The bottleneck nobody could engineer around

CAR-T cell therapy has produced some of modern oncology's most striking results, driving certain advanced blood cancers into deep, lasting remission after every other option had failed. But the therapy has a supply problem that no amount of clever receptor engineering has solved: each treatment typically requires collecting and modifying a patient's own T cells, a process that's expensive, slow, and doesn't scale into anything resembling an off-the-shelf product. Extending similar cell-based approaches to solid tumors, which make up the vast majority of cancer diagnoses, has proven even harder.

A team at the University of Southern California may have found a way around that bottleneck entirely, and the mechanism they uncovered upends a long-standing assumption in stem cell biology. In a study published June 19, 2026 in the journal Cell, researchers led by Dr. Qi-Long Ying at the Keck School of Medicine of USC showed that a class of immune cell precursors called granulocyte-monocyte progenitors, or GMPs, can be coaxed into dividing indefinitely in the lab while retaining their full biological identity — a capability scientists had assumed belonged exclusively to stem cells.

Cells that were never supposed to renew themselves

The conventional wisdom in hematopoiesis, the study of how blood and immune cells form, has held that long-term self-renewal is a defining property of hematopoietic stem cells specifically. Progenitor cells like GMPs sit further along the developmental pathway — they're already partially specialized, on their way to becoming mature macrophages or other immune cells, and were thought to have only limited capacity to keep dividing before running out of steam.

Ying's team found that assumption doesn't hold under the right laboratory conditions. Using what the researchers describe as a defined culture system — a precisely tailored chemical environment — they coaxed GMPs to keep dividing extensively without maturing out or losing their identity, essentially unlocking a self-renewal capacity nobody expected these cells to possess. "The prevailing view has been that long-term self-renewal in the blood system is primarily a property of the hematopoietic stem cells that can generate any type of blood or immune cell," Ying said in comments distributed through USC's press materials. "We found that, under the right conditions, GMPs can also self-renew, dividing extensively while keeping their identity and ability to produce functional immune cells."

From lab-grown cells to cancer-hunting machines

Discovering that GMPs could self-renew was only the first half of the study. The more consequential test came next: could these expanded progenitors actually be turned into a cancer-fighting tool? The researchers engineered the self-renewing GMPs with a chimeric antigen receptor, or CAR — the same basic technology underlying CAR-T therapy — designed to recognize specific cancer markers. They then added a second engineered signal intended to recruit and activate the patient's own T cells, layering two distinct anti-cancer mechanisms onto a single cell platform.

When the engineered GMPs were injected into tumor-bearing mice, the cells engrafted successfully in the bone marrow and continuously replenished a supply of anti-tumor macrophages over time, rather than being used up in a single wave and requiring re-dosing. According to reporting from Jerry Cards' coverage of the study, the combination of both engineered features together outperformed either modification used alone, and the approach slowed progression in both blood cancers and solid tumors in the animal models tested — the solid tumor result being particularly notable, given how much harder solid cancers have historically been for cell-based immunotherapies to reach effectively.

A second use nobody was originally looking for

The study turned up a further application that wasn't the original focus of the research. In a separate set of experiments involving an inherited immune disease model, the engineered GMPs restored the animals' ability to fight off infection, according to a summary of the findings circulated by Jerry Cards. That's a meaningful expansion of what this platform might eventually be used for — not just hunting cancer cells, but potentially reconstituting functional immunity in patients whose immune systems are compromised by inherited conditions, not only by malignancy.

That dual utility matters because it suggests the underlying platform — renewable, engineerable myeloid progenitors — isn't a narrow fix for one cancer-therapy supply problem. It looks more like a foundational cell-manufacturing technology that could feed into multiple distinct treatment categories, depending on how the cells are engineered downstream.

Why "off-the-shelf" is the phrase that matters most here

The practical implication driving most of the excitement around this study is manufacturing, not biology for its own sake. Current CAR-T therapies are personalized by necessity — manufactured from each individual patient's own cells, which makes the process slow, costly, and logistically difficult to scale for widespread use. A cell source that self-renews indefinitely, tolerates freezing, and appears tolerant of donor-recipient mismatch, as researchers described in coverage of the study, opens the door to a genuinely different manufacturing model: cells produced once, in bulk, and distributed to many patients rather than custom-built for each one.

Ying framed the broader implication succinctly in comments reported by TechExplorist: the future of immunotherapy may depend not just on designing better CAR receptors, but on choosing the right developmental stage of cell to build them on. That's a subtly different research direction than most of the field has pursued over the past decade, which has focused heavily on receptor design and T-cell engineering specifically. This study suggests the earlier-stage progenitor cell itself, previously overlooked because nobody thought it could self-renew, may turn out to be just as important a lever. The work remains preclinical — confined to mouse models rather than human trials — but it reframes a manufacturing constraint that has limited cell therapy's reach for years as a solvable engineering problem rather than a biological ceiling.

*This article was researched using publicly available reporting from Cell, USC Stem Cell, EurekAlert, ScienceDaily, TechTimes, TechExplorist, and Stanford Medicine's coverage of the peer-reviewed study. It is intended for informational purposes and is not medical advice.*