Increasedvulnerabilitytokainicacid-inducedepileptic seizuresinmiceunderexpressingthescaffoldprotein Islet-Brain1/JIP-1 Fulvio Magara, 1,2 Jacques-AntoineHae¯iger, 1 Nancy Thompson, 1 BeatRiederer, 2 EgbertWelker, 2 Pascal Nicod 1 and Ge Ârard Waeber 1,2 1 Department of Internal Medicine, BH 10±640, University Hospital CHUV, 1011 Lausanne, Switzerland 2 Institute of Cell Biology and Morphology (IBCM), University of Lausanne, Switzerland Keywords: c-Jun±N-terminal-kinase, epilepsy, excitotoxicity, gene targeting, hippocampus, JNK-interacting-protein 1, kainic acid Abstract Islet-Brain 1, also known as JNK-interacting protein-1 (IB1/JIP-1) is a scaffold protein mainly involved in the regulation of the pro- apoptoticsignallingcascademediatedbyc-Jun±N-terminalkinase(JNK).IB1/JIP-1organizesJNKandupstreamkinasesinacomplex thatfacilitatesJNKactivation.However,overexpressionofIB1/JIP-1inneurons in vitro hasbeenreportedtoresultininhibitionofJNK activation and protection against cellular stress and apoptosis. The occurrence and the functional signi®cance of stress-induced modulationsofIB1/JIP-1levels in vivo arenotknown.WeinvestigatedtheregulationofIB1/JIP-1inmousehippocampusaftersystemic administrationofkainicacid(KA),inwild-typemiceaswellasinmicehemizygousforthegene MAPK8IP1,encodingforIB1/JIP-1.We showherethatIB1/JIP-1isupregulatedtransientlyinthehippocampusofnormalmice,reachingapeak8hafterseizureinduction. HeterozygousmutantmiceunderexpressingIB1/JIP-1showedahighervulnerabilitytotheepileptogenicpropertiesofKA,whereas hippocampal IB1/JIP-1 levels remained unchanged after seizure induction. Subsequently, an increasing activation of JNK in the 8h followingseizureinductionwasobservedinIB1/JIP-1haploinsuf®cientmice,whichalsounderwentmoresevereexcitotoxiclesionsin hippocampal CA3, as assessed histologically 3days after KA administration. Taken together, these data indicate that IB1/JIP-1 in hippocampus participates in the regulation of the neuronal response to excitotoxic stress in a level-dependent fashion. Introduction The c-Jun±N-terminal kinases (JNKs) are transducers of the signalling cascades eventually leading to cell death, both during development (Kuan etal.,1999)andinresponsetochemicalorphysicalstress(Yang et al., 1997). JNK activation is operated by the mitogen-activated protein kinase kinases (MAPKK) MKK4 and MKK7, which are in turn activated by other upstream MAPKK kinases. The scaffold protein c- Jun±N-terminal-kinase-interacting-protein-1 (JIP-1) controls the spe- ci®city of this signalling cascade. It organizes the kinases MLK1, MKK7 and JNK into a structural complex that potentiates JNK activation (Davis, 2000). Mice carrying a targeted deletion of the JNK-binding domain of JIP-1 have been found to be less vulnerable to kainate-induced cell death in the hippocampus (Whitmarsh et al., 2001). Human (Mooser et al., 1999) and rat (Bonny et al., 1998) homo- logues of the murine JIP-1 were independently cloned and termed Islet-Brain 1 (IB1/JIP-1) as its expression is high in pancreatic islets and in the brain. A single missense mutation in the human gene encoding IB1/JIP-1 was found to be associated with accelerated apoptosis in insulin-secreting cells in vitro (Waeber et al., 2000). Overexpression of IB1/JIP-1, or of its JNK-binding domain, has been found to prevent JNK activation and apoptosis induced by mechanical stress (Tawadros et al., 2002), cytokines (Bonny et al., 2000) or nerve growth factor withdrawal (Eilers et al., 2001; Harding et al., 2001). These studies suggest that, whereas IB1/JIP-1 participates in JNK activation, high levels of this protein can also exert an inhibitory function on JNK cascade. Whether and when IB1/JIP-1 is expressed invivo atlevelsresultinginaninhibitoryfunctionontheJNK-mediated cascade, remains to be established. Recently, the interactions between IB1/JIP-1 and JNK have been found to be modulated by other proteins, such as the anti-apoptotic kinase Akt1 (Kim et al., 2002) and a- synuclein (Hashimoto et al., 2002). In addition, IB1/JIP-1 can form complexes with several other signal transduction proteins, such as the low density lipoprotein receptor-related proteins LRP, megalin, ApoER2, the cytoskeletal-interacting protein p190RhoGEF and amy- loid precursor protein (Meyer et al., 1999; Gotthardt et al., 2000). Finally, IB1/JIP-1 has been found to be a cargo of kinesin, a protein mediating axoplasmic transport (Verhey et al., 2001). Therefore, because of the pleiotropic roles of IB1/JIP-1 in vitro, its physiological functions in vivo are dif®cult to predict, and could be at best inves- tigated by means of pharmacological tools in a model of haploinsuf- ®ciency, heuristically more desirable than a full gene knockout (Chapman, 2002). We previously generated mice carrying a targeted disruption of the MAPK8IP1 gene, spanning from exon 3 to exon 8 (Thompson et al., 2001). The null mutation resulted in early embryonic lethality, and IB1/JIP-1 expression in the brain of mutant heterozygous mice (IB1/JIP-1 / ) was 45% lower in comparison with their wild-type littermates, indicating lack of compensatory mechanisms on the transcription of the hemizygous allele. This prompted us to investigate European Journal of Neuroscience, Vol. 17, pp. 2602±2610, 2003 ß Federation of European Neuroscience Societies doi:10.1046/j.1460-9568.2003.02701.x Correspondence: Dr Ge Ârard Waeber, as above. E-mail: gwaeber@chuv.hospvd.ch Received 16 December 2002, revised 4 April 2003, accepted 15 April 2003