Macrophages are attracting increasing attention as candidates for cancer immunotherapy because they can infiltrate tumors, engulf abnormal cells, and reshape the tumor microenvironment. These properties distinguish them from conventional T cell-based therapies and make them particularly appealing for treating solid tumors, where immune cell access and local immune suppression often limit therapeutic effectiveness. However, generating macrophage products from patient-derived monocytes remains challenging due to donor variability, limited expansion capacity, and difficulties in genetic modification. Researchers have therefore sought alternative platforms that can provide a consistent, scalable source of therapeutic macrophages.

To address this need, the researchers investigated pluripotent stem cell-derived myeloid cell lines, a renewable platform that can be expanded indefinitely and readily engineered. Using two independent human pluripotent stem cell lines, they introduced a doxycycline-inducible anti-HER2 CAR and generated macrophage-like cells through directed differentiation. The resulting cells displayed characteristic macrophage morphology and expressed a range of macrophage-associated markers, demonstrating the feasibility of generating a renewable, genetically engineered myeloid cell product.

The research team then evaluated whether these cells could function as direct tumor-targeting effectors. In laboratory co-culture experiments with HER2-expressing cancer cells, the macrophage-like cells reduced tumor cell viability and actively engaged with cancer targets. Live-cell imaging visualized macrophage-like cells interacting and engulfing tumor cells, confirming their phagocytic activity. Although CAR induction appeared to enhance anti-tumor responses in some settings, this effect was not consistently observed across all tumor models, suggesting that CAR-mediated activity alone may not fully explain the observed tumor control.

To better understand how the cells responded to tumor engagement, the researchers analyzed changes in gene expression. Contact with tumor cells triggered broad activation of inflammatory pathways, as well as programs involved in cellular trafficking, antigen presentation, and cellular uptake. These changes reflected a shift from a resting state toward a more activated immune phenotype. When the CAR was induced, the transcriptional response shifted toward pathways involved in intracellular trafficking and lysosomal processing, indicating that CAR signaling may influence how macrophages respond after encountering tumor cells. These findings provide new insight into how engineered macrophages respond at the molecular level during interactions with cancer cells.

The researchers then assessed the cells in a mouse co-implantation model in which tumor cells and macrophage-like cells were introduced together, allowing evaluation of early tumor-cell engagement. While treated animals showed signs of short-term tumor control, cell-tracking studies revealed a striking limitation. The administered macrophage-like cells declined rapidly after transplantation, with detectable signals falling sharply within three days and disappearing entirely by seven days. Repeated administration was feasible but did not clearly overcome this durability problem.

These findings highlight a critical distinction between manufacturing potential and therapeutic persistence. Pluripotent stem cell-derived myeloid cell lines provide a renewable and highly engineerable source of macrophage-based cell products, but unlimited expansion of the source cells does not ensure long-term survival of the final therapeutic product after administration. By identifying limited persistence as a key bottleneck, this study provides an important framework for future efforts to improve cell survival, retention, and functional durability. Such advances could help unlock the promise of scalable, off-the-shelf macrophage therapies for patients with solid tumors.

• doi: 10.1016/j.jcyt.2026.102976