A protein bridge made of ferritin helps CAR T cells find leukemia even when the tumor hides

More than half of leukemia patients treated with CAR T cell therapy eventually relapse. The reason is elegant and frustrating: leukemia cells learn to hide. Under therapeutic pressure, they reduce or eliminate expression of the surface antigen that CAR T cells are engineered to recognize. Once the target disappears, the hunter becomes blind.

Previous attempts to solve this problem have focused on redesigning the CAR itself, adding new targeting domains through additional rounds of genetic engineering. That approach is technically complex, expensive, and slow. A team from the Institute of Process Engineering (IPE) at the Chinese Academy of Sciences has taken a fundamentally different approach, and the results, published in Cell, suggest it works.

Building a molecular bridge from ferritin

The researchers started with a biological observation. By analyzing a large number of clinical samples in collaboration with Zhujiang Hospital and the Institute of Hematology and Blood Diseases Hospital, they found that a protein called CD71, the transferrin receptor responsible for transporting iron into cells, is highly expressed on leukemia cells across different disease types and stages. CD71 also appears on the CAR T cells themselves.

This dual expression suggested an opportunity. Ferritin, the body's natural iron-storage protein, is the biological ligand for CD71. By controlling the assembly conditions, the researchers induced ferritin molecules to self-assemble into an ordered aggregate they call FACE (ferritin aggregation cell engager), a molecular bridge that can bind CD71 on both the CAR T cell and the leukemia cell simultaneously.

During CAR T cell preparation, FACE attaches firmly to CD71 on the surface of the CAR T cells. After infusion into the patient, the FACE molecules on the CAR T cells also bind to CD71 on leukemia cells, creating a physical link that strengthens the interaction between hunter and target.

One-fifth the dose, same effect

In patient-derived xenograft (PDX) mouse models with normal antigen expression, FACE-CAR T cells achieved the same therapeutic effect as conventional CAR T cells at one-fifth the cell dose. This is clinically significant because lower doses mean less manufacturing time, lower cost, and importantly, a reduced risk of cytokine release syndrome, the sometimes-fatal inflammatory response that is one of CAR T therapy's most dangerous side effects.

When antigens drop below 10%

The more dramatic results came in models simulating relapsed disease. When leukemia antigen levels dropped below 10% of normal, a condition under which conventional CAR T cells were largely ineffective, FACE-CAR T cells still eliminated the cancer cells. In PDX models of this antigen-low condition, 100% of mice survived.

The researchers went further, developing a drug-loaded version called FACED, which exploits ferritin's cage-like structure to carry therapeutic cargo. FACED-CAR T cells effectively treated PDX models with initial leukemia burdens up to 40% and low antigen expression, and they also eliminated antigen-negative leukemia cells, the population most responsible for relapse.

No additional genetic engineering required

What distinguishes this approach from other strategies to improve CAR T therapy is what it does not require. FACE is composed of an endogenous protein (ferritin) and FDA-approved polymer derivatives. It integrates into existing CAR T cell manufacturing workflows as a culture supplement, co-incubated with CAR T cells before infusion. No additional genetic modification of the CAR construct is needed.

This matters for scalability. Redesigning CAR constructs requires new clinical-grade vector production, additional safety testing, and regulatory approval for each modification. A culture supplement that enhances existing approved CAR T products faces a substantially simpler path to clinical use.

Predictive framework

The team also established an efficacy database across diverse patient-derived samples and developed an AI-assisted predictive framework that can forecast FACE-mediated enhancement before treatment. If validated in clinical settings, this could help identify which patients are most likely to benefit.

Limitations and open questions

All results so far come from mouse models and patient-derived samples tested in vitro. Human clinical trials have not begun. The durability of the FACE coating on CAR T cells in a human body, where immune surveillance and protein turnover operate differently than in mice, remains unknown. And while CD71 is broadly expressed on leukemia cells, its expression on healthy tissues raises questions about on-target, off-tumor toxicity that will need to be carefully evaluated.

The approach is promising precisely because it is practical. It does not replace CAR T therapy. It augments it, using the body's own iron-transport biology to solve one of the therapy's most persistent problems.