Selective vulnerability in Parkinson's disease: contributions from neuronal connectivity, lineage and functional identity

In Parkinson's disease (PD) patients, a specific set of neuronal populations are known to present Lewy pathology, cell death or both. This selective distribution of vulnerability is observed during disease progression and it has been hypothesised to result from both brain connectivity and shared cell-autonomous characteristics. Herein we simulate the contribution of the connectome alone to the spread of arbitrary aggregates using a computational model of temporal spread within an abstract representation of the mouse mesoscale connectome. Our simulations are compared with the spread of alpha-synuclein aggregation within in vivo mouse models. This comparison highlights that neuronal connectivity appears to be required for the presence of alpha-synuclein aggregation, however, it is insufficient to fully describe the pattern of distribution of pathology throughout the mouse brain. Therefore, we further investigate the role of the shared cell-autonomous characteristics by analysing the developmental lineages, mature transcriptomic profiles and metabolic capabilities of a subset of 28 synaptically connected neuronal populations that are vulnerable to PD, to varying degrees. We highlight a set of developmental transcription factors that are commonly required for specification of many vulnerable brainstem neurons, e.g. Ascl1 and Lmx1b, and transcription factors that are able to distinguish between differently vulnerable neuronal populations, e.g. Pitx3. Using genome-wide microarray data from six brains of healthy donors, we further investigate the existence of a shared transcriptomic profile between vulnerable neuronal populations. We highlight a list of genes whose expression levels seem to selectivity identify the set of vulnerable neuronal populations. Interestingly, many of these genes are involved in biological processes previously implicated in PD, some of which have also variants previously highlighted in genome-wide association studies of PD. Since the vulnerable set of neurons are known to have a high metabolic burden, we used a constraint-based modelling approach to generate metabolic models and further investigate their ability to generate mitochondrial ATP. Interestingly, most of the metabolic models of vulnerable brain structures seem to specifically require the same set of metabolic reactions in order to generate mitochondrial ATP. This thesis, therefore, highlights the importance of a multidisciplinary approach to understanding the pathogenesis of PD.