Serum testing for NMDAR and other autoantibodies is fraught with pitfalls, and careful clinical consideration is important.
As I’ve discussed in this blog, there is increasing recognition of the role of anti-brain-specific autoantibodies in human disease. There are a host of syndromes associated with antinuclear neuronal antibodies-eg, anti-Hu and anti-Yo-and other syndromes associated with cell surface directed autoantibodies-eg, anti–N-methyl-D-aspartate receptor (anti-NMDAR) and anti-aquaporin 4.
A considerable proliferation of diagnostic tests are being ordered to search for these various antibodies in neurologic diseases of various types. This is leading to a process of drift, where autoantibodies are being found in cases without the core classic features associated with the syndromes where these antibodies were described.
I believe that a couple of sensible questions need to be asked when trying to interpret the utility of these tests:
• How frequently are these antibodies found in patients without neurologic disease?
• How frequently are these antibodies seen in patients with other neurologic diseases without any suspicion of an autoantibody-mediated mechanism?
Fortunately, a recent study by Dahm and associates1 is shedding some light on these questions. The authors found that serum from a large cohort of healthy controls (1703) and an equivalently large group of neuropsychiatrically ill patients (2533) with a range of syndromes (schizophrenia, affective disorders, stroke, Parkinson disease, amyotrophic lateral sclerosis, personality disorder) had fairly similar frequencies of antibodies to a panel of antigens in serum-including NMDAR, CASPR2, MOG, and GAD65. In most cases, these were positive in about 1% of each group. However, anti-NMDAR antibodies were seen in 10% of the patients in both groups!
There are some very interesting features of this finding with NMDAR-only a small number of patients had IgG directed against NMDAR; most had IgA or IgM responses.
What should we extract from this? Does this mean that these antibodies are nonpathogenic?
I think this is clearly the wrong conclusion to reach. As we now know, there is molecular mimicry between herpes simplex virus 1 (HSV1) and NMDAR and a very high fraction of the population has been infected with HSV1 and, in fact, a fair number probably have IgA titers against HSV1. This seems a clear possible inciting factor.
Also, the presence of IgG in only a small percentage and the likelihood of IgG being associated with chronic immunological phenomena seems more reasonable.
Also, this study examined serum only-clearly CSF antibodies are far more likely to be associated with brain disease.
One conclusion to extract is that serum testing for NMDAR and other autoantibodies is fraught with pitfalls and careful clinical consideration is important. Also, I think it is quite interesting to consider whether patients with active serum titers for NMDAR might be more likely to develop autoimmune encephalitis in conditions where the systemic immune system gains enhanced access to the CNS.
1. Dahm L, Ott C, Steiner J, et al. Seroprevalence of autoantibodies against brain antigens in health and disease. Ann Neurol. 2014;76:82-94.