How Biomarkers are Changing Knowledge and Management of Alzheimer's Disease

As research into Alzheimer's disease biomarkers intensifies, the accumulating evidence suggests that through a combination of genetic screening and imaging, clinicians may eventually be able to detect AD early on and perhaps tailor their treatments to delay dementia progression. These tantalizing findings also have important implications for new therapeutic development. Practical Neurology interviewed researchers and clinicians about recent developments.

Liana G. Apostolova, MD is Assistant Professor of Neurology at University of California, Los Angeles School of Medicine and on the staff of the Brain Research Institute at UCLA. Her recent research is in Imaging Genetics, and she is an investigator in the ImaGene study now underway.

Which biomarkers or types of biomarkers are showing the greatest value in detecting familial AD? What about sporadic AD?

Dr. Apostolova: Unlike familial autosomal dominant AD, where we can determine with absolute certainty which at-risk individuals will develop AD with genetic testing, determining one's risk for late onset (sporadic) AD (LOAD) is far more challenging. Several types of LOAD biomarkers are being actively developed with the hope that any one or a combination of them would be able to reliably predict one's risk for dementia. The most promising biomarkers are structural (MRI), functional (PET) and molecular neuroimaging (PIB-PET, AV45-PET) and CSF levels of abeta and tau. Other under development, but far from being near prime time, are other molecules in the CSF or in blood.

Which biomarkers do you think have the most potential clinical utility and are most likely to lead to practical clinical uses? Which are least likely to have clinical relevance?

Dr. Apostolova: Structural MRI, FDG PET and amyloid PET are the imaging biomarkers that already have or, in the case of amyloid PET, will likely soon have clinical utility. CSF abeta and tau are likewise useful. Protein levels in blood are, at least for now, least helpful.

In recent years we have whitnessed an exciting spurt in human genetics and many reports of genes that are related to human diseases have been published. While they are not exacltly biomarkers, genes influencing one's risk for developing AD are enourmusly important. Such risk genes will one day become very helpful in determining one's likelihood for developing AD down the road. Preventative therapies would be then prescribed to those at high risk to develop AD.

A good deal of research has been emerging on biomarkers in AD in recent years. What are the next steps for your research personally? What are the key questions you'd like to see answered by yourself or others?

Dr. Apostolova: My research presently lies at the crossroads of imaging and genetics. The new field of Imaging Genetics is aiming to define the effect of individual genes or traits on disease symptoms and characteristics as well as on the brain structural and functional changes that occur in AD.

I am the Principal Investigator of the ImaGene project at UCLA—a prospective five-year long study that follows subjects who are either cognitively normal or meet criteria for mild cognitive impairment. These subjects provide detailed clinical and neuropsychological evaluation, structural MRI and peripheral blood, on an annual basis. We do full genome genotyping at baseline (already completed) and study the changes in gene expression annually. The importance of structural imaging and genetics were already discussed.

Gene expression will provide us with critical insight into important critical downstream disease pathways and might help guide therapeutic development. This project is well positioned to improve our understanding of the genetic map of sporadic AD and bring us closer to iindividualized prevention and pharmacogenomics.

Many specialists have suggested that biomarkerbased diagnosis of AD will rely on assessment of multiple markers, including CSF markers, genetic markers, and imaging? Do you concur? If so, what are the greatest obstacles to implementing this approach to diagnosis?

Dr. Apostolova: Yes, I absolutely concur. This is exactly what my lab is doing; We are now looking into combining imaging, genetic and gene expression data into powerful unbiased diagnostic and prognostic classifiers. We have demonstrated that the combined classifiers lead to substantial improvement in diagnostic accuracy (e.g., from 60 percent when only one biomarker is used to 84 percent as three different types of biomarkers are used).

Just a few years ago, most experts were deterring patients from undergoing ApoE status testing, given that findings did not provide a clear risk profile, nor was there any therapeutic implication to these findings. What is the current status of genetic profiling for patients with a family history of early-onset AD? Is there any role for genetic testing in the general population yet?

Dr. Apostolova: The attitude towards ApoE4 testing is unchanged. Yet it is likely to change in future years once we understand better the role of genetics for AD. Currently only 50 percent of AD's heritability is explained; Many more genes are yet to be discovered. Once we get there, we will be capable of pre-symptomatic risk assessment, which will be tremendously valuable if we had a cure or a disease-modifying agent. But in order to get there we need at least 10 or more years of rigorous genetic research of two types:

1. large multi-site genetic consortia which will discover new risk genes and
2. Imaging Genetics (or other gene/phenotype) studies like the ImaGene study I am currently leading at UCLA that will demonstrate unequivocally the effect of these newly discovered risk genes on the human brain in the symptomatic and presymptomatic stages and define the downstream events—the genetically triggered abnormal pathways.

This will ultimately lead to discovery of new therapeutic targets.

John Ringman, MD, MS is an Associate Clinical Professor of Neurology and board-certified neurologist at the David Geffen School of Medicine at UCLA and is the interim Director of the UCLA California Alzheimer's Disease Center. His recent research has shown that in tau and p-tau in the CSF are increased early in familial AD.

Which biomarkers or types of biomarkers are showing the greatest value in detecting familial or sporadic AD?

Dr. Ringman: In familial AD, our studies are showing that increases in tau and p-tau in the CSF are the earliest changes occurring in familial AD. Abeta is also going down, but it appears that it may be falling from a previously elevated level in this population (preliminary data); it is in the normal range when tau and p-tau are already up. Increases in PIB binding are occurring somewhat later, with MRI atrophy occurring later still.

With sporadic AD, the most commonly held notion is that abeta going down in the CSF is coincident with positive amyloid binding followed by characteristic hypometabolism on FDG-PET and then increased tau in the CSF and atrophy on MRI.

Which biomarkers do you think have the most potential clinical utility and are most likely to lead to practical clinical uses? Which are least likely to have clinical relevance?

Dr. Ringman: I think amyloid imaging and CSF for abeta and tau will have the most utility in diagnosing AD early. They will be used in the context of persons who want to know what their future holds and/or in the event we have disease-modifying treatments.

I find FDG-PET is less useful for diagnosis.

What are the next steps for your research personally? What are the key questions you'd like to see answered by yourselves or others?

Dr. Ringman: We're identifying interesting new markers in presymptomatic persons with mutations causing familial AD. I hope to evaluate their utility in persons developing late-onset AD.

Do you agree that biomarker-based diagnosis of AD will rely on assessment of multiple markers, including CSF markers, genetic markers, and imaging? If so, what are the greatest obstacles to implementing this approach to diagnosis?

Dr. Ringman: I would agree that right now no one marker is 100 percent diagnostic. However, I anticipate that we will be able to get enough information from, say, the CSF, such that other markers are not necessary. The lack of sufficiently effective interventions is the main obstacle to implementation of this and more importantly is the main obstacle to helping people!

What is the current status of genetic profiling for patients with a family history of early-onset AD? Is there any role for genetic testing in the general population yet?

Dr. Ringman: I think it is still unhelpful to genotype people at this time, but this might change— specifically in regard to Apo-E testing—soon as certain treatments that may be approved may be differentially effective and risky depending on the individual's Apo-E genotype.

Lidia Glodzik, MD, PhD, is research assistant professor in the Department of Psychiatry and on staff at the Center of Excellence on Brain Aging at NYU Langone Medical Center. She recently investigated the relationship between Alzheimer's disease markers, hypertension, and gray matter damage.

What did your study show regarding CSF biomarkers, hypertension, and gray matter density?

Dr. Glodzik: We found that cerebrospinal fluid biomarkers reflecting tau pathology—total tau and phosphorylated tau—correlate with how much gray matter we have in regions which are known to be affected in Alzheimer's disease.¹ Hypertension was related to atrophy in different brain regions. Subjects with both hypertension and higher levels of phosphorylated tau presented, not surprisingly, more global gray matter atrophy.

When we analyzed gray matter in the medial temporal lobe using less stringent statistical criteria than the ones used in statistical parametric mapping, we observed that both tau markers and hypertension contributed to medial temporal lobe atrophy. The medial temporal lobe is the earliest site where Alzheimer's disease related pathology can be observed.

Your findings seem to suggest that using biomarkers to identify pre-clinical AD may be a multi-factorial process. Do you agree with this assessment? What does it mean for the utility of biomarkers for the diagnosis or detection of AD?

Dr. Glodzik: We tried to show that different modalities—CSF and MRI—correlate with each other. It makes sense, since both are used to detect different aspects of the same pathological process. Correlations are not perfect, but this on the other hand is a rationale for using multiple markers (CSF, PET, MRI) in detection, since each of them will bring some information missing from the other. Both the work from our Center and from other labs shows that this multimodal approach helps better predict, for example, who among MCI subjects will develop AD.

In subjects who are cognitively healthy (like in the current paper) the issue is much more complicated. The signal is “more diluted.” We were able to show that subjects who have more tau markers show more atrophy in AD-specific brain regions, but we still don't know if and when they will develop AD. However, I think it is safe to say that if we will be able to predict the disease it will most likely be based on many biomarkers, not just one.

Do your findings suggest any implications for the approach to AD management?

Dr. Glodzik: Our findings confirm that hypertension is related to brain atrophy. It is still a matter of debate whether hypertension contributes to AD pathology or simply diminishes brain reserve. Either way, there is no doubt it should be treated.

What next steps do you expect to come from your research findings?

Dr. Glodzik: I hope with our next projects we will be able to show longitudinal dynamics of multimodal AD biomarkers and possibly some influence of interventions, not just correlations. To determine their true predictive value in cognitively healthy middle aged and older subjects, a really long follow-up time is necessary. We hope to be able to continue our observation.

What do you see for the role of biomarkers in clinical AD management in the next five to 10 years? Will AD be diagnosable through biomarkers?

Dr. Glodzik: I hope so. Probably it will be more than one biomarker. A lot of work has been done in the domains of neuroimaging and CSF biomarkers, but the field of blood biomarkers is growing. I believe with time and refinement of detection technique it will contribute both to detection and understanding of pathomechanisms.

Joshua D. Grill, PhD is Adjunct Instructor; Director of the Katherine and Benjamin Kagan Alzheimer's Disease Treatment Development Program; Director of Recruitment and Education Core; and Associate Director, Deane F. Johnson Center for Neurotherapeutics at UCLA.

As an AD researcher, what do you see as the most important recent developments in the realm of AD biomarker research?

Dr. Grill: AD biomarkers will create an opportunity to ensure enriched populations in clinical trials. For example, trials of anti-amyloid therapies may be designed to include only persons with positive amyloid PET scans or decreased CSF beta amyloid levels. Biomarkers can also be used as outcome measures, especially in early stage clinical trials. Examining drug effect on target biomarkers can be instructive in decisions whether to move an agent forward in clinical development. This is important for a variety of reasons, including that later stage trials are expensive and utilize a wide variety of resources, including volunteer patient participants.

Is knowledge of biomarkers influencing the design and conduct of clinical trials in AD?

Dr. Grill: The increased understanding of AD biomarkers and the time course that precedes dementia onset has led to a general consensus that intervention can be attempted earlier in the disease and may be necessary in order to be successful. Thus, studies that enroll only participants who demonstrate biological evidence of disease but do not meet criteria for dementia are ongoing.

Do you believe biomarkers research will lead to new therapies in the near term or is it more likely to improve outcomes with existing therapies by allowing clinicians to better match therapy to patient needs?

Dr. Grill: Biomarkers can greatly facilitate drug development by helping make early stage choices, enriching enrolled populations, and potentially by allowing for smaller studies if used as outcome measures.

What do you foresee as the greatest potential clinical benefit of biomarker research in AD?

Dr. Grill: If biomarkers facilitate the development of disease-slowing therapies, the availability of novel therapies will be of the greatest clinical benefit.

What do you perceive to be the greatest limitations of biomarker research in AD?

Dr. Grill: As yet, no biomarker has been validated as a surrogate marker for a clinical outcome (i.e. cognition). Trials that utilize biomarkers as outcome measures can enroll smaller populations and allow for more interventions to be tested at any given time. But because the burden of demonstrating efficacy to regulatory bodies currently necessitates clinical outcomes, registration trials will remain very large and very long studies, for now.

1. Glodzik L, Mosconi L, Tsui W, de Santi S, Zinkowski R, Pirraglia E, Rich KE,McHugh P, Li Y, Williams S, Ali F, Zetterberg H, Blennow K, Mehta P, de Leon MJ. Alzheimer's disease markers, hypertension, and gray matter damage in normal elderly. Neurobiol Aging. 2011 Apr 27, Epub