Single-cell RNA sequencing identifies stem-like subpopulations and their gene signature in lung adenocarcinoma progression

Lung adenocarcinoma (LUAD) demonstrates significant intra-tumor heterogeneity (ITH), which plays a critical role in disease progression and complicates therapeutic interventions. In the present study, we employed single-cell RNA sequencing (scRNA-seq) to analyze cellular populations derived from stage I and stage III LUAD tumors, alongside matched adjacent normal tissues, with the objective of delineating malignant cell states and identifying tumor-propagating subpopulations. Unsupervised clustering analysis identified eight principal cell lineages, with epithelial cells predominating within tumor samples. Notably, advanced-stage tumors demonstrated a shift toward an immunosuppressive microenvironment, characterized by elevated T cell exhaustion and an increased presence of pro-angiogenic macrophages. Further stratification of malignant cells revealed eleven distinct subclusters, with stage III tumors enriched for invasive and immunosuppressive phenotypes. Pseudotime trajectory analysis identified evolutionary transitions from alveolar-like states toward squamous and stress-responsive lineages. Employing CytoTRACE to evaluate stemness independently of predefined markers, we identified subpopulations of cancer stem cells (sub-CSCs) predominantly in stage III tumors. These sub-CSCs exhibited heightened transcriptional heterogeneity and a core gene signature associated with cytoskeletal motility, immune regulation, metabolic reprogramming, and protein homeostasis. A four-gene panel derived from this signature demonstrated robust prognostic significance in independent LUAD cohorts. Collectively, these findings enhance the current understanding of stemness-driven tumor progression in LUAD and suggest potential therapeutic targets to address tumor heterogeneity.