Image: A newly developed molecular test could detect Alzheimer\'s disease at a pre-symptomatic state. The results indicated that multivariate analysis of transcripts in blood samples can provide an accurate and minimally invasive strategy for diagnosis of AD and early detection of AD risk (Image courtesy of Arizona State University; Graphic by Jason Drees).
A new multivariate molecular test is being developed that includes testing for genes to identify people at risk of developing Alzheimer’s disease (AD) and testing of peripheral blood leukocyte RNA transcripts to provide early diagnosis of people with AD. The new method distinguished between Alzheimer’s, Parkinson’s, and healthy controls, indicating that it can pick out AD from other degenerative brain conditions.
The enormous lag between the disease’s inception and the appearance of clinical symptoms presents one of the main challenges for diagnostics and treatment. By the time the first outward manifestations of Alzheimer’s appear (e.g. confusion, memory loss), Alzheimer’s has been ravaging the brain for decades. If the disease could be identified much earlier, even close to its origin, then perhaps it could be greatly slowed or even halted. Ideally, such a method should be appropriate for primary care settings, allowing a broad swath of the public to be accurately and regularly tested.
Given the need for a safe and reliable early diagnostic, many previous efforts have taken aim at the problem. However, not only have these efforts run aground, the accuracy of diagnosis even after the disease has entered its clinical phase remains poor.
In their new study, Paul Coleman, Alzheimer’s researcher at Arizona State University (Tempe, AZ, USA) and colleagues demonstrated that their new test has potential as an early diagnostic for AD. The results suggest that AD can be detected even before the onset of symptoms in persons at genetic risk for AD. Study collaborators include researchers from ASU, Mayo Clinic, University of Rochester, Banner Alzheimer Institute, and Barrow Neurological Institute.
“What we’ve done in our paper is to replicate our own work multiple times with different populations and even using different technologies,” said Prof. Coleman, “We also presented data showing the ability to detect people at risk of a future diagnosis for Alzheimer’s disease.”
The method accomplishes this by examining RNA in white blood cells (leucocytes). Alzheimer’s produces changes in the brain that can stimulate genes relating to conditions like stress and inflammation. Specific RNA transcripts from these genes appear in the blood. These transcripts were combined to form an early diagnostic biomarker panel.
The diagnostic precision of the new test is significant. Existing diagnostic screening results for known AD cases (identified through clinical and neuropathological factors), showed diagnostic sensitivity was between 71-87%, while specificity ranged from 44-70%. Such diagnoses are typically conducted in specialized facilities devoted to the study of Alzheimer’s, but the accuracy of standard diagnosis falls significantly in primary care settings. The result is that Alzheimer’s is generally detected (if at all) very late, a blueprint for treatment failure because the illness has already irreparably damaged the brain. Also, the high rate of misdiagnosis leads to frequently unnecessary and ineffective treatment.
In a fresh approach, the authors identify RNA transcripts in blood using two different RNA-analysis techniques: cDNA array and reverse transcriptase-RT-PCR. Results of the two methods were in close agreement and were further shown to be replicable across multiple sample populations. This allowed the researchers to design a consistent suite of transcripts that could be used to diagnose AD. This multivariate analysis demonstrated impressive accuracy in a number of critical experiments described in the paper.
The test identifies people carrying 2 copies of the APOE4 gen, a severe risk factor for developing Alzheimer’s. The test also includes transcript screening to identify those at risk for future cognitive impairment due to having at least one direct relative with AD. Careful analysis of RNA transcripts in blood samples distinguished early clinical AD, Parkinson’s disease (PD), and cognitively healthy patients. Both cDNA and RT-PCR methods managed to distinguish probable AD from normal controls with an accuracy of 93.8% when analyzing only 5 RNA transcripts. The author’s note that the accuracy may be even higher as some of the samples labeled as “false positives” may be from subjects who are actually positive for pre-symptomatic manifestations of Alzheimer’s.
In addition to RNA transcripts linked with inflammation and stress, the study examines a series of epigenetic transcripts (RNA sequences that have undergone post-transcriptional modification). Results showed a strong correlation between the presence of these epigenetic markers and AD, implying they too may provide a diagnostic tool.
Future refinements should sharpen the method’s ability to accurately identify AD at an early stage prior to the onset of clinical symptoms, and in a primary care setting with a simple blood extraction. Early diagnostics should eventually be combined with testing of new therapeutics aimed at early intervention. Intriguingly, one or more of the many existing drugs for AD that have failed in clinical trials, may later succeed in slowing or arresting AD if they can be delivered early enough in the disease process. Further, trials for new drugs targeting at-risk patients can be ramped up significantly if a simple, non-invasive blood test can replace costly imaging like PET scan.
The study, by Delvaux E et al, was published June 19, 2017, in the journal Neurobiology of Aging.
Arizona State University