Anti-NMDA (N-methyl D-aspartate) receptor antibody encephalitis, also termed NMDA receptor antibody encephalitis, is an acute form of encephalitis which is potentially lethal but has high probability for recovery. It is caused by an autoimmunereaction primarily against the NR1 subunit of the NMDA receptor.^[1] Different descriptions and syndromal designations for this disease existed in medical literature prior to 2007 but it was then that the disease was officially categorized and named by Dalmau and colleagues.^[2]

Movement disorder relapses after herpes simplex virus 1 (HSV1) encephalitis have been hypothesized to be secondary to postviral autoimmunity. Recently, a proportion of patients with HSV1 encephalitis (HSE) were shown to produce autoantibodies against N-methyl-D-aspartate receptor (NMDAR).^[3]

The condition is associated with tumours, mostly teratomas of the ovaries, and is thus considered a paraneoplastic syndrome. However, there are a substantial number of cases with no detectable tumour and in fact it is emerging that most cases do not have tumour.^[4] This is an especially important point as the search for a tumour should not occupy attention when immune treatments are being planned.

The overall incidence of the condition is unknown .^[5] More recent figures produced by the California Encephalitis Project showed that the disease had a higher incidence than its individual viral counterparts in patients <30 years.^[6] The largest cases series to date characterized 577 patients with anti-NMDA receptor encephalitis. The epidemiologic data were limited but this study provides the best approximation of disease distribution. This case series found that women are disproportionally affected, with 81% of cases reported as female. Disease onset is skewed towards children, with a median age of diagnosis of 21 years. Over a third of cases were children, while only 5% of cases were patients over the age of 45. This same review found showed that 394 out of 501 patients (79%) had a good outcome by 24 months.^[7] 30 patients (6%) died and the rest were left with mild to severe deficits. This study also confirmed that patients with the condition are more likely to be of Asian or African origin.

Prior to the development of a symptom complex that is specific to anti-NMDA receptor encephalitis, patients may experience prodromal symptoms. These could include headaches, flu-like illness, or symptoms similar to an upper respiratory infection. These symptoms may be present for weeks or months prior to disease onset.^[1] Beyond the prodromal symptoms the disease progresses at varying rates and patients may present with a variety of neurologic symptoms. During the initial stage of the disease symptoms vary slightly between children and adults- however behavior changes are a common first symptom within both groups. These changes often include increased agitation, paranoia, psychosis, and violent behaviors. Other common first manifestations include seizures and bizarre, often rhythmic, movements mostly of the lips and mouth but also including pedaling motions with legs or hand movements resembling playing a piano. Some other symptoms typical during the disease onset include impaired cognition, memory deficits, and speech problems (including aphasia or mutism).^[7][8]

As the disease progresses the symptoms become medically urgent and often include autonomic dysfunction, hypoventilation, cerebellar ataxia, hemiparesis, loss of consciousness, or catatonia. During this acute phase most patients require care in an Intensive Care Unit to stabilize breathing, heart rate, and blood pressure. It is important to note one distinguishing characteristic of Anti-NMDA Receptor Encephalitis is the concurrent presence of many of the above listed symptoms. It is extremely unusual that patients have only one or two symptoms- the majority of patients experience at least four symptoms with many experiencing six or seven over the course of the disease.^[7][8]

The presence of antibodies in the serum and cerebrospinal fluid (CSF)

The condition is mediated by autoantibodies that target NMDA receptors in the brain. These are produced by cross reactivitywith NMDA receptors in the teratoma; these tumours contain many different cell types, including brain cells, and thus present a window in which a breakdown in immunological tolerance can occur. Other autoimmune mechanisms are suspected for patients that lack any form of neoplasm. Whilst the exact pathophysiology of the disease is still debated, empirical evaluation of the origin of anti-NMDA antibodies in the serum and the CSF leads to the consideration of two possible mechanisms.

These mechanisms may be informed by some simple observations. Firstly, serum NMDAR-antibodies are consistently found at higher concentrations than CSF antibodies, on average 10 fold higher.^[9][10] This strongly suggests the antibody production is systemic rather than in the brain / CSF. When concentrations are normalised for total IgG, ‘intrathecal synthesis’ is detected. This implies that there is more NMDAR-antibody in the CSF than would be predicted given the expected quantities of total IgG.

1. Passive access involves the diffusion of antibodies from the blood across a pathologically disrupted blood-brain barrier(BBB).^[11] This cellular filter, separating the central nervous system from the circulatory system, normally prevents larger molecules from entering the brain. A variety of reasons for such a collapse in integrity have been suggested, with the most likely answer being the effects of acute inflammation of the nervous system. Likewise, the involvement of corticotropin releasing hormone on mast cells in acute stress has been shown to facilitate BBB penetration.^[12] However, it is also possible that the autonomic dysfunction manifested in many patients during the later phases of the condition aids antibody entry. For example, an increase in blood pressure would force larger proteins, such as antibodies, to extravasate into the CSF.
2. Intrathecal production (production of antibodies in the intrathecal space) may also be a possible mechanism. Dalmau et al. demonstrated that 53 out of 58 patients with the condition had at least partially preserved BBBs, whilst having a high concentration of antibodies in the CSF. Furthermore cyclophosphamide and rituximab,^[13] drugs used to eliminate dysfunctional immune cells, have been shown to be successful second line treatments in patients where first line immunotherapy has failed.^[14] These destroy excess antibody producing cells in the thecal space, thus alleviating the symptoms.

A more sophisticated analysis of the processes involved in antibody presence in the CSF hint at a combination of these two mechanisms in tandem.

The binding of antibodies to NMDA receptors

Once the antibodies have entered the CSF, they bind to the NR1 subunit of the NMDA receptor. There are 3 possible methods in which neuronal damage occurs.

1. A reduction in the density of NMDA receptors on the post synaptic knob,due to receptor internalisation once the antibody has bound. This is dependent on antibodies cross linking.^[15]
2. The direct antagonism of the NMDA receptor by the antibody, similar to the action of typical pharmacological blockers of the receptor, such as phencyclidine and ketamine.
3. The recruitment of the complement cascade via the classical pathway (antibody-antigen interaction). Membrane attack complex (MAC) is one of the end products of this cascade^[16] and can insert into neurons as a molecular barrel, allowing water to enter. The cell subsequently lyses. Notably, this mechanism is unlikely as it causes cell to die, which is inconsistent with current evidence.

If patients are found to have a tumour, the long term prognosis is generally better and the chance of relapse is much lower. This is because the tumour can be removed surgically, thus eradicating the source of autoantibodies. In general, early diagnosis and aggressive treatment is believed to improve patient outcomes, but this remains impossible to know without data from randomized controlled trials.^[7] Given that the majority of patients are initially seen by psychiatrists (not neurologists) due to the development of psychiatric symptoms, it is critical that all physicians (especially psychiatrists) consider anti-NMDA receptor encephalitis as a cause for acute psychosis in adolescents with no past neuropsychiatric history.

• If a tumour is detected, its removal should occur in conjunction with first line immunotherapy . This involves steroids to suppress the immune system, intravenous immunoglobulins and plasmapheresis to physically remove autoantibodies. The study of 577 patients showed that over 4 weeks, about half the patients improved after receiving first line immunotherapy.
• Second line immunotherapy includes rituximab, a monoclonal antibody that targets the CD20 receptor on the surface of B cells, thus destroying the self-reactive B cells. Cyclophosphamide, an alkylating agent that cross-links DNA and is used to treat both cancer and autoimmune diseases, has sometimes been proven to be useful when other therapies have failed.
• Other medications, such as alemtuzumab remain experimental.^[17]

New York Post reporter Susannah Cahalan wrote a book called Brain on Fire: My Month of Madness about her experience with the disease.^[18]

Dallas Cowboys defensive lineman Amobi Okoye spent 17 months battling Anti-NMDA receptor encephalitis. In addition to 3 months in a medically-induced coma, he experienced a 145-day memory gap and lost 78 pounds. He returned to practice on October 23, 2014.^[19]

Anti-glutamate receptor antibodies

• “My mysterious lost month of madness“, article by Susannah Cahalan about her experience with the disease (New York Post, October 4, 2009)
• “Quick Diagnosis of Rare Disease Leads to Remarkable Recovery“, article by Susan Conova (Columbia University Medical Center, September 1, 2010)
• Autoimmune Encephalitis Alliance – Resources for patients and doctors about Anti-NMDA and other forms of autoimmune encephalitis