The challenge
Cerebrospinal fluid is central to neurological diagnostics, but its biomarker potential is still underused in routine care. In multiple sclerosis, diagnosis still relies heavily on clinical findings, MRI, and oligoclonal bands, yet misdiagnosis and delayed diagnosis remain common. Earlier mass-spectrometry based CSF biomarker studies often examined too few samples, measured too few proteins, or used control groups that did not reflect the real clinical differential-diagnosis problem. Another major obstacle is that biological confounders such as blood-CSF barrier impairment, age, and sex can strongly shape the CSF proteome and influence mass-spectrometry based proteomics.
Our approach
The team established a scalable mass-spectrometry workflow for CSF proteomics and applied it across a very large neurological cohort. They then used improved instrumentation, machine learning, and targeted validation to focus on multiple sclerosis, derive a clinically relevant protein panel, and test it in an independent replication cohort.
Our findings
The study showed that blood-CSF barrier impairment is a dominant source of variation in CSF mass-spectrometry based proteomics and must be accounted for when searching for disease biomarkers. After correcting for these effects, the authors identified shared and disease-specific protein signatures across neurological disorders. In MS, they derived a 22-protein panel that matched standard markers overall but outperformed them in diagnostically difficult oligoclonal band-negative cases. They also translated the discovery into a targeted assay and showed that proteome-based staging correlates with disability, progression, and conversion along the relapsing-to-progressive MS spectrum.
The implications
This work moves CSF proteomics closer to routine clinical use by linking large-scale discovery, biomarker selection, assay translation, and replication in one study. It could improve MS diagnosis and help identify patients at risk of progression earlier.
Creating SyNergies
This study brought together large-scale proteomics, computational biology, and clinical neurology within the Munich systems-neurology environment. SyNergy members Fabian J. Theis contributed computational supervision and data-analysis expertise, while Bernhard Hemmer co-led the study’s clinical and translational MS focus, including cohort design and interpretation. Their collaboration helped connect biomarker discovery with clinically meaningful validation.
