UCLA

Parkinson's disease (PD) is the second most prevalent progressive neurodegenerative disorder, characterized by the loss of dopaminergic neurons and a wide range of motor and non-motor symptoms. Among the non-motor symptoms, depression is one of the most prevalent and disabling, affecting approximately 30 - 50% of patients and contributing to accelerated cognitive decline, reduced quality of life, and increased caregiver burden. Despite its clinical importance, the biological mechanisms underlying depression in PD remain poorly understood, and it is unclear whether depression in this population shares etiological pathways with major depression in the general population or reflects distinct, disease-specific processes. This dissertation addresses that gap by applying epidemiological methods across multiple omics levels - genomic, epigenomic, and metabolomic – using data from population-based PD cohorts to characterize the genetic, immune-inflammatory, and psychosocial-metabolic contributors to depression and symptom progression in PD.The first aim investigated the genetic architecture of depression in PD using gene-based association analyses (SKAT, Burden test, and SKAT-O) applied to single nucleotide polymorphisms (SNPs) within 171 PD-related and 315 depression-related gene regions in European ancestry PD patients from three cohorts: the Parkinson’s in Denmark (PASIDA) study, the Parkinson’s Environment and Genes (PEG) study, and the Parkinson’s Progression Markers Initiative (PPMI) study. We identified KRTCAP2, an immune and N-glycosylation pathway gene, as the most consistently statistically significant result which survived false discovery rate (FDR) correction, where a higher burden of variants was associated with 20% increased odds of lifetime depression diagnosis. While PD polygenic risk score (PRS) showed no association with depression in PD, a higher depression PRS was significantly associated with increased odds of depression diagnosis, suggesting that general depression genetic liability remains relevant even in a PD context. Together, these findings suggest that depression in PD reflects a combination of shared depression genetic architecture and disease-specific neuroinflammatory pathways. The second aim built on this immune signal at the epigenetic level, examining DNA methylation–derived proxies of 75 immune cell and cytokine markers in 506 PEG PD patients of European and Hispanic ancestry. Using weighted gene co-expression network analysis (WGCNA), these proxies were clustered into eight co-methylation modules, whose eigengenes were tested against three depression outcomes: lifetime diagnosis, continuous Geriatric Depression Scale (GDS-15) scores, and a three-level GDS severity category. The turquoise module — anchored by inflammatory protein proxies including CRP, GDF15, and B2M — showed the most robust and consistent associations across all outcomes, while additional modules implicated pro-inflammatory cytokines (IL-6, OSM), chemokine networks (CXCL10, CXCL11), and adaptive immune markers (L-selectin) in outcome-specific patterns. Together, these findings converge with Aim 1 to suggest that immune dysregulation is a central and biologically coherent feature of depression in PD, detectable at both the genomic and epigenomic levels. The third aim extended the investigation to the psychosocial and metabolic dimensions of PD symptom progression, examining whether stressful life events (SLE) were associated with depressive symptoms, functional status, motor severity, and cognition among PEG PD patients followed across repeated visits. Using linear mixed-effects models, a higher burden of SLE — particularly events appraised as upsetting — was associated with worse depressive symptoms (GDS-15, PHQ-9) and poorer patient-reported function and quality of life (MDS-UPDRS patient questionnaire, SF-36), but not with clinician-rated motor severity (MDS-UPDRS-III) or cognition (MMSE), suggesting a dissociation between the affective–functional and objective motor–cognitive burden of PD. In parallel, a metabolome-wide association study using high-resolution serum metabolomics found no robust feature-level associations with SLE, though exploratory pathway analysis pointed to tryptophan (serotonin and kynurenine), tyrosine (catecholamine), and steroid metabolism — systems consistent with known stress biology. Together, these findings indicate that psychosocial stress tracks the affective and self-perceived functional dimensions of PD more closely than its neurodegenerative motor course. Collectively, this dissertation advances understanding of depression in PD as a multifactorial phenotype shaped by immune-related genetic variants, epigenetic inflammation signatures, and psychosocial stressors that may leave detectable metabolic traces. The convergence of findings across omics layers points toward neuroinflammatory pathways as a promising target for future mechanistic investigation and, ultimately, for the development of biologically informed strategies to identify and treat depression in this vulnerable population.