Randall Bateman of Washington University, St. Louis, Missouri, updated the audience on the latest DIAN data. As of 13 October, 106 people had completed their first visit. Nearly all complied with the clinical assessments, four out of five with imaging, and two-thirds with the requested lumbar puncture. Twenty people are within plus or minus three years from their expected age at onset; a majority is younger. The non-carriers show no cognitive deficit on the MMSE or the CDR sum of boxes assessments, whereas carriers who are five, even as far back as 10 years, younger than their expected age at onset do show subtle decrements. On structural MRI, DIAN replicates Fox’s and Rossor’s findings that the rate of atrophy accelerates before age at onset, although those curves separated only by a little. PIB-PET showed a wider separation in uptake in the precuneus and caudate brain areas.
Importantly, CSF measurements in these volunteers show that carriers overproduce Aβ42 relative to Aβ40. Mutation carriers produce more Aβ42 than do non-carriers. “We believe this reflects the pathogenic mechanism. Carriers produce more Aβ42. As the disease progresses, their CSF levels drop, but 20 years prior, carriers start out higher than non-carriers. We think this is of great interest,” Bateman said. Overall, the pathobiology emerging from this cohort is similar to that of late-onset or sporadic AD, and can be detected 15 years before symptoms. “This group appears well suited for proof-of-concept studies, and eventually secondary prevention studies,” Bateman concluded.
At the EMA, three European experts who are not involved in DIAN endorsed the premise of treating pre-symptomatic carriers of AD mutations. First, Bruno Dubois of Paris’s Hospital de la Salpêtrière, France, placed DIAN’s study population in the broader context of a shift in terminology that is spreading through the field as clinicians grapple with how to define pre-dementia stages for research and clinical trial purposes. Last July, a set of expert panels largely from the U.S. had proposed revised diagnostic guidelines (see ARF related news story). In October, a different (though partly overlapping) group of clinical leaders from Europe and the U.S. published their own attempt at clarifying exactly what to call the stages preceding overt AD (Dubois et al., 2010).
As per this new lexicon, people who have deficits in episodic memory tests such as word recall, and who also have abnormal biomarkers, would be told they have “prodromal AD.” This is a symptomatic stage that reaches back three to five years into the pre-dementia phase. People at earlier stages—that is, those long five to 15 years prior during which they have no clinical expression of the underlying disease—would fall into two groups. Asymptomatic people who have biomarker changes should be told they are “at risk for AD,” not that they have “preclinical AD.” Using the latter term would consign people to fear even while scientists do not yet know for certain how these biomarkers predict future clinical disease. “This is very important,” said Dubois. In contrast, people with autosomal-dominant mutations and biomarker changes do have “pre-symptomatic AD,” Dubois said. The population for which DIAN proposes therapeutic trials falls into this latter category. To date, 83 percent of carriers in DIAN have mutations in presenilin-1, 3 percent in presenilin-2, and the rest in APP.
Second, Caroline Graff from the Karolinska Institute in Stockholm, Sweden, emphasized the appeal of clinical trials for families with these dire genetics. At her center, researchers know about 50 at-risk relatives. Recent data on some of them confirm biomarker findings from other studies that CSF Aβ42 drops significantly in carriers at least nine years prior to their expected age of onset. Graff disputed a statement that sometimes crops up in these discussions, namely that autosomal-dominant AD is different from sporadic AD, or even that each mutation generates its own disease. In her experience working with AD patients, “with ADAD the signs and symptoms are the same as sporadic AD, except the onset is usually younger and it progresses faster,” Graff said. Some clinical variability occurs in all neurodegenerative diseases even within a family; it is not especially pronounced in ADAD, other AD clinician-researchers agreed.
Graff cited the example of a family that came to her center in 2005 with a suspicion of HD but turned out to have a presenilin mutation that brings on disease in a person’s thirties. “We now have contact with eight individuals of this family who are at 50 percent risk, seven of them in the age range for clinical trials,” Graff said. DIAN scientists are exploring how either the entire observational biomarker protocol, or a clinical trials portion of it, can be expanded to include participation from European families at Graff’s and other centers in Sweden and other countries, said Bateman.
Third, Harald Hampel of the University of Frankfurt, Germany, endorsed the initiative of biomarker-driven drug trials for mutation carriers. In a review of existing data on CSF biomarkers in FAD mutation carriers, he noted that it is highly concordant (e.g., Almkvist et al., 2003; Moonis et al., 2005; Scheuner et al., 1996). Like DIAN and Graff’s data now, these older papers showed an Aβ42 decrease in CSF a decade before symptoms and an increase in tau some five years prior to symptoms. Likewise, imaging research has produced a signature for prodromal AD, Hampel said (Reiman et al., 2010). Regional cerebral glucose metabolism by FDG-PET, fibrillar amyloid deposition by PET, and regional gray matter atrophy by MRI all yield strongly concordant results that together identify prodromal AD. Research by Fox and Martin Rossor at University College London has over the years shown that global and regional atrophy accelerates some five to three years prior to symptoms, with annual rates of change that apply consistently, Hampel noted (Ridha et al., 2006). As in sporadic AD, in ADAD metabolic abnormalities precede shrinkage by some years (Mosconi et al., 2006).
Hampel said that organized, comprehensive studies such as DIAN were the logical next step to follow on single-center studies, and he applauded DIAN for adopting ADNI protocols so data can be directly compared. DIAN’s preliminary biomarker and imaging data conform closely not only to prior, smaller ADAD studies but also to the preclinical staging graphs that grew out of ADNI. “That opens a considerable window. The biochemical and imaging data suggest abnormal APP metabolism, and axonal/synaptic/neuron losses are evident in a pre-symptomatic phase years before clinical onset. This is useful in clinical trials,” Hampel said.
Hampel showed a staging slide from a recent review that overlaps closely with those that many researchers are showing at conferences these days (Hampel et al., 2010; also see Perrin et al., 2009 and ARF Webinar with Cliff Jack). On a time axis, the commonly accepted staging scheme starts the disease process out with amyloid deposition/CSF Aβ42 lowering around 15 years prior to symptoms, followed by FDG-PET reduction, then by MRI atrophy, then CSF tau increase five years prior, cognitive performance some three years prior, and then a functional decline around the time of conventional diagnosis.
General Consensus: Check. Now for Specifics
So far, all agreed. How, then, can scientists actually conduct preclinical trials? Paul Aisen of the University of California, San Diego, and the Alzheimer’s Disease Cooperative Study, put an example design on the table for the regulators and pharma scientists to discuss.
To Aisen’s mind, the optimal time of intervention may be an early stage of amyloid deposition. A trial of an anti-amyloid drug could be designed for a broad population across ADAD families and enroll participants whose brains have brain amyloid. Such a trial would necessarily have to use a biomarker outcome, because people at this stage remain years away from clinical symptoms. This outcome could again be amyloid imaging or CSF, or it could be MRI. MRI makes sense to Aisen, because the normal subjects in the ADNI cohort show marked atrophy if they also have brain amyloid. The same is true for FDG-PET and even the crude MMSE test: The decline in these markers among the normal ADNI group is disproportionately due to those people who are amyloid positive, Aisen said.
But do these biomarkers respond to drugs? Little is known. Amyloid imaging did decrease with drug, and increase with placebo in one study (Rinne et al., 2010).
Specifically, Aisen proposed this design for regulatory discussion: a randomized, parallel group study with a select anti-amyloid treatment given for two years, followed by an open-label extension to ensure every person gets access to treatment. Biomarkers would constitute primary outcome measures, including MRI, FDG-PET, amyloid PET, CSF, cognition, and patient-reported outcomes. Clinical outcomes would follow post-study. “We can power such a study in ADAD with group sizes ranging from 25 to 250,” Aisen said. Dubois added to this that the results of French observational studies suggest follow-up clinical outcomes should choose tasks that flag highly AD-specific deficits, such as free recall and cued recall tests stressing hippocampal function. In Part 3 of this series, read how regulators from Europe and the U.S. chewed on this proposed trial design.—Gabrielle Strobel.
This is Part 2 of a four-part series. See also Part 1, Part 3, Part 4. View PDF of entire series.