Lipid Metabolism and Multiple Sclerosis Severity: Genetic Evidence for a Causal Role of FADS1
Multiple sclerosis (MS) is characterized by pronounced heterogeneity in its clinical course: some individuals experience relatively limited neurological impairment, whereas others develop progressive disability despite comparable disease duration. Metabolomic studies have repeatedly identified alterations in circulating lipids, amino acids, and energy-related metabolites among people with MS, but conventional observational analyses cannot readily determine whether these metabolic changes contribute to disease progression or instead arise as consequences of inflammation, disability, medication exposure, diet, or reduced physical activity. The study by Noroozi and colleagues addresses this problem through an integrative genetic framework designed to distinguish correlation from potential causation. Specifically, the investigators combined metabolome-wide Mendelian randomization, genetic colocalization, pathway-focused cis-Mendelian randomization, and brain cell-type-specific expression analyses. Their principal conclusion is that genetically regulated polyunsaturated fatty acid metabolism—particularly reduced activity of the enzyme encoded by FADS1—may contribute to greater MS severity. The analysis also implicates CYP4F2, another lipid-metabolizing gene, and identifies additional signals involving amino acid and mitochondrial pathways.
A Multi-Layered Mendelian Randomization Strategy
The investigation used two-sample Mendelian randomization, an instrumental-variable approach in which inherited genetic variants associated with an exposure are used to estimate that exposure’s effect on an outcome. Because alleles are allocated at conception and remain largely unaffected by later clinical or environmental conditions, Mendelian randomization can reduce—but not completely eliminate—confounding and reverse causation. The exposure data originated from a genome-wide association study of 1,091 plasma metabolites and 309 metabolite ratios measured in 8,299 participants of European ancestry from the Canadian Longitudinal Study on Aging. MS severity was assessed using genome-wide association summary statistics from 12,584 people with MS, with the age-related MS severity score serving as the outcome. This score integrates age with disability measured using the Expanded Disability Status Scale, thereby capturing whether an individual’s neurological impairment is greater or lower than expected for their age. The workflow illustrated on page 5 of the article shows a sequential design: genetic instrument selection, metabolome-wide causal screening, sensitivity analyses, multivariable modeling, colocalization, functional cis-MR, and finally brain single-cell expression analysis.
Metabolome-Wide Signals Involving Lipids and Amino Acids
The primary inverse-variance-weighted analysis identified 45 metabolites with nominal evidence of an effect on MS severity, including 36 annotated metabolites or metabolite ratios and nine uncharacterized compounds. The signals extended across amino acid, lipid, energy, peptide, xenobiotic, and metabolite-ratio categories, illustrating that MS progression may reflect interconnected metabolic processes rather than a single isolated biochemical abnormality. Particularly notable findings included associations involving arginine, betaine, succinoyltaurine, α-ketoglutarate, phospholipids, acylcarnitines, sphingomyelins, and ratios reflecting arachidonic acid metabolism. Higher genetically predicted arginine was associated with greater severity, whereas betaine and succinoyltaurine showed inverse associations in pathway-resolved analyses. Several lipid measures pointed toward altered conversion of linoleic acid and other fatty-acid precursors into longer-chain products such as arachidonic acid. These findings are biologically plausible because polyunsaturated fatty acids contribute to cellular membrane architecture, myelin biology, mitochondrial function, and the synthesis of inflammatory or inflammation-resolving lipid mediators. Nevertheless, nominal metabolome-wide associations alone cannot establish that an individual metabolite is the direct causal agent, especially when genetically and biochemically correlated metabolites share instruments.
Robustness Testing and Multivariable Prioritization
To address the principal vulnerabilities of Mendelian randomization, the investigators applied MR-Egger regression, weighted-median and mode-based methods, Cochran’s Q tests, MR-PRESSO, leave-one-out analyses, Steiger filtering, and reverse-direction Mendelian randomization. Forty of the nominally significant metabolites retained concordant effect directions under MR-Egger analysis, while most showed no strong evidence of heterogeneity. Excluding variants associated with potential MS-severity confounders, including body mass index, smoking, and educational attainment, did not materially change the estimates. Reverse analyses generally did not support the interpretation that greater MS severity caused the observed metabolite differences, although decadienedioic acid was an exception. The authors then performed pathway-specific multivariable Mendelian randomization to estimate the direct effect of each metabolite after accounting for correlated members of the same biochemical class. This analysis prioritized three amino acid-related measures and two fatty-acid-related lipids: arginine and propionylglycine were associated with greater severity, whereas betaine, succinoyltaurine, and arachidonate were associated with lower severity. Of these, arachidonate displayed the strongest statistical signal, focusing subsequent mechanistic analyses on polyunsaturated fatty-acid metabolism.
The FADS1/FADS2 Locus as a Central Metabolic Node
Genetic colocalization was used to determine whether metabolite levels and MS severity were influenced by the same causal variant within a genomic region rather than by different variants that happened to be correlated through linkage disequilibrium. Strong colocalization was detected on chromosome 11 at a region containing FADS1, FADS2, MYRF, and TMEM258. The relevant metabolite ratios reflected conversion between linoleate, oleate or vaccenate, and arachidonate, processes closely connected to the fatty-acid desaturase pathway. Although several genes occupy this highly correlated locus, functional evidence favored FADS1. The authors selected rs174546, a variant in the 3′ untranslated region of FADS1, as a cis-acting genetic instrument. Its T allele is associated with repression of FADS1 expression and reduced Δ5-desaturase activity. Across multiple downstream lipid biomarkers, genetically proxied reduction of FADS1 activity produced a coherent biochemical signature: accumulation of upstream products generated before the FADS1-catalyzed step and depletion of downstream long-chain polyunsaturated fatty acids. Both components of this signature were associated with greater MS severity, providing pathway-level rather than single-metabolite evidence for impaired FADS1-mediated lipid conversion.
Convergence Across Circulation, Brain Cells, and CYP4F2
A particularly important feature of the study is the extension of circulating-metabolite findings into cell-type-specific brain biology. Using rs174546 as an expression quantitative trait locus, the investigators examined FADS1 expression across seven brain cell populations. The severity-associated allele was linked to reduced FADS1 expression in astrocytes, oligodendrocytes, excitatory neurons, and inhibitory neurons. Single-instrument cis-MR estimates indicated that lower expression in each of these populations was associated with greater MS severity, while colocalization supported a shared underlying genetic signal. This convergence suggests that altered PUFA processing may operate within both systemic and central nervous system compartments, potentially affecting glial support, neuronal membrane composition, myelin maintenance, and inflammatory signaling. A second locus implicated CYP4F2, which encodes an enzyme involved in ω-hydroxylation and turnover of long-chain fatty acids and eicosanoids. The missense variant rs2108622 was associated with reduced CYP4F2 protein abundance, changes in downstream metabolites, and greater MS severity. However, the colocalization evidence for CYP4F2 was less definitive than that observed at the FADS locus, making it a supportive rather than primary mechanistic finding.
Translational Significance, Limitations, and Research Priorities
The study advances a biologically coherent model in which inherited variation affecting long-chain polyunsaturated fatty-acid synthesis and degradation contributes to differences in MS severity. This model may eventually support metabolic biomarker development, genetically informed patient stratification, or targeted nutritional and pharmacological interventions. However, the findings do not demonstrate that omega-3 or omega-6 supplementation will reduce disability, nor do they establish that increasing arachidonic acid is universally beneficial. FADS-dependent metabolic effects are pathway-, tissue-, genotype-, and substrate-specific, and circulating lipid measurements may not fully represent lipid handling within the brain. The analysis was also restricted primarily to participants of European ancestry, limiting generalizability across populations with different FADS haplotypes and dietary backgrounds. Additional limitations include the modest size of the metabolomics GWAS, the use of relaxed instrument-selection thresholds, dependence on single-variant cis-MR for key analyses, and the possibility of residual pleiotropy within complex genomic regions. Most importantly, the article is a preprint that has not undergone peer review and should not guide clinical practice. Replication in diverse cohorts, longitudinal metabolomics, experimental perturbation of FADS1 and CYP4F2, and genotype-stratified clinical studies will be essential before therapeutic conclusions can be drawn.
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.
