Untargeted serum metabolomics reveals novel metabolite associations and disruptions in amino acid and lipid metabolism in Parkinson’s disease

• Published in: Molecular neurodegeneration Vol. 18; no. 1; pp. 100 - 16
• Main Authors: Paul, Kimberly C., Zhang, Keren, Walker, Douglas I., Sinsheimer, Janet, Yu, Yu, Kusters, Cynthia, Del Rosario, Irish, Folle, Aline Duarte, Keener, Adrienne M., Bronstein, Jeff, Jones, Dean P., Ritz, Beate
• Format: Journal Article
• Language: English
• Published: England BioMed Central Ltd 19.12.2023; BioMed Central; BMC
• Subjects: Amino acids; Amino Acids - metabolism; Biliverdin; Chromatography; Cresol; Cysteine; Dopa; Glucuronides; Glutamine; Humans; Independent study; Inflammation; Levodopa; Lipid Metabolism; Liquid chromatography; Mass spectrometry; Mass spectroscopy; Metabolic networks; Metabolism; Metabolites; Metabolomics; Movement disorders; Neurodegenerative diseases; p-Cresol; Pantothenic acid; Parkinson Disease - metabolism; Parkinson's disease; Physiological aspects; Population studies; Quality control; Regression analysis; Sulfates; United States > US; California
• ISSN: 1750-1326; 1750-1326
• DOI: 10.1186/s13024-023-00694-5

Untargeted high-resolution metabolomic profiling provides simultaneous measurement of thousands of metabolites. Metabolic networks based on these data can help uncover disease-related perturbations across interconnected pathways. Identify metabolic disturbances associated with Parkinson's disease (PD) in two population-based studies using untargeted metabolomics. We performed a metabolome-wide association study (MWAS) of PD using serum-based untargeted metabolomics data derived from liquid chromatography with high-resolution mass spectrometry (LC-HRMS) using two distinct population-based case-control populations. We also combined our results with a previous publication of 34 metabolites linked to PD in a large-scale, untargeted MWAS to assess external validation. LC-HRMS detected 4,762 metabolites for analysis (HILIC: 2716 metabolites; C18: 2046 metabolites). We identified 296 features associated with PD at FDR<0.05, 134 having a log fold change (FC) beyond ±0.5 (228 beyond ±0.25). Of these, 104 were independently associated with PD in both discovery and replication studies at p<0.05 (170 at p<0.10), while 27 were associated with levodopa-equivalent dose among the PD patients. Intriguingly, among the externally validated features were the microbial-related metabolites, p-cresol glucuronide (FC=2.52, 95% CI=1.67, 3.81, FDR=7.8e-04) and p-cresol sulfate. P-cresol glucuronide was also associated with motor symptoms among patients. Additional externally validated metabolites associated with PD include phenylacetyl-L-glutamine, trigonelline, kynurenine, biliverdin, and pantothenic acid. Novel associations include the anti-inflammatory metabolite itaconate (FC=0.79, 95% CI=0.73, 0.86; FDR=2.17E-06) and cysteine-S-sulfate (FC=1.56, 95% CI=1.39, 1.75; FDR=3.43E-11). Seventeen pathways were enriched, including several related to amino acid and lipid metabolism. Our results revealed PD-associated metabolites, confirming several previous observations, including for p-cresol glucuronide, and newly implicating interesting metabolites, such as itaconate. Our data also suggests metabolic disturbances in amino acid and lipid metabolism and inflammatory processes in PD.