August 2009: Volume 35, Number 8
The Search for Alzheimer’s Diagnostics
Are They Close or Still Years Away?
By Genna Rollins
Despite significant research efforts, Alzheimer’s disease is poised to remain one of the most pressing healthcare problems of the 21st century. The neurodegenerative condition, which has devastating consequences for its victims and their loved ones, affects about 5.3 million people in the U.S. today, with an estimated 500,000 new cases each year. But with the aging of the baby boom population, the incidence is expected to rise to nearly 1 million cases per year by 2050. In contrast to other leading causes of death such as heart disease, breast and prostate cancer, which have declining death rates, deaths from Alzheimer’s disease rose approximately 47% between 2000 and 2006. At the same time healthcare costs are three times higher for patients with Alzheimer’s than for other people age 65 and older. Without effective treatments and a means of diagnosing patients earlier in the disease process, care associated with Alzheimer’s has the potential to wreak havoc on the U.S. healthcare system.
“As the number of people with Alzheimer’s disease grows, it’s simply going to sop up the total available budget for healthcare,” predicted William Thies, PhD, chief medical and scientific officer of the Alzheimer’s Association. “Any kind of reorganization in healthcare or progress we hope to make is entirely dependent on being successful at doing a better job of managing people with Alzheimer’s disease. And that’s ultimately what’s at stake in finding better therapies and diagnostics.”
Thies’s clarion call has been heard loudly and clearly by the research community. There are several major clinical trials in progress, as well as numerous lines of investigation aimed at understanding the biochemistry and pathogenesis of the disease. The ultimate goals of these efforts are to develop effective treatments and to identify and validate biomarkers that can be used to diagnose the disease, assess the risk for developing Alzheimer’s, and monitor the hoped-for treatments. Investigators in the field report that science is edging closer to both biofluid and imaging markers that will accomplish these aims. “The idea is that certain biomarkers are already abnormal before the onset of symptoms or impairments, so the earlier you go in the disease process, the more important the biomarkers become,” explained Pieter Visser, MD, PhD, assistant professor at the University of Maastricht and VU Medical Center (The Netherlands) and coordinator of the Development of Screening Guidelines and Diagnostic Criteria for Predementia Alzheimer’s disease (DESCRIPA) trial.
For the time being, a presumptive diagnosis of Alzheimer’s can be made clinically using various cognition tests, neurological exams, and patient history. A definitive diagnosis is possible only through post-mortem brain analysis, but clinical diagnosis has sensitivity as high as 90%. Unfortunately, by the time symptoms appear and a clinical diagnosis is made, the disease has been simmering for decades and intractable neurological damage occurred. Worse still, clinicians have little to offer in terms of reversing or even stabilizing the inevitable further decline. In the words of one leading Alzheimer’s researcher, “this neurodegenerative dementia is the most devastating major human disorder for which there is still no effective long-term treatment” (J Practical Neurology 2004;3:30–34). Five drugs have been approved by the Food and Drug Administration for Alzheimer’s treatment to address symptoms of the disease but have no indications for delaying or halting its progression.
Age is the major risk factor for Alzheimer’s, with the risk doubling approximately every 5 years after age 65. By the age of 85, the risk is close to 50%. The vast majority of cases are sporadic, but inheritance of the apolipoprotein E (APOE) ε4 allele is the strongest genetic risk factor discovered so far and may play a role in more than half of late-onset cases. However, the APOE ε3/ε3 genotype is the most common, carried by about half the population. The rarer APOE ε2/ε2 genotype has been associated with reduced risk.
The Amyloid β-Tau Connection
While intensive research efforts haven’t completely unearthed the mystery of Alzheimer’s disease, considerable progress still has been made. Scientists have established and thoroughly documented that the brain tissue of people with Alzheimer’s is filled with senile plaques and neurofibrillary tangles, and that over time there is decreased brain activity and brain shrinkage. Volumetric structural imaging now can capture the brain atrophy, and a relatively new technique, Pittsburgh Compound B (PiB) PET imaging, uses a tracer that binds selectively to amyloid ß deposits, so the extent of disease can be seen in vivo. Even with such breakthroughs, however, many unanswered questions remain about the mechanisms behind these hallmark manifestations of the disease.
The The amyloid ß 1–42 peptide (Aß1-42), which is the major component of the plaques, is produced in neurons and other brain cells by the amyloid precursor protein via sequential ß- and γ-secretase enzyme cleavage. Amyloid ß is a normal product of the brain, but Aß1-42 is the least soluble of the Aß peptides and it tends to clump together, forming plaques that impair the synapses and neuronal dendrites, eventually causing cell death.
Years of research lead to the discovery that tau, a protein involved in microtubule assembly and stabilization, is the primary component of Alzheimer’s characteristic neurofibrillary tangles. When tau functions normally, it supports microtubule processes involved in transporting molecules within the cells. But in Alzheimer’s disease, tau undergoes hyperphosphorylation, which leads to disruption of microtubules, loss of neuronal transport mechanisms, neuronal cell death and formation of paired helical filaments that are the major constituent of the tangles.
Initially the Aß1-42 and tau pathologies seem to appear independently, with tau developing first in the hippocampus and spreading to the neocortex, and Aß plaques appearing in the neocortex. Extensive plaque build-up may precede tangle pathology.
Recent research indicates that as Alzheimer’s disease progresses, levels of Aß1-42 decrease in cerebrospinal fluid (CSF) and that the ratios of Aß1-42 to total tau and hyperphosphorylated tau are bellwethers of the disease. New research by Visser and his colleagues has found that an abnormal Aß1-42:tau ratio predicted development of Alzheimer’s-type dementia in patients with amnestic mild cognitive impairment (Lancet Neurol 2009;8:619–27). “The rationale that reflects our best current understanding is that as the plaque burden builds up, the Aß1-42 is depleted from the surrounding brain tissue, thereby depleting the CSF of Aß1-42,” explained Leslie Shaw, PhD, professor of pathology and laboratory medicine and director of the toxicology laboratory and biomarker research laboratory in the Department of Pathology and Laboratory Medicine at the University of Pennsylvania. Shaw also is biomarker core director of the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and a collaborator of the AACC 2010 Wallace H. Coulter Lectureship Award winner, John Q. Trojanowski, MD, PhD, a leading researcher in the field.
Curiously, the mere presence of Aß1-42 and plaques doesn’t necessarily signal disease. “The amyloid story is a bit complex,” said Thies. “The major species elaborated in the brain is elaborated all your life, but things change subtly as you age, so that you’re making slightly longer amino acid chains—42 amino acids instead of 40—with the ratio changing as the disease progresses. However, one of the things that’s fascinating is that the amount of amyloid accumulating in Alzheimer’s doesn’t track perfectly with the clinical course.”
ADNI is one of the major research initiatives currently underway that is exploring the Aß-tau pathology. The $67 million, 6-year trial, which is a public-private partnership between NIH, 15 pharmaceutical and biotech companies, and two non-profits, including the Alzheimer's Association, has produced intriguing results so far. This prospective multisite trial is the most extensive of its kind to collect imaging, fluid biomarkers, and clinical data to identify the best markers for following disease progression and monitoring response to treatment. “ADNI grew out of conversations I had with several pharmaceutical companies that are trying to develop drugs for modifying the Alzheimer's disease process,” explained Neil Buckholtz, PhD, chief of the Dementias of Aging Branch at the National Institute of Aging. “The idea was that by using neuroimaging modalities and fluid biomarkers this would be a way of distinguishing between the symptomatic benefit and disease-modifying effect of any drugs that might be developed.”
The initial and primary focus of ADNI is collection of Aß1-42, total tau, phosphorylated tau, homocysteine, and isoprostanes in CSF, and MRI and PET scanning of the brain, with the aim of developing standardized imaging and biochemical biomarker methods and clinically validating them. Four add-on studies to the original project include PiB PET testing, genome-wide genotyping, genetic analysis, and development of serum-based biochemical biomarkers. Most analyses from the main ADNI trial are expected in 2010; ADNI investigators announced in March 2009 that the high-density genome-wide analysis was more than 95% complete. A separate public-private partnership, The Biomarkers Consortium, is spearheading an effort to use an existing multiplex assay of approximately 100 analytes and mass spectroscopy analysis from plasma and CSF.
ADNI has enrolled more than 800 subjects at 57 sites, nearly 400 with amnestic mild cognitive impairment, 229 with normal cognition, and 193 with probable Alzheimer's. A recently published report of 1-year follow-up data on CSF markers found Aß1-42 to be the most sensitive biomarker, with receiver operator characteristic area under the curve of 0.913 and sensitivity of 96.4% (Ann Neurol 2009;65:403–413.). A logistic regression model of Aß1-42, total tau, and APOε4 allele count provided the best discernment of mild Alzheimer's disease. Perhaps most significantly, 33 of 37 subjects with mild cognitive impairment who converted to probable Alzheimer's disease had an Alzheimer's-like total tau:Aß1-42 ratio at baseline.
Changing Definitions, Procedures
Even as ADNI follow-up is continuing, the strength of the first-year results and accumulated evidence from other studies could change the diagnostic landscape. “We’ve recently had discussions about changing the criteria for the definition of Alzheimer’s disease or earlier states like mild cognitive impairment,” said Buckholtz. “The proposal has been put forth to utilize biomarkers in the definition of the disease, so people are thinking about it. I don’t know when it’s going to happen, but it’s definitely being considered by a lot of groups.”
ADNI also has been ground-breaking in other ways. In one of the first such arrangements, data from the trial are being posted and regularly updated on a publicly accessible website available to researchers worldwide. ADNI organizers expect the open data policy will speed additional Alzheimer’s-related research. Feedback from non-ADNI researchers using the data already has informed Shaw’s analysis of CSF markers. “Most of us scientists did not have the experience of posting data before it was published, so it was a real commitment that we all made,” he said. “But to have someone looking at the data differently and having insights that you didn’t have has been enormously rewarding.”
In addition, ADNI may lead to better procedures and analytic standardization in future trials involving CSF analysis. “One of the problems in the field is that various sites use different assay procedures and the range of tau, for example, in CSF can be orders of magnitude different depending on where the study was done,” noted Buckholtz. But three rounds of inter-laboratory validation studies and round robin discussions by ADNI investigators lead to establishment of common assay and CSF collection procedures. “In the earliest days of the study the standard lumbar puncture trays used CSF collection tubes made of polystyrene. But we learned that if the CSF stands in the tubes any length of time, gradually the amyloid beta sticks to the walls and to a smaller extent the tau does as well,” explained Shaw. The discovery lead to restocking all lumbar puncture trays used in the study with polypropylene tubes.
ADNI organizers also went to great lengths to educate participating centers about best practice lumbar puncture procedures and how to counsel subjects about the process. The effort was a success, as organizers expected to collect CSF from only 20% of subjects, but have 51% participation, according to Shaw. ADNI also benefitted from having access to an ADNI-independent set of CSF specimens, collected from a cohort of presumptive Alzheimer’s patients whose diagnosis was confirmed through post-mortem autopsy, along with age-matched samples from cognitively normal subjects.
The structure and success of ADNI has lead to similar efforts in Japan, China, and Australia. Lead investigators of the four initiatives routinely hold conference calls to discuss progress, challenges and next steps, according to Shaw. See box, below, for a list of major Alzheimer’s trials.
DESCRIPA is on a parallel track with ADNI. The European Commission-funded study is using a battery of clinical tools along with neuroimaging, genotyping, and biomarkers to develop clinical criteria for predementia Alzheimer’s (Neuroepidemiology 2008;30:254–265). There are 20 participating centers in 11 countries. Aß1-42 and Aß1-40 will be measured in plasma; Aß1-42, total tau and phosphorylated tau will be measured in CSF. In addition to ADNI and DESCRIPA, numerous other major longitudinal studies are in progress, some expressly designed to investigate markers of predementia Alzheimer’s and others which could be used for such purposes, according to Visser.
While the Aß-tau line of investigation is the most advanced, numerous other theories are being pursued. In all, as many as 150 biomarkers have been implicated as having relevancy to Alzheimer’s, and more than 400 genes have been evaluated for a link to late-onset disease. In addition, drug trials now in progress are examining at least 20 different treatment strategies.
Hypotheses abound about the exact pathogenesis of Alzheimer’s. Early research and drug therapies focused on the impact of decreased synthesis of the neurotransmitter acetylcholine. Other theories involve inflammation, oxidative stress, and even herpes simplex virus type 1 interacting with APOEε4 susceptibilities. There also are numerous variations off the Aß theme. For example, the Beta amyloid oligomers in the Early Diagnosis of Alzheimer’s disease and as Marker for Treatment Response (EDAR) study is exploring the role of Aß oligomers as the key pathogenic agent in the disease. Investigators plan to measure low-concentration oligomers in serum and CSF using immunoprecipitation in combination with ELISA and immune-PCR with nano-structured DNA-protein conjugates. Visser is the coordinator of this study.
In another line of investigation, researchers at McGill University in Montreal recently reported the use of short-wavelength, near-infrared spectrophotometry to identify changes in oxidative stress levels in blood plasma (J Alzheimer’s Dis 2009;17:391–397.) They found that the technique differentiated Alzheimer’s disease from normal controls with a sensitivity of 80% and specificity of 77%. On the tau side, a separate team of McGill researchers recently reported that Ser202 phosphorylation is a principal pathogenic event in Alzheimer’s (J Biol Chem 2009;284:13422–13433).
Finding Serum Markers
Until recently, attempts to use urine or serum markers have not panned out for various reasons, and CSF has been the focus of most biomarker investigations. For instance, in 2005 FDA denied approval of a kit version of the AlzheimAlert™ urine test for neural thread protein citing the need for better demonstration of the test’s performance. The manufacturer, Nymox, offers the test in its own CLIA-certified lab, and the test has received the European CE Mark. The company is continuing to work on improving the test, according to Brian Doyle, director of business development. Still, it has not gained widespread acceptance, according to leading Alzheimer researchers.
Isoprostanes also had been touted as strong biomarker candidates for Alzheimer’s, based on their role in inflammation and oxidative stress. However, Shaw recently found no significant difference in isoprostane levels between Alzheimer’s and cognitively normal subjects. This study, which used liquid chromatography tandem mass spectroscopy, has not been published yet.
Identification of serum or urine markers for Alzheimer’s remains a high priority because collection of these biofluids is less invasive and more practical for wide-spread screening efforts than CSF. Yet challenges remain. Even though Aß and Aß precursor protein are present in various tissues throughout the body, and there is evidence of a dynamic exchange across the blood-brain barrier, efforts to validate Aß peptides in plasma have been “fraught with frustration,” according to researchers. One recent analysis found no statistically significant differences in Aß40, Aß42, total Aß levels, and Aß42:Aß40 ratios between Alzheimer’s subjects and non-demented controls (Alzheimer’s & Dementia 2009:5:18–29). However, plasma Aß levels were found to fluctuate significantly over time within individuals, due in part to Aß’s ability to bind to a variety of plasma and membrane proteins.
One particularly intriguing line of investigation involves patterns of secreted intercellular signaling proteins (Nature Med 2007;13(11):1359–1362). Researchers found that 18 such proteins were able to distinguish between patients with Alzheimer’s and various controls with nearly 90% accuracy. “Alzheimer’s disease is known to have a strong inflammatory component, and inflammatory molecules such as cytokines, chemokines, growth factors and other signaling proteins all and are used by cells to communicate with each other,” explained co-author Markus Britschgi, PhD, instructor in the department of neurology and neurological sciences at Stanford University School of Medicine in Palo Alto, Calif. “We wondered if we could listen in to what these cells were telling each other.” The study’s principal investigator, Tony Wyss-Coray, PhD, has coined the term “communicome” to describe the universe of intercellular communication factors.
A follow-up study also found that the network between these proteins was significantly different in Alzheimer’s patients versus normal controls (Arch Neurol 2009;66:161–165). “The network of cross-correlation between these markers is changed in Alzheimer’s disease and this reflects again changes in biological pathways,” noted Britschgi. For example, the interleukin 1, α protein, which is a pro-inflammatory marker, was not connected with any of the other 17 proteins in healthy controls, but was correlated with 7 proteins in Alzheimer’s patients.
As intriguing as the research involving intercellular communication is, it still has to pass validation muster, which may prove to be a tall order, according to Cris McReynolds, CEO of Satoris. Satoris has licensed the technology used in Wyss-Coray’s lab from Stanford and the VA Palo Alto Health Care System with the goal of developing a commercialized test. In June 2009 the company received a notice of allowance from the U.S. Patent and Trademark Office for claims relating to the use of specific blood biomarkers as an aid in diagnosing Alzheimer’s disease.
Still, commercialization of the technology with an exact replication of the initial findings will be no mean feat. “One of the challenges we face is the platform upon which the original data was generated and the study was based is a platform that has changed over time and therefore can’t be reproduced in the original form,” said McReynolds. The original study used a filter membrane platform, but Satoris is working with a bead-based multiplex antibody assay on the assumption that it will be easier to use in clinical settings.
The hurdles facing Satoris are no different than those awaiting the myriad of biomarkers and drugs under investigation in Alzheimer’s disease, according to Shaw. “Their hypothesis makes logical sense, and we hope to see further studies that can inform us about which tests could possibly contribute to the field,” he said. “There are a number of leads that just have to confirmed. That’s the progression of science.”
Fits and starts may lie ahead for the various lines of investigation, but overall the Alzheimer’s community sees biomarkers and effective medications on, not over, the horizon, Thies observed. “Many people who started in Alzheimer’s research many years ago thought it was an interesting intellectual area but didn’t have much thought of it coming to definitive therapies. But those people now are enthusiastic and optimistic about seeing therapies and diagnostics in the foreseeable future."