26 February 2010. Last month, a one-day gathering of some 50 representatives of a range of international pharmaceutical companies, academic research leaders, funding representatives, and regulatory advisers ended with verbal support for the Alzheimer’s Prevention Initiative (API) proposed by Eric Reiman and Pierre Tariot, both at the Banner Alzheimer’s Institute in Phoenix, Arizona (see Part 1 of this series). What, then, is this initiative?

API is a proposal to get serious about Alzheimer disease prevention research. Arguing that the time is right, the API spells out a vision for launching a period of broad-based, parallel drug testing in pre-symptomatic AD throughout the field. Initially, the API begins with trials in two groups of people who, by virtue of their genes and age, face a high risk of developing AD symptoms in the next five years.

Reiman was not alone in thinking about such trials for many years. The idea had been around ever since AD genetics had its first successes. Some physician-researchers who have cared for and studied genetic families have long dreamed about trying to protect the adult children and younger siblings of their patients from this disease with drug trials, as indeed did scientists who run observational studies of cognitively normal people carrying two copies of the risk gene ApoE4. As Reiman recalls it, what jolted him into action was a particular encounter with a PS1 mutation carrier who had sought to know his genetic status years before from another scientist (see ARF Interview). A proactive, no-nonsense, can-do businessman, this man had been participating in multiple observational studies by Reiman’s and other research groups ever since finding out his genotype, all the while insisting that the scientists show him his data. One day in 2007, he walked into Reiman’s office, slammed onto his desk PIB-PET scans of his brain that Bill Klunk of the University of Pittsburgh had run, and said, simply “Look. Here’s my disease. Enough with just observation. When will someone try to find a prevention therapy for it?” “Speak of the urgent need!” Reiman told this reporter. He pledged to find a way to bring the makers of the most promising disease-modifying treatments to the table and gather the evidence needed to conduct prevention trials as soon as possible in people at the highest imminent risk of symptomatic AD. “I think of our conversation, this gentleman, and others like him every single day,” Reiman said.

Reiman then drafted a proposal, and he and Tariot spoke privately with colleagues in science and related stakeholder communities such as the Alzheimer’s Study Group to field advice, invite vetting of ideas, and build support. Converting the long-held ideal of prevention trials in high-risk folks into an action plan required an intense internal effort by a team of people at Banner. “The process was lengthy and laborious even before anything was made public,” Tariot recalls. Along the way, the scientists fine-tuned the API to combine a general call to arms with two specific drug trials they hope to get going as early as next year. Now, Reiman, Tariot, and Jessica Langbaum of Banner are presenting the API publicly with an article in this month’s issue of Biomarkers in Medicine. Here is a summary.

At present, the scientists argue, the greatest roadblock in AD research is not the discovery of new candidate treatments, but the absence of effective means to evaluate those treatments rigorously and rapidly. Conventional prevention trial paradigms don’t work for that. They require too many cognitively normal people (thousands in a trial), too many years (longer than a treatment’s patent protection), and too many dollars for the whole field to evaluate more than one or two treatments at a time in this way. In fact, recent history of AD prevention research has featured but three large prevention trials, the ADAPT trial of two NSAIDs (ADAPT Research Group, 2008), the Gingko Evaluation of Memory Study (DeKosky et al., 2008), and the Women’s Health Initiative Memory study of hormone replacement therapy (Shumaker et al., 2003; Shumaker et al., 2004). What’s needed are nimbler trials so more treatments can be evaluated more quickly.

This need invokes imaging and biochemical markers, but these are caught in a strange Catch-22, the authors argue. Observational studies such as the Alzheimer's Disease Neuroimaging Initiative (ADNI) and others are providing a strong scientific rationale for biomarker use in pre-symptomatic trials, but that is not enough. Regulatory agencies at present are reluctant to approve pre-symptomatic treatments based on biomarker endpoints alone; rather, they expect sponsors to show that drug-induced changes in such markers reasonably predict that a person will benefit clinically, i.e., that the biomarker change truly helps. This evidence can come from separate randomized clinical trials, and it can be pieced together from across different trials of different treatments. Such biomarkers-in-trials type of evidence is needed to establish which markers budge in response to which treatment, whether they change in the expected or in the opposite direction, and which biomarker changes portend a meaningful benefit. Alas, not nearly enough of this evidence is being generated these days because many drug sponsors hesitate to shoulder the added cost of assessing AD imaging and biomarkers in clinical trials of a given drug. Conventionally, such clinical trials programs focus narrowly on advancing the drug at hand, not on gathering broadly relevant surrogate marker data along the way. That none of the biomarkers in question are by themselves FDA-approved (or “qualified,” in FDA lingo) adds to the state of limbo. The upshot of it all is that drug sponsors are passing on costly pre-symptomatic trials while they see neither a clear path to approval nor a financial incentive. “This dilemma may at first seem...[insurmountable], leading to a sense of nihilism,” the scientists write.

But inaction is not an option, they continue. Here are some reasons for why biomarker-driven pre-symptomatic trials should break the deadlock now.

• Recent projections predict 65 million AD cases worldwide by 2030, even more in the early pre-diagnosis stages (see ARF related news story).
• Promising interventions spanning lifestyle, dietary factors, and approved medications exist; a good way to test them does not.
• Investigational treatments (anti-amyloid, anti-tau, others) are gathering safety and efficacy data in randomized clinical trials—enough, perhaps, to test them in prevention.
• Some of these investigational treatments may perform better in the disease’s pre-symptomatic stages, especially those whose hypothesis is tied to the earliest-moving biomarkers.
• People at the highest risk for AD offer a stepping stone into broader prevention studies. Their risk could be genetic or evidenced by abnormal biomarker profiles.
• Imaging and CSF biomarkers to detect and track the brain changes of pre-symptomatic AD are in hand.

Regarding this last point, ADNI results and attendant power calculations support future trials. ADNI is not the only source of such data. Reiman and colleagues point out a 14-year-old, ongoing study in Phoenix that does essentially the same thing, sans CSF biomarkers, for cognitively normal, middle-aged people who carry either zero, one, or two copies of the AD risk allele ApoE4. Banner scientists and their colleagues at the Mayo Clinic Arizona have measured declining synaptic activity, shrinkage of the brain, and amyloid deposition in people with copies of ApoE4, whose lifetime risk of developing AD is 91 percent. This observational study and others like it have generated data that support the design of trials in which these imaging tools would evaluate an investigational treatment. In some cases, this could be done directly against a particular person’s documented history of biomarker change. By now, a range of studies have documented abnormal changes in brain activity, volume, or CSF levels of Aβ/tau in ApoE4 carriers, PS or APP mutation carriers, or even in people without a defined genetic burden. Together, this body of evidence makes a strong case for using those same markers to evaluate treatments in people whose markers are changing, according to the Banner authors.

Here’s how Reiman and colleagues intend to break the deadlock. In general, they define pre-symptomatic AD treatment as one that slows or stops an existing disease process and delays or prevents its symptoms. This shifts the balance of competing risks—those of impending clinical onset versus harm from an experimental medication with limited safety data—toward treatment in the eyes of trial participants, physicians, and regulatory agencies. The API proposal does not focus on primary prevention in the general population.

Specifically, the authors propose a new scientific paradigm and some public policy reforms. The science piece includes a series of preclinical treatment trials in defined groups of people (see Part 3 of this series), a call on all drug sponsors to embed into their ongoing and new drug trials measurement of the imaging and biomarkers that are performing well in ADNI and other longitudinal studies, and the building of national registries of cognitively normal people who express interest in participating in pre-symptomatic AD trials. The API’s policy piece includes some ideas previously formulated in the report of the Alzheimer’s Study Group. For example, resetting the clock for patent protection to the time of completion of lengthy prevention trials might galvanize industry interest in such trials. Extending market exclusivity has proven successful for orphan drugs, the authors note, and could be tried as a financial incentive for AD prevention, too. Other ideas include federal funding for prevention trials of off-patent drugs and for inclusion of biomarker assessments into proprietary drug trials in return for full publication of the biomarker results.—Gabrielle Strobel.

This is Part 2 of a five-part series. See also Part 1, Part 3, Part 4, Part 5. See also a PDF of the entire series.