The first day of the Alzheimer’s Disease Research Summit 2015, hosted by the National Institutes of Health February 9-10 in Washington, D.C., focused on finding therapeutic targets for Alzheimer’s and testing them in clinical trials (see Part 1 of this series). On the second day, speakers switched gears and brainstormed creative ways to share and analyze big data sets, attract more research participants, and use developing technology to glean data from novel sources. These talks seemed to inspire a renewed sense of energy and excitement from audience members. “A number of the topics this year were quite different than we had in 2012,” said Neil Buckholtz, National Institute on Aging (NIA), Bethesda, Maryland. Buckholtz especially highlighted discussions on patient-oriented groups and technological developments. Steve Estus, University of Kentucky, Lexington, wrote to Alzforum that he considered these talks the most interesting at the summit. Laurie Ryan, NIA, also called attention to the potential for citizen science to expedite data analysis, and easier informed consent to draw in new participants.
Maximizing the Available Data
A prominent theme at the summit focused on means of sharing data. Several speakers noted that data sets could gain new life in the hands of more researchers, who could reanalyze the information for new knowledge about AD. A prime example comes from the Accelerating Medicines Partnership (AMP), which aims to find new biomarkers and therapeutic targets for AD. It will help pay for four research groups to generate data on molecular pathways involved in the disease and make it publicly available online prior to publication (see Feb 2014 news; Sep 2013 news). The four projects—one headed jointly by Philip De Jager, Brigham and Women’s Hospital in Boston, and David Bennett, Rush University Medical Center in Chicago, the others by Eric Schadt, Icahn School of Medicine at Mount Sinai, New York, Todd Golde, University of Florida, Gainesville, and Allan Levey, Emory University, Atlanta—are generating clinical, pathological, genomic, epigenomic, transcriptomic, and proteomic data on both human brain tissue and animal models of AD (for details, see the FNIH AMP webpage. So far, these researchers have been collaborating every other week to look across the raw data sets on a shared online platform and draw connections between them.
This month, the first of this network analysis data will be released publicly. It will be available upon request on the AMP-AD Knowledge Portal, accessible through the Synapse online registry hosted by Sage Bionetworks in Seattle. As more data are generated, periodic additions will follow this first release, said Stephen Friend of Sage Bionetworks, who collaborates with the AMP on this project. “The ultimate goal is to encourage as many users as possible to query [the data], develop new methods, and look for AD targets,” Bennett told Alzforum. He has already seen others use his data in ways that would not have occurred to him. For instance, one research group used his data to validate their integrated-systems approach to finding gene networks involved in late-onset AD (see Zhang et al., 2013). Bennett has since hired an author on that paper to develop the technique further and apply it more systematically to his own data.
Golde added, “If the public paid for these data sets to be generated, they should be shared to do the greatest good.” Golde said the investigators had felt some reservation about sharing the data before publication. They were especially concerned about junior investigators on the team receiving proper credit for generating the data. Other speakers at the summit, too, urged that ways be found for junior investigators in large, shared group projects to be recognized in ways that support their careers. Meanwhile, the AMP project has given Golde new opportunities to collaborate with people outside his specialty who offer new perspectives and suggest ways to analyze his data. Both Bennett and Golde said that to undertake these projects, they needed no preconceived notion of particular targets to seek, allowing them to take an unbiased, data-driven approach.
Continuing the theme of aggressively expanding data collection and sharing, Simon Lovestone, University of Oxford, described initiatives in Britain. He announced that Alzheimer’s Research UK is funding a new, £30 million ($46.3 million) Drug Discovery Alliance. It will launch three drug-discovery institutes at Oxford and Cambridge universities, as well as University College London. The aim is to hire up to 90 research scientists. They will build on ideas from academic researchers and identify and develop AD targets, by, for example, solving their structures, generating assays, and screening for early stage compounds. All this data will be made freely available.
Lovestone also highlighted the Dementias Platform UK. It aims to repurpose 22 current research cohorts, as well as their data and samples. This initiative brings together data on more than 2 million people aged 50 and older, he said. One cohort in particular, the UK Biobank, is being enhanced.Half a million volunteers have given genomic data and biological samples to the biobank, 100,000 will have whole-body MRIs, and 10,000 will have repeat scans. In a separate pilot study, a cohort of 24 people with preclinical AD will undergo amyloid and tau imaging once, followed by structural and functional MRI, magnetoencephalography, electroencephalography, optical tomography, assessment of gait and cognition, as well as blood and CSF collection, about every two months for half a year. After researchers determine how often patients are willing and able to undergo this kind of intensive, frequent testing, they will enroll for a larger trial of 300 people. All of that data will be made freely accessible, so that researchers can look for a signature of change in preclinical AD, Lovestone said.
While these initiatives will obtain and share network analysis and drug-discovery data, the Alzheimer’s Association is funding an effort to link results from clinical studies, said Maria Carrillo of the Association. Called the Global Alzheimer's Association Interactive Network, GAAIN is a big-data community. This platform so far includes clinical data about Alzheimer’s and other neurodegenerative diseases on more than 280,000 patients from 11 so-called data partners, Carrillo said. The partners include the Alzheimer’s Disease Neuroimaging Initiative (ADNI), the National Alzheimer’s Coordinating Center (NACC), and the French National Alzheimer Disease Database (see the GAAIN scoreboard). GAAIN is a federated platform where each data partner retains rights and access rules to their data, but registered researchers can query the combined data sets via interactive graphs. With a feature called the GAAIN Interrogator, researchers can define their own patient groups and variables of interest. GAAIN is meant to bridge silos of existing data sets and allow other researchers to mine them for new insights, said Carrillo.
The Alzheimer’s Drug Discovery Foundation (ADDF) has developed its own project to bridge the gap between basic and translational science. Called ADDF Access, the new, Web-based platform operates a bit like Match.com for drug development scientists, quipped Diana Shineman, director of scientific affairs at ADDF. It connects scientists in academia, biotech, or pharma who need particular services for translational or drug-discovery research, with providers of such services, primarily contract research organizations (CROs). In essence, service providers can register and describe their services at no cost, academic or pharma scientists can find those services, and when they contract with a service provider, ADDF charges a broker’s fee. The site also provides project management tools and links for other useful research services. Through ADDF Access, scientists can message providers directly, issue requests for proposals, compare bids, and even provide anonymous feedback about these communications. Ultimately, ADDF hopes this will speed the development of therapies for dementia.
Engaging More Patients
Meanwhile, several speakers showcased innovative ways by which other fields engage more people and patients in research projects. Sharon Terry of Genetic Alliance, Washington, D.C., launched the Platform for Engaging Everyone Responsibly. PEER originated in the orphan disease field and helps build cohorts for participant tracking. Participants upload health data and genomic information, answer research questions, and set up privacy layers. Sally Okun, of Patients Like Me in Cambridge, Massachusetts, talked about her company’s similar platform. Patients sign up for their condition to connect with others who have that disease and learn from them. They can track their symptoms and responses to various treatments, and share and discuss that data with others.
While the main purpose of both sites is to help connect patients, they also allow their members to participate in research. Most participants in Patients Like Me want to do that, said Okun. They generate a rich set of de-identified and aggregated longitudinal health data, which can be used to answer research questions, she said, noting that 50 publications have come out of Patients Like Me data (e.g., Tran et al., 2014). Okun recommended that Alzheimer’s researchers engage their patients through these types of platform to help develop active communities and harness their interest in research.
Another theme that emerged at the summit was the need for more diversity in clinical trials. “Racial, ethnic, and educational factors are not just nuisance variables to get out of way,” said Jennifer Manly, Columbia University Medical Center, New York. “They offer rich, important variability critical for understanding the biology of Alzheimer’s disease.” Minorities will make up the majority of the U.S. population in the coming decades, said Lisa Barnes of Rush University. They carry greater risk for Alzheimer’s, and their health will increasingly determine that of the nation, she said. Based on successful recruitment strategies at Rush, Barnes recommended that researchers cultivate relationships with communities before asking them to participate in research. Building networks with leaders, creating culturally tailored programs, and including minority staff in their studies all work to build trust. Stephanie Monroe of the African American Network Against Alzheimer’s added that institutions should employ a community-outreach representative and actively engage primary physicians, who are often the first people potential participants ask when they are interested in joining a study. Contradicting common notions about distrust of medical research among African-Americans, Monroe cited surveys indicating that 80 to 90 percent of African-Americans say they would indeed be willing to participate in studies if only they were asked.
To make it easier for more people to enroll in studies, John Wilbanks at Sage Bionetworks is tackling the problem of onerous consent forms. Informed consent is legally required, but the language used is often shrouded in legalese and can stop some people from signing up for a study. Wilbanks is working on digitizing the process into a mobile app. Using graphic icons to represent items on the consent form, and simple language to explain each point, he aims to break down a complicated set of information into simple concepts. A toolkit to adapt this consent process for other research studies is freely available.
Mining New and Improved Sources of Data
Besides broadening data sharing and attracting new interest in research participation, innovation is also needed in the type of behavioral data researchers collect in clinical trials, said Jeffrey Kaye, Oregon Health and Science University, Portland. The cognitive performance of research participants varies day to day, he said, so the sparse data points gathered in quarterly tests during clinic visits capture, at best, a coarse trajectory of change. By contrast, passive sensors built into the home or wearable technology can generate data on movement between rooms, computer use, gait speed, and sleep. This enables scientists to track behavioral factors continuously in a natural setting, Kaye said. By outfitting several hundred homes in Portland with such sensors, Kaye and colleagues have found that physical activity is much more variable in people with MCI than in people who are aging normally. They walk more slowly, wake less during the night, forget their medication more often, and gradually use their computers less. The latter variable correlates with atrophy in the temporal lobe, Kaye’s group has found. Using these types of measures to build models of behavior in clinical trials could track disease progress at high resolution, lower required sample sizes, and reduce the time it takes trials to read out, he proposed.
Barry Greenberg, Toronto Dementia Research Alliance, asked how Kaye would cross-validate those outcomes against standard clinical trial measures. Kaye responded that he would prefer to see proof-of-concept trials that test whether a treatment improves sleep, physical activity, or medication use. In response to an audience question about when this technology will be ready for routine home health care, Kaye said that it requires more work but may be useful for clinical trials in the near future.
Given all the data that will pour in from big cohorts and continuous monitoring, how will researchers analyze it? Pietro Michelucci, of the Human Computation Institute in Fairfax, Virginia, touted the power of citizen science to tackle the challenge. The general idea is that instead of a small team analyzing all the data, the task of analysis gets broken down into manageable parts and distributed among thousands of people, each of whom spends a small amount of time on them. Michelucci highlighted one project in which his institute is crowd-sourcing the analysis of brain microvessels in AD. This particular analysis is a time-consuming, manual task, and Michelucci estimates that his distributed strategy will reduce analysis time from 60 years to two. “Citizen-science projects give the general public the feeling of being part of something bigger, and of making a real difference,” Michelucci said.
The NIH also has a citizen-science working group. Its co-coordinator, Jennifer Couch of the National Cancer Institute, said that much of the data that will be coming from sensors described by Kaye will lend itself to this kind of analysis. By turning data analysis into a “game” and giving people the tools they need to answer a question, researchers can benefit from the problem-solving skills of the public. “People are creative and interested in helping,” Couch told the audience. “There are opportunities for data collection and also insights that wouldn’t be obtainable through conventional research approaches.” Couch cited examples in protein-folding and cancer, where research has benefitted from this approach. She noted that the group is still working out the ethical, legal, social implications of citizen science.
Scientific research journals can also play a role in improving the research landscape, said Diane Stephenson of the Critical Path Institute in Tucson, Arizona. She suggested that when a journal publishes a paper about a new biomarker, the NIA could fund independent replication of the results, and the same journal that ran the original report should commit to publishing the replication attempt as well, regardless of whether it is positive or negative. That will give the field and regulatory agencies confidence that the biomarker finding is robust enough for use in clinical trials, Stephenson said. Preclinical treatment studies sometimes are followed up in this way (see May 2013 news). Stephenson further recommended that funders require researchers to share their raw biomarker data, as well as encourage researchers to use data standards created by the Clinical Data Interchange Standards Consortium (CDISC) in ongoing and prospective clinical trials (see Nov 2011 news). Starting in 2016, the FDA will require new drug applications to conform to CDISC standards.
The Way Forward
Immediately following the summit, world health leaders from the G7 countries, Alzheimer’s advocates, and others held follow-up meetings in Washington, D.C. A writing group of AD experts, NIH staff, and representatives from other funding agencies and the National Alzheimer’s Project Act Council distilled recommendations from the summit meeting to refine the NAPA research milestones and assign a dollar value to research needed to fulfill the requirements of the Alzheimer’s Accountability Act (see Apr 2014 news). Separately, the health ministers of the G7 countries gathered for an update on their respective countries’ research efforts in AD, while the World Dementia Council met to review progress in 2014 and strategies to maintain their momentum in 2015.
On February 11, the Alzheimer’s Association, Weston Brain Institute in Canada, and Alzheimer’s Research UK announced their launch of an $1.25 million international research funding initiative called MEND, short for MEchanisms of cellular death in NeuroDegeneration. It will support new projects, especially research collaborations that examine the reasons behind brain cell death underlying multiple types of dementia.
Click here to view the second day of talks at the AD Research Summit 2015.—Gwyneth Dickey Zakaib
- New Initiative AMPs Up Alzheimer’s Research
- NIH Funds Prevention Trials and Translational Studies
- Bexarotene Revisited: Improves Mouse Memory But No Effect on Plaques
- New AD Data Standard: FDA Wants It; Will Trial Groups Use It?
- Alzheimer’s Accountability Act Puts Price Tag on NAPA Goals
- Zhang B, Gaiteri C, Bodea LG, Wang Z, McElwee J, Podtelezhnikov AA, Zhang C, Xie T, Tran L, Dobrin R, Fluder E, Clurman B, Melquist S, Narayanan M, Suver C, Shah H, Mahajan M, Gillis T, Mysore J, MacDonald ME, Lamb JR, Bennett DA, Molony C, Stone DJ, Gudnason V, Myers AJ, Schadt EE, Neumann H, Zhu J, Emilsson V. Integrated systems approach identifies genetic nodes and networks in late-onset Alzheimer's disease. Cell. 2013 Apr 25;153(3):707-20. PubMed.
- Tran VT, Harrington M, Montori VM, Barnes C, Wicks P, Ravaud P. Adaptation and validation of the Treatment Burden Questionnaire (TBQ) in English using an internet platform. BMC Med. 2014 Jul 2;12:109. PubMed.