For the growing number of cancer patients receiving CAR T-cell therapy, neurotoxicity remains one of the most dangerous and treatment-limiting complications — and current corticosteroid-based management is far from reliable. A mechanistic framework proposing a fundamentally different intervention target could reshape how clinicians approach this problem.
Published in Frontiers in Pharmacology, this theoretical-mechanistic paper argues that mitochondrial dysfunction acts as a feedforward amplifier of neuroinflammation during immune effector cell-associated neurotoxicity syndrome (ICANS) — the serious neurological complication that can emerge after CAR T-cell infusion. Rather than defaulting to corticosteroids, which carry significant risks at high doses including hyperglycemia, acute myopathy, and psychosis, the authors construct a structured drug-repurposing framework centered on agents with mitochondrial pharmacology, CNS bioavailability, and demonstrated oncological safety. The antirheumatic drug leflunomide emerges as a lead candidate, primarily through its inhibition of dihydroorotate dehydrogenase (DHODH), a mitochondrial enzyme involved in pyrimidine synthesis and immune cell activation. Evidence from adjacent neuroinflammatory conditions is marshaled to support translational plausibility.
This is a hypothesis-generating framework paper rather than a clinical trial, which is an important caveat. No patient outcome data is presented, and leflunomide's role in ICANS remains entirely preclinical and inferential at this stage. That said, the mechanistic rationale is substantive: mitochondrial stress is an established amplifier of inflammatory cascades in multiple CNS disease contexts, and DHODH inhibition has shown immunomodulatory effects in multiple sclerosis research. The steroid-sparing framing is clinically urgent — ICANS can be severe enough to halt therapy in otherwise responding patients. This paper is best classified as an incremental but conceptually meaningful contribution: it synthesizes existing mechanistic knowledge into an actionable repurposing logic, which is precisely the kind of translational scaffolding that tends to precede formal clinical investigation. Oncologists and neurologists managing CAR T patients may find the candidate selection framework useful even absent trial data.