The issue: Measuring disease progression

Parkinson’s is an unpredictable disease. It rarely unfolds in a straight line, and it’s difficult to know which symptoms someone will get, how severe they’ll be, or how quickly they’ll get worse. The disease also tends to progress relatively slowly, which makes it difficult to assess new treatments. “In order to detect whether a drug is able to modify the course of the disease,” explains Professor Ken Marek of the Institute for Neurodegenerative Disorders in New Haven, Connecticut, “one needs to evaluate individuals for a long time and with very large sample sizes.”

Identifying accurate biomarkers for Parkinson’s disease would enable researchers to track its progression and assess the effects of potential therapies without waiting for patients’ health to deteriorate—which is why the Michael J. Fox Foundation launched the Parkinson’s Progression Markers Initiative (PPMI) in 2010. PPMI aims to identify and validate biomarkers of Parkinson’s disease progression. This, in turn, can help spur the development of new and better treatments for the disease.

A new approach: Parkinson’s biobank

PPMI investigators aimed to recruit around 200 researchers from across the world to gather fresh longitudinal data on 1,400 recently diagnosed Parkinson’s patients and control volunteers—including motor and non-motor clinical data; dopamine scans, which measure the density of dopamine neurons in the brain; and data collected from blood, urine and cerebral spinal fluid (CSF). The goal was to build the world’s largest collection of Parkinson’s-related clinical and imaging data, biological specimens, and to make those data, images, and samples freely available to any qualified researcher, anywhere.

“The decision was that large-scale data sharing was optimal and would enhance the project for a couple of reasons,” says Professor Marek, who led the initiative. “The first is that this data is theoretically of value to anyone interested in developing new therapies for Parkinson’s disease, so we wanted to make it available. The second is that even though this project was large and involved many people who have expertise in Parkinson’s disease, there are many others in the world with expertise, and I think having as many eyes on the data as possible creates the situation which is most likely to end up in success.”

Barriers to data-sharing

Even though all participants agreed to rapid data-sharing in principle, researchers worried about how the principle would work in practice. “Some of the investigators were concerned that either they would not be credited for the data properly, or that others would take the data and run with it when they hadn’t made the effort to produce it,” says Marek. Others worried that researchers who used PPMI data might misrepresent or misinterpret it in their own work.

These are legitimate worries. For instance, Professor Marek says, misinterpretation is easy when people don’t fully understand how data was gathered and what the limitations of these methods are. On the other hand, “being involved in the project puts you at an advantage, because you’re more familiar with the data and can therefore process it more effectively. I think we convinced investigators that, on balance, the advantages of making these data widely available were much greater than any of these concerns.”

Solutions: Honesty and transparency

“The key issue, in my view, is to be very clear about the requirements for participation in the study and what the goals of the project are,” says Marek. “In the past people would have collected data on patients and kept it to themselves, whereas here we were asking them to put every patient they collected data on into the database. We were recruiting researchers to be part of this larger collaborative effort, because we believe that the sum is greater than the parts.”

Professor Marek believes that having the support of the Michael J. Fox Foundation, a high-profile organization, also gave the project a boost. “[The Foundation’s] backing creates a natural anchor for the community, but also attaches a certain status to being part of the project,” he said. “It encourages people to want to join.”

Impact: Imaging tools

To date, PPMI data has been downloaded more than 5 million times via an online system which can be accessed from anywhere in the world. The process is fairly straightforward. Researchers seeking to access the data must first register and sign a data-use agreement; then, the PPMI’s data and publications committee reviews the request and checks to be sure the researcher is affiliated with a scientific or educational institute. The committee approves most requests the day they are submitted.

Already, the impact of the project is being felt. For instance, in 2018, Professor Marek and his colleagues published five-year longitudinal data in Movement Disorders using dopamine transporter (DAT) imaging to track changes in patients’ clinical symptoms. “We now have evidence of a fairly robust reduction [in dopamine transporter binding] that we can measure over time as a marker of disease progression,” Marek says. Such data will be highly valuable for the design of future disease-modification trials.

The initiative also recently published another study in Movement Disorders suggesting that reduced levels of α-synuclein—a protein which misfolds and forms clumps within the brains of people with Parkinson’s disease—can be detected in cerebrospinal fluid before motor symptoms of Parkinson’s develop. However, α-synuclein levels do not generally correlate with progression of the disease, they found: higher α -synuclein levels do not necessarily mean a person’s disease will progress more slowly, or vice versa. “I can’t say that these insights wouldn’t have been possible without this approach, but it has provided the sample size necessary to detect these changes, and to evaluate these measures in relationship to all of the other measures that are being collected, meaning the data is much richer as a result,” says Marek.

The PPMI project is now being expanded, with the goal of recruiting some 4000 individuals – including people at high risk of Parkinson’s disease but who haven’t yet developed it. The investigators are also expanding the type of data they are collecting, to include measurements from wearable sensors, as well as from additional types of brain scans.

All this data could be overwhelming, but Marek says that an important lesson researchers have learned is the value of collaborating with data analysts from the outset of their projects to work out how best to extract meaning from it. “Unless the data analysts and the content experts are working very closely together from the start, you typically end up with a lot of effort without much success,” he says. Professor Marek and his colleagues hope all this data holds the keys to identifying biomarkers that will make it possible to track the progression of Parkinson’s before patients experience new or worsening symptoms. These tools will help bring predictability to a currently unpredictable situation.