Large-scale CSF and plasma proteomics across AD, PD, DLB, and FTD (n≈2.7–3.0k) reveals shared immune dysregulation and disease-specific pathway changes, with PD showing ATF4/PERK (ER stress) signaling alterations and high-performing CSF/plasma diagnostic models.

Identifies ER stress (ATF4/PERK) and immune pathways as actionable mechanisms in PD and delivers robust biomarker panels for patient stratification, supporting target prioritization and translational therapeutic efforts.

Neurodegenerative diseases (NDs), including Alzheimer's disease (AD), Parkinson's disease (PD), dementia with Lewy bodies (DLB), and frontotemporal dementia (FTD), share overlapping clinical and pathological features. We analyzed cerebrospinal fluid (CSF) and plasma proteomes from 2,705 and 3,009 samples, respectively, across these NDs, identifying disease-specific and shared molecular signatures. CSF showed more disease-associated proteins than plasma, with AD and DLB exhibiting the strongest cross-tissue similarity. Pathway analyses revealed shared dysregulation of immune-related processes in CSF and plasma across the NDs, as well as disease-specific impairment of glycosylation and apoptotic pathways in AD; ATF4 and PERK signaling in PD; fibroblast growth factor receptor (FGFR) and interleukin signaling in DLB; and glycoprotein hormones disruption in FTD. We developed disease-specific predictive models showing high accuracy (area under the curve [AUC]: 0.81-0.95 in CSF and 0.80-0.89 in plasma). These findings reveal distinct and convergent mechanisms across NDs, highlighting potential biomarkers and pathways for diagnostic and therapeutic strategies in neurodegeneration.