Our study objective was testing for anti-neuronal autoantibodies within commercially available intravenous immunoglobulin (IVIg) preparations. Sixteen samples from 5 different commercially available IVIg preparations were tested with cell-based assays (CBA) and enzyme-linked immunosorbent assay (ELISA) to detect and characterize common neuronal autoantibodies, and with immunohistochemistry on teased fibers from mouse sciatic nerve and on mouse brain sections to screen for nodal and not yet identified neuronal antigens. In 15/16 IVIg preparations, anti-GAD antibodies were detected in titers ranging from 40 to 1507 IU/mL, as typically seen in type 1 diabetes, but not in the range (> 2000 IU/mL) seen in GAD-positive neurological patients. None of the preparations was however positive with anti-GAD CBA. Antibodies to AQP4 were also detected by ELISA in 15/16 IVIg preparations with titers comparable to those seen in AQP4-seropositive NMO patients; with CBA, however, all IVIg samples were AQP4-negative. IVIg preparations contained IgG-anti-MAG antibodies by ELISA at statistically significant higher titers compared to controls. Two of the 16 IVIg samples were positive for human 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR) antibodies. All IVIg preparations were negative for antibodies to MOG, NMDAR, anti-nodal, and other neuronal-specific proteins. IVIg preparations contain antibodies against GAD and AQP4 in titers comparable to those seen in autoimmune patients when tested by ELISA, but not by CBA or tissue immunohistochemistry, suggesting that the autoantibodies within the IVIg are against linear rather than structural epitopes, as part of the natural antibody immune repertoire. The information is clinically important for diagnosis when testing patients' sera after they have received therapy with IVIg to avoid false interpretation.
Read full abstract