Identification of distinct tumor cell patterns with single-cell RNA sequencing integrating primary lung adenocarcinoma and brain metastasis tumor

Background: Lung adenocarcinoma (LUAD) is the most common form of lung cancer and is often accompanied by brain metastasis (BM). The heterogeneity of the tumor renders all current conventional treatments less effective. This study aims to dissect tumor cell heterogeneity and identify potential therapeutic targets. Methods: We conducted single- cell RNA-sequencing (scRNA-seq) in 8 patients with treatment-naive LUAD BM and included scRNA-seq data of 10 primary LUAD samples and their matched adjacent normal tissue from GSE131907 to determine the tumor cell heterogeneity. Results: Our analyses revealed tumor cells derived from brain metastases were more heterogeneous. Tumor cells from BM harbored significantly more copy number variants (CNVs), and cells of magnoid subtype were the critical source of malignant cells both in BM and the primary lung tumor. Pseudo-time trajectory analysis revealed that malignant cells had upregulated genes enriched for cell cycle and cell division. Integrated analysis of tumor cells revealed 2 distinct malignant cell clusters (cluster 4 and cluster 6) and their marker genes. The signatures identified in the single-cell profile had prognostic value in the bulk tumor profiles. Moreover, the signature of cluster 4 had significant prognostic value in predicting patients surviving longer than 3.5 years, while the signature of cluster 6 showed better predictive ability within 1 year. Magnoid-type cells are most likely to develop into the riskiest cell type and potentially promote tumor progression. Conclusions: scRNA profiling that integrates LUAD BM and primary LUAD can provide information on those malignant cells with BM potential, offering additional prognostic information at cellular level, and may serve as a foundational resource for further tumor cell dissection and therapeutic target exploration.