Metabolomic Clues to Multiple Sclerosis: A Mendelian Randomization Study of Potential Causal Blood Metabolites
Multiple sclerosis is a chronic autoimmune disease of the central nervous system, marked by neuroinflammation, demyelination, and neurodegeneration. Although established risk factors such as Epstein–Barr virus infection, smoking, vitamin D deficiency, obesity, and genetic susceptibility have advanced the field, the molecular pathways linking these risks to disease onset remain incompletely resolved. The article by Ge and colleagues addresses this gap by focusing on the circulating metabolome: the measurable set of small molecules in blood that reflects lipid metabolism, amino acid turnover, energy balance, and inflammatory physiology. Rather than treating metabolites only as disease biomarkers, the study asks whether some may contribute causally to multiple sclerosis risk.
Study Design: A Metabolome-Wide Mendelian Randomization Framework
The authors used a two-sample Mendelian randomization design, which leverages genetic variants as instrumental variables to infer whether lifelong differences in metabolite levels may influence disease risk. This approach helps reduce two major limitations of observational metabolomics: reverse causation and residual confounding. The study integrated genetic instruments from three blood-metabolome GWAS datasets and multiple sclerosis associations from the International Multiple Sclerosis Genetics Consortium, including 14,802 cases and 26,703 controls. As illustrated in the workflow diagram on page 3, the analysis evaluated metabolites under two SNP-selection thresholds and applied several sensitivity methods, including inverse variance-weighted MR, MR-Egger, weighted median, weighted mode, MR-PRESSO, and MR Steiger analyses.
Main Findings: Twenty-Nine Candidate Causal Metabolites
Across 571 circulating metabolites, the study prioritized 29 metabolites with suggestive evidence of causal association with multiple sclerosis. Six metabolites showed nominally significant and directionally consistent results under both genetic-instrument thresholds, while another 23 reached significance under one threshold. The forest plot on page 4 summarizes these associations, showing both risk-increasing and risk-decreasing metabolite effects. Notably, genetically predicted higher levels of serine, lysine, acetone, acetoacetate, and uridine were associated with increased multiple sclerosis risk, whereas several lipid fractions in large very-low-density lipoprotein particles were associated with lower risk.
Lipid Metabolism: Lipoprotein Subclasses as Disease-Relevant Signals
One of the most important contributions of the article is its emphasis on lipoprotein subclass biology rather than broad lipid categories alone. The authors found that cholesterol, phospholipids, and triglycerides in large VLDL particles were associated with reduced multiple sclerosis risk, while cholesterol, cholesterol esters, and phospholipids in very large HDL particles were associated with increased risk. This distinction is biologically meaningful because lipoprotein subclasses differ in size, density, molecular cargo, and inflammatory behavior. The findings suggest that the role of lipid metabolism in multiple sclerosis cannot be adequately interpreted through conventional lipid markers alone, such as total HDL-C or LDL-C, but may require a more granular metabolomic and lipoproteomic framework.
Amino Acids and Ketone Bodies: Links to Immunometabolism and Neurobiology
The study also implicates amino acid and energy metabolism in multiple sclerosis susceptibility. Genetically predicted higher levels of serine and lysine were associated with elevated disease risk, aligning with prior observational evidence that amino acid profiles differ in patients with multiple sclerosis. Serine is especially notable because it participates in one-carbon metabolism and serves as a precursor for lipids relevant to myelin biology, including phosphatidylserine and sphingomyelin. The study further identified acetone and acetoacetate, both ketone-related metabolites, as risk-increasing candidates. On page 5, scatter plots, forest plots, and leave-one-out analyses for serine, acetoacetate, and acetone visually support the robustness of these associations while also showing the importance of checking whether individual SNPs disproportionately influence MR estimates.
Interpretation: Causality, Robustness, and Biological Caution
The article’s strength lies in its systematic design and its use of multiple sensitivity analyses to test whether the inferred causal effects were robust to pleiotropy and instrument-related bias. The authors also applied MR Steiger analysis to support the direction of causality from metabolite to multiple sclerosis rather than the reverse. However, the findings should be interpreted as prioritization rather than definitive proof of mechanism. Mendelian randomization estimates lifelong genetically influenced exposure, not the effect of short-term dietary, pharmacological, or disease-stage-specific changes in metabolite levels. The study also focuses on individuals of European ancestry, does not distinguish among multiple sclerosis subtypes, and cannot fully exclude horizontal pleiotropy despite extensive sensitivity testing.
Future Directions: From Metabolite Prioritization to Translational Research
Overall, this metabolome-wide Mendelian randomization study provides a valuable causal hypothesis map for multiple sclerosis research. By identifying metabolites in lipid, amino acid, nucleotide, and energy metabolism, it offers candidate biomarkers for risk stratification and potential molecular entry points for therapeutic investigation. Future work should test these metabolites in longitudinal cohorts, evaluate their relevance across ancestry groups and multiple sclerosis subtypes, and integrate metabolomics with transcriptomics, proteomics, immune profiling, and neuroimaging. The most important implication is not that any single metabolite fully explains multiple sclerosis pathogenesis, but that metabolic regulation may be an active component of disease susceptibility rather than a passive consequence of established neuroinflammation.
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.
