Subscribe to eTOC

Autoantibody Immunotherapy Yields Promising Results in First Subjects with Unclassified Dementia


In a new study, autoimmune antibodies to N-methyl-D-aspartate receptor (NMDAR), primarily immunoglobulin A (IgA), were found to be elevated in the sera but not in the CSF of patients with several neuropsychiatric disorders.

A retrospective study of serum and cerebrospinal fluid (CSF) samples from 660 patients with dementia and other neuropsychiatric disorders found elevated levels of a subtype of a brain-specific autoantibody in the sera of some patients, especially those with unclassified disease, compared with healthy controls, although the number of such patients was small.

However, prospective testing of immunotherapy in two unclassified dementia patients — whose mean age was 65 — reduced autoantibody levels, stabilizing dementia and improving functional magnetic resonance imaging markers of neurodegeneration at 12 months follow-up. Ongoing testing in an additional 13 patients has shown similarly promising results, one of the study authors told Neurology Today.

In the study, autoimmune antibodies to N-methyl-D-aspartate (NMDA) receptor, primarily immunoglobulin A (IgA), were found to be elevated in the sera but not in the CSF of patients with several neuropsychiatric disorders, according to corresponding author Harold Prüss, MD, of the department of neurology and experimental neurology at the Charité Universitätsmedizin in Berlin. In addition, various epitope determinants were identified in patients with NMDA-receptor (NMDAR) IgA-associated cognitive decline.

Dr. Prüss told Neurology Today in a telephone interview that the initial two patients responded well to immunotherapy with high-dose methylprednisolone and plasma exchange, followed by treatment in one patient with cyclophosphamide. One patient showed marked improvement in the first few weeks, with positive signs in cognitive tests, alertness, aphasia, and motivation. Neither patient showed any signs of progression during the eight-to-12 month follow-up.

Ongoing immunotherapy in the other 13 patients has found similar improvement in some, Dr. Prüss said. The results will be reported once a full cohort of 20 patients has been recruited for a prospective study.

“We have now treated a total of 15 patients and while not all of them have improved, none appear to be getting any worse after six to nine months. Our findings to date reveal a subgroup of patients with high antibody levels who can potentially benefit from immunotherapy.”

Serum NMDAR IgA and immunoglobulin M (IgM) antibodies occur in a significant number of patients with dementia and other neurodegenerative disorders, but whether they contribute to or are caused by these disorders remains unknown, he said.


Published in the October issue of Annals of Clinical and Translational Neurology, an open-access journal, researchers in Germany and the United States retrospectively assessed and compared NMDAR autoantibody levels in 286 patients with dementia and 90 neurodegenerative disease patients without dementia, 26 patients with schizophrenia, depression, or dissociative disorders, and 47 healthy age-matched individuals, averaging 66.7 years of age.

The study participants had behavioral variant frontotemporal dementia (bvFTD), Lewy body dementia (LBD), Creutzfeldt-Jakob disease (CJD), Parkinson's disease (PD) with dementia, corticobasal syndrome (CBS), progressive supranuclear palsy (PSP), primary progressive aphasia (PPA), Huntington's disease (HD), vascular dementia, and unclassified dementia.

In total, NMDAR autoantibodies were found in 16.1 percent of 286 dementia patients but only 2.8 percent of 217 cognitively healthy controls. While the number of unclassified dementia patients was small — just 20 patients — 60.1 percent had NMDAR antibodies as well as more frequent CSF abnormalities and subacute or fluctuating disease progression. Higher levels were also found in some patients with PSP, CBS, PD-related dementia, and PPA. Antibody binding to receptor mutants also allowed epitope mapping in a subgroup of patients.

“Our findings revealed a subgroup of patients with high antibody levels who can potentially benefit from immunotherapy. If confirmed in a larger prospective study, I think this could really change the playing field,” Dr. Prüss told Neurology Today.

It is also the first study to map epitopes in NMDAR IgA-positive patients, he noted, adding that the observed epitope specificity of the antibodies is likely related to their pathogenic potential.

“The paper can also be viewed as a pilot study on individual antibody diagnostics to better understand why some patients do not respond to immunotherapy, and this might be linked to which epitope that IgA antibodies bind themselves to.”

Imaging and in vitro data as well as analyses of the first few immunotherapy cases also suggest that the presence and levels of NMDAR-IgA could possibly predict partial reversibility of the disease, he added.

Which form(s) of immunotherapy or how long treatment should be given remains unclear, however. Whether or not a patient responds to steroids alone might not be enough to predict their response to immunotherapy, or whether additional treatment with plasma exchange, intravenous immunoglobulin, rituximab, or cyclophosphamide might be better, he said. Also, patients with higher NMDAR antibody levels might need longer treatment, although the risk and severity of side effects increases with longer and higher doses of these agents.


In a study published last July in Annals of Neurology, Hannelore Ehrenreich, MD, DVM, a professor of clinical neuroscience at the Max Planck Institute of Experimental Medicine in Göttingen, Germany, and colleagues reported that screening of over 4,200 subjects for serum autoantibodies directed against brain antigens revealed comparable autoantibody seroprevalence, immunoglobulin classes, and titers in patients with various neuropsychiatric disorders and in healthy individuals. They concluded that serum autoantibodies like the NMDAR1 subunit on their own are definitely not disease indicators. (Read the Neurology Today story about this study, “Brain Antigen-Directed Autoantibodies Found Comparably Detectable in Healthy and Diseased Groups,” here:


DR. HAROLD PRÜSS: “We have now treated a total of 15 patients and while not all of them have improved, none appear to be getting any worse after six to nine months. Our findings to date reveal a subgroup of patients with high antibody levels who can potentially benefit from immunotherapy.”

Commenting on the new study, Dr. Ehrenreich told Neurology Today that the new report is potentially very interesting but still very preliminary. To be convincing, she said, the findings of abnormally high levels of IgA/IgM autoantibodies in just 20 patients with unclassified dementia out of 660 total subjects would have to be replicated. “The seroprevalence in the remaining 640 individuals, healthy and ill, are essentially in the range of that reported in our study for both healthy and ill subjects of comparable age.

“Therefore they are as expected and no surprise,” she said.

In addition, she said the number of healthy individuals with NMDAR autoantibodies in the new paper was much too low, considering almost all the recent literature on reasonable amounts in healthy individuals — a bias that could be corrected by including a larger number of healthy subjects.

Because only two case reports were included as a basis for the study's conclusions and therapeutic suggestions, and because epitope mapping was based on only two patients, conclusions cannot really be drawn, she said.

“If there is no CSF antibody titer detectable, how can the authors claim any connection with dementia? Or with the other CNS symptoms they describe? Even though most antibodies, once they penetrate the blood-brain-barrier, may end up being bound to brain tissue, the lack of the slightest ‘spillover’ into the CSF — if the amounts are believed to be symptom-causing — is strange and should at least have been discussed.”

“I think the findings are potentially very important if confirmed in larger trials, but we have to be careful how we interpret them,” said Sean J. Pittock, MD, a professor of neurology, consultant in the Autoimmune Neurology Clinic, and co-director of the Mayo Clinic's Neuroimmunology Laboratory in Rochester, MN.

“Overall, these are new and intriguing findings that further open the conversation regarding the concept of autoimmune dementia,” he told Neurology Today in a telephone interview.

“Although the potential role of neural receptor-specific IgA and IgM as biomarkers or pathogenic effectors in neurological disease remains to be elucidated, the paper provides evidence that some patients with dementia may have an autoimmune etiology for their cognitive disorder that might respond favorably to immunotherapy. Immunotherapy responsiveness is widely acknowledged in autoimmune encephalitis, but this concept has generally not been considered for patients with dementia.”

Dr. Pittock said he and his colleagues do not advocate immunotherapy as a diagnostic test in all cases of dementia.

“Additional serological biomarkers are needed to assist clinicians in identifying patients most likely to have an autoimmune etiology.”

Clinical clues helpful in identifying patients most likely to have a favorable response to immunotherapy include rapid progression, fluctuating course, a concomitant movement disorder such as myoclonus or tremor, headache, inflammatory spinal fluid, and MRI considered atypical for a neurodegenerative disease, he said. An additional very helpful clue is the detection of neural immunoglobulin G (IgG) autoantibodies in serum or CSF, such as voltage-gated potassium channel complex (VGKC) or DPPX antibodies.

“If NMDAR IgA or IgM may provide a novel clue to an autoimmune etiology, as this study suggests, and if this is indeed a robust finding, and the clinical sensitivity and specificity of NMDAR IgA and IgM hold true in future studies, the therapeutic implications for dementia are profound and will lead to a seismic shift in neuroimmunology laboratory testing, which has for decades been purely focused on IgG.”

At this time, however, Dr. Pittock advised caution in how this preliminary data is interpreted.

“In autoimmune NMDAR encephalitis, where IgG targeting NMDARs has been shown to be pathogenic, the identification of NMDAR IgG in CSF is more specific than when identified in serum alone, and patients who do not have encephalitis can harbor NMDAR IgG in serum but not in CSF. In this situation, the NMDAR IgG does not appear clinically relevant. The same may hold true for serum NMDAR IgA and IgM. In the new study, most patients in whose serum NMDAR IgA or IgM was detected lacked detectable NMDAR autoantibody in CSF, suggesting that the antibodies did not cross the blood-brain barrier.”


•. Doss S, Wandinger KP, Hyman BT, Prüss H, et al. High prevalence of NMDA receptor IgA/IgM antibodies in different dementia types. Ann Clin Transl Neurol 2014;10:822–832.
    •. Busse S, Brix B, Kunschmann R, et al. N-methyl-D-aspartate glutamate receptor (NMDA-R) antibodies in mild cognitive impairment and dementias. Neurosci Res 2014;85:58–64.
    •. Prüss H, Höltje M, Maier N, et al. IgA NMDA receptor antibodies are markers of synaptic immunity in slow cognitive impairment. Neurology 2012;78:1743–1753.
    •. Pittock SJ, Balice-Gordon R. Emerging synaptic biomarkers in autoimmune dementia: NMDARs targeted by IgA: Editorial. Neurology 2012;78:1710–1711.
    •. Flanagan EP, McKeon A, Pittock SJ, et al. Autoimmune dementia: clinical course and predictors of immunotherapy response. Mayo Clin Proc 2010;85:881–897.
    •. Dahm L, Ott C, Steiner J, Ehrenreich H, et al. Seroprevalence of autoantibodies against brain antigens in health and disease. Ann Neurol 2014;76:82–94.
    •. Ehrenreich H, Steiner J. Serum autoantibodies against N-methyl-D-aspartate receptor subunit NR1 are no disease indicator. Ann Neurol 2014; Epub 2014 Jul 23.