Epilepsy and the Immune System

NEUROLOGICAL REVIEW

Johan A. Aarli, MD

Objective: To discuss evidence that immune mecha- data to prove that immune mechanisms are involved in nisms are involved in the pathogenesis of some forms of the pathogenesis of intractable childhood epilepsies other epilepsy. than Rasmussen encephalitis. Epilepsy is more common in patients with systemic lupus erythematosus who Data Sources: Computerized data sources and pub- have antiphospholipid antibodies, and it is possible that lished indexes and articles. these antibodies can lead to immune-mediated cortical damage. Immune defects in patients with epilepsy may Study Selection: Published reports showing disor- occur as a consequence of long-term antiepileptic treat- ders of the immune system in patients with epilepsy and ment or may represent a genetic coupling to the convul- in animals with experimental epilepsy. sive disorder.

Data Synthesis: Rasmussen encephalitis is an ex- Conclusion: The finding of an immunological basis may ample of an autoimmune disorder of the central ner- offer new modalities for the treatment of selected cases vous system. Serum samples of patients with this dis- of intractable partial epilepsies. ease contain antibodies to the glutamate receptor GluR3, and immunization of animals with GluR3 induces a disorder resembling the human disease. There are still few Arch Neurol. 2000;57:1689-1692

HE FINDING of immune sys- Rogers and coworkers3 demonstrated in tem activation in patients 1994 that immunization of rabbits with the with a seizure disorder has glutamate receptor GluR3 produces a con- lead to the suggestion that dition resembling Rasmussen encephali- immune mechanisms may tis. They also showed that serum samples play a role in the pathogenesis of some of patients with this disease indeed con- T 1 forms of epilepsy. For a long time, such tain antibodies to GluR3 and postulated theories lacked an experimental basis. The that Rasmussen encephalitis is an auto- results of recent studies, especially in pa- immune disorder.3 tients with Rasmussen encephalitis, have To establish that a disease has an au- given new information about the pos- toimmune etiology, Witebsky postulates sible relation between epileptic disorders require that an autoimmune response be and the immune system. recognized in the form of an autoanti- body or cell-mediated immunity, that the RASMUSSEN ENCEPHALITIS—AN corresponding antigen be identified, and AUTOIMMUNE DISORDER? that an analogous autoimmune response be induced in experimental animals. Fi- The clinical setting in Rasmussen encepha- nally, the immunized animals must de- litis is an apparently healthy child or young velop a similar disease.4 adult developing focal seizures that ini- The most straightforward evidence tially appear to be benign, but that gradu- for an autoimmune etiology of a disease ally develop into epilepsia partialis con- is to reproduce it in a normal recipient tinua with hemiparesis and mental by direct transfer of autoantibodies or From the Department retardation. A neuropathological exami- sensitized cells. This is seen as nature’s of Neurology, University nation reveals the typical picture of en- own experiment in neonatal myasthenia of Bergen, Haukeland Hospital, cephalitis with perivascular lymphocyte gravis, where placental transfer of mater- Bergen, Norway. cuffs and scattered microglial nodules.2 nal IgG anti–acetylcholine receptor anti-

(REPRINTED) ARCH NEUROL / VOL 57, DEC 2000 WWW.ARCHNEUROL.COM 1689

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 bodies produces muscular weakness in the newborn. effect. Surprisingly, intraventricular infusion of inter- However, the blood-brain barrier represents an impedi- feron alfa reduced the seizure frequency dramatically in ment for antibodies to reach target antigens in the cen- a child with intractable Rasmussen encephalitis.11 This, tral nervous system. Opportunities for direct transfer of however, may relate to the demonstration that GluR3 an- autoimmune central nervous system disease are, there- tibodies exhibit greater immunoreactivity toward the fore, limited. interferon alfa receptor than toward GluR3.12 The classic strategy for documenting an autoim- These new data on Rasmussen encephalitis confirm mune etiology has been to identify and isolate the of- that immune factors may be central in the pathogenesis fending antigen and reproduce the essential features of of an epileptic disorder and may lead to an increased un- the disease by experimental immunization. This has been derstanding of the basis of other seizure disorders with im- accomplished with Rasmussen encephalitis. In raising an- munological concomitants. tibodies to glutamate receptors, Rogers et al3 demonstrated that 2 animals immunized with GluR3 devel- EPILEPSY IN PATIENTS WITH SYSTEMIC oped behavior typical of seizures and histopathological LUPUS ERYTHEMATOSUS (SLE) features mimicking Rasmussen encephalitis. This obser- vation has since been confirmed in other species.5,6 Between 10% and 20% of patients with SLE develop epi- Additional support that the autoantibodies are patho- leptic seizures at some stage of their disease. This is genic comes from the demonstration that IgG from rab- nearly 8 times the prevalence of epilepsy in the general bit antiserum to GluR3 binds to and activates a subpopu- population; epilepsy is, therefore, much more common lation of fetal murine cortical neurons in culture and that in patients with SLE than would be expected. Between this binding could be blocked by a competitive antago- 5% and 10% have onset of seizures several years before nist, CNQX.7 the clinical onset of SLE.13-15 This may mean that long- The experimental disease in these animals mirrors term treatment with antiepileptic drugs may precipitate several features of the human disease, such as the epi- SLE, or that epilepsy and SLE occur together as manifes- leptic seizures, the histopathological findings, and the se- tations of a genetically determined predisposition. lectivity of the immune response to GluR3. It differs, how- According to Mackworth-Young and Hughes,14 epilepsy ever, from Rasmussen encephalitis in humans in that the developing in patients before the other manifestations of latter is unilateral while the experimental disease in- SLE differs from that developing after the other manifes- volves both hemispheres. tations of SLE. Epileptic seizures occurring before the The neuronal cell death observed in vitro appears onset of other manifestations of SLE were more often not to be caused by excitotoxicity. IgG antibodies to GluR3 primary generalized, but seizures that occurred after the isolated from immunized animals (both ill and healthy) clinical onset of SLE were either focal or generalized- promote death of cultured cortical cells by a complement- tonic.14 dependent mechanism. IgG and complement factors of Seizures may occur in patients with SLE based on the membrane attack complex were found on neurons immune-mediated neuronal damage, because of throm- and their processes in the cortex of brains from patients botic events in cortical blood vessels, or they may be with Rasmussen encephalitis. Antibodies seem, there- secondary to hypertensive encephalopathy or renal fail- fore, to gain access to the central nervous system and trig- ure. Antibodies to transmitter receptors such as GluR3 ger complement-mediated neuronal damage. This pro- have never been convincingly demonstrated in patients cess may be of special importance in the initial, active with SLE. Epilepsy in patients with SLE is significantly phase of the disease.8,9 Genetic and species differences associated with antiphospholipid antibodies (aPLs).16 may exist, as Levite et al5 recently reported that murine Chapman et al17 reported recently that purified IgG GluR3 antibodies can mimic excess glutamate effect and containing aPLs depolarized synaptoneurosomes from induce neuronal death via activation of the receptor ion rat brainstem and suggested this as an additional channel, apparently independent of complement. mechanism in nonthromboembolic central nervous There is additional circumstantial evidence that au- system manifestations. toimmune mechanisms operate in Rasmussen encepha- Patients with aPLs are at risk of thromboembolic litis. Li et al10 have demonstrated restricted T-lymph- manifestations, intrauterine fetal loss, and thrombocy- ocyte populations in the brains of patients with Rasmussen topenia (antiphospholipid syndrome). Such antibodies encephalitis. Removal of antibodies by plasma ex- are commonly found in serum samples of patients with change transiently reduces the seizure frequency and im- SLE and are generally heterogeneous, and the different proves the neurologic function as the serum concentra- specificities may play causal roles in different clinical mani- 3 18,19 ␤ tions of GluR3 antibodies decrease. As antibodies to festations. For example, anti– 2-glycoprotein I anti- GluR3 are found in serum samples from immunized ani- bodies may have a direct pathogenic role in thrombosis. ␤ mals without apparent disease, a focal or a general dis- They recognize a 2-glycoprotein I structure on cellular ruption of the blood-brain barrier is essential for serum structures and are also significantly elevated in patients antibodies to reach the brain.1 with epilepsy.20,21 IgG antibodies to cardiolipin have been The finding of an immunological background for an found in 30% to 60% of unselected patients with SLE.13,22 intractable epilepsy offers new modalities for treatment. Liou et al23 found that epilepsy was 3.7 times as fre- For example, Rogers et al3 demonstrated the effect, al- quent among patients with SLE who had antibodies to beit transient, of plasma exchange.This has since been cardiolipin than among patients with SLE who did not confirmed by other groups. Corticosteroids also have some have antibodies to cardiolipin.

(REPRINTED) ARCH NEUROL / VOL 57, DEC 2000 WWW.ARCHNEUROL.COM 1690

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 Using magnetic resonance imaging and spectros- controlled. That study31 gave a positive trend in favor of copy, Sabet-Arman and coworkers24 found that epilepsy intravenous immunoglobulin treatment, but the differ- (and stroke) was more common in patients with SLE and ence was not statistically significant. aPLs and suggested that these antibodies exacerbate SLE, A direct effect of immunoglobulin on the brain is resulting in increased thrombotic and nonthrombotic possible only if the blood-brain barrier is abnormal. It brain injuries. Angelini and coworkers25 studied 23 chil- has been demonstrated in experimental animals that the dren with partial epileptic seizures and no clinical or se- blood-brain barrier opens transitorily during seizures. rological evidence of SLE. None of them had magnetic Since the effect of immunoglobulin treatment is seen only Ј resonance imaging evidence of focal ischemic lesions, but with intact IgG and not with F(ab )2 fragments, the ef- 3 of them, all with frontal lobe epilepsies, had aPLs. These fect requires an intact Fc part of the molecule.32 researchers25 speculated that the antibodies could lead to immune-mediated damage, which could be a patho- IMMUNOLOGICAL CONSEQUENCES genic mechanism for partial epilepsy. OF LONG-TERM TREATMENT WITH ANTIEPILEPTIC DRUGS OTHER CHILDHOOD EPILEPSIES Phenytoin treatment may induce a reversible IgA defi- There are few data that indicate that immune mecha- ciency involving serum and secretory IgA.1 This has been nisms are involved in the pathogenesis of intractable demonstrated by monitoring IgA levels before and dur- childhood epilepsies other than Rasmussen encephali- ing initiation of phenytoin treatment and measuring IgA tis. Most interest has been focused on infantile spasms levels after withdrawal of the drug. (West syndrome) and the Lennox-Gastaut syndrome. Carbamazepine may also induce a reduction of se- These are heterogeneous conditions with uncertain rum IgA levels, but never as low as with phenytoin. An- pathophysiological characteristics and can be precipi- other antiepileptic drug, zonisamide, has recently been tated by various causes. Corticotropin and corticoste- reported to induce IgA and IgG2 deficiency.33 roids are widely used in their treatment, and some Some cases of epilepsy are, however, associated with patients, especially with West syndrome, respond to the primary IgA deficiency. This is the most common im- treatment far better than they do to conventional anti- mune defect in humans. Most subjects with primary IgA epileptic drugs. This does not imply that immune deficiency are healthy, but patients with selective IgA de- mechanisms are involved, and it has never been demon- ficiency often have an increased susceptibility to upper strated how corticotropin (or corticosteroids) exerts its respiratory tract infections.34 Seizures may also occur in therapeutic effect in this group of disorders. patients with ataxia telangiectasia, which is associated with Neuropathological examinations of brains from chil- IgA (and occasionally IgG) deficiency. While primary IgA dren who have died of Lennox-Gastaut syndrome have deficiency is associated with the haplotype HLA–A1-B8 revealed cerebellar atrophy with loss of Purkinje cells, and DR3, patients with drug-induced defect are haplo- but antibodies to Purkinje cells have not been demon- type HLA-A2. The mechanisms responsible for depres- strated. Van Engelen et al26 found an increased serum con- sion of humoral and cellular immunity by phenytoin are centration of IgG, but the humoral immune response to unknown, but studies in mice indicate a CD8+ cell– hemocyanin was impaired. Otherwise, no specific im- mediated inhibitory effect.35 munological abnormalities were detected. Using peroxidase immunostaining with slices of Accepted for publication March 1, 2000. human temporal cortex, Connolly and coworkers27 re- Reprints: Johan A. Aarli, MD, Department of Neu- cently demonstrated that children with the Landau- rology, University of Bergen, Haukeland Hospital, Jonas Kleffner variant syndrome (acquired epileptic aphasia) Lies v 65, 5021 Bergen, Norway (e-mail: johan.a.aarli have antibodies that react with brain endothelial cells @haukeland.no). and with cell nuclei. Their definition included children with autistic regression as long as they had epileptoid REFERENCES electroencephalographic results. There is one report of a consistent response to intravenous immunoglobulin in a child with Landau-Kleffner syndrome, and im- 1. Aarli JA. Immunological aspects of epilepsy. Brain Dev. 1993;15:41-50. provement in language function after immunomodul- 2. Andermann F, ed. Chronic Encephalitis and Epilepsy: Rasmussen’s Syndrome. Woburn, Mass: Butterworth-Heinemann; 1991:296. ating therapy has been reported by some, but not all, 3. Rogers SW, Andrews PI, Gahring LC, et al. Autoantibodies to glutamate recep- 28,29 researchers. tor GluR3 in Rasmussen’s encephalitis. Science. 1994;265:648-651. 4. Rose NR, Bona C. Defining criteria for autoimmune diseases (Witebsky’s pos- IMMUNOGLOBULIN TREATMENT tulates revisited). Immunol Today. 1993;14:426-430. 5. Levite M, Fleidervish IA, Schwarz A, Pelled D, Futerman AH. Autoantibodies to IN PATIENTS WITH EPILEPSY the glutamate receptor kill neurons via activation of the receptor ion channel. J Autoimmun. 1999;13:61-72. 30 In 1977, Pechadre et al reported that children with epi- 6. Levite M, Hermelin A. Autoimmunity to the glutamate receptor in mice: a model lepsy who were treated with intramuscular injections of for Rasmussen’s encephalitis? J Autoimmun. 1999;13:73-82. immunoglobulin for recurrent upper respiratory tract in- 7. Twyman RE, Gahring LC, Spiess J, Rogers SW. Glutamate receptor antibodies activate a subset of receptors and reveal an agonist binding site. Neuron. 1995; fections had a decrease in the frequency and severity of 14:755-762. their seizures. This has been supported by several re- 8. He XP, Patel M, Whitney KD, Janumpalli S, Tenner A, McNamara JO. Glutamate ports, but only one has been double-blind and placebo- receptor GluR3 antibodies and death of cortical cells. Neuron. 1998;20:153-163.

(REPRINTED) ARCH NEUROL / VOL 57, DEC 2000 WWW.ARCHNEUROL.COM 1691

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 9. Whitney KD, Andrews PI, McNamara JO. Immunoglobulin G and complement unselected patients with systemic lupus erythematosus: clinical and laboratory immunoreactivity in the cerebral cortex of patients with Rasmussen’s encepha- associations. Rheumatol Int. 1995;15:23-29. litis. Neurology. 1999;11:699-708. 23. Liou HH, Wang CR, Chen CJ, Chuang CY, Chiang IP, Tsai MC. Elevated levels of 10. Li Y, Uccelli A, Laxer KD, et al. Local-clonal expansion of infiltrating T lym- anticardiolipin antibodies and epilepsy in lupus patients. Lupus. 1996;5:307-312. phocytes in chronic encephalitis of Rasmussen. J Immunol. 1997;158:1428- 24. Sabet-Arman SWL Jr, Stidley CA, Danska J, Brooks WM. Neurometabolite mark- 1437. ers of cerebral injury in the antiphospholipid antibody syndrome of systemic lu- 11. Dabbagh O, Gascon G, Crowell J, Bamoggadam F. Intraventricular interferon-␣ pus erythematosus. Stroke. 1998;29:2254-2260. stops seizures in Rasmussen’s encephalitis: a case report. Epilepsia. 1997;38: 25. Angelini L, Granata T, Zbordi F, Binelli S, Zorzi G, Besana C. Partial seizures as- 1045-1049. sociated with antiphospholipid antibodies in childhood. Neuropediatrics. 1998; 12. Gahring LC, Carlson NG, Rogers SW. Antibodies prepared to neuronal gluta- 29:249-253. mate receptor subunit3 bind IFN ␣ receptors: implications for an autoimmune 26. Van Engelen BG, Weemaes CM, Renier WO, Bakkeren JA, Borm GF, Strengers process. Autoimmunity. 1998;28:243-248. PF. A dysbalanced immune system in cryptogenic Lennox-Gastaut syndrome. 13. Toubi E, Khamashta HA, Fanarra A, Hughes GRV. Association of antiphospho- Scand J Immunol. 1995;41:209-213. lipid antibodies with central nervous disease in systemic lupus erythemathosus. 27. Connolly AM, Chez MG, Pestronk A, Arnold ST, Mehta S, Deuel RK. Serum au- Am J Med. 1995;99:397-401. toantibodies to brain in Landau-Kleffner variant, autism, and other neurologic 14. Mackworth-Young CG, Hughes GRV. Epilepsy: an early symptom of systemic disorders. J Pediatr. 1999;134:607-613. lupus erythematosus [letter]. J Neurol Neurosurg Psychiatry. 1985;48:185. 28. Fayad MN, Choueiri R, Mikati M. Landau-Kleffner syndrome: consistent re- 15. Glanz BI, Schur PH, Khosbin S. EEG abnormalities in systemic lupus erythem- sponse to repeated intravenous ␥-globulin: a case report. Epilepsia. 1997;38: atosus. Clin Electroencephalogr. 1998;29:128-131. 489-494. 16. Herranz MT, Rivier G, Khamashta MA, Blaser KU, Hughes GRV. Association be- 29. Paquier PF, Van Dongen HR, Loonen CB. The Landau-Kleffner syndrome or “ac- tween antiphospholipid antibodies and epilepsy in patients with systemic lupus quired aphasia with convulsive disorder”: long-term follow-up of six children and erythematosus. Arthritis Rheum. 1994;37:568-571. a review of the relevant literature. Arch Neurol. 1992;49:354-359. 17. Chapman J, Cohen-Armon M, Shoenfeld Y, Korczyn AD. Antiphospholipid anti- 30. Pechadre JC, Sauvezie B, Osier C, Gibert J. Traitement des ence´phalopathies e´pi- bodies permeabilize and depolarize brain synaptoneurosomes. Lupus. 1999;8: leptiques de l’enfant par les gamma globulines. Rev Electroencephalogr Neuro- 127-133. physiol Clin. 1977;7:443-447. 18. Hughes GRV, Harris NN, Gharavi AE. The anticardiolipin syndrome. J Rheuma- 31. Van Rijckevorsel-Harmant K, Delire M, Schmitz-Moorman W, Wieser HG. Treat- tol. 1986;13:486-489. ment of refractory epilepsy with intravenous immunoglobulins. Int J Clin Lab Res. 19. Rand JH. Antiphospholipid antibody syndrome: new insights on thromboge- 1994;24:162-166. netic mechanisms. Am J Med Sci. 1989;316:142-151. 32. Ariizumi M, Hibio S, Ogawa K, Michihiro N, Shiihara H, Utsumi Y. Comparative 20. Lakos G, Kiss E, Regeczy N, et al. Isotype distribution and clinical relevance of study of nontreated and pepsin-treated immunoglobulin therapy for idiopathic

anti ␤2-glycoprotein I (beta2-GPI) antibodies: importance of IgA isotype? Clin Lennox syndrome. Brain Dev. 1984;6:422-425. Exp Immunol. 1999;117:574-579. 33. Maeoka Y, Hara T, Dejima S, Takeshita K. IgA and IgG2 deficiency associated 21. Day HM, Thiagarajan P, Ahn C, Reveille JD, Tinker KF, Arnett FC. Autoantibodies with zonisamide therapy: a case report. Epilepsia. 1997;38:611-613. to ␤2-glycoprotein I in systemic lupus erythematosus and primary antiphos- 34. Gilhus NE, Aarli JA. Respiratory disease and nasal immunoglobulin concentra- pholipid antibody syndrome: clinical correlations in comparison with other an- tions in phenytoin-treated epileptic patients. Acta Neurol Scand. 1981;63:34-43. tiphospholipid antibody tests. J Rheumatol. 1998;25:667-674. 35. Andrade-Mena CE, Sardo-Olmedo JAJ, Ramirez-Lizardo EJ. Effect of phenytoin 22. Sachse C, Luethke K, Hartung K, et al. Significance of antibodies to cardiolipin in administration on murine immune function. J Neuroimmunol. 1994;50:3-7.

(REPRINTED) ARCH NEUROL / VOL 57, DEC 2000 WWW.ARCHNEUROL.COM 1692