Optimizing animal models of autoimmune encephalitis using active immunization

Encephalitis is a devastating neurologic disorder with high morbidity and mortality. Autoimmune causes are roughly as common as infectious ones. N-methyl-D-aspartic acid receptor (NMDAR) encephalitis (NMDARE), characterized by serum and/or spinal fluid NMDAR antibodies, is the most common form of autoimmune encephalitis (AE). A translational rodent NMDARE model would allow for pathophysiologic studies of AE, leading to advances in the diagnosis and treatment of this debilitating neuropsychiatric disorder. The main objective of this work was to identify optimal active immunization conditions for NMDARE in mice.

Female C57BL/6J mice aged 8 weeks old were injected subcutaneously with an emulsion of complete Freund's adjuvant, killed and dessicated Mycobacterium tuberculosis, and a 30 amino acid peptide flanking the NMDAR GluN1 subunit N368/G369 residue targeted by NMDARE patients' antibodies. Three different induction methods were examined using subcutaneous injection of the peptide emulsion mixture into mice in 1) the ventral surface, 2) the dorsal surface, or 3) the dorsal surface with reimmunization at 4 and 8 weeks (boosted). Mice were bled biweekly and sacrificed at 2, 4, 6, 8, and 14 weeks. Serum and CSF NMDAR antibody titer, mouse behavior, hippocampal cell surface and postsynaptic NMDAR cluster density, and brain immune cell entry and cytokine content were examined.

All immunized mice produced serum and CSF NMDAR antibodies, which peaked at 6 weeks in the serum and at 6 (ventral and dorsal boosted) or 8 weeks (dorsal unboosted) post-immunization in the CSF, and demonstrated decreased hippocampal NMDAR cluster density by 6 weeks post-immunization. In contrast to dorsally-immunized mice, ventrally-induced mice displayed a translationally-relevant phenotype including memory deficits and depressive behavior, changes in cerebral cytokines, and entry of T-cells into the brain at the 4-week timepoint. A similar phenotype of memory dysfunction and anxiety was seen in dorsally-immunized mice only when they were serially boosted, which also resulted in higher antibody titers.

Our study revealed induction method-dependent differences in active immunization mouse models of NMDARE disease. A novel ventrally-induced NMDARE model demonstrated characteristics of AE earlier compared to dorsally-induced animals and is likely suitable for most short-term studies. However, boosting and improving the durability of the immune response might be preferred in prolonged longitudinal studies.