A protein scaffold tripled the stem cell content of CAR-T cells - and they kept killing cancer for months

CAR-T therapy is often described as a living drug. Doctors extract a patient's T cells, genetically reprogram them to target cancer, and infuse them back. The initial results can be dramatic - tumors vanish, remission is declared. Then, in roughly half of treated patients, the cancer comes back. The living drug dies before the job is done.

A study published March 13 in Science Advances by researchers at Albert Einstein College of Medicine describes a manufacturing change that may substantially extend how long these engineered cells survive and fight.

The persistence problem

Standard CAR-T manufacturing uses chemical signals to activate T cells before genetic reprogramming. The process works, but it produces cells that are heavily skewed toward short-lived effector types - potent killers that burn out quickly. Fewer than 5% of conventionally manufactured CAR-T cells belong to the T memory stem cell (Tscm) population, a rare subset capable of self-renewal and generating fresh waves of active immune fighters over time.

Harris Goldstein, the study's senior author and professor of pediatrics and microbiology at Einstein, wanted to change that ratio. His team developed an alternative activation approach using a protein scaffold called HCW9206 that links three naturally occurring cytokines: IL-7, IL-15, and IL-21. Each of these signaling molecules is known to promote T cell survival and immune memory. Linking them together created a sustained, coordinated signal that pushed T cells toward the stem-like phenotype during manufacturing.

From 5% to more than 50%

The results were stark. More than half of CAR-T cells produced with the multi-cytokine scaffold displayed the T memory stem cell profile, compared to less than 5% with conventional methods. The scaffold-produced cells retained strong disease-fighting abilities while gaining the capacity for self-renewal - exactly the combination needed for long-term persistence in the body.

In a mouse model of human leukemia, both conventional and scaffold-produced CAR-T cells eliminated cancer after initial treatment. The difference appeared during relapse simulation. When researchers re-infused leukemia cells weeks later, only the scaffold-engineered CAR-T cells mounted a strong recall response - expanding in number and preventing tumor recurrence. Conventional CAR-T cells, their numbers depleted, failed to respond.

An HIV application

The team also tested the approach against HIV. Current antiretroviral therapy (ART) suppresses HIV to undetectable levels but cannot eliminate cells harboring dormant virus. If ART is stopped, the virus rebounds from these long-lived reservoirs. A functional cure would require immune cells capable of patrolling the body for years, hunting down and destroying residual infected cells.

In a humanized mouse model, scaffold-generated CAR-T cells eliminated significantly more HIV-infected cells than conventionally manufactured ones. The team also produced CAR-T cells from patients already living with HIV and demonstrated that those cells could eradicate HIV-infected cells in the lab.

Mouse models, human questions

These are animal studies, and the gap between mouse efficacy and human outcomes is wide in cell therapy. Humanized mouse models approximate but do not fully reproduce the complexity of a functioning human immune system. The long-term safety of infusing a higher proportion of stem-like T cells - which by definition persist and self-renew - requires careful evaluation. There is a theoretical concern that extremely long-lived engineered cells could themselves become problematic, though no evidence of that emerged in this study.

Manufacturing scalability is another open question. The HCW9206 scaffold is a proprietary protein produced by HCW Biologics, and translating the approach from research-scale to clinical-scale production involves regulatory, logistical, and cost hurdles that the study does not address.

For cancer patients, the immediate promise is a potential reduction in relapse rates after CAR-T therapy. For people living with HIV, it represents a step toward the elusive goal of sustained viral control without lifelong medication - though that endpoint remains years of clinical testing away.