Résumé
Background: Current neuromyelitis optica (NMO) spectrum disorder criteria stratify patients according to the aquaporin-4 (AQP4) antibody status. Among seronegative NMO (neg-NMO), some patients have antibodies against myelin oligodendrocyte glycoprotein (MOG) (MOG-NMO). Distinct molecular and cellular pathways (T-cells vs. B-cells) could explain clinical, therapeutic and pathological differences reported recently among these 3 groups.Objectives: To evaluate the level of cerebrospinal fluid (CSF) biomarkers in these 3 NMO subtypes as well as in multiple sclerosis, other inflammatory neurological disorders (OIND) and symptomatic controls (SC).Methods: Retrospective multicentric study including 21 MOG-NMO, 21 AQP4-NMO and 18 neg-NMO, 26 relapsing-remitting MS (RRMS), 17 primary progressive MS (PPMS), 25 OIND and 37 SC. We determined CSF levels of IL-1b, IL-6, IL-8, IL-10, IL-17a, CXCL-13, CD27, MCP-1, MMP-9, sICAM-1, sVCAM1 (multiplex ELISA, Thermo Fisher), chitinase 3-like protein 1 (CHI3L1, Microvue) and neurofilament-light chain (NFL, Uman Diagnostics). Descriptive statistical analyses were performed with PRISM.Results: CSF levels of CHI3L1 and MCP-1 were similar in the three NMO groups while other biomarkers were increased in AQP4-NMO compared to MOG-NMO and neg-NMO, except for NFL, with highest levels in neg-NMO. Increased CSF levels of CXCL-13, CD27 and IL-17a discriminated RRMS and AQP4-NMO from MOG-NMO and neg-NMO (p< 0.001, Mann Whitney test). MOG-NMO and neg-NMO were only differentiated by increased NFL in neg-NMO (p< 0.05). IL-6 (p< 0.001), MCP-1 (p< 0.01) and sVCAM-1 (p< 0.01) CSF levels were increased in all NMO groups compared to RRMS and PPMS (p< 0.001).Discussion: The CSF profiles of MOG-NMO and neg-NMO are different from AQP4-NMO, confirming the clinical and pathological differences recently reported. The only significant difference found between MOG-NMO and neg-NMO was a low NFL level in MOG-NMO, suggesting that, in this condition, axons are preserved that could explain the relatively good outcome associated to MOG-NMO patients. Most interestingly, we found an increase of B-cell markers (CXCL-13, CD27) in AQP4-NMO and in MS, but not in MOG-NMO or in neg-NMO, suggesting that B-cell therapies could be less effective in MOG-NMO and neg-NMO.Conclusion: Our study confirms that mechanisms in MOG and AQP4-associated disease are probably different and that B-cell depletion therapies might not be the best treatment in MOG-NMO and neg-NMO.