From Inflammation to Metabolism: Genetic Insights into Multiple Sclerosis
Multiple sclerosis (MS) is a complex immune-mediated and neurodegenerative disorder characterized by immune-cell infiltration into the central nervous system, demyelination, axonal injury, and persistent neuroinflammation. Although magnetic resonance imaging, neurological examination, and cerebrospinal fluid analysis contribute to diagnosis, no single blood- or cerebrospinal fluid–based biomarker currently provides sufficient disease specificity. This diagnostic uncertainty is clinically important because earlier initiation of effective disease-modifying therapy is associated with better long-term outcomes. The study by Mao and colleagues addresses this unmet need by investigating whether circulating inflammatory proteins and cerebrospinal fluid metabolites are causally involved in MS rather than merely changing as secondary consequences of the disease. The researchers further examined whether alterations in cerebrospinal fluid metabolism constitute an intermediate biological mechanism through which peripheral inflammation influences MS susceptibility.
Integrating Inflammatory Proteomics, Metabolomics, and Human Genetics
The investigation combined genome-wide association study data for three interconnected biological domains: 91 circulating inflammatory proteins, 338 cerebrospinal fluid metabolites, and MS susceptibility. The MS dataset was derived from the International Multiple Sclerosis Genetics Consortium and included 47,429 cases and 68,374 controls of European ancestry. Genetic associations with inflammatory protein concentrations were obtained from 14,824 participants assessed using the Olink Target Inflammation Panel, whereas cerebrospinal fluid metabolite data originated from a previously published metabolomic GWAS. As illustrated by the workflow diagram on page 2, the analysis consisted of three stages: bidirectional testing between inflammatory proteins and MS, bidirectional testing between cerebrospinal fluid metabolites and MS, and mediation analysis linking proteins, metabolites, and disease. This design allowed the investigators to distinguish direct causal associations from a hypothetical sequential pathway in which systemic inflammatory activity first alters central nervous system metabolism and subsequently modifies MS risk.
Mendelian Randomization as a Tool for Causal Inference
Mendelian randomization uses genetic variants associated with an exposure as instrumental variables to estimate the effect of that exposure on an outcome. Because genetic variants are assigned at conception, this approach can reduce—but not completely eliminate—the confounding and reverse-causation problems that affect conventional observational studies. The authors selected independent single-nucleotide polymorphisms associated with inflammatory proteins or cerebrospinal fluid metabolites and used inverse-variance weighting as the principal analytical method. MR-Egger regression, MR-PRESSO, Cochran’s Q testing, leave-one-out analysis, and Steiger directionality testing were applied to evaluate horizontal pleiotropy, heterogeneity, influential variants, and the direction of causation. Nevertheless, the instruments were selected using relatively permissive association thresholds of P < 1×10−5 P < 1×10−5 for inflammatory proteins and P < 5×10−5 P < 5×10−5 for metabolites. These thresholds increased the number of available instruments but may also have introduced weaker genetic proxies, making replication and instrument-strength evaluation essential when interpreting the findings.
Inflammatory Proteins Associated with MS Susceptibility
Five circulating inflammatory proteins showed evidence of a causal relationship with MS. Genetically predicted higher concentrations of C-C motif chemokine 4, interleukin-17C, and leukemia inhibitory factor receptor were associated with increased disease risk. The estimated odds ratios were 1.126, 1.189, and 1.211, respectively, indicating moderate increases in the odds of MS per genetically predicted increment in each protein. Conversely, higher genetically predicted levels of C-X-C motif chemokine 6 and neurturin were associated with lower risk, with odds ratios of 0.530 and 0.745. These results are biologically noteworthy because chemokines regulate leukocyte migration, IL-17 family cytokines participate in Th17-mediated inflammatory responses, and neurturin is a neurotrophic factor involved in neuronal survival. The leukemia inhibitory factor receptor finding is particularly complex because experimental literature has often assigned protective or immunoregulatory functions to leukemia inhibitory factor signaling. Moreover, the discussion section appears to contain inconsistencies in the naming and direction of certain protein associations; therefore, the numerical results section should be regarded as the more reliable account of the study’s primary findings.
A Distinct Cerebrospinal Fluid Metabolic Signature
The metabolomic analysis identified ten cerebrospinal fluid metabolites associated with MS. Six were linked to increased susceptibility: 1-stearoyl-2-arachidonoyl-glycerophosphocholine, allantoin, N-acetylglutamate, 1-palmitoyl-2-stearoyl-glycerophosphocholine, N-acetylaspartate, and 1-stearoyl-glycerophosphocholine. Four metabolites—N,N,N-trimethyl-alanylproline betaine, N-acetylglycine, N-acetyl-aspartyl-glutamate, and N6-methyllysine—were associated with reduced risk. The forest plot on page 4 demonstrates that most effect sizes were modest, with odds ratios relatively close to 1, but their biological distribution is informative. Several risk-associated molecules are glycerophosphocholines, suggesting possible involvement of membrane lipid turnover, myelin biology, or inflammatory lipid signaling. N-acetylaspartate and N-acetyl-aspartyl-glutamate are especially relevant to neuronal and axonal physiology. However, the positive association between genetically predicted N-acetylaspartate and MS risk differs from observational studies reporting reduced concentrations during established disease. This apparent discrepancy may reflect differences between lifelong genetically influenced metabolite levels and metabolite depletion occurring after axonal damage, emphasizing that disease susceptibility biomarkers and disease-progression biomarkers are not necessarily equivalent.
Why the Proposed Metabolic Mediation Pathway Was Not Supported
Although both inflammatory proteins and cerebrospinal fluid metabolites were individually associated with MS, the mediation analysis did not support a sequential pathway from peripheral inflammation to altered cerebrospinal fluid metabolism and then to disease. For mediation to be demonstrated, the inflammatory protein must causally influence the proposed metabolite mediator, and the metabolite must subsequently influence MS. The study found no convincing causal association between the relevant inflammatory proteins and cerebrospinal fluid metabolites; consequently, the essential first component of the mediation pathway was absent. Reverse-direction analysis also found no evidence that MS causally altered the identified cerebrospinal fluid metabolites, although MS was associated with modest changes in several inflammatory proteins, including artemin, monocyte chemoattractant protein-4, and vascular endothelial growth factor A. Sensitivity analyses did not detect significant horizontal pleiotropy or heterogeneity, and leave-one-out analyses suggested that the principal estimates were not driven by a single genetic variant. The findings therefore support parallel protein-related and metabolite-related pathways rather than a simple linear inflammatory protein–metabolite–MS cascade.
Scientific Significance, Limitations, and Future Directions
This study expands the molecular landscape of MS by highlighting potential contributions from chemokine signaling, IL-17 biology, neurotrophic pathways, membrane-associated lipids, amino-acid derivatives, and neuronal metabolites. Nevertheless, the results should be interpreted as hypothesis-generating rather than immediately clinically actionable. The reported probability values for the highlighted associations do not appear to reach strict Bonferroni thresholds calculated across all 91 proteins or 338 metabolites, meaning that many findings would be classified as suggestive under the authors’ stated framework. Additional limitations include restriction to participants of European ancestry, reliance on summary-level genetic data, potential weak-instrument bias, incomplete assessment of tissue-specific protein activity, and the possibility that circulating concentrations do not accurately represent signaling within the central nervous system. Mendelian randomization also estimates the consequences of lifelong genetically influenced exposure and cannot directly predict the effects of pharmacologically increasing or suppressing a biomolecule over a shorter period. The most valuable outcome of the study is therefore a prioritized set of molecular candidates for replication in independent genetic datasets, longitudinal patient cohorts, experimental models, and integrated blood–cerebrospinal fluid multi-omics studies.
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:
Mao, X., Lu, X., Liu, Y., Wu, H., Li, B., & Bi, X. (2025). Exploring the mediating role of cerebrospinal fluid metabolites in the pathway from circulating inflammatory proteins to multiple sclerosis: A Mendelian randomization study. Multiple Sclerosis and Related Disorders, 98, 106440.
