Genetic Links Between Lipid Metabolism and Multiple Sclerosis Severity
Multiple sclerosis is characterized not only by considerable clinical heterogeneity but also by substantial uncertainty regarding the biological processes that determine long-term disability. Individuals with apparently similar disease onset, demographic characteristics, and treatment histories may experience markedly different rates of neurological deterioration. Metabolomic studies have identified alterations in lipids, amino acids, and energy-related intermediates among people with multiple sclerosis; however, conventional observational analyses cannot easily determine whether these alterations contribute to disease progression or arise as consequences of inflammation, disability, medication exposure, dietary change, or reduced physical activity. The study by Noroozi and colleagues addresses this problem through an integrative genetic framework designed to distinguish potentially causal metabolic pathways from secondary disease-associated changes. Its central conclusion is that genetically regulated polyunsaturated fatty acid metabolism—particularly pathways controlled by FADS1 and, to a lesser extent, CYP4F2—may contribute to variation in multiple sclerosis severity. Because the article is a preprint that has not undergone peer review, its findings should be regarded as hypothesis-strengthening rather than clinically definitive.
A Multi-Layered Mendelian Randomization Strategy
The investigators applied two-sample Mendelian randomization, an instrumental-variable approach that uses inherited genetic variants as proxies for modifiable biological exposures. The analysis integrated genome-wide association data for 1,091 circulating metabolites and 309 metabolite ratios measured in 8,299 participants from the Canadian Longitudinal Study on Aging with genetic data on age-related multiple sclerosis severity from 12,584 people with multiple sclerosis. Severity was quantified using the Age-Related Multiple Sclerosis Severity Score, which contextualizes disability according to age. Metabolites were retained when their circulating concentrations showed sufficient SNP-based heritability, and instrumental variants were selected according to association strength, allele frequency, linkage disequilibrium, and instrument-strength criteria. Variants in the major histocompatibility complex were excluded because of the region’s extensive linkage disequilibrium and complex immunological effects. As illustrated in the study workflow on page 5, the authors then combined conventional metabolome-wide Mendelian randomization with sensitivity testing, multivariable models, genetic colocalization, pathway-resolved cis-Mendelian randomization, and brain single-cell expression analyses.
Metabolome-Wide Signals Extend Beyond a Single Biochemical Class
The primary inverse-variance-weighted analysis identified 45 metabolites with nominal evidence of an effect on multiple sclerosis severity, including 36 annotated metabolites or ratios and nine uncharacterized compounds. Forty of these signals showed concordant effect directions in MR-Egger analyses, providing a degree of robustness across methods with different assumptions about horizontal pleiotropy. The implicated compounds encompassed amino acids, fatty acid-related lipids, mitochondrial intermediates, peptides, xenobiotics, and metabolite ratios. Pathway-specific multivariable Mendelian randomization, which attempted to separate the direct effects of correlated metabolites, prioritized three amino acid-related exposures and two fatty acid-related exposures. Higher genetically predicted arginine and propionylglycine were associated with greater severity, whereas betaine, succinoyltaurine, and arachidonate were associated with lower severity in the multivariable models. These observations suggest that multiple sclerosis progression may reflect interacting disturbances in lipid remodeling, one-carbon metabolism, amino acid utilization, and mitochondrial bioenergetics rather than a single isolated metabolic defect.
FADS1 Emerges as a Central Regulator of Polyunsaturated Fatty Acid Effects
The most coherent mechanistic evidence arose from the chromosome 11 region containing FADS1 and FADS2, which encode desaturase enzymes responsible for key steps in long-chain polyunsaturated fatty acid synthesis. Genetic colocalization indicated that several fatty acid-related metabolite ratios and multiple sclerosis severity were likely influenced by a shared signal within this locus. The relevant ratios included arachidonate-to-linoleate, arachidonate-to-oleate or vaccenate, and a glycerolipid ratio reflecting linoleic- and arachidonic-acid-containing species. Although the colocalized region also includes genes such as MYRF and TMEM258, the metabolic direction of the associations strongly supported the biological relevance of the FADS pathway. The authors therefore used the functional variant rs174546 as a cis-instrument for reduced FADS1 activity. This allele has been associated with lower FADS1 expression and a characteristic biochemical profile: accumulation of upstream FADS2-derived intermediates accompanied by depletion of downstream FADS1-derived products. Across numerous phospholipid and fatty acid biomarkers, this genetically proxied reduction in Δ5-desaturase activity was consistently associated with greater multiple sclerosis severity, making pathway-level coherence one of the study’s principal strengths.
Brain Cell-Specific Analyses Connect Systemic Lipid Biology to the Central Nervous System
A particularly important component of the study was its extension from circulating metabolites to cell-specific gene expression in the brain. Using single-cell expression quantitative trait locus data from 474 individuals, the researchers evaluated rs174546 as an instrument for FADS1 expression across seven central nervous system cell types. The severity-associated allele was linked to reduced FADS1 expression in astrocytes, oligodendrocytes, excitatory neurons, and inhibitory neurons, but not significantly in microglia, endothelial cells, or oligodendrocyte precursor cells. Cis-Mendelian randomization estimates indicated that genetically predicted reductions in FADS1 expression within the four affected cell populations were associated with increased multiple sclerosis severity. These findings are biologically plausible because polyunsaturated fatty acids influence membrane architecture, myelin biology, neuronal signaling, oxidative balance, and the production of inflammatory lipid mediators. Nevertheless, the results do not establish that reduced FADS1 expression directly damages these cell types; rather, they identify a genetically anchored association that should be examined through experimental manipulation in human cellular systems, organoids, and animal models.
CYP4F2 and Mitochondrial Metabolism Define Complementary Mechanistic Axes
The analysis also identified a secondary signal at CYP4F2, a cytochrome P450 enzyme involved in the ω-hydroxylation and turnover of long-chain polyunsaturated fatty acids, including arachidonic acid. Moderate colocalization was observed between succinoyltaurine levels and multiple sclerosis severity at the CYP4F2 locus, with rs2108622—a functional missense variant associated with reduced CYP4F2 protein abundance—prioritized as the shared candidate variant. Cis-Mendelian randomization using metabolites downstream of CYP4F2 activity suggested that genetically proxied enzyme reduction was associated with greater disease severity. Although weaker than the FADS1 evidence, this result reinforces the broader interpretation that dysregulated polyunsaturated fatty acid processing and bioactive lipid signaling may influence neurological outcomes. Additional metabolome-wide associations involving α-ketoglutarate, the α-ketoglutarate-to-proline ratio, and succinyl-CoA-related compounds further implicated mitochondrial metabolism. Such findings raise the possibility that lipid remodeling, inflammatory signaling, and cellular energy production form an interconnected network contributing to neuroaxonal injury and disability accumulation.
Translational Significance, Limitations, and Research Priorities
The study provides a compelling example of how metabolomics, human genetics, and cell-specific regulatory data can be integrated to move from broad biochemical association toward testable molecular mechanisms. Its results nominate FADS1-regulated polyunsaturated fatty acid metabolism as a candidate determinant of multiple sclerosis severity and suggest that low-activity FADS haplotypes could contribute to clinically meaningful heterogeneity. However, the findings do not yet justify personalized fatty acid supplementation, dietary prescriptions, or pharmacological targeting of FADS1 or CYP4F2. Mendelian randomization depends on assumptions regarding instrument validity, pleiotropy, linearity, and the comparability of genetic effects with therapeutic intervention; single-variant cis-analyses may also remain vulnerable to linkage disequilibrium and locus complexity. The datasets were largely restricted to individuals of European ancestry, and the metabolomics cohort was modest relative to contemporary genomic studies. Replication in diverse populations, sex- and age-stratified analyses, longitudinal clinical validation, experimental perturbation of the implicated enzymes, and randomized intervention studies will therefore be essential. Despite these constraints, the article substantially narrows the mechanistic search space by identifying a genetically supported connection between polyunsaturated fatty acid regulation, brain-cell FADS1 expression, and the severity of multiple sclerosis.
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:
Noroozi, R., Higgins Tejera, C., Chen, M., Briggs, F. B., Bhargava, P., & Fitzgerald, K. C. (2026). Integrative Genetic Analyses of Lipid Metabolism and Multiple Sclerosis Severity Using Metabolome-Wide and Cis-Mendelian Randomization. medRxiv, 2026-05.
