Stem-Cell Dopamine Neurons Cut Parkinson’s Off Scores by 23 Points — Grafts Still Lit Up After Immune Drugs Stopped

Subtitle: BlueRock / Memorial Sloan Kettering’s Nature Phase I exPDite trial of bemdaneprocel — off-the-shelf hES-derived dopaminergic progenitors grafted into the putamen — met its safety bar at 12 months, showed dose-tilted motor gains at 18 months, and still showed graft survival and clinical trends at 36 months after immunosuppression ended.

Parkinson’s disease is not a shortage of pills. It is a shortage of midbrain dopamine neurons whose axons once bathed the striatum in the right chemical at the right time. Levodopa can stand in for the missing transmitter — until the therapeutic window narrows, dyskinesias arrive, and “OFF” hours steal the day. Deep brain stimulation and focused ultrasound can rescue selected motor features. None of them replace the cells that died.

On April 16, 2025, Nature published the first peer-reviewed clinical chapter of a product built to do exactly that. Viviane Tabar, Lorenz Studer, Claire Henchcliffe, and collaborators at Memorial Sloan Kettering, Weill Cornell, UC Irvine, Toronto Western, and BlueRock Therapeutics (a Bayer subsidiary) reported an open-label Phase I trial of bemdaneprocel (formerly MSK-DA01 / BRT-DA01) — a cryopreserved, off-the-shelf midbrain dopaminergic neuron progenitor product derived from human embryonic stem cells and grafted bilaterally into the post-commissural putamen (DOI 10.1038/s41586-025-08845-y; ClinicalTrials.gov NCT04802733). An October 2025 company update stretched the same cohort to 36 months.

Twelve people, two doses, one year of immune cover

Twelve adults with Parkinson’s and motor fluctuations enrolled sequentially: low dose 0.9 million cells per putamen (n = 5), high dose 2.7 million cells per putamen (n = 7). Median age 67; 75% male; median time since diagnosis 9 years. Surgery under general anesthesia delivered every intended cell deposit. To protect the allogeneic graft, participants received a solid-organ-style immunosuppression cocktail for one year only: intraoperative/postoperative basiliximab, tapered prednisone, and tacrolimus aimed at a 4–7 ng/mL trough — then the drugs stopped.

That design is the quiet hinge of the story. Fetal midbrain grafts from earlier decades sometimes looked miraculous in open-label series and then stumbled in blinded trials — partly on tissue logistics, partly on graft-induced dyskinesias blamed on serotonergic contaminants. Bemdaneprocel’s pitch is industrial: GMP floor-plate–derived DA progenitors, release criteria that demand midbrain identity and no leftover pluripotent cells, and a manufacturing path that does not depend on aborted fetal tissue for every patient.

Safety first — and the ghost of graft-induced dyskinesia stayed quiet

The primary endpoint was safety and tolerability at 12 months. It cleared. No adverse events were attributed to the cell product. No deaths. No tumors, abnormal tissue overgrowth, or intracerebral hemorrhages on MRI. Two serious adverse events in the first year — overnight COVID hospitalization in the low-dose arm; a single perioperative seizure in the high-dose arm that added one hospital day — were not charged to the cells or to immunosuppression. Through 18 months, investigators tallied 78 treatment-emergent adverse events, nearly all mild or moderate; none related to bemdaneprocel.

The historical scare word — graft-induced dyskinesia — did not appear as a clinical signal. For anyone who remembers why fetal-cell enthusiasm cooled in the 2000s, that sentence is half the shareable story.

The numbers that travel

Secondary and exploratory motor readouts at 18 months (six months after immunosuppression ended) are where Sunday audiences lean in — with the honest caveat that this was open-label, uncontrolled, and not powered for efficacy:

MDS-UPDRS Part III OFF (motor exam off medication): mean improvement 23.0 points in the high-dose cohort (baseline ~49 → ~25) versus 8.6 points in the low-dose cohort.
Good ON time (Hauser / PD diary, without troublesome dyskinesia): +2.7 hours per day high-dose versus +0.2 hours low-dose.
OFF time: −2.7 hours high-dose versus −0.8 hours low-dose.
Antiparkinsonian medication use stayed roughly at baseline — so the diary and scale gains were not an artifact of simply piling on more levodopa.

Imaging backed the biology. Serial ¹⁸F-DOPA PET showed increased putaminal uptake bilaterally across cohorts — the radioligand signature of surviving dopaminergic terminals — including at the 18-month scan, half a year after immune drugs stopped. High dose was sized to approach the estimated ~300,000 healthy substantia nigra DA neurons; low dose targeted the historical ~100,000 grafted survivors associated with clinical benefit.

Three years later

On October 6–7, 2025, BlueRock/Bayer presented 36-month exPDite follow-up at the International Congress of Parkinson’s Disease and Movement Disorders. Safety remained consistent; no therapy- or procedure-related adverse events newly attributed to the product. F-DOPA imaging still suggested engrafted cells after the year-one immunosuppression washout. High-dose MDS-UPDRS Part III OFF improvement settled at a mean 17.9 points from baseline (n = 7); low-dose (n = 4 at this cut) 13.5 points — both framed as clinically meaningful shifts on published MCID yardsticks, still without a sham control. Good ON gains compressed toward roughly +1 hour in the high-dose diary subset. Participants continue in long-term evaluation; a pivotal Phase III (exPDite-2) with sham-surgery control is the next honesty filter.

What this is — and what it is not

It is not a cure. It is not permission to stop levodopa. It is not the hypoimmune islet story — those grafts were edited to hide from immunity; these progenitors used a finite immunosuppressant course and still lit up on PET afterward. It is not deep brain stimulation with a prettier press release.

What it is: peer-reviewed Phase I evidence that an off-the-shelf hES-derived DA progenitor product can be stereotactically delivered, tolerated, survive past immunosuppression withdrawal, avoid the classic graft-induced dyskinesia trap in this small cohort, and associate with dose-tilted motor improvements large enough that a blinded Phase III is worth running. Nature news coverage (April 2025) and Science News / Parkinson’s News Today explainers framed the same balance: big leap in manufacturing credibility, early clinical signal, efficacy still unproven.

The global Parkinson’s census is already past 10 million and climbing toward projections above 14 million by 2040. The unfinished bargain of modern PD care is that we replace a missing chemical with a schedule — and hope the schedule holds. Bemdaneprocel asks whether we can put living midbrain circuitry back into the putamen at industrial scale.

Twelve brains. Two doses. One year of tacrolimus and steroids, then none. A 23-point OFF-score swing at 18 months in the higher-dose arm. PET signal that did not vanish when the immune drugs did. Not a pharmacy shelf tomorrow. A credible, checkable step from fetal-era aspiration to a cryopreserved product that Phase III can actually falsify — which is how Sunday science should travel.

Sources

Tabar, V., Sarva, H., Lozano, A. M., et al. (2025). Phase I trial of hES cell-derived dopaminergic neurons for Parkinson’s disease. Nature, 641, 978–986. https://doi.org/10.1038/s41586-025-08845-y
Nature news. (2025, April). ‘Big leap’ for Parkinson’s treatment: symptoms improve in stem-cell trials. Nature. https://doi.org/10.1038/d41586-025-01208-7
BlueRock Therapeutics / Bayer. (2025, April 16). Publication in Nature of 18-month Phase I data for bemdaneprocel. Company / GlobeNewswire release.
BlueRock Therapeutics / Bayer. (2025, October 7). Positive 36-month results from Phase I exPDite trial of bemdaneprocel. Company press release / MDS Congress presentation.
ClinicalTrials.gov. NCT04802733 — Phase 1 study of bemdaneprocel (BRT-DA01) in subjects with Parkinson’s disease.