Reprogramming tumor-associated macrophages in breast cancer: from myeloid plasticity and tumor-promoting programs to CAR-macrophage therapy.

Breast cancer develops within spatially organized, multicellular microenvironments in which particular tumor-associated macrophage (TAM) states can influence immune access and stromal remodeling. Conventional M1/M2 classification and total macrophage density fail to capture the developmental, transcriptional, metabolic, and spatial heterogeneity that determines TAM function. This macrophage-focused review examines breast cancer within its multicellular microenvironment, integrating lineage-tracing, single-cell, spatial, mechanistic, and clinical evidence while distinguishing breast-specific findings from hypotheses extrapolated from other tumors. We examine how tissue-resident and monocyte-derived macrophages are recruited, retained, expanded, and reprogrammed across primary and metastatic sites, emphasizing interactions with malignant cells, cancer-associated fibroblasts, extracellular matrix, vasculature, adipose tissue, and lymphoid populations. In defined breast-tumor contexts, TAM programs have been implicated in immune exclusion, invasion, metastatic dissemination, and adaptation to systemic treatment, although evidence strength varies across mechanisms and therapeutic modalities. Small and heterogeneous clinical studies have not established a benefit from broad CSF1/CSF1R-axis targeting in breast cancer, despite encouraging preclinical findings. Accordingly, we highlight a shift toward state-selective, spatially resolved, and temporally coordinated targeting. Emerging approaches include restoring phagocytosis, disrupting chemokine and fibrotic circuits, metabolic reprogramming, and converting suppressive TAMs into inflammatory, antigen-presenting effectors. A major focus is chimeric antigen receptor macrophage (CAR-M) therapy, which combines antigen-directed phagocytosis with tumor infiltration, matrix remodeling, antigen presentation, and recruitment of endogenous immunity. Although early clinical data support manufacturing feasibility, tumor trafficking, and acceptable initial safety, limited antitumor activity underscores challenges involving antigen heterogeneity, persistence, metabolic fitness, biodistribution, and scalable production. We propose a translational roadmap integrating spatial biomarkers, quantitative target selection, macrophage-adapted CAR design, checkpoint-resistant phagocytosis, metabolic engineering, safety switches, and biomarker-guided combination therapy. This framework positions macrophage reprogramming as a context-dependent strategy for modifying selected breast-tumor niches, with therapeutic targets chosen according to the macrophage program and its surrounding cellular architecture.