One of the great challenges with developing treatments for neurodegenerative disorders stems from the capricious nature of disease progression, where symptoms remain hidden or unchanged for years, perhaps even decades, despite steady accumulation of cellular wreckage within the body. Looking for ways to reliably gauge this internal damage in Alzheimer’s, researchers from industry and academia launched an ambitious $64 million study in 2004. For three years, the Alzheimer’s Disease Neuroimaging Initiative (ADNI) followed more than 800 participants at nearly 60 sites across North America. Analysis of this massive longitudinal dataset has drawn consensus around a handful of brain imaging and fluid biomarkers that could potentially substitute for the slower-changing cognitive measures typically used in AD drug trials (see ARF related news story and ARF ADNI series).

Now, recruitment is underway for a similar effort to identify biomarkers for Parkinson disease progression. Spearheaded and sponsored by the Michael J. Fox Foundation for Parkinson’s Research, the Parkinson’s Progression Markers Initiative (PPMI) is looking to enroll 400 newly diagnosed PD patients and 200 healthy controls for its study tracking three- to five-year longitudinal change in various imaging and biologic measures. With a smaller cohort than ADNI, and fewer sites (18 in the U.S. and Europe), PPMI carries an estimated price tag of $40 million. So far, GE Healthcare and Pfizer have each contributed several million dollars toward the initiative, and the Foundation hopes additional partners will chip in to jointly cover at least half the anticipated cost. Enrollment began in July, with six-month data on the initial wave of participants expected by early 2011 (see study timeline).

Just this month, an article summarizing PPMI appears in a special PD-themed issue of Biomarkers in Medicine. Anchored by a freely available overview by Michael Schlossmacher of the University of Ottawa Health Research Institute and Brit Mollenhauer at the Paracelsus- Elena Clinic, Kassel, Germany, the issue also contains individual articles on most measures included in PPMI.

The idea for PPMI emerged over many years. Motivated in part by a string of failures in clinical trials of disease-modifying agents (e.g., Parkinson Study PRECEPT Investigators, 2007), the Fox Foundation rallied leading experts for a series of workshops and discussions on the need for a large-scale PD biomarker study. “The consensus became that without more objective measures of disease progression, it was going to be difficult for other pharma companies to put their toe in the water or have a chance of success,” said PPMI’s principal investigator Ken Marek, Institute of Neurodegenerative Disorders, New Haven, Connecticut, in an interview with ARF.

The other force that propelled PPMI forward was ADNI. Based on one-year data, this public-private consortium has by and large exceeded the expectations of its organizers, who are now conducting an extension study (ADNI-GO) while also gearing up for a second phase (ADNI 2) that will focus on AD biomarkers in the earliest stages of disease (see ARF related news story). ADNI “has really set the benchmark for how to do multisite coordinated studies where data are made publicly available,” Arthur Toga, University of California, Los Angeles, told ARF. “We had the advantage of learning from ADNI how to do things in an efficient way. We took those lessons and applied them in PPMI.” Toga heads ADNI’s informatics core and now does the same for PPMI.

Like ADNI, PPMI has a “hub and spoke” design, where raw data from multiple sites get channeled to a central database and, after passing quality control, become available for access by other investigators. Toga pictures this arrangement like a bike tire, with data coming toward the center along different spokes, and then going out again. “Some spokes might connect to a quality control group, others to the biorepository core, or to a coordinating center,” Toga said. “It’s a communication mechanism that allows everyone to know what’s going on.”

For example, magnetic resonance imaging (MRI) data from a PPMI participant scanned in Boston would get sent to the study database maintained by Toga and colleagues at UCLA’s Laboratory of Neuro Imaging (LONI). Once uploaded, the data are placed in quarantine, prompting automated e-mails that alert quality control (QC) teams to a new dataset in the system. By clicking a link within the e-mail, QC personnel can view the data and, if they meet specified criteria, flag it “acceptable.” That releases the data from quarantine and makes them available on the study’s website. Though ADNI did have a few cases where data on a patient scanned in the morning became publicly available that same afternoon, Toga said, more typically it will take a few days for data to be released from quarantine. PPMI quality control teams consist of outside parties with expertise in analyzing a particular type of data, for example, MRI, cognitive batteries, fluid assessments.

Though similar in overall structure, the informatics core plays a more central role in PPMI than it did in ADNI. Not all data from ADNI were stored in the UCLA system built by Toga and colleagues. Information on cognitive assessments, for example, was kept in a database maintained by the Alzheimer’s Disease Cooperative Study (ADCS) at UC San Diego. The ADCS did not send these raw data to the UCLA system, but only shared updates based on analysis of the data. Compared to ADNI, “we have our fingers in more pies,” Toga said of PPMI’s bioinformatics core. “We’re acting not only as the central database, but also the communication system.”

One extra layer of coordination in the PPMI database arises from the study’s biorepository core, which is headed by Alison Ansbach of Coriell Institute for Medical Research in Camden, New Jersey. Coriell is a nonprofit research institution that maintains one of the largest cell repositories for studies of aging-related diseases, including PD (see ARF iPS series). Unlike PPMI, ADNI had no bona fide biorepository division. The burden of receiving, storing and distributing biosamples fell largely to the directors of the study’s fluid biomarker core—John Trojanowski and Leslie Shaw—who took this on alongside analyzing the samples and heading up their own research teams at the University of Pennsylvania in Philadelphia.

In PPMI, Trojanowski and Shaw are leading the bioanalytics team, which will validate and standardize cerebrospinal fluid (CSF) assays. However, Coriell will take care of packaging, shipping, inventory, and other nitty-gritty details that added some stress and initial chaos to ADNI (see ARF ADNI series). In particular, Coriell will provide all study sites with supplies for drawing blood, spinal fluid, and collecting urine, as well as with materials for shipping the respective samples back to Coriell for storage and distribution out to groups wanting to analyze them. “Obviously, one of the jewels of this study is the biosamples. We have to treat them carefully,” said PPMI steering committee member Mark Frasier of the Fox Foundation. “It’s nice to have a seamless organization that, from start to finish, puts the kits together and then receives and houses the samples, all in a one-stop shop.”

Though Coriell will store and distribute the specimens, sample requests will go through the LONI website, which syncs with Coriell’s database to stay up to date with how much of each sample is available. Because the samples are nonrenewable, “we are looking to use them on biomarkers that already have some data,” Frasier said. “We do not anticipate using the samples for novel biomarker discovery.” For example, if researchers see a certain protein changing in their PD versus control samples, they could verify that discovery in a larger cohort using PPMI samples. The data should be solid but don’t need to be published, Frasier noted.

All proposals will go to a panel of five to 10 biomarker experts, including one steering committee member, for review. Why the range? “We want to prevent conflict of interest where someone reviewing requests would also be wanting to use the samples themselves,” said Frasier. Other considerations will be more practical—such as whether enough of the requested material is available, and whether the sample collection methods are compatible with the techniques and assays in the proposed research.

For starters, Frasier said, PPMI will focus on five potential fluid biomarkers—α-synuclein, DJ-1, Aβ, tau in CSF, and urate in blood. α-synuclein is the prime component of the Lewy body inclusions that gum up the brains of people with PD, and has been shown to be reduced in CSF of these patients (e.g., Mollenhauer et al., 2008; Tokuda et al., 2006). Mutations in DJ-1 (aka PARK7) cause early onset PD (Cookson, 2003), and CSF levels of this protein appear to be decreased in PD as well (Hong et al., 2010). Aβ42 and tau are heavy hitters among CSF biomarkers for Alzheimer disease, and have been found to be low and high, respectively, in CSF of PD with dementia (PDD) patients as well (Mollenhauer et al., 2006). In addition, a recent longitudinal study found that low CSF Aβ42 in PD patients predicted cognitive decline over the next two years (Siderowf et al., 2010 and ARF related news story). And other research suggests an association between high plasma concentrations of the antioxidant urate and decreased PD risk (Weisskopf et al., 2007).

A big problem in the PD field, as with AD, is that different single-center CSF studies on the same biomarker don’t necessarily agree with each other. Take α-synuclein, for example. Brit Mollenhauer, a member of PPMI’s steering committee and site leader at Paracelsus-Elena-Klinik, Kassel, Germany, has worked on measuring CSF α-synuclein in PD for nearly a decade. She teamed up with Michael Schlossmacher, then at Harvard Medical School, Boston, to build and test a new ELISA for measuring total α-synuclein in unconcentrated CSF and found, quite convincingly, lower CSF α-synuclein concentrations in PD patients than in healthy controls or people with AD (Mollenhauer et al., 2008). However, others have reported elevated CSF synuclein in people with synucleinopathies (e.g., Mukaetova-Ladinska et al., 2008), and some find that CSF levels of α-synuclein do not differ between those with and without synucleinopathies (e.g., Reesink et al., 2010). “The results are not consistent,” Mollenhauer said in an interview with ARF. “The problem is, [the studies] use different ELISAs, different antibodies, different standards. The variety of measurement is huge, and you cannot compare the results.” (See also ARF AD/PD story.)

Hence, one of the immediate priorities of PPMI’s bioanalytics core is to conduct a side-by-side comparison of different CSF α-synuclein assays in hopes of defining a standard method for subsequent PPMI analyses. For this evaluation, four labs will each perform their particular method on a common set of samples consisting of CSF from PD patients and age-matched controls, and standard α-synuclein reference material. “We hope to start the study by the end of this calendar year,” Shaw told ARF. Alzheimer disease fluid biomarker research has gone through similar trials and tribulations in recent years, and a much larger standardization and quality control initiative is well underway in that field (see ARF related news story).

The multicenter evaluation study within PPMI is an essential step toward the eventual goal of measuring CSF α-synuclein in a longitudinal cohort of newly diagnosed, drug-naïve patients. It is fairly common for PD patients with more advanced disease to take three or four different drugs, Mollenhauer noted. To rule out possible medication effects, PPMI is requiring PD patients to start the study drug-free. “We want to chart the natural course of disease without treatment,” Mollenhauer said.

The scientists anticipate that recruitment for PPMI will be difficult. Though most PD patients do not go on medication immediately, many will probably start drug treatment within six to 12 months of entering the study, said steering committee member Sohini Chowdhury of the Fox Foundation. “There is a window before their symptoms progress to a point where they require medication, but it is definitely challenging to find these people.”

In fact, PPMI chose its sites in part because of their prior success in that department. “They know how to recruit de novo patients,” Chowdhury said. Mollenhauer is doing double duty in this regard. While trying to find subjects for PPMI, she and colleagues have also been recruiting for a similar German study called DeNoPa (“de novo Parkinson’s”), which aims to improve early diagnosis of PD. DeNoPa participants are followed for 15 years, but come just once every two years for analyses including spinal tap, brain imaging by MRI and ultrasound, and observation in a sleep laboratory. Each visit lasts several days, during which the participant and an accompanying person (often a spouse enrolled as a control) are admitted to the clinic’s research hospital situated in a historic park. PPMI follow-ups are far less extensive, but more frequent. “We’ll probably just get the most motivated people for PPMI, since they get spinal taps more often,” Mollenhauer said. PPMI participants have a CSF sample taken at enrollment, then at six and 12 months, and once yearly thereafter. Whereas about two-thirds of ADNI participants consented to spinal taps—far exceeding initial expectations—baseline and longitudinal CSF sampling have been made a requirement of PPMI (see ARF Webinar).

Subjects enrolling in PPMI will also get genotyped for a set of PD-associated mutations, including LRRK2, GBA, synuclein, tau, and ApoE. Andrew Singleton of the National Institute on Aging in Bethesda, Maryland, leads the genetics core.

PPMI enrollment kicked off in July, with 10 of 18 sites recruiting at present. Fourteen people had consented for the study as of 13 October 2010. The Kassel clinic may not open for recruitment until early 2011, Mollenhauer said. “We are a bit behind because we first had to translate all the protocols. Now we have put them to the institutional review board, and they are discussing them. We still need to go through the radiation safety office in Germany.” None of the other European sites—Tuebingen (Germany), Innsbruck (Austria), and Naples (Italy)—have begun recruiting.

The fact that PPMI is a biomarker study, not a clinical trial, leaves investigators in somewhat of a recruitment quandary. “Patients are used to trials that test new therapies,” said Andrew Siderowf, steering committee member and co-investigator for the University of Pennsylvania site. “PPMI is different. The products of the study are validated biomarkers—tools for doing future clinical trials more rapidly.” Viewed in this light, PPMI may require more altruism from the patients, but on the upside, it has a much higher chance of success, he said. “When you go into a therapeutic trial, there’s an awfully good chance that whatever gets tested will be part of the process of learning, but not necessarily proven efficacious. A biomarker study, on the other hand, will produce a tool that can be used over and over again to test different sorts of therapeutics more rapidly and effectively in future trials.”—Esther Landhuis.

This is Part 1 of a two-part series. See also Part 2. View PDF of entire series.


  1. Access the official PPMI study website. Sign up on the homepage for weekly updates from the study blog, which includes news on scientific developments, publications, media and events, and recruitment.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Sorrento: ADNI Imagines the Future of AD Imaging
  2. ADNI: One-year Data Narrow Field of MRI, FDG-PET Approaches
  3. In Alzheimer Disease Research, iPS Cells Catch On Slowly
  4. Common Ground: Is Aβ the Foundation for Multiple Dementias?
  5. Still Early Days for α-synuclein Fluid Marker
  6. Worldwide Quality Control Set to Tame Biomarker Variation
  7. PPMI: Brain Imaging To Reveal Preclinical Parkinson’s Signature?

Webinar Citations

  1. Untapped Resource? New Study to Boost Acceptance of CSF Analysis

Paper Citations

  1. Mixed lineage kinase inhibitor CEP-1347 fails to delay disability in early Parkinson disease. Neurology. 2007 Oct 9;69(15):1480-90. PubMed.
  2. . Direct quantification of CSF alpha-synuclein by ELISA and first cross-sectional study in patients with neurodegeneration. Exp Neurol. 2008 Oct;213(2):315-25. PubMed.
  3. . Decreased alpha-synuclein in cerebrospinal fluid of aged individuals and subjects with Parkinson's disease. Biochem Biophys Res Commun. 2006 Oct 13;349(1):162-6. PubMed.
  4. . Pathways to Parkinsonism. Neuron. 2003 Jan 9;37(1):7-10. PubMed.
  5. . DJ-1 and alpha-synuclein in human cerebrospinal fluid as biomarkers of Parkinson's disease. Brain. 2010 Mar;133(Pt 3):713-26. PubMed.
  6. . Beta-amlyoid 1-42 and tau-protein in cerebrospinal fluid of patients with Parkinson's disease dementia. Dement Geriatr Cogn Disord. 2006;22(3):200-8. PubMed.
  7. . CSF amyloid {beta} 1-42 predicts cognitive decline in Parkinson disease. Neurology. 2010 Sep 21;75(12):1055-61. PubMed.
  8. . Plasma urate and risk of Parkinson's disease. Am J Epidemiol. 2007 Sep 1;166(5):561-7. PubMed.
  9. . Alpha- and gamma-synuclein proteins are present in cerebrospinal fluid and are increased in aged subjects with neurodegenerative and vascular changes. Dement Geriatr Cogn Disord. 2008;26(1):32-42. PubMed.
  10. . CSF α-synuclein does not discriminate dementia with Lewy bodies from Alzheimer's disease. J Alzheimers Dis. 2010;22(1):87-95. PubMed.

Other Citations

  1. ARF ADNI series

External Citations

  1. Alzheimer’s Disease Neuroimaging Initiative
  2. Parkinson’s Progression Markers Initiative
  3. fewer sites
  4. study timeline
  5. article summarizing PPMI
  6. special PD-themed issue
  7. overview
  8. study’s website
  9. Alzheimer’s Disease Cooperative Study
  10. LONI website
  11. DeNoPa

Further Reading