In the second half of 2011, scientists driving the Alzheimer's Prevention Initiative have been reporting at scientific conferences the first emerging biomarker findings from their human volunteers. These data provide tantalizing glimpses of what happens in the brains of young people carrying a deterministic Alzheimer's disease mutation when they are still in their twenties and thirties. While these imaging and fluid data at present represent but small snapshots of the disease 25 years before dementia, they nonetheless suggest that a quiet drama unfolds in the Alzheimer's-bound brain years before amyloid. “At present, it looks as if functional and structural changes may occur prior to fibrillar amyloid deposition,” Adam Fleisher of the Banner Alzheimer’s Institute said in a talk at the Clinical Trials in Alzheimer’s Disease (CTAD) conference held 3-5 November 2011 in San Diego, California. If further data substantiate those initial findings, and if the findings generalize to late-onset Alzheimer’s, they would then call for a refinement of the proposed biomarker staging diagrams that have captured the imagination of Alzheimer’s disease researchers worldwide.

Fleisher belongs to a large collaborative team of scientists who have been developing the API as a program meant to pioneer secondary prevention trials in people who are at high risk of developing Alzheimer’s disease. Led jointly by Eric Reiman and Pierre Tariot at the Banner Alzheimer’s Institute in Phoenix, Arizona, and Francisco Lopera at the Universidad de Antioquia, Medellin, Colombia, the API has been doing the groundwork preparing for such trials in people who carry autosomal-dominant mutations that will give them the disease with near certainty. (The API also prepares for trials in aging people who carry the ApoE4 risk allele.) “There are many people who are at very high risk of AD who are clamoring for therapeutic trials,” Tariot said.

The Initiative’s autosomal-dominant half is complementary to the Dominantly Inherited Alzheimer Network (DIAN, ARF related story), and its late-onset half is complementary to the A4 initiative. Together, the three programs share the goal of breaking ground on secondary prevention drug trials across the AD spectrum. That is, they range from rare, deterministic AD genetics on one end to risk genetics in the middle, and to the most common forms of late-onset AD on the other end. Success in any and all of these trials could energize earlier-stage trials throughout the field, the scientists believe. However, each program is also unique in some aspects. DIAN has fewer patients than the API, but subsumes all APP and presenilin mutations; A4 is potentially the largest study, but further behind in terms of funding and driven by biomarkers, not genetics. Along the way of gathering observational data and planning their respective programs, the leaders of all three meet frequently to work out where they can coordinate to enhance each other’s goals and ensure that their respective datasets can be analyzed together.

So what’s new with API since its last update on Alzforum (see ARF API series)? In 2011, the researchers have enrolled some 1,300 relatives of the Colombian families afflicted with the E280A Paisa mutation in presenilin 1 into the observational biomarker and cognitive study phase meant to precede treatment trials. About a third are carriers. The scientists hope to bring the number of participants to 3,000 and the number of carriers close to a thousand by 2013.

That goal—as indeed all key goals of API, DIAN, and A4—hinges on new funding coming forward. In the case of API, its leaders are currently awaiting final review by the National Institute on Aging of a pending grant proposal for the first treatment trial with an identified (but undisclosed) experimental drug while simultaneously stitching together a funding coalition of company money and private philanthropy.

In the meantime, the scientists have expanded their original biomarker studies with the Colombian participants that started in 2010. In 2011, the scientists, led by Fleisher and Yakeel Quiroz, currently at Boston University, added new cohorts of cognitively older adults in age brackets from age 35 and up, all the way back to children aged eight to 17. The children are not undergoing spinal taps, but they are donating a blood sample and, importantly, lying still in the scanner for various modalities of magnetic resonance imaging.

Why children? The scientists want to chronicle the entire natural history of this form of AD from its beginning, meaning they will trace back at what age biomarker measurements begin to diverge between carriers and their non-carrying siblings. In the next-older age bracket—the 18- to 26-year-olds—mutation carriers already show distinct differences in brain function and even structure. Hence, Quiroz and colleagues reached back with the less invasive tests into even younger ages.

To date, MRI has been taken from some 200 volunteers age eight and up. This happens on a Siemens 1.5T scanner at the Hospital Pablo Tobón Uribe in Colombia. “MRI capability is very good there for API studies,” Tariot told the audience at CTAD. Plasma has been taken from some 130 volunteers age eight and up, CSF from some 90 people age 18 and older. Fluids are being drawn in Medellin following standard acquisition, preparation, storage, and shipping directions developed for DIAN. They are analyzed in the lab of Anne Fagan at Washington University, St. Louis, Missouri, to ensure that data are comparable with CSF measures in the DIAN and, indeed, the Alzheimer’s Disease Neuroimaging Study (ADNI). PET imaging with florbetapir started up in September, when the first of what will be 50 participants flew to Bogotá, and from there to Miami and then Phoenix for FDG metabolic and amyloid imaging with florbetapir (see NYT coverage). These people will travel to Phoenix in small groups to get PET studies going until a cyclotron that is currently under construction near Medellin can start providing labeled ligand for a local PET scanner that began operating in October 2011. “This travel is logistically challenging, and the team in Medellin is absolutely amazing in coordinating it,” Fleisher said.

All the above measures are also being taken in a much smaller group of relatives already affected with mild cognitive impairment or AD. The goal is to take sufficient biomarker measurements to pinpoint the earliest divergence between carrier and non-carrier for each of them, trace them forward into symptomatic AD, and integrate this information into a staged natural history of this form of Alzheimer’s. This information can then serve as a foundation for treatment trials, first in this population, but also, together with similar data from DIAN longitudinal biomarker studies of ApoE4 cohorts and ADNI and AIBL cohorts, for prevention trials in late-onset AD (LOAD). “Ultimately, we want to use treatment trials in early onset AD as models for late-onset AD,” Fleisher said.

What are the results so far? The data for the children and adolescents are not available yet. But as shown at conferences, data for people in their twenties are trickling in, and they show functional and even subtle structural brain changes that appear to precede amyloid deposition. Specifically, carriers had abnormalities compared to their non-carrying siblings and cousins in their brain activation patterns when they performed an established fMRI task asking them to associate and subsequently remember face-name pairs (Sperling et al., 2001). Carriers performed the task as well as non-carriers, but in doing so, they activated their hippocampi more strongly and deactivated their precuneus brain area less strongly. This is essentially the same pattern of change as previously reported for the later preclinical stages of other forms of early onset AD and, indeed, late-onset AD. Quiroz and colleagues presented these data at the Alzheimer’s Association International Conference (AAIC) in Paris in July 2011.

Also at this conference, Fleisher and colleagues presented a poster suggesting that this same group of twenty-somethings already have subtle morphological changes—meaning atrophy—in their brains. In a whole-brain comparison of gray matter volume between carriers and non-carriers, the 20 carriers had less gray matter in their temporoparietal and parahippocampal brain areas than 24 non-carriers who were otherwise matched in age, sex, education, and cognitive test scores. It’s well established that atrophy accelerates three to five years before dementia onset (e.g., Ridha et al., 2006). In this earlier work, the new signature may not have come up because the group was smaller, the imaging was not generally done in people this young, and what was done used more global measures of how the boundaries of regions of interest shift. The new API research uses voxel-by-voxel comparisons independent of regions of interest in people twenty years younger than their expected age at onset.

MRI offers a growing number of increasingly sensitive measures for AD research, and the API team put one more to the test. In a cohort of 18 mutation carriers and 22 controls in their thirties to early forties, Quiroz and colleagues worked with Brad Dickerson at Massachusetts General Hospital, Boston, to look for the cortical thinning signature Dickerson had developed in four previous studies in mild LOAD, MCI conversion to AD, and cognitively normal people who have amyloid and are being followed longitudinally. Dickerson had pinpointed nine regions of interest per hemisphere and found that atrophy, as measured by a thinner cortex in those regions, predicted that a cognitively normal person would develop dementia some eight years prior (Dickerson et al., 2011).

This is the first study of cortical thinning in the API population. In this cohort, mutation carriers on average had a 4.75 percent thinner cortex in these regions, Quiroz reported at AAIC in Paris. Most shrunken, by 6 to 8 percent, were the angular gyrus, the superior parietal lobule, and the precuneus regions. All nine regions showed a trend in the same direction, though not all are statistically significant, Quiroz said. Consistent with previous studies in other populations, these results point to neurodegeneration well underway by this stage, which in this population corresponds to what is generally called pre-MCI. With the Paisa AD mutation, affected carriers generally meet MCI criteria by age 44, a bit older than this cohort. In neuropsychological testing, this cohort, whose average age was 38, performed similarly overall to the non-carriers, though trends toward subtle decrements in word recall, verbal fluency, and recall of drawings were apparent. The earliest known cognitive deficit clearly demonstrated in this form of AD—in carriers in their thirties just like the ones studied for cortical thinning—is the visual binding memory deficit reported by Mario Parra and colleagues (see ARF related news story; Acosta-Baena et al., 2011 and Parra et al., 2010).

How do all these findings on brain imaging relate to Aβ? Amyloid PET results from API are unavailable as yet, but the first CSF and plasma data are beginning to roll in. At AAIC, Reiman presented the first cut, on 10 carriers and 10 non-carriers in the 18-26 age bracket. At this age, cognitive tests detect no difference, but brain function and structure measurements do. So far, Reiman reported, it looks as if carriers have elevated plasma Aβ42 but not Aβ40, suggesting that the presenilin 1 E280A mutation raises systemic absolute levels of this more aggregation-prone form of the peptide, as well as the Aβ42/40 ratio. (To some audience members, this finding hinted that middle-age elevated plasma Aβ42 might prove to be a risk factor in the general population as well.)

In CSF at this age, Aβ42 but not Aβ40 is elevated as well in carriers over non-carriers, Reiman reported at AAIC. This is consistent with the DIAN’s prior finding of elevated Aβ42 in carriers of a variety of early onset AD mutations in their twenties (see ARF DIAN London story; see ARF DIAN Honolulu story). Scientists generally assume that this reflects overproduction of Aβ, implying elevated levels of the peptide in the brain at an age where there is no fibrillar amyloid deposition yet. Not everything fits neatly, though: The same study finds a paradoxical reduction of CSF tau in carriers at this young age, upwards of 20 years prior to dementia, Reiman noted at AAIC.

What does this mean? It’s too early to make a strong statement, and it’s not proven that this form of early onset AD models LOAD, both Fleisher and Reiman cautioned in separate conversations. “Even so, at present it looks as if the functional and structural brain changes precede fibrillar amyloid deposition,” Reiman said, noting that this would be consistent with published work on reduction on FDG PET or in mitochondrial glucose metabolism in young adult ApoE4 carriers. Some studies are beginning to hint that fibrillar amyloid deposition, as visible by PET, happens soon after CSF Aβ42 has begun to drop. It is tempting, then, to speculate that the early functional and structural changes that Quiroz, Fleishman, and colleagues see might be happening in a situation of years of elevated Aβ levels but prior to when the brain deposits and, presumably, sequesters. This could imply that fibrillar amyloid deposition is an attempt by the brain to mitigate damage to synapses from an overabundance of prefibrillar forms of AD, Reiman said.

Both API and DIAN are pressing to add both cross-sectional and longitudinal data so they can address at what ages CSF Aβ42 starts dropping and how all markers the studies are tracking fit together. “More data on larger numbers of volunteers will sort this out,” Fleisher said. In the process, the currently proposed staging diagrams of preclinical (e.g., Perrin et al., 2009; Jack et al., 2010; Weiner et al., 2010; Frisoni et al., 2010) may get updated as some curves change their shape and slope or even trade places.

The result will be a knowledge base on the natural history of AD as a foundation for better clinical trials. For now, the API scientists are planning a first clinical trial as outlined in its pending grant proposal to the NIA, provided they can secure an appropriate compound, funding, and regulatory and ethical approval. At CTAD, Tariot emphasized that this trial is designed without pre-formed assumptions on which biomarker patterns will prove to be good outcome markers. Instead, it is designed precisely to address this question. “We must be humble about what we know,” Tariot said at CTAD, noting that regulators had advised API in previous planning meetings that their first trial should use a cognitive endpoint and include many biomarker readouts as secondary endpoints in order to learn as much as possible about them. Because the field does not know which biomarkers will prove to be outcome measures and how they will behave in response to a drug, the current trial is primarily frequentist with some adaptive elements. “We lack sufficient natural history data to build the computer models for a true Bayesian trial, and we have to be agnostic about the ability of biomarkers to predict treatment response. This is why we are not ready to use a Bayesian model yet,” Tariot said.

The proposed API trial, then, would use a change in a composite cognition measure as the primary outcome, looking for a slower rate of decline on drug versus placebo. Jessica Langbaum at the Banner Institute and colleagues elsewhere are developing this measure (see ARF related news story). Because this change will emerge slowly, the trial needs to be large and long. As proposed, the trial would enroll 300 participants. Two hundred carriers would be randomized 1:1 to treatment or placebo so no one would have to find out his or her mutation status; 100 non-carriers would be on placebo. The trial would feature an interim analysis after two years, guided by rules that assume biomarkers will change before cognition does. If the trial shows a positive biomarker pattern and/or clinical trends, then it will continue to five years, long enough to learn whether favorable cognitive changes are detectable.

Overall, the Alzheimer’s research field went from thinking a few years ago that this is too out-of-the-box to multiple groups now doing the same thing. In particular, industry scientists previously pointed to the absence of a regulatory path (see ARF eFAD essays). That path is clearer now, and involvement and support on the part of regulators have been evident. “The feedback from the regulatory scientists to API and DIAN has been incredibly valuable,” Tariot told the audience at CTAD (see ARF related news story; ARF news story). With an emerging regulatory path, the patients, the protocols, the tools, and some biomarker data in hand, researchers know the fate of those initial trials at this point would seem to lie squarely in the hands of funders.—Gabrielle Strobel.

This is Part 2 of a three-part series. See also Part 1 and Part 3. Download a PDF of the entire series.


No Available Comments

Make a Comment

To make a comment you must login or register.


News Citations

  1. DIAN Forms Pharma Consortium, Submits Treatment Trial Grant
  2. Colombians Come to Fore in Alzheimer’s Research, Mass Media
  3. News Flash: Colombian Families Come to Phoenix for Amyloid PET
  4. Detecting Familial AD Ever Earlier: Subtle Memory Signs 15 Years Before
  5. London: What, No Argument? Speakers Agree on Trials for Familial AD
  6. DIAN Dispatch from Hawaii: Glimpse at Data, Push for Trials
  7. Scientists and Regulators Discuss Preclinical AD Trials
  8. London: What Regulators Say About Trials in Familial AD
  9. Anti-Amyloid Treatment in Asymptomatic AD Trial

Paper Citations

  1. . Encoding novel face-name associations: a functional MRI study. Hum Brain Mapp. 2001 Nov;14(3):129-39. PubMed.
  2. . Tracking atrophy progression in familial Alzheimer's disease: a serial MRI study. Lancet Neurol. 2006 Oct;5(10):828-34. PubMed.
  3. . Alzheimer-signature MRI biomarker predicts AD dementia in cognitively normal adults. Neurology. 2011 Apr 19;76(16):1395-402. PubMed.
  4. . Pre-dementia clinical stages in presenilin 1 E280A familial early-onset Alzheimer's disease: a retrospective cohort study. Lancet Neurol. 2011 Mar;10(3):213-20. PubMed.
  5. . Visual short-term memory binding deficits in familial Alzheimer's disease. Brain. 2010 Sep;133(9):2702-13. PubMed.
  6. . Multimodal techniques for diagnosis and prognosis of Alzheimer's disease. Nature. 2009 Oct 15;461(7266):916-22. PubMed.
  7. . Hypothetical model of dynamic biomarkers of the Alzheimer's pathological cascade. Lancet Neurol. 2010 Jan;9(1):119-28. PubMed.
  8. . The Alzheimer's disease neuroimaging initiative: progress report and future plans. Alzheimers Dement. 2010 May;6(3):202-11.e7. PubMed.
  9. . The clinical use of structural MRI in Alzheimer disease. Nat Rev Neurol. 2010 Feb;6(2):67-77. PubMed.

Other Citations

  1. ARF eFAD essays

Further Reading