Prioritizing MMP8 as an exploratory therapeutic candidate in Parkinson's disease through proteome-wide Mendelian randomization and computational multi-omics integration.
TL;DR
The 32 proteins with nominal MR associations represent candidates for follow-up investigation and should be interpreted as exploratory because a verified proteome-wide FDR correction could not be established.
Abstract
Background
Parkinson's disease (PD) lacks disease-modifying therapies. We aimed to prioritize circulating proteins using multi-omics integration combining proteome-wide Mendelian randomization (MR) for discovery with network, structural, and single-cell evidence for prioritization.
Methods
We performed proteome-wide Mendelian randomization using deCODE plasma protein data (4907 proteins) and FinnGen PD genome-wide association data (4681 cases, 407,500 controls). Instrument strength and causal direction were assessed; candidate selection used nominal IVW P < 0.05 with multi-criteria prioritization because a verified proteome-wide FDR correction could not be established from the archived summary statistics. Prioritized candidates were further evaluated by network analysis, drug repurposing, molecular docking, 100-ns molecular dynamics simulations, and single-cell transcriptomics.
Results
Proteome-wide MR identified 32 proteins with nominal associations with PD risk at IVW P < 0.05 (12 protective and 20 risk-increasing). MMP8 showed a nominal risk-increasing association (OR = 1.166, 95% CI 1.027-1.322, P = 0.018) and was prioritized as an exploratory candidate. Drug repurposing screening and 100-ns molecular dynamics simulations indicated that medroxyprogesterone acetate showed greater binding persistence with MMP8 than aldosterone in single exploratory trajectories; all structural findings are exploratory in silico predictions. In a non-disease mouse whole-brain single-cell atlas, MMP8 expression was enriched in myeloid-lineage cells, providing cell-type context requiring validation in human PD single-nucleus data; these findings are discovery-oriented and hypothesis-generating.
Conclusions
The 32 proteins with nominal MR associations represent candidates for follow-up investigation and should be interpreted as exploratory because a verified proteome-wide FDR correction could not be established. As a nominal exploratory candidate, MMP8 is an inflammation-related molecule with supportive network and structural evidence, warranting experimental validation. Medroxyprogesterone acetate showed greater binding persistence in exploratory single-trajectory simulations and emerged as a candidate for follow-up evaluation based on suggestive in silico evidence. Our multi-omics framework provides a reusable resource for therapeutic target discovery in neurodegenerative diseases.