AINEWS
Two Labs Found the Same Cancer Target. The Convergence Is the Point.
The surface protein GPNMB sits on both cancer cells and the immune cells that shield them, which is why separate research groups landed on it months apart.
The surface protein GPNMB sits on both cancer cells and the immune cells that shield them, which is why separate research groups landed on it months apart.
Two Nature papers published months apart reached the same destination through independent routes: the surface protein GPNMB, expressed on both tumor cells and the myeloid cells surrounding them, is a viable CAR T target in solid tumors. That convergence is not coincidence. It is how new drug targets become credible in oncology.
CAR T therapy, which engineers a patient's own T cells to attack a chosen surface marker, has transformed blood cancers since the FDA approved the first CD19-directed therapy in 2017. Subsequent CD19 and BCMA approvals have produced durable remissions in leukemia, lymphoma, and multiple myeloma, the arc the field wants to repeat in solid tumors. Solid tumors have been harder, and the failures share a shape. They are heterogeneous, so a single marker rarely covers every cancer cell. They sit behind an immunosuppressive tumor microenvironment called the TME, where myeloid cells and other suppressors blunt the T cells' punch. Targets that worked in blood cancers usually break against one of those walls.
The protein appears on tumor cells themselves and on tumor-associated macrophages, the myeloid cells that make up much of the protective shield. A CAR T directed at GPNMB can attack the cancer and the shield from one receptor, a different biological problem than targeting either alone. That dual expression is also why the target held up in two different tumor types instead of failing the way most solid-tumor candidates do.
The two papers chose those tumor types separately. One team, writing in Nature, aimed at glioblastoma, the aggressive brain cancer with a five-year survival rate that has barely moved in two decades and no clearly curative option after surgery. Glioblastoma sits behind the blood-brain barrier, where T-cell trafficking is limited, on top of the standard immunosuppressive microenvironment. A separate group, publishing in Nature Cancer, focused on solid tumors driven by MiT/TFE-family gene fusions, a rare subset of renal, skin, and other cancers in which TFE translocations drive the disease. Neither team knew about the other's work when it started, according to a Singularity Hub explainer that drew on both papers.
The MiT/TFE track also has a longer pedigree. A 2025 MedRxiv preprint (not peer-reviewed) described first-in-human development of a GPNMB-directed CAR T for the same fusion-driven tumors. The 2026 Nature Cancer paper is the peer-reviewed confirmation. Together they form a continuous line rather than a single flash, and a Medical Xpress summary of the glioblastoma work echoes the same two-pronged framing for the brain tumor side.
When two separate groups, working in different tumor types, name the same molecule within months, the underlying biology is usually telling them something real. GPNMB now has that kind of validation: two peer-reviewed papers and a preprint lineage.
What this is not yet: a survival benefit for anyone. Both Nature papers and the preprint work in preclinical or early-human settings, where "activity" means the engineered T cells engage the target and the tumors respond in dish or animal models. Proving that GPNMB CAR T extends lives in solid-tumor patients will take the kind of trials that took CD19-directed therapy nearly a decade from first approval to curative option in many patients.
The clear next move is the glioblastoma team's push into clinical testing. If a CAR T directed at GPNMB shows measurable activity in patients against a brain tumor that has resisted nearly every modern therapy, the field will treat 2026 as a turning point. If it does not, GPNMB joins the list of solid-tumor targets that looked unbeatable in mice and faded in humans.