Does Aquaporin-4–Seronegative Neuromyelitis Optica Exist?

Abstract

With the improvement in anti–aquaporin-4 (AQP4) assay sensitivity to near 90%,1 the concept of seronegative neuromyelitis optica (NMO) is being challenged. The widely accepted criteria forNMOproposed2 in 1999 and revised3 in 2006define NMO as a disease localized to the optic nerves and spinal cord, which is not multiple sclerosis (MS). The antiAQP4 antibody has always been a very highly specific biomarker forNMObut, owing to insensitive assays, seropositivity was initially considered only a supportive criterion. Seronegative NMO was intended to represent the same diseaseprocessas seropositiveNMObut, forwhatever reason, the anti-AQP4antibodywasnotdetectable at the time.During the past 8 years since the anti-AQP4 antibody biomarker was discovered and linked toNMO, the research field has focused intensely on the role of AQP4 as an immunopathogenic target.4 Supportedby incremental advancements in theanti-AQP4biomarker assay, theunderstanding of trueNMO is expanding its localization to brainstem and cortical disease, while narrowing its population base to those who are seropositive for the anti-AQP4antibody.5That leavesneuroimmunologistswith the question of whether a seronegative NMO disease actually exists? The problem with seronegative NMO is that it is not well characterized. The scientific purists prefer to keep this group confined to those whose disease is clinically, radiographically, and pathologically identical to seropositive NMO. Their concern is that contaminating the cohort of patientswith true NMOwill make the diseasemore difficult to study. However, the seasoned clinicians prefer to expand the definition of seronegative NMO to include all patients who have recurrent non-MS inflammatory disease of the central nervous system that should be treated empirically like NMO with immunosuppression. They fear that neglecting seronegativeNMOwill lead tomistreatmentwithMSmedications that couldmake the disease worse. In this issueof JAMANeurology, theNMOexpertsatOxford University, including scientificpuristsWaters andVincentplus a solid groupof seasoned clinicians ledbyPalace, are attempting to tackle the problem of seronegative NMOby identifying a subpopulation that tests positive for antibodies againstmyelin-oligodendrocyte glycoprotein (MOG).6 Anti-MOG– positive patients with clinical features of NMO (actually almosthalfof themmeet the2006clinical criteria) appear tohave aslightlydifferentdiseasephenotype,withademographicand clinical profile more similar to patients with acute demyelinating encephalomyelitis (ADEM). While older females dominated the AQP4-seropositive group, MOG seropositivity was more common in youngmales; episodes weremore severe in the MOG-positive group but these patients healed better and weremore likely to bemonophasic comparedwith thosewith AQP4-positiveNMO. Therewere no patientswho tested positive for both anti-AQP4 and anti-MOG antibodies, suggesting theymightbedistinct immunopathologicdiseasesdespite their similar presentations. As this study shows, there may be diseases that present like NMO with spinal cord and optic nerve inflammation but immunologicallymaybe related toanotherdisease likeADEM. In the case of ADEM, the anti-MOG antibody has previously been associatedwith up to 40%of cases.7 Other diseases that present in a similar fashion to NMO include MS, systemic lupuserythematosis, Sjogren syndrome,neurosarcoidosis, varicella zoster infection, andparaneoplasticdiseases suchasantiMa2 encephalomyelitis. Workup for these conditions should be considered inpatientswith suspectedNMOwho test negative for the anti-AQP4 antibody, and this study smartly recommends testing for the anti-MOG antibodies in these situations too (currently available at theOxfordNeuroimmunology Testing Service, Oxford, England, for a fair £30). There may be patients with seronegative NMO who truly have NMO. These patients may have the same immunopathologic disease process but they test seronegative either because their titer is below the detection limit or because they do not yet have circulating anti-AQP4 antibodies. The latter presents an interesting group of patients to study because understanding why, when, and how patients develop anti-AQP4 antibodies could explain the early steps of NMO disease and provide drug-treatable targets for treatment. A third possible explanation for why some patients with true NMO are seronegative is because they have a slightly different immune process that ultimately causes the same downstream cascade of inflammatory events. Consider the paradigm of myasthenia gravis, where 38 years after discovery of the anti-acetylcholine receptor (AChR) antibody, a subset of seronegative patients were found to have autoantibodies against the muscle-specific receptor tyrosine kinase (MuSK) causing a phenotypically similar disease by involving a target physically linked to AChRs.8 A newly discovered myasthenia autoantibody against lipoprotein-related protein 4, another receptor in the AChR/MuSK complex, may further reduce the prevalence of seronegative myasthenia.9 Aquaporin-4 is a membrane protein linked to several other possible immune targets on the astrocyte membrane.10 Recently, the inward rectifying potassium channel 4.1 (Kir4.1) has been implicated in MS in a study showing that 46.9% of patients tested positive for anti-Kir4.1 antibodies. Patients with Kir4.1-seropositive MS share some pathologic Related article page 276 Opinion