Meta-Analysis Identifying Epithelial-Derived Transcriptomes Predicts Poor Clinical Outcome And Immune Infiltrations In Ovarian Cancer
Dong-Feng Li,Aisikeer Tulahong,Md Nazim Uddin,Huan Zhao,Hua Zhang
DOI: https://doi.org/10.3934/mbe.2021324
2021-01-01
Mathematical Biosciences and Engineering
Abstract:Background: Previous studies revealed that the epithelial component is associated with the modulation of the ovarian tumor microenvironment (TME). However, the identification of key transcriptional signatures of laser capture microdissected human ovarian cancer epithelia remains lacking. Methods: We identified the differentially expressed transcriptional signatures of human ovarian cancer epithelia by meta-analysis of GSE14407, GSE2765, GSE38666, GSE40595, and GSE54388. Then we investigated the enrichment of KEGG pathways that are associated with epithelia-derived transcriptomes. Finally, we investigated the correlation of key epithelia-hub genes with the survival prognosis and immune infiltrations. Finally, we investigated the genetic alterations of key prognostic hub genes and their diagnostic efficacy in ovarian cancer epithelia. Results: We identified 1339 differentially expressed genes (DEGs) in ovarian cancer epithelia including 541upregulated and 798 downregulated genes. We identified 21 (such as E2F4, FOXM1, TFDP1, E2F1, and SIN3A) and 11 (such as JUN, DDX4, FOSL1, NOC2L, and HMGA1) master transcriptional regulators (MTRs) that are interacted with upregulated and the downregulated genes in ovarian tumor epithelium, respectively. The STRING-based analysis identified hub genes (such as CDK1, CCNB1, AURKA, CDC20, and CCNA2) in ovarian cancer epithelia. The significant clusters of identified hub genes are associated with the enrichment of KEGG pathways including cell cycle, DNA replication, cytokine-cytokine receptor interaction, pathways in cancer, and focal adhesion. The upregulation of SCNN1A and CDCA3 and the downregulation of SOX6 are correlated with a shorter survival prognosis in ovarian cancer (OV). The expression level of SOX6 is negatively correlated with immune score and positively correlated with tumor purity in OV. Moreover, SOX6 is negatively correlated with the infiltration of TILs, CD8+ T cells, CD4+ Regulatory T cells, cytolytic activity, T cell activation, pDC, neutrophils, and macrophages in OV. Also, SOX6 is negatively correlated with various immune markers including CD8A, PRF1, GZMA, GZMB, NKG7, CCL3, and CCL4, indicating the immune regulatory efficiency of SOX6 in the TME of OV. Furthermore, SCNN1A, CDCA3, and SOX6 genes are genetically altered in OV and the expression levels of SCNN1A and SOX6 genes showed diagnostic efficacy in ovarian cancer epithelia. Conclusions: The identified ovarian cancer epithelial-derived key transcriptional signatures are significantly correlated with survival prognosis and immune infiltrations, and may provide new insight into the diagnosis and treatment of epithelial ovarian cancer.