Researchers at the Keck School of Medicine of USC have grown immune cell precursors called granulocyte-monocyte progenitors, or GMPs, in the lab for long periods, then engineered them to slow leukemia and solid tumors in mice. The results are preclinical. No person has received the cells.
The work appeared in the journal Cell on June 19, 2026, according to a USC Stem Cell release. The paper is titled “Expansion and CAR engineering of granulocyte-monocyte progenitors for cellular immunotherapy”.
Problem with Engineered Macrophages
Macrophages are immune cells that engulf pathogens and dead cells, and researchers want to equip them with chimeric antigen receptors, or CARs, to attack tumors. The USC release says mature macrophages are hard to expand, hard to engineer and hard to freeze. After infusion, they tend to settle in the lungs and liver rather than spread through the body.
The Cell abstract states the problem in similar terms: poor expansion outside the body, difficult genetic engineering and uncertain distribution after transfer. The USC team went one step earlier in the cell’s development. GMPs are the precursors that give rise to macrophages and other myeloid immune cells. If the precursors can be multiplied and engineered, the body can produce macrophages from them, which the release says may avoid the rapid clearance seen with mature macrophage therapies.
What Team Did
The team used a defined chemical mixture to keep mouse and human GMPs multiplying for long periods without maturing. The Cell abstract says the cells kept their identity and their ability to make myeloid cells, and the authors call the cells “a renewable engineering platform.” The abstract also names myeloperoxidase, an enzyme, as a regulator of GMP proliferation.
Once transferred into mice, the expanded GMPs settled in the bone marrow niches where blood cells form and produced cells derived from the donor. In mice with chronic granulomatous disease, an inherited immune deficiency, those cells restored the ability to fight bacterial infection. In tumor models they produced abundant macrophages inside tumors.
The USC release adds a point about self-renewal. Qi-Long Ying, the corresponding author and a professor of stem cell biology and regenerative medicine at the Keck School, said the team found that GMPs can also self-renew under the right conditions. The release frames that as a challenge to the view that long-term self-renewal in the blood system belongs mainly to hematopoietic stem cells.
Cancer Results, Tissue Mismatch
The team added CARs to GMPs. According to the abstract, the engineered cells suppressed CD19-positive leukemia and HER2-positive solid tumors in mice. CD19 is a marker on certain blood cancer cells. HER2 is a marker found on some solid tumors. A second design carried an added signal in the form of a CAR containing an IgG Fc domain.
The abstract says it recruited host immune cells that carry Fc receptors, allowed T cells to be primed even when donor and recipient tissue types did not match, and improved results in mouse cancer models with intact immune systems. The release says results were strongest with both the CAR and the added signal.
The mismatch finding matters for manufacturing. If cells can work across tissue types, one batch could serve many patients, which is what researchers mean by off-the-shelf. The release says this raises that possibility. It does not say a product exists or that human testing has been scheduled.
Ravi Majeti, director of the Institute for Stem Cell Biology and Regenerative Medicine at Stanford, said in the release that the method opens the door to many translational applications. Majeti’s lab independently reproduced the long-term maintenance and engineering of GMPs, the release says.
Funding, Financial Ties
The release lists funding from the Chen Yong Foundation of the Zhongmei Group, the L.K. Whittier Foundation, an Eli and Edythe Broad Innovation Award, a Ming Hsieh Institute cancer award, a USC small business planning award, the Xia Research Fund and the Wu & Jiang Research Fund.
It also lists sponsored research from a company, Myelogene Inc. Majeti reports support from the Ludwig Institute for Cancer Research, and co-author Zheng Guo was supported by a California Institute for Regenerative Medicine fellowship. The disclosures link the company to the authors. Several authors, including Ying, first author Shi Yue, Majeti and Chao Zhang, are co-inventors on patents that USC licensed to Myelogene.
The release says Ying, Yue, Zhang and Majeti co-founded the company. The release says Majeti also serves on several advisory boards and co-founded Pheast Therapeutics. The release’s main funding list has eight entries, and one of them is that company. That count is a calculation from the release. The release states these ties openly. The lead investigators could benefit if the platform becomes a product, and the lab that independently reproduced the work belongs to Majeti, a company co-founder.
What Remains Unproven
The release says the cancer and disease findings come from mice. It says safety and effectiveness in people, particularly cell persistence over the long term, still need to be established. The mouse work covered one blood cancer marker, CD19, and one solid tumor marker, HER2. The release does not report survival figures, effect sizes or durations of benefit, and the Cell paper is behind a subscription, so those details are not available for this report.
Ying put the claim in these terms in the release: “The study establishes a scalable and engineerable GMP platform for cellular immunotherapy.” Before any trial in people, the team would need data on how long the engineered cells persist and how they behave in human bone marrow, which are the questions the release names as open.
