Abstract

Background Although optic neuritis (ON) is a defining feature of neuromyelitis optica (NMO), appropriate animal models of NMO ON are lacking. Most NMO patients are seropositive for immunoglobulin G autoantibodies (NMO-IgG) against the astrocyte water channel aquaporin-4 (AQP4). Methods Several approaches were tested to develop a robust, passive-transfer mouse model of NMO ON, including NMO-IgG and complement delivery by: (i) retrobulbar infusion; (ii) intravitreal injection; (iii) a single intracranial injection near the optic chiasm; and (iv) 3-days continuous intracranial infusion near the optic chiasm. Results Little ON or retinal pathology was seen using approaches (i) to (iii). Using approach (iv), however, optic nerves showed characteristic NMO pathology, with loss of AQP4 and glial fibrillary acidic protein immunoreactivity, granulocyte and macrophage infiltration, deposition of activated complement, demyelination and axonal injury. Even more extensive pathology was created in mice lacking complement inhibitor protein CD59, or using a genetically modified NMO-IgG with enhanced complement effector function, including significant loss of retinal ganglion cells. In control studies, optic nerve pathology was absent in treated AQP4-deficient mice, or in wild-type mice receiving control (non-NMO) IgG and complement. Conclusion Passive transfer of NMO-IgG and complement by continuous infusion near the optic chiasm in mice is sufficient to produce ON with characteristic NMO pathology. The mouse model of NMO ON should be useful in further studies of NMO pathogenesis mechanisms and therapeutics.