Predictive test for Alzheimer's disease

26 October 2007

US researchers have developed a blood test that measures levels of proteins associated with cell signalling, indicating whether people have the condition, or are likely to develop Alzheimer’s disease (AD) in the next few years [Ray S et al. (2007) Nat Med. Oct 14 (Epub ahead of print); doi:10.1038/nm1653], published in the journal Nature Medicine.

In seeking molecular biomarkers of disease, they reasoned that the levels of cell signalling molecules in the blood plasma might be altered as part of the disease process. Microarray analysis of the expression levels of 120 proteins in blood from cases (patients with AD) and controls (without AD), identified nineteen proteins that showed significantly different expression profiles in the two groups. Mathematical analysis produced a predictive model based on the levels of eighteen of these proteins, which was tested using samples from people with symptoms of dementia ranging from mild cognitive impairment to severe AD, and found to be consistent with previous clinical diagnosis in 89% of cases. Importantly, the test was found to perform with 81% accuracy when used to analyse blood samples taken from a group of 47 patients 2-6 years previously, and compared with current clinical status.

The eighteen proteins on which testing was based were found to be involved in two independent regulatory networks, one involving tumor necrosis factor alpha and monocyte-colony stimulating factor, and the other involving epidermal growth factor; these pathways are involved in inflammation, haematopoiesis (blood cell production) and apoptosis (programmed cell death). The authors conclude that these protein expression levels, which they dub the ‘cellular communicome’ may be a valuable target for therapeutic intervention as well as diagnostic testing, and further suggest that similar gene expression signatures may exist for other diseases of the central nervous system (CNS).

Comment: According to the Alzheimer’s Society, over 700,000 people in the UK have dementia of which Alzheimer’s Disease is thought to account for around half. A proteomic blood test for AD would be beneficial because it is often difficult to distinguish the condition from other forms of dementia and currently a definitive diagnosis can only be achieved post-mortem. Dr Susanne Sorensen of the Alzheimer's Society told the BBC: "Early diagnosis is essential if we are ever to develop treatments that can change the course or halt the progression of dementia rather than just treat the symptoms" (see BBC news). Similar excitement also surrounds Pittsburgh compound B, a molecular diagnostic that binds to amyloid plaques present in the brains of AD patients and is detectable using a PET scanner. However, the specificity and clinical validity of this compound is currently unclear.

Using a test to predict which individuals are likely to develop AD in coming years could be controversial; although it would probably be used only in patients already affected by mild forms of dementia and could therefore help to direct the most appropriate form of early treatment, current UK policy only permits the use of drug treatment in patients where the disease has reached moderately severe levels. In particular, the widespread implementation of a diagnostic test may be hindered by the current lack of treatments to cure AD or significantly halt progression in the early stages of the disease.

Further analysis with much larger samples will be required to verify how accurate the new test really is, but it nevertheless presents a promising area of research for both diagnostics and therapeutics for AD.

More from us