Comprehensive characterization of T cell subtypes in lung adenocarcinoma: Prognostic, predictive, and therapeutic implications

Abstract

Background

T cells are crucial for immunosurveillance and tumor eradication, with their dysregulation or absence in the tumor microenvironment linked to immunotherapy resistance. In lung adenocarcinoma (LUAD), this resistance is a significant barrier to effective treatment, highlighting the need for robust biomarkers and therapeutic targets to improve clinical outcomes.

Methods

T cell-related markers were identified through single-cell RNA sequencing analysis. The TCGA dataset was used for consensus clustering to define molecular subtypes associated with distinct survival outcomes and immune profiles. A T cell-related prognostic signature was developed by integrating LUAD datasets from TCGA, GSE31210, GSE50081, and GSE68465 using 10 machine learning algorithms. Further analysis linked risk scores to immune infiltration and drug sensitivity. The role of a hub gene in CD4+ T cell function and its involvement in tumor immunity was explored through in vitro experiments and molecular biology techniques.

Results

Cluster analysis identified three LUAD subtypes, with cluster1 showing the best prognosis and immune characteristics. A Lasso + PLSRcox-based signature was a significant risk factor for predicting LUAD patient outcomes, outperforming traditional clinicopathological factors. The risk score correlated with immune microenvironment features, immune cell infiltration, and sensitivity to immunotherapy and chemotherapy. CPA3 expression was elevated in activated CD4+ T cells, particularly in Th1 cells, promoting differentiation and IFN-γ secretion. Overexpression of CPA3 enhanced tumor cell apoptosis and increased Granzyme B and IFN-γ levels, highlighting its role in immune responses.

Conclusion

We developed a powerful prognostic signature in LUAD that accurately predicts clinical outcomes and can guide immunotherapy and chemotherapy responses.