Molecular subtyping of esophageal squamous cell carcinoma by large-scale transcriptional profiling: Characterization, therapeutic targets, and prognostic value

Abstract

The tumor heterogeneity of the transcriptional profiles is independent of genetic variation. Several studies have successfully identified esophageal squamous cell carcinoma (ESCC) subtypes based on the somatic mutation profile and copy number variations on the genome. However, transcriptome-based classification is limited. In this study, we classified 141 patients with ESCC into three subtypes (Subtype 1, Subtype 2, and Subtype 3) via tumor sample gene expression profiling. Differential gene expression (DGE) analysis of paired tumor and normal samples for each subtype revealed significant difference among subtypes. Moreover, the degree of change in the expression levels of most genes gradually increased from Subtype 1 to Subtype 3. Gene set enrichment analysis (GSEA) identified the representative pathways in each subtype: Subtype 1, abnormal Wnt signaling pathway activation; Subtype 2, inhibition of glycogen metabolism; and Subtype 3, downregulation of neutrophil degranulation process. Weighted gene co-expression network analysis (WGCNA) was used to elucidate the finer regulation of biological pathways and discover hub genes. Subsequently, nine hub genes (CORO1A, CD180, SASH3, CD52, CD300A, CD14, DUSP1, KIF14, and MCM2) were validated to be associated with survival in ESCC based on the RNA sequencing (RNA-seq) data from The Cancer Genome Atlas (TCGA) database. The clustering analysis of ESCC granted better understanding of the molecular characteristics of ESCC and led to the discover of new potential therapeutic targets that may contribute to the clinical treatment of ESCC.