Get access

Ex vivo spinal cord slice model of neuromyelitis optica reveals novel immunopathogenic mechanisms

Authors

  • Hua Zhang PhD,

    1. Departments of Physiology, University of California at San Francisco, San Francisco, CA
    2. Departments of Medicine, University of California at San Francisco, San Francisco, CA
    Search for more papers by this author
  • Jeffrey L. Bennett MD, PhD,

    1. Departments of Neurology and Ophthalmology, University of Colorado Denver School of Medicine, Aurora, CO
    Search for more papers by this author
  • A. S. Verkman MD, PhD

    Corresponding author
    1. Departments of Physiology, University of California at San Francisco, San Francisco, CA
    2. Departments of Medicine, University of California at San Francisco, San Francisco, CA
    • 1246 Health Sciences East Tower, University of California, San Francisco, CA 94143-0521
    Search for more papers by this author

Abstract

Objective:

Neuromyelitis optica (NMO) is a neuroinflammatory disease of spinal cord and optic nerve associated with serum autoantibodies (NMO–immunoglobulin G [IgG]) against astrocyte water channel aquaporin-4 (AQP4). Recent studies suggest that AQP4 autoantibodies are pathogenic. The objectives of this study were to establish an ex vivo spinal cord slice model in which NMO-IgG exposure produces lesions with characteristic NMO pathology, and to test the involvement of specific inflammatory cell types and soluble factors.

Methods:

Vibratome-cut transverse spinal cord slices were cultured on transwell porous supports. After 7 days in culture, spinal cord slices were exposed to NMO-IgG and complement for 1 to 3 days. In some studies inflammatory cells or factors were added. Slices were examined for glial fibrillary acidic protein (GFAP), AQP4, and myelin immunoreactivity.

Results:

Spinal cord cellular structure, including astrocytes, microglia, neurons, and myelin, was preserved in culture. NMO-IgG bound strongly to astrocytes in the spinal cord slices. Slices exposed to NMO-IgG and complement showed marked loss of GFAP, AQP4, and myelin. Lesions were not seen in the absence of complement or in spinal cord slices from AQP4 null mice. In cultures treated with submaximal NMO-IgG, the severity of NMO lesions was increased with inclusion of neutrophils, natural killer cells, or macrophages, or the soluble factors tumor necrosis factor α (TNFα), interleukin-6 (IL-6), IL-1β, or interferon-γ. Lesions were also produced in ex vivo optic nerve and hippocampal slice cultures.

Interpretation:

These results provide evidence for AQP4, complement- and NMO-IgG–dependent NMO pathogenesis in spinal cord, and implicate the involvement of specific immune cells and cytokines. Our ex vivo model allows for direct manipulation of putative effectors of NMO disease pathogenesis in a disease-relevant tissue. ANN NEUROL 2011

Ancillary