The strategy rests on a molecular tug-of-war that determines whether a macrophage will engulf a tumor cell. Healthy cells display the surface protein CD47, which engages a receptor on macrophages and delivers a ‘don’t eat me’ signal, protecting normal tissue from collateral destruction. The researchers had previously shown that Ewing sarcoma cells exploit this system by upregulating CD47 while simultaneously downregulating cell-surface calreticulin, an ‘eat me’ signal that normally flags stressed or malignant cells for removal. The tumor, in other words, hides its own alarm while waving a shield. The team’s earlier work showed that blocking CD47 with the antibody magrolimab, while using the chemotherapy drug doxorubicin to push calreticulin back onto the tumor cell surface, dramatically enhanced macrophage phagocytosis of Ewing sarcoma cells in laboratory dishes and suppressed tumor growth and metastasis in an orthotopic mouse model.

That earlier combination, however, carried serious liabilities. Doxorubicin is a cornerstone of Ewing sarcoma treatment, but it is notorious for causing cardiotoxicity, a cumulative injury to heart muscle that can surface years after treatment, a particularly troubling prospect for pediatric survivors. Magrolimab, for its part, binds CD47 on red blood cells, triggering their clearance by the spleen and producing anemia, a side effect that has complicated the clinical development of first-generation anti-CD47 antibodies across the field. The new study set out to answer a deceptively simple question: could the therapeutic principle be preserved while swapping out the components responsible for the worst toxicities?

The first step was diagnostic. The researchers performed single-cell RNA sequencing on cells isolated from Ewing sarcoma xenograft tumors that had been treated with a phosphate buffer control, doxorubicin alone, magrolimab alone, or the doxorubicin plus magrolimab combination. Single-cell transcriptomics allows investigators to profile gene expression in individual cells rather than in bulk tissue, revealing how each cell type within the tumor microenvironment responds to therapy. The results were sobering: doxorubicin, whether given alone or with magrolimab, upregulated cellular pathways and functions associated with dilated cardiomyopathy in macrophages. The sequencing data suggested that the very drug used to sensitize the tumor to immune attack was simultaneously driving molecular changes in the tumor-associated immune compartment consistent with heart muscle disease, reinforcing the need for an alternative partner drug.

The team turned to cyclophosphamide, another standard chemotherapy agent for Ewing sarcoma, as a candidate replacement for doxorubicin. In a series of dose-response experiments, cyclophosphamide significantly enhanced cell-surface calreticulin expression on Ewing sarcoma cells in a dose-dependent manner, replicating the key sensitizing effect of doxorubicin without its cardiomyopathy-associated transcriptional signature. This matters because calreticulin exposure is the mechanistic linchpin of the approach: when the ‘eat me’ signal reappears on the tumor surface, CD47 blockade tips the balance decisively toward phagocytosis. The finding established that the dual-targeting logic of the therapy could be maintained with a chemotherapy agent already familiar to pediatric oncologists.

The second substitution addressed the anemia problem. Rather than magrolimab, the researchers tested lemzoparlimab, a next-generation anti-CD47 antibody engineered to spare erythrocytes. First-generation antibodies such as magrolimab bind CD47 indiscriminately, and because red blood cells are among the most CD47-abundant cells in the body, they are cleared en masse, producing the anemia that has limited dosing in clinical trials. Lemzoparlimab was designed with modified binding properties that reduce its interaction with red blood cells while preserving high-affinity blockade of CD47 on tumor cells. In in vitro phagocytosis assays, combining cyclophosphamide with either lemzoparlimab or magrolimab markedly increased macrophage engulfment of Ewing sarcoma cells, confirming that the newer antibody retained full functional potency in the presence of the calreticulin-inducing chemotherapy.

The decisive tests came in living animals. Using orthotopic Ewing sarcoma xenograft mouse models, in which tumors are established in the anatomically relevant site rather than under the skin, the researchers evaluated cyclophosphamide combined with magrolimab or lemzoparlimab in immunodeficient NSG hosts. Both combinations significantly reduced tumor growth and lung metastasis while prolonging animal survival compared with controls. These results demonstrated that swapping doxorubicin for cyclophosphamide did not sacrifice the anti-tumor and anti-metastatic efficacy that had made the original combination so compelling, and that either anti-CD47 antibody could deliver the phagocytic signal when paired with the calreticulin-inducing chemotherapy.

Safety, however, could only be properly assessed in a model with a functioning human hematopoietic system. In humanized NSG mice, mice engrafted with human immune and blood-forming cells, the difference between the two antibodies became stark. The cyclophosphamide plus magrolimab combination induced severe anemia and animal death, mirroring the red blood cell toxicity predicted from the antibody’s mechanism. The cyclophosphamide plus lemzoparlimab combination, by contrast, was well tolerated: treated animals maintained their blood counts, showed significantly reduced tumor burden, and enjoyed extended survival. The side-by-side comparison in humanized hosts provides unusually direct preclinical evidence that the erythrocyte-sparing design of lemzoparlimab translates into a genuine safety advantage in the context of this combination therapy.

The significance of the work extends beyond Ewing sarcoma itself. The CD47-calreticulin axis is exploited by a wide range of malignancies, and the field has struggled to reconcile the potent anti-tumor activity of CD47 blockade with its hematologic toxicity. By demonstrating that a chemotherapy agent can be selected not only for tumor-killing activity but for its ability to induce calreticulin without cardiomyopathy-associated transcriptional changes, and that an erythrocyte-sparing antibody can substitute for a first-generation blocker without loss of efficacy, the study offers a template for rational optimization of innate immunotherapy combinations. The use of single-cell RNA sequencing to monitor off-target pathway activation in the tumor microenvironment also illustrates how transcriptomic surveillance can flag toxic liabilities before they reach patients.

The authors caution that the findings remain preclinical, and the road from humanized mouse models to pediatric clinical trials involves regulatory, dosing and scheduling questions that animal studies cannot fully answer. Still, the combination of cyclophosphamide and lemzoparlimab brings together two agents with existing clinical pedigrees, one a decades-old standard of care in Ewing sarcoma and the other an antibody already advancing through oncology trials, which may ease the path to translation. For a disease in which new options for relapsed and metastatic patients are desperately needed, a safe and effective way to turn macrophages against the tumor represents a meaningful step forward. The research was supported in part by the National Cancer Institute Cancer Moonshot and the Children’s Cancer Fund, and the published data, the authors conclude, position the cyclophosphamide plus lemzoparlimab regimen as a therapeutic strategy with high potential for clinical translation in patients with Ewing sarcoma.

Subject of Research: Combinatorial macrophage-mediated innate immunotherapy with cyclophosphamide and lemzoparlimab for Ewing sarcoma

Article Title: Optimizing combinatorial macrophage induced innate immunotherapy against Ewing sarcoma

Article References: Luo, W., Zhu, H., Cannon, M. V., Gross, A., Rosenblum, J. M., Bellantoni, A. J., Mo, X., Roberts, R., Cripe, T. P., & Cairo, M. S. (2026). Optimizing combinatorial macrophage induced innate immunotherapy against Ewing sarcoma. Journal of Experimental & Clinical Cancer Research. https://doi.org/10.1186/s13046-026-03840-1

Image Credits: AI Generated

DOI: 10.1186/s13046-026-03840-1

Keywords: Ewing sarcoma, macrophages, CD47 blockade, lemzoparlimab, cyclophosphamide, calreticulin, magrolimab, immunotherapy, single-cell RNA sequencing, humanized mouse model, pediatric cancer, anemia