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

Abstract

Objectives: Recent advances in high-resolution metabolomics platforms allow the simultaneous measurement of thousands of small molecules produced from metabolism (metabolites), providing a map of disease-related perturbations across interconnected pathways. We used high performance, untargeted metabolomics to identify metabolic disturbances and molecular events associated with Parkinson's disease (PD) in two population-based studies. Methods: We performed a metabolome-wide association study (MWAS) on PD, using serum-based untargeted metabolomics data derived from high resolution liquid chromatography, mass spectrometry (LCMS). We used two independent, case-control populations for discovery and replication (n=642 PD patients, n=277 controls). Results: From the LCMS, 5,145 metabolites were detected across the two study populations in ≥50% of the samples (HILIC: 2913 metabolites; C18: 2063 metabolites). Using logistic regression and an FDR to correct for multiple testing, we determined 236 metabolites were associated with PD in a meta-analysis at an FDR<0.05. Of these, 110 metabolites were independently associated with PD in both discovery and replication studies at p<0.05 (187 at p<0.10), while 24 were associated with levodopa-equivalent dose among the PD patients. Intriguingly, the microbial related metabolite, p-cresol (meta-OR=1.29, 95% CI=1.13, 1.47, FDR=0.01), and it's two metabolites, p-cresol sulfate and p-cresol glucuronide, were found at higher intensity among the PD patients relative to controls. P-cresol glucuronide was also associated with motor symptoms among patients. Pyroglutamic acid (meta OR=3.79, 95% CI=2.60, 5.54; FDR=5.30E-09), the anti-inflammatory metabolite itaconate (meta OR=0.47, 95% CI=0.36, 0.61; FDR=8.44E-06), and cysteine-S-sulfate (meta OR=1.56, 95% CI=1.32, 1.83; FDR=1.66E-05) were also among the most strongly associated metabolites. Seventeen pathways were also enriched, including several related to amino acid and lipid metabolism. Conclusions: Our results revealed PD-associated metabolites in two independent study populations, implicating individual metabolites including p-cresol and itaconate, as well as suggesting metabolic disturbances in amino acid and lipid metabolism and inflammatory processes.

Untargeted Plasma Metabolomics Unveils Distinct Metabolite Profiles in Parkinson’s Disease Subtypes: A Focus on idiopathic REM Sleep Behavior Disorders

“…Further comparisons between all RBD groups versus control (Figure 2B and D, Table S4) and PD_Only group by generalized fold changes and Wilcoxon rank-sum tests (Figure 2C and E, Table S5) revealed RBD-specific enriched metabolites, including secondary bile acids (lithocholate sulfate, glycolithocholate), p-cresol sulfate, and phenylacetylglutamine (PAG), which are of gut microbial origin (10), Conversely, enriched metabolites in PD_Only included glucose and cortisol, while caffeine levels were decreased. Although previous studies have reported increased plasma levels of microbial p-cresol sulfate and PAG in PD (14,16), we did not observe these in our PD_Only group but found them in the PD_RBD+ group, supporting the importance of considering PD subtypes for accurate metabolite profiling. Reductions in caffeine metabolites, including caffeine, theobromine, and 1-methylurate, were consistent features of PD, aligning with other studies (14,(26)(27)(28).…”

“…On the other hand, 4-ethylphenylsulfate (4-EPS), a microbiota-derived metabolite found at increased levels in the plasma of individuals with autism spectrum disorder, has been shown to induce anxiety-like behaviors in mice when derived from the gut and crossing the blood-brain barrier (58–60), although its impact in RBD and PD has not been investigated. p-cresol sulfate, another metabolite with a microbial origin, has been detected at higher concentrations in PD across different cohorts (14, 16). p-cresol sulfate is known to accumulate with aging and influence oxidative stress and inflammation, which may play a role in behavior disorders and neurodegeneration (61).…”

“…Extensive metabolomic profiling of blood samples from PD patients has been conducted to identify metabolic biomarkers for better characterization or early-stage diagnosis of PD (12)(13)(14)(15)(16). Some studies have provided insights into circulating blood metabolites of gut microbial origin, such as p-cresol sulfate and phenylacetylglutamine (14,16). However, it is not clear whether PD subjects used in these studies represent mixed subtypes of PD (brain-first or bodyfirst).…”

Localized Pantothenic Acid (Vitamin B5) Reductions Present Throughout the Dementia with Lewy Bodies Brain

“…As pantothenic acid levels cannot be determined in vivo in the brain with current methods, investigations of peripheral levels, such as in the CSF or plasma, would be necessary to determine the temporality of alterations; this, however, would only be possible in the event that cerebral pantothenic acid changes are reflected in the periphery. There is some evidence that this may be the case, as has been observed in some studies on PD plasma and serum [27,28]; however, this has yet to be determined in DLB. As such, assessing peripheral pantothenic acid levels in a longitudinal cohort would be an intuitive and informative investigation to follow from the current study.…”

“…Non-CNS disturbances in pantothenic acid and related pathways have also been observed in neurodegenerative diseases; for instance, pantothenate and CoA biosynthesis has been found to be disturbed in LC-MS metabolomics analyses of PD and healthy plasma and serum samples, with decreased pantothenic acid levels observed even in the early stages of PD [27,28]. The presence of pantothenic alterations in the PD gut is disputed, with some studies showing no changes in stool sample levels [29] while others have shown decreases [30], and with another study showing positive associations between fecal pantothenic acid levels and non-motor symptoms in PD [31].…”