune cell-related genes were obtained by taking intersection with differentially expressed genes (DEGs) between DN and normal samples in the two datasets. Moreover, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed to investigate the biological functions of differentially expressed immune cell-related genes. Furthermore, multiple machines learning analyses, including Least absolute shrinkage and selection operator (LASSO) regression algorithm, XGBoost algorithm, and random forest algorithm, and ROC analyses were used to screen diagnostic genes. Finally, Gene Set Enrichment Analysis (GSEA) was performed to investigate the functions of diagnostic genes. A competing endogenous RNA (ceRNA) network was constructed and the target drugs were queried in the Drug Gene Interaction Database (DGIdb). Results The nine immune cells and six immune cells with significant differences between DN and normal samples in the GSE96804 dataset and GSE30528 dataset were intersected to obtain five co-regulated immune cells. In addition, the 321 immune cell-related genes were intersected with 65 DEGs between DN and normal samples to obtain 13 differentially expressed immune cell-related genes, including one down-regulated gene and 12 up-regulated gene in DN samples compared with normal samples. These 13 differentially expressed immune cell-related genes were mainly associated with extracellular matrix, Protein digestion and absorption, and ECM-receptor Interaction pathway responses. Furthermore, NAP1L2, MOXD1, COL1A2, COL15A1, and LUM were identified as diagnostic genes by multiple machine learning analysis and AUC evaluation. Finally, GSEA revealed that NAP1L2, MOXD1, COL1A2, COL15A1, and LUM were mainly related to immune response, amino acid metabolic, EMC-receptor interaction. Based on the diagnostic genes, 647 lncRNA-miRNA pairs were created and used to build the ceRNA network. Two targeted drugs, COL1A2 and COL15A1, were acquired in the DGIdb database. Conclusion In conclusion, NAP1L2, MOXD1, COL1A2, COL15A1, and LUM might be used as diagnostic biomarkers and therapeutic targets.