Up to two decades before people develop the characteristic memory loss and confusion of Alzheimer’s disease, damaging clumps of protein start to build up in their brains.
Now, a blood test to detect such early brain changes has moved one step closer to clinical use.
Researchers from Washington University School of Medicine in St. Louis report that they can measure levels of the Alzheimer’s protein amyloid beta in the blood and use such levels to predict whether the protein has accumulated in the brain.
When blood amyloid levels are combined with two other major Alzheimer’s risk factors – age and the presence of the genetic variant APOE4 – people with early Alzheimer’s brain changes can be identified with 94% accuracy, the study found.
The findings, published Aug. 1 in the journal Neurology, represent another step toward a blood test to identify people on track to develop Alzheimer’s before symptoms arise.
Surprisingly, the test may be even more sensitive than the gold standard – a PET brain scan – at detecting the beginnings of amyloid deposition in the brain.
Such a test may become available at doctors’ offices within a few years, but its benefits will be much greater once there are treatments to halt the disease process and forestall dementia.
Clinical trials of preventive drug candidates have been hampered by the difficulty of identifying participants who have Alzheimer’s brain changes but no cognitive problems.
The blood test could provide a way to efficiently screen for people with early signs of disease so they can participate in clinical trials evaluating whether drugs can prevent Alzheimer’s dementia.
“Right now we screen people for clinical trials with brain scans, which is time-consuming and expensive, and enrolling participants takes years,” said senior author Randall J. Bateman, MD, the Charles F. and Joanne Knight Distinguished Professor of Neurology.
“But with a blood test, we could potentially screen thousands of people a month.
That means we can more efficiently enroll participants in clinical trials, which will help us find treatments faster, and could have an enormous impact on the cost of the disease as well as the human suffering that goes with it.”
The test, an earlier version of which first was reported two years ago, uses a technique called mass spectrometry to precisely measure the amounts of two forms of amyloid beta in the blood: amyloid beta 42 and amyloid beta 40.
The ratio of the two forms goes down as the amount of amyloid beta deposits in the brain goes up.
The current study involved 158 adults over age 50. All but 10 of the participants in the new study were cognitively normal, and each provided at least one blood sample and underwent one PET brain scan.
The researchers classified each blood sample and PET scan as amyloid positive or negative, and found that the blood test from each participant agreed with his or her PET scan 88 percent of the time, which is promising but not accurate enough for a clinical diagnostic test.
In an effort to improve the test’s accuracy, the researchers incorporated several major risk factors for Alzheimer’s.
Age is the largest known risk factor; after age 65, the chance of developing the disease doubles every five years.
A genetic variant called APOE4 raises the risk of developing Alzheimer’s three- to fivefold.
And gender also plays a role: Two out of three Alzheimer’s patients are women.
When the researchers included these risk factors in the analysis, they found that age and APOE4 status raised the accuracy of the blood test to 94%. Sex did not significantly affected the analysis.
“Sex did affect the amyloid beta ratio, but not enough to change whether people were classified as amyloid positive or amyloid negative, so including it didn’t improve the accuracy of the analysis,” said first author Suzanne Schindler, MD, PhD, an assistant professor of neurology.
Further, the results of some people’s blood tests initially were considered false positives because the blood test was positive for amyloid beta but the brain scan came back negative.
But some people with mismatched results tested positive on subsequent brain scans taken an average of four years later.
The finding suggests that, far from being wrong, the initial blood tests had flagged early signs of disease missed by the gold-standard brain scan.
There is growing consensus among neurologists that Alzheimer’s treatment needs to begin as early as possible, ideally before any cognitive symptoms arise.
By the time people become forgetful, their brains are so severely damaged no therapy is likely to fully heal them.
But testing preventive treatments requires screening thousands of healthy people to find a study population of people with amyloid build-up and no cognitive problems, a slow and expensive process.
Surprisingly, the test may be even more sensitive than the gold standard – a PET brain scan – at detecting the beginnings of amyloid deposition in the brain. The image is in the public domain.
As part of the study, the researchers analyzed the enrollment process for a prominent Alzheimer’s prevention trial called the A4 study that used PET scans to confirm the presence of early Alzheimer’s brain changes in potential participants.
They concluded that prescreening with a blood test followed by a PET scan for confirmation would have reduced the number of PET scans needed by two thirds.
Unlike blood tests, which cost a few hundred dollars, each PET scan costs upward of $4,000.
A single site can only run a few dozen PET brain scans a month, because PET scanners are primarily reserved for patient care, not research studies.
“If you want to screen an asymptomatic population for a prevention trial, you would have to screen, say, 10,000 people just to get 1,500 or 2,000 that would qualify,” Bateman said.
“Reducing the number of PET scans could enable us to conduct twice as many clinical trials for the same amount of time and money.
It’s not the $4,000 per PET scan that we’re worried about.
It’s the millions of patients that are suffering while we don’t have a treatment.
If we can run these trials faster, that will get us closer to ending this disease.”
Alzheimer’s disease (AD) is pathologically characterized by the accumulation of amyloid-β (Aβ) plaques, neurofibrillary tangles and widespread neuronal loss in the brain.
In recent years, blood biomarkers have emerged as a realistic prospect to highlight accumulating pathology for secondary prevention trials. Neurofilament light chain (NfL), a marker of axonal degeneration, is robustly elevated in the blood of many neurological and neurodegenerative conditions, including AD.
A strong relationship with cerebrospinal fluid (CSF) NfL suggests that these biomarker modalities reflect the same pathological process. Yet, the connection between blood NfL and brain tissue pathology has not been directly compared.
In this study, longitudinal plasma NfL from cognitively healthy controls (n = 12) and AD participants (n = 57) were quantified by the Simoa platform. On reaching post-mortem, neuropathological assessment was performed on all participants, with additional frozen and paraffin-embedded tissue acquired from 26 participants for further biochemical (Aβ1–42, Aβ1–40, tau) and histological (NfL) evaluation.
Plasma NfL concentrations were significantly increased in AD and correlated with cognitive decline, independent of age. Retrospective stratification based on Braak staging revealed that baseline plasma NfL concentrations were associated with higher neurofibrillary tangle pathology at post-mortem.
Longitudinal increases in plasma NfL were observed in all Braak groupings; a significant negative association, however, was found between plasma NfL at time point 1 and both its rate of change and annual percentage increase.
Immunohistochemical evaluation of NfL in the medial temporal gyrus (MTG) demonstrated an inverse relationship between Braak stages and NfL staining. Importantly, a significant negative correlation was found between the plasma NfL measurement closest to death and the level of NfL staining in the MTG at post-mortem. For the first time, we demonstrate that plasma NfL associates with the severity of neurofibrillary tangle pathology and neurodegeneration in the post-mortem brain.
Electronic supplementary material
The online version of this article (10.1186/s40478-018-0649-3) contains supplementary material, which is available to authorized users.
In recent years, there has been an increasing emphasis on developing a blood biomarker (plasma or serum) to predict the clinical onset of Alzheimer’s disease (AD) or to identify the underlying pathophysiology at its earliest stage.
A blood-based measure has substantial practical and economic advantages over the most well-established AD biomarkers. Structural imaging (MRI) and positron emission tomography (PET) techniques using 18F-fluorodeoxyglucose, amyloid tracers, and, more recently, tau ligands are costly, and access is limited to specialised centres.
On the other hand, cerebrospinal fluid (CSF) sampling is becoming routine in neurology clinics and the cost for the core AD CSF immunological assays (Aβ1–42, phosphorylated tau [P-tau] and total tau [T-tau]) are much lower per patient than for PET scans.
Yet, there remains a level of perceived invasiveness or complexity attached to a lumbar puncture in many countries.
Therefore, a blood-based marker would complement CSF and molecular imaging biomarkers as a simplified initial triage step in a multi-stage assessment for early diagnosis, secondary prevention trial participant selection or monitoring of response to intervention over time (for review see, (Lewczuk et al. 2017)).
It has proven challenging to establish a robust blood biomarker for AD.
Candidate proteins are influenced by peripheral expression from extra-cerebral tissues and the measurements are confounded by matrix effects from plasma proteins (Zetterberg 2015).
More recently, ultra-sensitive immunological assays (Single Molecule Array) and mass spectrometric studies have overcome these complications and recently reported plasma amyloid-beta peptide ratios (Aβ1–42/Aβ1–40 and APP699–711/Aβ1–42) that can identify cerebral amyloidosis with high accuracy (Janelidze et al. 2016; Nakamura et al. 2018). Furthermore, using the Simoa platform, plasma T-tau has been shown to be consistently elevated in AD (Mattsson et al. 2016) and to correlate with cognitive decline (Mielke et al. 2017).
One promising avenue for AD is the neuroaxonal injury marker neurofilament light (NfL). Patients with AD have increased NfL concentrations in the CSF (Sjogren et al. 2000) and this is also reflected in the blood despite being more than 50-fold lower in concentration (Gisslen et al. 2016; Lewczuk et al. 2018).
Increases in blood NfL have also been described in frontotemporal dementia (Rohrer et al. 2016), Huntington’s disease (Byrne et al. 2017) and atypical parkinsonian disorders (Hansson et al. 2017), making blood NfL a global marker of neurodegeneration rather than a disease-specific biomarker per se.
Intriguingly, plasma NfL correlates highly with CSF NfL in the same individuals (correlation coefficients > 0.75 (Zetterberg and Blennow 2018)), further supporting the notion that plasma NfL indeed reflects central nervous system damage and could be used as a proxy measure for CSF NfL. However, at this time, no evidence has been gathered to suggest if blood NfL reflects the degree of neuropathology at post-mortem.
In this study, we sought to investigate this relationship by measuring longitudinal plasma NfL concentrations in cognitively unimpaired elderly controls (CTL) and patients with a clinical diagnosis of probable AD. For all participants, Braak staging of tau pathology was performed at post-mortem. Biochemical measures of brain Aβ (Aβ1–42, Aβ1–40 and Aβ1–42/ Aβ1–40) and tau (P-tau and T-tau), as well as immunohistochemical assessment of NfL, were performed in a subset.
Washington University School of Medicine
Judy Martin Finch – Washington University School of Medicine
The image is in the public domain.
Original Research: Closed access
“High-precision plasma β-amyloid 42/40 predicts current and future brain amyloidosis”. Suzanne E. Schindler, James G. Bollinger, Vitaliy Ovod, Kwasi G. Mawuenyega, Yan Li, Brian A. Gordon, David M. Holtzman, John C. Morris, Tammie L.S. Benzinger, Chengjie Xiong, Anne M. Fagan, Randall J. Bateman.