Integrated Single-Cell Transcriptomics Identifies Epi_C4 and Supports a Pro-Tumor Role for FAM20C in Lung Adenocarcinoma.

Abstract

Intratumoral epithelial heterogeneity contributes to lung adenocarcinoma (LUAD) progression and may limit the interpretability of prognostic models derived from bulk tissue. To characterize this heterogeneity and assess its clinical relevance, four public single-cell RNA-sequencing datasets comprising 193,280 quality-controlled cells were integrated. Five epithelial states were identified and characterized based on inferred aneuploidy, transcriptional programs, differentiation potential, trajectory structure, and intercellular communication. State-specific signatures were projected onto the Cancer Genome Atlas LUAD cohort and six independent Gene Expression Omnibus cohorts. Epi_C4 exhibited high inferred aneuploidy, reduced differentiation, elevated malignancy-associated program scores, and enrichment in LUAD and was associated with worse survival. A 12-gene prognostic model derived from Epi_C4 stratified patients according to overall survival in the training cohort and most validation cohorts, but not in GSE29013. High-risk tumors were enriched for extracellular matrix, PI3K-AKT, HIF-1, glycolytic, inflammatory, and immunoregulatory programs. FAM20C was enriched in Epi_C4, overexpressed in LUAD, and associated with worse survival. In NCI-H838 cells, siRNA-mediated FAM20C depletion reduced cell viability, DNA synthesis, Matrigel invasion, and wound closure while increasing apoptosis. Collectively, these findings identify Epi_C4 as a prognostically adverse epithelial state in LUAD and implicate FAM20C in tumor-associated cellular phenotypes.