Proteomic Signatures of Conversion Risk and Disease Severity in Multiple Sclerosis

Abstract Despite important advances in understanding the etiopathology of multiple sclerosis, factors determining disease progression remain partially understood and often difficult to predict. Specific diagnostic and prognostic biomarkers are needed to optimize the risk-benefit ratio of treatment for each patient. The aim of our study was to identify a cerebrospinal fluid proteomic signature associated with diagnosis and short- to mid-term prognosis across the multiple sclerosis continuum. Our multicentric cohort study analyzed CSF samples from 120 patients using a proteomics data-independent acquisition strategy. Differentially expressed proteins were identified across diagnostic groups: 62 patients with multiple sclerosis, 15 patients with clinically isolated syndrome, and 43 healthy controls. We also compared the CSF of patients with no evidence of disease activity with those with disease activity at 2 and 5 years of follow-up. A diagnostic and prognostic classification model was built using iterative cross-validated logistic regression models on shared differentially expressed proteins across these two comparisons. A total of 1,257 proteins were quantified, and 162 differentially expressed proteins were identified across comparisons. We identified a set of ten proteins associated with the diagnosis and prognosis of multiple sclerosis, including previously identified potential biomarkers (CH3L2, IGHG1, IGKC, LAMP2, ADA2), proteins known to be involved in the pathophysiology of multiple sclerosis (A0A8J8YUT9, AT2A2, CO3A1) and two yet unreported proteins (DSC2 and MMRN2). Multivariate models based on these proteins achieved good accuracy for the diagnosis of MS compared with CIS (area under the receiver operating characteristics curve [AUROC] up to 80% using 3 proteins) and prognosis (NEDA vs. EDA; AUROC up to 96% at 2 and 5 years; using 5 proteins). These results, which will require further investigation to validate the new biomarkers, open new perspectives on multiple sclerosis pathophysiology and therapeutic targets.