The François Jacob Institute of Biology brings together five departments and three services
The last two years in scientific news
Group leader: Marc Dhenain (DR1CNRS, French Veterinary Academy, National Academy of Medicine)
Our Team is dedicated to i. the characterization of biological mechanisms involved in the development of Alzheimer's disease and its therapies. ii. the development of innovative imaging tools to follow-up cerebral pathologies. We focus on 3D microscopic imaging methods based on high performing computing (HPC) algorithms, and on resting state fMRI to evaluate neuronal networks.
Our group has demonstrated that the key lesions of Alzheimer's disease (β-amyloid and tau lesions) are transmissible (Gary, 2019). This transmission makes it possible to explore the pathophysiogenic mechanisms of Alzheimer's disease and to test new therapies. Using novel models based on AD lesion transmission, we have shown that synapse loss is linked to tau pathology and that activation of inflammatory brain cells called microglia limits synaptic loss (Lam, 2021). We are now seeking to understand the mechanisms underlying the heterogeneity of Alzheimer's disease through an integrative and holistic approach that considers all aspects of the disease (amyloid / tau / inflammation / functional impairment).
In particular, we focus on the role of various nucleating factors that regulate AD pathophysiology. Either nucleating factors extracted from human brain samples (Collaboration with Stephane Haik and Susana Boluda from ICM, Paris) or nucleating factors based on very purified forms of amyloid with very specific mutations (collaboration with Alain Buisson from Grenoble Institute of Neurosciences).
In addition to studies in mice, one of our strength is to be able to work in primates and in particular with mouse lemur primates. The latter is a model of neurodegenerative pathologies linked to aging. This animal presents, as it ages, cognitive alterations, alterations in cerebral metabolism, cerebral atrophy and amyloid deposits.
Amyloid and tau lesions induced in the brain of
experimental models following inoculation of amyloid and tau nucleating
Neurodegenerative diseases are related to many different "small scale events" (pathological protein accumulation, neuroinflammation, cellular alterations) that lead to large-scale events (tissue loss, neuronal networks alterations, cognitive impairments). Our team develop tools to integrate events occurring at different scales. These new tools require advanced imaging skills combined with big data management and high performance computing.
Our group implemented several image-processing pipelines to perform 3D post mortem reconstruction program in primates and rodents, while focusing on the exploitation of 3D information in rodents. 3D-reconstructed brain samples can be analyzed using semi-automatic manual analysis, digital atlas-based analysis (Lebenberg, 2011) or voxel-wise SPM approach without a priori (Vandenberghe, 2018). The method can be used to detect lesions as amyloid plaque related to Alzheimer (Vandenberghe, 2018).
New methods were then implemented to quantify cells at the level of the whole brain. They are based on virtual microscopy performed after digitization of brain sections stained for cells (NeuN antibody) and on the segmentation of each cell (Random Forest -RF- machine learning algorithm) and individualization methods (You, 2019 & 2021). Performing this analysis at the level of the whole brain allows to create parametric maps synthesizing for example the morphology and distribution of individualized neurons. Then, it is possible to synthesize this information in the form of lower-resolution parametric maps at the level of anatomical regions, sections and even, eventually, the entire brain. This step converts qualitative color microscopic images to quantitative mesoscopic images, more informative and easier to analyse, to statistically assess neuronal death or relationships between neuronal densities and brain function assessed from in vivo images. New research to develop digital twins of brains has been initiated recently to exploit information from 3D histology to produce realistic mathematical models of cytoarchitecture. These approaches will open perspectives to validate new imaging techniques (MRI) and to better understand and decipher the mechanisms at the cellular level. High performance computing (HPC) resources are integrated in our research projects to deal with massive data analysis and increasing algorithm complexity. This requires strong methodological development, performed though a collaboration with the TGCC of the CEA (Bruyères-le-Châtel). This structure hosts a supercomputer that is one of the ten most powerful machines in the world.
These tools are mainly developed using in-house software platform BrainVISA (http://brainvisa.info).
Methods used to
evaluate neuronal density and other neuronal parameters based on high
performance computing. Brain sections stained for cells (NeuN antibody)
are segmented and various parameters reflecting neuronal characteristics (e.g.
their density, size, orientation, etc…) are calculated. Based on this analysis,
we can produce parametric maps reflecting neuronal states at the level of the
whole brain. These maps can then be compared with other maps (lesion, brain
2) Quantification of neuronal networks
Individual cells function in a harmonized way that leads to harmonious brain activity through functional networks. These networks can be assessed by resting state functional MRI and sophisticated image processing tools. Our group studies brain activity with resting state functional imaging.We participated in the development of the reference article, which, through a multicentre evaluation, defines the optimal conditions for carrying out network imaging in rodents (Grandjean, 2020).We have developed software for analysing neural networks (Celestine, 2020; https://sammba-mri.github.io/).We compare neural networks in primates and humans and seek to define how evolution has modulated these networks (Garin, 2021).
Example of detection of neural networks in
humans and in the smallest primate in the world (mouse lemur) by magnetic
Evaluation of automated segmentation algorithms for neurons in macaque cerebral microscopic images.You, Z., M. Jiang, Z. Shi, X. Ning, C. Shi, S. Du, A. S. Hérard, C. Jan, N. Souedet, Delzescaux T.Microscopy Research and Technique. 2021: 27 Apr 2021: https://doi.org/2010.1002/jemt.23786.
CEA is a French government-funded technological research organisation in four main areas: low-carbon energies, defense and security, information technologies and health technologies. A prominent player in the European Research Area, it is involved in setting up collaborative projects with many partners around the world.