Abstract
A large number of colon adenocarcinoma (COAD) patients are already advanced when diagnosed. In this study, we aimed to further understand the mechanism of tumor development in early COAD by focusing on epithelial-mesenchymal transition (EMT) and long non-coding RNAs (lncRNAs). Expression profiles of early COAD patients were obtained from public databases. EMT-related lncRNAs were used as a basis for constructing molecular subtypes through unsupervised consensus clustering. Genomic features, pathways and tumor microenvironment (TME) were compared between two subtypes. LncATLAS database was applied to analyze the relation between lncRNAs and transcription factors (TFs). First order partial correlation analysis was conducted to identify key EMT-related lncRNAs.C1 and C2 subtypes with distinct prognosis were constructed. Oncogenic pathways such as EMT, KRAS signaling, JAK-STAT signaling, and TGF-β signaling were significantly enriched in C2 subtype. Higher immune infiltration and expression of immune checkpoints were also observed in C2 subtype, suggesting the key EMT-related lncRNAs may play a critical role in the modulation of TME. In addition, JAK-STAT signaling pathway was obviously enriched in upregulated TFs in C2 subtype, which indicated a link between key lncRNAs and JAK-STAT signaling that may regulate TME. The study further expanded the research on the role of EMT-related lncRNAs in the early COAD. The six identified EMT-related lncRNAs could serve as biomarkers for early screening COAD.