Metabolic Pathways and Multiple Sclerosis: What Mendelian Randomization Reveals
Multiple sclerosis (MS) is a chronic autoimmune disorder of the central nervous system characterized by neuroinflammation, demyelination, axonal injury, and progressive neurodegeneration. Although genetic susceptibility and environmental exposures—including Epstein–Barr virus infection, smoking, obesity, and low vitamin D status—are established contributors, the biochemical mechanisms connecting these factors to disease initiation remain incompletely understood. The study by Ge and colleagues, entitled “A metabolome-wide Mendelian randomization study prioritizes potential causal circulating metabolites for multiple sclerosis,” addresses this gap by examining whether genetically influenced differences in circulating metabolites contribute directly to MS risk. Metabolomics studies have previously identified altered amino acids, ketone bodies, nucleotides, lipids, and lipoproteins in patients with MS. However, conventional case-control comparisons cannot determine whether such alterations precede disease onset, result from disease activity, or reflect treatment, diet, disability, or other confounding factors. The central contribution of this investigation is therefore not merely the identification of metabolic differences, but the prioritization of metabolites that may occupy causal positions in MS pathogenesis.
Mendelian Randomization as a Framework for Causal Inference
The investigators used two-sample Mendelian randomization, a genetic epidemiological approach that treats inherited genetic variants as instrumental variables for modifiable exposures. Because alleles are allocated at conception and generally precede disease development, genetic variants associated with metabolite concentrations can be used to estimate the long-term effect of those metabolites on MS susceptibility. In principle, this design reduces the influence of reverse causation and many environmental confounders that complicate observational metabolomics. The analogy is often made to a randomized clinical trial: individuals are effectively assigned different genetically influenced levels of an exposure before birth, although Mendelian randomization remains dependent on several assumptions. The genetic instruments must be strongly associated with the metabolite, must not be associated with confounders, and must influence MS only through the metabolite being studied. The authors selected independent single-nucleotide polymorphisms at two association thresholds, required at least three instruments per metabolite, and confirmed that instrument strength exceeded the conventional F-statistic threshold of 10. This dual-threshold strategy increased metabolome coverage while allowing the consistency of causal estimates to be examined under different instrument-selection criteria.
A Large-Scale Metabolome-Wide Analytical Strategy
Genetic associations with circulating metabolites were obtained from three genome-wide association studies involving 7,824, up to 24,925, and 115,078 participants of European ancestry. Associations with MS were derived from an International Multiple Sclerosis Genetics Consortium dataset containing 14,802 cases and 26,703 controls. After excluding metabolites with fewer than three genetic instruments, the authors evaluated 404 metabolites using genome-wide significant variants and expanded the analysis to 571 metabolites when the more permissive instrument threshold was applied. The workflow diagram on page 3 illustrates the progression from metabolite GWAS datasets through variant clumping, harmonization, primary inverse-variance-weighted analysis, and multiple sensitivity tests. The primary model combined variant-specific causal estimates using a multiplicative random-effects inverse-variance-weighted method. Robustness was then evaluated using MR-Egger regression, weighted median and weighted mode estimators, MR-PRESSO outlier detection, heterogeneity testing, leave-one-out analyses, and the MR Steiger directionality test. This multilayered design was intended to identify findings that remained credible despite potential pleiotropy, influential variants, or methodological assumptions.
Twenty-Nine Metabolites Emerge as Potential Causal Factors
The analysis prioritized 29 circulating metabolites with evidence of causal associations with MS. Six metabolites produced nominally significant and directionally consistent estimates at both genetic-instrument thresholds, whereas 23 showed suggestive evidence under at least one threshold. Sensitivity analyses generally supported the direction of the primary estimates, and Steiger testing favored a pathway from metabolite levels to MS rather than from MS liability to metabolite concentrations. Among the most prominent risk-associated metabolites were serine, lysine, O-sulfo-L-tyrosine, uridine, acetone, and acetoacetate. A genetically predicted one-standard-deviation increase in serine was associated with approximately 56% higher odds of MS, while lysine was associated with an 18% increase. Acetone and acetoacetate each showed odds ratios of approximately 2.5, although their wide confidence intervals indicate lower precision than the serine estimate. Uridine was associated with approximately 45% higher odds of disease. The forest plot on page 4 demonstrates that these associations extend across several metabolic classes rather than clustering within a single biochemical pathway.
Lipoprotein Subclasses Reveal a Complex Lipid Architecture
One of the most informative findings was that the relationship between lipid metabolism and MS depended strongly on lipoprotein particle subclass. Genetically predicted total cholesterol, phospholipids, and triglycerides within large very-low-density lipoprotein particles were associated with lower MS risk. Phospholipids in small VLDL and in chylomicrons or the largest VLDL particles also showed inverse associations. In contrast, total cholesterol, cholesterol esters, and phospholipids carried in very large high-density lipoprotein particles were associated with increased risk. These results caution against treating total circulating cholesterol or aggregate HDL cholesterol as biologically uniform exposures. Lipoprotein particles differ in size, composition, trafficking, inflammatory activity, and interactions with immune and vascular cells. Accordingly, the same lipid class may have different biological consequences depending on its carrier particle. The findings are compatible with previous Mendelian randomization evidence linking higher HDL cholesterol to MS risk, while also refining that relationship to specific HDL subclasses. They further suggest that lipid abnormalities observed after MS diagnosis may partly reflect disease-related metabolic adaptation rather than causal injury, particularly when observational and genetically inferred associations point in different directions.
Amino Acids and Ketone Bodies Connect Immunity, Myelin, and Energy Metabolism
The associations involving serine, lysine, acetone, and acetoacetate provide plausible links between systemic metabolism and central nervous system pathology. Serine contributes to one-carbon metabolism, nucleotide synthesis, redox regulation, and the production of phosphatidylserine and sphingolipids, which are important components of cellular membranes and myelin. The authors note that Epstein–Barr virus can increase serine uptake and biosynthesis in B cells, suggesting a possible intersection between viral biology, immune-cell metabolism, and MS susceptibility. Lysine may similarly reflect altered amino-acid utilization or immune activation, although its mechanistic role remains less clearly defined. The ketone bodies acetone and acetoacetate are particularly intriguing because they were associated with higher disease risk, whereas ketogenic interventions have sometimes been proposed as beneficial after MS onset. The distinction between disease initiation and disease treatment is critical: lifelong genetically elevated ketone-body concentrations may have different consequences from a time-limited dietary intervention in established disease. The study therefore does not demonstrate that ketogenic diets cause MS. Instead, it indicates that ketone-body metabolism may participate differently across disease stages and warrants targeted experimental investigation.
Clinical Significance, Limitations, and Future Directions
This study provides a prioritized catalogue of metabolic candidates rather than immediate diagnostic tests or therapeutic prescriptions. Mendelian randomization estimates represent the consequences of lifelong genetically influenced exposure and do not specify the critical developmental period, tissue compartment, dose-response shape, or effects of short-term clinical manipulation. Horizontal pleiotropy also cannot be eliminated completely, some metabolites lacked sufficient genetic instruments, and the analysis could not distinguish relapsing-remitting from progressive MS. Furthermore, the exclusive use of European-ancestry datasets limits generalizability, and the assumption of linear metabolite–disease relationships may not hold for all exposures. Nevertheless, the integration of large GWAS datasets with multiple sensitivity methods makes the identified metabolites valuable candidates for replication, functional validation, and drug-target research. Future work should incorporate ancestry-diverse cohorts, MS subtype-specific GWAS, multivariable Mendelian randomization, tissue-resolved metabolomics, and experimental models capable of testing whether modification of serine metabolism, ketone-body pathways, or specific lipoprotein subclasses changes immune activation or demyelination. The principal scientific advance is therefore a shift from cataloguing metabolic consequences of MS toward constructing testable causal models of its biochemical origins.
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.
