Single-cell transcriptomics reveals the drivers and therapeutic targets of lymph node metastasis in lung adenocarcinoma.

Lymph node metastasis (LNM) is usually the most common metastatic pathway in lung adenocarcinoma (LUAD) and is associated with a poorer prognosis and higher possibility of recurrence. Therefore, discovering the drivers and therapeutic targets of LNM is important for early and non-invasive detection of patients with a high risk of LNM and guiding individualized therapy. Various cell constitutions of the primary tumor and lymph node microenvironment was characterized based on scRNA-seq data. The copy number variation (CNV) analysis was performed to probe clonal structures and origins of metastatic lymph nodes, and found 6q loss and 20q gain may drive LNM in LUAD. Then a LNM-related cell subset, named Scissor+ cells, was identified using the Scissor algorithm. And cell-cell communication network among Scissor+ cells and microenvironment was further analyzed. Besides, a pro-LNM signature was subsequently constructed based on 27 genes using pseudotime trajectory analysis and gene set variation analysis. The pro-LNM signature showed a significant correlation with stage and a good predictive ability of LUAD survival. At last, we identified that erastin and gefitinib could potentially inhibit LNM by targeting Scissor+ cells based on the drug sensitivity data of the cancer cell lines, which provided new insights for LUAD therapy.