Stem Cell Therapy for Parkinson’s: Inside the Bemdaneprocel Phase 3 Trial

In 2001, a team of surgeons did something brave.

Open-label studies had shown that transplanting fetal dopamine tissue into the brains of people with Parkinson's looked promising. So they ran the test that matters. Forty patients. Half got the transplant. Half got sham surgery: small holes drilled in the skull, nothing implanted. Nobody knew who got what for a year.

The primary result was flat. On the trial's global rating, the transplant group scored 0.0 and the sham group scored −0.4. The difference was not significant. A second sham-controlled trial, published in 2003, also missed its main endpoint, and 56% of transplanted patients developed dyskinesias that persisted even off medication.

That history is the best lens for today's news.

A new kind of cell product, bemdaneprocel, made from human embryonic stem cells, is now in a Phase 3 trial with the same tough design: sham surgery, double-blind, randomized. Its Phase 1 data look encouraging. Four-year follow-up was presented at the International Congress of Parkinson's Disease and Movement Disorders in Seoul, October 4–8, 2026.

Encouraging is not proven. Here is the plain-English version of where things stand.

What Is Bemdaneprocel?

Bemdaneprocel is a cell product designed to replace the dopamine-making brain cells that Parkinson's destroys.

It starts as human embryonic stem cells. In the lab, those cells are guided step by step into dopamine neuron progenitors. Those are young cells on their way to becoming mature dopamine neurons. The batch is frozen, stored, and shipped as an "off-the-shelf" product. It was developed at Memorial Sloan Kettering Cancer Center (formerly called MSK-DA01) and is now run by BlueRock Therapeutics, a wholly owned subsidiary of Bayer.

A neurosurgeon places the cells into the posterior putamen on both sides of the brain. That is the area that loses its dopamine supply in Parkinson's. In Phase 1, it was done in a single session through one burr hole on each side.

The FDA granted bemdaneprocel Fast Track designation in 2021 and Regenerative Medicine Advanced Therapy (RMAT) designation in 2024. Those designations are meant to speed up development and review. They are not approvals. BlueRock states plainly that bemdaneprocel has not been approved by any health authority.

Why Try to Replace Dopamine Neurons at All?

Parkinson's disease slowly kills a specific group of neurons in the midbrain. Those neurons send dopamine to the putamen, which helps control movement. Lose enough of them and you get slowness, stiffness, and tremor.

Levodopa and related drugs replace the missing dopamine signal. They often work well for years. Then many people start to "wear off" between doses. Patients call these swings ON time (meds working) and OFF time (symptoms back). Higher doses can cause dyskinesia, which is extra, involuntary movement.

Deep brain stimulation and drug pumps can help. None of them replace the lost cells. That is the idea behind cell replacement: put new dopamine cells where the old ones used to send their signal. The old problem was the cells. Fetal tissue was hard to get, varied from batch to batch, and contained other cell types. Stem cell–derived products like bemdaneprocel try to fix that with a standardized, manufactured cell.

What Did the Phase 1 Trial Show?

The design

The first-in-human trial, called exPDite (NCT04802733), treated 12 people at sites in the United States and Canada. Results were published in Nature in April 2025 by Tabar and colleagues.

• Who: People with Parkinson's for a median of 9 years. Median age 67. 75% male.
• Doses: Low dose, 0.9 million cells per side (5 people). High dose, 2.7 million cells per side (7 people).
• Immune drugs: All participants took immunosuppressants for one year, starting at surgery. The regimen included basiliximab, steroids, and tacrolimus. The cells come from a donor line, so the goal is to keep the immune system from attacking the graft while it settles in.
• Main goal: Safety and tolerability at one year. Not effectiveness.

Safety in year one

The trial met its pre-set safety criteria. There were two serious adverse events in the first year. One person in the low-dose group was hospitalized overnight with COVID-19. One person in the high-dose group had a single seizure within 24 hours of surgery, which was attributed to the surgical procedure. No serious events were attributed to the cells or the immune drugs.

MRI showed no tumors, no abnormal tissue overgrowth, and no brain bleeds. There were no graft-induced dyskinesias. That last point matters because uncontrolled movements were a major problem in the old fetal tissue trials.

Motor scores at 18 months

The main motor measure was the MDS-UPDRS Part III score, tested in the OFF state, which means after holding Parkinson's medication. Lower is better.

• High-dose group: OFF score improved by an average of 23.0 points (SD 7.9). Good ON time rose by about 2.7 hours a day.
• Low-dose group: OFF score improved by an average of 8.6 points, with very wide variation (SD 29.2). Good ON time rose by about 0.2 hours.

PET scans using ¹⁸F-DOPA, a tracer that shows dopamine activity, showed increased uptake in the putamen at 18 months. The authors read that as a sign that grafted cells survived after immunosuppression stopped.

Three and four years later

All 12 participants moved into a long-term follow-up study (NCT05897957) that runs to five years after surgery. BlueRock-funded investigators have reported updates at medical meetings.

At three years, the high-dose group's OFF score was 17.9 points better than baseline on average, with about 2.1 more hours of good ON time a day (abstract).

The four-year poster from MDS 2026 reports:

• No deaths, no dropouts due to side effects, and no adverse events judged related to bemdaneprocel.
• Ten serious adverse events in total over four years (including hip and femur fractures, a seizure, a GI bleed, and heart rhythm and heart failure events). None was attributed to the cells, the delivery device, or the immune drugs.
• One low-dose participant withdrew from the study after the four-year visit.
• No graft-induced dyskinesia and no mass lesions or tumors seen on MRI.
• OFF motor score versus baseline: −15.6 points (SD 12.7) in the high-dose group and −6.5 points (SD 40.1) in the low-dose group.
• Good ON time versus baseline: about +1.0 hour a day in the high-dose group and roughly no change (−0.1 hour) in the low-dose group.
• Parkinson's medication use went up over four years, by an average of 331 mg levodopa-equivalent a day in the high-dose group and 74 mg in the low-dose group.
• On PET, dopamine signal in the posterior putamen (where cells were placed) stayed about stable through three years, while the caudate, which was not grafted, declined by about 25%.

Read that list slowly. The high-dose motor benefit versus baseline is still there at four years, but smaller than at 18 months. Medication use went up. And the untreated part of the brain kept losing dopamine signal, which is what you would expect from a disease that keeps progressing.

The poster itself says these data "should be interpreted carefully given the small study population and open-label design." We agree.

Why Does a Sham-Surgery Trial Matter So Much?

Go back to the 2001 trial. Freed and colleagues reported it in the New England Journal of Medicine. Open-label fetal transplants had looked good. Under blinded, sham-controlled conditions, the primary result did not separate from sham. Only a subgroup of younger patients showed a significant motor benefit on standardized testing.

Then Olanow and colleagues ran a 34-patient, 24-month double-blind trial, published in Annals of Neurology in 2003. No significant overall treatment effect (p=0.244). PET scans showed the grafts survived. Patients still did not do clearly better than sham as a group, and more than half of the transplanted patients developed off-medication dyskinesias.

So surviving cells and good open-label numbers did not guarantee a win in a blinded trial. That is the lesson.

In October 2026, Cell Transplantation published a small randomized, double-blind, placebo-controlled trial of intranasal human neural stem cells in 24 people with early-onset Parkinson's. It missed its primary endpoint (p=0.853). The authors pointed to some exploratory motor and PET signals and called for larger trials. Different product, same reminder: randomized trials are where hopeful ideas get sorted.

What Is the exPDite-2 Phase 3 Trial Testing?

exPDite-2 (NCT06944522) is the trial that could decide bemdaneprocel's future. Here is what the ClinicalTrials.gov record and BlueRock's announcements say, as checked on October 19, 2026.

• Design: Phase 3, multicenter, randomized, sham-surgery-controlled, double-blind.
• Size: About 102 participants, randomized 2:1 to bemdaneprocel or sham surgery.
• Dose: 3 million cells per side.
• Blinding: The protocol includes an immunosuppression regimen and placebo equivalents to keep the blinding intact.
• Who: Adults 45 to 75, diagnosed 4 to under 12 years earlier, with at least 2.5 hours of OFF time a day and a clear levodopa response. People with prior deep brain stimulation or prior cell or gene therapy are excluded.
• Primary endpoint: Change in daily ON time without troublesome dyskinesia, from patient diaries, at week 78.
• Secondary endpoints: OFF time, MDS-UPDRS Parts II and III, PDQ-39 quality of life, and side effects through month 60.
• Status: Recruiting. Start date June 17, 2025. BlueRock announced the first randomized treatment on September 22, 2025. The registry lists sites in the United States, Canada, and Australia, an estimated primary completion of March 2027, and study completion in March 2032.

One thing worth noticing. The Phase 3 headline measure is different from the Phase 1 headline. Phase 1 was a safety trial, and the news focused on the OFF motor score. Phase 3 is built around good ON time, which is how patients themselves experience the day. That is a sensible, patient-centered choice. It also means the Phase 1 "23 points" does not directly predict the Phase 3 result. And registry dates are estimates. They move.

BlueRock says that, depending on the outcome, results are intended to support future applications for marketing approval. If the trial misses, that plan changes.

How Does Bemdaneprocel Compare With Japan's iPS Product and With MSCs?

In March 2026, Japan granted conditional, time-limited approval to Amchepry, a Parkinson's product made from induced pluripotent stem (iPS) cells. We covered that in detail in Japan's first iPS cell therapy approvals. Bemdaneprocel is a cousin, not a twin. And both are very different from the mesenchymal stromal cells (MSCs) used in clinics.

The point of the table is simple. "Stem cells for Parkinson's" is not one thing. Data from a brain-implanted dopamine cell product do not transfer to an IV vial of cord-tissue cells, and the reverse is also true.

What Are the Risks and Open Questions?

Honest list. Some of these are known. Some are unknown until Phase 3 reads out.

• It is brain surgery. Burr holes, a cannula passed into deep brain tissue, general anesthesia. One Phase 1 participant had a seizure attributed to surgery.
• A year of immunosuppression. Tacrolimus and steroids raise infection risk and have their own side effects. This is a real trade-off, similar in spirit to the one we discussed for stem cell–derived islets in our zimislecel post.
• Long-term graft behavior. No tumors or mass lesions were seen on MRI in 12 people through four years. That is reassuring for a small group. It is not proof for thousands of future patients. Follow-up continues to five years.
• Durability. High-dose motor benefit versus baseline was smaller at four years than at 18 months, and medication use rose.
• The rest of the brain. Parkinson's affects more than the putamen. Cell grafts target motor symptoms. They are not designed to stop the disease elsewhere.
• Placebo. The single biggest unknown. Phase 3 exists to measure it.

For a general look at how side effects are reported in cell therapy, see our page on stem cell therapy side effects. For the difference between FDA approval, designations, and registration, see stem cell therapy and FDA approval.

Where The Stem Cell Club Fits

Straight answer: we do not offer bemdaneprocel. We do not offer embryonic stem cell products. We do not do brain cell transplants.

We offer umbilical cord Wharton's Jelly MSC options sourced from an FDA-registered U.S. lab. Registration covers facility and manufacturing standards. It is not FDA approval of a therapy. Our offerings are not FDA-approved to diagnose, treat, cure, or prevent any disease, including Parkinson's.

We will not borrow BlueRock's numbers to sell you anything. Different cells. Different delivery. Different regulatory road. Bemdaneprocel is on the same Biologics License Application path that the FDA requires for drugs like Ryoncil, the first FDA-approved MSC therapy.

If you or someone you love has Parkinson's and this news caught your eye, the right first call is a movement disorder neurologist. If you want to explore a clinical trial, start at ClinicalTrials.gov and bring the listing to that appointment.

Questions to Ask Before You Believe Any "Stem Cells for Parkinson's" Claim

• Which exact product? Bemdaneprocel is a specific, manufactured cell product in a registered trial. A clinic vial is not that.
• Was there a control group? Open-label improvement in Parkinson's is common. Ask for sham- or placebo-controlled data.
• How many people? Twelve is a start. It is not an answer.
• What are the trade-offs? Surgery, immunosuppression, side effects, cost.

We keep a longer checklist here: questions to ask any stem cell clinic.

Frequently Asked Questions