Decoding Multiple Sclerosis Genetics Through Shared Autoimmune Signals: Fourteen New Risk Loci from a Correlation-Informed Meta-analysis
Multiple sclerosis (MS) is characterized by inflammatory damage to oligodendrocytes and myelin in the central nervous system, and genetic evidence increasingly supports an autoimmune (rather than purely neurodegenerative) etiology, given that many implicated loci map to immune pathways and overlap with other autoimmune diseases (ADs). Building on this premise, Olafsson and colleagues leverage the observation that ADs share components of polygenic architecture—sometimes with concordant effects, sometimes antagonistic—to increase discovery power beyond conventional single-trait genome-wide association studies (GWAS).
Study Design: Three-Phase Strategy to Increase Statistical Power
The authors implement a three-step approach: (i) a meta-analysis combining public MS summary statistics with three Nordic cohorts to discover new MS-associated variants, (ii) polygenic risk score (PRS) analyses across multiple ADs to quantify genetic overlap while excluding the major histocompatibility complex (MHC) region due to its unusually complex linkage disequilibrium (LD), and (iii) a proxy-phenotype strategy in which primary biliary cirrhosis (PBC)—identified as strongly genetically correlated with MS—is used to prioritize variants for targeted MS association testing. Across Nordic cohorts and public resources, the combined case–control scale is substantial (reported as 21,079 cases and 371,198 controls), enabling detection of associations that previously fell short of stringent GWAS thresholds.
Meta-analysis Findings: Seven Previously Unreported MS-Associating Variants
In the initial meta-analytic phase, the investigators perform inverse-variance weighted analyses that (a) maximize power for Immunochip-covered loci and (b) broaden genome-wide coverage through imputed data in Nordic cohorts. Excluding the MHC region, they identify seven association signals not previously reported at genome-wide significance, including a notable coding variant in MTHFR (rs1801133; A222V missense) with a protective effect (odds ratio approximately 0.88) and several intronic/intergenic signals near immune-relevant or regulatory genes (e.g., near CD6/CD5, ETS1, and loci containing immune-related noncoding elements). The paper further contextualizes these loci with functional annotation efforts (eQTL lookups and regulatory annotations), although direct, high-confidence transcriptional links in whole blood are limited for most variants.
Polygenic Risk Profiling: Mapping Shared Architecture Across Autoimmune Diseases
A central contribution is the systematic PRS-based comparison across ten autoimmune phenotypes (plus asthma as a non-autoimmune inflammatory comparator), using public summary statistics and testing scores in a large Icelandic target set while excluding the extended MHC region to avoid spurious pleiotropy driven by dense LD. The results recapitulate a biologically intuitive partition of ADs into clusters that broadly correspond to autoantibody presence: “seronegative” diseases (e.g., Crohn’s disease, ulcerative colitis, psoriasis, ankylosing spondylitis) and “seropositive” diseases (e.g., rheumatoid arthritis, type 1 diabetes, systemic lupus erythematosus, autoimmune thyroiditis). Notably, MS and PBC do not fit cleanly into either cluster, and the MS PRS shows its strongest cross-disease association with PBC, highlighting a distinctive shared genetic component between these two conditions.
Proxy-Phenotype Innovation: Using PBC to Prioritize MS Risk Variants
Motivated by the strong MS–PBC PRS reciprocity (PBC-PRS increasing MS risk and MS-PRS increasing PBC risk), the authors treat PBC as a proxy phenotype for MS, reasoning that variants contributing to PBC polygenic liability have an elevated prior probability of influencing MS. Rather than restricting to only genome-wide significant PBC loci (as in classic proxy-phenotype designs), they test the set of variants forming the most predictive PBC PRS and apply multiple-testing correction across the screened markers. This strategy yields seven additional MS-associated signals not explained by previously established MS loci, demonstrating how correlation-informed priors can expose pleiotropic risk mechanisms that are difficult to capture through single-phenotype discovery alone.
Mechanistic Leads: Cytokine Signaling, TYK2, and One-Carbon Metabolism
Among the newly implicated variants, two coding changes stand out for biological interpretability. First, a low-frequency missense variant in TYK2 (rs35018800; A928V) shows a relatively large protective effect on MS risk (odds ratio ~0.68 in the MS meta-analysis), strengthening the case for JAK/STAT-linked cytokine signaling as a core axis in autoimmune susceptibility. Second, MTHFR rs1801133 (A222V) is a function-reducing variant previously associated with elevated homocysteine and altered folate/one-carbon metabolism; the authors connect this to longstanding hypotheses involving vitamin B12–dependent methylation and metabolic dysregulation observed in MS tissues. Beyond coding variation, several proxy-phenotype hits localize to genes plausibly relevant to T-cell differentiation and interferon/IL-12 pathways (e.g., IL12RB2, TXK, proximity to IRF5), providing specific hypotheses for downstream functional follow-up.
Interpretation and Limitations: What These Results Do—and Do Not—Establish
The study illustrates a pragmatic route to incremental locus discovery in highly polygenic immune traits: combine large-scale meta-analysis with cross-trait PRS mapping and then formalize genetic overlap into a proxy-phenotype prior that reduces the effective multiple-testing burden. At the same time, the authors appropriately caution that some signals do not reach the strictest “genome-wide” thresholds and that interpretation is complicated by factors such as Immunochip ascertainment, possible inflation in some contributing summary statistics, and partial control overlap between studies used for PRS derivation and MS meta-analysis. Moreover, PRS correlations primarily reflect shared inherited susceptibility to disease onset and, consistent with their analyses, do not necessarily translate into determinants of clinical severity (e.g., limited association with the Multiple Sclerosis Severity Score). Taken together, the work provides 14 candidate loci that refine the genetic map of MS, strengthen mechanistic emphasis on immune signaling and immunometabolism, and exemplify how pleiotropy-aware designs can accelerate discovery in complex disease genomics.
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:
Olafsson, S., Stridh, P., Bos, S.D. et al. Fourteen sequence variants that associate with multiple sclerosis discovered by meta-analysis informed by genetic correlations. npj Genomic Med 2, 24 (2017). https://doi.org/10.1038/s41525-017-0027-2
