Metabolic Pathways and Multiple Sclerosis: Genetic Evidence for Potential Causal Metabolites
Multiple sclerosis (MS) is a chronic immune-mediated disease of the central nervous system characterized by inflammation, demyelination, and progressive neurodegeneration. Although genetic susceptibility and environmental factors such as Epstein–Barr virus infection, smoking, obesity, and low vitamin D status are established contributors to MS risk, the biochemical processes connecting these factors to disease initiation remain incompletely understood. Metabolomics—the systematic measurement of small molecules involved in cellular metabolism—has revealed numerous differences between people with MS and healthy individuals, including alterations in amino acids, lipids, nucleotides, and compounds associated with energy metabolism. A central difficulty, however, is determining whether these metabolic abnormalities contribute to the development of MS or arise as consequences of inflammation, neurological injury, medication, or changes in lifestyle following diagnosis. The 2023 study by Ge and colleagues, published in the Journal of Neuroimmunology, addresses this problem through a metabolome-wide Mendelian randomization analysis designed to prioritize circulating metabolites that may have causal relationships with MS rather than merely correlating with the disease.
Using Genetics as a Natural Experiment
The investigators applied two-sample Mendelian randomization (MR), a genetic epidemiological approach in which inherited genetic variants associated with an exposure are used as instrumental variables to estimate that exposure's effect on an outcome. Because alleles are allocated before disease develops, MR can reduce some of the reverse-causation and confounding problems that complicate conventional observational studies. The researchers obtained genetic associations with circulating metabolites from three genome-wide association studies involving up to 115,078 participants of European ancestry and compared these with genetic associations for MS from the International Multiple Sclerosis Genetics Consortium, comprising 14,802 cases and 26,703 controls. As illustrated by the study workflow on page 3, metabolites were required to have at least three independent genetic instruments, and the investigators employed two thresholds for selecting variants: the conventional genome-wide threshold of P < 5 × 10⁻⁸ and a more permissive threshold of P < 1 × 10⁻⁶. The stricter analysis included 404 metabolites, whereas the broader approach increased coverage to 571. The principal causal estimates were obtained using an inverse variance-weighted model, with MR-Egger regression, weighted median, weighted mode, MR-PRESSO, leave-one-out analyses, and MR Steiger testing used to evaluate robustness, pleiotropy, influential variants, and the likely direction of causation.
Twenty-Nine Metabolites Emerge as Candidates
Across this large-scale screen, the researchers identified 29 circulating metabolites with evidence suggesting a potential causal relationship with MS. Six demonstrated nominally significant and directionally consistent associations under both genetic-instrument thresholds, while another 23 showed significant evidence under one threshold. Importantly, the authors described these molecules as candidates for prioritization rather than established causes, an appropriate distinction because many of the reported associations were based on nominal significance and Mendelian randomization itself depends on assumptions concerning instrument validity and horizontal pleiotropy. Nevertheless, several biological pathways appeared repeatedly, particularly lipid metabolism, amino-acid metabolism, and cellular energy metabolism. The forest plot presented on page 4 demonstrates that the direction of association was not uniform even within the same broad biochemical class: some metabolites were associated with higher predicted MS risk, whereas others appeared protective. This complexity indicates that interpreting metabolic contributions to MS requires attention not simply to the overall concentration of lipids or amino acids but also to molecular subtype, carrier particle, metabolic pathway, and physiological context.
Lipoprotein Subclasses Reveal a Striking Metabolic Contrast
One of the most notable findings concerned lipids contained within different lipoprotein subclasses. Genetically predicted increases in total cholesterol, phospholipids, and triglycerides within large very-low-density lipoprotein (VLDL) particles were associated with lower MS risk, with odds ratios of approximately 0.83, 0.80, and 0.81, respectively. Phospholipids in small VLDL and in chylomicrons or the largest VLDL particles also showed inverse associations. In contrast, several lipid components carried by very large high-density lipoprotein (HDL) particles were associated with greater MS risk: total cholesterol in very large HDL produced an odds ratio of 1.20, phospholipids an odds ratio of 1.13, and cholesterol esters an odds ratio of 1.14. These findings are scientifically important because they challenge the oversimplified assumption that a lipid such as cholesterol has a uniform biological effect regardless of its transport environment. Previous observational work had reported elevated lipid concentrations in large VLDL and HDL subclasses among patients with relapsing-remitting MS, but such observations could reflect disease-induced metabolic changes. The MR results suggest instead that the specific lipoprotein context may matter: very large HDL-associated lipids could contribute to susceptibility, whereas certain VLDL-associated lipids may be inversely related to risk.
Serine, Lysine and Ketone Bodies Link MS to Core Cellular Metabolism
The analysis also highlighted several metabolites with direct relevance to cellular biosynthesis and energy production. Genetically predicted serine showed one of the strongest associations, with each genetically predicted one-standard-deviation increase corresponding to an estimated 56% increase in MS odds (OR 1.56, 95% CI 1.25–1.95). Lysine was similarly associated with higher risk, although more modestly (OR 1.18), as was O-sulfo-L-tyrosine. Serine is particularly intriguing because it participates in one-carbon metabolism and serves as a precursor for phosphatidylserine and sphingomyelin, lipids with important roles in neural membranes and myelin. Epstein–Barr virus, itself strongly implicated in MS susceptibility, can also stimulate serine uptake and biosynthesis in B cells, providing a plausible mechanistic bridge worthy of experimental investigation. The study additionally identified two ketone bodies—acetoacetate and acetone—with comparatively large risk estimates: OR 2.47 and OR 2.45, respectively. The detailed scatter, forest, and leave-one-out plots on page 5 show how individual genetic variants contributed to the estimates for serine, acetoacetate, and acetone. Although potential outliers were considered, MR-PRESSO did not detect significant outliers for acetoacetate, and the acetone association remained significant after identified outliers were removed. These results are particularly thought-provoking because ketone bodies may have different biological implications before MS develops than during treatment, where ketogenic interventions have sometimes been investigated for possible therapeutic benefit.
Methodological Strength Does Not Eliminate Important Uncertainties
The study has several methodological advantages, including its systematic examination of hundreds of metabolites, the use of strong genetic instruments with F-statistics above 10, concordant analyses under two instrument-selection thresholds, and the application of multiple complementary MR sensitivity methods. These procedures make the conclusions more resistant to weak-instrument bias and provide checks against certain forms of horizontal pleiotropy. Nevertheless, Mendelian randomization cannot reproduce every feature of a randomized clinical trial. The authors acknowledge that residual horizontal pleiotropy cannot be completely excluded, that some metabolites could not be analyzed because too few suitable genetic variants were available, and that differences among metabolomics GWAS may influence statistical power. The analysis also treats metabolite–disease relationships as approximately linear and reflects the consequences of lifelong genetically influenced differences in metabolite levels rather than the effect of changing a metabolite concentration for a defined period in adulthood. Furthermore, the study could not distinguish between different MS subtypes and was restricted to participants of European ancestry, limiting immediate generalization to more genetically diverse populations. Consequently, the findings should be interpreted as a map of promising biological hypotheses rather than as evidence that altering any identified metabolite will necessarily prevent or treat MS.
From Metabolic Association to Mechanistic Medicine
The broader significance of this research lies in shifting MS metabolomics from descriptive biomarker discovery toward causal prioritization. Serine, lysine, uridine, acetone, acetoacetate, gamma-glutamyl amino acids, and lipid components of specific VLDL and HDL subclasses now represent candidates for detailed mechanistic investigation. Future research should determine how these metabolites influence immune-cell activation, blood–brain barrier function, oligodendrocyte biology, myelin synthesis and degradation, mitochondrial metabolism, and neuroinflammatory signaling. Replication in larger and ancestrally diverse genetic datasets will be essential, as will studies capable of distinguishing relapsing-remitting, secondary progressive, and primary progressive MS. Experimental manipulation in cellular and animal models could then establish whether the relationships predicted by MR correspond to biologically modifiable pathways. Ultimately, the importance of this study is not that it identifies a single metabolic cause of multiple sclerosis, but that it provides a systematically derived shortlist of biochemical pathways that may participate in disease development. Such prioritization can help direct future research toward more informative biomarkers, preventive strategies, and potentially novel therapeutic targets while emphasizing that translation from genetic causal inference to clinical intervention requires substantial further validation.
Disclaimer: This blog post is based on the provided research article and is intended for informational purposes only. It is not intended to provide medical advice. Please consult with a healthcare professional for any health concerns.
References:
Ge, A., Sun, Y., Kiker, T., Zhou, Y., & Ye, K. (2023). A metabolome-wide Mendelian randomization study prioritizes potential causal circulating metabolites for multiple sclerosis. Journal of neuroimmunology, 379, 578105.
