MOG cell-based assay detects non-MS patients with inflammatory neurologic disease

Objective: To optimize sensitivity and disease specificity of a myelin oligodendrocyte glycoprotein (MOG) antibody assay. Methods: Consecutive sera (n 5 1,109) sent for aquaporin-4 (AQP4) antibody testing were screened for MOG antibodies (Abs) by cell-based assays using either full-length human MOG (FL-MOG) or the short-length form (SL-MOG). The Abs were initially detected by Alexa Fluor goat anti-human IgG (H 1 L) and subsequently by Alexa Fluor mouse antibodies to human IgG1. Results: When tested at 1:20 dilution, 40/1,109 sera were positive for AQP4-Abs, 21 for SLMOG, and 180 for FL-MOG. Only one of the 40 AQP4-Ab–positive sera was positive for SLMOG-Abs, but 10 (25%) were positive for FL-MOG-Abs (p 5 0.0069). Of equal concern, 48% (42/88) of sera from controls (patients with epilepsy) were positive by FL-MOG assay. However, using an IgG1-specific secondary antibody, only 65/1,109 (5.8%) sera were positive on FL-MOG, and AQP4-Ab– positive and control sera were negative. IgM reactivity accounted for the remaining anti-human IgG (H 1 L) positivity toward FL-MOG. The clinical diagnoses were obtained in 33 FL-MOG–positive patients, blinded to the antibody data. IgG1-Abs to FL-MOG were associated with optic neuritis (n 5 11), AQP4-seronegative neuromyelitis optica spectrum disorder (n 5 4), and acute disseminated encephalomyelitis (n 5 1). All 7 patients with probable multiple sclerosis (MS) were MOG-IgG1 negative. Conclusions: The limited disease specificity of FL-MOG-Abs identified using Alexa Fluor goat antihuman IgG (H 1 L) is due in part to detection of IgM-Abs. Use of the FL-MOG and restricting to IgG1-Abs substantially improves specificity for non-MS demyelinating diseases. Classification of evidence: This study provides Class II evidence that the presence of serum IgG1- MOG-Abs in AQP4-Ab–negative patients distinguishes non-MS CNS demyelinating disorders from MS (sensitivity 24%, 95% confidence interval [CI] 9%–45%; specificity 100%, 95% CI 88%–100%). Neurol Neuroimmunol Neuroinflamm 2015;2:e89; doi: 10.1212/ NXI.0000000000000089