Eleven genes associated with progression and prognosis of endometrial cancer (EC) identified by comprehensive bioinformatics analysis.

Background: Endometrial cancer (EC) is one of the female malignant tumors. Endometrial cancer predominately affects post-menopausal women. Bioinformatics analysis has been widely applied to screen and analyze genes in linkage to various types of cancer progression.

Methods: Download the gene expression profile from Gene Expression Omnibus (GEO). Calculate raw expression data according to pre-processing procedures. We performed the "limma" R language package to screen DEGs between Endometrial cancer tissue samples and normal uterus tissue samples. Enrichment of the functions and pathways was analyzed by using clusterprofiler. We utilized Search Tool for the Retrieval of Interacting Genes Database (STRING) to assess protein-protein interaction (PPI) information, and then we used plug-in Molecular Complex Detection (MCODE) to screen hub modules of PPI network in Cytoscape. We also performed functional analysis on the genes in the hub module by using clusterprofiler. Next, we utilized the "WGCNA" package in R to establish co-expression network for the DEGs. The Venn diagram was performed to overlap the gene in key module and hub PPI cluster. We validated the key genes in TCGA, GEPIA, UALCAN and Immunohistochemistry staining obtained from The Human Protein Atlas database. And then we did ROC curve analysis by SPSS. Gene set enrichment analysis (GSEA) and mutation analysis were also performed for hub genes.

Results: Functional and pathway enrichment analysis demonstrated that the upregulated differentially expressed genes (DEGs) were significantly enriched in CXCR chemokine receptor binding, chemokine activity, chemokine receptor binding, G-protein coupled receptor binding, RAGE receptor binding, cytokine activity, microtubule binding, receptor regulator activity and microtubule motor activity, and the down-regulated genes were highly enriched in collagen binding. After using STRING software to construct PPI network, 30 prominent proteins were identified and the first two significant modules were selected. In co-expression network, 5 EC-related modules were identified. Among them, the turquoise module has the highest correlation with the EC. We further analyzed the genes in the PPI and turquoise module, and selected eleven key genes related to EC after validation of TCGA database, GEPIA, UALCAN and immunohistochemistry. Six of them had mutation significance.

Conclusions: In summary, these 11 genes may become new therapy targets for EC treatment.