IgG in serum of patient with thymoma-related paraneoplastic encephalitis binds to α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors in mouse brain tissue.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Zekeridou A, McKeon A, Lennon VA. Frequency of Synaptic Autoantibody Accompaniments and Neurological Manifestations of Thymoma. JAMA Neurol. 2016;73(7):853–859. doi:10.1001/jamaneurol.2016.0603
Thymoma is commonly recognized in association with paraneoplastic autoimmune myasthenia gravis (MG), an IgG–mediated impairment of synaptic transmission targeting the nicotinic acetylcholine receptor of muscle. Newly identified synaptic autoantibodies may expand the serological profile of thymoma.
To investigate the frequency of potentially pathogenic neural synaptic autoantibodies in patients with thymoma.
Design, Setting, and Participants
We retrospectively identified patients with histopathologically confirmed thymoma and serum available to test for synaptic autoantibodies (collected 1986-2014) at the Mayo Clinic Neuroimmunology Laboratory. We identified and classified 193 patients with thymoma into 4 groups: (1) lacking neurological autoimmunity (n = 43); (2) isolated MG (n = 98); (3) MG plus additional autoimmune neurological manifestations (n = 26); and (4) neurological autoimmunity other than MG (n = 26).
Main Outcomes and Measures
Clinical presentation and serum profile of autoantibodies reactive with molecularly defined synaptic plasma membrane proteins of muscle, peripheral, and central nervous systems.
Of the 193 patients with thymoma, mean patient age was 52 years and did not significantly differ by sex (106 women) or group. Myasthenia gravis was the most prevalent clinical manifestation (64%) followed by dysautonomia (16 patients [8%]) and encephalopathy (15 patients [8%]); 164 patients (85%) had at least 1 synaptic autoantibody, and 63 of these patients (38%) had at least 1 more. Muscle acetylcholine receptor was most frequent (78%), followed by ganglionic acetylcholine receptor (20%), voltage-gated Kv1 potassium channel-complex (13%), and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (5%). Less frequent were aquaporin-4, voltage-gated Kv1 potassium channel-complex related proteins (leucine-rich glioma-inactivated 1 and contactin-associated protein-like 2), glycine receptor, and γ-aminobutyric acid-A receptor. Synaptic autoantibodies were significantly more frequent in patients with neurological autoimmunity than in those without and were most frequent in patients with neurological manifestations other than or in addition to MG.
Conclusions and Relevance
Synaptic autoantibodies, particularly those reactive with ion channels of the ligand-gated nicotinic acetylcholine receptor superfamily (namely α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid, glycine, and γ-aminobutyric acid-A receptors), were prevalent in patients with thymoma. Autoantibodies of this extended spectrum may enhance autoimmune serological testing as an aid to preoperative thymoma diagnosis. Detection of currently known synaptic autoantibody specificities absent from this profile have potential algorithmic usefulness as negative predictors for thymoma (as recognized for neuronal voltage-gated calcium channel autoantibodies).
Paraneoplastic neurological autoimmunity is a recognized accompaniment of thymic epithelial neoplasia. Myasthenia gravis (MG), the most frequent association, is predicted to develop in 25% of patients with thymoma. It is further estimated that thymoma will be found in 21% of patients who have MG.1,2 Myasthenia gravis is an IgG-mediated neuromuscular synaptic disorder that targets the muscle nicotinic acetylcholine receptor (AChR). In 2004, the Mayo Clinic Neuroimmunology Laboratory reported the frequency of a limited repertoire of neuronal autoantibodies in 201 patients with thymoma. Glutamic acid decarboxylase 65 was the most frequent, followed by voltage-gated Kv1 potassium channel-complex (VGKC-complex), collapsin response-mediator protein 5 (CRMP5), and antineuronal nuclear antibody 1, all more frequent than ganglionic AChR.3 Peripheral nerve hyperexcitability (neuromyotonia), a manifestation of VGKC-complex autoimmunity, is known to associate with thymoma. When accompanied by central hyperexcitability (seizures, hyperhidrosis, and insomnia), the disorder is known as Morvan syndrome.3,4 Additionally recognized neurological accompaniments of thymoma include dysautonomia (especially gastrointestinal dysmotility), myelopathy (including stiff-man manifestations), encephalitis, cranial and other peripheral neuropathies, and myositis (sometimes with myocarditis).3,5,6 Regardless of MG, the most commonly detected autoantibody accompaniments of thymoma are specific for muscle AChR and skeletal and cardiac sarcomeric (“striational”) proteins.3
Several neural autoantibodies discovered in the past decade in both paraneoplastic and idiopathic contexts are potentially pathogenic for neuronal or astrocytic synapses.7-9 As in MG, their neurological manifestations (eg, encephalitis and neuromyelitis optica) generally respond favorably to antibody-depleting immunotherapy.8,9 Seven of the new synaptic autoantibodies described in case reports of thymoma are specific for the aquaporin-4 (AQP4) water channel, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR), γ-aminobutyric acid-A receptor (GABAAR), glycine receptor (GlyR), N-methyl-d-aspartate receptor (NMDAR), leucine-rich glioma inactivated 1 (LGI1), and contactin-associated protein-like 2 (Caspr2).4,10-15 The objective of this study was to investigate in patients with thymoma the frequency of autoantibodies potentially pathogenic for neuronal, astrocytic, and muscle synapses. Although glutamic acid decarboxylase 65 was the most frequently detected neuronal autoantibody in our 2004 report (22% seropositive overall), it was excluded from this study because the antigen’s cytoplasmic vesicular location in GABAergic nerve terminals precludes accessibility to pathogenic autoantibodies.3,16 Likewise, we did not test for voltage-gated calcium channel antibodies (P/Q-type or N-type) because neither was found among the 201 patients studied in 2004.3
Question How frequently does thymoma associate with potentially pathogenic autoantibodies targeting muscle, peripheral, and central nervous system synapses?
Findings In this study, at least 1 synaptic autoantibody was detected in 85% of 193 patients with thymoma, regardless of neurological manifestations. Autoantibody specificities included muscle acetylcholine receptor, ganglionic acetylcholine receptor, voltage-gated Kv1-potassium channel-complex, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor, and others.
Meaning An extended spectrum of synaptic autoantibodies, many reactive with ligand-gated nicotinic muscle acetylcholine receptor superfamily ion-channel receptors, enhance autoimmune serological testing as an aid to thymoma diagnosis, regardless of neurological presentation.
This study was approved by the Mayo Clinic’s institutional review board. Patients with histopathologically confirmed thymoma (6 had thymic carcinoma) whose records indicated research consent (Mayo Clinic patients) or whose physicians had obtained verbal consent (non–Mayo Clinic patients) were identified in the Mayo Clinic Neuroimmunology Laboratory database.
Sera (n = 193) were submitted in the period from 1986 to 2014 for service autoantibody testing (n = 150) or collected under an institutional review board–approved research protocol (n = 43). Available serum from patients in our 2004 report (collected through 2003)3 were supplemented by 120 patient specimens collected after 2003. For patients with multiple serum specimens available, we tested when possible the earliest specimen received.
We reviewed clinical records for patients seen at the Mayo Clinic (84%) and physician-provided information for patients not seen at the Mayo Clinic (16%). We classified the patients into 4 diagnostic groups: (1) lacking neurological autoimmunity; (2) isolated MG; (3) MG plus additional neurological manifestations; and (4) autoimmune neurological disorder other than MG. In view of the retrospective nature of this study and the multifocality of documented neurological findings, we categorized manifestations in a descriptive rostro-caudal neuroanatomical order: (1) encephalopathy (ie, subacute onset of rapid cognitive decline with altered sensorium or consciousness and/or cognitive abnormalities including memory loss, personality and behavioral changes [eg, agitation, catatonia, and psychosis], and aphasia); (2) epilepsy; (3) movement disorder; (4) cranial neuropathy; (5) myelopathy; (6) peripheral nerve hyperexcitability (with or without central manifestations); (7) peripheral neuropathy; (8) dysautonomia; (8) myopathy; or (10) other.
All sera were stored long term at −30°C. Assays for autoantibodies other than GABAAR and GlyR were performed in the neuroimmunology laboratory using clinically validated contemporary methods.17,18 Radioimmunoprecipitation assays for muscle and ganglionic AChR antibodies were performed at serum receipt or retrospectively from 2005 (when service ganglionic AChR antibody testing was implemented). When serum volume was limiting, testing was prioritized as follows: (1) VGKC-complex antibodies: radioimmunoprecipitation screening with antigenic protein solubilized from pig brain membranes and ligated with 125I-labeled α-dendrotoxin; (2) IgG-specific for NMDAR, AMPAR, γ-aminobutyric acid-B receptor, metabotropic glutamate receptor-1 (mGluR1), metabotropic glutamate receptor-5 (mGluR5), LGI1, and Caspr2: cell-binding assays on fixed human embryonic kidney cells transfected with plasmids encoding the respective antigens (Euroimmun AG); (3) AQP4-IgG: flow cytometry using live human embryonic kidney cells transiently cotransfected with human AQP4-M1 isoform and AcGFP18 (insufficient serum available to test 5 patients); and (4) GABAAR and GlyR IgGs: all available sera from groups 1, 3, and 4 and sera from 11 group 2 patients who were positive for an autoantibody other than muscle and ganglionic AChR were tested by in-house live cell-binding assays in the Laboratory of Experimental Neuroimmunology, Biomedical Research Institute August Pi I Sunyer, University of Barcelona, Spain.
Two-tailed t test and 1-way analysis of variance were used to compare continuous variables that were normally distributed. Mann-Whitney test was used to compare the mean frequency of autoantibodies between non-normally distributed groups (nonparametric statistics). Fisher exact test was used to evaluate differences in antibody frequencies between groups. A P value less than .05 was considered significant; we adjusted the value to less than .008 for multiple comparisons between subgroups (frequencies of autoantibodies in each group) according to the Bonferroni method. Data were analyzed using IBM SPSS statistics version 2.0 (IBM).
Forty-three of the 193 patients with thymoma were neurologically asymptomatic (group 1); 98 had isolated MG (group 2); 26 had MG plus other neurological manifestations (group 3); and 26 had 1 or more neurological diagnoses other than MG (group 4). The mean age (52 years; median, 51.5 years; range, 7-86 years) did not differ significantly for male and female patients (106 patients were women) nor between the 4 groups; the sex ratio did not significantly differ for patients with and without MG. At least 1 synaptic autoantibody was found in 164 patients (85%) (Table 1).
Autoantibody frequencies differed in each of the 4 groups (P < .001). Patients without clinically evident neurological autoimmunity had, as anticipated, significantly fewer autoantibodies than patients in any other group (P < .001). Comparison of groups 1 and 2 (neurologically asymptomatic and patients with isolated MG) with groups 3 and 4 (patients with additional manifestations of neurological autoimmunity) revealed a significantly higher mean frequency of autoantibodies in the latter groups (P = .003).
The frequencies of AQP4, VGKC-complex, Caspr2, AMPAR, muscle AChR, and ganglionic AChR autoantibody specificities differed in the 4 groups (P < .05; Table 2). Muscle AChR autoantibody was most frequent (78% positive); all patients who had MG were seropositive, and 39% of patients without MG were seropositive (Table 1). Ganglionic AChR antibody was always found in the company of another autoantibody, with a single exception.
Voltage-gated Kv1 potassium channel-complex autoantibody was found in 25 of 193 patients (13%; mean value, 0.14 nmol/L; normal range, 0.00-0.02 nmol/L). Values for patients with MG (0.17 nmol/L) and without MG (0.12 nmol/L) did not differ significantly. Voltage-gated Kv1 potassium channel-complex–IgG of Caspr2 subspecificity was detected in only 7 patients (3 negative by radioimmunoprecipitation, 1 with isolated MG, 1 with MG plus gastrointestinal hypermotility, and 1 with hyperexcitability of both central and peripheral nervous systems [Morvan syndrome]). Voltage-gated Kv1 potassium channel-complex–IgG of LGI1 subspecificity was detected in 2 patients. The frequency of VGKC-complex-specific IgG by immunoprecipitation in the 193 patients was significantly higher than the frequency of LGI1 and Caspr2 autoantibodies combined (P = .003).
The next most frequently detected synaptic autoantibody, AMPAR-IgG, was detected in 9 sera by both transfected cell-binding assay and by its characteristic hippocampus-predominant staining pattern in tissue-based immunofluorescence assay (Figure). Five of the 9 cases had encephalopathy with limbic or cortical manifestations. One of those 5 patients additionally had catatonia and another had rigidity, insomnia, and peripheral neuropathy. A sixth AMPAR-IgG–positive patient had isolated hemichorea (a syndromic manifestation of her coexisting CRMP5-IgG).19 The 3 remaining AMPAR-IgG-positive cases had no neurological symptoms or signs other than MG; 1 (followed up for 3 years) had coexisting gastrointestinal hypermotility.
Aquaporin-4–IgG was found in 3 cases. One patient had optic neuritis and clinical signs consistent with neuromyelitis optica spectrum disorder (encephalopathy, seizures, and signs of circumventricular organ involvement).20 The second patient had longitudinally extensive transverse myelitis, and the third had no neurological manifestations at presentation (non–Mayo Clinic patient without follow-up information).
GABAAR-IgG was detected in 2 of 98 sera tested. One was from a 45-year-old man with peripheral neuropathy; his serum was additionally positive for VGKC-complex-IgG (0.04 nmol/L). The second was from the woman with MG plus gastrointestinal hypermotility who had multiple coexisting autoantibodies (Caspr2-IgG, muscle AChR, AMPAR-IgG, and CRMP5-IgG).
GlyR-IgG was detected in 2 women, both neurologically asymptomatic; 1 was 39 years old (followed up for 1 year) and the other was 84 years old (no follow-up information).
We did not detect NMDAR, dipeptidyl-peptidase-like protein-6, mGluR1, or mGluR5 IgG specificities in any serum. Interestingly, γ-aminobutyric acid-B receptor–IgG was detected in a serum specimen from the aforementioned female patient with gastrointestinal hypermotility and MG who had multiple coexisting autoantibodies (drawn 2 years later than the specimen tested in this study).
Myasthenia gravis diagnosis was made in 124 patients (64% of all thymoma cases). The next most frequent neurological presentations were dysautonomia in 16 patients (8%; half had gastrointestinal dysmotility), encephalopathy in 15 patients (8%), and seizures in 12 patients (6%) (Tables 3 and 4). Less common manifestations were peripheral nerve hyperexcitability in 9 patients (5%; one-third had coexisting central and autonomic hyperexcitability [Morvan syndrome]), peripheral neuropathy in 9 patients (5%), myopathy in 5 patients (3%), cranial neuropathies in 3 patients (2%), myelopathy in 2 patients (1%), and movement disorders in 2 patients (1%). Rarer neurological manifestations were rigidity, pain, and uveitis.
Patients with dysautonomia commonly had MG (frequency was not significantly different in patients without MG). Of the 15 patients with encephalopathy, 9 had limbic encephalitis and 3 had Morvan syndrome (peripheral, central, and autonomic hyperexcitability); the remaining 3 did not fulfill criteria for a classic encephalopathic paraneoplastic syndrome. Eleven of these 15 patients had seizures and 3 had a psychiatric disorder (psychosis, catatonia, or delusions). The single patient with isolated epilepsy was VGKC-complex-IgG–positive by radioimmunoprecipitation assay but lacked detectable LGI1-IgG or Caspr2-IgG. Six of 9 patients with peripheral nerve hyperexcitability were VGKC-complex-IgG positive (one had Caspr2-reactive IgG). A seventh patient (male) had Morvan syndrome and Caspr2-IgG without VGKC-complex-IgG detectable by radioimmunoprecipitation. Four of 5 patients stated to have myopathy had coexisting MG. No histopathological data were available; 1 had dermatomyositis by history.
Muscle AChR autoantibody, the pathogenic determinant of paraneoplastic MG, was found in 14 (33%) of the neurologically asymptomatic patients and also in 13 (50%) of the neurologically symptomatic patients without evidence of MG. One or more synaptic autoantibody specificities other than muscle AChR were detected in serum of 6 (14%) of the 43 neurologically asymptomatic patients (VGKC-complex, 4; ganglionic AChR, 1; AQP4, 1; LGI1, 1; Caspr2, 1; and GlyR, 2). Serum from 32 of the 98 patients (33%) with isolated MG (all muscle AChR autoantibody-positive) had 1 or more additional synaptic autoantibodies (ganglionic AChR, 26; VGKC-complex, 7; Caspr2, 1; and AMPAR, 1).
This study extends the onconeural autoantibody profile of thymoma beyond that previously recognized (striational, muscle AChR, ganglionic AChR, glutamic acid decarboxylase 65, CRMP5, and VGKC-complex).3 The 4 additional synaptic autoantibodies that we have identified as potential markers of thymoma per se (regardless of neurological manifestation) enhance the clinical predictive value of comprehensive serological evaluation in the preoperative investigation of patients presenting with an anterior mediastinal mass.21 It is noteworthy that 3 are specific for ionotropic channels of the nicotinic AChR superfamily (AMPAR, GlyR, and GABAAR), and the fourth is specific for the AQP4 water channel, a key component of astroneuronal, astroendothelial, and astroglial synapses.7 The lack of NMDA, DPPX, mGluR1, and mGluR5 receptor–specific autoantibody detection suggests that IgGs specific for those ionotropic and metabotropic receptors may have algorithmic usefulness as negative predictors for thymoma, as is recognized for neuronal voltage-gated calcium channel autoantibodies.3,21
Despite their high prevalence, the frequencies of neurological disorders we report cannot be considered representative for patients with thymoma because most sera were submitted by neurologists requesting autoantibody testing.
Acetylcholine-receptor–specific autoantibodies, both muscle-type and ganglionic-type, and VGKC-complex autoantibodies have been reported previously in patients with neurologically asymptomatic thymoma.3 A novel finding of this study is the high prevalence of other synaptic autoantibodies in asymptomatic patients. Of the synaptic specificities found, AMPAR-IgG was the fourth most frequent. All 9 previously reported patients with AMPAR-IgG in the setting of thymoma had paraneoplastic autoimmune encephalitis, with poorer outcome when AMPAR-IgG coexisted with other onconeural autoantibodies.10,14,22,23 Three AMPAR-IgG–seropositive patients in this study had MG: 2 as the sole neurological diagnosis and 1 with dysautonomia without any manifestation of encephalopathy (3 years of follow-up). The few patients who were seropositive for GlyR or GABAAR autoantibodies were neurologically asymptomatic or had MG (1 had coexisting dysautonomia).
The detection of synaptic autoantibodies in patients with thymoma lacking neurological symptoms supports the role of the thymic neoplasm as initiator of the immune response against neural autoantigens. Ascertainment of the likelihood of neurological symptom development would necessitate longer-term follow-up. Aquaporin-4–IgG positivity has been documented with MG and thymoma up to 16 years before the clinical onset of neuromyelitis optica.24 Factors determining the timing of neurological symptom onset in patients who are seropositive for synaptic autoantibodies might include the titer or fine specificity of an evolving humoral immune response. For example, the dominant plasma membrane antigen specificity of the antibody products could change from a cytoplasmic or oncofetal-restricted epitope specificity (ie, nonpathogenic) to an extracellular or mature neural isoform specificity (ie, potentially pathogenic).
Radioimmunoprecipitation evidence of VGKC-complex antibody was lacking in 3 patients in whom Caspr2-IgG was detected. Two patients had MG and 1 patient had symptoms of peripheral nerve hyperexcitability with central and autonomic manifestions (Morvan syndrome). We conclude that the seropositivity yield for VGKC-complex autoantibody detection increases incrementally with testing for subtype specificities.
The limitations of this study are its retrospective design, the long period of serum storage, and the short follow-up period for asymptomatic cases. We confined our analysis to the clinical presentation at the time of thymoma diagnosis and tested the earliest sample available for each patient.
These findings raise intriguing immunobiological questions. What determines the exceptional potency of thymoma as an initiator of paraneoplastic neurological autoimmunity, and why are channel proteins highly immunogenic in the context of thymoma, in particular proteins of the nicotinic AChR superfamily?
Corresponding Author: Vanda A. Lennon, MD, PhD, Immunology and Neurology, Neuroimmunology Laboratory, Mayo Clinic College of Medicine, 200 First St SW, Rochester, MN 55905 (firstname.lastname@example.org).
Accepted for Publication: February 18, 2016.
Published Online: May 2, 2016. doi:10.1001/jamaneurol.2016.0603.
Author Contributions: Drs Zekeridou and Lennon had full access to all the study data and take responsibility for its integrity and the accuracy of the data analysis.
Study concept and design: Zekeridou, Lennon.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Zekeridou.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Zekeridou.
Administrative, technical, or material support: Zekeridou.
Study supervision: Lennon.
Conflict of Interest Disclosures: Dr Lennon reports royalties from marketing of aquaporin-4–IgG diagnostic kits, performance of aquaporin-4–IgG diagnostic tests outside of Mayo Clinic, and US patents pending for functional assays for aquaporin-4 autoantibodies and aquaporin-4 autoantibody as a cancer marker. No other disclosures were reported.
Funding/Support: Dr McKeon receives research support from Medimmune.
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank Nancy Peters and Patrick Ingvaldson (Neuroimmunology Laboratory, Mayo Clinic, Rochester, Minnesota) for performing immunofluorescence and immunoprecipitation assays and Masoud Majed, MD (Neuroimmunology Laboratory, Mayo Clinic, Rochester, Minnesota), and Ross Dierkhising (Center for Clinical and Translational Science, Mayo Clinic, Rochester, Minnesota), for consultation in statistical analyses. We also thank Josep Dalmau, MD, PhD (Laboratory of Experimental Neuroimmunology, 1 Biomedical Research Institute August Pi I Sunyer, University of 2 Barcelona, Spain), and his laboratory staff (Makoto Hara, MD, and Marianna Spatola, MD) for performing assays for γ-aminobutyric acid-A and glycine receptor autoantibodies and for collegially facilitating the study on a blinded basis and critically reviewing the manuscript. No compensation was provided.