Investigating Plasma Metabolomics and Gut Microbiota Changes Associated With Parkinson Disease: A Focus on Caffeine Metabolism

Abstract

Background and objectives: Coffee intake is linked to a reduced risk of Parkinson disease (PD), but whether this effect is mediated by gut microbiota and metabolomic changes remains unclear. This study examines PD-associated metabolomic shifts, caffeine metabolism, and their connection to gut microbiome alterations in a multicenter study.

Methods: We conducted an untargeted serum metabolomic assay using liquid chromatography with high-resolution mass spectrometry on an exploratory cohort recruited from National Taiwan University Hospital (NTUH). A targeted metabolomic assay focusing on caffeine and its 12 downstream metabolites was conducted and validated in an independent cohort from University Malaya Medical Centre (UMMC). In the exploratory cohort, the association of each caffeine metabolite with gut microbiota changes was investigated by metagenomic shotgun sequencing. A clustering-based approach was used to correlate microbiome changes with plasma caffeine metabolite level and clinical severity. Body mass index, antiparkinsonism medication use, and dietary habits (including coffee and tea intake) were recorded.

Results: Sixty-three patients with PD and 54 controls from NTUH formed the exploratory cohort while 36 patients with PD and 20 controls from UMMC served as an validation cohort to replicate the plasma caffeine findings. A total of 5,158 metabolites were detected from untargeted metabolomic analysis, with 3,131 having high confidence for analysis. Compared with controls, the abundance of 56 metabolites was significantly higher and that of 7 metabolites was significantly lower (adjusted p < 0.05 and log2 fold change >1) in patients with PD. Caffeine metabolism was significantly lower in patients with PD (p = 0.0013), and serum levels of caffeine and its metabolites negatively correlated with motor severity (p < 0.01). Targeted metabolomic analysis confirmed reduced levels of caffeine and its metabolites, including theophylline, paraxanthine, 1,7-dimethyluric acid, and 5-acetylamino-6-amino-3-methyluracil, in patients with PD; these findings were replicated in the validation cohort (p < 0.05). A clustering approach found that 56 microbiome species enriched in patients with PD negatively correlated with caffeine and its metabolites paraxanthine and theophylline (both p < 0.05), notably Clostridium sp000435655, Acetatifactor sp900066565, Oliverpabstia intestinalis, and Ruminiclostridium siraeum.

Discussion: This study identifies PD-related changes in microbial-caffeine metabolism compared with controls. Our findings offer insights for future functional research on caffeine-microbiome interactions in PD.

Figure 1. Serum Untargeted Metabolomics Analysis in Patients With PD and Healthy Controls in the Exploratory Cohort

(A) Flowchart and participants enrolled in the study. (B) The volcano plot reveals the metabolites with differential abundance between patients with PD and controls in the exploratory cohort. The orange and green points indicate metabolites that were in significantly higher and lower abundance in the PD group compared with the control group (adj. p value <0.05 and |log2(fold-change)| > 1). (C) Pathway enrichment analysis identified the most relevant metabolic pathways through pathway impact (x-axis) and p value (y-axis). Pathway impact represents a combination of centrality and pathway enrichment results: higher impact values indicate the relative importance of the pathway. The size and color of the points represent the impact and significance of the pathway, respectively, with more intense red colors indicating lower p values. PD = Parkinson disease.

Figure 2. KEGG Reference Caffeine Metabolism Pathway

KEGG = Kyoto Encyclopedia of Genes and Genomes. The metabolites highlighted by orange and blue dots are observed in humans, with orange dots indicating those detected in the metabolomics analysis.

Figure 3. Comparative Abundance of Caffeine Metabolites Between the PD and Control Groups From the Untargeted Metabolomics Analysis

PD = Parkinson disease.

Figure 4. Heatmap Shows the Correlation Profiles Between Gut Microbiome Abundance and the Serum Concentrations of Caffeine and Its Downstream Metabolites

Figure 5. Violin Plots Reveal the Distribution of Correlation Values Between the Abundance of Species in a Given Microbiome Cluster and a Specific Metabolite in Patients With PD

PD = Parkinson disease.

Figure 6. Gut Microbiome Clusters Identified Based on the Correlation Profiles Between Species Abundance and Caffeine Metabolite Concentrations

(A) Comparative abundance of the microbiome clusters in the PD and control groups. (B) The scatter plots show the correlation between the concentration of caffeine metabolite and the abundance of microbiome C3. PD = Parkinson disease.