Abstract

Background Lung adenocarcinoma (LUAD) frequently leads to brain metastasis (BM), representing a significant clinical challenge associated with poor patient outcomes. While the tumor microenvironment (TME) is known to influence progression, the specific immune cell programs and mechanisms underpinning cross-organ metastatic evolution remain incompletely defined. In this study, we systematically mapped cellular composition and macrophage-state transitions to investigate the functional remodeling of tumor-associated macrophages (TAMs), specifically dissecting the CD74High TAM subpopulation throughout the metastatic process. Methods We reanalyzed a published single-cell RNA-sequencing dataset (GSE131907) comprising LUAD primary lung tumors (LTs) and LUAD with BMs. Following standardized preprocessing, cell-type annotation, and inferCNV-based inference of the copy number variation (CNV)—somatic gains or losses of large DNA segments used to identify malignant cells—we reconstructed ligand-receptor networks using CellChat. The TAM population was integrated across organs to perform differential expression analysis, pathway enrichment, AUCell activity scoring, GeneNMF-based meta-program discovery, and pseudotime trajectory reconstruction to delineate TAM evolution. Results Our analysis revealed that the TME retained several conserved communication modules across organs, among which the APP-CD74 axis remained highly active and preferentially associated with CD74High TAMs. However, functional states diverged significantly by tissue site. CD74High TAMs in LT exhibited immune-clearance features enriched for phagosome, antigen processing, and transendothelial migration programs. In contrast, brain-metastatic CD74High TAMs displayed a coordinated attenuation of phagocytosis-associated programs, accompanied by heightened inflammatory signaling, metabolic reprogramming, and stress-adaptation states.