Do Multiple Sclerosis Risk Genes Also Shape Disease Severity? Evidence from 7,125 Patients Across 10 Cohorts
Multiple sclerosis (MS) is a chronic inflammatory disease of the central nervous system characterized by demyelination, axonal injury, and accumulating neurologic disability. While genome-wide association studies (GWAS) have been highly productive for identifying genetic determinants of MS susceptibility, much less is known about whether those same loci also influence how severe MS becomes over time. The article “Multiple sclerosis risk loci and disease severity in 7,125 individuals from 10 studies” addresses this translational gap by testing whether 52 established MS risk variants—robustly associated with disease risk—also correlate with clinical severity in a large international sample.
Why severity genetics is difficult in MS
The clinical course of MS is heterogeneous: some individuals have relatively mild disability for decades, while others progress rapidly. Established susceptibility factors, including the strong major histocompatibility complex (MHC) association at HLA-DRB1*15:01, clearly increase the probability of developing MS, yet prior studies have not convincingly linked these risk alleles to progression or disability accumulation. In parallel, environmental exposures (e.g., Epstein–Barr virus infection, smoking, low vitamin D, and adolescent obesity) are well supported for MS risk, but their relationship to severity is less established. This context motivates a direct, well-powered test of whether the genetic architecture of susceptibility overlaps meaningfully with that of severity.
Cohorts and phenotype: operationalizing “severity” with MSSS
The investigators aggregated 10 independent MS cohorts totaling 7,125 cases, spanning multiple countries and genotyping platforms, and quantified severity using the Multiple Sclerosis Severity Score (MSSS). MSSS is derived from the Expanded Disability Status Scale (EDSS) while incorporating disease duration, with the intention of providing a time-adjusted measure of disability. The study analyzed MSSS (i) as a continuous trait, (ii) as a median split (≤5 vs >5), and (iii) as an “extremes” comparison (< 2.5 vs ≥7.5) to reduce phenotype ambiguity by excluding intermediate values. Importantly, MSSS values were assigned using a large independent “Global MSSS” reference dataset to improve stability across cohorts.
Genetic predictors: from single variants to polygenic risk scores
The genetic exposure of interest comprised 52 non-MHC GWAS variants previously replicated for MS susceptibility. Each variant was tested individually, and also aggregated into two summary measures: an unweighted genetic risk score (GRS) (simple allele count) and a weighted genetic risk score (wGRS) that applied published discovery odds ratios as variant-specific weights. Because cohorts differed in genotyping coverage, the study used validated tagging SNPs for key markers (including for HLA-DRB1*15:01 where classical typing was unavailable) and performed imputation for missing variants in some datasets, followed by careful handling of remaining missingness. This design allowed both locus-level and polygenic-level hypotheses to be evaluated under a consistent analytic framework.
Core finding: susceptibility burden does not predict MSSS severity
Across random-effects meta-analyses adjusted for cohort, sex, age at onset, and HLA-DRB115:01, neither wGRS nor GRS showed a robust association with MSSS once key confounders were considered and analyses were restricted to individuals with disease duration ≥10 years (a step intended to improve MSSS stability). In contrast, male sex and later age at onset were consistently associated with higher MSSS (more severe disease) across models, whereas **HLA-DRB115:01 status**—despite being a major susceptibility locus—was not associated with severity in this study. Taken together, these results argue against a substantial overlap between the strongest known susceptibility alleles and disability severity as measured by MSSS.
Single-variant analysis: nominal signals fail multiple-testing correction
When examining the 52 variants individually, the authors observed limited nominal evidence for association at a small number of loci (e.g., rs874628 within/near MPV17L2 and rs650258 upstream of CD6) across MSSS outcomes; however, no SNP remained significant after adjusting for multiple comparisons. The use of random-effects meta-analysis was methodologically appropriate given differences in cohort composition and ascertainment, but the overall pattern—weak, non-replicating signals that do not survive correction—supports the interpretation that established risk loci with comparatively larger effects on susceptibility do not translate into detectable effects on MSSS-defined severity in a sample of this size.
Interpretation, limitations, and the path forward
A central implication is that MS may have partially distinct genetic architectures for initiating disease versus driving progression and disability. The paper also highlights limitations that plausibly dilute true severity associations: MSSS depends on EDSS, which is heavily weighted toward ambulatory function and may be insensitive to cognitive impairment or other domains; different combinations of EDSS and disease duration can yield similar MSSS values; disease duration is anchored to symptom onset (which may lag subclinical disease); and the study lacked longitudinal progression trajectories, relapse metrics, treatment histories, and high-resolution MRI phenotypes that could provide more biologically proximal outcomes. The authors therefore advocate larger, prospective cohorts with sensitive clinical endpoints and quantitative MRI measures—potentially enabling GWAS or sequencing studies focused specifically on progression biology rather than susceptibility alone.
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:
George, M. F., Briggs, F. B., Shao, X., Gianfrancesco, M. A., Kockum, I., Harbo, H. F., Celius, E. G., Bos, S. D., Hedström, A., Shen, L., Bernstein, A., Alfredsson, L., Hillert, J., Olsson, T., Patsopoulos, N. A., De Jager, P. L., Oturai, A. B., Søndergaard, H. B., Sellebjerg, F., Sorensen, P. S., … Barcellos, L. F. (2016). Multiple sclerosis risk loci and disease severity in 7,125 individuals from 10 studies. Neurology. Genetics, 2(4), e87. https://doi.org/10.1212/NXG.0000000000000087
