In Reply Takahashi and colleagues detected in the cerebrospinal fluid of a patient with Creutzfeldt-Jakob disease antibodies against the NR2b subunit of the N-methyl-D-aspartate receptor (NMDAR) confirmed by immunoblot and immunofluorescence on human embryonic kidney 293 cells transfected with NR1 and NR2B subunits of the NMDAR.1 They concluded that antibodies against NMDAR do not rule out the diagnosis of Creutzfeldt-Jakob disease. Takahashi and colleagues do not understand that our article was focused on determining antibodies to neuronal cell-surface proteins including NMDARs and other synaptic receptors.2 These antibodies are against conformation-dependent epitopes and do not react with immunoblots of the protein. In addition, Takahashi and colleagues do not understand that anti-NMDAR encephalitis associates with antibodies against a well-defined epitope in the N-terminal region of the NR1 subunit of the NMDAR.3 These antibodies cause capping and internalization of the receptor, leading to a decrease of the density of cell surface and synaptic NMDAR and a reduction of NMDAR-synaptic currents.4 These antibodies show intense reactivity with the neuropil of brain along with robust immunolabeling of the neuronal cell surface and are demonstrated with human embryonic kidney cells transfected with NR1 alone or combined with NR2B.5 These data5 have been in the literature since 2008. The lack of significance of antibodies against NR2B was also reported in 2008, when we assessed samples (all supposedly with NMDAR antibodies) from Takahashi6 showing that those we identified (masked of clinical information) with NR1 antibodies were found to be highly specific for anti-NMDAR encephalitis, while those that were NR1 antibody negative (supposedly NR2B positive) had no syndrome specificity, lacking clinical utility. Moreover, we did not find any reactivity with the neuropil of brain or cultures of neurons with those supposedly NR2B antibody positive. Therefore, it is difficult to understand why Takahashi and colleagues still confuse NR1 with NR2B antibodies. Indeed, most reports by Takahashi refer to antibodies reacting with NR2B linear epitopes or peptides that have never been convincingly shown to react with the neuropil of brain or cell surface of neurons. In their article1 where they suggest binding of NR2B antibodies with human embryonic kidney cells, the unusual staining pattern and absence of colocalization studies raise serious concerns about their interpretation. These authors have never demonstrated any pathogenic effect of NR2B antibodies and therefore attributing to these antibodies1 the findings obtained by other investigators with NR14 is highly misleading. The numerous disorders that Takahashi and colleagues have reported in association with NR2B antibodies, including Rasmussen encephalitis, epilepsia partialis continua, viral encephalitis, neurodegenerative diseases, paraneoplastic opsoclonus, and encephalitis6 (to which now Creutzfeldt-Jakob disease has been added), leads to a modified interpretation of their conclusion: cerebrospinal fluid antibodies against NR2B (NMDAR) do not rule out the diagnosis of any disease.
Graus F, Dalmau J. Neuronal Antibodies in Creutzfeldt-Jakob Disease—Reply. JAMA Neurol. 2014;71(4):514-515. doi:10.1001/jamaneurol.2014.30