Single-cell and bulk RNA analysis identifies EMT-associated CAF signatures and prognostic model in lung adenocarcinoma.

Background: Cancer-associated fibroblasts (CAFs) have gained increasing attention for their role in lung adenocarcinoma (LUAD), particularly in epithelial-mesenchymal transition (EMT), which may play a key role in tumor microenvironment remodeling and cancer progression. This study aims to identify CAF signature genes associated with EMT through the integration of single-cell RNA sequencing and bulk transcriptomic data, and to construct a prognostic model for LUAD to further understand the molecular mechanisms of CAFs in LUAD and their clinical significance.

Methods: We integrated RNA sequencing data and clinical information of LUAD patients from the TCGA and GEO databases, along with single-cell RNA sequencing data for cell subtype identification and EMT activity analysis. The Seurat package was used for quality control, clustering, and EMT pathway activity quantification. Differential gene expression analysis was performed to identify EMT-CAF-related genes, and univariate Cox regression and LASSO regression were applied to construct a risk score model. Kaplan-Meier survival analysis and ROC curve analysis were used to evaluate the predictive performance of the model, along with further assessments of the tumor microenvironment, immune infiltration features, tumor mutational burden (TMB), and drug sensitivity.

Results: Single-cell RNA sequencing analysis identified eight major cell populations, with CAFs showing significant EMT activation. Differential expression analysis identified 84 EMT-CAF-related genes, from which eight key genes were selected to construct a risk score model. High-risk patients showed significantly worse survival outcomes compared to low-risk patients. Tumor microenvironment analysis indicated that the high-risk group had lower immune and stromal scores, with higher tumor purity. Immune cell infiltration analysis revealed a lower proportion of anti-tumor immune cells (such as activated B cells and CD8 + T cells) in the tumor tissue of the high-risk group, with a higher proportion of immune-suppressive cells. TMB analysis demonstrated significantly higher tumor mutational burden in the high-risk group compared to the low-risk group. Additionally, drug sensitivity analysis indicated that the high-risk group exhibited greater sensitivity to common anti-lung cancer drugs.

Conclusions: This study integrates single-cell and bulk transcriptomic data to uncover the potential role of EMT-CAF-related genes in LUAD and constructs a prognostic risk score model with significant clinical value. The model effectively predicts patient survival risk and provides new insights for risk stratification and personalized treatment. Moreover, we observed that high-risk patients have higher tumor mutational burden and immune suppression characteristics, suggesting that these features may influence disease progression. Further studies are needed to validate the clinical applicability and biological mechanisms of these findings.