Gut Microbes and Blood Chemistry Reveal Hidden Metabolic Signature of Parkinson’s Disease

Parkinson’s disease has long been framed as a disorder of the brain, a slow erosion of dopamine-producing neurons that manifests as tremor, rigidity, and slowness of movement. Yet over the past decade, an increasingly persuasive body of evidence has pointed somewhere unexpected: the gut. Now, a new cross-sectional study published in npj Parkinson’s Disease has taken that idea further than most, deploying a multi-omics strategy to map, in unusual detail, the chemical conversation between the gut microbiome and the rest of the body in people living with the condition. The results suggest that the disease is written not only in neurons but also in the lipids and amino acids circulating through the bloodstream, many of them shaped by microbial metabolism.

The research team, led by Tzi Shin Toh, Chun Wie Chong, and Ai Huey Tan, with colleagues spanning neurology, medical microbiology, physiology, and computer science at Universiti Malaya, Monash University Malaysia, and Korea University, enrolled 67 people with Parkinson’s disease and 64 non-Parkinson’s controls. Rather than examining a single molecular layer in isolation, they integrated three complementary datasets: untargeted serum metabolomics and untargeted fecal metabolomics, both performed using liquid chromatography-mass spectrometry, alongside 16S rRNA gene sequencing of the gut microbiome. This combination allowed them to simultaneously survey what microbes are present in the gut, what metabolic products accumulate in feces, and what small molecules travel through the blood.

The most striking finding concerned lipids. Parkinson’s disease patients exhibited widespread lipid dysregulation, with measurable alterations in serum isoprenoids, a class of molecules built from repeating five-carbon units that feed into cholesterol synthesis and protein modification pathways. Intriguingly, the researchers also detected changes in a bacteria-derived precursor of isoprenoids, 2-C-methyl-D-erythritol 4-phosphate, a compound produced through the MEP pathway used by many bacteria rather than the classical mevalonate pathway of human cells. Its presence and perturbation in the serum of Parkinson’s patients hints that some of the lipid abnormalities in the disease may originate not from human biochemistry but from the metabolic machinery of gut microbes themselves.

Beyond isoprenoids, the team documented disturbances in two other lipid families with well-established roles in cellular life: sphingolipids and glycerophospholipids. Sphingolipids are structural components of cell membranes and key signaling molecules, and their disruption has been implicated in neurodegeneration, including the alpha-synuclein aggregation that defines Parkinson’s pathology. Glycerophospholipids, meanwhile, form the backbone of every cellular membrane and participate in inflammatory signaling. The study also found perturbations in fatty acid biosynthesis pathways, suggesting that the lipid economy of Parkinson’s patients is broadly, rather than narrowly, remodeled. Together, these findings position lipid metabolism as a central arena in which host and microbial biology interact in the disease.

To make sense of the three data layers at once, the researchers applied multi-omics integration methods, asking which of the molecular and microbial datasets carried the strongest statistical association with Parkinson’s disease status. The answer was clear: the disease was more closely associated with the gut microbiome and serum metabolomics than with the fecal metabolome. In other words, the combination of who lives in the gut and what circulates in the blood proved more informative about disease state than the raw metabolic contents of stool. This kind of integrative conclusion is difficult to reach from any single omics platform and underscores the value of measuring multiple molecular domains in the same individuals.

One specific microbial-metabolite pairing stood out. Serum levels of N-hydroxy-L-phenylalanine, a modified amino acid, were linked to Desulfovibrio, a genus of sulfate-reducing bacteria known to produce hydrogen sulfide. Desulfovibrio has attracted attention in Parkinson’s research before, partly because hydrogen sulfide can influence mitochondrial function and gut barrier integrity, and partly because the genus has been repeatedly reported as enriched in the gut microbiomes of patients. The new association implicates microbial amino acid metabolism as a plausible contributor to the systemic metabolic shifts seen in the disease, adding a concrete biochemical thread to what has often been a correlational literature.

Perhaps the most clinically provocative result came from machine learning. The team trained predictive models to estimate clinical features of Parkinson’s disease from the multi-omics data and found that cognition was predicted more robustly than motor impairment. Cognitive decline is one of the most feared and difficult-to-manage aspects of Parkinson’s disease, and objective biomarkers that track it remain scarce. The fact that a model built from molecular and microbial features could predict cognitive performance more reliably than motor scores suggests that the gut-metabolic axis may be more tightly coupled to the non-motor dimensions of the disease than to its classical movement symptoms.

Equally notable was the composition of the predictive features. The models predicting cognition were driven exclusively by gut-derived features, including short-chain fatty acid-producing taxa, opportunistic pathogens, and fecal lipid metabolites. No serum metabolites contributed to the cognitive predictions. Short-chain fatty acids and their producing bacteria are central players in gut-brain signaling, influencing immune tone, barrier function, and even microglial activity in the brain. Opportunistic pathogens, by contrast, are organisms that typically lurk harmlessly but can flourish under dysbiotic conditions, and their presence alongside beneficial taxa in the predictive set paints a picture in which both the loss of protective functions and the gain of potentially harmful ones shape the cognitive landscape of Parkinson’s disease.

The authors emphasize that these findings highlight novel microbial-driven lipid and amino acid pathways with potential biomarker and therapeutic relevance in Parkinson’s disease. The implications run in two directions. As biomarkers, gut microbial taxa and fecal lipid metabolites could eventually complement clinical assessment, particularly for tracking cognitive trajectory, where current tools are limited. As therapeutic targets, the MEP pathway and the metabolic activities of organisms like Desulfovibrio offer possibilities that conventional neurology has never considered: modulating specific bacterial metabolic outputs rather than neurons directly. Neither application is imminent, and the study’s cross-sectional design means that causality remains unresolved, but the specificity of the molecular signals gives both avenues a concrete starting point.

The study also carries methodological weight. By combining untargeted metabolomics in two compartments with microbiome sequencing and machine learning in a single cohort, the researchers demonstrated a template for how gut-brain research can move beyond simple association studies toward mechanistically annotated, clinically oriented signatures. The work was funded by the Ministry of Higher Education Malaysia through a Fundamental Research Grant Scheme awarded to Ai Huey Tan, and it reflects a growing recognition that Parkinson’s disease research is a genuinely multi-system endeavor. For patients and clinicians, the message is that the gut is not merely a bystander in Parkinson’s disease; it is an active metabolic participant whose chemical output, measurable in blood and stool, appears to track with some of the most burdensome features of the illness. As larger and longitudinal studies build on these findings, the microbial and metabolic markers identified here may help define which biological pathways deserve the next generation of therapeutic attention.

Subject of Research: Host–gut microbiome metabolic crosstalk in Parkinson's disease

Article Title: Multi-omics of host–gut metabolic crosstalk in Parkinson’s disease

Article References: Multi-omics of host–gut metabolic crosstalk in Parkinson’s disease. (n.d.). https://doi.org/10.1038/s41531-026-01561-y

Image Credits: AI Generated

DOI: 10.1038/s41531-026-01561-y

Keywords: Parkinson's disease, gut microbiome, metabolomics, multi-omics, lipid metabolism, Desulfovibrio, sphingolipids, short-chain fatty acids, machine learning, cognition, 16S rRNA sequencing, gut-brain axis