Tissue-Specific Epigenetic Drift in Parkinson’s Disease: A Multi-Region DNA Methylation Meta-Analysis

Abstract

Introduction
Parkinson’s disease (PD) is a neurodegenerative disorder whose genetics explains only part of the disease risk, suggesting the involvement of non-genetic factors. Epigenetic-drift, defined as the accumulation of stochastic epigenetic mutations (SEMs) in DNA, has been implicated in other neurodegenerative diseases but remains poorly characterized in PD.

Objectives
To investigate the role of global and gene-specific epigenetic-drift in PD and assess its relevance across different tissues and brain regions affected by the disease.

Methods

A meta-analytic study design was adopted by integrating public DNA methylation datasets. Eligibility criteria for dataset inclusion required the use of Illumina 450K or EPIC arrays, phenotypic metadata, clinical or neuropathological diagnosis of PD, at least 15 subjects per study group and cross-sectional study design. Applying these criteria yielded a final cohort comprising 3,425 peripheral blood and 378 brain tissue samples from case-control designs, alongside a multi-region dataset containing intra-subject repeated measures across caudate nucleus, prefrontal cortex and substantia nigra from PD individuals (N=97). SEMs were identified as extreme outlier using Tukey’s rule, defining reference ranges from control beta values at each CpG site or through a leave-one-out resampling procedure for the multi-region dataset. Differences in global SEM burden between cases and controls were evaluated using regression models and were meta-analyzed via a random-effects meta-analysis. Differences in methylation deviation distributions between cases and controls were evaluated using regression models with subject as a random effect. The multi-region dataset was analyzed using mixed-effects models, accounting intra-subject correlation, followed by post-hoc comparisons to identify region-specific increases in epigenetic drift. Gene-specific associations were explored using kernel-based methods (SKAT), treating SEMs as rare variants within an aggregation framework. All models were adjusted for standard EWAS covariates and validated via sensitivity analyses.

Results
Meta-analysis revealed a significant increase in global SEM burden in the frontal lobe of PD patients compared with HCs (β = 0.6; p = 0.03), whereas no significant differences were observed in peripheral blood or the parietal cortex. In the frontal lobe, PD patients exhibited an enrichment of SEMs characterized by smaller methylation deviations, while HCs showed a relative enrichment of high-delta SEMs (p < 0.01). Gene-based analysis in blood identified genes differentially enriched for SEMs (Bonferroni-corrected p-value < 0.01), involved in axon guidance pathway. In the multi-region dataset, SEM burden increased progressively from the caudate nucleus to the prefrontal cortex, reaching its highest level in the substantia nigra, where it was significantly greater than in the other regions.

Conclusions
Our findings indicate that PD is associated with a tissue-specific increase in epigenetic-drift, particularly in brain regions most affected by neurodegeneration. The enrichment of low-delta SEMs suggests increased local epigenetic heterogeneity, consistent with the accumulation of SEMs that are poorly shared among cells. These findings support a potential involvement of epigenetic-drift in PD pathophysiology, although they do not allow us to determine whether this phenomenon contributes to disease development or arises as a consequence of neurodegenerative processes. In particular, increased cell epigenetic heterogeneity may either enhance the vulnerability of brain tissue to neurodegeneration or represent a downstream consequence of disease-associated pathological processes.