Parkinson disease (PD), one of most common neurodegenerative disorders in the United States and worldwide, continues to affect the elderly population. Medication and surgical intervention can alleviate the motor symptoms of PD and maintain some quality of life for the patient; however, progression is inevitable, and the etiology and a definitive cure of PD remain elusive. The current understanding is that PD is a multi-factorial disease attributed to multiple risk factors including genetic and environmental exposures. Our group has long demonstrated connections between ambient pesticides exposure and its interactions with genetic susceptibility and PD risk. Recent developments in omics research opened up exciting avenues for researchers to investigate metabolic and gut bacterial changes in PD patients and provide novel insights into the PD pathogenesis and opportunities for treatment. In this dissertation, we connect ambient exposure to organophosphorus pesticides (OPs), the gut microbiome, and the serum metabolomics in PD patients and community controls. All study participants are residing in Freso, Tulare and Kern counties of California were heavy agricultural pesticides use is common. Our aims are to gain a deeper understanding about the interplay between pesticide exposure, gut bacteria, and human metabolism in PD.First, we investigated the differences in gut bacterial composition and predicted metabolic function in PD patients and controls. For 96 PD patients and 74 controls, microbiome data were obtained with 16S rRNA gene sequencing of fecal samples. These were then analyzed for microbial diversity, taxa abundance, and predicted functional pathways in association with PD. We also examined the bacterial composition and predicted function associated with PD-specific features (disease duration, motor subtypes, L-Dopa daily dose, and motor function). We found that PD patients’ gut microbiome showed lower species diversity (p = 0.04) and were compositionally different (p = 0.002) compared to controls but had a higher abundance in three phyla (Proteobacteria, Verrucomicrobiota, Actinobacteria) and five genera (Akkermansia, Enterococcus, Hungatella, and two Ruminococcaceae) controlling for sex, race, age, and sequencing platform. Also, 35 Metacyc pathways were predicted to be differentially expressed in PD patients including biosynthesis, compound degradation/utilization/assimilation, generation of metabolites and energy, and glycan pathways. Additionally, the postural instability gait dysfunction subtype was associated with three phyla and the NAD biosynthesis pathway. PD duration was associated with the Synergistota phylum, six genera, and the aromatic compound degradation pathways. Two genera were associated with motor function.
We then focused on the perturbations of gut microbiota in response to long term ambient OP exposure. Similarly, the gut bacterial abundance and the predicted metagenome of 190 participants were obtained from 16S rRNA gene sequencing of fecal samples. Ambient long-term OP exposures were assessed based on pesticide application records combined with residential addresses via a geographic information system. We examined gut microbiome differences due to OP exposures, specifically differences in microbial diversity indices (Shannon diversity and Bray-Curtis dissimilarity), differential taxa abundance, and predicted Metacyc pathway abundance. Linear regression revealed no association between OP exposure and bacterial diversity after controlling for potential confounders. However, the predicted metagenes were sparser and less even among those highly exposed to OPs (p=0.02). Additionally, we found that the abundance of two families, 22 genera, and the predicted expression of 34 Metacyc pathways were associated with long-term OP exposure. These pathways included perturbed processes related to cellular respiration, increased biosynthesis and degradation of compounds related to bacterial wall structure, increased biosynthesis of RNA/DNA precursors, and decreased synthesis of Vitamin B1 and B6.
Lastly, we took a multi-omics approach to integrate the gut microbiome and serum metabolomic data collected from 113 PD patients and 46 controls. In addition to the microbiome data from 16s rRNA sequencing of fecal samples, high-resolution metabolomics analysis of serum samples was used to generate metabolomic profiles. Using principal component analysis of identified genera associated with PD, we constructed a summary PD-associated bacterial score using the first principal component. We then conducted partial least square (PLS) regression to identify metabolic features associated with the PD bacterial score. Pathway enrichment analysis was then performed on metabolic features selected from PLS. We identified 266 features and annotated 29 metabolic compounds as being associated with PD-related microbes at p<0.1. Pathway enrichment analysis indicated multiple perturbed pathways involving lipids metabolism including fatty acid activation and metabolism, linoleate metabolism, and glycerophospholipids metabolism, as well as carbohydrate metabolism such as hexose phosphorylation, starch and sucrose metabolism.
Our studies confirmed current literature on changes in the gut microbiome profile in PD patients compared to controls, and reported on some novel bacterial groups related to PD. In addition, we found evidence that the gut microbiome is associated with PD-specific features including disease duration and motor subtypes. Our findings of a disturbed gut microbiome in response to long-term OPs exposure provide evidence that the gut microbiome is a potential pathway for OPs causing an increased risk for PD. Linking the gut microbiome and serum metabolome, we found that gut microbiota associated with PD are involved in lipid metabolism and immune response in humans, adding information on the possible interplay between gut bacterial and human metabolism in PD. n conclusion, this dissertation enhances our understanding of risk factors contributing to Parkinson’s disease and provides evidence supporting the use of omics in investigating human disease, particularly for hypothesis validation and generation.
