Abstract
Lewy body dementia (LBD), which encompasses Parkinson's disease dementia (PDD) and Dementia with Lewy bodies (DLB), currently lacks established biofluid markers reflective of its complex molecular pathophysiology. Multiplex proteomic tools, such as the recently released NUcleic acid Linked ImmunoSandwich Assay (NULISA(™)), can simultaneously assess a diverse array of neurodegenerative targets and address this critical biomarker gap. We used the NULISA CNS disease panel to analyze 852 baseline and longitudinal cerebrospinal fluid (CSF) samples from control, PD without cognitive impairment (PD-NCI), PD with cognitive impairment (PD-CI), DLB, and Alzheimer's disease (AD) subjects. A subset of control and DLB cases also had alpha-synuclein seed amplification assay (αSyn-SAA) results. Of the 124 panel markers analyzed, nearly half (n=53) were significantly altered in PD-CI and/or DLB compared to controls. Of these LBD-associated targets, 12 demonstrated divergent abundance trends in AD. Alpha-synuclein (SNCA) as measured by NULISA did not change across clinical LBD diagnoses but was among proteins significantly altered in subgroup comparisons of SAA+ and SAA- cases. Longitudinal analysis revealed five proteins (NEFL, NRGN, CCL26, CRH, PGF) with baseline changes indicative of subsequent cognitive progression. Machine learning identified a series of panels, each comprising only 10 markers, capable of discriminating clinical diagnoses and SAA positivity with high accuracy (AUC ≥ 0.89). Overall, our results highlight the utility of NULISA multiplex proteomic analysis in the identification of diagnostic and prognostic CSF biomarkers of LBD that reflect its diverse molecular pathophysiology.