Protein aggregates such as Amyloid β plaques and tau tangles are thought to be causative of Alzheimer’s disease, the most common form of dementia. Each of these two protein can serve as a biomarker for the disease, meaning that detection of the protein can help to qualitatively (Do I have Alzheimer’s?) and quantitatively (How advanced is the disease already?) characterize patients. First of all, biomarkers are important to facilitate therapy. The idea is: The earlier a disease is diagnosed, the more damage can be prevented. Unfortunately, in the case of Alzheimer’s disease currently no drugs are available to cure the disease. However, several options are available for patients and their families to decrease the disease burden either pharmacologically, by psychotherapy or by adapting the living environment into a form that prolongs autonomy and decreases stress. Here biomarkers can significantly decrease the disease burden since an early adaption to the disease avoids distress, wrong diagnosis and might even add several years that are experienced as ‘positive’. Secondly, biomarkers are important for the development of effective anti-Alzheimer’s drugs. Patients for clinical trials need to be identified early, so that the effects of drug candidates on disease progression can be judged appropriately. Both Alzheimer’s-linked proteins Amyloid β and tau can be detected in patients as fluid and imaging biomarkers. Unfortunately, detecting both proteins in the cerebral fluid is not risk-free and a complicated procedure. Also imaging is not straightforward and can not be applied frequently in large patient groups.

Because of the high medical need, but current obstacles for efficient Alzheimer’s disease biomarkers, a recent publication by Oliver Preische, Matthias Jucker and many more scientists received a lot of attention. In this publication, they describe a protein that can be used as a blood-based biomarker for Alzheimer’s disease. Blood can of course be easily, safely and repeatedly drawn from large groups of patients. But what did the study find exactly? And how precise and accurate would this test be?

In this blog post, I will briefly describe some of the key methodologies, the principle of the test, some of the implications, and potential weaknesses. In case you are interested in the full detail of the original paper, please check here.

First of all, I should mention that the biochemical basis for the test itself is not completely new. Similar tests, using the same
neurofilament light (NfL) protein as biomarker, have been developed before, for example for Huntington’s disease or Parkinson’s disease. NfL is a protein which is found in neurons. Once these neurons become damaged, NfL ‘leaks out’ and can be identified in the blood of patients by an anti-body based test. Because NfL principally only indicates neuron damage, but is not specific for one type of disease, careful controls are necessary when using this tests in patients. In the present study, the researchers have used this existing test, but quite cleverly so and above all, they chose a good study group that allowed a good validation of the test.

So how did the researchers apply and validate the already known NfL-biomarker test in Alzheimer’s disease?

Patients with known Alzheimer’s disease mutations (but still without the manifested disease) were compared to healthy members of their own family. The figure below shows how this data looks like: Each red dot represents a human with an Alzheimer’s-causing mutation and each blue dot represents a control person without this mutation. In people without Alzheimer’s mutation you never know when and if the disease ever breaks out. That’s why its is very hard to use such people to test a new detection method. In the case of a well-studied mutation, it is quite clear when the disease will break out, and therefore, in the first step, the years up to the first onset of symptoms (= estimated years to symptom onset (EYO)) could be calculated.

The researchers determined the estimated years to symptom onset (EYO) for people carrying known genetic Alzheimer’s mutations (red dots) and compared this data to members of their family without the mutation (blue dots). Next they measured the amount of NfL protein in the blood of the two groups. The data shows that the closer a person approaches the measurable outbreak of the disease, the higher the NfL levels become. Importantly, their is a difference between both groups. People with Alzheimer’s mutation seem to have higher NfL values, especially in the early phase, when the disease has just started. Therefore, NfL could be potentially used a biomarker in Alzheimer’s disease.

In the second step, the actual biomarker test was carried out. The amount of the biomarker protein NfL was displayed as a function of the years until the onset of the disease, even though people are not yet sick (check the figure above).

There are however some open questions to this approach:

The theoretical assumptions concerning the duration until disease outbreak might not be very accurate in practice and overall epidemiological data might not be accurate for a particular person in the data cloud.
It is also important that there are significant differences between the groups only from -2 EYO onwards (strong increase of the red compared to the blue curve). This means: The biomarker levels become really different
only 2 years before the onset of the disease. Whether two years are actually enough to treat the disease well or prevent it, is currently unclear. This test is therefore not suited for the long-term prediction of disease outbreak. NfL levels are simply not high enough very early on.

What is certain, however, is that this test will be very important for clinical trials of potential drugs. Surprisingly, one of the major problems with drug trials is that there is often no real evidence to suggest that you can prevent a disease. How would you ever know if you have prevented a disease although it might not break out for a particular person anyway?

This is where this novel NfL Alzheimer’s test comes into play. Now that we know that it works in principle, it can be applied to the majority of studies with patients who do not carry a clear Alzheimer’s mutation or in which the disease breaks out later in life. Patients who test positive because they have another neurological disorder that also releases NfL proteins because nerve cells are dying (for example, Parkinson’s disease or multiple sclerosis) need to be adjusted for afterwards. Sadly, at the moment, there is no other option.

So in essence, the test described here is relatively sensitive (shortly before Alzheimer’s breaks out), but not very specific. A good test should be both. However, at the moment this is technically not feasible because we do not know any specific biomarkers yet. Despite this and by using a clever study design, the researchers have now at least identified NfL as a sensitive biomarker that might help to find a long thought-after curing or anti-progressive Alzheimer’s drug.


During my PhD we used multi-color live cell and fixed cell single molecule imaging to light up the unknown fate of mRNA molecules during the stress response, which spans from the bright transcription site into the dark corners of the cytosol where mRNAs can encounter P-bodies and Stress Granules.

You can learn more about this exciting work by reading the most recent publication of the Jeffrey Chao lab: