Reading the Blood for Clues: How Metabolites Reveal the Hidden Severity of Multiple Sclerosis
Multiple sclerosis (MS) is usually described in terms of MRI scans, immune cells, and demyelinated lesions—but underneath all of that biology sits chemistry. Every cell in an inflamed brain is constantly transforming nutrients into energy, lipids, and signaling molecules. Those small molecules, collectively called the metabolome, can leak clues into the bloodstream. In this paper, Villoslada and colleagues asked a simple but powerful question: can the pattern of metabolites in blood tell us who has MS and how severe their disease is? Using high-resolution mass spectrometry, they mapped hundreds of lipids and amino acids in the serum of people with MS and healthy controls, then linked those chemical fingerprints to disease activity and disability over time.
Two cohorts, one question: can blood chemistry reflect disease severity?
To tackle this, the team assembled two independent Spanish cohorts. The first was a large, retrospective longitudinal cohort: 238 people with MS and 74 healthy controls from several hospitals, followed clinically for two years. These patients represented a typical MS mix—clinically isolated syndrome, relapsing–remitting MS, and some progressive cases—with standardized disability scoring using the Expanded Disability Status Scale (EDSS). The second cohort was smaller but much richer in time series: 61 MS patients and 41 healthy controls with serial serum samples every three months over two years, plus parallel clinical follow-up. In both cohorts, blood was drawn in the morning after overnight fasting, then analyzed with ultra-high-performance liquid chromatography coupled to mass spectrometry (UHPLC-MS), using platforms optimized for lipids and amino acids. This design let them ask both cross-sectional and longitudinal questions: who has MS, and whose disease is getting worse?
A metabolomic signature that separates MS from healthy controls
The first result was reassuringly binary: people with MS do not look metabolically “normal” in serum. Using multivariate analysis (principal component analysis and partial least squares discriminant analysis), the authors derived metabolomic signatures that distinguished MS patients from healthy controls with good accuracy in both cohorts. In the time-series cohort, the separation held up across multiple sampling time points, suggesting that the “MS fingerprint” in blood is relatively stable over two years rather than just a noisy snapshot. When they moved to the larger validation cohort, they could identify specific metabolite classes driving this separation: shifts in sphingomyelins, phosphatidylethanolamines, phosphatidylcholines, cholesteryl esters, ceramides, and related lipids. Together, these findings argue that MS is not only an immune disease of the CNS—it is also a systemic metabolic state that leaves a trace in peripheral blood.
Metabolites that track relapses and disability
Beyond diagnosis, the clinically important question is prognosis: who will remain stable and who will accumulate disability? The authors looked at two dimensions of severity. First, they classified patients as having “stable” or “active” disease over two years, based on relapses and confirmed EDSS worsening. Second, they compared patients who remained at low disability (EDSS < 3) with those who progressed to more severe disability (EDSS > 4.5). In the retrospective cohort, they found over a hundred metabolites that differed between stable and active disease at a nominal level, and a smaller subset survived correction for multiple testing. These included inflammatory lipids such as arachidonic acid and oxidized derivatives like 13-hydroxyoctadecadienoic acid (13-HODE), as well as several lysophosphatidylcholines (LysoPC species). For disability progression, the signature pulled in a somewhat different set: cortisol (hydrocortisone), glutamic acid, tryptophan, eicosapentaenoic acid (EPA), 13-HODE, and several lyso-phospholipids (LysoPC and LysoPE). The predictive performance was modest (AUC around 0.6, with high specificity but low sensitivity), but the key point is that disease activity and disability are mirrored—at least partially—in serum chemistry.
What these lipids and amino acids are telling us about MS biology
The metabolite list is more than a statistical curiosity; it actually makes biological sense. Many of the altered lipids are central to myelin structure and immune signaling. Sphingomyelins and ceramides are major components of myelin and precursors to sphingosine-1-phosphate, a signaling lipid that controls lymphocyte trafficking and is already a drug target in MS (e.g., fingolimod). Phosphatidylethanolamines and phosphatidylcholines are core membrane phospholipids; their balance regulates cell proliferation, apoptosis, and membrane dynamics. Lysophosphatidylcholines are particularly interesting: they can act as potent inflammatory mediators and are capable of inducing demyelination in experimental models, fitting well with their association with active disease and disability. On the amino acid side, glutamate is a classic player in excitotoxicity and oligodendrocyte damage, while tryptophan sits at a critical checkpoint in immune regulation via the indoleamine 2,3-dioxygenase (IDO) pathway. Cortisol reflects hypothalamic–pituitary–adrenal axis activity and chronic stress/inflammation. Put together, the pattern points to a re-wiring of lipid and amino acid metabolism that unites immune activation, myelin biology, and neuronal stress.
MS as a systemic metabolic reprogramming, not just a brain-only disease
One striking theme from this work is that MS leaves a consistent, systemic metabolic imprint, not just local changes in brain tissue. Previous MR spectroscopy studies had already shown altered brain metabolites in MS (e.g., changes in choline, N-acetylaspartate, and glutamate), and phospholipid/sphingolipid imbalances in CNS tissue have been reported in autopsy studies. This serum metabolomics study echoes that story at the systemic level: relative increases in certain phospholipids and decreases in sphingolipids suggest that lipid synthesis and turnover are globally shifted, perhaps as a consequence of chronic demyelination, gliosis, and ongoing immune activation. The authors even speculate that increased phospholipids and decreased sphingolipids might reflect a kind of adaptive remodeling of membranes and myelin, or a compensatory response to inflammatory damage. For clinicians and researchers, the takeaway is that MS should be thought of as a whole-body metabolic state, not just a focal demyelinating process seen on MRI.
Limitations and why this isn’t a ready-to-use blood test (yet)
Despite its appeal, this work is not ready to be translated into a clinical diagnostic kit. The first cohort used a platform that detected spectral peaks without full metabolite identification, meaning that some features remain anonymous. Treatment is a potential confounder: most patients in the retrospective cohort were on interferon-β, and while that reduces heterogeneity somewhat, it also makes it hard to fully disentangle drug effects from disease biology. Disease activity was defined clinically without systematic MRI, so “active” disease may be under- or over-called relative to modern imaging-based standards. Statistical performance, especially for predicting disability worsening, is significant but not strong; an AUC of ~0.6 with low sensitivity is not enough for individualized prediction. And like all omics studies, there is the risk of overfitting in complex multivariate models, even with cross-validation and external cohort testing. So this should be viewed as a proof-of-concept map of promising metabolic pathways, not a finished biomarker panel.
What this means for the future of MS biomarkers and therapy
Even with these caveats, the study is an important step toward metabolomics-informed MS medicine. It highlights specific lipid and amino acid pathways—sphingolipids, phospholipids, oxidized fatty acids, tryptophan and glutamate metabolism, stress hormones—that are consistently linked to disease presence and severity. Future work can layer these metabolomic signatures with genetics, transcriptomics, and imaging to build more robust predictive models, or focus on refining a small set of highly reproducible metabolites for clinical use. Therapeutically, the data reinforce the idea that targeting lipid metabolism and oxidative pathways (beyond S1P alone) might be fruitful, and that dietary or pharmacologic manipulation of omega-3 fatty acids and tryptophan pathways might influence disease course. For now, the message is clear: if we want to understand and predict MS progression, we should not look only at lesions on MRI, but also at the chemical stories written in the blood.
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:
Villoslada, P., Alonso, C., Agirrezabal, I., Kotelnikova, E., Zubizarreta, I., Pulido-Valdeolivas, I., ... & Castro, A. (2017). Metabolomic signatures associated with disease severity in multiple sclerosis. Neurology: Neuroimmunology & Neuroinflammation, 4(2), e321.
