Autoimmune Neurology
The study of autoimmune neurologic diseases is an emerging and rapidly evolving subspecialty
that encompasses the diagnosis and treatment of neurologic disorders with an autoimmune
basis. The last decade has seen a dramatic increase in the discovery of neural and
glial-specific antibodies and their target antigens.[1]
[2]
[3] The molecular identification of these antigenic targets provides insights into the
pathogenic mechanisms underlying many autoimmune neurologic disorders, including epilepsy.
Historically, the first autoantibodies were detected by binding to brain tissue sections
and targeted intracellular antigens (nuclear and cytoplasmic enzymes, transcription
factors, and RNA binding proteins).[4]
[5] These antibodies were frequently associated with underlying neoplasms and were termed
paraneoplastic. The identification of paraneoplastic or onconeural autoantibodies
can help direct the search for cancer and provides prognostic information. However,
paraneoplastic disorders tend to be poorly responsive to immunosuppression. Antigenic
proteins inside intact cells are inaccessible to circulating antibodies; thus, these
antibodies are not generally considered pathogenic. Rather, it is thought that CD8+
T cell-mediated inflammatory responses are the primary mechanism of neuronal destruction
in these disorders ([Fig. 1]).[6]
[7]
Fig. 1 Immunopathogenic mechanisms of paraneoplastic and nonparaneoplastic (idiopathic)
neural autoantibodies. In cases of paraneoplastic autoimmunity, tumor-targeted immune
responses are initiated by onconeural proteins expressed in the plasma membrane (red diamond) or in the cytoplasm, nucleus or nucleolus (green triangle) of certain tumors. These antigens are also expressed in neural cells and thus are
coincidental targets. Although there is evidence to support an analogous infectious-induced
mimicry in nonparaneoplastic autoimmunity (e.g., NMDAR encephalitis after HSV encephalitis),[108] the source of the antigen remains elusive in most cases. Antibodies targeting plasma
membrane antigens are effectors of injury (red): antibodies (red) directed at neural cell plasma membrane antigens (e.g., voltage-gated potassium
channels complex, NMDA, AMPA, GABA-B receptor) are effectors of cellular dysfunction
or injury through multiple effector mechanisms. These mechanisms include receptor
agonist or antagonist effects, activation of the complement cascades, activation of
Fc receptors leading to antibody-dependent cell-mediated cytotoxicity (ADCC), and
antigen internalization (antigenic modulation), thereby altering antigen density on
the cell surface. Antibodies targeting nuclear or cytoplasmic antigens are serum markers
of a T-cell effector mediated injury (green): Intracellular antigens (green triangles) are not accessible to immune attack in situ, but peptides derived from intracellular
proteins are displayed on upregulated MHC class-I molecules in a proinflammatory cytokine
milieu after proteasomal degradation, and are then accessible to peptide-specific
cytotoxic T cells. Antibodies (green, e.g., ANNA-1, CRMP-5) targeting these intracellular antigens (green) are detected in both serum and cerebrospinal fluid, but are not pathogenic. In clinical
practice, these antibodies serve as diagnostic markers of a T cell predominant effector
process. Modified with permission (Nature Publishing Group) from [Fig. 1] (antibodies can have a range of effector functions) from Diamond et al. Losing your
nerves? Maybe it's the antibodies. Nature Reviews Immunology 2009;9:449–456.
Not all autoantibodies targeting intracellular antigens have a strong association
with malignancy; however some like glutamic acid decarboxylase (GAD65) antibodies,
can respond to immunotherapy.[8]
[9] We explore the reasons for this further in later sections.
Autoantibodies can also target plasma membrane proteins (neurotransmitter receptors,
ion channels, water channels, and channel-complex proteins). By contrast, these antibodies
are probably pathogenic, as they can access their target proteins in vivo and potentially
alter their number or function ([Fig. 1]). Neurologic diseases associated with plasma membrane autoantibodies tend to be
immunoresponsive and are less frequently associated with malignancies.[1]
[3]
Autoimmunity and Epilepsy
Patients presenting with new-onset epilepsy pose a diagnostic and therapeutic challenge.
Despite exhaustive investigations, no underlying cause is found in ∼ 40% of adult-onset
epilepsies, and one-third of cases are intractable to antiepileptic drug therapy.[10] A link between immunity and inflammatory processes in epilepsy has long been recognized.
Evidence for this link was first suggested by the anticonvulsant activity of adrenocorticotropic
hormone (ACTH) and corticosteroids in some of the childhood epilepsies,[11]
[12]
[13] as well as the presence of chronic inflammation and partial response to immunotherapy
in patients with Rasmussen encephalitis.[14]
[15] The demonstration of proinflammatory molecules (such as interleukin- (IL-) 1β) in
the serum of patients with febrile seizures,[16] and the increased frequency of seizures in patients with systemic autoimmune disorders
such a systemic lupus erythematosus has provided further evidence for this link.[17] Inflammation also appears to be a central mechanism of seizure generation in some
experimental models of epilepsy, although the extent to which this applies in human
epilepsies remains to be elucidated.[18]
[19]
Perhaps one of the most promising developments in this field has come from the discovery
of multiple neural-specific autoantibodies occurring in patients with seizures and
status epilepticus intractable to antiepileptic drug therapy. As with other autoimmune
disorders affecting the nervous system, the first such antibodies discovered targeted
intracellular proteins ([Table 1]).[ 20] In recent years, antibodies targeting plasma membrane proteins have been identified
that have broadened the phenotypic spectrum of autoimmune epilepsies ([Table 2]).[21]
[22]
[23]
[24]
[25]
[26] Unlike patients with antibodies to intracellular targets, these patients respond
remarkably well to immunotherapy. Many neurologists still suspect a paraneoplastic
or autoimmune etiology for seizures only in the presence of syndromic features of
limbic or extralimbic encephalitis. This is unfortunate, as a growing body of evidence
supports an autoimmune basis for seizures in the absence these syndromic manifestations
for a subset of patients with drug-resistant epilepsy.[21]
[22]
[27]
[28]
[29]
[30]
[31]
Table 1
Neuronal nuclear cytoplasmic antibodies
|
Antibody
|
Oncological association
|
Frequency of tumor
|
Response to immunotherapy
|
Clinical relevance
|
Neurologic manifestations
|
|
ANNA-1 (anti-Hu)
|
Small-cell carcinoma
|
> 90%
|
Poor
|
High
|
Limbic/cortical encephalitis autonomic neuropathies, sensory neuronopathy, other peripheral
neuropathies
|
|
Ma1, Ma2
|
Testicular (Ma2); breast, colon, testicular (Ma1)
|
> 90%
|
Moderate
|
High
|
Limbic encephalitis, encephalomyelitis brainstem encephalitis, peripheral neuropathy
|
|
CRMP-5
|
Small-cell carcinoma, thymoma
|
> 90%
|
Poor
|
|
Encephalitis, optic neuritis and retinitis, myelopathy, neuropathy, Lambert–Eaton
myasthenic syndrome
|
|
Amphiphysin
|
Small-cell carcinoma, breast adenocarcinoma
|
> 90%
|
Poor
|
High
|
Limbic encephalitis myelopathy, stiff-person syndrome, cerebellar degeneration
|
|
GAD65
|
Thymoma; renal cell, breast or colon adenocarcinoma
|
< 5%
|
Moderate
|
High
|
Limbic/cortical encephalitis, stiff-person syndrome, stiff-person phenomena, brainstem
encephalitis, cerebellar degeneration
|
Abbreviations: ANNA-1, Antineuronal nuclear antibody type 1; CRMP-5, collapsin response
mediator protein-5; GAD65, glutamic acid decarboxylase 65.
Table 2
Antibodies target neural plasma membrane ion channels, receptors, and synaptic proteins
|
Antibody
|
Oncological association
|
Frequency of tumor
|
Response to immunotherapy
|
Clinical relevance
|
Neurologic manifestations
|
|
VGKC complex
LGI1+
Caspr2+
LGI1-;Caspr2-
|
Small-cell lung carcinoma, thymoma or adenocarcinoma of breast or prostate
|
< 20%
|
Good
Good
Moderate
|
High
High
Uncertain
|
Limbic encephalitis, dementia, hyponatremia, faciobrachial dystonic seizures, peripheral
nerve hyperexcitability, or both (Morvan syndrome)
|
|
NMDAR
|
Ovarian Teratomas, Testicular germinoma, neuroblastoma
|
Varies with age, gender, and ethnicity
|
Good
|
High
|
Psychiatric disturbances, dyskinesias, catatonia, central hypoventilation and autonomic
instability, opsoclonus–myoclonus
|
|
AMPAR
|
Thymic tumors, lung carcinoma, breast adenocarcinoma
|
70%
|
Good
|
High
|
Limbic encephalitis, nystagmus
|
|
GABA-B receptor
|
Small-cell lung carcinoma, other neuroendocrine neoplasia
|
70%
|
Good
|
High
|
Limbic encephalitis, orolingual dyskinesias
|
|
mGluR5 receptor
|
Hodgkin lymphoma
|
> 90%
|
Good
|
High
|
Cerebellar ataxia and limbic encephalitis (Ophelia syndrome)
|
|
DPPX
|
None described to date
|
|
Moderate
|
Uncertain
|
Encephalitis, sleep disturbances, myoclonus, hyperekplexia, dysautonomia, gastrointestinal
dysmotility
|
|
P/Q- and N-type VGCC
|
Small-cell carcinoma, breast
|
∼ 50%
|
Moderate
|
Uncertain
|
Encephalopathy, myelopathy, neuropathy, Lambert–Eaton myasthenic syndrome
|
|
gAChR
|
Adenocarcinoma, thymoma, small-cell carcinoma
|
|
Moderate
|
Uncertain
|
Dysautonomia, peripheral neuropathy
|
Abbreviations: AMPAR, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor;
Caspr2, contactin-associated protein-like 2; DPPX, dipeptidyl-peptidase-like protein-6;
GABA-B, γ-aminobutyric acid-B; gAChR, neuronal ganglionic nicotinic acetylcholine
receptor; LGI1, leucine rich glioma inactivated protein I; mGluR5, metabotropic glutamate
receptor 5; NMDAR, N-methyl-D-aspartate receptor; VGCC, voltage gated calcium channel;
VGKC, voltage gated potassium channel.
Clinical clues that help identify these patients include subacute onset (evolving
over days to weeks), an unusually high seizure frequency, intraindividual seizure
variability or multifocality, antiepileptic drug resistance, personal or family history
of autoimmunity, or history of recent or past neoplasia.[28]
[30]
[32]
[33] Rapidly evolving cognitive impairment, neuropsychiatric symptoms, evidence of multilevel
involvement of the central nervous system (CNS), or new-onset movement disorder, suggest,
but are not necessary, for the diagnosis. The detection of neural-specific autoantibodies
in serum or cerebrospinal fluid (CSF) can help to establish the diagnosis and guide
management. Other helpful paraclinical aids include the presence of inflammation on
magnetic resonance imaging (MRI) or fluorodeoxyglucose positron-emission tomography
(FDG-PET), as well as evidence of neuroinflammation in the CSF.
In this article, we summarize the clinical presentation, pathophysiology, and management
of autoimmune epilepsies for which neural antigen-specific autoantibodies serve as
diagnostic aids. Epilepsies where inflammation occurs, but the full pathogenic cascade
or role of antibodies is either not clear (e.g., Rasmussen encephalitis) or only hypothesized
fever-induced refractory epilepsy in school-age children (FIRES), are not discussed.
Clinical Characteristics and Pathophysiological Mechanisms of Neural-Specific Autoimmune
Epilepsy
Autoantibodies Specific for Intracellular Antigens
ANNA-1 (Anti-Hu)
Antineuronal nuclear antibody type 1 (ANNA-1, also known as anti-Hu) binds to the
Hu family of RNA binding proteins, which participate in posttranscriptional regulation
of RNA in postmitotic neurons.[34]
[35]
[36] ANNA-1 is highly associated with small-cell carcinoma, but can also occur with thymoma
and neuroblastoma.[33] Neurologic manifestations include, in decreasing order of frequency: peripheral
neuropathy, limbic encephalitis, encephalomyelitis, and gastrointestinal dysmotility.[33]
[34] Seizures, epilepsia partialis continua, and status epilepticus may occur in the
absence of other syndromic manifestations of limbic encephalitis.[37]
[38]
Seizures are probably caused by cytotoxic T-cell-mediated damage to both mesial temporal
and extralimbic structures. Autopsy studies of ANNA-1 seropositive patients with paraneoplastic
encephalomyelitis showed inflammatory infiltrates, gliosis, microglial nodules, and
neuronophagia.[39]
[40] Although perivascular infiltrates contained both B and T cells, it was T cells that
predominated in parenchymal infiltrates[39]: CD4+ T cells predominated in the perivascular regions, whereas CD8+ T cells were
pervasive in the interstitial spaces. Consistent with these observations of a cytotoxic
T-cell-mediated process, T cells expressing TIA-1 (a component of cytotoxic granules)
were observed in clusters around neurons, whereas C9neo, a marker of antibody-mediated
complement activation, was absent.[41]
Ma-1 and Ma-2 (Ta)
Ma-1 and Ma-2 (Ta) are neuronal nuclear proteins thought to play a role in RNA transcription
and regulation of apoptosis.[42]
[43] Dual Ma-1/Ma-2 positivity (also known as anti-Ma) is more common in females and
is associated with breast, ovarian, and colon cancer. Ma-2 positivity (also known
as anti-Ta) is associated with testicular germ-line cancers in males. Autoantibodies
binding to these antigens are associated with limbic or brainstem encephalitides.
Neurologic manifestations are likely cytotoxic T-cell mediated.[42]
CRMP-5/CV2-IgG
Collapsin response mediator protein-5 antibodies bind to a protein of the same name
involved in axonal development in the early nervous system.[44]
[45] The most common associated malignancies are small-cell-lung carcinoma and thymoma.[46] Common manifestations include cerebellar degeneration, chorea as a “basal ganglionitis,”
optic neuropathy, retinopathy, myelopathy, radiculoneuropathy, and autonomic dysfunction.
It can also present with limbic encephalitis and seizures. Autopsy findings have been
reported in one CRMP-5 IgG seropositive patient with optic neuritis, retinitis, and
encephalomyelopathy, and this demonstrated CD8+ T cell predominance.[47]
Amphiphysin
Amphiphysin antibodies target a synaptic vesicle-bound protein that works with dynamin
to retrieve membrane constituents after neurotransmitter exocytosis.[48] The most commonly associated malignancies are breast and small-cell-lung carcinoma.[49] Initially described in paraneoplastic stiff-person-like syndrome, its spectrum is
now appreciated to be much wider, including limbic and diffuse encephalitis.[50]
Neuropathological autopsy specimens of patients with amphiphysin-seropositive patients
demonstrate CD8+ T cell predominance.[50]
[51] Contrary to this hypothesis, one group reported electrophysiology in an animal model
of amphiphysin autoimmunity, as well as in vitro findings in mouse neuronal cell culture
that they interpreted as evidence of a direct functional effect of amphiphysin antibody.[52] The authors reported that purified amphiphysin IgG induced a stiff-person-like disorder
in rats when injected intrathecally. They also reported internalization of fluorescent
nanocrystal-tagged amphiphysin antibody into mouse hippocampal neurons.[52] The mechanism by which IgG is purported to be endocytosed and then become pathogenic
has not been demonstrated.
GAD65
GAD65 is the synaptic vesicle-associated antigen that catalyzes synthesis of γ-aminobutyric
acid (GABA) from L-glutamate. The 65 kDa isoform of GAD has been identified as an
autoreactive T-cell target in autoimmune diabetes mellitus type 1.[53] GAD65 is also often detected in patients with autoimmune neurologic disease, but
usually on an order of magnitude higher than in those with type 1 diabetes mellitus.[54]
[55] When occurring on its own it is rarely paraneoplastic; however, it can frequently
present in association with other onconeural antibodies, which should raise suspicion
for an underlying malignancy. Neurologically, it is associated with stiff-person syndrome,[54]
[56] cerebellar ataxia, encephalomyelitis, and extrapyramidal disorders,[57] but can also present with epilepsy.[27]
Because GAD65 is an intraneuronal antigen, GAD65 autoantibodies are unlikely to be
directly pathogenic, but may be associated with T-cell-mediated autoimmunity. Findings
in neuropathological studies of three patients with GAD65 antibody seropositive encephalitis
demonstrated multiple apposition of GrB+ cytotoxic T cells to neurons and a higher
CD8/CD3 ratio than patients with antibodies to plasma membrane antigens.[58] There was no evidence of IgG or complement deposition.[58] Contrary to this, 30 to 50% of patients with GAD65-associated autoimmunity respond
favorably to immunosuppression, including intravenous immunoglobulins,[30]
[59]
[60] suggesting similarities with disease associated with plasma membrane targets. A
likely explanation for this could be the coexistence of pathogenic autoantibodies
targeting as of yet unrecognized plasma membrane antigens, such as the glycine receptor
autoantibody recently demonstrated to coexist with GAD65 antibody in some patients
with stiff-person syndrome,[9] or the detection of GABA-B receptor antibodies in patients with GAD65-associated
encephalitis.[61]
Autoantibodies Specific for Plasma Membrane Antigens
Voltage-Gated Potassium Channel Complex Antibodies
The voltage-gated potassium channels (VGKCs) modulate neuronal excitability, axonal
conduction, and neurotransmitter release in the central, peripheral, and autonomic
system.[62] Only the Shaker Kv1 VGKCs sensitive to α-dendrotoxin (Kv 1.1, Kv 1.2, and Kv 1.6)
appear pertinent to neurologic autoimmunity.[63] Voltage-gated potassium channels form macromolecular complexes interacting with
cell-adhesion molecules and scaffolding proteins, including metalloproteinase-22 (ADAM22),
and a soluble binding partner of ADAM22-leucine-rich glioma-inactivated (LGI1) protein,
contactin-associated protein-2 (Caspr2), membrane-associated guanylate kinases, and
disintegrin.[64]
[65] Recent evidence suggests that some of the serum autoantibodies that bind macromolecular
complexes containing VGKCs, measured by radioimmunoprecipitation from solubilized
mammalian brain membranes, are actually directed at the extracellular domains of these
associated proteins.[66]
[67]
LGI1 and Caspr2 are the two antigenic targets in the potassium channel complex that
have been well characterized, but up to 54% of patients positive for potassium channel
complex antibodies have no positivity for either of these, suggesting that there is
at least one further target yet to be identified.[68] Although LGI1 antibodies are more frequently associated with limbic encephalitis,
and Caspr2 antibodies with peripheral nervous system manifestations, both antibodies
can affect all levels of the nervous system.[68]
Neurologic manifestations of VGKC autoimmunity include limbic encephalitis, as well
as peripheral nervous system hyperexcitability disorders.[69]
[70]
[71] More recently, a wider spectrum has been appreciated including reversible dementia-like
syndromes,[72]
[73]
[74] autonomic and peripheral neuropathy, pain syndromes,[75] and autoimmune epilepsy.[28]
[70]
[76] Seizures tend to be focal at onset, with mesial temporal or hippocampal onset more
common than extratemporal.[28]
[30]
[66] A new seizure type termed “faciobrachial dystonic seizure” has been described in
VGKC-complex antibody encephalitis, and, if present, it is virtually pathognomonic.[77] The incidence of cancer detection in patients with VGKC complex antibodies is relatively
low (< 20%), and the types of cancer diverse (small cell carcinoma, thyroid carcinoma,
thymoma, neuroblastoma, and adenocarcinoma).[70]
The pathophysiological mechanisms by which VGKC antibodies cause seizures remain ill-defined.
Presumably, these are partly due to alterations in the function of the VGKC caused
by antibody-mediated disruption of specific associated proteins such as LGI1. Linkage
analysis studies revealed that mutations of LGI1 are associated with autosomal dominant
lateral temporal epilepsy (ADLTE), an inherited epileptic syndrome characterized by
focal seizures with predominant auditory symptoms originating from the temporal lobe
cortex.[78]
[79] Knockout models of LGI1, ADAM22, or Kv1 result in severe epileptic phenotypes suggesting
a functional association between these proteins.[80]
[81]
[82] Similarly, mutations of CNTNAP2, the gene that codes for Caspr2, have been linked
to psychosis, autism, mental retardation, and intractable focal seizures.[83]
[84] In a recent study, investigators incubated rat hippocampi with IgG from a patient
with LGI1 positive limbic encephalitis and showed increased afterdischarges upon extracellular
stimulation in the stratum lucidum of the CA3 subregion compared with control IgG.[85] At the single cell level, IgG from the test subject was associated with CA3 pyramidal
cell excitability and dyssynchronization of excitatory stimulus coming from mossy
fibers unto these cells.[85] These effects could be reproduced when utilizing the VGKC antagonist α-dendrotoxin,
leading the authors to hypothesize that the effects observed were secondary to reduction
in VGKC function.[85]
Another possible mechanism of seizure generation may relate to tissue injury secondary
to inflammation. Neuropathologic evaluation of brain specimens of patients with VGKC
autoantibodies reveals perivascular and parenchymal lymphocytes, astrogliosis, and
diffuse microglial activation in biopsied or resected mesial temporal regions.[71]
[86]
[87] One detailed report of a postmortem specimen showed focal inflammation of both hippocampi
and amygdalae with loss of pyramidal neurons in the CA4 region.[88] Consistent with an antibody- and complement-mediated process, a more recent study
of both postmortem and biopsy specimens noted C9neo deposition in the cytoplasm and
surface of hippocampal CA4 neurons, as well as in dentate gyrus and cortical neurons.[58] These areas of complement deposition co-localized with regions of acute neuronal
death.[58] The above findings correlate well with changes observed in imaging. Up to 50% of
patients with VGKC-associated encephalitis have MRI evidence of inflammation and apoptosis
manifested as enlargement, T2 hyperintensity, enhancement, and restricted diffusion
of the mesial temporal lobe structures in the acute phase.[89] Serial MR imaging frequently shows mesial temporal sclerosis, even in successfully
treated patients ([Fig. 2]).[89] Neuronal cell loss in mesial temporal structures may account in part for the seizures
in this patient population and may explain why some patients need to remain on antiepileptic
drug therapy in spite of response to immunotherapy.
Fig. 2 A 25-year-old man with autoimmune voltage-gated potassium channels epilepsy. Imaging
at presentation shows enlargement and increased signal intensity in the bilateral
hippocampi (A) and bilateral amygdalae (B) on coronal fluid-attenuated inversion-recovery (FLAIR), with faint ill-defined enhancement
(arrowheads) of the hippocampi (C) on coronal contrast-enhanced T1. Follow-up coronal FLAIR imaging (D) at 4 years shows progression to bilateral mesial temporal sclerosis (arrows). From Kotsenas AL, Watson RE, Pittock SJ, et al. MRI findings in autoimmune voltage-gated
potassium channel complex encephalitis with seizures: one potential etiology for mesial
temporal sclerosis. AJNR Am J Neuroradiol 2014;35(1):84–89.[89] Copyrighted and used with permission from the American Society of Neuroradiology.
Ionotropic Glutamate Receptor Antibodies
N-Methyl-D-aspartate receptors (NMDARs) are glutamate-gated cation channels involved
in hippocampal synaptic transmission, neuronal plasticity, and long-term potentiation.[90]
[91] The receptors are heterotetrameric complexes formed by subunits derived from three
related families: NR1, NR2, and NR3. NMDAR-specific antibodies target the NR1/NR2
subunits.[92]
NMDAR encephalitis was first described in young women with ovarian teratomas,[93] but the disease is now known to affect men, infants, and patients without tumors
as well.[23]
[94] About one-third of women over 18 years with this disorder have a teratoma, but the
likelihood of finding a tumor varies according to age, sex, and ethnicity.[23] Approximately 5% of men have testicular germ-cell tumors, and ovarian teratomas
are more frequent in African Americans.[23] The classic presentation is characterized by a viral-like prodrome followed by psychiatric
disturbances with associated oro-lingual-facial dyskinesias and seizures. If untreated,
central hypoventilation, autonomic dysfunction, and even coma can follow, requiring
intensive care in most cases.[23] Seizures can occur early and then disappear later in the course of the disease either
due to rapid response to immunotherapy or due to progression to a more severe stage.
Unlike in patients with VGKC-associated epilepsy, there is no predilection for mesial
temporal structures, and seizures can arise temporally, extratemporally, or multifocally.[95] In a recent case series of continuous electroencephalogram (EEG) recording of 23
patients with NMDAR encephalitis, 60% had electrographic seizures without clinical
correlate, and 30% had a unique electrographic pattern that the investigators named
“extreme delta brush” because of similarities to waveforms seen in premature infants.[96] The specificity of this pattern is not yet determined, but its presence should raise
suspicion for NMDAR encephalitis.
Data from in vitro experiments are consistent with the hypothesis that NMDAR encephalitis
is antibody-mediated. Rat hippocampal neurons treated with patient CSF or NMDAR IgG
cause cross-linking and internalization of the target receptors, accompanied by reduced
synaptic NMDAR-mediated currents.[97] Antibody-depleting therapies and tumor removal optimize recovery, which can be complete
in up to 75% of patients. Cases in whom there is no underlying tumor identified have
a worse prognosis and a higher rate of relapse.
AMPA Receptors Antibodies
The α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors mediate
most fast excitatory neurotransmission in the brain. Antibodies directed at one or
both of the GluR1 and GluR2 subunits of the AMPA receptors have been associated with
limbic encephalitis.[24] In the initial case report only 4 of 10 patients had seizures, and 8 had MRI findings
of temporal lobe inflammation. Seven out of 10 patients had cancers subsequently identified
(small-cell and non-small-cell lung cancer, thymoma, and breast cancer).[24]
Similarly to NMDR antibodies, AMPA receptors are internalized following AMPA receptor
antibody binding. Patients tend to respond to immunotherapy and tumor removal when
applicable, but relapses appear to be common and patients may require long-term immunosuppression.[24]
GABA-B Receptor Antibodies
The metabotropic gamma-amino butyric Acid (GABA-B) receptors are G-protein coupled
receptors that are functionally linked to potassium channels and elicit both presynaptic
and slow postsynaptic inhibition. They are heterodimers of a GABA-B1 subunit (ligand
binding) and a GABA-B2 subunit (responsible for signaling and membrane targeting).
GABA-B receptors antibodies have been recently recognized as a cause of epilepsy associated
with limbic encephalitis. In the initial case series, all 15 patients had seizures
with 13 of them complaining of seizures as their presenting symptom.[25] Furthermore, three patients developed status epilepticus. The vast majority of seizures
in these patients were determined to be of temporal lobe onset, and most had MRI evidence
of inflammation in the temporal lobes.[25]
GABA-B antibodies have been associated with small-cell lung and breast cancer. They
also frequently coexist with GAD-65 antibodies and voltage-gated calcium channel antibodies.[3] GABA-B antibodies are thought to impair receptor function, but unlike NMDAR antibodies
they do not cause receptor internalization. Patients respond to immunosuppression
and tumor removal, if applicable; relapses are rare.
mGluR5 Antibodies
Metabotropic glutamate receptors (mGluR) are G-protein receptors that modulate neuronal
activity by activating intracellular signaling pathways. Eight different receptor
types are divided into three groups on the basis of structure and function: I (mGluR1
and mGluR5), II (mGluR2 and mGluR3), and III (mGluR4, mGluR6, mGluR7, and mGluR8).
Only mGluR1 and mGluR5 have been documented to be pertinent autoantigens. mGluR5 is
particularly prevalent in the hippocampus; autoantibodies to this receptor have been
identified in two patients with Hodgkin lymphoma and limbic encephalitis (Ophelia
syndrome).[26]
Other Antibodies
Both neuronal ganglionic nicotinic acetylcholine receptor (gAChR) and voltage-gated
calcium channel N-type and P/Q-type (VGCC N/VGCC P/Q) antibodies have been described
in patients suspected of having autoimmune epilepsy in a nonparaneoplastic context.[30] Up to 30% of patients with gAChR have cancer, usually adenocarcinoma, although some
have small-cell lung cancer and thymoma.[98] Clinical presentations most commonly include dysautonomia and peripheral neuropathy,
although occasionally encephalopathy.[98] P/Q-type VGCCs are detected in 85% of cases of Lambert–Eaton myasthenic syndrome,
and 50% of these cases are associated with small-cell lung cancer.[99] Both paraneoplastic and nonparaneoplastic encephalomyelopathy and cerebellar ataxia
have been described with coexisting N-and PQ-type VGCCs.[99] The pathogenic role of these antibodies in autoimmune epilepsy remains uncertain,
and may be secondary to coexisting autoantibodies such a GABA-B, which is known to
co-occur with VGCC antibodies.[25]
Recently, seizures have been described in patients harboring antibodies targeting
dipeptidyl-peptidase-like protein-6 (DPPX).[100]
[101] DPPX is a regulatory subunit of the voltage-gated A-type (rapidly inactivating)
Kv4.2, and is the principal channel responsible for transient, inhibitory currents
in the central and peripheral nervous systems. These currents regulate repetitive
firing rates and back-propagation of action potentials into neuronal dendrites.[102]
[103] There is a broad clinical presentation including encephalopathy, symptoms of central
hyperexcitability, myelopathy, dysautonomia, and seizures.[100]
[101]
Algorithmic Approach to Diagnosis and Treatment
Diagnosis
Early recognition of autoimmune epilepsy is paramount as prompt initiation of treatment
is associated with better outcomes,[28]
[30]
[104] but establishing the diagnosis can be challenging. Although suggestive, syndromic
manifestations of limbic or extralimbic encephalitis are not always present, and new-onset
epilepsy may be the sole presenting manifestation. Moreover, the presence of a neural
autoantibody does not always suffice to establish the diagnosis or determine prognosis.
Several studies have confirmed the presence of VGKC-complex antibodies (usually low
titers and without LGI1 or CASPR2 reactivity), and GAD65 antibodies in around 10%
of adults with longstanding epilepsies.[21]
[27]
[105] The pathogenic role of the antibodies in these cases remains unclear. Also, the
presence of GAD65 antibodies, even in cases of classic limbic encephalitis, does not
always predict response to immunotherapy.[30] Conversely, failure to detect an autoantibody in patients presenting with a clinical
picture suggestive of autoimmune epilepsy does not rule out the diagnosis, and some
of these patients can respond to immunosuppression.[30] The reasons for this are unclear, although it is possible that they harbor as yet
to be discovered pathogenic autoantibodies. Response to an immunotherapy trial can
support the diagnosis in these cases, and can help to identify those most likely to
respond to maintenance immunosuppressive therapy.[30] Such positive responses need to be interpreted with caution, however, as immunotherapy
is sometimes used to treat intractable epilepsies not proven to be autoimmune.[12]
[13]
Acknowledging the above difficulties, here we propose a diagnostic and therapeutic
approach based on available data and our own clinical experience ([Fig. 3]).
Fig. 3 Diagnostic and therapeutic approach to autoimmune epilepsy. AMPAR, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic
acid receptor; ANNA-1, Antineuronal nuclear antibody type 1; Caspr2, contactin-associated
protein-like 2; CRMP-5, Collapsin response mediator protein-5; DPPX, dipeptidyl-peptidase-like
protein-6; GABA-B, γ-aminobutyric acid-B; gAChR, neuronal ganglionic nicotinic acetylcholine
receptor; GAD65, glutamic acid decarboxylase 65; IVIG, intravenous immunoglobulin;
IVMP, intravenous methylprednisolone; LGI1, leucine rich glioma inactivated protein
I; mGluR5, metabotropic glutamate receptor 5; NMDAR, N-methyl-D-aspartate receptor;
PLEX, plasma exchange; VGCC, voltage gated calcium channel; VGKC, voltage gated potassium
channel.
When to Suspect Autoimmune Epilepsy
Clinicians should suspect autoimmune epilepsy in patients presenting with
-
Recent-onset cryptogenic epilepsy arising in the presence of a well-defined clinical
syndrome such as limbic encephalitis, faciobrachial dystonic seizures, or NMDAR encephalitis
or
-
Cryptogenic status epilepticus (including nonconvulsive status epilepticus) or
-
Subacute onset (maximal seizure frequency < 3 months) of cryptogenic epilepsy
Supportive clinical features include
-
Viral prodrome
-
Antecedent psychiatric symptoms
-
Antiepileptic drug resistance
-
History of systemic autoimmunity
-
History of recent or past neoplasia, particularly with a tumor known to be associated
with autoimmune epilepsy
These patients should undergo a thorough evaluation looking for paraclinical biomarkers
supportive of the diagnosis of autoimmune epilepsy, including neural-specific autoantibodies.
Care should be taken to rule out infective, metabolic, neoplastic, or structural causes
of epilepsy. Selective antibody testing is not advised because no single neural antibody
is definitively associated with seizures, and markers of occult cancer may be missed.[1]
[106]
Supportive paraclinical biomarkers include
-
Evidence of CNS inflammation on:
-
CSF (elevated protein, pleocytosis, oligoclonal bands, elevated IgG index or synthesis
rate
-
MRI brain scan (mesial temporal or parenchymal fluid-attenuated inversion-recovery
(FLAIR)/T2-weighted hyperintensity)
-
Functional imaging (FDG-PET) (hypermetabolism)
-
EEG showing extreme delta brush
-
Serological markers of systemic autoimmunity such as antinuclear antibody (ANA) or
thyroid peroxidase (TPO) antibody positivity
Once the diagnosis of autoimmune epilepsy is suspected based on clinical features
and presence of paraclinical biomarkers, neural specific autoantibody status dictates
management, and, together with response to an immunotherapy trial, helps to determine
prognosis.
Tumor Screening Based on the Results of Antibody Testing
Finding an autoantibody to intracellular onconeural proteins should prompt a thorough
search for an associated malignancy. If this reveals a neoplasm atypical for the paraneoplastic
antibody, then clinicians should consider the possibility of a second more typical
occult malignancy.[106] Computed tomography (CT) of the chest and pelvis is recommended as a screening tool,
but if this is negative FDG PET-CT is the next investigation.[107] Antibodies against plasma membrane proteins (NMDAR, AMPAR, GABA-B, mGluR5) can also
be paraneoplastic, and tumor surveillance may be indicated in these cases even if
an initial search for a malignancy is negative. Fluorodeoxyglucose positron-emission
tomography is not appropriate in a female with NMDAR encephalitis or in other patients
suspected of having a germ-cell tumor. Ultrasound scanning or MRI are preferred modalities
in these cases.
Treatment
Despite a paucity of formal evidence, a rational therapeutic approach can be devised
based on treatments that have been successfully applied previously in a variety of
autoimmune disorders. We typically use a protocol divided into acute and chronic therapeutic
phases ([Fig. 3]). Our standardized approach is guided by what we term the three “M's” of therapy:
Acute Therapy: Diagnostic Test
Response to immunotherapy in the acute treatment phase can have both diagnostic and
therapeutic value. We generally give a trial of high-dose intravenous (IV) methylprednisolone
or IV immunoglobulin ([Fig. 3], [Table 3]). We tend to reserve IV immunoglobulin for children (due to its perceived favorable
side-effect profile in children compared with corticosteroids), and for patients who
either have, or are at risk for, diabetes mellitus (i.e., patients seropositive for
GAD65 or IA-2 autoantibodies). Plasma exchange is also a useful first-line acute treatment
generally reserved for critically ill patients or when IV methylprednisolone or IV
immunoglobulin is poorly tolerated. After an initial 4- to 6-week trial of therapy,
patients should be re-evaluated for subjective and objective clinical improvement.
Repeat, imaging, EEG and/or CSF analysis may help if these had been abnormal at first
evaluation. If there is a strong suspicion, a trial with a different first-line agent
may be warranted even if a patient fails to respond to the first agent tried.[30] Rituximab and cyclophosphamide can be considered as second-line agents when there
is either no, or incomplete, response to first-line treatments ([Fig. 3], [Table 3]). The duration of the trial and timing of starting a second-line agent may vary
according to the severity of presentation and the degree of confidence in the diagnosis.
Status epilepticus in a patient with NMDAR antibody positivity or in a patient with
limbic encephalitis and LGI1 positivity may warrant more rapid escalation of immunotherapy
([Fig. 3]).
Table 3
Therapeutic options in patients with autoimmune epilepsy
|
Drug
|
Dose
|
Route
|
Frequency
|
Some common and severe side effects encountered
|
Therapeutic phase
|
|
Methylprednisolone
|
1000 mg
|
IV
|
Daily for 3–5 d,
then weekly for 4–8 wk
|
Insomnia, increased appetite, psychiatric disturbance, Cushing syndrome, diabetes,
cataracts, osteoporosis, hip avascular necrosis, skin thinning
Addisonian crisis on rapid withdrawal of physiologic doses of corticosteroid
|
Acute and chronic, then taper
|
|
Immunoglobulin
|
0.4 g/kg
|
IV
|
Daily for 3 d, then alternate weeks for 6–8 wk
|
Aseptic meningitis, deep venous thrombosis, headache, anaphylaxis, renal failure
|
Acute and chronic, then taper
|
|
Azathioprine
|
1 mg/kg/d to 2 mg/kg/d
|
PO
|
Two daily divided doses
|
Myelotoxicity, liver toxicity, hypersensitivity reaction, rash
|
Chronic
|
|
Mycophenolate mofetil
|
500 mg/d to
2000 mg/d
|
PO
|
Two daily divided doses
|
Myelotoxicity, CNS lymphoma, diarrhea, hypertension, renal failure
|
Chronic
|
|
Rituximab
|
1000 mg once,
then again 2 wk
later
|
IV
|
Every 6 mo
|
Infusion reactions, edema, hypertension, fever, fatigue, chills, headache, insomnia,
rash, pruritus, nausea, diarrhea, weight gain, cytopenias, neutropenic fever, liver
toxicity, hepatitis B reactivation
|
Acute (2nd line) and chronic
|
|
Cyclophosphamide
|
500 mg/m2/mo to
1000 mg/m2/mo (IV)
1 mg/kg/d to
2 mg/k/d (PO)
|
IV or PO
|
Monthly (IV)
Daily (PO)
|
Chronic infertility, alopecia mucositis, hemorrhagic cystitis, myelotoxicity
|
Acute (2nd line) and chronic
|
Abbreviations: CNS, central nervous system; IV, intravenous; mo, month; PO, by mouth.
A recent retrospective study looked at the use of an immunotherapy trial in evaluating
patients with suspected autoimmune epilepsy.[30] Sixty-two percent of patients improved overall, and of those receiving a second
agent after not responding to the first, 43% improved. Responders included 93% patients
with antibodies to plasma membrane antigens, 33% of patients seropositive for GAD
65 antibodies, and 33% of patients without detectable antibodies. Beyond the detection
of plasma membrane autoantibodies, the strongest predictor of response was a shorter
interval between symptom onset and starting treatment, highlighting the importance
of prompt initiation of immunotherapy when suspecting an autoimmune cause. Other reported
predictors of response include subacute onset, multiple seizure types, and CSF findings
of inflammation (elevated protein, oligoclonal bands or pleocytosis).[28]
[31]
Patients with antibodies to intracellular onconeural antigens tend to carry a worse
prognosis, but should still have the underlying malignancy treated. Clinicians should
consider a trial of immunotherapy as described above; some patients may respond preferentially
to agents targeting T-cell cytotoxic mechanisms, such as cyclophosphamide. However,
fewer than 10% of these patients make a substantial recovery.[1]
Long-Term Therapy
Objective improvements (> 50% reduction in seizure frequency) should prompt consideration
of a long-term plan for immunotherapy because symptoms relapse in most patients on
withdrawal of acute therapies ([Fig. 3], [Table 3]). Medium- to long-term treatment with corticosteroids or IV immunoglobulin is sometimes
required in these patients, but the overall the goal is eventually to stop these.
This may be achieved by adding an oral long-term immunosuppressant such as azathioprine
or mycophenolate mofetil, each of which has been used widely in treating organ-specific
autoimmune diseases such as myasthenia gravis. In our practice, we gradually extend
the interval between infusions of IV methylprednisolone or IV immunoglobulin over
a period of 4 to 6 months from weekly to fortnightly, every 3 weeks, and then monthly.
A faster taper of IV therapy can result in a relapse. When using daily oral prednisone,
a slow reduction from 60 mg of prednisone over months is preferable. It is important
to overlap corticosteroid or IV immunoglobulin treatment with the oral long-term immunosuppressant
(∼12 weeks for azathioprine and 8 weeks for mycophenolate mofetil). Some patients,
however, remain dependent on corticosteroids or IV immunoglobulin, despite of optimization
of oral long-term immunosuppression. In general, we prefer IV corticosteroid “pulse
therapy” over long-term oral corticosteroids, as the evidence suggests a more benign
side-effect profile and safer drug cessation. Rituximab and cyclophosphamide can be
considered as long-term therapies in patients who required these agents during the
acute phase or patients who relapse and fail to respond to first-line therapies.
There are no data to guide the duration of long-term immunosuppression in autoimmune
epilepsy. Some patients experience spontaneous remission, whereas others depend upon
lifelong immunosuppression to maintain remission. We generally start a trial of immunosuppressant
medication withdrawal after 2 years.
Antiepileptic Drugs in Patients with Autoimmune Epilepsy
Patients with autoimmune epilepsy are typically resistant to antiepileptic drugs;
generally, by the time immunotherapy is started, many are on multiple antiepileptic
drugs. There are no data comparing immunotherapy alone versus combined treatment in
autoimmune epilepsy. We generally keep patients on at least one antiepileptic drug
during the acute phase for symptomatic treatment, with a goal of eventually stopping
these if feasible. Some patients require maintenance therapy with an antiepileptic
drug; this may be related to chronic mesial temporal atrophy secondary to the initial
immunomediated process.
