Decoding Multiple Sclerosis: How Genetics Could Predict Disease Severity and Treatment Response
Multiple sclerosis (MS) has long puzzled both scientists and patients. Why do some people live for decades with only mild symptoms, while others progress rapidly to severe disability? And why do treatments that work wonders for some fall flat for others?
A new study led by researchers from Cardiff University and the University of Groningen offers an intriguing answer: the difference may lie in our genes.
The Challenge of Predicting MS Outcomes
MS is an unpredictable disease. Despite more than 15 approved disease-modifying treatments (DMTs), doctors still struggle to predict how each person’s disease will unfold. Clinical risk factors—like age at onset, sex, or early symptom type—only go so far in explaining outcomes.
Meanwhile, genetic research has uncovered over 230 variants linked to getting MS, and a handful associated with disease severity. But looking at single genes hasn’t given a reliable way to forecast progression.
This study takes a fresh approach: instead of focusing on individual risk variants, the team combined them into broader “genomic risk scores” and then grouped patients with similar genetic profiles using machine learning.
Three Genetic Clusters, Three Disease Trajectories
The researchers analyzed data from 1,455 Welsh MS patients and validated their results in 272 post-mortem cases from the Netherlands Brain Bank.
By applying unsupervised clustering, they identified three distinct genetic subtypes of MS:
Cluster 1: Carried the highest burden of HLA-related risk genes but surprisingly showed the slowest disease progression. Patients in this group reached disability milestones much later than others.
Cluster 2: Had more genetic variants tied to faster progression. These patients experienced earlier disability, more severe symptoms like swallowing difficulties and spasticity, and faster lesion growth on MRI. However, they also responded the best to MS treatments.
Cluster 3: Shared some features with Cluster 2, showing faster progression than Cluster 1, but without the same strong treatment response.
Genetics as an Independent Predictor
One striking finding: genetics predicted outcomes beyond traditional clinical markers. For example, patients who would normally be expected to do well—like those presenting with optic neuritis—still had a higher risk of disability if they belonged to Cluster 2 or 3.
This suggests that genetic testing could one day be a valuable addition to early prognostic assessments, helping patients and doctors make more personalized treatment decisions.
Treatment Response: A Silver Lining
Perhaps the most hopeful discovery is that Cluster 2 patients—despite having more aggressive disease—benefited the most from DMTs. In other words, while their genetic makeup predisposes them to faster decline, modern therapies can substantially slow the process if used appropriately.
This opens the door to a precision medicine approach: identifying which genetic subtype a patient belongs to, then tailoring treatment intensity from the start.
Why This Matters
If validated in larger studies, this genetic clustering approach could transform MS care by:
Offering patients a clearer picture of their likely disease course.
Guiding neurologists toward the right treatment strategy earlier.
Providing insights into the biology of MS progression, which could fuel new drug development.
Caveats and Next Steps
Like all good science, this study raises new questions. The researchers note that their sample sizes, particularly in the replication cohort, were relatively small. They also couldn’t yet tease apart how specific drugs interact with genetic profiles.
Future research will need larger, trial-based cohorts and more detailed genetic analysis to confirm which variants drive these effects.
Still, the message is clear: MS is not one disease, but many—shaped by our genes. And understanding those differences could finally give doctors the tools they need to predict, and change, the course of this complex condition.
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:
Kreft, K. L., Mekkes, N. J., Uzochukwu, E., Loveless, S., Wynford-Thomas, R., Harding, K. E., ... & Robertson, N. P. (2025). Genetic subtypes predict multiple sclerosis severity and response to treatment. medRxiv, 2025-04.
