Frontotemporal degeneration and motor neurone disease: biomarkers for studying the presymptomatic phase

Symptoms appearing between the age of 50 and 70

Frontotemporal degeneration (FTD) and motor neurone disease (or amyotrophic lateral sclerosis - ALS) are terminal neurodegenerative diseases for which there is currently no cure. By the time symptoms first appear - ordinarily between the age of 50 and 60 for FTD and between 50 and 70 for MND - it is likely to be too late for doctors to be able to do anything. 

Prior to this, the diseases develop in silence over a period of decades, with no indication as to when symptoms will appear. This is true even for patients in whom a mutation in the c9orf72 gene has been clearly identified, who are almost certain to develop the disease. 

Biomarkers for each phase of the disease

These are the patients that our study focused on, explains Olivier Colliot, director of the Aramis project team at the Inria Paris research centre. Our aim was to find biomarkers that would enable us to distinguish presymptomatic patients (those with no symptoms) from control patients and confirmed cases. Not only were we successful in doing so, but we did it with markers that could be obtained via a simple blood test.

Symptoms appearing between the age of 50 and 70

Frontotemporal degeneration (FTD) and motor neurone disease (or amyotrophic lateral sclerosis - ALS) are terminal neurodegenerative diseases for which there is currently no cure. By the time symptoms first appear - ordinarily between the age of 50 and 60 for FTD and between 50 and 70 for MND - it is likely to be too late for doctors to be able to do anything. 

Prior to this, the diseases develop in silence over a period of decades, with no indication as to when symptoms will appear. This is true even for patients in whom a mutation in the c9orf72 gene has been clearly identified, who are almost certain to develop the disease. 

Biomarkers for each phase of the disease

These are the patients that our study focused on, explains Olivier Colliot, director of the Aramis project team at the Inria Paris research centre. Our aim was to find biomarkers that would enable us to distinguish presymptomatic patients (those with no symptoms) from control patients and confirmed cases. Not only were we successful in doing so, but we did it with markers that could be obtained via a simple blood test.

Four microRNAs selected out of 589 possible candidates using artificial intelligence

The two Inria teams (Dyliss in Rennes and Aramis in Paris) worked to develop a machine learning model - an artificial intelligence tool - as part of Virgilio Kmetzsch's PhD, which they jointly supervised. Out of a possible 589 microRNAs, the machine learning model identified those four felt to be the most suitable, an enormous sorting operation which underlined the benefits of having biologists and computer scientists working together.

We ran into a few stumbling blocks when designing the model, explains Virgilio Kmetzsch. We had never used data on microRNAs before, as this is a very recent field of research. Out of the research cohort, there were only 67 people who had the c9orf72 gene mutation:  frontotemporal degeneration and motor neurone disease are rare conditions.

© Inria / Photo C. Morel

 

The goal: to treat patients before the first symptoms emerge

Despite the difficulties, the researchers achieved what they set out to do. Their model is capable of determining if a subject is at the presymptomatic stage or if they already have symptoms, based on the level of expression of the four microRNAs.

This is the first time that a microRNA signature for the presymptomatic and symptomatic stages has ever been identified, explains Olivier Colliot. Our new tool will enable us to develop a deeper understanding of the mechanisms involved in both frontotemporal degeneration and motor neurone disease, in addition to supporting research into treatments capable of being administered prior to the first symptoms appearing.

 

Biomarkers for determining the disease's “clock”

However, we should strike a note of caution when discussing this new development: the patients in the study have only been monitored for three years, which is a very short period of time for these diseases. That said, knowledge and understanding of microRNAs could well improve over time. 

As we track patients for another few years, we hope to be able to find markers that evolve with the disease, explains Olivier Colliot, a sort of clock that would indicate the optimal time to begin treatment.

 

© Pixabay

The next step: adding microRNAs from imaging

Previous research carried out by the Paris Brain Institute had revealed other biomarkers through the use of brain imaging. In subjects carrying the c9orf72 mutation, a degeneration of certain regions of the brain was observed. These lesions, which could be seen from 40 years of age onwards, shed additional light on the presymptomatic stage of both frontotemporal degeneration and motor neurone disease.

How can these medical images be combined with the expression levels of four microRNAs? Once again, the answer lies in biologists and computer scientists working together.

Our researchers in Paris and Rennes have been working with the Paris Brain Institute on the integration of multimodal data, explains Olivier Colliot. The machine learning model will evolve in order to factor in imaging, which will enable us to deepen our understanding of the different stages of the two diseases.

Inria and the Paris Brain Institute - eight years of working together

Inria and the Paris Brain Institute have been working together since 2012. It was then that Aramis, a joint project team, was set up to study neurodegenerative diseases through a combination of neuroscience and computer science, with a particular focus on medical imaging. In 2017, Inria and the Paris Brain Institute expanded into genetic and genome-related aspects, as part of the Project Lab Neuromarkers. Combining different methods gives researchers more data to work with - the difficulty then lies in using them simultaneously, with new methods and algorithms.