Cancer as a Neural Circuit Disease

# From cancer neuroscience to new therapies *Evidence guide: Preclinical mechanisms, retrospective associations, and early clinical observations. Efficacy, safety, and nervous-system function require separate evaluation.* Levetiracetam is an established antiseizure medicine. A 2026 Monje Lab study found that, in brain slices, it reduced GABAergic currents onto DMG cells, but not onto healthy neurons, and surprisingly did not depend on SV2A, its usual target. The exact mechanism is unknown. In mice, levetiracetam reduced DMG proliferation and tumor burden and extended survival, but not in hemispheric high-grade glioma models or in DMG cells grown without neurons. Phenytoin and ethosuximide, antiseizure drugs that act differently, had no such effect. In patient records, children with DMG who happened to take levetiracetam had longer median survival, not seen in hemispheric high-grade glioma. But the main treated group was about fifteen children, not randomly assigned. The authors call for prospective trials: this is a repurposing hypothesis, not an established treatment or advice for any patient. A second strategy uses the immune system. T cells recognize targets, called antigens, through their receptors. A chimeric antigen receptor, or CAR, is synthetic: an antibody-like binder for a chosen surface molecule, joined to parts that activate the T cell. The target is GD2, a sugar-containing lipid on the cell surface. In 2018, the Monje and Mackall labs found that H3K27M DMG cells express high GD2. In five patient-derived mouse models, GD2 CAR T cells cleared most of the tumor, but some GD2-low cells remained, and inflammation around tumors was fatal in some mice. The phase one trial was published in Nature in 2025. Eleven patients with DIPG or spinal DMG received intravenous GD2 CAR T cells, and nine later received infusions into the brain. The goals were feasibility and safety: manufacturing worked for every patient, and the lower intravenous dose was set as the maximum tolerated. Side effects were serious. Every patient had tumor inflammation-associated neurotoxicity, swelling and symptoms at the tumor site, and three at the higher dose had severe cytokine release syndrome. In all patients the neurotoxicity was reversed with intensive care. Responses varied: four patients had major tumor shrinkage, fifty-two to one hundred percent, and three had smaller reductions. One had a complete response lasting over thirty months; others progressed. With eleven patients and no control group, the authors describe cautious optimism, not a cure. Effective cancer treatment can also harm the brain, a long-standing Monje Lab theme. In mice, the chemotherapy drug methotrexate caused persistent microglial activation, reactive astrocytes, and disrupted oligodendrocyte-lineage cells and myelin. Depleting microglia restored myelin and cognitive behavior. In 2025, the lab showed in mice that CAR T-cell therapy, even for cancers outside the brain, impaired cognition, with lasting microglial reactivity and disrupted oligodendrocytes. Brain tissue from patients treated with CAR T cells for brainstem tumors showed similar reactive states. In mice, depleting microglia or blocking the receptor CCR3 rescued cognition. So cancer therapy must treat the nervous system not merely as tissue around a tumor, but as an organ whose function matters. Sources: [barron2026](https://doi.org/10.1038/s41591-026-04646-6), [majzner2022](https://doi.org/10.1038/s41586-022-04489-4), [mount2018](https://doi.org/10.1038/s41591-018-0006-x), [monje2025](https://doi.org/10.1038/s41586-024-08171-9), [gibson2019](https://doi.org/10.1016/j.cell.2018.10.049), [geraghty2019](https://doi.org/10.1016/j.neuron.2019.04.032), [geraghty2025](https://doi.org/10.1016/j.cell.2025.03.041), [stanford-cns](https://med.stanford.edu/neurology/divisions/cancer-neuroscience.html), [mm2023](https://doi.org/10.1038/s41586-023-05968-y).