Construction of EMT related prognostic signature for kidney renal clear cell carcinoma, through integrating bulk and single-cell gene expression profiles

Introduction: Kidney renal clear cell carcinoma (KIRC), as a main type of malignant kidney cancers, has a poor prognosis. Epithelial-mesenchymal transformation (EMT) exerts indispensable role in tumor progression and metastasis, including in KIRC. This study aimed to mine more EMT related details and build prognostic signature for KIRC. Methods: The KIRC scRNA-seq data and bulk data were downloaded from GEO and TCGA databases, respectively. The cell composition in KIRC was calculated using CIBERSORT. Univariate Cox regression analysis and LASSO Cox regression analysis were combined to determine the prognostic genes. Gene set variation analysis and cell-cell communication analysis were conducted to obtain more functional information. Additionally, functional analyses were conducted to determine the biological roles of si-LGALS1 in vitro. Results: We totally identified 2,249 significant differentially expressed genes (DEGs) in KIRC samples, meanwhile a significant distinct expression pattern was found in KIRC, involving Epithelial Mesenchymal Transition pathway. Among all cell types, significantly higher proportion of epithelial cells were observed in KIRC, and 289 DEGs were identified in epithelial cells. After cross analysis of all DEGs and 970 EMT related genes, SPARC, TMSB10, LGALS1, and VEGFA were optimal to build prognostic model. Our EMT related showed good predictive performance in KIRC. Remarkably, si-LGALS1 could inhibit migration and invasion ability of KIRC cells, which might be involved in suppressing EMT process. Conclusion: A novel powerful EMT related prognostic signature was built for KIRC patients, based on SPARC, TMSB10, LGALS1, and VEGFA. Of which, si-LGALS1 could inhibit migration and invasion ability of KIRC cells, which might be involved in suppressing EMT process.