In people with Parkinson’s disease, neurons in the brain that produce dopamine are continuously lost. Without the neurotransmitter, which is crucial for movement, control of motor function begins to break down, resulting in symptoms such as tremors — one of the hallmarks of Parkinson’s.
Replacing dopamine is the goal of many treatments. Since the 1960s, that’s been done with a drug called levodopa, an amino acid that the brain converts to dopamine. But that conversion is carried out by the dopamine neurons that Parkinson’s eats away at. Lose too many, and the medication can no longer reliably control symptoms.
Some researchers have therefore long been eyeing the possibility of restoring those neurons. That entails growing them outside the brain and then injecting them through holes drilled into each side of the skull. The hope is that the cells will take root and begin to replenish the dwindling supply of dopamine.
Nature Outlook: Parkinson’s disease
Nearly four decades after such a transplant was first attempted, the work is beginning to pay off. Since early 2025, four phase I and II clinical trials have reported early results showing that the treatment is largely safe, with some people reporting improved motor skills and less ‘off’ time, during which their medication is ineffective. Off time can result from small daily changes, such as to diet, as well as the loss of dopamine neurons over time. Some of those treatments are moving into phase III trials for efficacy. More trials are in the works, raising cautious hopes that a new treatment for the disease is on the horizon. One of the first trials to test the cells in people with Parkinson’s was successful enough that the treatment is already available in Japan, albeit under strict regulatory oversight.
“This is such an exciting place to be,” says Claire Henchcliffe, a neurologist at the University of California, Irvine. “These are first-in-human trials. We’ve never had any products like these before.”
Despite the momentum, no one thinks that the treatment is ready for large-scale use. The trials so far have been small, with no control groups, and were designed to measure safety, not effectiveness. They’ve opened up a host of questions for researchers to investigate, including which type of cell to transplant, how many to deliver and where to inject them. There’s also work to be done to understand how the brain responds to its new members.
Transplanting foreign cells requires around one year of immunosuppression, and the brain surgery itself carries risks, including haemorrhage and infection. But the trials move cell transplants from a theory requiring proof to an early-stage treatment needing development, says molecular neurobiologist Kwang-Soo Kim at McLean Hospital in Belmont, Massachusetts, who completed the first transplant of neurons derived from the cells of a person with Parkinson’s in 20171. The question now, Kim says, is, “How can we optimize them?”
Progress, at last
To begin down the road to approval, researchers led by Jun Takahashi at Kyoto University, Japan, injected cells grown from donor tissue — known as allogenic cells — into the brains of seven people with Parkinson’s. The procedure seemed to be safe: none developed tumours, a concern of transferring immature cells into the body2.
Positron emission tomography (PET) scans showed that, two years after surgery and around 9 months after immunosuppression was stopped, on average, participants had more dopamine in the putamen — the part of the brain into which the neurons are injected — than they did before the procedure. This showed that the transplanted cells remained biologically active, even after immunosuppression was stopped. The movement symptoms of some of the participants also showed signs of improvement.
On the basis of the results, the Japanese government gave the team conditional approval for the cell transplantations. This regulatory stage gives people access to the therapy, and the costs are covered by the national health insurance system, while researchers closely observe and document their progress — similar to being in a phase III clinical trial in the United States. Full approval will take seven years, Takahashi says, during which his team will monitor safety and efficacy.
The concept behind stem-cell therapy for Parkinson’s was first tested in people in the 1980s, when researchers transplanted dopaminergic neurons taken from 8–10-week-old aborted fetuses into the brains of people with Parkinson’s. The participants showed improvements in their motor symptoms. After three years, one participant had improved enough to stop taking levodopa. Nearly two decades after treatment, two individuals still had the motor improvements seen in the years immediately after surgery and had stopped taking medication more than a decade earlier.
Dopamine neurons are lost in Parkinson’s.Credit: Janko Kajtez/Parmar lab
But this approach has a major flaw: fetal tissue is rare and difficult to obtain, and using cells derived from it in research is legally and ethically complex, says Malin Parmar, a neurobiologist at Lund University in Sweden. Parmar has worked on the TransEuro study, which used fetal tissue for the treatment of Parkinson’s3. “You can never upscale them into reliable therapies,” Parmar says.
But a 2006 discovery4 of how to revert any mature tissue into stem cells — dubbed induced pluripotent stem cells (iPSCs) — and then coax them into becoming any other cell was a game-changer. Suddenly it was possible to imagine creating vast amounts of cells for research and medicine, and it didn’t take long for researchers to start probing the best ways to turn iPSCs into dopaminergic neurons — work that is now making its way to people. “It has turned from, ‘This has been done before, it won’t work,” to suddenly, ‘Oh boy, I can’t keep up with all the clinical trials,” Henchcliffe says.
A rush of results
In 2023, the US Food and Drug Administration (FDA) cleared the way for Sumitomo Pharma in Osaka, Japan, which makes the cell product used by Takahashi, to conduct trials in the United States. Three people have already received transplants, says Joseph Ciacci, a neurosurgeon at the University of California, San Diego, who is leading the trial. His team aims to treat eight people.
Ciacci is hopeful that the brains of people he has treated can recover lost function after the transplants. The new neurons could be a buffer against treatments such as levodopa becoming less effective over time. “I feel very excited about it,” he says.
Sumitomo says it has also begun a larger trial that will enrol dozens of people, including a control group that will receive sham surgery. The company aims to compare the cells used in both trials and launch its therapy in the United States by 2032.
Three other trials have reported results so far, using a mix of cell types. All have a small sample size — fewer than 15 people — and none of them has a control group. The results should therefore be interpreted cautiously, researchers say. Still, all of the studies that have reported data met their safety endpoints and showed early signs of clinical benefit.
A trial involving 12 people, co-led by Henchcliffe, reported similar results to those of Takahashi’s team, including increased dopamine uptake at 18 months measured using PET and improvements in motor scores5. The trial was sponsored by stem-cell specialist BlueRock Therapeutics in Cambridge, Massachusetts, which is a subsidiary of multinational pharmaceutical company Bayer.
Blue Rock’s cell product, called bemdaneprocel, received the FDA’s Regenerative Medicine Advanced Therapy (RMAT) designation in 2024 and the Japanese health ministry’s Pioneering Regenerative Medical Product designation in December 2025. The team is currently enrolling participants in a phase III trial that will include a sham-surgery control group; it aims to transplant cells into 102 people in the United States, Canada and Australia and monitor their motor function for 18 months.
A third trial in eight individuals, run by the biotech company Aspen Neuroscience in San Diego, California, used dopaminergic neurons generated from the participants’ own cells, known as autologous cells. The cell recipients showed motor improvements, and several were able to reduce their dose of levodopa, according to results presented at the International Conference on Alzheimer’s and Parkinson’s Diseases in March.
Dopaminergic progenitor cells are injected into the brain of a person with Parkinson’s.Credit: Kyoto University Hospital
The company has now completed transplants in another seven people with its cell product, called sasineprocel, and received RMAT designation in July. It plans to begin a phase III trial of the treatment this year.
In the same month, Parmar and her colleagues reported the first results of a trial, called STEM-PD, in which they transplanted neurons derived from embryonic stem cells into eight patients. The participants showed “good early signs” of symptom improvement, Parmar says. Six of the seven people who completed the follow-up period opted to reduce their dose of levodopa as a result. Similar to the other trials, none grew tumours or had other serious reactions to the cells, but one participant died ten weeks after the transplant owing to an infection they developed while taking immunosuppressive drugs6.
Taken together, Parmar says, the trial results show that the transplanted cells survive and don’t generate tumours. But more people need to be followed up for longer periods before the treatment can truly be called effective, especially given that treatments such as levodopa already exist, she says. She recommends that researchers adopt an approach of skynda långsamt, a Swedish phrase meaning to rush slowly. “What we have is really hopeful and very promising,” Parmar says. “But the expectations are super high.”
Open questions
While trials seek evidence that the approach works, researchers are also trying to get a better understanding of how the brain responds to allogeneic and autologous transplants, with an eye to further boosting safety and efficacy.
Allogenic transplants require that people undergo immunosuppression, generally for one year, to prevent the brain from attacking or rejecting the cells. But some researchers are working on a way to do away with that need by genetically ‘cloaking’ allogenic cells — editing certain genes to hide the cells from the immune system7.
Last year, a team led by researchers at the Florey Institute in Melbourne, Australia, coaxed human stem cells into overexpressing genes that modulate the immune system. They then grew the cells into neurons and implanted them into a rat model of Parkinson’s. The rats that received the edited cells mounted less of an immune response than did those that received unedited neurons, and showed greater improvement in motor symptoms8.
Autologous cells don’t require such immunosuppression. But they do have their own drawbacks. Because they need to be generated for each person, individuals need to wait months before the cells can be implanted — a particular problem for people with Parkinson’s, whose neurons are continuously being destroyed. Using individualized cells for each transplant also introduces variation that makes measuring outcomes in clinical trials difficult, Parmar says. It’s more expensive than a treatment using allogenic cells, too, which researchers hope will eventually be available in large, off-the-shelf batches.
It’s even possible, says Parmar, that autologous cells from a person with Parkinson’s might be innately susceptible to developing the disease after transplantation, meaning that the treatment’s success could be shorter-lived. In 2020, Parmar and her team transplanted neurons derived from either an embryonic stem-cell line or a person with a gene mutation linked to Parkinson’s into a rat model of the disease. Both types of cell survived and rats receiving either type had improved symptoms. But in the rats that received the cells with the Parkinson’s gene, the transplanted grafts showed early signs of disease development9. Regardless of the source, turning immature brain cells into a viable, accessible treatment will require improvements to the processes for generating them en masse. Researchers are hoping that automation and standardization will bring down both the cost and the time needed to produce them. Some draw comparisons to full-genome sequencing: the first full human genome cost hundreds of millions of dollars to sequence, but today it can be done for about US$1,000.
Work is also needed to improve the survival of implanted cells. Although the trials — which are testing both high and low cell doses — show that some cells do take root in the brain and become productive dopamine-producing neurons, many more do not. In the BlueRock trial, for example, researchers delivered 900,000 cells to each side of the brain of people in the low-dose group, but they expected that only 150,000 cells per side would survive. In the high-dose group, participants received 2.7 million cells on each side of the brain in an attempt to have 300,000 survive5.
“After the cells are put into brain, you have no more control,” says Su-Chun Zhang, director of the Center for Neurologic Diseases at Sanford Burnham Prebys, a non-profit research institute in La Jolla, California. “They are on their own.”
Some researchers who are trying to improve cell survival think that the environment into which cells are transplanted might matter — just as the quality of a plot of land can affect a gardener’s outcomes. “We have spent many, many years to improve the seed,” Kim says. “But we didn’t pay attention to the soil.”
Like fertilizing soil, preparing the brain before the transplant could help. Some groups, for example, are exploring whether injecting specific growth factors along with the cells might coax them into becoming mature dopamine neurons. And in 2023, Kim and his team discovered a phenomenon they called ‘needle trauma’ in mice, in which the brain responds to placebo injections with an immune response that causes inflammation and kills brain cells. Transplanting immune cells called T cells along with the placebo injection suppressed the response10.
Kim is also concerned about the potential for cells to acquire mutations during the reprogramming and manufacturing process — and the potential long-term effects of such genetic changes, given that the cells might remain in a person’s brain for decades. “Stem-cell therapies are living products,” Kim says. Questions about such genomic integrity highlight the careful balance researchers must strike when pursuing a desperately sought after and potentially life-changing treatment but one that researchers still know so little about. That was obvious to Kim in 2020, when he was inundated with calls, e-mails and letters from people with Parkinson’s and their families after the publication of preliminary results for the first person to receive his transplant.
“We have to keep optimistic, and we have to give them hope,” Kim says. “But very informed hope.”
