By Professor Simon Lovestone
Professor of Old Age Psychiatry, NIHR Biomedical Research Centre for Mental Health, King's College London, Institute of Psychiatry, De Crespigny Park, London SE5 8AF
Suppose you have memory problems and go to your doctor - what happens next? At best, and the best is not always achieved, then an assessment of memory and other cognitive function is made and, in some cases, a referral is made to a memory clinic where there may be more memory tests and perhaps a brain scan. If the outcomes of these tests are not severe enough to warrant a clear-cut diagnosis then, usually, the cognitive tests are repeated after a year to see if they have got worse. This is unsatisfactory as the wait must seem interminable to patients and their relatives. For professionals too, not being able to make an early diagnosis is frustrating.
When treatments for Alzheimer's disease (AD) go beyond symptomatic treatment to therapies for the illness itself, this wait will be unacceptable since it is in this early phase, before dementia is established, that the drugs are most likely to be effective. This, then, is one of the most important drivers for research into 'biomarkers' of AD. A biomarker is a biological signal that can be detected for diagnosis, ideally very early and before doctors are currently able to diagnose the condition. Biomarkers are also useful as the basis of tests for measuring how a disease is progressing. This latter use of biomarkers would be especially useful in research to find new treatments for dementia. Currently clinical trials rely almost entirely on memory tests which are sometimes less reliable than we would like as measures of disease. A biomarker that reflected the disease progression in the brain would be immensely useful as a measure against which to judge new treatments.
Spinal taps and brain scansToday, we do not have any biomarkers widely accepted either as diagnostic markers or as progression markers. But there is a huge effort underway to discover markers either in spinal fluid, using brain scans or in blood. Spinal fluid is perhaps the best place to start and nearly ten years ago researchers found that two proteins - amyloid and tau - are changed in the spinal fluid of people with Alzheimer's. Amyloid is the protein that forms plaques in the brain in people with Alzheimer's disease and tau is the protein that forms tangles. Interestingly in the spinal fluid amyloid levels are reduced in Alzheimer's disease and tau levels increase. It is thought that this is because less amyloid reaches the spinal fluid as it gets bound up in the plaques, and more tau reaches the fluid because the nerve cells are dying and releasing their tau protein. Whatever the explanation, the finding has been very widely reproduced and a test has been produced that clinicians can use to measure amyloid and tau in spinal fluid. This test looks very promising as an early biomarker of dementia. It is now being incorporated into clinical trials of new drugs to prevent amyloid build-up and tangle formation and some evidence is beginning to suggest that it is proving to be a useful marker in these drug trials. This is all hugely encouraging and an international collaborative group is now working to ensure that the test is used in the same way, and gives the same results, in laboratories across the world.
However, collection of spinal fluid has disadvantages due to patient discomfort and some minor side-effects. An alternative approach is brain scans. These are used today mostly to exclude other brain disease that might cause memory problems. However, new scans and techniques for analysing data have been introduced. The new analytical techniques use computer algorithms to measure shrinkage in the brain in different regions. This is a hugely complex task and in the past could only be done manually. It is now possible to do it in an automated fashion and such analysis, pioneered by Professor Nick Fox and others in London, is now being used to measure the effects of novel treatments in clinical trials. At the Maudsley Hospital in London we have just begun to use these automated analyses of brain scans in our routine clinical work.
Other approaches to biomarkers are using imaging technologies to directly measure the amount of amyloid plaques in the brain in living patients. This is known as amyloid-PET (positron emission tomography) and involves the injection into the bloodstream of a chemical that is radioactive for a very short time and seeks out and binds to amyloid deposits. Other kinds of PET imaging have revolutionised the measurement of some cancers and are routinely used in diagnosis and in evaluating the success of treatments. It is very exciting seeing this technology applied to the brain in Alzheimer's disease. However, the limitation is that PET scanners are few and far between and the short-lived radioactive isotopes currently used are so short-lived that the chemists need to make them no more than half an hour or so before they are used. Research is underway to find other, slightly longer-lived probes that would make amyloid-PET imaging more widely available.
A blood biomarker for Alzheimer'sOur own research has focused on trying to find a biomarker for Alzheimer's disease in blood. When we started this work, nearly ten years ago, it seemed unlikely that there would be such a marker; after all, the brain is separated from the blood by a nearly impenetrable membrane. Despite this, because most biomarkers in medicine are blood-based and because getting blood samples analysed is so routine in healthcare, we persevered and were funded to do so by both Alzheimer's Society and the Alzheimer's Research Trust. We employed proteomics to look for protein differences in the blood of people with Alzheimer's and elderly controls. Somewhat to our surprise we did find some differences and went on to confirm these in quite a large group of people. We published this finding and then an Alzheimer's Society fellow, Dr Madhav Thambisetty, showed that the changes we see in the blood in proteins called complement factor H (CFH) and alpha-2-macroglobulin (A2M) matched changes in the function of the brain in people not only with Alzheimer's but also those with pre-Alzheimer states such as mild cognitive impairment. This is very encouraging as it suggests that these proteins act as very early markers of Alzheimer's disease pathology.
Encouraged by this finding we have concentrated all our efforts in the last year or two on finding more very early markers of Alzheimer's pathology. We have found one protein that is consistently increased in blood very early in the disease process and is increased especially in people who go on to decline quickly. We have been able to study mice with Alzheimer's pathology and find this protein is increased in the blood here also and, in collaboration with Madhav, who is now working at the National Institutes of Ageing in the US, we have shown our protein is increased in people with large amounts of amyloid as shown by PET. This takes us ever closer to a blood-based biomarker for Alzheimer's. Our preliminary findings suggest we can be more than 90 per cent accurate in predicting which patients with mild memory problems go on to get Alzheimer's in the next year, based on a blood test and a brain scan.
The future of biomarker researchThese findings are very encouraging indeed. They suggest that in the future the person worried about their memory who goes to the doctor will not face the prospect of a year's wait to see if it gets worse before a diagnosis is made, but will instead be able to have a blood test, a spinal fluid test or a brain scan (or perhaps all three) which will help the doctor come to a much faster diagnosis very early in the course of the illness. However, we are not quite there yet and more research needs to be done. Specifically, we need larger, longer term studies involving more people donating blood and spinal fluid samples. We need to discover ever better and more accurate biomarkers and in particular we need to work harder to convert scientific discoveries into useable tests. This has been something of a gap in research until recently but large funding agencies, including the Medical Research Council and the National Institute of Health Research in the UK, have recognised this and have put more effort and funding into this 'translation gap'. Not just funding but volunteers are essential for this research and we all have reason to be grateful to the many who have contributed to our cohort studies over the years and agree, year after year, to give up their time and come and have another assessment and another blood test.
There is one particular kind of cohort study that is hugely valuable in the search for biomarkers and that is the so-called prospective study where volunteers are assessed over time in life and then donate their brain after death. Through the Brains for Dementia Research programme we have been able to show that one of our biomarkers (lets call it biomarker A) is specifically associated with plaques in the brain and that the amount of this protein in blood in life matches the amount of the protein in the brain after death. In another study we find a different protein in the blood (biomarker B), although clearly associated with plaques in the brain, does not show the same blood-protein matching. These are rare and special studies. They allow us not only to find biomarkers but to begin the difficult process of asking why they are biomarkers. We are able to conclude that biomarker A is acting in the brain but biomarker B is acting in the blood. As both affect amyloid we think biomarker A stops plaques from forming but biomarker B draws amyloid out of the brain and into blood where it can do no harm. Armed with this knowledge we can begin to look for other proteins that do the same and then even plan studies designed to find new treatments based on this knowledge.
There are very few studies anywhere in the world that are able to combine approaches in this way. By establishing cohorts of people assessed in research in life and then donating their brains to research after death, Brains for Dementia Research looks well set to make a huge contribution to the search for biomarkers for better and earlier diagnosis as well as contributing to our understanding, and hopefully one day, to treatment of Alzheimer's disease.
ReferencesHuang, H.C., Jiang, Z.F. (2009). Accumulated amyloid-beta peptide and hyperphosphorylated tau protein: relationship and links in Alzheimer's disease. J Alzheimers Dis, 16(1), pp.15-27.
Mattsson, N., Blennow, K., Zetterberg, H. (2009). CSF biomarkers: pinpointing Alzheimer pathogenesis. Ann N Y Acad Sci, 1180, pp.28-35.
Barnes, J., Ourselin, S., Fox, N.C. (2009). Clinical application of measurement of hippocampal atrophy in degenerative dementias. Hippocampus, 19(6), pp.510-6.
Lovestone, S. et al. (2007). Proteomics of Alzheimer's disease: understanding mechanisms and seeking biomarkers. Expert Rev Proteomics, 4(2), pp.227-38.
Lovestone, S., Francis, P., Strandgaard, K. (2007) Biomarkers for disease modification trials--the innovative medicines initiative and AddNeuroMed. J Nutr Health Aging, Jul-Aug;11(4), pp.359-61.
Thambisetty, M. et al. (2008). Proteome-based identification of plasma proteins associated with hippocampal metabolism in early Alzheimer's disease. J Neurol, 255(11), pp.1712-20.
Hye, A. et al. (2006). Proteome-based plasma biomarkers for Alzheimer's disease. Brain, 129(11), pp.3042-50.