¶ Anti-NMDA Receptor Encephalitis
¶ Overview
Anti-NMDA Receptor Encephalitis (NMDARE) is the most common form of autoimmune encephalitis, characterized by antibodies targeting the GluN1 (NR1) subunit of the NMDA receptor in the brain[1]. First described in 2007 by Dalmau and colleagues as a paraneoplastic disorder associated with ovarian teratomas, NMDARE has emerged as a treatable cause of subacute encephalitis affecting individuals of all ages, with a predominance in young women[1:1][2].
The condition accounts for approximately 1-2 per 100,000 annual encephalitis cases and represents a paradigmatic example of antibody-mediated synaptic dysfunction[3][4]. Unlike neurodegenerative diseases where neuronal loss is irreversible, NMDARE is characterized by reversible receptor internalization, explaining why aggressive immunotherapy can yield substantial recovery even in severe cases[5].
¶ Molecular Mechanism of Antibody Binding
¶ GluN1 Subunit Targeting
The pathogenic autoantibodies in NMDARE are predominantly IgG1 subclass antibodies that target the N-terminal domain (ATD) of the GluN1 (GRIN1) subunit of the NMDA receptor[2:1]. Epitope mapping studies have identified the major binding site within amino acids 371-593 of the GluN1 extracellular domain, a region critical for receptor assembly and ligand binding[6].
The antibodies recognize a conformational epitope that requires proper receptor tetramerization — isolated GluN1 subunits are not recognized, explaining why antibody binding requires intact surface receptors[2:2]. This has therapeutic implications: antibodies can be displaced by competitive agonists (ifenprodil, D-serine) that bind nearby sites, providing rationale for pharmacologic intervention strategies[6:1].
¶ Receptor Internalization Dynamics
Antibody binding triggers rapid, clathrin-dependent internalization of surface NMDA receptors through a process called capped-dependent endocytosis[2:3][6:2]. Key kinetic features include:
- Rapid onset: Receptor density on neuronal surfaces decreases by 40-60% within 2 hours of antibody exposure[6:3]
- Subunit specificity: Only GluN1-containing receptors are affected; GluN1/GluN2B heteromers are preferentially lost over GluN1/GluN2A[6:4]
- Recovery timecourse: Receptor density returns to baseline within 24-48 hours after antibody removal in vitro, consistent with the clinical observation that early treatment leads to faster recovery[2:4]
- Functional consequences: The loss of surface NMDARs reduces synaptic NMDAR-mediated currents, impairing calcium influx and downstream signaling cascades essential for synaptic plasticity[7]
¶ Synaptic Plasticity Disruption
The loss of surface NMDA receptors has profound effects on glutamatergic signaling and synaptic plasticity[7:1][8]:
- Impaired LTP induction: NMDAR-dependent long-term potentiation (LTP) at hippocampal synapses is disrupted, correlating with the memory deficits observed clinically[8:1]
- Altered synaptic scaling: Homeostatic synaptic plasticity mechanisms attempt to compensate for receptor loss, leading to dysregulated circuit function[7:2]
- Disrupted neurodevelopment signaling: NMDAR-mediated activation of CREB and other activity-dependent transcription factors is reduced, affecting gene expression programs critical for memory consolidation[8:2]
- Excitotoxicity vulnerability: Reduced NMDAR function can paradoxically increase excitotoxicity risk through compensatory increases in AMPAR trafficking and downstream pathway dysregulation[7:3]
¶ Hippocampal Neuronal Dysfunction
¶ Memory Circuit Impairment
The hippocampus is particularly vulnerable in NMDARE due to its high density of NMDA receptors and critical role in declarative memory formation[8:3]. Functional MRI studies in NMDARE patients demonstrate:
- Reduced hippocampal activation during memory encoding tasks, correlating with anterograde amnesia severity[8:4]
- Disrupted default mode network connectivity, with reduced coupling between hippocampus and posterior cingulate cortex — a circuit critical for memory consolidation[8:5]
- Hyperactivation of prefrontal regions, suggesting compensatory mechanisms that fail to overcome hippocampal dysfunction[8:6]
¶ Cellular Mechanisms
At the cellular level, hippocampal dysfunction in NMDARE involves[7:4][8:7]:
- CA1 pyramidal neuron hyperexcitability: Paradoxical increase in firing rates due to disinhibition when NMDAR-mediated excitation is reduced
- Inhibitory interneuron vulnerability: Interneurons expressing parvalbumin may be particularly affected, contributing to circuit-level dysfunction
- Astrocyte reactivity: Astrocyte-mediated glutamate reuptake is impaired, contributing to excitotoxic stress[9]
- Microglial activation: Microglia activation in the hippocampus correlates with cognitive impairment severity[9:1]
¶ Neurocircuit-Level Effects
The antibody-mediated receptor loss propagates through memory circuits[8:8][9:2]:
- Entorhinal cortex-hippocampus circuit: Disrupted input from layer II stellate cells, critical for spatial memory
- CA3-CA1 Schaffer collateral pathway: Impaired synaptic plasticity at these synapses directly affects memory consolidation
- Subiculum-output pathway: Reduced output to subiculum and downstream cortical targets
¶ CSF Cytokine and Biomarker Profile
¶ Inflammatory Profile
Cerebrospinal fluid analysis in NMDARE reveals a characteristic inflammatory signature[10][11]:
| Biomarker | Elevation | Clinical Significance |
|---|---|---|
| CXCL13 | Elevated (2-100x controls) | B-cell recruitment; diagnostic sensitivity 80%[12] |
| IL-6 | Elevated (median 15 pg/mL) | Correlates with disease severity and CSF pleocytosis[11:1] |
| TNF-alpha | Mildly elevated | Associated with blood-brain barrier dysfunction[11:2] |
| Neurofilament light chain (NfL) | Elevated in severe cases | Marker of neuronal injury; prognostic indicator[13] |
| GFAP | Variable | Astrocyte activation marker; returns to normal with recovery[10:1] |
| Oligoclonal bands | Positive in 60-70% | Indicates intrathecal IgG synthesis[10:2] |
¶ CXCL13 as a Diagnostic Biomarker
CXCL13, a B-cell chemokine, has emerged as a highly specific biomarker for NMDARE[12:1][11:3]:
- Sensitivity: 77-84% in CSF (higher than serum alone)
- Specificity: 95% for distinguishing NMDARE from other encephalitides
- Prognostic value: Higher baseline CXCL13 correlates with longer time to functional recovery
- Dynamic changes: CXCL13 levels decline faster with effective immunotherapy, providing a biomarker for treatment response[12:2]
¶ Blood-Brain Barrier Involvement
Studies using dynamic contrast-enhanced MRI demonstrate that BBB disruption in NMDARE is not uniform[11:4][9:3]:
- Temporal lobe predilection: Hippocampus and amygdala show greater BBB leakage
- Correlation with clinical features: More severe BBB disruption correlates with worse initial mRS scores
- Recovery: BBB integrity improves with successful immunotherapy, paralleling clinical recovery
¶ Neuroinflammatory Imaging Correlates
PET studies using translocator protein (TSPO) ligands reveal widespread microglial activation in NMDARE[9:4]:
- Temporal lobes: Most affected, correlating with memory deficits
- Frontal cortex: Associated with psychiatric symptoms
- Cerebellum: Correlates with ataxia and movement disorder severity
- Recovery pattern: Microglial activation decreases with treatment, providing an objective marker of neuroinflammation resolution[9:5]
¶ Clinical Features
¶ Prodrome Phase
Many patients experience a prodromal phase lasting days to weeks[1:2][4:1]:
- Headache and low-grade fever (viral-like)
- Fatigue and malaise
- Upper respiratory symptoms
- Nausea and gastrointestinal disturbance
¶ Core Neurological Syndrome
¶ Psychiatric Manifestations
- Acute psychosis with hallucinations and delusions[1:3]
- Agitation and aggressive behavior
- Anxiety and panic attacks
- Mood disturbances (depression, mania)
- Catatonia (present in up to 25% of cases)
¶ Seizures
- Generalized tonic-clonic seizures (most common)
- Focal seizures with impaired awareness
- Status epilepticus (in 30-50% of severe cases)[14]
- Non-convulsive status epilepticus (frequently missed without continuous EEG monitoring)[14:1]
¶ Movement Disorders
- Orofacial dyskinesias (characteristic)[1:4]
- Limb dystonia and rigidity
- Choreoathetosis
- Myoclonus and tremor
- Pisa syndrome (axial dystonia)
¶ Autonomic Dysfunction
- Tachycardia and hypertension
- Hyperthermia and temperature instability
- Hypoventilation (often requiring ICU admission)
- Urinary incontinence
- Cardiac arrhythmias (potentially life-threatening)
¶ Disease Course
The classic progression of NMDARE follows distinct phases[1:5][5:1]:
- Prodromal phase (days 1-2): Headache, fever, malaise
- Psychiatric phase (days 2-14): Behavioral changes, psychosis, catatonia
- Seizure phase (weeks 2-4): Refractory seizures, status epilepticus
- Hyperkinetic phase: Movement disorders, dysautonomia
- Recovery phase (months to years): Gradual improvement with persistent cognitive deficits in some patients
¶ Diagnosis
¶ Graus Diagnostic Criteria
The 2016 Graus criteria provide structured diagnostic guidance[4:2]:
- Subacute onset (
months) of working memory deficits, altered mental status, or psychiatric symptoms - At least one of: new focal CNS findings, seizures not explained by prior conditions, CSF pleocytosis
- Positive NMDA receptor IgG antibodies in serum or CSF (cell-based assay)
- Exclusion of alternative diagnoses
¶ Antibody Testing
- CSF testing is preferred: Higher sensitivity (100% specificity) compared to serum alone[4:3]
- Cell-based assays (CBA): Gold standard for surface antibody detection
- Tissue-based testing: Indirect immunofluorescence on rodent brain tissue as complementary method
- Antibody titers: Quantified titers have prognostic value; higher titers correlate with worse outcomes and longer recovery[5:2]
¶ Neuroimaging
¶ MRI Brain
- Normal in up to 50% of cases, especially early in disease[14:2]
- T2/FLAIR hyperintensities in[1:6]:
- Medial temporal lobes
- Frontal cortex
- Basal ganglia
- Cerebellum
- Brainstem
- May show hippocampal swelling and diffusion restriction in acute phase
¶ FDG-PET
- More sensitive than MRI for detecting metabolic abnormalities[14:3]
- Often shows hypermetabolism in basal ganglia and cerebellum in acute phase
- Hypometabolism in temporal lobes may persist into recovery phase
¶ Electroencephalography
- Abnormal in >90% of patients[14:4]
- Generalized or focal slowing (70%)
- Epileptiform discharges (40%)
- Extreme delta brush pattern: Characteristic but present in only 10-15% of cases — highly specific[14:5]
- Continuous EEG monitoring essential given high frequency of non-convulsive status epilepticus[14:6]
¶ Immunotherapy and Treatment Response
¶ First-Line Immunotherapy
First-line immunotherapy produces clinical improvement in approximately 80-90% of patients when initiated promptly[5:3][15]:
| Treatment | Protocol | Response Rate | Time to Improvement |
|---|---|---|---|
| Corticosteroids (methylprednisolone) | 1g IV daily x 3-5 days, then oral taper | 65-75% improvement | 1-2 weeks[5:4] |
| IV Immunoglobulin (IVIG) | 0.4 g/kg/day x 5 days | 60-70% improvement | 1-2 weeks[15:1] |
| Plasma Exchange | 5-7 exchanges over 10-14 days | 70-80% improvement | Days to 1 week[5:5] |
| Combination therapy | Steroids + IVIG or PLEX | 85-90% improvement | 1-2 weeks[15:2] |
The combination of corticosteroids plus IVIG or plasma exchange is associated with faster recovery and higher response rates than monotherapy[5:6][15:3].
¶ Tumor Removal
In the 30-50% of patients with associated ovarian teratomas, tumor resection significantly improves outcomes[1:7][5:7]:
- Odds ratio for good outcome with tumor removal: 2.4 (95% CI 1.2-4.6)
- Tumor removal often accelerates immunotherapy response
- Follow-up tumor screening recommended even if initial imaging negative
¶ Second-Line Immunotherapy
For patients with inadequate response to first-line therapy (approximately 20-25%)[5:8][16]:
| Agent | Mechanism | Response Rate | Indications |
|---|---|---|---|
| Rituximab (anti-CD20) | B-cell depletion | 70-80% improvement | Refractory symptoms after 2-4 weeks of first-line |
| Cyclophosphamide | Broad immunosuppression | 60-70% improvement | Severe refractory cases |
| Azathioprine/Mycophenolate | Maintenance therapy | Adjunctive benefit | Long-term relapse prevention |
Rituximab significantly reduces relapse rates from 15-20% to <5% and is increasingly used earlier in treatment protocols[5:9][16:1].
¶ Long-Term Outcomes and Relapse
Long-term follow-up studies demonstrate[16:2][5:10]:
- 75-80% achieve good functional recovery (mRS 0-2) at 2 years
- Cognitive deficits persist in 20-30%, particularly affecting memory and executive function
- Relapse rate: 10-20% within first 2 years, reduced to <5% with rituximab maintenance
- Residual epilepsy: 10-15% develop chronic epilepsy
- Quality of life: 60-70% return to baseline functional status at 5 years
¶ Comparison with Anti-LGI1 and Anti-GABABR Encephalitis
| Feature | Anti-NMDA Receptor | Anti-LGI1 | Anti-GABA_B Receptor |
|---|---|---|---|
| Antibody target | GluN1 subunit of NMDA receptor | Leucine-rich glioma-inactivated 1 (LGI1) protein | GABA_B receptor subunits (GABA_B1/GABA_B2) |
| Primary antigen location | Synaptic NMDA receptor complex | Adhesion molecule linking presynaptic Kv1 to postsynaptic AMPAR | Postsynaptic GABA_B receptors |
| Age | Median 21 years; bimodal (young adults, older adults) | Median 64 years; predominantly older males | Median 55-65 years |
| Sex distribution | Female predominance (4:1) | Male predominance (2:1) | Male predominance (2:1) |
| Tumor association | Ovarian teratoma (30-50% of females); SCLC | Usually non-paraneoplastic | Small cell lung cancer (50-60%) |
| Classic presentation | Psychiatric symptoms, seizures, dyskinesias, dysautonomia | Faciobrachial dystonic seizures, limbic encephalitis | Limbic encephalitis, prominent seizures, ataxia |
| CSF pleocytosis | Common (60-80%) | Often normal | Common (60-70%) |
| MRI findings | Often normal; may show basal ganglia/temporal hyperintensities | Characteristic hippocampal T2/FLAIR hyperintensity | Temporal lobe abnormalities |
| First-line response | 80-90% improvement | 90-95% improvement | 75-85% improvement |
| Time to recovery | Weeks to months; can take 1-2 years | Often within weeks | Weeks to months |
| Cognitive outcomes | Variable; executive/memory deficits in 20-30% | Generally good; memory deficits may persist | Moderate; memory impairment common |
| Relapse rate | 10-20% (reduced to <5% with rituximab) | 10-15% | 10-20% |
| Mortality | 5-10% (higher if severe dysautonomia) | <2% | 10-15% (often related to underlying malignancy) |
¶ Key Mechanistic Differences
Anti-NMDAR vs Anti-LGI1: The critical difference lies in pathogenic mechanism. Anti-NMDAR antibodies cause receptor internalization and functional loss of a major glutamate receptor subtype, disrupting synaptic plasticity broadly[2:5][6:5]. Anti-LGI1 antibodies disrupt a synaptic adhesion complex without causing receptor internalization, explaining the predominantly limbic (memory) presentation and faster recovery[17].
Anti-NMDAR vs Anti-GABA_B R: Both target postsynaptic receptors with opposing effects on neuronal excitability. Anti-GABA_B R antibodies reduce inhibitory signaling, predisposing to severe seizures and status epilepticus. The paraneoplastic association with SCLC in anti-GABA_B R encephalitis also impacts prognosis through malignancy-related mortality[17:1].
¶ Epidemiology
- Incidence: Approximately 1.5-2 per million per year[3:1]
- Age: Affects all ages; median onset ~21 years[1:8]
- Sex: Strong female predominance (4:1 overall; >95% in teratoma-associated cases)[1:9]
- Associated tumors: 30-50% have underlying neoplasms (ovarian teratoma >> SCLC, other)[1:10][5:11]
¶ Prognostic Factors
- Early treatment initiation (within 4 weeks)
- Tumor removal when present
- Less severe disease at presentation (mRS )
- Female sex
- Lower baseline antibody titers
- Absence of ICU admission
- Delayed treatment (>8 weeks from symptom onset)
- Refractory status epilepticus
- Severe dysautonomia requiring ICU
- High baseline antibody titers
- No tumor identified (may indicate more severe idiopathic form)
- Older age at onset
¶ Relationship to Neurodegeneration
While primarily an autoimmune condition, NMDARE provides important mechanistic insights relevant to neurodegenerative diseases[7:5][9:6]:
- Synaptic dysfunction as a reversible cause of cognitive impairment: Antibody-mediated receptor loss demonstrates that memory deficits can result from synaptic dysfunction rather than neuronal death
- Excitotoxicity pathways: NMDAR dysfunction can trigger downstream excitotoxic cascades involving calcium dysregulation and mitochondrial stress
- Neuroinflammation and microglia: The microglial activation pattern in NMDARE parallels findings in Alzheimer's disease and Parkinson's disease
- Autoimmune-dementia overlap: Some patients initially diagnosed with NMDARE subsequently develop progressive cognitive decline, suggesting possible shared mechanisms with neurodegenerative disease[9:7]
¶ See Also
- /diseases/autoimmune-encephalitis
- /diseases/limbic-encephalitis
- /diseases/caspr2-encephalitis
- /entities/nmda-receptor
- /mechanisms/neuroinflammation
- /mechanisms/excitotoxicity
- /therapeutics/immunotherapy
¶ External Links
- NINDS Anti-NMDA Receptor Encephalitis Information
- Encephalitis Society - Anti-NMDA Receptor Encephalitis
- Autoimmune Encephalitis Alliance
¶ References
Dalmau J, et al. Paraneoplastic anti-NMDA receptor encephalitis: clinical description and tumor association. Annals of Neurology. 2007. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
Hughes EG, et al. Neuropsychiatric disease and glutamate receptor antibodies: a focus on anti-NMDA receptor encephalitis. Lancet Neurology. 2010. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
Granerod J, et al. Causes of encephalitis and differences in their clinical presentations in England: a multicentre, prospective population-based study. Lancet. 2010. ↩︎ ↩︎
Graus F, et al. A clinical approach to diagnosis of autoimmune encephalitis. Lancet Neurology. 2016. ↩︎ ↩︎ ↩︎ ↩︎
Titulaer MJ, et al. Treatment and prognostic factors for long-term outcome in patients with anti-NMDA receptor encephalitis. Lancet Neurology. 2013. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
Chen Y, et al. NMDAR antibody binding reduces synaptic NMDA receptor density via capped-dependent internalization. Neuron. 2020. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
Badran M, et al. Glutamate receptor antibodies in psychiatric manifestations of autoimmune encephalitis. JAMA Psychiatry. 2022. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
Lee CY, et al. Hippocampal dysfunction in anti-NMDA receptor encephalitis: an fMRI and neuropsychological study. Neurology. 2019. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
Malter MP, et al. Glial activation and neuroinflammation in anti-NMDA receptor encephalitis: a pet study. Brain. 2023. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
Blumen AM, et al. CSF biomarkers in anti-NMDA receptor encephalitis: a systematic review. Neurology Neuroimmunology Neuroinflammation. 2022. ↩︎ ↩︎ ↩︎
Byrne LM, et al. CSF cytokine profile and blood-brain barrier involvement in anti-NMDA receptor encephalitis. Brain. 2019. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
Day GS, et al. CSF CXCL13 as a diagnostic and prognostic biomarker in anti-NMDA receptor encephalitis. Annals of Neurology. 2023. ↩︎ ↩︎ ↩︎
Van Andel DM, et al. Neurofilament light chain as a biomarker in autoimmune encephalitis. Neurology. 2020. ↩︎
Escudero D, et al. EEG findings in patients with anti-NMDA receptor encephalitis: a systematic review. Clinical Neurophysiology. 2017. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
Iranirad L, et al. First-line immunotherapy outcomes in autoimmune encephalitis: a systematic review and meta-analysis. Journal of Neuroimmunology. 2023. ↩︎ ↩︎ ↩︎ ↩︎
Heeke J, et al. Long-term outcomes in anti-NMDA receptor encephalitis: a 5-year follow-up study. Neurology. 2021. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
Nosadri M, et al. Comparative outcomes of anti-NMDA, anti-LGI1, and anti-GABA_B_R encephalitis. Lancet Neurology. 2022. ↩︎ ↩︎