The pathogenesis and clinical heterogeneity of Parkinson’s disease have been evaluated from molecular, pathophysiological, and clinical perspectives. High-throughput proteomic analysis of CSF has opened new opportunities for scrutinizing this heterogeneity. To date, this is the most comprehensive CSF-based proteomics profiling study in Parkinson’s disease (1103 patients, 4135 proteins). Combining CSF aptamer-based proteomics with genetics we determined protein quantitative trait loci (pQTLs). Analyses of pQTLs together with summary statistics from the largest Parkinson’s disease genome wide association study (GWAS) identified 68 potential causal proteins by Mendelian randomization. The top causal protein, GPNMB was previously reported to be upregulated in the substantia nigra of Parkinson’s disease patients.
We also compared the CSF proteomes of patients and controls. The Parkinson’s disease cohort comprised not only LRRK2+ and GBA+ mutation carriers but also idiopathic patients. Proteome differences between GBA+ patients and unaffected GBA+ controls suggest degeneration of dopaminergic neurons, altered dopamine metabolism and increased brain inflammation. The proteins discriminating LRRK2+ patients from unaffected LRRK2+ controls, revealed dysregulated lysosomal degradation, as well as altered alpha-synuclein processing, and neurotransmission. Proteome differences between idiopathic patients and controls suggest increased neuroinflammation, mitochondrial dysfunction / oxidative stress, altered iron metabolism and potential neuroprotection mediated by vasoactive substances.
Finally, we used proteomic data to stratify idiopathic patients into "endotypes". The identified endotypes show differences in cognitive and motor disease progression based on the use of previously reported protein-based risk scores.
In summary, we: i) identified causal proteins for Parkinson’s disease, ii) assessed CSF proteome differences in Parkinson’s disease patients of genetic and idiopathic etiology, and. iii) stratified idiopathic patients into robust clinically relevant subtypes. Our findings not only contribute to the identification of new therapeutic targets but also to shaping personalized medicine in CNS neurodegeneration.