Dissecting the single-cell transcriptome network of macrophage and identifies a signature to predict prognosis in lung adenocarcinoma

In conclusion, we investigated the composition, expression differences, and phenotypic changes of macrophages from different origins in lung adenocarcinoma. In addition, we developed a prognostic prediction model based on different macrophage subtype infiltration, which can be used as a valid prognostic biomarker. New insights were provided into the role of macrophages in the prognosis and potential treatment of LUAD patients.

We performed single-cell analysis using 21 lung adenocarcinoma (LUAD), 12 normal, and four peripheral blood samples from our data and public databases. A prognostic prediction model was then constructed using 502 TCGA patients and explored the potential factors affecting prognosis. The model was validated using data from 4 different GEO datasets with 544 patients after integration.

The tumor immune microenvironment (TME) plays a vital role in tumorigenesis, progression, and treatment. Macrophages, as an important component of the tumor microenvironment, play an essential role in antitumor immunity and TME remodeling. In this study, we aimed to explore the different functions of different origins macrophages in TME and their value as potential predictive markers of prognosis and treatment.

According to the source of macrophages, we classified macrophages into alveolar macrophages (AMs) and interstitial macrophages (IMs). AMs mainly infiltrated in normal lung tissue and expressed proliferative, antigen-presenting, scavenger receptors genes, while IMs occupied the majority in TME and expressed anti-inflammatory, lipid metabolism-related genes. Trajectory analysis revealed that AMs rely on self-renew, whereas IMs originated from monocytes in the blood. Cell-to-cell communication showed that AMs interacted mainly with T cells through the MHC I/II signaling pathway, while IMs mostly interacted with tumor-associated fibrocytes and tumor cells. We then constructed a risk model based on macrophage infiltration and showed an excellent predictive power. We further revealed the possible reasons for its potential prognosis prediction by differential genes, immune cell infiltration, and mutational differences.