– HISTOPARK

Early Parkinson's disease (PD) is characterized by propagation of alpha-synuclein inclusion and neuronal loss in many brain regions including small brainstem nuclei, such as substantia nigra and locus coeruleus, and hypothalamic nuclei regulating sleep, later leading to dysfunction of the striatum and dysregulation of basal ganglia circuits. However, the precise pattern and spread of PD pathology during the preclinical stages, along with its association with the progression of early symptoms, remain largely elusive. This knowledge gap is due to the complex and poorly understood anatomy of the small nuclei responsible for early symptoms and the complex interactions between them. In addition, the low resolution and low specificity of non-invasive neuroimaging techniques make it difficult to detect and monitor neurodegeneration of these small structures in early preclinical stages, as well as to capture subtle atrophy-related changes throughout the brain. Furthermore, the link between neurodegenerative alternation at the microscopic scale and neuroimaging biomarkers is poorly understood.
To address this gap, we recently integrated ultra-high-resolution multiparametric quantitative neuroimaging at 7 Tesla (7T) with an automated brainstem segmentation technique informed by a new multimodal histological atlas of small nuclei anatomy, including the locus coeruleus and the nigrosomes of the substantia nigra. The new level of precision of our approach led to a reinterpretation of nigral radiological markers in the context of PD. This neuroimaging protocol providing longitudinal quantitative measures of dopaminergic cell density and iron load, subcortical myelination, tissue water and iron content, along with sleep measurements and genetic characterization, was used in a multi-center study of both healthy participants and individuals at risk for PD to characterize the progression of early-stage PD with this unprecedented level of precision. The resulting data set will provide precise characterizations of brain anatomy at sub-millimeter resolution and quantitative MRI parameters across the entire brain and particularly over the subcortex, providing insight into microstructural changes due to the disease progression that would go undetected by typical clinical neuroimaging.
This proposal aims to extend and disseminate the methodology developed. We will (i) apply machine learning and mechanistic modeling to analyze the longitudinal ultra-high resolution MRI data and to create a generative framework of disease progression and heterogeneity, accounting for interaction between biological factors across subcortical and cortical structures (e.g. disease spreading), and establishing the link between the symptoms and neurodegeneration in the early stage. We will (ii) integrate quantitative MRI-based biomarkers of nigrosome, locus coeruleus, and hypothalamic nuclei integrity, as well as whole-brain myelin and iron measurements to capture atrophy; (iii) extend the dataset by including a larger cohort of patients and individuals at risk of developing PD; (iv) enhance the dataset by including an additional time point to follow neurodegeneration at an early stage; (v) enrich the subcortical atlas by quantitative 3D histology of small hypothalamic nuclei responsible for both motor and non-motor PD symptoms and biophysical model linking between tissue histology and in vivo quantitative MRI; (vi) translate our existing imaging protocol for practical use in clinical settings by using advanced imaging acceleration routines. The project will deepen our understanding of the dynamics of PD in its early stages, paving the way for more effective diagnostic and intervention strategies.