Longitudinal blood-based proteomics reveal progression-specific molecular signatures in Parkinson’s disease

Parkinson disease (PD) shows marked heterogeneity in clinical progression and single baseline biomarkers have shown limited value for predicting these changes. We examined whether repeated serum proteomics can identify protein-clinical associations linked to worsening across motor, cognitive, non-motor, and treatment-related outcomes. We analysed 166 patients with PD from the prospective PROMOVE registry with repeated OFF-state clinical assessments and longitudinal serum proteomics covering 297 proteins. For each outcome, participants were classified as worsening or non-worsening based on individual longitudinal slopes. In worsening participants, all proteins were tested using linear mixed-effects models that separated within- from between-subject protein effects. Six associations remained significant after FDR correction. Higher DDC was associated with increasing levodopa equivalent daily dose. Higher CD276 and POSTN were associated with worse MDS-UPDRS II scores, while higher VEGF-C and IL19 were associated with worse MDS-UPDRS III scores. Higher IL12 was associated with lower MMSE scores. The five non-treatment associations remained significant after additional adjustment for time-varying dopaminergic treatment. Repeated serum proteomics identified distinct protein-clinical associations in patients showing clinical worsening. The findings point to treatment-related, immune, vascular, and extracellular-matrix processes and suggest that longitudinal proteomics may complement clinical measures when monitoring PD progression.