Editorial illustration of the core trial question: local inhibitory interneuron cell therapy in mesial temporal lobe epilepsy moving toward a harder sham-controlled Phase 3 threshold. Credit: CellXperience generated editorial illustration.
The next test is finally visible
Neurona's latest rezanecel update has the kind of numbers that make people stop skimming.
In April 2026, the company reported updated AAN data from ongoing open-label Phase 1/2 trials of rezanecel, formerly NRTX-1001, in drug-resistant mesial temporal lobe epilepsy. The headline figures are large: 89% median disabling seizure reduction in the low-dose unilateral MTLE with MTS cohort during months 7-12, 100% median reduction in a small unilateral non-MTS interim group, and 64% median reduction in a small bilateral MTLE interim group. Neurona also reported no serious adverse events attributed to the cells or administration procedure, no substantial group cognitive decline, and maintained seizure reductions in participants followed beyond the 12-month primary endpoint.
Those data deserve attention. Final language has to wait.
The stronger reason to watch is that the public record no longer stops at encouraging open-label follow-up. The current ClinicalTrials.gov record for NCT05135091 lists a recruiting Phase 3 study with estimated enrollment of 88, randomized parallel assignment, triple masking for the Phase 3 portion, NRTX-1001 and sham-comparator arms, and no posted results. The AES Phase 3 EPIC design abstract describes the same kind of harder test: randomized, double-blind, sham-controlled, with a seizure-frequency endpoint after surgery.
That is where the claim belongs. Open-label seizure reductions can make a program worth the field's attention. A sham-controlled Phase 3 trial can tell whether the field should change its mind.
The signal is large enough to deserve pressure
It would be too easy to flatten the open-label data into a shrug. Small, uncontrolled, sponsor-reported, no registry-posted results: all true. Also incomplete.
Drug-resistant MTLE carries enough clinical weight for the signal to matter. Existing options can include resection, ablation, or neuromodulation, each with its own tradeoffs. A cell therapy that could durably reduce disabling seizures without removing epileptogenic tissue would be a serious clinical idea if the effect survives a controlled test. That conditional matters, but so does the idea.
The unilateral low-dose MTS cohort is the cleanest-looking piece of the current sponsor-reported update because it has reached the 7-12 month primary efficacy window in nine participants. The high-dose cohort did not simply dominate the low dose in the reported 2026 update, which is also useful. It makes the story less like a dose-escalation script and more like a real biologic and procedural question.
The non-MTS and bilateral MTLE signals are earlier. The bilateral registry record, NCT06422923, is still a small Phase 1/2, single-group, open-label study with estimated enrollment of 10, serious adverse events at 12 months as the primary endpoint, and no posted results. The April 2026 numbers from bilateral participants are interesting because they push the concept beyond the simplest unilateral MTS population. They are not the same quality of evidence as a mature controlled result.
The right response is pressure: let the signal face the design now being built around it. If the effect is real, the controlled trial is where it should become harder to explain away.
This is a brain-delivered cell product
Rezanecel asks a different question than an added antiseizure medication.
The product is described as an allogeneic stem-cell-derived inhibitory interneuron therapy. CIRM's program description frames NRTX-1001 as human embryonic stem cell-derived inhibitory neurons, delivered as a single intracranial injection into the seizure focus, intended to distribute locally, functionally integrate, and release GABA. The Phase 3 design abstract places that in clinical-operational terms: subjects in the treatment arm initiate immunosuppression before stereotactic, imaging-guided administration into the affected hippocampus, followed by immune monitoring, EEG, brain imaging, neuropsychological testing, and seizure diaries.
This matters because the evidence burden is different from a pill trial. A brain-delivered cell product has to justify procedure, immune management, durability, and cognitive safety alongside seizure reduction. It has to show that the benefit, if confirmed, is worth the complexity.
Complexity is not an argument against the program. It is part of what Phase 3 has to measure. Many serious therapies earn their place by asking medicine to accept complexity: deep brain stimulation, CAR-T, transplant medicine. But complexity changes what readers should watch. "No serious adverse events attributed to the cells or administration procedure" is encouraging sponsor-reported safety language. The more durable question is whether that safety profile, cognitive profile, and immune-management burden hold up when the Phase 3 design exposes the program to a larger, blinded, more operationally demanding setting.
UCB's bet matters, but it cannot answer the trial
UCB's April 2026 agreement to acquire Neurona for up to US $1.15 billion changes the commercial seriousness around rezanecel. UCB brings a different kind of attention than a generalist biotech investor. It is an established epilepsy company choosing to buy its way into regenerative neural cell therapy.
It is a serious company signal, with a clinical boundary around it.
UCB's own release keeps the necessary boundary visible: NRTX-1001 is investigational and has not been approved by the FDA or other health authorities. The same release describes FDA RMAT and EMA PRIME designations, while the official EMA annex lists NRTX-1001 as PRIME-eligible for adults with drug-resistant MTLE based on non-clinical plus clinical exploratory data. FDA's RMAT framework also makes clear why this kind of designation matters: it is meant for regenerative therapies for serious conditions where preliminary clinical evidence indicates potential to address unmet need.
The word potential is doing real work there. PRIME and RMAT can speed dialogue, development planning, and regulatory interaction. They help shape the route through development; the controlled trial still has to supply the result.
For investors, that distinction is useful rather than deflating. The acquisition says a specialized CNS company sees enough probability and strategic fit to make a serious move. The trial still has to decide whether that probability was priced around evidence or hope.
What would change the conversation
The next readout with the most weight is the controlled Phase 3 comparison.
A convincing result would need to show a separation from sham on disabling seizure frequency, preserve an acceptable safety and cognitive profile, and make the immunosuppression and procedure burden feel medically proportionate to the benefit. It would also need enough transparency to understand which patients are driving the result: unilateral MTS, non-MTS, bilateral disease, and eventually other focal epilepsy populations are not automatically the same evidence bucket.
This is why rezanecel belongs on a regenerative-neuro watchlist even though epilepsy is not a neurodegenerative disease. The program is testing questions that Parkinson's and other CNS cell-therapy programs also have to face: whether transplanted neural cells can produce durable functional benefit, whether brain delivery can be standardized, whether cognition and safety stay clean enough, and whether a controlled trial can be built around an invasive cellular intervention without losing interpretability.
The hopeful version of this story is narrower and better: the program has moved close enough to a real test that the answer may become interpretable.
That is a meaningful step. It is also a harder place to stand. Open-label promise can survive for years on selective attention. A randomized, sham-controlled trial is less forgiving. Good. If a brain cell therapy is going to change epilepsy care, that is the kind of trial it should be willing to meet.