Brain-sensing synNotch-CAR T cells enable precise targeting and clearance of diffuse midline glioma in multiple preclinical models
A brain-gated CAR-T platform cleared tumors in mouse models; researchers are preparing a clinical trial.
TL;DR
- UCSF researchers reported preclinical results for brain-sensing B-SYNC CAR-T cells against diffuse midline glioma.
- The cells cleared tumors in two mouse models after delivery into the brain ventricle.
- A different target combination caused severe toxicity in one mouse model, and B-SYNC’s human safety remains unproven.
Researchers published the brain-sensing CAR-T study on 2026-10-28 in Science Translational Medicine. The B-SYNC design uses the brain-associated protein brevican to trigger expression of a receptor targeting tumor antigens. [1] [2] [3]
With ventricular delivery, the EphA2 and IL-13Rα2-targeting version cleared tumors in all 19 mice across two human tumor models, according to independent coverage. UCSF said the team also tested intravenous delivery and found that the cells could reach the brain. [2] [3]
A different B7-H3 and IL-13Rα2-targeting version caused severe inflammation in a mouse model after ventricular delivery, according to Peptide & Protein. Its report said the EphA2 and IL-13Rα2 version did not produce that toxicity by either route in the reported experiments. [2]
UCSF said it is preparing clinical trials of B-SYNC that may begin as early as April 2027. The reported tumor results are preclinical, so whether this approach is safe or effective in people remains open. [3]
Why it matters
The platform separates a tissue-location signal from tumor recognition, giving cell-therapy developers a possible way to control where an engineered receptor is expressed. The differing toxicity between constructs also shows why target selection and delivery require separate evaluation.
Editor's note
The findings are from preclinical models, not a human efficacy result. Medical Xpress carries UCSF-provided material; Peptide & Protein is independent coverage of the same underlying research, not an independent replication.