Contribution of Aberrant GluK2-Containing Kainate Receptors to Chronic Seizures in Temporal Lobe Epilepsy

Transcription

1 Cell Reports Report Contribution of Aberrant GluK2-Containing Kainate Receptors to Chronic Seizures in Temporal Lobe Epilepsy Angélique Peret, 1,2,4 Louisa A. Christie, 1,2,4 David W. Ouedraogo, 1,2 Adam Gorlewicz, 3 Jérôme Epsztein, 1,2 Christophe Mulle, 3 and Valérie Crépel 1,2, * 1 INSERM, INMED, U91, 139 Marseille, France 2 Aix-Marseille Université, UMR 91, 139 Marseille, France 3 Interdisciplinary Institute for Neuroscience, CNRS UMR 297, University of Bordeaux, 33 Bordeaux, France 4 Co-first author *Correspondence: valerie.crepel@inserm.fr This is an open access article under the CC BY-NC-ND license ( SUMMARY Kainate is a potent neurotoxin known to induce acute seizures. However, whether kainate receptors (KARs) play any role in the pathophysiology of temporal lobe epilepsy (TLE) is not known. In TLE, recurrent mossy fiber (rmf) axons form abnormal excitatory synapses onto other dentate granule cells that operate via KARs. The present study explores the pathophysiological implications of KARs in generating recurrent seizures in chronic epilepsy. In an in vitro model of TLE, seizure-like activity was minimized in mice lacking the GluK2 subunit, which is a main component of aberrant synaptic KARs at rmf synapses. In vivo, the frequency of interictal spikes and ictal discharges was strongly reduced in GluK2 / mice or in the presence of a GluK2/GluK receptor antagonist. Our data show that aberrant GluK2-containing KARs play a major role in the chronic seizures that characterize TLE and thus constitute a promising antiepileptic target. INTRODUCTION Kainate (KA) has long been known to induce behavioral and electrophysiological acute seizures reminiscent of those found in patients with temporal lobe epilepsy (TLE) (Ben-Ari and Cossart, 2; Vincent and Mulle, 29; Lerma and Marques, 213). However, whether kainate receptors (KARs) activated by the endogenous agonist glutamate play any role in the etiology of TLE is yet unknown (Vincent and Mulle, 29; Lerma and Marques, 213). Indeed, previous studies addressing the role of KARs have been restricted to pharmacologically induced acute epileptiform activities in naive animals (Mulle et al., 1998; Vissel et al., 21; Smolders et al., 22; Khalilov et al., 22). Therefore, the role of KARs in chronic and recurrent seizures, which is a cardinal pathological feature of TLE, remains to be established. In animal models of chronic epilepsy, as in human TLE, the hippocampus displays major network reorganization (Coulter et al., 22; Noebels et al., 21). In particular, sprouting of hippocampal mossy fibers (Represa et al., 1989a; Sutula et al., 1989; Isokawa et al., 1993; Franck et al., 199; Okazaki et al., 199) leads to the formation of powerful recurrent excitatory circuits between dentate granule cells (DGCs), which accounts for, in part, the enhanced ability of the hippocampus to generate epileptiform activity in human patients and animal models of TLE (Tauck and Nadler, 198; Patrylo and Dudek, 1998; Lynch and Sutula, 2; Buckmaster et al., 22; Scharfman et al., 23; Gabriel et al., 24). At the aberrant recurrent excitatory synapses in DGCs, mossy fiber inputs impinging on DGCs operate mostly via ectopic KARs and drive synaptic events with abnormal long-lasting kinetics not present in naive conditions (Epsztein et al., 2, 21; Artinian et al., 211). In keeping with this, an increased density of kainate binding sites was reported in the dentate gyrus (DG) of epileptic patients (Represa et al., 1989b). The present study explores the pathophysiological implications of KARs in the generation of chronic and recurrent seizures in TLE through the use of KARsubunit-deficient mice and selected pharmacological agents. In an animal model of TLE, we observe a strong reduction of both interictal and ictal activities in the DG recorded in vitro and in vivo in mice lacking the GluK2 subunit (GluK2 / ), or with the use of a pharmacological agent inhibiting GluK2/ GluK receptors. Therefore, we demonstrate that aberrant GluK2-containing KARs at recurrent mossy fiber (rmf) synapses play a major role in chronic seizures in TLE. RESULTS Removal of KARs Reduces Epileptiform Activity in the DG in Organotypic Slices In an attempt to elucidate the role of KARs in spontaneous epileptiform discharges in the DG, we used organotypic hippocampal slice cultures treated with pilocarpine that display mossy fiber sprouting and spontaneous epileptiform activity (Figures 1A 1C and S1A) (Zimmer and Gähwiler, 1984; McBain et al., 1989; Gutiérrez and Heinemann, 1999; Thomas et al., 2; Dyhrfjeld-Johnsen et al., 21; Grabs et al., 1994; Albus et al., 213). We observed a similar sprouting of rmfs in wild-type Cell Reports 8, , July 24, 214 ª214 The Authors 347

2 A C E R B-Power ( μv 2 /Hz ) CA3 Highpassfiltered (>1Hz) D H G CA3 PD RBs RBs Time (s) (Hz) Frequency UBP31 H Maximum G R B-Power ( μv 2 /Hz ) B Events/min + UBP31 + UBP Time (s) SYM Frequency (Hz) * * * *.2. Time (s) * * * * * * * + UBP31 + UBP31 + SYM281 2 μv + SYM281 Figure 1. Reduced Epileptiform Activity in the Dentate Gyrus of Organotypic Slices from GluK2 / Mice and in Presence of KAR Antagonists (A) Synaptoporin (purple) shows similar mossy fiber sprouting in and GluK2 / slices. Granule cell layer and pyramidal cell layer of the CA3 (Nissl, in green) is indicated by the dotted line. Inset highlights mossy fiber sprouting (arrow); scale bar, mm H, hilus; G, granule cell layer; scale bar, 1 mm. (B) Number of epileptiform activity recorded in and GluK2 / slices. ***p <.1 (n = 32 and n = 2 slices, respectively) and in the presence of mm UBP31 (*p <., n = 32 and n = 11) or 1 mm SYM281 (*p <., n = 32 and n = 1). (C) Extracellular field recordings of epileptiform activity (ACSF containing mm K + and mm gabazine) in DG from (top) and after band-pass filtering (bottom) before (left) and after application of mm UBP31 (middle) and in GluK2 / slices (right). (D) Time-frequency spectrograms using wavelet analysis of recurrent bursts (RBs) in before and after application of UBP31 and in GluK2 / slices from high-pass-filtered traces; field power is coded in color with red corresponding to the highest power. (E) Power spectrum analysis (left) of filtered RBs and maximum power in each group (right). ***p <.1 (n = 32 in and n = 2 in GluK2 / slices), **p <.1 (n = 11 with UBP31) and *p <. (n = 9 with SYM281). All statistical tests were performed by Mann-Whitney U test. In this and following figures: error bars = SEM. See also Figures S1 S3..1 s max () (n = 32) and GluK2 / (n = 27) slices following incubation with pilocarpine as revealed by synaptoporin (SPO) staining (Grabs et al., 1994) (see Supplemental Experimental Procedures) (optical density [OD]: 26.1 ± 2.9 in and 2. ± 3. in GluK2 /, p >., Figure 1A). We confirmed that DGCs (Figure S1B) in organotypic slices from mice displayed evoked and miniature slow excitatory postsynaptic currents (EPSCs) mediated by KARs (EPSC KA ) (see Experimental Procedures; Figures S1C S1E), as observed in chronically epileptic rodents but not in naive animals (Epsztein et al., 2, 21; Artinian et al., 211). EPSC KA were absent in GluK2 / slices highlighting the role of GluK2 in aberrant synaptic transmission; the remaining fast EPSCs were mediated by AMPARs (Figures S1D and S1E). EPSC KA originated from rmf synapses because they were strongly inhibited by the group II mglur agonist DCGIV (Feng et al., 23; Epsztein et al., 2) (Figures S1F and S1G). Reliable stereotyped spontaneous interictal-like activity (Sabolek et al., 212) was recorded in cultured slices using extracellular field recordings in mm K + -containing ACSF and mm gabazine (Figures 1B and 1C). The interical-like activity consisted of two distinct phases, the paroxysmal discharge (PD) followed by a late phase containing recurrent bursts (RBs) (McBain et al., 1989; Gutiérrez and Heinemann, 1999; Thomas et al., 2; Dyhrfjeld-Johnsen et al., 21) (Figure 1C). The number of interictal-like events was markedly reduced in cultured slices from GluK2 / mice in comparison with (events/min: 3.2 ±.4 in and 1.7 ±.3 in GluK2 / )(Figure 1B). The interictallike activity nested pathological high-frequency oscillations with a peak frequency at around 2 Hz (Figures 1D and 1E) in keeping with previous observations (Bragin et al., 24; Dyhrfjeld-Johnsen et al., 21). Interictal-like activity was visualized by high-pass filtering (Figure 1C) and by time-frequency spectrograms (Figure 1D). Analysis revealed a marked reduction of power during RBs in GluK2 / mice (Figures 1D and 1E) without a significant effect on the PD phase (Figures S2A and S2B). In slices from mice, the number of interictal-like events was significantly reduced in the presence of SYM281 a broad-spectrum KAR inhibitor (Epsztein et al., 2, 21; Artinian et al., 211) (events/min: 1.69 ±.3 in SYM281) or UBP31, an antagonist of postsynaptic GluK2/GluK receptors at rmf synapses (Pinheiro et al., 213) (1.47 ±.4 in UBP31) (Figures 1B and 1C). SYM281 or UBP31 markedly decreased the power of RBs (Figures 1D and 1E) without a significant effect on the PD phase (Figures S2A and S2B). A similar reduction of interictal-like activity was observed in the presence of UBP31 in slices in which cuts were made to isolate the DG (Figures S2C S2F) indicating that epileptiform activity originated from and propagated within the DG itself. In support of a role of GluK2, (1) there was no effect on interictal-like activity in the presence of UBP31 in GluK2 / mice (p >., n = 7); (2) the number of interictal-like activity was similar in slices from GluK1 / and mice (p >., n = 6); (3) UBP31 was similarly effective in reducing the number of interictal-like events in and in GluK1 / mice (p >., n = 6). Moreover, application of cyclothiazide (CTZ) to slow the decay kinetics of EPSC AMPA, hence mimicking the kinetics of aberrant EPSC KA (Figure S3A), restored the number of events and power of interictal-like activity in DGCs of GluK2 / mice (Figures S3B S3D). Therefore, slow 348 Cell Reports 8, , July 24, 214 ª214 The Authors

3 A -/ - G luk2 C E AD B B 6 * * * +GYKI36 pa 1 1 ms (pa) EPSC amplitude iml g h iml g h 4 2 Gabazine + APV + GYKI 36 +UBP31 EPSCs mediated by KARs containing GluK2 are key players in interictal-like activity in this in vitro model of TLE. Aberrant GluK2-Containing KARs Are Involved in rmf Network-Driven Bursts in DG in Slices from Mice with TLE To test the role of KARs in an animal model of TLE, experiments were performed in slices from pilocarpine-treated or +UBP31 7 μv 1 pa D 1 ms F 1 ms Coastline Index AS CA3 (pa) EPSC amplitude 2, 2, 1, 1, DG + GYKI 36 * * * * * * * + UBP31 + UBP31 LFP Figure 2. rmf Network-Driven Bursts Are Reduced in the Dentate Gyrus of GluK2 / Mice and by UBP31 in Mice (A) Synaptoporin (purple) shows similar mossy fiber sprouting in pilocarpinetreated and GluK2 / mice; granule cell (g) layer labeled with Nissl staining (green); h, hilus; iml, inner molecular layer. Scale bar, 2 mm. Higher magnification (right) highlights mossy fiber sprouting; scale bar, mm. (B and C) On the left, averaged EPSC (n = 1) evoked by stimulating in the inner molecular layer of the dentate gyrus in the presence of mm gabazine and 4 mm D-APV. In slices (B), note that GYKI36-resistant EPSC (black, 3mM) is abolished by mm UBP31 (gray). (C) In GluK2 / slices, note that EPSC (black) is abolished by 3 mm GYKI36. On the right, EPSC amplitude before and after UBP31 in slices (B,**p <.1 by Wilcoxon test, n = 9) or before and after GYKI36 in GluK2 / slices (C,***p <.1 by Wilcoxon test, n = 11). (D) Illustration of a hippocampal slice depicting the experimental design used to evoke rmf network-driven bursts by antidromic stimulation (AS) of mossy fibers. (E) Local field potential (LFP) evoked in the presence of 6. mm K + and mm gabazine showing an antidromic population spike (AD) followed by a burst (B) in slices before and after mm UBP31 or in GluK2 / slices. (F) Mean coastline burst index in slices before and after UBP31 or in GluK2 / slices. **p <.1, ***p <.1 by Mann-Whitney U test (n = 11 slices per group). GluK2 / mice several months after the inaugurating status epilepticus (see Experimental Procedures). Under these conditions, similar sprouting of rmfs was observed in both (n = 33) and GluK2 / (n = 2) mice (OD: 1.9 ± 1.1 in and 2.9 ± 3.9 in GluK2 /, p >., Figure 2A) but not in naive conditions (data not shown) as revealed by SPO staining. Hence, the lack of GluK2 does not affect the establishment of rmf connections in the pilocarpine model of TLE. By stimulating in the inner third of the molecular layer and thereby targeting rmfs, EPSC KA were recorded in DGCs from but not GluK2 / slices (Figures 2B and 2C). EPSC KA originated from rmf synapses because they were strongly inhibited by DCGIV (by 84%, p <., n = 7, data not shown). Epileptiform activities were then evoked by antidromic stimulation of the mossy fiber pathway in the CA3b area while recording local field potentials in the DGC layer (Patrylo and Dudek, 1998; Epsztein et al., 2) (Figures 2D and 2E). We observed that rmf- network-driven bursts were strongly reduced (see Experimental Procedures) in mice in the presence of UBP31, and in GluK2 / mice (Figures 2E and 2F). Thus, KARs containing GluK2 play a major role in burst activity driven by rmfs in the DG. Interictal Spikes and Ictal Discharges Are Reduced in the Absence of GluK2-Containing KARs In Vivo in Mice with TLE Interictal spikes (ISs) represent one of the best documented biomarkers of epilepsy in patient and animal models of TLE (Sabolek et al., 212). To test whether KARs are involved in ISs, local field potential recordings were performed in DG in vivo on anesthetized naive and pilocarpine-treated mice. As previously reported (Curia et al., 28), epileptic mice displayed numerous ISs (137.2 ± 27.9 ISs/min, Figures 3A and 3B) several months after inaugurating status epilepticus, but this type of activity was never observed in naive animals (n = ). In GluK2 / mice, the number of ISs was strongly reduced by 84% (21.6 ± 7. ISs/min, Figures 3A and 3B) without a significant change in IS amplitude (p >., n = 8, data not shown). Intraperitoneal injection of UBP31 (6 mg/kg) strongly reduced the number of ISs (Figures 3C and 3D) in mice, but not in GluK2 / mice (n = 3, data not shown). Interestingly, we did not observe any significant effect of UBP31 on the typical physiological slow oscillations (1. 2 Hz, Figures S4A S4C) driven by glutamatergic cortical synaptic inputs (Hahn et al., 27). No significant effect on IS number was observed with a lower dose of UBP31 (6 mg/kg, p >., n = 6, data not shown) or with injection of vehicle (p >., n = 13, data not shown). Ictal discharges, due to abnormal and long-lasting (tens of seconds) hypersynchronous neuronal activity, constitute the cardinal pathophysiological electroencephalographic (EEG) activity in TLE. These spontaneous and recurrent ictal events originate from limbic structures including the hippocampus and lead to profound disabling clinical manifestations. Intrahippocampal EEG recordings were performed in the DG of freely moving and GluK2 / pilocarpine-treated mice several months after inaugurating status epilepticus. EEG was continuously monitored for several days using a telemetric system (Figure 4A; see Experimental Procedures). mice displayed numerous ictal discharges (9.6 ± 2.3 per day) lasting Cell Reports 8, , July 24, 214 ª214 The Authors 349

4 A. mv C. mv 2 s 2 s 36.9 ± 1.8 s (n = ) and including tonic and clonic phases (Figure 4B). These ictal events corresponded to the well-described generalized tonic-clonic seizures with motor convulsions and loss of postural control (data not shown). We observed a strong reduction of the number of ictal events in GluK2 / mice (by 72%) (Figures 4C and 4E) without a significant change in their duration (p >., Figures S4D and S4E). Moreover, intraperitoneal injections of UBP31 (6 mg/kg, see Experimental Procedures) in mice for 3 days significantly reduced the number of ictal discharges (Figure 4F) without a significant change in their duration (p >., n = ). No effect was observed with injection of vehicle (n = 3, data not shown). Taken together, these experiments demonstrate that aberrant GluK2-containing KARs play a major role in the generation of both ISs and ictal discharges in vivo in TLE. DISCUSSION +UBP31 The well-characterized convulsive effects of exogenous kainate application are often confused with the physiological functions of KARs activated by the endogenous ligand glutamate (Vincent and Mulle, 29; Lerma and Marques, 213). It is well-established that KARs are involved in major physiological functions in the CNS (Contractor et al., 211), but paradoxically whether. mv. mv 1 ms 1 ms B D # of ISs/min # of ISs/min * * + UBP31 Figure 3. The Number of Interictal Spikes Is Decreased in GluK2 / Mice and by UBP31 in Mice (A) Local field potential (LFP) recordings performed in vivo in dentate gyrus of pilocarpine-treated (top) and GluK2 / mice (bottom); the dashed line indicates the threshold for Interictal Spikes (IS) detection. Right traces show magnified interictal spikes (ISs). (B) Number of ISs/min in (n = 14) and GluK2 / mice (n = 8). *p <. by unpaired t test. (C) LFP recordings performed in dentate gyrus of pilocarpine-treated mice in the absence and presence of UBP31 (6 mg/kg injected intraperitoneally). Right traces show magnified ISs. (D) Number of ISs/min before (n = 6) and after 6 mg/kg UBP31 (n = 6). *p <. by Wilcoxon test. See also Figure S4. A Radiofrequency transmitter Receiver Telemetric hippocampal EEG recording 24h per day C D # Animal 77.8% Total number of seizures: 248 B Ictal activity Tonic phase Seizure/day 1 Clonic phase Hours E F 1 * % KAR activated by the endogenous agonist glutamate play any role in the etiology of chronic epilepsy remains to be established (Vincent and Mulle, 29; Lerma and Marques, 213). Previous studies addressing the role of KARs have been restricted to pharmacologically induced acute epileptiform activities in naive animals (Mulle et al., 1998; Vissel et al., 21; Smolders et al., 22; Khalilov et al., 22). In TLE, DGCs operate via aberrant KARs at rmf synapses not present in naive conditions (Epsztein et al., 2, 21; Artinian et al., 211). In keeping with this, an increased density of kainate binding sites was previously reported in DG of epileptic patients (Represa et al., 1989b). We thus explored the pathophysiological implications of KARs in chronic and spontaneous seizures in DG of mice, and we now show that KARs play a major role in chronic and spontaneous interictal and ictal discharges in TLE. Seizure/day 1 1 * * mv mv + UBP31 Figure 4. The Number of Ictal Discharges Is Decreased in GluK2 / Mice and by UBP31 in Mice (A) Illustration of the experimental design used for telemetric EEG recordings (4 days, 24 hr per day). (B) Example of a typical ictal event recorded in dentate gyrus of GluK2 / mice. Note the tonic and clonic phases. (C) Distribution of seizures over time (12 hr periods) for each and GluK2 / mouse. (D) Pie chart showing the fraction of ictal events recorded in (n = ) and GluK2 / (n = ) mice (193 and ictal events in and GluK2 / mice, respectively). (E) Number of ictal events in (n = ) and GluK2 / mice (n = ) per day. *p <. by Mann-Whitney U test. (F) Number of ictal events before and during treatment with 6 mg/kg UBP31 (n = ). **p <.1 by paired t test. See also Figure S4. s 2 s 3 Cell Reports 8, , July 24, 214 ª214 The Authors

5 In the present study, several lines of evidence support the role of KARs containing GluK2 in chronic epileptic activities. (1) Spontaneous and recurrent interictal-like activity measured in vitro was less frequent during application of SYM281, or UBP31, a GluK2/GluK antagonist (Pinheiro et al., 213) as well as in GluK2 / but not in GluK1 / mice. The power of epileptiform activity was reduced to the same degree by the antagonists and in slices from GluK2 / mice indicating the specific importance of GluK2; similar data were obtained in slices with isolated DG confirming the key role of GluK2-containing KARs in this area. The pharmacological conversion of EPSC AMPA kinetics hence mimicking the slow kinetics of aberrant EPSC KA restored interictal-like activity in GluK2 / slices. This implies that the slow kinetics of EPSC KA is a key element in the generation of epileptiform activities presumably because this feature eases temporal summation of EPSPs and neuronal firing (Artinian et al., 211). (2) Network-driven bursts evoked by antidromic mossy fiber stimulation and recorded in the DG of slices from chronic pilocarpine-treated mice were strongly reduced in the absence of KARs. This result is a clear indication that KARs are involved in the rmf network-driven activity in TLE. (3) In vivo recordings revealed a drastic reduction of ISs and ictal discharges in GluK2 / compared with mice. Accordingly, UBP31, which blocks synaptic KARs at recurrent synapses in the DG (Pinheiro et al., 213) (this study), acts as an antiepileptic agent reducing the frequency of ISs and ictal discharges in mice. At mossy fiber CA3 synapses, postsynaptic KARs that mediate a small component of the EPSC with slow kinetics (Castillo et al., 1997; Frerking et al., 1998; Bureau et al., 2; Ben-Ari and Cossart, 2; Epsztein et al., 2) in part due to the accessory Neto protein (Copits et al., 211; Zhang et al., 29; Copits and Swanson, 212) and comprise GluK2 (Mulle et al., 1998), GluK (Contractor et al., 21; Ruiz et al., 2), and GluK4 (Fernandes et al., 29). At present, the subunit composition of KARs at rmf synapses remains unknown. Nevertheless, we propose that the reduced seizure activity in the presence of UBP31 or in GluK2 / mice is likely due to an effect on postsynaptic GluK2/GluK receptors because (1) both GluK2 and GluK subunits are expressed in the DG but not GluK1 (Bahn et al., 1994; Bureau et al., 1999); (2) the presence of GluK critically depends on that of GluK2 (Christensen et al., 24; Ruiz et al., 2), providing a reasonable explanation why EPSC KA are not observed in DGCs of GluK2 / mice; (3) UBP31 was originally developed as a GluK1 and GluK3 antagonist (Dolman et al., 27; Perrais et al., 29), but the observed effect of UBP31 on epileptic activity is likely not mediated by these subunits (Dolman et al., 27; Perrais et al., 29), because this compound is ineffective in GluK2 / mice and fully effective in GluK1 / mice; (4) UBP31 is effective on postsynaptic GluK2/GluK receptors, but not on homomeric GluK2 receptors (Pinheiro et al., 213), blocking EPSC KA in epileptic mice and rats (Pinheiro et al., 213); and () presynaptic KARs that modulate glutamate release at both mossy fiber CA3 (Schmitz et al., 21; Contractor et al., 21; Lauri et al., 21; Sachidhanandam et al., 29) and at rmf-dgc synapses (Feng et al., 23) are probably not involved in the reduction of seizures by UBP31, because this compound is not effective on presynaptic receptors. Synaptic activation of KARs can also decrease Ca 2+ -activated K + currents (I sahp ) via a metabotropic signaling pathway, leading to an increase of excitability (Melyan et al., 22; Fisahn et al., 2; Ruiz et al., 2). In our experimental conditions, we could hypothesize that antagonism of KARs by UBP31 could reduce excitability by acting on I sahp. This is unlikely because UBP31 inhibits the ionotropic function of KARs at mossy fiber CA3 synapses without affecting their metabotropic action (Pinheiro et al., 213). In human and animal models of epilepsy, interictal activities as well as pathological high-frequency oscillations are a hallmark of epilepsy and are used as biomarkers to diagnose the pathology (Sabolek et al., 212; Jefferys et al., 212). Moreover, ictal discharges constitute the cardinal pathophysiological EEG manifestation of TLE leading to profound disabling clinical symptoms. Besides strategy to prevent development of epilepsy (Liu et al., 213), blockade of chronic and recurrent seizures is therefore one of the main challenges in the treatment of TLE, which often displays pharmaco-resistant features. Over the past 3 years, several new antiepileptic drugs have been developed. These clinically approved compounds display a spectrum of mechanisms of action, with effects on both inhibitory and excitatory signaling (Löscher and Schmidt, 212). Perampanel, a noncompetitive selective AMPA receptor antagonist, has been recently shown to decrease seizure frequency in patients with refractory focal epilepsy (Löscher and Schmidt, 212; Rogawski and Hanada, 213). This therapeutic compound, which is now clinically available, is administered at low concentration in order to avoid side effects and neurological impairment because it targets the predominant receptors responsible for excitatory neurotransmission in the brain (Löscher and Schmidt, 212; Rogawski and Hanada, 213). Our present work discloses an antiepileptic strategy that specifically targets different receptors (KARs) and fully spares AMPA receptor-mediated excitatory neurotransmission in TLE in the chronic phase of epilepsy. GluK2/GluK KARs constitute a more restricted target, present at only a limited number of synaptic sites, compared with AMPARs that are present at all excitatory synapses. In addition to the different type of target, our results provide a clear rationale for a specific pathophysiological mechanism (KARs at aberrant recurrent mossy fiber dentate granule cell synapses) in contrast to the broad action of perampanel on excitatory glutamatergic transmission. All in all, we show that aberrant GluK2-containing kainate receptors contribute to chronic seizures in TLE, urging for the development of an antiepileptic strategy targeting these KARs. EXPERIMENTAL PROCEDURES All experiments were approved by the Institut National de la Santé et de la Recherche Médicale (INSERM) animal care and use agreement (B ) and the European community council directive (21/63/UE). Mice (FVB/N background) had access to food and water ad libitum and were housed under a 12-hr-light/dark cycle at 22 C 24 C. For detailed methods, see Supplemental Experimental Procedures. In Vivo and In Vitro Models of Temporal Lobe Epilepsy Mouse: Male GluK2 / and mice (P6-8) were administered pilocarpine subcutaneously (2-6 mg/kg) using a ramp protocol. Slice culture: P9-1 hippocampal/entorhinal cortex slices were prepared from GluK2 /, GluK1 / Cell Reports 8, , July 24, 214 ª214 The Authors 31

6 and mice. Pilocarpine (. mm) was added to the medium at days in vitro (DIV) and removed at 7 DIV; slices were utilized for experiments from 9 DIV to 11 DIV. Electrophysiological Recordings and Analysis In Vitro Transverse hippocampal slices were prepared from extracted hippocampi of pilocarpine-treated GluK2 / and mice, at 9 months of age. Cultured or acute slices were individually transferred to a recording chamber maintained at 3 32 C and continuously perfused (2 ml/min) with oxygenated normal or adapted ACSF containing mm gabazine. Local field potential recordings in cultured and acute slices were made in the granule cell layer of the DG. Evoked synaptic transmission was achieved utilizing a bipolar NiCh electrode. In organotypic slices, whole-cell voltage-clamp recordings of spontaneous EPSCs were recorded at 7 mv. To characterize the nature and kinetic properties of fast and slow events, miniature EPSCs (mepscs) evoked in Sr 2+ conditions were recorded in DGCs from pilocarpine-treated GluK2 / and slices. Electrical stimulation was made in the molecular layer in the continued presence of mm gabazine and 4 mm D-APV. cultured slices displayed slowly decaying spontaneous and mepscs mediated by KARs (Epsztein et al., 2)(Figures S1C S1E). KAR-mediated mepscs were insensitive to the AMPAR antagonist GYKI36 (1 mm) (Figure S1D) and abolished in the presence of the broad-spectrum KAR antagonist SYM281 (1 mm) (Epsztein et al., 2, 21; Artinian et al., 211) or mm UBP31, a potent antagonist of KARs at mossy fiber CA3 synapses (Pinheiro et al., 213). These events were absent in GluK2 / slices (Figure S1D); the remaining fast EPSCs were fully blocked by AMPAR antagonist GYKI36 (1 mm) in GluK2 / slices (data not shown). Accordingly, mepscs recorded in slices in the presence of mm UBP31 were fully blocked by 1 mm GYKI36 (data not shown). Therefore, in organotypic slices 1 mm GYKI36 is fully effective in abolishing AMPAR-mediated EPSCs. The frequency of ongoing excitatory postsynaptic currents (EPSCs) was similar in DGCs from and GluK2 / slices in organotypic slice (p >., n =, data not shown) and in acute slices (p >., n = 7, data not shown). In cultured slices, visual representation of time-frequency spectrogram of local field potentials was made using Autosignal software (Seasolve 1.7). In acute hippocampal slices, intensity of evoked network-driven bursts was quantified using coastline index. Electrophysiological Recordings and Analysis In Vivo Naive wild-type () and pilocarpine-treated or GluK2 / mice ( 6 months old) were used for local field potentials or electroencephalographic (EEG) recordings in DG. UBP31 or vehicle (saline/1% DMSO) was subsequently injected intraperitoneally. Morphological Analysis See Supplemental Experimental Procedures for recurrent mossy fiber staining and quantification and for biocytin and Prox1 revelation of DGCs. Statistics Analyses All values are given as means ± SEM. Statistical analyses were performed using GraphPad Prism (GraphPad software.1). SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures and four figures and can be found with this article online at /j.celrep AUTHOR CONTRIBUTIONS V.C. and C.M. conceptualized the study. A.P., L.A.C., and V.C. designed and performed experiments. Some in vivo recordings in anaesthetized animals were also achieved by D.W.O. under the supervision of J.E. Some mossy fiber staining and experiments in organotypic slices were done by A.G. V.C., A.P., L.A.C., and C.M. discussed and wrote the manuscript. The co-first authors A.P. and L.A.C designed, performed, and analyzed experiments and wrote the manuscript. V.C. supervised the project. ACKNOWLEDGMENTS This work was supported by Institut National de la Santé et de la Recherche Médicale (INSERM), Centre National de la Recherche Scientifique (CNRS), Conseil Régional PACA, Université Aix-Marseille, the Agence Nationale de la Recherche (ANR) (KAREP, ANR-21-BLAN to C.M. and V.C. and ANR-9-BLAN-29-1 to V.C.), the Ministère de l Enseignement Supérieur et de la Recherche (MESR to A.P. and D.W.O.), and the Ligue Française contre l Epilepsie (LFCE to A.P.). We thank Dr. S. Feldt-Muldoon, T. Tressard, F. Michel, and S. Varpula for their technical help, and Drs. A. Represa and J. Artinian for critical reading of the manuscript. Received: December, 213 Revised: May 8, 214 Accepted: June 19, 214 Published: July 17, 214 REFERENCES Albus, K., Heinemann, U., and Kovács, R. (213). Network activity in hippocampal slice cultures revealed by long-term in vitro recordings. J. Neurosci. Methods 217, 1 8. Artinian, J., Peret, A., Marti, G., Epsztein, J., and Crépel, V. (211). Synaptic kainate receptors in interplay with INaP shift the sparse firing of dentate granule cells to a sustained rhythmic mode in temporal lobe epilepsy. J. Neurosci. 31, Bahn, S., Volk, B., and Wisden, W. (1994). Kainate receptor gene expression in the developing rat brain. J. Neurosci. 14, Ben-Ari, Y., and Cossart, R. (2). Kainate, a double agent that generates seizures: two decades of progress. Trends Neurosci. 23, Bragin, A., Wilson, C.L., Almajano, J., Mody, I., and Engel, J., Jr. (24). Highfrequency oscillations after status epilepticus: epileptogenesis and seizure genesis. Epilepsia 4, Buckmaster, P.S., Zhang, G.F., and Yamawaki, R. (22). Axon sprouting in a model of temporal lobe epilepsy creates a predominantly excitatory feedback circuit. J. Neurosci. 22, Bureau, I., Bischoff, S., Heinemann, S.F., and Mulle, C. (1999). Kainate receptor-mediated responses in the CA1 field of wild-type and GluR6-deficient mice. J. Neurosci. 19, Bureau, I., Dieudonne, S., Coussen, F., and Mulle, C. (2). Kainate receptormediated synaptic currents in cerebellar Golgi cells are not shaped by diffusion of glutamate. Proc. Natl. Acad. Sci. USA 97, Castillo, P.E., Malenka, R.C., and Nicoll, R.A. (1997). Kainate receptors mediate a slow postsynaptic current in hippocampal CA3 neurons. Nature 388, Christensen, J.K., Paternain, A.V., Selak, S., Ahring, P.K., and Lerma, J. (24). A mosaic of functional kainate receptors in hippocampal interneurons. J. Neurosci. 24, Contractor, A., Swanson, G., and Heinemann, S.F. (21). Kainate receptors are involved in short- and long-term plasticity at mossy fiber synapses in the hippocampus. Neuron 29, Contractor, A., Mulle, C., and Swanson, G.T. (211). Kainate receptors coming of age: milestones of two decades of research. Trends Neurosci. 34, Copits, B.A., and Swanson, G.T. (212). Dancing partners at the synapse: auxiliary subunits that shape kainate receptor function. Nat. Rev. Neurosci. 13, Copits, B.A., Robbins, J.S., Frausto, S., and Swanson, G.T. (211). Synaptic targeting and functional modulation of GluK1 kainate receptors by the auxiliary neuropilin and tolloid-like (NETO) proteins. J. Neurosci. 31, Coulter, D.A., McIntyre, D.C., and Löscher, W. (22). Animal models of limbic epilepsies: what can they tell us? Brain Pathol. 12, Curia, G., Longo, D., Biagini, G., Jones, R.S., and Avoli, M. (28). The pilocarpine model of temporal lobe epilepsy. J. Neurosci. Methods 172, Cell Reports 8, , July 24, 214 ª214 The Authors

7 Dolman, N.P., More, J.C., Alt, A., Knauss, J.L., Pentikäinen, O.T., Glasser, C.R., Bleakman, D., Mayer, M.L., Collingridge, G.L., and Jane, D.E. (27). Synthesis and pharmacological characterization of N3-substituted willardiine derivatives: role of the substituent at the -position of the uracil ring in the development of highly potent and selective GLUK kainate receptor antagonists. J. Med. Chem., Dyhrfjeld-Johnsen, J., Berdichevsky, Y., Swiercz, W., Sabolek, H., and Staley, K.J. (21). Interictal spikes precede ictal discharges in an organotypic hippocampal slice culture model of epileptogenesis. J. Clin. Neurophysiol. 27, Epsztein, J., Represa, A., Jorquera, I., Ben-Ari, Y., and Crépel, V. (2). Recurrent mossy fibers establish aberrant kainate receptor-operated synapses on granule cells from epileptic rats. J. Neurosci. 2, Epsztein, J., Sola, E., Represa, A., Ben-Ari, Y., and Crépel, V. (21). A selective interplay between aberrant EPSPKA and INaP reduces spike timing precision in dentate granule cells of epileptic rats. Cereb. Cortex 2, Feng, L., Molnár, P., and Nadler, J.V. (23). Short-term frequency-dependent plasticity at recurrent mossy fiber synapses of the epileptic brain. J. Neurosci. 23, Fernandes, H.B., Catches, J.S., Petralia, R.S., Copits, B.A., Xu, J., Russell, T.A., Swanson, G.T., and Contractor, A. (29). High-affinity kainate receptor subunits are necessary for ionotropic but not metabotropic signaling. Neuron 63, Fisahn, A., Heinemann, S.F., and McBain, C.J. (2). The kainate receptor subunit GluR6 mediates metabotropic regulation of the slow and medium AHP currents in mouse hippocampal neurones. J. Physiol. 62, Franck, J.E., Pokorny, J., Kunkel, D.D., and Schwartzkroin, P.A. (199). Physiologic and morphologic characteristics of granule cell circuitry in human epileptic hippocampus. Epilepsia 36, Frerking, M., Malenka, R.C., and Nicoll, R.A. (1998). Synaptic activation of kainate receptors on hippocampal interneurons. Nat. Neurosci. 1, Gabriel, S., Njunting, M., Pomper, J.K., Merschhemke, M., Sanabria, E.R., Eilers, A., Kivi, A., Zeller, M., Meencke, H.J., Cavalheiro, E.A., et al. (24). Stimulus and potassium-induced epileptiform activity in the human dentate gyrus from patients with and without hippocampal sclerosis. J. Neurosci. 24, Grabs, D., Bergmann, M., Schuster, T., Fox, P.A., Brich, M., and Gratz, M. (1994). Differential expression of synaptophysin and synaptoporin during pre- and postnatal development of the rat hippocampal network. Eur. J. Neurosci. 6, Gutiérrez, R., and Heinemann, U. (1999). Synaptic reorganization in explanted cultures of rat hippocampus. Brain Res. 81, Hahn, T.T., Sakmann, B., and Mehta, M.R. (27). Differential responses of hippocampal subfields to cortical up-down states. Proc. Natl. Acad. Sci. USA 14, Isokawa, M., Levesque, M.F., Babb, T.L., and Engel, J., Jr. (1993). Single mossy fiber axonal systems of human dentate granule cells studied in hippocampal slices from patients with temporal lobe epilepsy. J. Neurosci. 13, Jefferys, J.G., Menendez de la Prida, L., Wendling, F., Bragin, A., Avoli, M., Timofeev, I., and Lopes da Silva, F.H. (212). Mechanisms of physiological and epileptic HFO generation. Prog. Neurobiol. 98, Khalilov, I., Hirsch, J., Cossart, R., and Ben-Ari, Y. (22). Paradoxical antiepileptic effects of a GluR agonist of kainate receptors. J. Neurophysiol. 88, Lauri, S.E., Bortolotto, Z.A., Bleakman, D., Ornstein, P.L., Lodge, D., Isaac, J.T.R., and Collingridge, G.L. (21). A critical role of a facilitatory presynaptic kainate receptor in mossy fiber LTP. Neuron 32, Lerma, J., and Marques, J.M. (213). Kainate receptors in health and disease. Neuron 8, Liu, G., Gu, B., He, X.P., Joshi, R.B., Wackerle, H.D., Rodriguiz, R.M., Wetsel, W.C., and McNamara, J.O. (213). Transient inhibition of TrkB kinase after status epilepticus prevents development of temporal lobe epilepsy. Neuron 79, Löscher, W., and Schmidt, D. (212). Epilepsy: perampanel-new promise for refractory epilepsy? Nat. Rev. Neurol. 8, Lynch, M., and Sutula, T. (2). Recurrent excitatory connectivity in the dentate gyrus of kindled and kainic acid-treated rats. J. Neurophysiol. 83, McBain, C.J., Boden, P., and Hill, R.G. (1989). Rat hippocampal slices in vitro display spontaneous epileptiform activity following long-term organotypic culture. J. Neurosci. Methods 27, Melyan, Z., Wheal, H.V., and Lancaster, B. (22). Metabotropic-mediated kainate receptor regulation of IsAHP and excitability in pyramidal cells. Neuron 34, Mulle, C., Sailer, A., Pérez-Otaño, I., Dickinson-Anson, H., Castillo, P.E., Bureau, I., Maron, C., Gage, F.H., Mann, J.R., Bettler, B., and Heinemann, S.F. (1998). Altered synaptic physiology and reduced susceptibility to kainate-induced seizures in GluR6-deficient mice. Nature 392, Noebels, J.L., Avoli, M., Rogawski, M., Olsen, R., and Delgado-Escueta, A.V. (21). Jasper s Basic Mechanisms of the Epilepsies Workshop. Epilepsia 1 (Suppl ), 1. Okazaki, M.M., Evenson, D.A., and Nadler, J.V. (199). Hippocampal mossy fiber sprouting and synapse formation after status epilepticus in rats: visualization after retrograde transport of biocytin. J. Comp. Neurol. 32, Patrylo, P.R., and Dudek, F.E. (1998). Physiological unmasking of new glutamatergic pathways in the dentate gyrus of hippocampal slices from kainateinduced epileptic rats. J. Neurophysiol. 79, Perrais, D., Pinheiro, P.S., Jane, D.E., and Mulle, C. (29). Antagonism of recombinant and native GluK3-containing kainate receptors. Neuropharmacology 6, Pinheiro, P.S., Lanore, F., Veran, J., Artinian, J., Blanchet, C., Crépel, V., Perrais, D., and Mulle, C. (213). Selective block of postsynaptic kainate receptors reveals their function at hippocampal mossy fiber synapses. Cereb. Cortex 23, Represa, A., Le Gall La Salle, G., and Ben-Ari, Y. (1989a). Hippocampal plasticity in the kindling model of epilepsy in rats. Neurosci. Lett. 99, Represa, A., Robain, O., Tremblay, E., and Ben-Ari, Y. (1989b). Hippocampal plasticity in childhood epilepsy. Neurosci. Lett. 99, Rogawski, M.A., and Hanada, T. (213). Preclinical pharmacology of perampanel, a selective non-competitive AMPA receptor antagonist. Acta Neurol. Scand. Suppl., Ruiz, A., Sachidhanandam, S., Utvik, J.K., Coussen, F., and Mulle, C. (2). Distinct subunits in heteromeric kainate receptors mediate ionotropic and metabotropic function at hippocampal mossy fiber synapses. J. Neurosci. 2, Sabolek, H.R., Swiercz, W.B., Lillis, K.P., Cash, S.S., Huberfeld, G., Zhao, G., Ste Marie, L., Clemenceau, S., Barsh, G., Miles, R., and Staley, K.J. (212). A candidate mechanism underlying the variance of interictal spike propagation. J. Neurosci. 32, Sachidhanandam, S., Blanchet, C., Jeantet, Y., Cho, Y.H., and Mulle, C. (29). Kainate receptors act as conditional amplifiers of spike transmission at hippocampal mossy fiber synapses. J. Neurosci. 29, 8. Scharfman, H.E., Sollas, A.L., Berger, R.E., and Goodman, J.H. (23). Electrophysiological evidence of monosynaptic excitatory transmission between granule cells after seizure-induced mossy fiber sprouting. J. Neurophysiol. 9, Schmitz, D., Mellor, J., and Nicoll, R.A. (21). Presynaptic kainate receptor mediation of frequency facilitation at hippocampal mossy fiber synapses. Science 291, Smolders, I., Bortolotto, Z.A., Clarke, V.R., Warre, R., Khan, G.M., O Neill, M.J., Ornstein, P.L., Bleakman, D., Ogden, A., Weiss, B., et al. (22). Antagonists of GLU(K)-containing kainate receptors prevent pilocarpine-induced limbic seizures. Nat. Neurosci., Cell Reports 8, , July 24, 214 ª214 The Authors 33

8 Sutula, T., Cascino, G., Cavazos, J., Parada, I., and Ramirez, L. (1989). Mossy fiber synaptic reorganization in the epileptic human temporal lobe. Ann. Neurol. 26, Tauck, D.L., and Nadler, J.V. (198). Evidence of functional mossy fiber sprouting in hippocampal formation of kainic acid-treated rats. J. Neurosci., Thomas, A.M., Corona-Morales, A.A., Ferraguti, F., and Capogna, M. (2). Sprouting of mossy fibers and presynaptic inhibition by group II metabotropic glutamate receptors in pilocarpine-treated rat hippocampal slice cultures. Neuroscience 131, Vincent, P., and Mulle, C. (29). Kainate receptors in epilepsy and excitotoxicity. Neuroscience 18, Vissel, B., Royle, G.A., Christie, B.R., Schiffer, H.H., Ghetti, A., Tritto, T., Perez- Otano, I., Radcliffe, R.A., Seamans, J., Sejnowski, T., et al. (21). The role of RNA editing of kainate receptors in synaptic plasticity and seizures. Neuron 29, Zhang, W., St-Gelais, F., Grabner, C.P., Trinidad, J.C., Sumioka, A., Morimoto- Tomita, M., Kim, K.S., Straub, C., Burlingame, A.L., Howe, J.R., and Tomita, S. (29). A transmembrane accessory subunit that modulates kainate-type glutamate receptors. Neuron 61, Zimmer, J., and Gähwiler, B.H. (1984). Cellular and connective organization of slice cultures of the rat hippocampus and fascia dentata. J. Comp. Neurol. 228, Cell Reports 8, , July 24, 214 ª214 The Authors

9 SUPPLEMENTAL INFORMATION Contribution of Aberrant GluK2-containing Kainate Receptors to Chronic Seizures in Temporal Lobe Epilepsy Angélique Peret, Louisa A. Christie, David W. Ouedraogo, Adam Gorlewicz, Jérôme Epsztein, Christophe Mulle and Valérie Crépel Supplemental Information contains 4 Supplemental Figures and Supplemental Experimental Procedures.

10 A B C Prox1 Streptavidin merged Peret et al. Figure S1 CA3 MF - Spontaneous EPSCs rmf DG 2 pa ms D +GYKI36 +UBP31 pa 2 ms E EPSC decay (ms) GYKI36 * * * * * ** * * + UBP31 + SYM281 F G 2 EPSC KA 1 + DCGIV 4 pa 1 ms EPSC KA amplitude (pa) 1 * * DCGIV Figure S1 related to Figure 1. Aberrant kainate receptor-mediated EPSCs in dentate granule cells from organotypic slices (A) Example reconstruction of granule cell after patch-clamp recording labelled with biocytin. Nissl staining of granule cell layer and pyramidal cell layer of the CA3 (green). Note the recurrent mossy fiber (rmf, arrow) passing through the dentate granule cell layer. Dentate gyrus (DG); scale bar: 1 μm. (B) Immunostaining for Prox1 in cells that were recorded and filled with biocytin; scale bar: 1 μm. (C) DGC recordings of spontaneous EPSCs (Vh= -7 mv) in hippocampal slice culture from mice in the presence of μm gabazine and 4 μm D-APV. Traces include fast ( ) and slow events ( ). (D) DGC recordings of strontium evoked mepscs (Vh= -7 mv) in slices after application of 1 μm GYKI36 or μm UBP31, or in slices. (E) Decay times of mepscs in slices in the presence of 1 μm GYKI36 (n=6 cells) or 1 μm SYM281 (n= cells) or μm UBP31 (n= cells), or in slices (n=8 cells). Respectively ***P<.1; ***P<.1 by Mann-Whitney U test and ***P<.1 unpaired t-test. (F) Average EPSC KA (n=1) evoked in slices by electrical stimulation (arrow) of the inner molecular layer of the dentate gyrus in the presence of μm gabazine, 4 μm D-APV and 1 μm GYKI36 before (black) and after (grey) bath application of 1 μm DCGIV. Note the strong reduction of EPSC KA in the presence of DCGIV confirming their recurrent mossy fiber origin. (G) Amplitude of EPSC KA before and after DCGIV (n=7 cells). **P<.1, paired t-test.

11 Peret et al. Figure S2 A Frequency (Hz) C PD CA3.2. Time (s) DG + UBP31.2. Time (s) D Events/min Time (s) * UBP31 E max Maximum PD-Power (μv 2 /Hz) B PD-Power (μv 2 /Hz) UBP31 SYM Frequency (Hz) ns UBP31 F Maximum RB-Power (μv 2 /Hz) Maximum PD-Power (μv 2 /Hz) * ns ns ns UBP31 UBP31 SYM281 Figure S2 related to Figure 1. Analysis and origin of interictal-like activity in the DG with or without KARs (A) Time-frequency spectrograms using wavelet analysis of paroxysmal discharge (PD) in before and after application of UBP31 and in slices from highpass-filtered traces. (B) Power spectrum analysis (left) of filtered PD and maximum power in each group (right). P>. (n=32 in and n=24 in slices), P>. (n=11 with UBP31) and P>. (n=9 with SYM281). All statistical tests were performed by Mann- Whitney U test. Note that the PD phase is not changed in the absence of KAR (C) Illustration of cultured hippocampal slice depicting the location of cuts (dotted lines). (D) Number of epileptiform activity in slices recorded after cuts, before and after application of μm UBP31. *P<. by Wilcoxon test (n=6 slices); (E) maximum value of the power of paroxysmal discharge (PD) in isolated DG. P>. by Wilcoxon test (n=6 slices); (F) maximum value of the power of recurrent bursts (RBs) in isolated DG. *P<. by Wilcoxon test (n=6 slices). Note that interictal-like activity originates from and propagates within the DG.

12 Peret et al. Figure S3 A GluK2 -/- + CTZ B pa 2 ms Highpassfiltered (>1Hz) Cumulative propbability CTZ Decay time (ms) + CTZ 2 μv.1 s C D RB-Power (μv 2 /Hz) Events/min * * * * + CTZ +CTZ Frequency (Hz) Figure S3 related to Figure 1. EPSCs with slow kinetics are important determinants of epileptiform activity (A) Left, DGC recordings of averaged mepscs (n=3) in the presence of 4 μm D-APV and μm gabazine in cultured slices before and after application of 1 μm cyclothiazide (CTZ), an allosteric inhibitor of AMPAR desensitization; Right, cumulative probability plot of the decay time constant of mepscs before and after CTZ (n=6 cells per group). Note that in the presence of CTZ, the decay time distribution is significantly shifted toward slower decay times: median from 2.6 to 7.1 ms before and after CTZ, respectively. (B) Extracellular field recordings of epileptiform activity (ACSF containing mm K+ and μm gabazine) in dentate gyrus from slices (top) and after bandpass filtering (bottom) before (left) and after application of CTZ (right). (C) Number of epileptiform activity. ***P<.1 by Mann-Whitney U test (n=32 in, n=2 in and n=1 in CTZ). (D) Power spectrum analysis of filtered RBs. *P<. by Mann-Whitney U test.

13 (mv 2 /Hz) A B C D +UBP31.4 mv.4 s Power (mv 2 /Hz) UBP Frequency (Hz) mv 1 s 1 Maximum Power E Duration of seizure (s) ns UBP 31 Peret et al. Figure S4 Figure S4 related to Figure 3 and 4. UBP31 has no effect on basal slow oscillations and similar duration of ictal events in and mice (A) Example of typical slow oscillations recorded in the dentate gyrus before and after injection of 6 mg/kg of UBP31. (B) Power spectrum analysis (left) and maximum power (right) before and after application of UBP31. P>. by Wilcoxon test (n=6). (C) Peak of frequency. P>. by Wilcoxon test (n=6). (D) Examples of typical ictal events recorded in dentate gyrus in (top) and mice (bottom). (E) Duration of ictal event in (193 ictal events, n= mice) and in conditions ( ictal events, n= mice). P>. by Mann- Whitney U test. Peak of Frequency (Hz) ns ns UBP 31

14 SUPPLEMENTAL EXPERIMENTAL PROCEDURES Slice culture Hippocampal/entorhinal cortex slices (3 µm) were prepared from pre-genotyped, GluK1 -/- and mice (P9-1) using a McIlwain tissue chopper. Slices were placed on mesh inserts (Millipore) inside culture dishes containing 1 ml of the following medium: MEM %, HS 2%, HBSS 2%, HEPES 1 mm, glucose 6. mg/ml and insulin.1 mg/ml. Culture medium was changed every 2-3 days and slices maintained in an incubator at 37 C/% CO 2. Mouse model of temporal lobe epilepsy Male and mice (P6-8) were given scopolamine (1 mg/kg) subcutaneously (s.c.) 1 minutes prior to s.c. administration of pilocarpine (2-6 mg/kg). A ramp protocol was used whereby animals were given an initial dose of 2 mg/kg followed by half-doses every 3 minutes until seizures appeared. Both and mice typically experienced at least two seizures prior to entering status epilepticus (SE). Diazepam (1 mg/kg) was administered 4 minutes after onset of SE. Acute slice preparation Hippocampal slices were prepared from chronic epileptic and control and mice, at -9 months of age. Animals were deeply anesthetized with xylazine 13 mg/kg / ketamine 66 mg/kg and transcardially perfused with a modified artificial cerebrospinal fluid (macsf) containing the following (in mm): 132 choline, 2. KCl, 1.2 NaH 2 PO 4, 2 NaHCO 3, 7 MgCl 2,. CaCl 2, and 8 D-glucose prior to decapitation. The brain was then removed rapidly, the hippocampi were dissected, and transverse 3 µm thick slices were cut using a Leica VT12S vibratome in ice-cold oxygenated (9% O 2 and % CO 2 ) macsf. Slices were transferred to rest at ~3 C for 3 min and then room temperature for a further 3 min, in