Background: Neuromyelitis optica spectrum disorders (NMOSD) are severe central nervous system inflammatory demyelinating disorders (CNS IDD) characterized by monophasic or relapsing, longitudinally extensive transverse myelitis (LETM) and/or optic neuritis (ON). A significant proportion of NMOSD patients are seropositive for aquaporin-4 (AQP4) autoantibodies. We compared the AQP4 autoantibody detection rates of tissue-based indirect immunofluorescence assay (IIFA) and cell-based IIFA.
Methods: Serum of Chinese CNS IDD patients were assayed for AQP4 autoantibodies by tissue-based IIFA using monkey cerebellum and cell-based IIFA using transfected HEK293 cells which express human AQP4 on their cell membranes.
Results: In total, 128 CNS IDD patients were studied. We found that 78% of NMO patients were seropositive for AQP4 autoantibodies by cell-based IIFA versus 61% by tissue-based IFA (p = 0.250), 75% of patients having relapsing myelitis (RM) with LETM were seropositive by cell-based IIFA versus 50% by tissue-based IIFA (p = 0.250), and 33% of relapsing ON patients were seropositive by cell-based IIFA versus 22% by tissue-based IIFA (p = 1.000); however the differences were not statistically significant. All patients seropositive by tissue-based IIFA were also seropositive for AQP4 autoantibodies by cell-based IIFA. Among 29 NMOSD patients seropositive for AQP4 autoantibodies by cell-based IIFA, 20 (69%) were seropositive by tissue-based IIFA. The 9 patients seropositive by cell-based IIFA while seronegative by tissue-based IIFA had NMO (3), RM with LETM (3), a single attack of LETM (1), relapsing ON (1) and a single ON attack (1). Among 23 NMO or RM patients seropositive for AQP4 autoantibodies by cell-based IIFA, comparison between those seropositive (n = 17) and seronegative (n = 6) by tissue-based IIFA revealed no differences in clinical and neuroradiological characteristics between the two groups.
Conclusion: Cell-based IIFA is slightly more sensitive than tissue-based IIFA in detection of AQP4 autoantibodies, which are highly specific for NMOSD.