Approximately 7 % to 9 % of people in western populations suffer from one of 80 inflammatory disorders that are collectively defined as autoimmune diseases. This translates into 24 million people in the United States of America alone. During the development of an autoimmune disease, the immune system is not able to distinguish self from foreign antigens and attacks its own healthy tissue. About a decade ago, a new autoimmune disease, the N-methyl-D-aspartate receptor (NMDAR) encephalitis, has been discovered. NMDARs are frequently expressed in nerve cells in the brain and autoantibodies targeting this receptor are thought to influence neuronal brain functions. The disease is associated with a prodromal phase which includes fever and headache and can lead to psychotic syndromes, seizures, decreased levels of consciousness, dyskinesia and often to intensive care treatment with autonomic instability and hypoventilation. A trigger for some cases of the disease might be an ovarian teratoma. Previously, in NMDAR encephalitis, the involvement of anti-NMDAR antibodies has been demonstrated in the disease process but an association with other neuronal autoantibodies could not be excluded. Therefore, we collected the cerebrospinal fluid (CSF) of eight women between 18 and 34 years, five with ovarian teratoma, either in the acute phase of the disease or in remission in order to study their antibody repertoire. With fluorescence-activated cell sorting (FACS) we separated CSF single memory B cells and antibody secreting cells and amplified their immunoglobulin heavy and light chains. We generated 170 human monoclonal antibodies of NMDAR encephalitis patients representing their antibody repertoire. We tested all antibodies on mouse brain sections and human embryonic kidney cells (HEK293T) transfected with the NR1 subunit of the NMDAR, which was identified as the major autoantigen in NMDAR encephalitis. However, only nine antibodies displayed a specific anti-NMDAR reactivity. Previously, it was suggested that the target region of the autoantibodies lies in the amino-terminal domain (ATD) and a mutation at the amino acid N368Q abolishes binding. We could confirm a loss of binding with the N368Q mutation in all nine monoclonal NMDAR autoantibodies. In vivo binding of NMDAR epitopes by the monoclonal autoantibodies was further demonstrated by intravenous injections of the purified antibodies into mice and subsequent immunohistochemistry (IHC) of brain sections. In addition, functional in vitro assays on primary hippocampal neurons revealed surface NMDAR downregulation and impairment of NMDAR-mediated glutamate-evoked currents and therefore proved the antibody’s inhibitory function on NMDARs. To investigate the origin and pathogenesis of NMDAR autoantibodies we analyzed whether somatic hypermutations and class switching had occurred. Here, we could show that in comparison to all non-NMDAR antibodies in the patients CSF, NMDAR autoantibodies presented less somatic hypermutations and some of them even had no mutations as compared to the germline. The latter are in general referred to as natural occurring antibodies, which indicate incomplete formation of tolerance against the NMDAR during B cell development. Besides antibody secreting cells, such as plasma cells and plasma blasts, we could also detect memory B cells, which are thought to be a risk factor for clinical relapses after therapy. Interestingly, only a small amount of autoantibodies is NMDAR specific, over 95 % of the non-NMDAR antibodies demonstrated binding to mouse brain epitopes including neuronal surface antigens. In mass spectrometry analysis, we could detect autoantibodies against a variety of brain epitopes, such as alpha-amino-3-hydroxy5-methyl-4-isoxazolepropionic acid receptor subtype 2 (AMPAR2), inositol 1,4,5-trisphosphate receptor type 1 (ITPR1), and glial fibrillary acidic protein (GFAP). Further investigations of these and additional autoantibodies could reveal novel targets, implying a contribution to the diverse clinical picture of NMDAR encephalitis. In the future, the generated recombinant human monoclonal autoantibodies, against NMDAR and non-NMDAR targets, could be used as tools for additional research, e.g. together with high resolution synaptic imaging to study molecular mechanisms of disease pathogenesis. As titer correlations between patients are ambiguous, using these monoclonal NMDAR autoantibodies for standardization of autoantibody titers could improve its relevance in clinical settings and ultimately improve therapeutic intervention
