Diffuse midline glioma kills almost every child it touches, usually within a year, and for decades nothing has meaningfully slowed it. Now researchers at UC San Francisco have built immune cells that sleep through the trip around the body, wake up only inside the brain, and completely eliminated the tumors in every mouse they tested.
A Cancer With Almost No Good News
Let's start with what this disease is, because the number matters. Diffuse midline glioma, or DMG, grows into the brainstem, the part of your head that tells your lungs to breathe, your heart to beat, and your throat to swallow. It doesn't sit in a neat lump a surgeon can cut out. It spreads into the surrounding healthy tissue like ink into a paper towel.
Nearly every child diagnosed with it dies within a year. Existing treatments have had little success in stopping its progression. That's not a rough patch in the research. That's a wall, and families have been hitting it for a very long time.
So when a lab reports a result this clean, it's worth paying attention. Cleanly eliminating a tumor in animals is not the same as curing a kid. Hold onto that. We'll come back to it. But the result is still extraordinary.
The Cells That Wait for Their Cue
The therapy is called B-SYNC CAR-T, according to a release from UC San Francisco published on ScienceDaily. CAR-T treatments work by taking a patient's T cells, a type of immune cell, and modifying them with special receptors that recognize proteins on cancer cells and trigger an attack. They've worked against certain cancers. Brain tumors have stayed stubbornly out of reach.
The UCSF team's trick is making the attack machinery conditional. The engineered cells don't build their cancer-fighting receptors until they detect brevican, a protein found naturally throughout the brain. Cells floating in the blood or sitting in the spleen never switch on. Cells that reach the brain do.
That was "just as we designed it," said Senthilnath Lakshmanachetty, the postdoctoral researcher who led the experiments. Anyone who has watched a clinical trial go sideways knows how rarely that sentence gets to be true.
Why a Brain Protein Is the Whole Point
This builds on earlier work from Hideho Okada and Wendell Lim, who in 2021 reported laboratory findings on a predecessor called E-SYNC, aimed at glioblastoma. That version looked for EGFRvIII, a mutated protein found on some tumor cells. Smart idea, real limitation.
Only about 20% of patients have tumors carrying EGFRvIII. Worse, tumors can shed the protein as treatment kills off the cells that carry it, and the survivors simply stop showing up on the radar. It's the cancer equivalent of a burglar changing coats.
So the team switched the trigger. Instead of a tumor mutation that can vanish, B-SYNC keys off brevican, which belongs to the brain itself. Once activated, the cells hunt for tumor cells carrying EphA2 or IL13Rα2, two markers common to both glioblastoma and DMG. No dependence on the one mutation the cancer can ditch.
Every Mouse, Every Time, and Then Some
In the first round, researchers gave B-SYNC cells intravenously to mice implanted with human DMG cells. Some of the engineered cells made it into the brain, started producing their receptors, and went to work. They tested the approach in two mouse models, and the tumors were completely eliminated in both.
"By limiting B-SYNC's CAR expression to when and where it matters, in the brain, we saw consistent, total shutdown of DMG tumors in the lab," Okada said. Consistent and total are not words cancer researchers throw around casually.
There's a second result that might matter even more. Weeks later, the researchers introduced fresh cancer cells into the same mice, and the immune cells kept fighting, blocking new tumors from forming. The therapy didn't just clear the problem. It stuck around to guard against a repeat.
What Nobody Can Say Yet
Here's the part where we keep our feet on the ground. This was done in mice. The study was published October 28 in Science Translational Medicine, and the researchers themselves say they don't fully understand why the targeted approach works as well as it does. They suspect the brain-only activation makes the cells more effective, but the precise reason remains unclear.
Mouse models have a long history of making promising treatments look better than they turn out to be in people. Plenty of therapies have erased tumors in rodents and then fizzled in humans. The brain is a harder, messier environment, and children are not small mice.
Still, the team is moving. Scientists are preparing clinical trials at UCSF Health, potentially beginning as early as April 2027. For a disease with a survival timeline measured in months, that is the right instinct.
The Dingo Take
Why does it take a story about a dying child's cancer vanishing in a lab mouse to make us remember what science is for? Not the grift economy, not the culture war over who gets to say the word "research." This. A team of people spending years figuring out how to make a cell sleep until it reaches the one organ where it needs to fight.
Now look at the funding picture this kind of work depends on. Childhood cancers are rare, which means they're unprofitable, which means they live and die on public research money and patient foundations. Every time a politician talks about gutting the agencies that pay for this stuff, picture a lab like this one losing a postdoc. Then picture a trial that never opens.
We don't know yet if B-SYNC will work in a single human child. Nobody does, and the researchers are honest about that. But a nearly always fatal cancer just lost every round it fought in mice, and a trial could start in under six months. After years of news that makes you want to lie face down on the carpet, we'll take this one and say so out loud.
Comments