• Epilepsy;
  • Autoantibodies;
  • Glycine receptor;
  • Voltage-gated potassium channel;
  • N-methyl-d-aspartate receptor


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Disclosure
  7. References


Neuronal antibodies have been identified in patients with seizures as the main or sole symptom. Our aim was to investigate the prevalence of these autoantibodies in patients with focal epilepsy of unknown cause (FEoUC) and in the group having mesial temporal lobe epilepsy with hippocampal sclerosis (MTLE-HS).


We studied anti-neuronal antibodies of consecutive adult patients diagnosed with FEoUC and MTLE-HS in our epilepsy center. The clinical and laboratory features of antibody-positive patients were compared with those of seronegative patients. The responses to therapy have also been investigated.


Sera from 81 patients with epilepsy were tested. We found antibodies against glycine receptor (GLY-R) in 5 (6.2%), contactin-associated protein 2 (CASPR-2) in 4 (4.9%), N-methyl-d-aspartate receptor (NMDA-R) in 2 (2.5%), and voltage-gated potassium channel (VGKC)-complex in 2 (2.5%) of our patients with epilepsy. Psychotic attacks and nonspecific magnetic resonance imaging (MRI) white matter changes (WMCs) showed significant associations in seropositive patients (p = 0.003 and p = 0.03, respectively). Poor drug-response rates and total seizure counts were also higher in the seropositive patients but without reaching statistical significance. Three seropositive patients with previous epilepsy surgery showed typical histopathologic results for MTLE-HS, but not inflammatory changes. Moreover, some patients harboring these antibodies partly benefited from immunotherapy.


We detected neuronal antibodies in one sixth of patients with focal epilepsy, GLY-R antibodies being the leading one. Psychosis or nonspecific MRI WMCs were frequent in the seropositive group. Our results suggested that relevant antibodies should be screened for a treatment possibility in these groups.

Thumbnail image of

Dr Ekizoglu has received her neurology training in UEMS Board certified, Istanbul Medical School.

Epilepsy is a common neurologic disorder showing resistance to antiepileptic drugs (AEDs) in at least 25–30% of the cases, and its etiology is still unknown.[1, 2] On the other hand, there is growing evidence that autoimmunity might play a role in epilepsy. A variety of serum antibodies to specific neuronal proteins has recently been identified in ordinary patients with epilepsy. Other than the first reported typical cases with limbic encephalitis or encephalopathy, the following antigenic targets have been shown in patients who present with seizures as the main or sole symptom in recent years: voltage-gated potassium channel (VGKC)-complex,[3-5] N-methyl-d-aspartate receptor (NMDA-R),[3, 6] glutamic acid decarboxylase (GAD),[3-5, 7] and glycine receptors (GLY-Rs).[3]

A recent study demonstrated that 11% of two large unselected epileptic cohorts of new onset and with chronic course had antibodies to one or more of VGKC, NMDA-R, GAD, or GLY-R.[3] Furthermore, an association between the autoimmune etiology and AED resistance has been reported, and a potential benefit of immunotherapy in improving seizure control has been suggested in some patients with antibodies to VGKC or GAD.[5, 8] Some of the reported seropositive patients also had mesial temporal lobe involvement.[8]

The exact clinical associations of these autoantibodies in various epilepsy syndromes are currently not well-established, and the clinicians want to know how to predict those patients who possibly harbor autoantibodies.

Our aim was to investigate the prevalence of these autoantibodies in consecutive patients diagnosed with established focal epilepsy of unknown cause (FEoUC) and also in the group having mesial temporal lobe epilepsy with hippocampal sclerosis (MTLE-HS), which is mostly an AED-resistant and clinically distinct constellation, showing some inflammatory findings.[1, 9] Furthermore, we aimed to disclose the epilepsy profile of the patients who are seropositive for one of these neuronal autoantibodies to guide the clinicians who are managing patients with established epilepsy.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Disclosure
  7. References


We included all consecutive adult patients diagnosed with FEoUC or MTLE-HS, followed by our epilepsy center for more than 1 year. Our epilepsy center accepts all kinds of epilepsy patients and is not specific for drug-refractory patients. As a result we also follow many patients with a relatively benign course.

The study was approved by the ethics committee. Informed consent was obtained from all participants before blood sampling. Seizures and syndromes were diagnosed according to the revised terminology, and concepts for organization of seizures and epilepsies of the International League Against Epilepsy (ILAE) Commission on Classification and Terminology,[1] and their auras were classified according to Report of the ILAE Task Force on Classification and Terminology.[10] The patients with obvious provoking factors or an apparent remote origin, such as a brain malformation or tumor, trauma, central nervous system infection with magnetic resonance imaging (MRI) evidence, or generalized epilepsy were excluded.

Clinical parameters such as current age, age at onset, sex, all neurologic findings, and complaints like movement disorders, autonomic disturbances, sleep disturbances, medical and family history, epilepsy duration, seizure frequency, semiology, total seizure count in the previous year, types of aura, medication at the time of serum sampling, electroencephalography (EEG), video-EEG, MRI, and other laboratory findings were collected from the files of the patients and by reexamining and querying them with a standard form systematically, at the day of serum sampling by one of the authors.

All MRI studies were performed with 1.5 T scanners with thin coronal in addition to sagittal and axial planes including T1, T2, and fluid-attenuated inversion recovery (FLAIR) images to visualize mesial temporal regions optimally. All EEG and video-EEG monitoring records were reviewed retrospectively by the authors.

AED response was defined as poor in case of experiencing more than one seizure per month (with the exception of isolated auras) despite reasonable trials of two or more AEDs. The patients with MTLE, fulfilling these criteria once, who had undergone epilepsy surgery afterwards, were also grouped in the “poor prognosis” category, despite being better after surgery. Some of our patients have also been evaluated by a psychiatrist, and formal neuropsychological tests were done when clinically needed during the course.

Autoantibody testing

Sera from all patients and from 30 healthy controls were kept at −80°C until assayed. NMDA-R; α-amino-3-hydroxy-5-methyl-4-isoxazoleproprionic acid receptor (AMPA-R); leucine rich, glioma-inactivated 1 (LGI1); contactin-associated protein-like 2 (CASPR-2); and GLY-R antibodies were detected by binding to HEK293 cells transfected with plasmids containing the NR1/NR2 subunits of the NMDA-R, GluR1/GluR2 subunits of the AMPA-R, LGI1, CASPR-2, or GLY-R α1 subunit, respectively. Transfected cells were incubated with patients' sera (1:20) and the appropriate Alexa Fluor secondary antibody, as described earlier.[3, 6, 11] The binding was scored visually on a range from 0 (negative) to 4 (very strong) as in previous studies.[11] Only scores greater than one were accepted as positive to avoid nonspecific low positivity. For VGKC-complex antibodies, radioimmunoassay (RIA) using brain extracts labeled with 125I-dendrotoxin (normal value <100 pm) were used.[11, 12] GAD antibodies were measured by immunoprecipitation of 125I-recombinant GAD.[13]

Immunotherapy was offered for seropositive patients with poor response to AEDs after their informed consent. Serum samples were obtained in a single patient with GLY-R antibody having improvement in seizure frequency after immunotherapy, at different time points and tested for this antibody at serial dilutions ranging from 1/10 to 1/320. In addition, the EEG spikes were counted before and after immune treatment in this patient.

Statistical analysis

Descriptive statistics were applied, and the two groups of patients with and without serum antibodies were compared with chi-square tests, Fisher's exact test, and independent samples t-test, where appropriate. No adjustments for multiple comparisons were done in this study. SPSS 15 software (SPSS Inc, Chicago, IL, U.S.A.) was used and the significance level was set at p < 0.05.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Disclosure
  7. References

A total of 81 consecutive patients with epilepsy, 55 with FEoUC (mean age: 39.0 ± 13.8 years), 26 with MTLE-HS (mean age ± standard deviation: 32.6 ± 9.4 years) and 30 healthy volunteers (mean age: 36.4 ± 8.1 years) were included in this study. The investigated autoantibodies were present in 13 patients (7 female, 6 male; mean age 35.5 ± 8.0 years) with a seropositivity level of 16%. Antibodies were detected against GLY-R in 5 (6.2%), CASPR-2 in 4 (4.9%), NMDA-R in 2 (2.5%), and VGKC-complex in 2 (2.5%) of our patients with epilepsy (Tables 1 and 2). None of the CASPR-2 antibody–positive patients had VGKC-complex antibodies and vice versa. The seropositivity level did not decrease (15.7%) after excluding 11 tested patients with previously known autoimmune conditions such as systemic lupus erythematosus (SLE) and thyroiditis (shown in Table 3), from both of the seropositive and seronegative groups. None of the healthy subjects had any of the investigated auto-antibodies.

Table 1. The clinical and laboratory features of the antibody-positive patients
NoAntibodyaAge sexEpilepsy syndromeHistoryFamily historyPsychiatric disorderMemory or cognitive comorbiditiesMRI findingsPET findings
  1. N, normal; M, male; F, female; FEoUC, focal epilepsy of unknown cause; MTLE-HS, mesial temporal lobe epilepsy with hippocampal sclerosis; VGKC, voltage-gated potassium channel; CASPR-2, contactin-associated protein-like 2; NMDAR, N-methyl-d-aspartate receptor; GLY-R, glycine receptor; FebS, febrile seizure; MRI, magnetic resonance imaging; PET, positron emission tomography; Fe, ferritin; CBZ, carbamazepine; F, frontal; T, temporal; MGUS, monoclonal gammopathy of unknown significance; BCS, bilateral convulsive seizure.

  2. a

    Numbers indicate the antibody binding intensity scored visually on a range from 0 (negative) to 4 (very strong).

  3. b

    Lumbar puncture results normal (infectious etiologies and cancer were excluded).

  4. c

    Patient 5 had head trauma 5 years before the onset of seizures, patient 7 had trauma 8 years before the onset of seizures, patient 13 had trauma 4 years after the onset of seizures.

  5. d

    This patient has the typical MRI findings compatible with mesial temporal lobe sclerosis but her clinical picture is rather benign, as seen in the second part.

  6. e

    Healthy control <100 pm.

1NMDA-R, 243 MFEoUCExplosive onset with unknown etiologybFebS in the sisterPostictal paranoid psychotic attacks 9 years after the onsetMemory disorderNonspecific white matter changes
2NMDA-R, 239 MFEoUCUnremarkableUnremarkableDepressionAttention disorder and secondary memory disorderRight F subcortical small nonspecific signal change
3GLY-R, 437 MFEoUCUnremarkableUnremarkableDepressionNot presentNormal
4GLY-R, 419 MFEoUCUnremarkableConsanguinity, FebS in the brotherPsychotic spells 10 years after the onset, borderline personality disorderDiffuse cognitive dysfunctionNormalLeft superior P hypometabolism
5GLY-R, 330 FFEoUCPenetrating head trauma,c migraineConsanguinity, epilepsy in three cousinsDepressionMild attention disorder and secondary memory disorderNormal
6GLY-R, 333 FMTLE-HS operatedAsthmaUnremarkableDepressionVerbal memory defect, frontal dysfunctionCompatible with left HSLeft T hypometabolism
7GLY-R, 233 FFEoUCBlunt head trauma,c migraine, hyperthyroidismConsanguinity, epilepsy and mental retardation in a cousinNoneNot presentNormal
8CASPR-2, 338 FMTLE-HSSystemic lupus erythematosusConsanguinity of the parentsDepressionNot presentCompatible with left HS and small white matter changes
9CASPR-2, 235 FMTLE-HS operatedUnremarkableUnremarkableDepressionFrontal lobe and left hemispheric dysfunctionCompatible with left HSBilateral mesial Thypometabolism
10CASPR-2, 340 FMTLE-HSdMigraineUnremarkableUnremarkableNot presentCompatible with right HS Nonspecific small white matter changes
11CASPR-2, 236 MMTLE-HS operatedUnremarkableUnremarkableNoneVerbal memory defectCompatible with left HS
12VGKC (201.4 pM)e53 FFEoUCHashimoto thyroiditis, MGUS, sleep problemsEpilepsy in her sister (BCS)NoneAttention difficulty, mild frontal dysfunction, hypergraphismNonspecific white matter changesBiT left > right hypometabolism
13VGKC (138.6 pm)e26 MMTLE-HS (bilateral)Blunt head traumacConsanguinity, epilepsy in the cousinPsychosis 8 years after the onset, suicidal ideationAttention disorder and secondary memory disorderCompatible with bilateral HSBiT right > left hypometabolism
Table 2. The seizure and EEG characteristics of the antibody positive patients
NoSeizure typesAge at onsetTypes of auraBackground EEG activityInterictal epileptic activity on routine EEGIctal EEG (video-EEG monitoring)Other EEG findingsAED at the samplingAED responseaOther findings
  1. FSwIoC, focal seizure with impairment of consciousness; F-BCS, focal seizure evolving to bilateral convulsive seizure; FebS, febrile seizure; L, left; R, right; C, central; F, frontal; P, parietal; T, temporal; FIRDA, frontal interictal rhythmic delta activity; TIRDA, temporal interictal rhythmic delta activity; AED, antiepileptic drug; CBZ, carbamazepine; LEV, levetiracetam; PRG, pregabalin; TOP, topiramate; VPA, valproic acid; ZON, zonisamide; VNS, vagus nerve stimulation.

  2. a

    Good prognosis implied that the patient has less than one convulsive seizure of any kind monthly (auras excluded): Those patients who have undergone selective amygdalohippocampectomy were also classified in the poor prognosis group even if they reach remission after epilepsy surgery.

1FSwIoC, F-BCS24Olfactory, autonomic, experientialDiffuse theta and delta waves to normalLeft and right T rare sharp wavesNot availableLEV 3500, TOP 300Poor 
2FebS, BCS32Cephalic, autonomicLeft FT theta, otherwise normalLeft FT spikeNot availableVPA 1000Good 
3F-BCS17NoneLeft FT theta, otherwise normalLeft FT rare sharp wavesNot availableVPA 600, LEV 1500Good 
4FSwIoC, F-BCS3NoneDiffuse thetaLeft CP >right CP spikesSeven seizures starting in left F area (also subclinical seizures)CBZ 1200, LEV 2000, TOP 200, VNSPoor 
5FSwIoC, F-BCS14Cephalic, autonomicParoxysmal slow waves, otherwise normalRight FT spikesRight FT subclinical seizuresRight FIRDACBZ 800GoodProminent hyponatremia with CBZ
6FSwIoC, F-BCS16Epigastric, autonomicLeft FT theta, otherwise normalLeft FT spikesThree seizures starting in left FT areaLeft TIRDAOXC 600, TOP 100Poor 
7FSwIoC, F-BCS16Auditory, affectiveNormalRight TP sharp waves and theta wavesNot availableVPA 750GoodFe deficiency anemia
8FebS, FSwIoC, F-BCS17Affective, autonomicFT theta-delta paroxysms otherwise normalRight FT spikesNot availableRight FIRDACBZ 1200, PRG 300PoorUse of immune-suppressants
9FebS, FSwIoC21Experiential, autonomicFT theta wavesBi FT left > right spikesInvasive recording 5 from left, two seizures from right FT areaCBZ 1200, LEV 2000Poor 
10FebS, FSwIoC37NoneFT Theta wavesNoneNot availableCBZ 400Good 
11FSwIoC, F-BCS20Cephalic, affectiveLeft FT theta-delta waves, otherwise normalLeft FT sharp waves with phase reversal, Right T rare sharp wavesTwo seizures left FT (with switch off to right side)Left TIRDACBZ 1200, LEV 1000, TOP 200Poor 
12FSwIoC46NoneBiFT L > R theta wavesLeft > right FT spikesFive seizures starting in left FT areaLEV 2000, ZON 200Poor 
13FebS, FSwIoC, F-BCS8Autonomic, affectiveRight FT theta-delta waves, otherwise normalRight T sharp wavesNo seizures during 5 days of VEMRight TIRDALEV 3000GoodFever triggered seizures
Table 3. Comparison of the clinical features of patients with and without autoantibodies
 Seropositive patients (n = 13)Seronegative patients (n = 68)p-Values
  1. AED, antiepileptic drugs; MTLE-HS, mesial temporal lobe epilepsy with hippocampal sclerosis; MRI, magnetic resonance imaging; n, number; NS, not significant; Q1, first quartile; Q3, third quartile.

  2. Please note that interquartile range is IQR = Q3−Q1.

  3. a

    Fisher's exact test two-sided.

Sex (female/male)6/728/40NS

Age at serum sampling (years)

Median; [Q3−Q1]

36; [39.5–31.5]36; [43.7–27.2]NS

Age at onset of epilepsy (years)

Median; [Q3−Q1]

17; [28–15]23; [31.7–15.2]NS

Epilepsy duration (years)

Median; [Q3−Q1]

16; [18.5–10.5]12; [18–5.2]NS

Total seizure count in the previous year

Median; [Q3−Q1]

18; [36–2.5]2; [24–0]NS
Poor AED response (n) (%)7 (53.8)31 (45.6)NS
History of status epilepticus (n) (%)0 (0)3 (4.4)NS
Autonomic aura (n) (%)7 (53.8)22 (32.3)NS
History of febrile seizures (n) (%)5 (38.5)13 (19.1)NS
Coexisting autoimmune diseases (n) (%)2 (15.4)9 (13.2)NS
Hashimoto's thyroiditis14 
Systemic lupus erythematosus (SLE)11 
SLE and antiphospholipid syndrome2 
Antiphospholipid syndrome1 
Behçet's disease1 
History of psychotic attacks (n) (%)3 (23.1)0 (0)0.003a
Nonspecific MRI white matter changes (n) (%)5 (38.5)8 (11.8)0.03a
Epilepsy surgery (n) (%)3 (23.1)8 (11.8)NS

In FEoUC group, we found GLY-R antibodies in 4, NMDA-R antibodies at low levels in 2 patients, and VGKC-complex antibodies at low levels in one patient, whereas in MTLE-HS group, 4 patients were positive for antibodies to CASPR-2, one had GLY-R antibodies, and one had VGKC-complex antibodies at low levels. Clinical and laboratory features of antibody-positive patients are given in Tables 1 and 2. Neurologic examinations were normal in the seropositive group, with the exception of two patients, who had mild essential tremor.

The clinical and laboratory features and some comorbidities of the patient groups with and without autoantibodies are shown in Table 3, comparatively. The prevalence of psychotic attacks and nonspecific MRI white matter changes (WMCs) were found to be significantly higher in seropositive patients (p = 0.003 and p = 0.03, respectively). History of febrile seizures was relatively more frequent in patients having autoantibodies, but this difference did not reach statistical significance (Table 3). Three seropositive patients had remote history of head trauma without MRI evidence (Table 1) versus seven patients in the seronegative group (p > 0.05).

Poor AED response and total seizure counts were higher in the seropositive group but without statistical significance. Epilepsy surgery was accomplished before serum sampling in three seropositive patients (3, 41, and 76 months ago) and also in 7 seronegative patients (median: 67, range: 32–124 months ago) with beneficial results. The remaining seronegative patient has undergone epilepsy surgery 7 months after antibody testing.

AED-resistant seropositive patients

Among the total seropositive group, seven patients (three with FEoUC and four with MTLE-HS) had poor response to AEDs (Tables 1 and 2).

We initiated immunotherapy for two of these FEoUC patients with poor response to AEDs. The remaining seropositive patient 1 (NMDA-R antibody positive) did not accept pulse steroid or intravenous immunoglobulin (IVIg) therapy. Seropositive patient 3 (GLY-R antibody positive) had early onset of seizures with a worsening course in years (5–15 seizures per day for more than 10 years), did not benefit from many AEDs or from vagus nerve stimulation. He was experiencing daily seizures at the time of first serum sampling. Intravenous methylprednisolone (IVMP) (1 g daily for 5 days followed by oral prednisolone 32 mg/day tapered off in 4 weeks) also remained ineffective, but afterward his seizure frequency improved with IVIg (0.4 g/kg/day given for consecutive 5 days); he had 0–2 seizures per month during the subsequent follow-up period of 3 months. In addition, his EEG also showed some improvement, with 49% decrease of spike counts after IVIg therapy. We analyzed several serum samples from this patient at different time points. Although GLY-R antibody titers were slightly decreased after steroid and IVIg therapies, the titers fluctuated between 1/120 and 1/180 (scores between 2 and 3) throughout the treatment period. In other words, we did not find a strong correlation between antibody titers and number of seizures during each treatment episode. Seropositive patient 12 (VGKC-complex antibody positive) did not benefit markedly from pulse steroid and IVIg therapies, except for being seizure free for 3 weeks.

One female patient with both MTLE and SLE (patient 8, CASPR-2 antibody positive) also had various immunotherapies. Her seizures did not benefit from steroid therapy and cyclophosphamide given during her previous follow-ups before antibody testing, and she remained seizure free for only 6 weeks under rituximab.

Although having benefited from epilepsy surgery in the first years, seropositive patients 6 and 11 both experienced late recurrences of rare seizures after 3 and 4 years. Patient 9 was still seizure free following epilepsy surgery for a follow-up period of 1 year. The neuropathologic examinations of this three seropositive MTLE-HS patients (Tables 1 and 2, patients 6, 9, 11) who had undergone amygdalohippocampectomy, did not show any inflammatory changes and were not different from the seronegative patients with epilepsy surgery. In one of them (GLY-R antibody positive) CA1 predominant neuronal cell loss and gliosis were reported, whereas the other two patients (both CASPR-2 antibody positive) had more diffuse neuronal cell loss, with preservation of CA2 region.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Disclosure
  7. References

There is growing interest for possible autoimmune mechanisms in clinical epilepsy research, and even a concept of “autoimmune epilepsy “was recently proposed by different authors.[8, 14] We found that a remarkable number of patients with focal epilepsy with either unknown cause or MTLE-HS harbor various neuronal antibodies, GLY-R antibodies being the first (6.2%), followed by CASPR-2 (4.9%), NMDA-R (2.5%), and VGKC-complex (2.5%) antibodies. Our study was based on a consecutive series with established epilepsy from an epilepsy center that accepted patients with all kinds epilepsy, not specifically drug-refractory patients only. There are previous studies showing the presence of neuronal antibodies in some patients having epilepsy but with different inclusion criteria.[3-5, 8] In our series with focal epilepsy, we found autoantibodies in one sixth of the patients with epilepsy, with a higher rate than all previous reports, emphasizing the importance of a syndromic approach.

Focal epilepsy of unknown cause and autoantibodies

It was remarkable that GLY-R antibodies were detected in four patients in the FEoUC group. Antibodies directed to GLY-R were first reported in patients with progressive encephalomyelitis with rigidity and myoclonus.[15, 16] The intriguing association of GLY-R antibodies with epilepsy was reported for the first time in a recent study where GLY-R antibodies were detected with 3% prevalence.[3] Our study supported the importance of this antibody with a prevalence of 6.2%. Our cohort included syndromes of FEoUC and MTLE-HS, whereas the first study included all kinds of epilepsies, including genetic generalized forms. Because GLY-R is a very recently discovered antibody, its exact incidence waits to be clarified. Future studies are needed to understand the contribution of GLY-R autoantibodies to the pathophysiology of epilepsy.

VGKC-complex antibodies were reported previously in patients with FEoUC,[3, 4] and in association with drug-resistant focal epilepsy.[5] Similarly, our VGKC-complex antibody-positive patient had a poor response to AEDs.

In addition, we have detected NMDA-R antibodies in two patients having FEoUC but with different clinical pictures; one had explosive onset of seizures similar to limbic encephalitis (LE) with memory problems and poor response to AEDs followed by very late onset episodic postictal paranoid psychotic symptoms, whereas the second one had a mild course and did not show these features. Furthermore, these two patients did not show the reported extreme delta brush activity on careful analysis of all EEG studies.[17] We suggest that clinicians should have a broader index of suspicion for NMDA-R antibody evaluation in patients with FEoUC, as this antibody might cause a wide clinical spectrum of disease.

MTLE-HS and autoantibodies

Antibodies directed to the VGKC complex, GAD, and NMDA-R have already been related to LE associated with seizures.[18-20] Furthermore the evolution from LE to TLE and to MTLE-HS, which cannot be distinguished from the MRI correlate of HS has been described.[21-23] Our consecutive group of 26 MTLE-HS patients was larger than the previous studies where the investigators also found neuronal antibodies.[4, 5, 8] Among our cases, four patients were positive for antibodies to CASPR-2 (but interestingly were negative for VGKC-complex antibodies assessed by RIA) and another one had VGKC-complex antibodies at low levels but no LGI1 or CASPR-2 antibodies. Hence 19.2% of our MTLE cases were shown to be associated with antibodies to the VGKC-complex, suggesting a potential role of VGKC-complex autoimmunity in the pathogenesis of a subgroup of MTLE-HS patients. Furthermore another intriguing MTLE case had GLY-R antibody. To our knowledge both CASPR-2 and GLY-R antibodies are reported for the first time in association with MTLE-HS syndrome. The implications of these findings remain to be elucidated in a larger group of patients with MTLE.

Other investigated neuronal autoantibodies undetected in our series

There have been some reports of GAD autoantibodies in patients with epilepsy.[3, 24] However, we could not detect any GAD antibody–positive case in this cohort. The frequency of GAD antibody in patients with epilepsy ranges from 0% to 7%.[4, 5, 7, 25] Therefore, the clinical role of GAD autoantibodies in epilepsy does not seem to be very clear currently.[4, 26] At the time of our study, we diagnosed a single case having GAD antibodies with nonconvulsive status epilepticus; she was not included because she was not admitted to the epilepsy center but was followed in the neurology ward and the details of this unique case are published elsewhere.[27]

Moreover, we could not detect any case with AMPA-R antibodies in our series, suggesting that this antibody does not play a pioneering role in epileptic seizures, confirming previous reports.[3, 4, 28] We also did not detect any antibodies to LGI antigen.[14]

Coincidence of autoantibodies with other autoimmune disorders

Some previous studies investigated autoantibodies in epileptic patients with coexisting autoimmune diseases.[4, 5] There is a well-known increased risk for patients with SLE to develop unprovoked seizures.[29] The mechanism of this association is unsolved in patients without any obvious cause for seizures. Some antibodies such as antiphospholipid antibodies have been reported to be associated with seizures in SLE, but there are controversial results.[5, 30, 31] Hashimoto's thyroiditis (HT) is another frequent autoimmune disease, in which the association with epilepsy is not very clear and the role of these antibodies in Hashimoto's encephalopathy showing higher association with seizures is also uncertain.[31]History of other antibody-mediated disorders or organ-specific autoimmunity has been proposed as a supportive feature for neuronal antibody screening by Zuliani et al.[32] We did not find an association between the presence of previously known autoimmune conditions and the seropositivity in our sample of consecutive patients, but two seropositive patients having coexisting autoimmune diseases had antibodies against potassium channels. Therefore, our findings may suggest that potassium channels could be another possible antigenic target in these patients with epilepsy.

The clinical and laboratory associations of seropositive patients

We compared the patients harboring neuronal antibodies with the seronegative patients and confirmed that the epilepsy duration did not seem to be important for the presence of autoantibodies as reported previously.[3, 4] The lack of association with epilepsy duration also partly contradicts the probability of the appearance of these antibodies as a result of neuronal destruction caused by chronic epilepsy.

Our findings showed that WMCs on the MRI and psychotic spells showed significant associations with seropositivity, as seen in Table 3. It should be noted, however, that our group could be considered small and no statistical adjustments for multiple comparisons were done. Therefore, further studies with large series are needed to draw firm conclusions. Antibodies against NMDA-R were reported previously in new-onset extratemporal epilepsy with psychiatric symptoms.[33] Two women in this report had WMCs on MRI, but this point was not further discussed. It is interesting to note that both of our patients with NMDA-R antibodies also had nonspecific WMCs. This radiologic finding was also seen in only one patient having established epilepsy with GLY-R antibodies.[3] Furthermore, subcortical bifrontal and left parietal hyperintensities were reported in one 9-year-old encephalitis case with VGKC antibodies.[23] The present study, however, extends previous anecdotal observations and suggests that WMCs seem to be a possible marker for the presence of autoantibodies in epilepsy patients and should also be included in the indication list for search of neuronal antibodies.[32] This association may partly be related to the coexistence of patients with SLE, migraine, and HT, since all of these disorders are known to cause WMCs.[31, 34] It should be noted, however, that there are also patients with SLE, migraine, and HT without any autoantibodies in our series. These antibodies might cause direct immune-mediated changes in the brain ending up with WMCs, but it seems also highly likely, that they could be innocent bystanders.

The association between anti-NMDA-R encephalitis and acute psychosis has been well documented.[35-37] The presence of VGKC antibodies in psychotic patients was also reported in a few patients.[38, 39] Therefore, the association of psychosis with neuronal antibodies in our epilepsy series is not a great surprise but indicates that patients with epilepsy and history of psychosis should be screened for these antibodies for a treatment possibility. On the other hand, one of our three cases with psychotic spells had GLY-R antibodies, adding this antibody to the list of antibodies associated with psychosis for the first time.

Responses to treatment

Poor AED response seemed higher in seropositive group without statistical significance. Other studies reported more patients with AED resistance, but they included epileptic patients with suspected autoimmune origin, which could have created a selection bias.[8] There were antibody-positive patients with a benign clinical course in our series, thus immunotherapeutic interventions should be based on the individual's AED response status.

Two AED-resistant patients with antibodies against VGKC showed no clear benefit from immunotherapy trials despite having short-term seizure reductions for 3–6 weeks. We also assessed prospectively the response to immunotherapy in a patient having FEoUC with seropositivity for GLY-R antibody for the first time. IV pulse steroid treatment did not show any benefit, whereas IVIg treatment showed reduction of seizure frequency unrelated to antibody titers and moderate EEG response, but it is not possible to exclude the consequences of natural fluctuations and placebo effect. There is also a report of a GLY-R antibody positive case with an immunotherapy responsive isolated mesial temporal lobe status epilepticus.[33]

We are only aware of two neuronal antibody-positive patients, who have undergone epilepsy surgery without any benefit, in the relevant literature.[8] Our three patients with MTLE-HS had undergone selective amydalohippocampectomy before antibody testing with favorable results and showed pathologic evidence of HS indistinguishable from the seronegative patients. Furthermore our results showed late relapses in two patients after successful epilepsy surgery. Moreover having previous epilepsy surgery also did not show any relation with antibody positivity excluding the appearance of antibodies in response to tissue destruction caused by surgery.

In conclusion, there is a substantial number (one sixth of all) of patients with focal epilepsy with either unknown cause or MTLE-HS who harbor various neuronal antibodies, the recently reported GLY-R antibodies being the leading cause, followed by VGKC-complex antibodies in our series. We recommend in the light of our data that epileptic individuals with psychosis or nonspecific MRI WMCs be assessed with the possibility of antibody-mediated disorder in mind. The patients showing these serum antibodies could be treated with immunotherapy, when there is a clinical indication. There are still many unanswered questions such as the selection of the candidates for antibody testing, natural history of autoimmune epilepsy possible immunotherapeutic intervention types, and duration and the role of epilepsy surgery. In addition, these autoantibodies could be the markers of currently unknown immunopathologic processes rather than having a major role in pathogenesis. Future studies with larger series of epilepsy syndromes are needed to draw further conclusions.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Disclosure
  7. References

None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Disclosure
  7. References