Innate and adaptive immunity in human epilepsies

Inna e and adap i e immuni y in human epilepsies *Jan Baue , †Albe J. Becke , ‡§Wassim Elyaman, ¶Jukka Pel ola, #S ephan R€ uegg, **Maa en J. Ti ulae , ††James A. Va ley, and ‡‡E o e Beghi Epilepsia, 58(Suppl. 3):57–68, 2017 doi: 10.1111/epi.13784 D . Jan Baue , Depa men o Neu oimmunology, Cen e o B ain Resea ch, Medical Uni e si y o Vienna, Vienna, Aus ia. SUMMARY In lamma o y mechanisms ha e been inc easingly implica ed in he o igin o sei- zu es and epilepsy. These mechanisms a e in ol ed in he genesis o encephali ides in which seizu es a e a common complain . Expe imen al and clinical e idence sugges s di e en in lamma o y esponses in he b ains o pa ien s wi h epilepsy depending on he e iology. In gene al, ac i a ion o bo h inna e and adap i e immuni y plays a ole in e ac o y o ms o epilepsy. Epilepsies in which seizu es de elop a e in il a ion o cells o he adap i e immune sys em in he cen al ne - ous sys em (CNS) include a b oad ange o epilep ic diso de s wi h di e en (known o unknown) e iologies. In il a ion o lymphocy es is obse ed in au oim- mune epilepsies, especially he classical pa aneoplas ic encephali ides wi h an ibod- ies agains in acellula umo an igens. The p esence o lymphocy es in he CNS also has been ound in ocal ce eb al dysplasia ype 2 and in co ical ube s. Va i- ous au oan ibodies ha e been shown o be associa ed wi h empo al lobe epilepsy (TLE) and hippocampal scle osis o unknown e iology, which may be due o he p esence o i al DNA. Du ing he las decade, an inc easing numbe o an ineu- onal au oan ibodies di ec ed agains memb anous epi opes ha e been disco e ed and a e associa ed wi h a ious neu ologic synd omes, including limbic encephali- is. A majo challenge in epilepsy is o de ine bioma ke s, which would allow he ecogni ion o pa ien popula ions who migh bene i om immune-modula o y he apies. Some pe iphe al in lamma o y ma ke s appea o be di e en ially exp essed in pa ien s wi h medically con olled and medically e ac o y and, as such, could be used o diagnos ic, p ognos ic, o he apeu ic pu poses. Es ablish- ing an au oimmune basis in pa ien s wi h d ug- esis an epilepsy allows o e ica- cious and a ge ed immuno he apy. Al hough cu en immuno he apies can gi e g ea bene i o he co ec ly iden i ied pa ien , he e a e limi a ions o hei e i- cacy and hey may ha e conside able side e ec s. Thus he iden i ica ion o new immunomodula o y compounds emains o u mos impo ance. KEY WORDS: Inna e immuni y, Adap i e immuni y, Epilepsy, Au oan ibodies, Encephali ides, Immunomodula o y d ugs. Accep ed Ma ch 24, 2017. *Depa men o Neu oimmunology, Cen e o B ain Resea ch Medical Uni e si y o Vienna, Vienna, Aus ia; †Sec ion o T ansla ional Epilepsy Resea ch, Depa men o Neu opa hology, Uni e si y o Bonn - Medical Cen e , Bonn, Ge many; ‡B igham and Women’s Hospi al and Ha a d Medical School, Bos on, Massachuse s, U.S.A.; §The B oad Ins i u e, Camb idge, Massachuse s, U.S.A.; ¶Depa men Neu ology, Tampe e Uni e si y Hospi al, Tampe e, Finland; #Depa men Neu ology, Uni e si y Hospi al Basel, Basel, Swi ze land; **Depa men Neu ology, E asmus Uni e si y Medical Cen e , Ro e dam, The Ne he lands; ††Nu ield Depa men Clinical Neu osciences, John Radcli e Hospi al, Ox o d, Uni ed Kingdom; and ‡‡IRCCS-Ma io Neg i Ins i u e o Pha macological Resea ch, Milano, I aly Add ess Co espondence o Jan Baue , Depa men o Neu oimmunology, Cen e o B ain Resea ch, Medical Uni e si y Vienna, Spi algasse 4, Vienna A-1090, Aus ia. E-mail: [email p o ec ed] and E o e Beghi, Depa men o Neu oscience, IRCCS-Ma io Neg i Ins i u e o Pha macological Resea ch, Via Giuseppe La Masa 19, Milan 20156, I aly. E-mail: e o e.beghi@ma ioneg i.i ©2017 The Au ho s. Epilepsia published by Wiley Pe iodicals, Inc. on behal o In e na ional League Agains Epilepsy. This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any med- ium, p o ided he o iginal wo k is p ope ly ci ed. 57 IMMUNITY AND INFLAMMATION IN EPILEPSY (IIE2016) Immune mechanisms ha e been disco e ed in se e al neu ologic diseases, some o hem associa ed wi h epi- lepsy. 1 These mechanisms a e no only p esen in epilepsies caused by in ec ious and cen al ne ous sys em (CNS) in lamma o y diseases, 2,3 bu also in epilep ic diso de s no associa ed wi h a clea in lamma o y pa hophysiology. 4 The exac ole o he in lamma o y phenomena (cause, e ec , o bo h) is a ma e o in ense in es iga ion. Rapid ac i a ion o p oin lamma o y cy okines and dange signals is obse ed a e acu e epilep ogenic b ain inju ies o a e sin- gle and ecu en seizu es in bo h expe imen al and clinical se ings. On he o he hand, he e is e idence o ch onic o e p oduc ion o cy okines and o he in lamma o y media- o s du ing epilep ogenesis in animal models, implica ing a neu omodula o y ole o in lamma ion and i s po en ial in ol emen in he gene a ion o spon aneous seizu es. The con ibu ions summa ized in his a icle in es iga e he ole o immune mechanisms based on he esul s o ecen and ongoing expe imen s o help imp o e ou unde s anding o hei ole in epilepsy. Inna e Immuni y in Epilep ic Diso de s (Jan Baue , Jukka Pel ola) The e is e idence suppo ing ha se e al in lamma o y media o s ha e a speci ic ole in empo al lobe epilepsy (TLE) o in neoco ical epilepsies associa ed wi h ocal mal o ma ions. Rapid ac i a ion o p oin lamma o y cy okines, such as in e leukin-1b(IL-1b), in e leukin-6 (IL-6), and umo nec osis ac o -a(TNF-a), and dange signals, such as high mobili y g oup box 1 (HMGB1)-ac i- a ing in lammasomes ia Toll-like ecep o s (TLRs), is obse ed a e acu e and ch onic seizu es in animal models o acqui ed epilepsies. 5 B ain- esiden inna e immune cells such as mic oglia as well as as ocy es a e pi o al gene a- o s o his in lamma o y esponse. The e is also e idence o ch onic o e p oduc ion o hese molecules and o he in lamma o y media o s (e.g., cyclooxygenase-2, p os aglandins, complemen sys em componen s, and immunop o easomes) in bo h glial cells and neu ons, and in cellula componen s o he blood–b ain ba ie , in pa ien s wi h TLE o mal o ma ions o co ical de elop- men , sugges ing a neu omodula o y ole o in lamma ion in epilepsy. Indeed, hese in lamma o y media o s we e shown o play a ole in he mechanisms o seizu es and epilep ogenesis in animal models. 6–10 In pa icula , he p esence o IL-1band HMGB1 has been shown in b ains o pa ien s wi h a ious epilep ic diso de s, such as TLE and ocal co ical dysplasia (FCD). 6,11,12 In addi ion o he local o ma ion o seizu e-inducing molecules in he CNS, molecules such as IL-1b, TNF-a, and IL-6 also migh en e he b ain om he blood. Fo his o happen, a b each in he blood–b ain ba ie (BBB) is necessa y, and hus sei- zu e induc ion by pe iphe al molecules only can play a ole in diso de s whe e he BBB is opened. In gene al, s udies on human epilep ic b ain show he p esence o ul as uc- u al changes and abno mal igh junc ions o ascula u e endo helial cells, indica ing a b each in he BBB and he possibili y ha se um p o eins may each he CNS pa - enchyma. 13 An example o his may be child en wi h eb- ile seizu es. Measu emen o he abo e-men ioned p ocon ulsan molecules in he plasma o hese pa ien s shows a clea inc ease. 14 In addi ion, in an animal model o TLE, he BBB opening has been shown in he p og es- sion o he disease. The au ho s ound albumin in he CNS ollowing s a us epilep icus (SE) and a posi i e co ela ion be ween he ex en o BBB opening and he numbe o sei- zu es. 15 On he o he hand, i is unclea o wha ex en hese molecules en e he CNS. Measu emen s o albumin and a2-mac oglobulin in he se um and ce eb ospinal luid (CSF) o child en wi h eb ile seizu es show ha mos chil- d en ha e he same le els o albumin and a2-mac oglobu- lin in he CSF as con ols bu o he s clea ly showed an inc ease o hese molecules in he CSF. 16 I is s ill unknown whe he and in which cell ypes IL-1bis exp essed in o ms o epilepsy cha ac e ized by ex ensi e in il a ion o adap i e immuni y cells, such as Rasmussen encephali is (RE). Baue and colleagues he e o e analyzed he p esence o IL-1bin RE using immune his ochemical (IHC) s udies, and hey ound selec i e p esence o IL-1bin mic oglial nodules in co ical g ay ma e as well as in sub- co ical whi e ma e , in ag eemen wi h a ecen epo by a Ramashwamy e al. 17 In he hippocampus, IL-1bwas no es ic ed o nodules, bu was also p esen in su ounding mic oglial cells. This immunohis ochemical da a we e sup- po ed by in si u hyb idiza ion (ISH) s udies wi h a p obe o IL-1bmessenge RNA (mRNA), which showed speci ic sig- nals in mic oglial nodules. As ocy es showed a weak immunos aining o IL-1bbu we e nega i e o IL-1b mRNA, aising he possibili y ha as ocy es may impo he cy okine om he ex acellula space, o example, ia ex a- cellula mic o esicles. In addi ion o using IHC and ISH s ud- ies, hese au ho s measu ed IL-1bby wes e n blo and Key Poin s •In lamma o y mechanisms ha e been implica ed in he o igin o seizu es in a numbe o encephali ides •Ac i a ion o bo h inna e and adap i e immuni y seems o occu in e ac o y o ms o epilepsy •An ineu onal au oan ibodies ha e been disco e ed and a e associa ed wi h a ious neu ologic synd omes, including limbic encephali is •Immunomodula ion demons a es incomple e e icacy wi h signi ican side e ec s •The iden i ica ion o new immunomodula o y com- pounds emains o u mos impo ance Epilepsia, 58(Suppl. 3):57–68, 2017 doi: 10.1111/epi.13784 58 J. Baue e al. enzyme-linked immunoso ben assay (ELISA) in ozen sec- ions om se en RE pa ien s. Bo h analyses, howe e , did no show posi i e signals o IL-1b, sugges ing ha he amoun o eco e ed IL-1bwas unde he de ec ion le el. Adap i e Immuni y in Epilep ic Diso de s (Jan Baue , Maa en J. Ti ulae ) In il a ion o T lymphocy es Expe imen al and clinical e idence sugges s a di e en in lamma o y esponse in he b ain o pa ien s wi h TLE, whe e he inna e immune componen appea s o be p e a- len , as compa ed o o he e ac o y o ms o epilepsy wi h p ominen ac i a ion o bo h inna e and adap i e immuni y. Because inna e and adap i e immuni y ecip ocally a ec each o he , i is impo an o ecognize he p esence and he composi ion o he in lamma o y milieu in epilep ic diso - de s. Epilepsies in which seizu es de elop a e in il a ion o cells o he adap i e immune sys em in he CNS a e an eme ging g oup. I is inc easingly ecognized ha his g oup consis s o a b oad ange o epilep ic diso de s wi h di e en known and unknown e iologies (Fig. 1). In il a ion can be ex emely abundan , in which case he epilep ic diso de is designa ed as encephali is. Examples o hese include he pes simplex i us encephali is 18 and RE. 19 A second g oup o epilep ic diso de s wi h domi- nan in il a ion o lymphocy es a e he au oimmune epilepsies (Fig. 1). These consis o classical pa aneo- plas ic encephali ides wi h an ibodies agains in acellula umo an igens such as Hu and Ma2, as well as pa aneo- plas ic and nonpa aneoplas ic cases wi h an ibodies agains a la ge and s ill-expanding ange o in acellula and ex acellula /memb anous an igens, as discussed la e . The numbe o in il a ing lymphocy es in pa aneo- plas ic cases wi h an ibodies agains he a ious oncoge- nes is abundan . Mo eo e , mos o hese in il a ing CD3 + T lymphocy es a e cy o oxic 20,21 and a ack neu- ons. 22–24 Whe eas in pa aneoplas ic cases in il a ion is se e e, lymphocy e in il a ion in o he au oimmune epilepsies, in pa icula , hose wi h an ibodies agains memb ane an igens, is less abundan . Indeed, mode a e in il a ion o cells in cases o leucine- ich glioma-inac i- a ed 1 (LGI1) encephali is has been obse ed and lym- phocy es a e p edominan ly seen in he pe i ascula and in e s i ial spaces in cases o N-me hyl-D-aspa a e ecep- o (NMDAR encephali is) 24,25 (Fig. 1). In il a ion o lymphocy es in he CNS has also been ound in a speci- ic subg oup o FCD, ha is, hose composed o dysplas- ic neu ons and balloon cells e med FCD ype IIb by he In e na ional League agains Epilepsy (ILAE) classi- ica ion sys em. 12 In mos o he FCD cases, in il a ion o T lymphocy es is ela i ely low. Baue and colleagues con i med ha cases o FCD IIb con ain mode a e o high numbe s o T cells and showed ha some o hese in il a ing lymphocy es we e cy o oxic in na u e. Finally, A onica and colleagues ecen ly epo ed ha in lamma- o y in il a es can also be p esen in a newly iden i ied subg oup ( ype C) o co ical ube s. 26 Baue and col- leagues suppo hese indings, showing ha CD3 + CD8 + T lymphocy es a e in close apposi ion o balloon cells, hus sugges ing ha cy o oxic T cells migh speci ically a ge hese abe an cell ypes. Ano he la ge g oup o epilep ic diso de s in which in il- a ion o lymphocy es can be ound a e he cases o TLE wi h hippocampal scle osis (HS) wi h unknown e iology. Baue and colleagues p esen ed da a con i ming ha in some o hese cases, in il a ing T cells a e as nume ous as in de ini i e limbic encephali is (LE). Thei p elimina y esul s also e ealed ha especially one subg oup, ha is, Figu e 1. Compa a i e T-cell in lamma ion in epilepsy. S aining o CD3 + T cells in (A, ba : 100 lm) HSV encephali is, (B) an i-Ma2 Ab pa ane- oplas ic encephali is, (C) Rasmussen encephali is, (D) an i-GAD65 Ab encephali is, (E) an i-LGI1 Ab Encephali is, (F) an i-NMDAR (NR1) Ab encephali is. (G) FCD IIb wi h ew T cells (H) TSC. This case con ains only e y ew in il a ing lymphocy es (I) idiopa hic TLE, (J) FCD IIb. He e he pa enchyma e eals he p esence o many T cells (K) TSC ype C, wi h many T cells (L) pos in ec ious TLE. Numbe s o in lamma o y T cells a e ex emely high. All images ha e he same magni ica ion. Single T lymphocy es a e indi- ca ed by a owheads. Epilepsia ILAE Epilepsia, 58(Suppl. 3):57–68, 2017 doi: 10.1111/epi.13784 59 Immuni y in human epilepsy pos encephali ic TLE cases, con ained high numbe s o in il a ing CD3 + CD8 + T lymphocy es (Fig. 1) wi h a smal- le subg oup o G anzyme-B + T cells. TheRoleo Vi alDNAinHuman TLE (Jan Baue , Albe J. Becke ) The eason o he high numbe s o lymphocy es in pos- encephali ic TLE cases is unknown. Such in lamma ion migh be due o he p esence o i al DNA in he CNS. High human he pes i us 6 (HHV-6) DNA load has been de ec ed in su gical issue o pa ien s wi h TLE, and i al DNA is mos commonly p esen in pa ien s wi h p e ious in lamma- o y b ain diseases. 27 Howe e , de ec ion a es ha e been ound o a y conside ably among s udies. Becke and col- leagues, in an ini ial limi ed se ies o 38 pha maco esis an TLE pa ien s e sus 10 au opsy con ols, de ec ed HHV-6 DNA in 55.6% o he TLE pa ien s wi h a his o y o encephali is, including mesial empo al scle osis (MTS) and glio ic hippocampi wi hou subs an ial neu odegene a- ion. 27 They did no ind HHV-6 DNA in lesion-associa ed TLE o nonlesional MTS wi h o wi hou a his o y o com- plex eb ile seizu es (CFS). These da a p omp ed hem o ca y ou a subsequen la ge-scale analysis o i al DNA/ RNA spec um in an ex ended se ies o TLE biopsies. 28 In addi ion o all He pes i idae, he g oup examined he p es- ence o po en ially ele an neu o opic RNA i uses. DNA and RNA we e ex ac ed om 346 esh- ozen epilepsy su ge y issue samples. F esh- ozen hippocampal issue samples om 62 pa ien s wi hou ch onic CNS disease se ed as con ols. Real- ime polyme ase chain eac ion (PCR) and nes ed PCR we e pe o med o He pes i idae and RNA i uses, espec i ely. In addi ion, hey analyzed he clinical eco ds o he pa ien s o he p esence o ea lie signs o in lamma o y b ain eac ions. Thei esul s e ealed HHV-6B DNA in 9.8% o he TLE pa ien s and in 12.9% o he con ol samples. In iguingly, howe e , he TLE sam- ples we e ound o ha e a highe i us concen a ion. In pa ien s wi h clinical signs o p e ious b ain in lamma ion, HHV-6B DNA was obse ed in 15.0%, whe eas only 6.3% o he samples om pa ien s wi hou eb ile seizu es o meningoencephali is we e posi i e o HHV-6B DNA. A me a-analysis o he eigh HHV-6 PCR s udies e ealed simila esul s. To summa ize, he biopsy-based s udy o Becke and colleagues e ealed no di e ences in equency o HHV-6B DNA de ec ion be ween TLE pa ien s and con- ols. Howe e , he highe i us load in TLE pa ien s and he inc eased de ec ion a e o HHV-6B DNA in pa ien s wi h p e ious in lamma o y b ain eac ions may be in line wi h a po en ial p omo ing ole o HHV-6 in TLE pa ien s wi h a his o y o encephali is. I is in e es ing o no e ha especially pos encephali ic TLE cases e ealed inc eased p esence o cy o oxic lymphocy es, as p esen ed by Baue and colleagues. As such, new coo dina ed mul icen e a emp s should be help ul o shed mo e ligh on he ole o HHV-6 as well as o he neu o opic i uses in he eme - gence o TLE. The abundance o i al nucleic acids in epilep ic hippocampal issue ep esen s an aspec ha equi es pa icula a en ion by u u e s udies. The Role o Immunologic Bioma ke s in Re ac o y Epilepsy (Wassim Elyaman) Conside able p og ess has been made in iden i ying ci - cula ing epilepsy bioma ke s in animal models o epi- lepsy. 29 Howe e , a majo challenge is o de ine bioma ke s in human epilepsy o allow he ecogni ion o app op ia e pa ien popula ions ha migh bene i om immunomodula- o y he apies. Elyaman and colleagues p oposed he hypo hesis ha pe iphe al in lamma o y media o s a e di - e en ially exp essed in pa ien s wi h medically con olled and medically e ac o y epilepsy and, as such, could be used as a bioma ke o diagnos ic, p ognos ic, o he apeu- ic pu poses. To cha ac e ize he egula ion o ci cula ing in lamma o y media o s in epilepsy, Elyaman and col- leagues assembled a mul idisciplina y eam ha includes epilep ologis s, immunologis s, epigene icis s, and compu- a ional biologis s o gene a e comp ehensi e p o eomic and ansc ip omic da a o pe iphe al (blood) and cen al (su gical issues) adap i e and inna e immune cells o a la ge pool o pa ien s wi h epilepsy. Findings om hese s udies may gi e insigh in o he complex ole o in lamma- ion in he gene a ion and exace ba ion o epilepsy. In addi- ion, i should yield new molecula a ge s o he design o an iepilep ic d ugs, which migh no only inhibi he symp- oms o his diso de , bu also p e en disease pa hogenesis. An ibodies and Epilepsy (Maa en J. Ti ulae ) In he pas 30 yea s >30 an ineu onal au oan ibodies (ANABs) ha e been disco e ed. 30 These ANABs a e associ- a ed wi h a ious neu ologic synd omes, which include LE, epilepsy, ce ebella degene a ion, pe iphe al (poly)neu opa- hy, Lambe -Ea on myas henic synd ome (LEMS), opso- clonus-myoclonus synd ome p og essi e encephalomyeli is wi h igidi y and myoclonus (PERM), and s i -pe son syn- d ome (SPS). A p esen , i is unclea how he o ma ion o hese an ibodies is igge ed. The induc ion o an ibodies agains he NMDAR has been associa ed wi h pos -he pes simplex i us encephali is, 31–33 bu o o he an ibodies such an associa ion is absen . The exac ole o hese an ibodies in seizu e induc ion emains he ocus o esea ch. These in es i- ga ions ha e al eady e ealed impo an mechanis ic insigh o a mino i y o an ibodies. Epilepsia, 58(Suppl. 3):57–68, 2017 doi: 10.1111/epi.13784 60 J. Baue e al. In he nine ies, ol age-ga ed po assium channel (VGKC) an ibodies we e desc ibed in associa ion wi h neu omyo o- nia. 34 In he ollowing yea s, hese an ibodies we e ound o be p esen in pa ien s wi h LE and in Mo an’s synd ome. 35 Mo e ecen ly, i was shown ha abou hal o hese an ibod- ies a e di ec ed agains p o eins complexed wi h he VGKC a he han he channel i sel . Mos pa ien s ha e been ound o ha e an ibodies agains LGI1. 36,37 A smalle g oup has an ibodies agains con ac in-associa ed p o ein-like 2 (Casp 2), which is si ua ed in he jux apa anodal loop o myelin su ounding axons. 37,38 Only a mino i y o pa ien s ha e an ibodies agains con ac in-2. 37 Recen ly, a p opo - ion o pa ien s wi h posi i e VGKC-complex an ibodies bu nega i e LGI1 and casp 2 an igen speci ic es s ha e been ound o con ain an ibodies agains he in e nal su ace o he VGKC. 39 The clinical signi icance o hese pa icula an ibodies is unclea , bu cu en ly he e is no e idence o a pa hogenic ole o a posi i e VGKC es in he absence o LGI1 o Casp 2 an ibodies. 40 The e o e, he e m VGKC is no use ul and diseases should be e e ed o by hei con- i med a ge an igen. Mechanis ically, an i-LGI1 an ibodies seem o a ec he binding o LGI1 o disin eg in and me allop o einase domain-con aining p o ein 22 (ADAM22), and he eby p e- en p ope signaling o he p esynap ic VGKC and he pos synap ic a-amino-3-hyd oxy-5-me hyl-4-isoxazolep o- pionic acid (AMPA) ecep o . 41 The mechanism o an i- casp 2 an ibodies has no been s udied ye , bu pa hologic in es iga ion o a casp 2 encephali is b ain showed immunoglobulin and complemen deposi ion associa ed wi h neu odegene a ion in he hippocampus. 42 This an i- body and complemen deposi ion–associa ed neu odegene - a ion was also ound in hippocampi o pa ien s wi h an i- LGI1 encephali is, sugges ing ha , a leas in hese pa icu- la cases, he casp 2 and LGI1 an ibodies can induce neu- odegene a ion ia an ibody-dependen complemen - media ed cy o oxici y. 24 All LGI1 and casp 2 se a con ain (non–complemen -ac i a ing) IgG4 subclass an ibod- ies. 43,44 In addi ion o IgG4, IgG1 o IgG2 subclass an ibod- ies we e ound o some ex en in 42% o LGI1 and 63% o casp 2 se a. These IgG1 and IgG2 an ibodies a e comple- men -ac i a ing and he e o e migh be esponsible o he hippocampal neu odegene a ion desc ibed in some pa ien s. 24,42 Because he an ibodies agains su ace an igens ha e been disco e ed only ecen ly, knowledge o he exac pa hophysiologic mechanisms o such an ibodies is s ill expanding. An i-NMDAR an ibody was he i s an ibody o be disco e ed eac ing o an ex acellula an igen. 45 An i-NMDAR encephali is a p esen is also he mos e- quen au oimmune epilepsy. Mos o he mechanis ic insigh o an i-NMDAR an ibodies has been desc ibed by Dalmau and colleagues. 46–48 In i o, he an ibodies seem o a ec he localiza ion o NMDAR on he synapses o inhibi o y neu ons, by b eaking he link wi h Eph inB2. 49 This change in localiza ion somehow leads o in e naliza- ion o he NMDAR and a ne dec ease in NMDAR. 48,50 In i o and in i o s udies in mice e eal ha he dec ease o NMDAR on he synapse is e e sible a e emo al o an ibodies. 51 Fu he mo e, in i o expe imen s sugges ha an o e load o Eph inB2 ligand can p e en he NMDAR in e naliza ion, and pa ially p e en he clinical pheno ype. This could o e al e na i e symp- oma ic ea men s in he u u e. Unlike in LGI1 and Casp 2 encephali is, no majo pa hologic e ec on indi- idual ecep o unc ion o cy o oxic o complemen - media ed e ec s ha e been ound. 24,25,52 A ques ion he e- o e emains why IgG complemen –ac i a ing subclass an ibodies om NMDAR encephali is pa ien s do no seem o induce complemen -media ed pa hology. Clinical Aspec s o Immuni y in Epilepsy (S ephan R€ uegg, Maa en J. Ti ulae ) F om an epilep ologic poin o iew, LE is he mos impo an immune synd ome associa ed wi h he newly dis- co e ed ANABs. The main ea u es o he synd ome a e subacu e onse o wo king memo y de ici s, al e ed men al s a us, con usion, deli ium, o psychia ic symp oms. Fu - he mo e, one can ind p og essi e impai men o con- sciousness up o coma. Fo he possibili y o an LE, a leas one o he ollowing obse a ions should be p esen : de no o seizu es o e en s a us epilep icus, a new ocal neu o- logic de ici , CSF pleocy osis, o MRI ea u es sugges i e o encephali is. Excluding al e na i e causes o he pa ien ’s condi ion is also impo an . Swi ecogni ion o he diso de is key because he e a e e ec i e he apies and da a inc easingly sugges ha ea ly s a o ea men is associa ed wi h be e ou come. 30,53 Excellen e iews on his opic ha e been published ecen ly. 54–58 Ano he impo an an ibody associa ed wi h epilepsy was he en i y di ec ed agains he cy osolic enzyme glu a- ma e deca boxylase (GAD). GAD ca alyzes he con e - sion om glu ama e o c-aminobu y ic acid (GABA), and he p esence o an ibodies agains GAD a e associa ed wi h epilepsy, 59 s i pe son synd ome (SPS), and LE 60,61 (Table 1). Mos o he ea ly de ec ed ANABs (de ec ed be o e 2007) ecognize in acellula epi opes and we e associa ed wi h umo s causing pa aneoplas ic synd omes, whe eas hose ANABs iden i ied mo e ecen ly a e less equen ly pa aneoplas ic and a e di ec ed mainly owa d an igens a he cell memb ane. An i-GAD an ibody encephali is is an excep ion because he an ibodies agains in acellula an igens a e induced in he absence o a umo . A con o e sial ansien ex acellula a ailabili y a he synapse has been sugges ed o explain hei pa hogenici y. 62 Since 2007, many new ANABs ha e been disco e ed (Table 2) and he clinical di e si y o he Epilepsia, 58(Suppl. 3):57–68, 2017 doi: 10.1111/epi.13784 61 Immuni y in human epilepsy ANAB-media ed synd omes has la gely expanded, main- aining he link wi h epilepsy in mos o hese ANABs. These aspec s a e discussed in b ie below. An i-NMDAR encephali is In 2013, Ti ulae e al. 53 p esen ed a la ge coho o 577 pa ien s, including 212 child en, wi h NMDAR encephali- is, and epo ed p ecise da a on clinical ea u es, ea men , and ou come. The majo i y (81%) had a good ou come (modi ied Rankin ou come scale 0–2), whe eas 6% died. P ognos ic ac o s p edisposing o a good ou come in bo h adul s and child en we e ea ly onse o ea men and nonad- mission o an in ensi e ca e uni . Relapse isk was 12% wi hin he i s 24 mon hs, bu elapses ended o be milde han he ini ial mani es a ion. The EEG phenomenon o “ex- eme del a b ush”(ex ensi e on al be a ac i i y supe im- posed on almos gene alized hy hmic high-ampli ude del a ac i i y) occu s in abou 30% o se e ely a ec ed in ensi e ca e uni (ICU)–bound pa ien s and is hough o be speci ic o he diso de . 63 Psychia ic symp oms a e impo an in he ea ly cou se o he disease, as 75% o adul pa ien s a e seen ini ially by a psychia is . 64 These symp oms we e he ini ial mani es a ion in 59% o he na ional F ench coho (n =111), wi h 40% expe iencing hallucina ions and 23% dep ession. A mino i y o hese pa ien s we e d ug esis an o an ipsycho ics o an idep essan s. Re e al o a neu o- logic (in ensi e ca e) uni was o many o he causes like e ol ing seizu es and s a us epilep icus and, in pa icula , o de elopmen o a neu olep ic malignan synd ome in 21 pa ien s. Fo y pe cen o he pa ien s we e i s hospi alized in psychia ic ins i u ions, howe e , mo e han hal o hese pa ien s also had a leas one neu ologic sign and u he 38% de eloped neu ologic symp oms wi hin days. These indings con i m p e ious s udies o an i-NMDAR encephali is cases a psychia ic wa ds 65 and ask o ca e ul neu ologic ( e-)examina ion o pa ien s wi h new-onse , el- a i ely acu e psychia ic illness. 66 In young child en, sei- zu es and dyskinesias a e he ini ial symp oms in hal o he pa ien s. In his g oup, psychia ic ea u es as p esen ing symp oms, howe e , a e less equen (abou one- hi d). 53 The symp oms o he diso de e ol e g adually and seem o esol e in e e se o de o onse . 67 An i-LGI1 encephali is Sho ly a e he disco e y o he LGI1-associa ed LE, I ani e al. 68 epo ed a unique epilep ic synd ome o “ acio- b achial dys onic seizu es”(FBDS), wi h o he g oups epo ing simila semiologies in he same con ex . 69 These seizu es a e e y sho , las ing abou 1–3 s, and consis o sho dys onic s i ening o one a m and g imacing o he ipsila e al mimic muscula u e, o , less equen ly, he leg. They may occu up o 50–100 imes pe day and, in 70% o cases, hey p ecede he onse o LE by an a e age o 35 days. Abou 10% o pa ien s do no de elop LE. Due o hei unusual and cha ac e is ic semiology, FBDS may be missed o misin e p e ed as “psychogenic”o “ ics.”Elec- oencephalog aphy (EEG) a ely shows epilep ic abno - mali ies. 68 LGI1 encephali is is he second mos equen au oimmune encephali is, wi h an incidence o sligh ly less han one pe million. 40 I is in equen ly (10–15%) associ- a ed wi h umo s ( hymoma, lung cance ), and is sligh ly mo e p e alen in men and in pa ien s >50 yea s. Seizu es a e ei he ocal in he ea ly s ages (FBDS o s e eo ypic ocal seizu es wi h dyscogni i e o au onomic ea u es) 68 o gene alized onic–clonic, a la e symp om as pa o an LE. Wi h he onse o LE, memo y loss, con usion, and pe son- ali y changes a e equen . 70 Mild o mode a e hypona- emia (115–130, mos o e 125 mmol/L) is p esen a onse in wo hi ds o pa ien s. The ou comes can be se ious, as 65% o pa ien s ha e pe sis en mild o se e e cogni i e de - ici s. 40,71 In addi ion, 13% o pa ien s de elop a cogni i e encephalopa hy wi hou seizu es. 43 A e adequa e ea - men , an i-LGI1 encephali is e ol es in o ch onic epilepsy in one in se en pa ien s. An i-Casp 2 encephali is Mo e a e han LGI1 encephali is, an i-Casp 2 encephali- is has been epo ed in abou 100 pa ien s o da e. E e y i h pa ien has pa aneoplas ic disease (lung, colon cance , Table 1. An ibodies associa ed wi h epilepsy: s a e o a in 2010 An ibody Epi ope loca ion (Non-)pa aneoplas ic Re s An i-Hu In acellula Pa aneoplas ic 118,119 An i-Ma1; Ma2/an i-Ta In acellula Pa aneoplas ic 120,121 An i-amphiphysin Memb anous Pa aneoplas ic 92 An i-Ri In acellula Pa aneoplas ic 122 An i-AMPAR (GluR1/2) Memb anous 2/3 Pa aneoplas ic 73 An i-GABA B R Memb anous 1/2 Pa aneoplas ic 123 An i-GAD Cy osolic ca. 10% Pa aneoplas ic 59 An i-AMPAR (GluR3) Memb anous Nonpa aneoplas ic 124 An i-NMDAR (NR1/2) Memb anous 1/3 Pa aneoplas ic 45 An i-LGI1 Memb anous 10% Pa aneoplas ic 36 An i-Casp 2 Memb anous 20% Pa aneoplas ic 37 Table 2. An ibodies associa ed wi h epilepsy disco e ed since 2010 An ibody Epi ope loca ion (Non)pa aneoplas ic Re s An i-mGluR5 Memb anous Pa aneoplas ic 79 An i-DPPX-6 Memb anous Ra ely pa aneoplas ic 82 An i-GABA A R Memb anous Me ely non-pa aneoplas ic 88 An i-glycineR Memb anous Non-pa aneoplas ic 95,125 An i-GABA A R Memb anous Me ely non-pa aneoplas ic 90 Epilepsia, 58(Suppl. 3):57–68, 2017 doi: 10.1111/epi.13784 62 J. Baue e al. hymoma). The e is s ong p eponde ance o male pa ien s (60–95% in di e en se ies), mainly olde han age 60. Pa ien s can p esen wi h LE, a axia, pe iphe al ne ous sys em in ol emen (neu omuscula hype ex- ci abili y, including neu omyo onia, o pain) o a combi- na ion o pe iphe al and cen al symp oms and signs, such as Mo an’s synd ome. 44,71,72 Seizu es occu in 50–90% o pa ien s. 44 An i-AMPAR encephali is A e a i s se ies o 10 pa ien s, 73 he e a e now >50 pa ien s epo ed wi h AMPA encephali is, which has a pa aneoplas ic ( hymoma, lung cance ) backg ound in wo hi ds o hese pa ien s. 73–75 The e is emale p eponde ance and pa ien s end o be olde han 40 yea s o age. The ol- lowing ou ypes ha e been epo ed: (1) a con usional ype wi h p edominan agi a ion up o psychosis; (2) a mainly amnes ic o m; (3) an epilep ic o m wi h gene alized and mo e a e ocal seizu es, wi h impai men o consciousness; and (4) a ulminan ype wi h a se e e cou se. Seizu es occu in 20–40% o pa ien s only, and pa ien s end o elapse al hough ea men esponse is good e en in hese ecu en episodes. 72,74 An i-GABA B R encephali is This o m o encephali is is inc easingly ecognized and among he mos p e alen a e NMDAR and LGI1 encephali is. Mo e han hal o cases a e pa aneoplas ic, mainly associa ed wi h small cell lung cance . O in e es , an i-GABA B ecep o an ibodies a e absen in la ge coho s o pa ien s wi h small cell lung cance , indica ing ha he p esence o he an ibodies can exe hei pa hogenic e ec in he absence o a umo . The e is p eponde ance o male pa ien s olde han 40 yea s who p esen wi h seizu es (pa - ial complex, seconda ily gene alized seizu es, o equen ly e ac o y s a us epilep icus [SE]). In addi ion, pa ien s ha e memo y loss, con usion, and se e e insomnia. Ra e symp- oms include a axia and opsoclonus/myoclonus syn- d ome. 76–78 An i-mGluR5 encephali is Encephali is media ed by an ibodies agains he me abo- opic glu ama e ecep o sub ype 5 (mGluR5) is always pa aneoplas ic and associa ed wi h Hodgkin’s lymphoma. 79 The signs o his ex emely a e condi ion (only se en cases epo ed) include ma ked and subacu e memo y loss, dep ession, delusions, hallucina ions, and psychosis oge he wi h umo symp oms o cachexia, e e , and nigh swea s. The combina ion o encephali is, dep ession, delusion, and cachexia led o he e m “Ophelia syn- d ome,”coined by he au ho o he i s case (his own daugh e ). 80 An ibodies agains he mGluR1 ecep o cause a comple ely di e en neu ologic, ce ebella (a axic) syn- d ome despi e la ge homology o he mGluR1 ecep o wi h he mGLuR5 ecep o . 79,81 An i-dipep idyl-pep idase-like p o ein-6 (DPPX) encephali is This is a a e synd ome whe e an ibodies agains he epi- ope DPPX o he ol age-ga ed po assium channel K 4.2 cause classical symp oms o LE wi h con usion, memo y loss, psychosis, o dep ession, bu addi ionally b ains em signs (eye mo emen dis u bances, a axia, dysa h ia, dys- phagia, and espi a o y ailu e), sleep di icul ies, and myo- clonus. As a unique clinical hallma k, p o use and di icul o ea , nonin ec ious dia hea p ecedes LE weeks o mon hs in wo hi ds o cases. 82–84 This may esul om he exp ession o he epi ope in bo h in es inal cells and neu- ons. Seizu es a e a ely obse ed (abou 15%). The an i- bodies ha e also been ound in cases o p og essi e encephalomyeli is wi h igidi y and myoclonus (PERM). 85 An i-GABA A(b3/c2) R-encephali is LE associa ed wi h an ibodies agains GABA A ecep o subuni s b3andc2 is a e and, in wo hi ds o cases, a ec s men be ween 3 and 74 yea s. The eason o his gende dis- pa i y is unknown bu in e es ing, since male p edominance in au oimmune diso de s is a e. One migh specula e ha emale sex ho mones, such as allop egnanolone, o o al con- acep i es may consolida e c2-subuni s 86 and be p o ec i e agains seizu es and SE. GABA A ecep o s, especially he ex asynap ic, can be modula ed by neu os e oids wi h sex- ho mone–like s uc u e. 87 Abou 20% o pa ien s ha e can- ce . Main symp oms a e he ones ypical o LE, bu 50–70% o pa ien s ha e se e e seizu es, up o almos in ac able SE. In addi ion, some pa ien s also ha e SPS o opsoclonus/my- oclonus synd ome. Response o seizu es and SE o immuno- supp ession is ema kable, whe eas classical an iseizu e d ugs a e ine ec i e agains epilep ic ac i i y. 88,89 An i-GABA A(a1/c2) R encephali is A second ype o an i-GABA A ecep o an ibody has been epo ed ecen ly. These an ibodies we e disco e ed in a la ge coho o pa ien s es ed o au oimmune CNS disease. The an ibodies we e di ec ed agains he ecep o sub ypes a1 and c2, and high i e s a e associa ed wi h bo h seizu es and memo y impai men in 47% o pa ien s, in 33% wi h hallucina ions, and 20% wi h anxie y. 90 A clea encephali ic synd ome could no be iden i ied and impo an clinical in o ma ion on hese pa ien s ( i e s in CSF, cou se o dis- ease, e ec o ea men , e c.) we e lacking; hus hese an i- bodies may no cause LE sensu s ic o. 91 An i-amphiphysin encephali is An i-amphiphysin encephali is has been e y a ely epo ed. Abou 40% o cases we e pa aneoplas ic and mainly caused by small lung cell and b eas cance . Men we e mo e equen ly a ec ed, and mean age o onse is 54 yea s. The classical o m includes memo y loss, cogni- i e decline, mood changes, dep ession, and some imes hal- lucina ions up o a ank psychosis. Seizu es occu in abou Epilepsia, 58(Suppl. 3):57–68, 2017 doi: 10.1111/epi.13784 63 Immuni y in human epilepsy 40% o pa ien s (gene alized > ocal wi h impai men o consciousness). Some pa ien s also expe ience b ains em signs, like e igo, c anial ne e palsies, and a axia. 92,93 An i-glycineR encephali is Al hough abou 3% o wo la ge coho s o adul pa ien s wi h newly diagnosed and es ablished epilepsy ha bo ed an ibodies agains glycine ecep o s, 94 only h ee young boys younge han age 6 yea s ha e been epo ed o ha e (sub-)acu e encephali is, ocal seizu es, and e en SE associ- a ed wi h hese an ibodies. 95–97 The an i-glycine ecep o an ibodies a e no mally linked o SPS and PERM. 98 In summa y, he incidence o all ypes o encephali is is es ima ed a 1:15,000 99 and he LE caused by ANABs is p obably a ound 1:100,000, al hough exac epidemiologic da a a e lacking. The incidence o sei- zu es in hese cases s ongly a ies acco ding o he speci ic ype o ANABs (Table 3). Con e sely, ANABs a e ound in abou 1–3% o pa ien s wi h newly es ab- lished epilepsy, bu in a highe p opo ion (10–15%) in pha maco esis an pa ien s. 94,100,101 This emains an impo an a ea o unce ain y because he signi icance o he p esence o ANABs in pa ien s wi h epilepsy alone is s ill poo ly unde s ood, as is he alue o an ibody es ing in high- h oughpu coho s, wi hou u he explo a ion o he signi icance o hese posi i e esul s. A e hey me ely bys ande o a eliable su oga e ma - ke o pha maco esis ance when hese pa ien s we e ea ed only wi h an iseizu e d ugs bu no immunomod- ula ing agen s? Such ea men s yielded ema kable he - apeu ic e ec s in one s udy o highly selec ed pa ien s, bu p ospec i e andomized con olled s udies a e s ill lacking. 102,103 T ea men s in Au oimmune Encephali is (James A. Va dley) The i s -line ea men o au oimmune encephali is com- bines s e oids wi h ei he plasma exchange o in a enous immunoglobulin (IVIG) o , occasionally, bo h. Second-line ea men s a e wi h cyclophosphamide and i uximab. 53 The bes e idence o ea men o au oimmune encephali is is p esen ed in NMDAR-Ab encephali is, as his is he mos common. 53 A ound 50% o NMDAR-Ab pa ien s espond o i s -line ea men wi h a “good ou come”o an modi ied Rankin scale (mRS) o 0–2 a 24-mon h ollow-up. In he emaining 50%, a good ou come can be achie ed by sec- ond-line ea men . The e a e nuances in his da a, as only 50% o pa ien s s udied e ospec i ely had ollow-up da a a 24 mon hs. Second-line ea men also hal ed he elapse incidence in 25% o pa ien s, and ea ly ea men was asso- cia ed wi h a highe chance o a be e ou come, some hing ha has been mi o ed in a small p ospec i e coho in LGI- Ab encephali is. 104 E idence o i uximab is equi ocal in LGI-Ab encephali is bu hese da a a e on a small sample size. 105 In o he an ibody encephali ides, because o hei a i y, he e is a less clea idea o ea men esponses. In gene al, a ound 70% o pa ien s espond o ea men . He e, ac o s associa ed wi h a poo ou come include coexis en umo , delay o ea men , and poo ini ial unc ional s a us. In addi- ion, he highe mo ali y seen in cases associa ed wi h an i- bodies such as GABA A , GABA B , and AMPA- ecep o an ibodies can be explained by he mo e equen occu - ence wi h umo s. 106 Nex -Gene a ion Ta ge ed Immunomodula o y The apies (James A. Va dley) Mo e speci ic a ge ed he apies a e an a ac i e p o- spec , as he b oad immunosupp essi e he apies ha e a wide ange o side e ec s ha con ibu e signi ican ly o mo bidi y. E o s ha e been made using he ollowing med- ica ions o in e en ions. Immunoabso p ion Plasma exchange is well es ablished as a ea men op ion, bu a ial o immunoabso p ion (IA) was ecen ly epo ed. 107 Eigh y-six pe cen o pa ien s wi h ANABs imp o ed, whe eas pa ien s wi h in acellula ly a ge ed an ibodies an igens did no . An ibody clea ance was also in e es ing, wi h i e s dec easing by 97% (se um) and 64% (CSF) 4 days a e IA and alling u he o 98% (se um) and 88% (CSF) a 4 weeks, showing a p olonged ea men e ec . I is wo h no ing ha his is highe han epo ed in Table 3. F equency o seizu es du ing/a e limbic encephali is (LE) An ineu onal an ibody F equency o seizu es Incidence o LE An i-GABA B R-IgG 1 90% Low o medium An i-GAD-IgG 25–100% a High An i-Hu-IgG 60–100% Low o medium An i-GABA A R-IgG 1>3 50–100% Medium An i-LGI1-IgG 4>2>1 90% Medium o high An i-NMDA1/2R-IgG 1 70% High An i-GABA A(b3/g2) R-IgG 1>3 47% Medium An i-Casp 2-IgG 4>1 20–65% Low o medium An i-Ma1/2-IgG 30–40% Low o medium An i-AMPAR-IgG 1 33% Low o medium An i-DPPX-IgG 15% Low An i-mGluR5-IgG 20% Ve y low An i-amphiphysin-IgG 10–20% Low An i-CV2/CRMP5-IgG Ra e Low o medium An i-glycineR-IgG Ra e Low a Depending on he de ini ion o acu e immune LE. Epilepsia, 58(Suppl. 3):57–68, 2017 doi: 10.1111/epi.13784 64 J. Baue e al. plasma exchange se ies 108 ; howe e , hose p io s udies we e published 20 yea s ago and ec ui ed pa ien s wi h in acellula an igens. The p ecise clinical e ec o IA is con ounded by he coadminis a ion o immuno he apy in his s udy, bu i does appea o be an e ec i e ea men alongside s anda d immuno he apy and spa es he use o p ecious ans usion p oduc s equi ed wi h plasma exchange and a oids he associa ed side e ec s. Tocilizumab B cells a e hough o be c ucial o he ongoing immune esponse in pa ien s wi h ANABs. IL-6 is a cy okine hough o be a key playe in B-cell ma u a ion and e minal di e - en ia ion in o plasma cells, which sec e e copious an i- body. 109 A s udy o neu omyeli is op ica (NMO) wi h aquapo in-4 an ibody–posi i e pa ien s iden i ied IL-6 as a key cy okine o B-cell ma u a ion and disease-speci ic an ibody p oduc ion. 110 P elimina y da a using Tocilizu- mab, an an i-IL-6 monoclonal an ibody, in NMO and au oimmune encephali is is encou aging, and mo e wo k mus be done o assess he e icacy o his as a s andalone second-line he apy. 111 Low-dose IL-2 A Ko ean g oup has ea ed pa ien s wi h ocilizumab in combina ion wi h low-dose IL-2 o s imula e T egula o y cells, which is known o educe T-cell ac i a ion and a enu- a e B-cell ma u a ion in ge minal cen e s, wi h encou aging esul s. 112 A ca ea is ha i is impossible o un angle he e ec o indi idual ea men s when mul iple ones a e gi en con empo aneously. This s udy is u he hampe ed by small sample size and he absence o con i ma ion o neu- onal su ace an ibodies in some o he pa ien s. Bo ezomib Plasma cells a e key componen s o he humo al immune esponse. Bo ezomib was o iginally designed as a ea - men o myeloma and unc ions by inhibi ing p o easome unc ion in cells. 113 This is hough o a ec he immune sys em in a a ie y o ways, bu plasma cells, due o hei p o ein syn hesis, a e s ongly dependen on p o easomal unc ion. Bo ezomib, he e o e, is hypo hesized o cause an accumula ion o p oapop o ic ac o s and a ge ed cell dea h, especially in plasma cells. This ea men has been ialed in NMDAR-Ab encephali is in wo e ac o y pa ien s and demons a ed p omising ini ial esul s. 107 Fu - he s udies wi h ca e ully selec ed pa ien s a e needed. Immunomodula o y ea men s a e no wi hou ad e se side e ec s. Ad e se e ec s o immunoabso p ion include coloniza ion o ca he e ip wi h coagulase-nega i e s aphy- lococci and enous ai embolism. 107 Uppe espi a o y ac in ec ion and pha yngi is o nasopha yngi is can be e- quen ly obse ed in pa ien s ecei ing acilizumab. 114 In hese pa ien s, clinical labo a o y abno mali ies include neu openia and ele a ed amino ans e ase le els. One o he mos common side e ec s o bo ezomib is pe iphe al neu opa hy ha is p edominan ly senso y wi h bu ning pa es hesia, hype es hesia–hypoes hesia, neu opa hic pain, and weakness. 115 Conclusions and Fu u e Di ec ions An i-in lamma o y d ugs ha e been epo ed o con ol seizu es in d ug- esis an epilepsy and in selec ed epilepsy synd omes e en in he absence o a clea in lamma o y basis. An au oimmune e iology is inc easingly iden i ied among pa ien s wi h epilepsy in whom no unequi ocal causes a e de ec ed wi h he p esen diagnos ic aids. In addi- ion, seizu es can ini ia e b ain in lamma ion in glial cells and p omo e BBB dis up ion independen o leukocy es o blood-bo ne in lamma o y molecules. 116 Howe e , wi h one excep ion, 117 he p esen e idence o seizu e educ ion by an i-in lamma o y d ugs in humans elies mos ly on case epo s o small se ies. In hese cases, a chance associa ion be ween he disease and biological ma ke s o al e ed immuni y is s ill possible. We canno e en con i m ha d ug esis ance has an au oimmune basis in hese pa ien s because a mo e igo ous in es iga ion o immune mecha- nisms is ypically pe o med in pa ien s wi h he mos se e e disease a ie ies. Resea ch in he ield o au oimmune epilepsies is a a as pace, wi h mul iple new au oan ibodies disco e ed e e y yea . Nume ous ques ions mus be answe ed. These can be di ided in hose ega ding basic mechanisms and hose ega ding clinical e alua ion and he apy. Fo ins ance, a p esen i is comple ely unclea why an ibodies such as an i-LGI mos ly a ge he limbic egions bu almos no pa hologic o clinical changes a e ound in egions ha also ha e high concen a ions o he a ge an igen such as he ce ebellum. Some o he mo e basic ques ions ha he e o e need o be add essed in he u u e a e he concep o an ibody o ma ion, he mecha- nisms by which he an ibody gains access o he CNS, and he eason ha he BBB ( o ins ance in an i-LGI encephali- is) seems o be b eached mos ly in limbic s uc u es. Fu - he mo e, he associa ion o he pes simplex i us wi h an i- NMDAR an ibodies sugges s ha i al in ec ion, a leas in some pa ien s wi h an i-NMDAR encephali is, migh acili- a e a b eak in ole ance gi ing ise o disease. I is s ill unclea whe he o he an ibody-associa ed limbic epilepsies a e also seconda y o a p ima y ( i al) in ec ion. F om a clinical poin o iew he sea ch o u he an i- bodies is equi ed because his may dec ease he numbe o pa ien s wi h clinical, MRI, and CSF indings o LE bu no iden i iable ANABs. A he same ime, i is impo an ha awa eness o cha ac e is ic clinical synd omes is imp o ed o allow o ea ly iden i ica ion and ea men , gi en he bene icial e ec o p omp ea men on ou come. In addi- ion, he ques ion should be asked whe he i is use ul ha Epilepsia, 58(Suppl. 3):57–68, 2017 doi: 10.1111/epi.13784 65 Immuni y in human epilepsy