To the Editor:

We read with interest the concise communication by Yoshio et al (1) describing an association of IgG anti-NR2 glutamate receptor antibodies in cerebrospinal fluid (CSF) with neuropsychiatric systemic lupus erythematosus (NPSLE). DeGiorgio et al first demonstrated that IgG anti-NR2 glutamate receptor antibodies caused neuronal death in mice, when injected into the brain (2). They also reported the presence of these antibodies in CSF from an SLE patient with progressive cognitive decline (2). It was further demonstrated by investigators in that group that the presence of IgG anti-NR2 glutamate receptor antibodies within the brain resulted in cognitive decline in mice (3). These findings suggested that these antibodies might cause diffuse psychiatric/neurologic syndromes in human SLE. However, Yoshio and colleagues found that, compared with lupus patients without NPSLE, CSF IgG anti-NR2 glutamate receptor antibody levels were increased in lupus patients with neurologic syndromes of the central nervous system, but not in those with diffuse psychiatric/neuropsychological syndromes alone (1).

To measure IgG anti-NR2 glutamate receptor antibodies in sera and CSF, Yoshio et al used an enzyme-linked immunosorbent assay (ELISA) with the synthetic DWEYSVWLSN peptide conjugated to bovine serum albumin (BSA) as antigen (1). Our group previously demonstrated that sera from SLE patients frequently express antibodies to human serum albumin (HSA), BSA, and ovalbumin (4). Therefore, subtracting binding activity to these carrier proteins would be mandatory in order to determine specific activities of antibodies to peptides conjugated to carrier proteins, including BSA and HSA (4). However, Yoshio and colleagues did not subtract BSA binding activities. It appears that all sera with anti-HSA also contain antibodies to BSA (4). We therefore reexamined whether CSF and sera from SLE patients contain antibodies to HSA, using the method described by Yoshio et al.

Wells of a 96-well microtiter plate (Falcon Pro-Bind; Becton Dickinson, Lincoln Park, NJ or Nunc-immuno module F8 Maxisorp; Nunc, Roskilde, Denmark) were coated with HSA (Miles, Elkhart, IN) or HSA conjugated (at a 1:1 weight ratio) with highly purified synthetic DWEYSVWLSN peptide (purity >95%) (HSA–NR2 peptide), at 20 μg/ml in phosphate buffered saline (PBS), overnight at 4°C. The wells were blocked with Block Ace (Dainippon, Osaka, Japan) for Falcon Pro-Bind plates or with PBS containing 1% BSA (Miles) for Nunc Maxisorp plates, for 2 hours at room temperature. Before being added to the wells, serum and CSF samples were diluted 1:200 and 1:2, respectively, in PBS containing 1% BSA. After incubation at 37°C for 1 hour, bound IgG anti-NR2 glutamate receptor antibody was detected with peroxidase-conjugated F(ab′)2 fragments of goat anti-human IgG (Cappel, Cochranville, PA). Binding activity was expressed as optical density at 492 nm (OD492) as measured in a 2-wavelength microplate photometer (MTP-450; Corona Electric, Ibaraki, Japan).

As seen in Table 1, all 16 samples exhibited positive binding to HSA–NR2 peptide in Falcon Pro-Bind plates (Yoshio and colleagues' method) as well as in Nunc Maxisorp plates. However, 8 of the 16 samples showed higher binding activity (OD492) to HSA alone than to HSA–NR2, indicating that those samples would yield false-positive results for IgG anti-NR2 glutamate receptor antibodies unless nonspecific binding to HSA alone were subtracted. In addition, the levels of binding activity obtained with Falcon Pro-Bind plates were ∼20–30% of those with Nunc Maxisorp plates, even though peroxidase-conjugated F(ab′)2 goat anti-human IgG was used at 1:5,000 in the former plates and at 1:10,000 in the latter. Therefore, it would be preferable to use Nunc Maxicorp plates. Nonetheless, these results confirm that ∼50% of serum and CSF samples contain antibodies to HSA (and presumably to BSA as well, since it was previously shown that all sera positive for anti-HSA contained antibodies to BSA [4]).

Table 1. Measurement of IgG anti-NR2 glutamate receptor antibodies (optical density at 492 nm) using enzyme-linked immunosorbent assay*
SampleCSF or serumMaxisorpPro-Bind
  • *

    Cerebrospinal fluid (CSF) or serum samples from systemic lupus erythematosus patients were assayed for IgG anti-NR2 glutamate receptor antibodies by enzyme-linked immunosorbent assay as described in the text. Peroxidase-conjugated F(ab′)2 goat anti-human IgG was used at 1:10,000 (Maxisorp plates) and at 1:5,000 (Pro-Bind plates). HSA–NR2 = human serum albumin conjugated with synthetic DWEYSVWLSN peptide.


These findings raise serious concern about the specificity of the ELISA used by Yoshio et al. It is highly likely that the presence of anti-BSA antibodies would have significantly influenced their results and conclusions. Therefore, their conclusion that IgG anti-NR2 glutamate receptor antibodies in CSF may cause focal neurologic damage such as seizure disorders, aseptic meningitis, and transverse myelopathy is not supported.

Shunsei Hirohata MD*, Yoshiyuki Arinuma MD*, Tamiko Yanagida PhD*, * Teikyo University School of Medicine, Tokyo, Japan.