Treat glioblastoma as a connected tumor-immune ecosystem
For decades, the war on glioblastoma has been fought against the wrong enemy — or at least an incomplete one. Researchers at McMaster University have now demonstrated that the tumor's true fortress includes not just cancer cells but the immune system's own macrophages, conscripted by the tumor to shield it from attack. By engineering CAR-T cells to recognize a protein called GPNMB present on both the cancer and its protectors, the team achieved complete tumor elimination in preclinical models — suggesting that treating cancer as an ecosystem, rather than a mass, may be the conceptual shift the field has long needed.
- Glioblastoma kills most patients within two years, and its resistance to treatment is partly engineered — the tumor hijacks the body's own immune cells to build a protective shield around itself.
- CAR-T therapy, a proven breakthrough in blood cancers, has repeatedly failed in solid brain tumors because targeting cancer cells alone leaves that immune fortress standing.
- The McMaster team discovered that a single protein, GPNMB, appears on both the tumor cells and the shielding macrophages — a molecular overlap that makes simultaneous attack possible.
- In preclinical models built from human patient tumors, the dual-targeting CAR-T cells eliminated tumors completely, with long-term disease-free survival and no recurrence.
- Parallel clinical trials in sarcoma patients at the University of Calgary suggest GPNMB is a viable target across cancer types, broadening the strategy's potential reach.
- Human trials in glioblastoma have not yet begun, and safety studies remain ahead — but the results, published in Nature, represent a rare and credible signal of hope for a disease that has resisted nearly everything.
Glioblastoma is among the most lethal cancers known — fast-growing, treatment-resistant, and fatal for most patients within two years. What makes it so formidable is not only the tumor itself, but the immune system it corrupts. The cancer recruits macrophages — normally the body's defenders — and turns them into accomplices, using them to suppress immune responses and create a protected environment in which the tumor can thrive. For decades, researchers focused on destroying cancer cells while largely ignoring this surrounding ecosystem.
The McMaster University team, led by surgeon Sheila Singh, found a way to strike both targets at once. They identified a protein called GPNMB expressed on both the cancer cells and the tumor-supporting macrophages. Engineering CAR-T cells — laboratory-modified immune cells designed to hunt specific targets — to recognize GPNMB allowed the therapy to dismantle both the tumor and its immune shield simultaneously.
The results in preclinical models grown from human patient tumors were striking: complete tumor elimination and long-term survival without recurrence. The findings were published in Nature on July 1, 2026. Complementary research from the University of Calgary, published simultaneously in Nature Cancer, showed GPNMB as a viable target in metastatic sarcoma patients already enrolled in clinical trials — suggesting the dual-targeting strategy may apply across multiple cancer types.
The work involved collaborators from King's College London, Northwestern University, the University of Toronto, and The Hospital for Sick Children, with support from the Terry Fox Research Institute and several cancer foundations. Human trials in glioblastoma patients have not yet begun, and additional safety and efficacy studies are required. But for a disease that has resisted nearly every therapeutic advance, the ability to dismantle both a tumor and the immune system protecting it offers something genuinely rare.
Glioblastoma is a relentless disease. The tumor grows fast, resists treatment, and kills most patients within two years of diagnosis. But what makes it so difficult to fight isn't just the cancer cells themselves—it's the immune system's own soldiers, turned against the body. Researchers at McMaster University have now identified a way to attack both at once.
The problem begins with macrophages, immune cells that normally patrol the body looking for infection and damage. In a healthy person, they're defenders. But glioblastoma has learned to recruit them, to turn them into accomplices. The tumor uses these hijacked macrophages to suppress the immune response, to shield itself from attack, and to create an environment where it can flourish. For decades, cancer researchers focused on killing the tumor cells directly. They largely ignored the immune ecosystem that was protecting the tumor from the inside.
The McMaster team, led by surgeon Sheila Singh, identified a protein called GPNMB that appears on both the cancer cells and on these tumor-supporting macrophages. That overlap was the key insight. If they could design a therapy to recognize GPNMB, they could strike at both targets simultaneously. They used CAR-T cells—immune cells engineered in the laboratory to hunt down specific targets—and showed that these modified cells could recognize GPNMB and destroy both the tumor and the macrophages shielding it.
In preclinical models grown from human patient tumors, the approach worked. Tumors disappeared completely. Mice that received the therapy survived long-term without the disease returning. The results were striking enough to publish in Nature, one of the world's most selective scientific journals, on July 1, 2026.
This builds on earlier work already moving toward human patients. Researchers at the University of Calgary have launched a clinical trial using CAR-T cells targeting GPNMB in patients with metastatic sarcoma, a cancer of connective tissue. Those results, published simultaneously in Nature Cancer, suggest that GPNMB is a viable target not just for brain tumors but across multiple cancer types. The dual strategy—treating the tumor not as an isolated mass but as an ecosystem—may be the missing piece that has kept CAR-T therapy from working well in solid tumors like glioblastoma.
CAR-T has been a genuine breakthrough for blood cancers, where it has saved lives. But translating that success to brain tumors has proven far harder. Most approaches have focused narrowly on killing cancer cells. The McMaster work suggests the reason: you can eliminate every cancer cell in the tumor, but if you leave the immune support system intact, the tumor can rebuild. By targeting both simultaneously, the researchers may have found a way around that limitation.
The work involved collaborators from King's College London, Northwestern University, the University of Calgary, the University of Toronto, and The Hospital for Sick Children. Funding came from the Terry Fox Research Institute, Brain Canada, the Cancer Research Society, and several brain cancer foundations. But the researchers are careful to note that much work remains. The therapy has not yet been tested in human patients with glioblastoma. Before that can happen, the team must conduct additional studies to ensure safety and efficacy. Still, for a disease that has resisted nearly every therapeutic approach, the ability to dismantle both the tumor and its immune protectors offers something that has been rare: genuine hope.
Citas Notables
Instead of treating the tumor as only a mass of cancer cells, we must treat glioblastoma as a connected tumor-immune ecosystem. Our approach attacks both the tumor and the environment that allows it to thrive.— Sheila Singh, senior author and professor of surgery at McMaster University
CAR-T therapy has been effective in some blood cancers but translating that success to brain tumors has been difficult. We may also need to dismantle the immune support system that helps the tumor survive.— Shan Grewal, co-lead author and MD/PhD candidate at McMaster
La Conversación del Hearth Otra perspectiva de la historia
Why has glioblastoma been so hard to treat compared to other cancers?
It's not just aggressive—it's politically smart. The tumor recruits the body's own immune cells and turns them into bodyguards. You can attack the cancer directly, but if those macrophages are still there, they rebuild the tumor and block new attacks.
So the macrophages are the real problem?
They're part of it. The tumor creates an entire ecosystem around itself. The macrophages suppress immune response, create inflammation that helps the tumor grow, and shield it from treatment. It's a partnership.
And GPNMB is on both the cancer cells and the macrophages?
Exactly. That's what makes this approach different. Instead of two separate therapies, one CAR-T cell can recognize GPNMB and attack both the tumor and the cells protecting it in a single strike.
The preclinical results sound almost too good—complete tumor elimination?
In the models, yes. But there's a reason we call them preclinical. These are lab conditions, often using mouse models or human tumor tissue grown in dishes. The human body is far more complex. That's why the Calgary team's sarcoma trial is so important—it's the first real test in actual patients.
What happens next?
More preclinical work to understand safety, then likely a clinical trial in glioblastoma patients. But that's years away. The science is promising, but we're still in the early stages of translation.