Integration of single-cell and bulk RNA-seq to establish a predictive signature based on the differentiation trajectory of M2 macrophages in lung adenocarcinoma

Background: Tumor-associated macrophages as important members of the tumor microenvironment, are highly plastic and heterogeneous. TAMs can be classified into two preliminary subtypes: M1 and M2 macrophages. M2 macrophages are significantly associated with the progression of lung adenocarcinoma. However, no study has investigated the heterogeneity among M2 macrophages and their differentiation-related genes at the single-cell level to guide the clinical treatment of lung adenocarcinoma. Methods: Using the available annotation information from the Tumor Immune Single-cell Hub database, we clustered and annotated 12 lung adenocarcinoma samples using the R package ‘Seurat’. Subsequently, we extracted M2 macrophages for secondary clustering analysis and performed cell trajectory analysis using the R package ‘monocle2’. Based on heterogeneous genes associated with the differentiation trajectory of M2 macrophages, we established a prognostic lung adenocarcinoma model using Lasso-Cox and multivariate stepwise regression. In addition, we also performed immunotherapy and chemotherapy predictions. Results: M2 macrophages exhibit heterogeneity among themselves. M2 macrophages in different differentiation states showed significant differences in pathway activation and immune cell communication. Prognostic signature based on heterogeneous genes can be used to classify the prognostic status and abundance of immune cell infiltration in lung adenocarcinoma patients. In addition, the calculation of the Tumor Immune Dysfunction and Exclusion (TIDE) algorithm and the validation of the GSE126044 database indicated that lung adenocarcinoma patients with high-risk scores had poorer treatment outcomes when receiving immune checkpoint inhibitors treatment. Conclusion: Based on scRNA-seq and Bulk-seq data, we identified M2 macrophage-associated prognostic signature with a potential clinical utility to improve precision therapy.