Cerebral Cortex Advance Access published June 7, 2010 Cerebral Cortex doi:10.1093/cercor/bhq100

Increased Dentate Gyrus Excitability in Neuroligin-2-Deficient Mice in Vivo Peter Jedlicka1, Mrinalini Hoon2, Theofilos Papadopoulos3, Andreas Vlachos1, Raphael Winkels1, Alexandros Poulopoulos2, Heinrich Betz3, Thomas Deller1, Nils Brose2, Fre´de´rique Varoqueaux2 and Stephan W. Schwarzacher1 1

Institute of Clinical Neuroanatomy, Goethe-University Frankfurt, NeuroScience Center, D-60590 Frankfurt am Main, Germany, Department of Molecular Neurobiology and Center for the Molecular Physiology of the Brain, Max-Planck Institute of Experimental Medicine, D-37075 Go¨ttingen, Germany and 3Department of Neurochemistry, Max-Planck Institute for Brain Research, D-60528 Frankfurt, Germany 2

The postsynaptic adhesion protein neuroligin-2 (NL2) is selectively localized at inhibitory synapses. Here, we studied network activity in the dentate gyrus of NL2-deficient mice following perforant path (PP) stimulation in vivo. We found a strong increase in granule cell (GC) excitability. Furthermore, paired-pulse inhibition (PPI) of the population spike, a measure for g-aminobutyric acid (GABA)ergic network inhibition, was severely impaired and associated with reduced GABAA receptor (GABAAR)--mediated miniature inhibitory postsynaptic currents recorded from NL2-deficient GCs. In agreement with these functional data, the number of gephyrin and GABAAR clusters was significantly reduced in the absence of NL2, indicating a loss of synaptic GABAARs from the somata of GCs. Computer simulations of the dentate network showed that impairment of perisomatic inhibition is able to explain the electrophysiological changes observed in the dentate circuitry of NL2 knockout animals. Collectively, our data demonstrate for the first time that deletion of NL2 increases excitability of cortical neurons in the hippocampus of intact animals, most likely through impaired GABAAR clustering. Keywords: GABA, GABAA receptor clustering, network activity, perisomatic inhibition, computational modeling

Introduction Neuroligins (NLs) form a family of postsynaptic cell-adhesion proteins interacting with presynaptic neurexins. They have been implicated as key players in the stabilization of excitatory glutamatergic and inhibitory c-aminobutyric acid (GABA)ergic synapses (Craig and Kang 2007). Expression levels and localization of different NLs have been proposed to control the balance between synaptic excitation and inhibition (E/I) (Cline 2005; Levinson and El-Husseini 2005; Lise´ and El-Husseini 2006). Initial in vitro studies suggested that NLs regulate the ratio of inhibitory and excitatory synapses on neuronal cells (Graf et al. 2004; Prange et al. 2004; Chih et al. 2005; Levinson et al. 2005). Recent analyses of NL knockout (KO) mice showed that, although not required for maintaining the number of synapses, NLs are essential for synaptic maturation and function (Varoqueaux et al. 2006). The significant role of NLs in neural information processing is supported by findings that NL deficits are associated with autism and other cognition disorders (Belmonte and Bourgeron 2006; Garber 2007; Su¨dhof 2008). NL2 is selectively targeted to inhibitory postsynapses as indicated by its colocalization with the inhibitory scaffold protein gephyrin and GABAARs (Graf et al. 2004; Varoqueaux Ó The Author 2010. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]

et al. 2004; Heine et al. 2008; Patrizi et al. 2008; Hoon et al. 2009). In vitro electrophysiological recordings revealed that deletion of NL2 affects inhibitory synaptic transmission in the central nervous system (Chubykin et al. 2007; Gibson et al. 2009; Hoon et al. 2009; Poulopoulos et al. 2009). Based on studies in cultured nonneuronal cells and in NL2-deficient neurons, NL2 has been proposed to facilitate the recruitment of GABAA receptors (GABAARs) (Dong et al. 2007; Hoon et al. 2009; Poulopoulos et al. 2009). Hence, increasing evidence suggests a pivotal role of NL2 in promoting the assembly and stability of functional inhibitory synapses. A key issue is whether NL2 regulates GABAergic transmission and thereby the ratio of E/I also in vivo (Craig and Kang 2007). Electroretinogram recordings in NL2-deficient mice showed a tendency toward smaller oscillatory potentials, suggesting that NL2 is important for in vivo retinal information processing (Hoon et al. 2009). Using a gain-of-function approach, a recent study has demonstrated that enhanced expression of NL2 in transgenic mice alters their behavior and leads to a significant increase in inhibitory synapse maturation and transmission in the frontal cortex (Hines et al. 2008). Apparently overexpression of NL2 is sufficient to shift the E/I ratio toward inhibition. However, it remains unclear whether NL2 is necessary for maintaining synaptic balance and neuronal excitability in cortical circuits of intact animals. To address this question, we employed a loss-of-function approach and investigated the functional and morphological consequences of NL2 removal in the dentate gyrus of adult NL2 KO mice. By combining field recordings in anesthetized mice, patch-clamp analysis in hippocampal slices, immunohistochemistry and computational modeling, we show that NL2 is required for normal GABAergic inhibition, neuronal excitability, and GABAAR clustering in the dentate circuit in vivo. Materials and Methods NL2 KO Mice Experiments were performed on 5- to 12-week-old adult wild-type (WT) and NL2 deletion--mutant (KO) littermate mice issued from the interbreeding of heterozygous pairs (Varoqueaux et al. 2006). All experiments were carried out in accordance with German laws governing the use of laboratory animals. All analyses were carried out on male age--matched littermates. Surgery and In Vivo Electrophysiology Electrophysiological recordings in the dentate gyrus of NL2 KO and WT mice were carried out as described before (Jedlicka, Papadopoulos, et al. 2009; Jedlicka, Schwarzacher, et al. 2009). Briefly, 8- to 12-weekold NL KO mice and their WT littermates were anesthetized with

Downloaded from cercor.oxfordjournals.org at Universitaetsbibliothek Johann Christian Senckenberg on December 14, 2010

Address correspondence to Peter Jedlicka, NeuroScience Center, Clinical Neuroanatomy, Goethe-University, Theodor-Stern-Kai 7, D-60590 Frankfurt am Main. Email: [email protected].

Electrophysiological Recordings in Hippocampal Slices For patch-clamp recordings, 5- to 6-week-old mice were anesthetized with isoflurane and decapitated. Transverse 350-lm thick hippocampal slices were cut using a vibratome tissue sectioner (Leica). Slices were dissected in a solution containing (in mM) 85 NaCl, 24 NaHCO3, 1.25 NaH2PO4, 2.5 KCl, 25 glucose, 75 saccharose, 0.5 CaCl2, 4 MgCl2 equilibrated with 95% O2 and 5% CO2. Individual slices were transferred to a temperature controlled submersion-type recording chamber (at 35 °C; Luigs and Neumann). The dentate gyrus was identified by differential interference contrast microscopy (Zeiss Axioscop 2FS). Whole-cell patch-clamp recording of GABAergic miniature inhibitory postsynaptic currents (mIPSCs) from granule cells (GCs) was performed at a holding potential of –80 mV, using a MultiClamp 700B amplifier (Axon Instruments). Patch pipettes were pulled from borosilicate (GC150TF-10) glass capillary tubing (1.5 mm outer diameter; Harvard Apparatus). The pipette solution contained (in mM): 140 CsCl, 2 MgCl2, 10 HEPES, and biocytin (0.3%). All mIPSCs recordings were performed at 37 °C in artificial cerebrospinal fluid (in mM; 126 NaCl, 26 NaHCO3, 1.25 NaH2PO4, 2.5 KCl, 10 glucose, 2 CaCl2, and 2 MgCl2), containing 0.5 lM tetrodotoxin (TTX), 10 lM 6cyano-7-nitroquinoxaline-2,3-dione (CNQX), and 10 lM D-2-Amino-5phosphonopentanoic acid (D-AP5; all from Tocris). Series resistance was continuously monitored, and the recordings were discarded if the series resistance reached 35 MX. The membrane currents were digitized at a sampling rate of 10 kHz using the DigiData 1440A interface (Axon Instruments). Data acquisition and analysis was performed using pClamp 10.2 (Axon Instruments) and MiniAnalysis (Synaptosoft) software. All mIPSC events were visually inspected and detected by an independent blind investigator. Values are shown as mean ± standard deviation (n is number of mice). The slope of mIPSCs was calculated as the change in current amplitude over time of onset in the interval between 10% and 90% of peak amplitude. All recorded neurons were post hoc identified by histochemistry to ensure that recordings were made from GCs. Slices were fixed in a solution of 4%

Page 2 of 11 Neuroligin-2 and Dentate Gyrus Excitability

d

Jedlicka et al.

paraformaldehyde in phosphate-buffered saline (PBS, 0.1 M, pH 7.4) and 4% sucrose for 1 h. The sections were washed thoroughly in PBS and blocked by a 1-h preincubation in blocking buffer (10% normal goat serum [NGS], 0.5% Triton X-100 diluted in PBS) followed by an incubation with Alexa568 conjugated streptavidin (Invitrogen; 1:500 in PBS, 1% NGS, and 0.2% Triton X-100) for 2 h. After several washes in PBS, sections were coverslipped in antifading mounting medium (DAKO Fluoromount) for microscopic analysis using a confocal microscope (Nikon A1). The characteristic dendritic tree and an axon projecting toward area CA3 (mossy fiber) served as morphological criteria for GCs. Sixteen GCs in 4 animals (4 cells per animal in 3--4 hippocampal slices) were analyzed in each group. Computational Modeling We used a detailed computational model of the dentate gyrus network, containing 4 major cell types (Santhakumar et al. 2005; see also Winkels et al. 2009): 500 GCs (cells 0--499), 15 mossy cells (MCs, cells 506--520), 6 basket cells (BCs, cells 500--505), and 6 hilar cells (HCs, cells 521-526). The model represents a 2000:1 scaled-down version of the dentate gyrus (Santhakumar et al. 2005). Simulation files were downloaded from the ModelDB website (Davison et al. 2004; Hines et al. 2004): http://senselab.med.yale.edu/modeldb/. All simulations were carried out with the NEURON simulation program (Hines and Carnevale 1997). For details regarding structural, passive and active properties of model cells, and synaptic and network parameters see Santhakumar et al. (2005). Parameters used in our simulations were identical to parameters in the published network model, including somatic GABAAR conductances (BC-GC: 1.6 nS, BC-MC: 1.5 nS, and BCBC: 7.6 nS). To explore the effects of the reduction of somatic GABAARs on dentate network activity, GABAAR conductance was gradually diminished selectively at BC-GC synapses or at all BC postsynapses. Similar results could be obtained in both kinds of simulation experiments. To study the effects of reduced dendritic inhibition on GC activation, GABAAR conductance was decreased at HC-GC synapses. To analyze PPI of GC firing, simulated network activity was initiated by paired-pulse synchronous activation of PP synaptic inputs to all postsynaptic cells with varying interpulse intervals. Similarly to the original model, PP synapses were modeled using strong synaptic conductance (GPPtoGC = 20 nS, GPPtoBC = 10 nS, GPPtoMC = 2.5 nS) to ensure that all GCs will fire after a single stimulus. PPDI is a more complex phenomenon which is thought to depend on various mechanisms including the reduction of GABAergic inhibition mediated by presynaptic GABABRs (Davies et al. 1991; Lambert and Wilson 1994; Brucato et al. 1995; Bliss et al. 2007; but see also Kraushaar and Jonas 2000) and rebound firing due to preceding hyperpolarization (Lomo 2009). Therefore, as the network model contains only GABAARs, we simulated only the PPI part of the PPI/PPDI curve (see Jedlicka et al. 2010 for the detailed discussion of model simplifications). For data analysis, the activity of the dentate gyrus network was visualized using spike time raster plots. The activity of GCs was presented as the percentage of maximal GC activation. Immunofluorescence Labeling Double immunostainings with antibodies against the c2 subunits of GABAARs (guinea pig polyclonal, 1:4000; the antibody was kindly provided by Dr Jean-Marc Fritschy, University of Zurich), and the gephyrin-specific antibody mAb7a (mouse monoclonal; 1:400), were performed as described previously (Papadopoulos et al. 2007; Jedlicka, Papadopoulos, et al. 2009). Briefly, mice were deeply anesthetized and decapitated. The brains were immediately removed and frozen on dry ice. Coronal hippocampal cryostat sections (14 lm) were fixed with 4% (w/v) paraformaldehyde for 10 min at 4 °C. Sections were preincubated for 30 min at 95 °C in SC buffer (10 mM sodium citrate, 0.05% (v/v) Tween-20, pH 8.0). Sections were permeabilized with 0.3% (w/v) Triton X-100 and incubated overnight at 4 °C with primary antibodies at appropriate dilutions in PBS/10% goat serum and for 1 h with secondary antibodies (Alexa488 and Alexa 546; Invitrogen; 1:1000). For double immunostaining with the antibody recognizing gephyrin (mAb7a, 1:400) and an antibody specific for the vesicular inhibitory amino acid transporter (VIAAT, rabbit polyclonal, 1:500, Synaptic Systems GmbH), mice were deeply anesthetized and decapitated. The

Downloaded from cercor.oxfordjournals.org at Universitaetsbibliothek Johann Christian Senckenberg on December 14, 2010

urethane (Sigma, 1.2 g/kg intraperitoneally; supplemental doses of 0.3--0.6 g/kg subcutaneously as needed) and kept at 37 °C. All recordings were made blind to the genotype. Recordings and stimulation were made in the granule cell layer (GCL) of the dentate gyrus (1.7 mm posterior and 0.9 mm lateral to bregma) and in the medial perforant path (PP) (3.8 mm posterior to bregma and 2.1 mm lateral to lambda), respectively. Stimulus--response relationships for field excitatory postsynaptic potentials (fEPSPs) and population spikes were determined using a range of stimulation intensities from 30 to 800 lA. The amplitude of the population spike was defined as the average of the amplitude from the first positive peak (a) to the succeeding negative peak (b) and the amplitude from the negative peak (b) to the second positive peak (c): ([a – b] + [c – b])/2. For the analysis of the slope of fEPSPs, only the early component of the response was measured to avoid contamination by the population spike. To measure paired-pulse facilitation of the fEPSP amplitude, a double-pulse stimulation at intensities subthreshold to a population spike was applied, with interpulse intervals of 15--100 ms. To study paired-pulse inhibition and disinhibition (PPI/PPDI) of the population spike, maximum (800 lA) and minimum (evoking 1-mV population spikes) stimulation intensities were used (interpulse intervals 15--1000 ms). PPI/PPDI curves were fitted using a Boltzmann equation to obtain the mean interpulse interval at which equal amplitudes of the first and second population spike could be observed. Repetitive stimulation of the PP with increasing stimulus frequencies (0.5--4 Hz) was used to determine the threshold for seizure-like responses in the form of GC multiple spikes. Long-term potentiation (LTP) was induced by theta-burst stimulation (TBS): 6 series of 6 trains of 6 stimuli at 400 Hz, 200 ms between trains, 20 s between series (Jones et al. 2001; Jedlicka, Papadopoulos, et al. 2009; Jedlicka, Schwarzacher, et al. 2009). Pulse width and stimulus intensity was doubled during the TBS in comparison with baseline recordings. A baseline fEPSP slope was calculated from the average of responses over the 10 min prior to the TBS. Baseline stimulus intensity was set to evoke a population spike of approximately 1 mV before the induction of LTP. The potentiation of the fEPSP slope was expressed as a percentage change relative to the baseline.

brains were immediately removed and incubated overnight at 4 °C in 50 mL fixative containing 4% (w/v) paraformaldehyde and 0.1% (w/v) glutaraldehyde in 0.1 M phosphate buffer, pH 7.4. Coronal 20-lm sections were prepared from frozen tissue and were mounted on SuperFrost Plus slides (Menzel GmbH). Sections were then postfixed for 10 min with 4% (w/v) paraformaldehyde at 4 °C, followed by immunostaining as described above.

Quantification Images were further processed with the AnalySIS software (Olympus). For quantification of synaptic markers, single-plane confocal images were smoothened with an open filter and objects (i.e., fluorescent clusters) were separated upon application of a separation filter (particle separator). Thresholded images were used to detect the total number of clusters throughout molecular versus GCLs. For VIAAT-gephyrin colocalization studies, regions of interest were also drawn for each object in a given channel, superimposed on the complementary channel, and the number of colocalized objects was determined manually (Fletcher et al. 1998). Statistical Analysis Differences between groups were statistically analyzed by an unpaired 2-tailed Student’s t-test. In cases where variance was inhomogeneous between groups (tested by Leven’s test, P < 0.05), the nonparametric Mann--Whitney U test was used. All statistical analyses were done using SPSS for Windows. Group values are reported as means ± standard error of the mean unless stated otherwise.

Results Basal Excitatory Transmission Is Unchanged at Perforant Path--Granule Cell Synapses in the Absence of NL2 First, we investigated the consequences of NL2 deficiency for the dentate gyrus network using electrophysiological recordings in anesthetized NL2 KO mice following PP stimulation. To examine synaptic transmission at excitatory perforant path-granule cell (PP-GC) synapses, we recorded evoked fEPSPs. Analysis of fEPSP slopes did not reveal significant differences between NL2 KO mice (n = 10, maximum slope 1.3 ± 0.1 mV/ ms) and their WT littermates (n = 12; 1.5 ± 0.1 mV/ms; P > 0.19; Fig. 1A). Since the slope of the fEPSP is a measure of synaptic efficacy, these data suggest that excitatory synapses are not significantly altered in NL2 KO animals. Consistent with these electrophysiological findings, double labeling for the excitatory presynaptic markers vesicular glutamate transporter 1 (VGLUT1) and vesicular glutamate transporter 2 (VGLUT2) showed that the overall density of glutamatergic nerve terminals in the dentate gyrus is similar in WT and NL2 KO mice (Supplementary Fig. 2) To evaluate the presynaptic function and short-term plasticity of PP-GC synapses, we compared paired-pulse facilitation of fEPSPs in NL2 KO and WT mice. Paired-pulse facilitation depends on presynaptic mechanisms (Zucker and Regehr 2002). It was determined as the ratio of fEPSP amplitudes following 2 successive stimuli at various interpulse intervals. The stimulation was subthreshold for the elicitation

GC Excitability Is Increased in NL2 KO Mice NL2 could be involved in the regulation of dentate GC excitability via GABAergic mechanisms (Varoqueaux et al. 2004). Therefore, we studied excitability of GCs in NL2deficient mice by comparing evoked population spikes across a range of PP stimulation intensities (50 -- 800 lA). The size of the population spike reflects the number and synchrony of firing GCs (Andersen et al. 1971). An input--output (IO) analysis revealed significantly higher amplitudes of the population spikes in NL2 KO mice (n = 10; maximum spike 5.4 ± 0.4 mV) as compared with their WT littermates (n = 12; 3.6 ± 0.3 mV; P < 0.003; Fig. 1C). Moreover, in mutant animals, more than one population spike could be observed as shown in the IO curve relating the stimulation intensity and the number of population spikes (P < 0.02; Supplementary Fig. 1A). Thus, the NL2 deficit greatly enhances the ability of GCs to fire action potentials. To further assess the excitability of GCs at the network level, we determined the frequency threshold for evoked multiple spike (seizure-like) responses. Seizure-like activity of GCs can be induced by repetitive stimulation of the PP with increasing frequencies (see Materials and Methods). NL2-deficient mice showed a significantly lower threshold frequency for the induction of epileptiform discharges (n = 7; 2.0 ± 0.1 Hz) in comparison with WT littermates (n = 7; P < 0.005; 2.9 ± 0.2 Hz; Fig. 1D). Taken together, these findings demonstrate a dramatic increase in GC excitability in the NL2-deficient dentate gyrus in vivo. GABAergic Network Inhibition Is Impaired in the Dentate Gyrus of NL2 KO Mice To investigate GABAergic network inhibition in the dentate gyrus of NL2 KO mice, we determined the paired-pulse ratio of the population spike. Paired-pulse stimulation of PP inputs results in population spike depression at short interstimulus intervals (PPI) followed by facilitation at longer intervals (PPDI). PPI and PPDI depend on GABAergic synaptic inhibition and disinhibition in the dentate network (Sloviter 1991; Bronzino et al. 1997; Jedlicka, Papadopoulos, et al. 2009; Lomo 2009). NL2 KO mice (n = 11) displayed markedly weaker PPI as indicated by a leftward shift of the PPI/PPDI curve as compared with WT curve (n = 10, Fig. 2A). The interval at which PPI turned to PPDI was significantly shortened in NL2 mutants (36.7 ± 1.6 ms) relative to WTs (48.0 ± 2.5 ms, P < 0.001; Fig 2A, right). To confirm that this difference was independent of stimulation intensity, we determined the paired-pulse ratio of the population spike also at minimal stimulation intensity (see Materials and Methods). Again, we found significantly diminished PPI in NL2-deficient mice (P < 0.05; Supplementary Fig. 1B). These data demonstrate that NL2-deficiency results in a substantial impairment of GABAergic inhibition and an increase of dentate network excitability in vivo. Inhibitory GABAergic Synaptic Transmission Is Reduced in the Dentate Gyrus of NL2 KO Mice To confirm that the reduced GABAergic network inhibition seen in the dentate gyrus of NL2 KO mice is of synaptic origin, Cerebral Cortex Page 3 of 11

Downloaded from cercor.oxfordjournals.org at Universitaetsbibliothek Johann Christian Senckenberg on December 14, 2010

Imaging Serial confocal images of immunostained slices were captured at a total magnification of either 4003 or 6303 on a confocal laser scanning microscope (Zeiss LSM 510 or Leica SP2 or Leica TCS-SP). All images were acquired as single optical slices. During acquisition, imaging parameters (gain and offset) were kept constant for a given labeling to allow for fluorescence intensity comparisons.

of a population spike. Paired-pulse facilitation was similar in WT (n = 12) and NL2 KO mice (n = 11, P > 0.2; Fig. 1B), indicating that the lack of NL2 does not affect presynaptic physiology at the PP-GC synapse in vivo.

we recorded GABAergic mIPSCs from GCs (Fig. 3B, see Materials and Methods) in slices prepared from NL2 KO mice and littermate WT controls. Pharmacologically isolated mIPSCs (0.5 lM TTX, 10 lM APV, and 1 lM CNQX) exhibited a strong reduction in amplitude in NL2-deficient neurons as compared with their WT counterparts (WT: 33.1 pA vs. NL2 KO: 22.9 pA; P < 0.001; Fig. 3A,C,D). In contrast, mean mIPSC frequency (WT: 3.2 Hz vs. NL2 KO: 3.1 Hz; P = 0.98) was not different in both groups. These data confirm that the loss of NL2 significantly reduces GABAergic synaptic inhibition of dentate GCs. GABAergic inhibition is known to modulate synaptic plasticity (Wigstro¨m and Gustafsson 1983; Bliss and Collingridge 1993; Paulsen and Moser 1998). Therefore, we tested the induction of LTP at PP-GC synapses in NL2 KO and WT mice, using a TBS protocol. LTP induction was not significantly changed in mutant animals (Fig. 2B; 0--10 min: 139.7 ± 9.1%) as compared with their WT littermates (137.2 ± 6.3%; P > 0.6). These data demonstrate that the lack of NL2 does not disrupt the ability of dentate PP-GC synapses to undergo LTP. Postsynaptic Gephyrin and GABAAR Cluster Numbers Are Reduced in the NL2 KO Dentate Gyrus In attempt to identify the anatomical substrate underlying the altered GABAergic transmission in the dentate gyrus of NL2deficient mice, we investigated the distribution of inhibitory Page 4 of 11 Neuroligin-2 and Dentate Gyrus Excitability

d

Jedlicka et al.

pre- and postsynaptic marker proteins in hippocampal sections derived from adult NL2 KO mice and their WT littermates. Staining with a gephyrin antibody revealed significant reductions in the density of gephyrin clusters in the dentate GCL (Fig. 4), whereas cluster density in the molecular layer (ML) was not detectably altered. Quantification of the number of gephyrin-immunoreactive puncta per 500 lm2 resulted in density values of 60.3 ± 2.2 versus 36.3 ± 4.5 for the GCL (P < 0.05) and of 126.2 ± 1 versus 124.7 ± 11.7 for the ML (not significant; ns) of the dentate gyrus in WT and NL2 KO sections, respectively (Fig. 4G). Notably, the number of VIAAT-immunoreactive puncta (GCL: WT: 114.0 ± 3.3 vs. KO: 109 ± 5.9 puncta per 500 lm2, ns; ML: WT: 140.7 ± 14.7 vs. KO: 147.2 ± 12.5 puncta per 500 lm2, ns) was not changed in the absence of NL2 (Fig. 4H). This indicates that the density of inhibitory nerve terminals was unaffected in the mutant mice. Since NL2 might be involved in the clustering of GABAARs at postsynaptic sites (Craig and Kang 2007; Hoon et al. 2009; Poulopoulos et al. 2009), we examined GABAAR localization in the NL2-deficient dentate gyrus. Using immunofluorescence staining, we found a significant reduction in the density of postsynaptic clusters containing the GABAAR c2-subunit in the GCL but not ML of NL2 KO mice, as compared with their WT littermates (Fig. 5A--F). Quantification of the number of immunoreactive clusters per 500 lm2 section area revealed

Downloaded from cercor.oxfordjournals.org at Universitaetsbibliothek Johann Christian Senckenberg on December 14, 2010

Figure 1. Increased GC excitability in anesthetized NL2 KO mice. (A) Analysis of field excitatory postsynaptic potentials (fEPSPs) in the dentate gyrus. Input-output (IO) analysis of fEPSP slopes revealed no significant differences between NL2 KO (n 5 10) and WT mice (n 5 12) (P [ 0.19). The fEPSP slope reflects the strength of PP-GC synapses. Sample responses at maximum stimulation intensity are displayed as Inset. (B) Paired-pulse facilitation of the fEPSP at various interstimulus intervals in the dentate gyrus of NL2 KO (n 5 11) and WT mice (n 5 12). The percentages denote the ratio of the second fEPSP amplitude to the first fEPSP amplitude. The difference between genotypes was not significant (P [ 0.2). Inset: Sample recordings from WT and NL2 KO mice show facilitation at 15 ms. (C): IO analysis of population spikes recorded in NL2 KO (n 5 10) and WT mice (n 5 12). The population spike amplitude is a measure for the number and the synchrony of firing GCs. Note significantly increased amplitude of population spikes in NL2 KO mice indicating strongly enhanced GC excitability (P \ 0.003). (D) NL2-deficient mice (n 5 7) showed significantly (*P \ 0.05) lower threshold frequency for the induction of epileptiform (multiple spike) discharges as compared with WT littermates (n 5 7). Inset: Sample multiple spike activity of GCs in a WT mouse at 2.5-Hz stimulation frequency.

density values of 73.1 ± 6.0 versus 44.6 ± 2.0 for the GCL (P < 0.001) and of 101.3 ± 7.3 versus 111.8 ± 5.0, for the ML (ns) in WT and NL2 KO sections, respectively (Fig. 5G--H). Correspondingly, the percentage of gephyrin clusters colocalized with the c2-subunit of GABAARs was significantly reduced in the GCL of NL2 KOs (GCL: WT: 90.0 ± 1.8% vs. KO: 33.8 ± 1.9%, P < 0.001; ML: WT: 93.3 ± 1.3% vs. KO: 91.3 ± 1.5%, of total gephyrin puncta, ns). In conclusion, NL2 deficiency causes a layer-specific reduction of gephyrin and GABAAR c2-subunit clusters in the somatic region of dentate GCs. Computational Modeling of Impaired Somatic GABAergic Inhibition To understand the network effects of NL2 deficiency, we addressed the question whether the observed reduction of somatic GABAAR clusters may be sufficient to account for the electrophysiological effects found in NL2 KO mice. The potential impact of a selective reduction in perisomatic

Discussion Using a combination of electrophysiological measurements and morphological analyses, this study shows that deletion of NL2 in the dentate gyrus results in a significant impairment of GABAergic inhibition, which is due to reduced clustering of GABAARs, and increased neuronal excitability in vivo. In Vivo Dentate Gyrus Excitability Is Increased in the Absence of NL2 Our recordings in intact mice reveal a dramatic shift in the synaptic E/I balance toward excitation in the NL2-deficient dentate network. This is in agreement with in vitro data suggesting that NLs are crucial determinants of the E/I ratio (Levinson et al. 2005; Chubykin et al. 2007; Hines et al. 2008). Our electrophysiological analysis provides 3 major findings. First, the lack of NL2 does not affect excitatory synaptic transmission at PP-GC synapses. This is demonstrated by the unchanged IO curve for the fEPSP slope, which is a measure of synaptic efficacy. Presynaptic properties are also not significantly altered, as indicated by the paired-pulse facilitation test. Consistent with our data, neither NL2-deficient neurons in acute cortical slices nor cultured hippocampal neurons overexpressing NL2 displayed changes in excitatory synaptic function (Chubykin et al. 2007). Likewise, pyramidal neurons in the prefrontal cortex of transgenic mice with increased NL2 expression did not show any reduction in miniature excitatory postsynaptic currents as compared with WT controls (Hines et al. 2008). Decreases in inhibitory tone and excitability may facilitate the induction of plastic changes at excitatory synapses (Wigstro¨m and Gustafsson 1983; Papadopoulos et al. 2007; Jedlicka, Papadopoulos, et al. 2009). However, similar to basal Cerebral Cortex Page 5 of 11

Downloaded from cercor.oxfordjournals.org at Universitaetsbibliothek Johann Christian Senckenberg on December 14, 2010

Figure 2. GABAergic network inhibition is impaired in the dentate gyrus of NL2 KO mice. (A) Paired-pulse inhibition/disinhibition (PPI/PPDI) of the population spike in the dentate gyrus of NL2 KO (n 5 9) and WT mice (n 5 10) at maximal stimulation intensity. PPI reflects GABAergic network inhibition. Note a significant leftward shift in the PPI/PPDI curves of NL2 KO mice. Top: Sample traces represent paired-pulse responses at 50-ms interstimulus interval, showing PPI in a WT mouse and PPDI in a NL2 KO mouse. Diagram: mean interpulse interval (in ms) at which equal amplitude of the first and second population spike was observed (i.e., duration of PPI). Note a significant (***P \ 0.001) decrease in the duration of PPI in NL2 KO mice. Similar effects were also observed at minimal stimulation intensity showing that the PPI reduction is independent from stimulus strength (Supplementary Fig. 1B). (B) Induction of LTP in the dentate gyrus of NL2 KO (n 5 4) and WT mice (n 5 5). Mean normalized fEPSP slope is plotted as a function of time. The potentiation is expressed as a percentage change relative to the mean response in the 10 min prior to TBS (6 series of 6 trains of 6 pulses at 400 Hz, 200 ms between trains, 20 s between series; arrow). NL2 KO mice show unimpaired long-term synaptic plasticity at PP-GC synapses.

GABAAR density on network activity was assessed using an established computational model of the dentate gyrus circuitry (Santhakumar et al. 2005). This network model comprises PP inputs and synaptic connections of GC, MC, BC, and HC cells based on realistic morphological and electrophysiological data (see Materials and Methods). To disclose the role of GABAARs in dentate network activity, we systematically varied their perisomatic densities. To simulate PPI data from NL2 KO mice, we studied the effect of GABAAR density changes on the network activity after paired-pulse stimulation of PP fibers (Winkels et al. 2009; see Materials and Methods). In the ‘‘WT’’ network, GC firing was suppressed after the second pulse, similarly to the experimentally observed PPI phenomenon (Fig. 6A,B, left panels; cf. Fig. 2A). Importantly, in the ‘‘KO’’ network model, PPI was significantly reduced (Fig. 6A,B, right panels) The reduction of perisomatic GABAAR densities in the KO simulations reproduced the leftward shift of the PPI curve observed in NL2-deficient mice (Fig. 6C; c.f. Fig. 2A). Thus, in the ‘‘NL2 KO’’ network, perisomatically targeting interneurons (BCs) were less effective in preventing action potential generation in GCs following paired-pulse stimulation. By contrast, selective reduction of dendritic GABAAR conductances did not result in significant changes of simulated PPI (Supplementary Fig. 3), suggesting that dendritically targeting interneurons (HCs) do not play a major role in mediating PPI. These computational results indicate that the loss of perisomatic GABAARs is sufficient to explain increased network excitability and decreased PPI of GCs in the dentate circuit of NL2 KO animals.

synaptic transmission, long-term plasticity of PP-GC synapses induced by a classical TBS protocol (e.g., Jones et al. 2001; Cooke et al. 2006) was not significantly altered by NL2deficiency and the associated increase in the E/I ratio. Second, and in contrast to the unaltered PP input, GABAergic inhibition of GCs is reduced in NL2-deficient mice, as indicated by the PPI test and the analysis of mIPSCs. PPI of the population spike is a readout of inhibitory circuit function, as it reflects GABAAR-mediated inhibition of GCs through local interneurons in the dentate gyrus (Sloviter 1991; Lomo 2009). The leftward shift of the PPI/PPDI curve indicates a decrease of GABAergic inhibition, giving rise to reduced suppression of GC population spikes in NL2 KO mice. The marked enhancement of dentate network activity seen in vivo was associated with a pronounced decrease of GABAergic synaptic inhibition in vitro, as disclosed by mIPSC recordings. The amplitudes of GABAergic mIPSCs were significantly diminished, most likely reflecting the loss of postsynaptic GABAARs (see below). Our data complement previous electrophysiological observations in acute slices prepared from NL2-deficient mice (Chubykin et al. 2007; Poulopoulos et al. 2009). Furthermore, enhanced inhibitory transmission was found in transgenic mice overexpressing NL2 (Hines et al. 2008; see also Fu and Vicini 2009), also supporting the close relationship of NL2 and inhibition. Taken together, both in vivo and in vitro data show that NL2 determines the efficacy of GABAergic synapses. Consistent with the reduced GABAergic inhibition found here, NL2-deficient animals have been recently shown to exhibit increased anxiety (Blundell et al. 2009), a behavioral trait known to be regulated by GABAergic mechanisms (Freund and Katona 2007). Third, deletion of NL2 significantly increases overall network excitability in the dentate gyrus. This is evident from both the strongly enhanced IO relationship between stimulation and population spikes and the decreased threshold frequency for Page 6 of 11 Neuroligin-2 and Dentate Gyrus Excitability

d

Jedlicka et al.

the induction of epileptiform GC discharges. Thus, GCs in the NL2-deficient dentate network have a lower threshold for generating action potentials in response to PP stimulation. As demonstrated by reduced GABAAR clustering (see below), decreased PPI and smaller mIPSC amplitudes, the enhanced excitability of GCs in NL2 KO mice is most likely attributed to reduced GABAergic inhibition. Although our data indicate that disrupted GABAA-mediated inhibition is sufficient to account for impaired PPI and increased GC excitability, we cannot completely rule out the possibility that additional indirect mechanisms triggered by the absence of NL2 contribute to the observed changes (Supplementary Fig. 4; cf. Gibson et al. 2009). However, it is unlikely that unspecific compensatory effects of NL2 deletion play a major part in augmented dentate network excitability since they would be expected to counteract the reduction of GABAergic inhibition (e.g., by homeostatic downregulation of voltage-gated channels or upregulation of presynaptic GABAB autoreceptors) thereby undermining (rather than mimicking) excitability changes in NL2 KO animals. Direct measurements of intrinsic GC properties in NL2deficient dentate gyrus might help to clarify whether their alteration diminishes (or participates in) GC excitability changes. Of note, disinhibition of the dentate gyrus might lead to temporal lobe epilepsy (Coulter and Carlson 2007). Further studies are needed to test whether the hyperexcitability of GCs resulting from NL2 deficiency has epileptogenic consequences. Disturbed GABAAR Clustering in NL2 KO Mice To investigate the molecular mechanism underlying the effects of NL2 deficiency, we performed immunohistochemical analyses of synaptic proteins in the NL2-deficient dentate network. Staining for VGLUT1/2 and VIAAT did not reveal significant differences in the distribution and density of excitatory and inhibitory presynaptic terminals between KO

Downloaded from cercor.oxfordjournals.org at Universitaetsbibliothek Johann Christian Senckenberg on December 14, 2010

Figure 3. Reduced GABAergic synaptic inhibition of dentate GCs in NL2-deficient hippocampal slices. (A) Traces of mIPSCs recorded from WT and NL2 KO neurons in the presence of 0.5 lM TTX, 10 lM CNQX, and 10 lM D-AP5 at a holding voltage of 80 mV. (B) Post hoc identified recorded GCs in the dentate gyrus. Scale bar 5 100 lm. (C) Cumulative distributions of mIPSC amplitudes and inter event intervals of 16 neurons per genotype (n 5 4 animals per group, 4 cells per animal). (D) Mean values for mIPSC amplitudes and frequency. The mIPSC amplitude is significantly reduced (P \ 0.001) in dentate GCs of NL2 KO mice.

and WT animals. This is consistent with previous findings showing that the density of synaptic contacts is not altered in the brain of NL 1--3 triple KOs (Varoqueaux et al. 2006). Similarly, a recent quantitative electron microscopy study found no change in synapse number in the NL2-deficient brain (Blundell et al. 2009). These data lend further support to the idea that NLs play a crucial role in the maturation of synapses but not in their initial formation. In contrast to presynaptic markers, the number of postsynaptic gephyrin and GABAAR c2-subunit clusters was reduced in NL2 KO mice as compared with their WT littermates. The c2-subunit is required for the postsynaptic clustering and functional integrity of GABAARs (Schweizer et al. 2003). Hence, our data indicate that NL2 is essential for the proper clustering of GABAARs and the stability of GABAergic postsynapses in the dentate gyrus. This is consistent with earlier studies suggesting a link between NL2 and GABAARs (Dong et al. 2007; Hoon et al. 2009; see also Fu and Vicini 2009). We observed a deficit in GABAAR clustering exclusively in the GCL of the dentate gyrus. Since this region is a target area for GABAergic interneurons controlling the perisomatic domain of principal neurons in the hippocampus, the layer-specific reduction of immunolabeling indicates that the loss of NL2 preferentially affects receptor

clustering at somatic GABAergic synapses. These findings are in accordance with a recent study that found a strikingly similar pattern of layer-specific reduction in gephyrin and GABAAR immunoreactivity in the hippocampal CA1 region in the absence of NL2 (Poulopoulos et al. 2009). Viewed together, disruption of gephyrin and GABAAR clusters both in the GCL (this study) and in the pyramidal cell layer (Poulopoulos et al. 2009) suggests that, in the hippocampus, NL2 is involved in the synaptic recruitment of perisomatic GABAARs in a gephyrindependent manner. Moreover, the functional PPI data also point to the dysfunction of perisomatic inhibition in NL2 KO mice. PPI is mediated mainly by inhibitory GABAergic interneurons targeting GC bodies (Sloviter 1991; Moser 1996; Zappone and Sloviter 2004; Lomo 2009). Thus, the impaired PPI seen in NL2-deficient animals is in agreement with the immunohistological findings indicating a reduction of perisomatic inhibition. Computer Modeling of Increased Excitability in the Dentate Gyrus of NL2 KOs Our computational data support the conclusion that reduced GABAergic transmission in NL2 KO animals accounts for the changes in network inhibition disclosed by the PPI paradigm. Cerebral Cortex Page 7 of 11

Downloaded from cercor.oxfordjournals.org at Universitaetsbibliothek Johann Christian Senckenberg on December 14, 2010

Figure 4. Selective impairment of gephyrin clustering in the GC layer (GCL) of NL2 KO mice. (A--F) Double immunolabeling for VIAAT (red) and gephyrin (green) was performed in WT and NL2 KO mice. (G, H) Fluorescent puncta for VIAAT and gephyrin were counted. The number of gephyrin puncta and the percentage of their apposition to VIAAT puncta were reduced in the GCL in sections from KO animals as compared with WT sections. In contrast, VIAAT immunoreactivity was unaffected by NL2-deficiency. For both genotypes, each bar corresponds to counts performed on 2 sections from 3 individual brains (*P \ 0.05; Student’s t-test). Scale bar, 10 lm. ML, molecular layer.

While selective weakening of perisomatic GABAAR conductances impaired simulated PPI of GCs, specific reduction of dendritic GABAAR conductances did not result in significant PPI changes. The results obtained from the data-driven dentate gyrus model reveal that a number of mechanisms may lead to the alteration of PPI, including changes in intrinsic properties (Winkels et al. 2009), GABA reversal potential (Jedlicka et al. 2010), or short-term synaptic plasticity (Thomas et al. 2005; our observations, see Supplementary Material). For example, simulations suggest that upregulation of dendritic voltage-gated sodium channels may lead to PPI reduction (Supplementary Fig. 4). However, as our goal was to explore whether PPI reduction can emerge purely from changes in perisomatic GABAA--mediated inhibition, the NL2 KO network model deliberately did not include modification of multiple factors that reportedly affect PPI. Hence, the simulations indicate that the loss of functional somatic GABAARs is sufficient to explain the increased network excitability and PPI changes in the dentate circuit of NL2 KO animals. Whereas dendritic inhibitory conductances have been reported to lower the threshold for action potentials and thereby shift the IO function of hippocampal neurons without changing their maximal firing frequency, somatic inhibitory Page 8 of 11 Neuroligin-2 and Dentate Gyrus Excitability

d

Jedlicka et al.

conductances not only shift the IO curve but also decrease the maximal firing frequency (Pouille et al. 2008). Consistently, the maximal population spike amplitude was substantially increased in NL2 KO mice. Notably, this parameter is not affected in collybistin KO mice, which exhibit deficits in dendritic inhibition (Papadopoulos et al. 2007; Jedlicka, Papadopoulos, et al. 2009). Further tests, including paired recordings in acute hippocampal slices, will be needed to provide direct evidence for selective deficits at perisomatic BCGC synapses in the NL2-deficient dentate gyrus. Interestingly, a recent study in the barrel cortex (Gibson et al. 2009) has reported that NL2 deletion decreased unitary IPSC amplitude evoked from fast-spiking (perisomatically targeting) interneurons while having no effect on IPSCs mediated by (dendritically targeting) somatostatin-positive interneurons. These data support our findings which suggest that NL2 is mainly involved in the regulation of perisomatic GABAergic inhibition (see also Poulopoulos et al. 2009). Implications for Neurological Diseases Disturbances of the E/I balance are thought to underlie neurological defects such as autism (Rubenstein and Merzenich 2003; Polleux and Lauder 2004), anxiety-related disorders, and

Downloaded from cercor.oxfordjournals.org at Universitaetsbibliothek Johann Christian Senckenberg on December 14, 2010

Figure 5. Selective reduction of GABAAR c2-subunit clustering in the GCL of NL2 KO mice. (A--F) Sections from adult NL2 KO mice and their WT littermates were stained with antibodies specific for gephyrin and c2-subunit of GABAARs. (G, H) Quantification of immunoreactivities. For both genotypes, each bar corresponds to counts performed on 2 sections from 3 individual brains (***P \ 0.001; Student’s t-test). The punctate staining of sections from NL2 KO animals for c2-subunit of GABAARs was significantly reduced in the dentate GCL as compared with WT sections (G). The colocalization of GABAAR c2-subunit and gephyrin was also substantially decreased in the GCL of NL2 KOs (H). Scale bar, 10 lm. ML, molecular layer.

epilepsy (Freund and Katona 2007; Fritschy 2008). NL2 KO mice display anxiety-like behavior, decreased pain sensitivity and decreased motor coordination (Blundell et al. 2009), and NL2 transgenic mice exhibit impaired social interactions

(Hines et al. 2008). Interestingly, in patients with autism-spectrum conditions, distinct mutations in NL3 and NL4 genes have been identified (Jamain et al. 2003; Laumonnier et al. 2004; Ylisaukko-oja et al. 2005; Lawson-Yuen et al. 2008; for Cerebral Cortex Page 9 of 11

Downloaded from cercor.oxfordjournals.org at Universitaetsbibliothek Johann Christian Senckenberg on December 14, 2010

Figure 6. PPI of GC discharges is impaired in the network model when reducing somatic GABAAR densities. (A) Simulated voltage traces of GCs, basket cells (BCs), mossy cells (MCs) and hilar cells (HCs) after paired-pulse stimulation of PP inputs (17-ms interpulse interval) in control and modified network model (left and right, respectively). Note that whereas all GCs fire synchronous action potentials after the first stimulus, fewer GCs fire after the second stimulus in the WT as well as in the KO network (PPI). In the KO network, after the second stimulus, action potentials are generated by more GCs than in the WT network (reduced PPI). Arrows: PP stimulation. (B) Spike raster plot of network activity after paired-pulse stimulation of PP inputs (17-ms interpulse interval) in the dentate gyrus network model. Time (in ms) is on the horizontal axis and index of neurons in the network on the vertical axis. Each point represents an action potential. Note the decreased number and synchronicity of GC discharges following the second pulse in the WT as well as in the KO network (PPI). PPI is weaker in the KO network with reduced (50%) GABAAR channel density at somatic inhibitory synapses. In contrast, selective reduction of dendritic GABAergic inhibition in silico leaves PPI intact (see Supplementary Fig. 3). Arrows: PP stimulation. (C) The density of perisomatic GABAARs in dentate cells was systematically reduced from 100% to 0% of the control value, and the interpulse intervals were varied from 10 to 25 ms. Plots represent averages of 3 runs obtained with randomized connectivity. Diagram shows the dependence of PPI on GABAAR channel density (17-ms interpulse interval). *P \ 0.05; ***P \ 0.001. (D) Basic dentate gyrus circuitry: PP: perforant path, GC: granule cells, Inh: GABAergic interneurons. Our data indicate that the reduction of perisomatic GABAergic inhibition caused by the loss of NL2 leads to enhanced ability of GCs to fire action potentials and to an overall increase of E/I ratio in the network.

Conclusions Based on functional and morphological findings, we conclude that NL2 is essential for intact GABAAR-dependent network inhibition in the dentate gyrus in vivo. Our data indicate that the absence of NL2 leads to the loss of gephyrin and GABAARs from postsynaptic sites and results in reduced somatic GABAergic transmission and increased GC excitability. Thus, NL2 is crucial for stabilizing inhibitory postsynaptic sites and the physiological E/I ratio in the hippocampus (Fig. 6D).

Supplementary Material Supplementary material .oxfordjournals.org/.

can

be

found

at:

http://www.cercor

Funding Deutsche Forschungsgemeinschaft (JE 528/1-1 to P.J.); MaxPlanck-Gesellschaft to H.B.; Fonds der Chemischen Industrie to H.B.; Cure Autism Now foundation to F.V.; European Community (NEUREST MEST-CT-2004-504193 to M.H.); Center for the Molecular Physiology of the Brain to N.B. and F.V. Notes The authors thank Dr Jean-Marc Fritschy for providing antibodies, and Ina Bartnik and Felix Weise for technical assistance in preparing and staining hippocampal slices. Conflict of interest : None declared.

References Andersen P, Bliss TV, Skrede KK. 1971. Unit analysis of hippocampal population spikes. Exp Brain Res. 13:208--221. Belmonte MK, Bourgeron T. 2006. Fragile X syndrome and autism at the intersection of genetic and neural networks. Nat Neurosci. 9:1221--1225. Bliss TV, Collingridge GL. 1993. A synaptic model of memory: long-term potentiation in the hippocampus. Nature. 361:31--39. Bliss TV, Collinridge G, Morris RG. 2007. Synaptic plasticity in the hippocampus. In: Andersen P, Morris RG, Amaral DG, Bliss TV, O’Keefe J, editors. The Hippocampus book. Oxford: Oxford University Press. p. 343--474. Blundell J, Tabuchi K, Bolliger MF, Blaiss CA, Brose N, Liu X, Sudhof TC, Powell CM. 2009. Increased anxiety-like behavior in mice lacking the inhibitory synapse cell adhesion molecule neuroligin 2. Genes Brain Behav. 8:114--126. Bronzino JD, Blaise JH, Morgane PJ. 1997. The paired-pulse index: a measure of hippocampal dentate granule cell modulation. Ann Biomed Eng. 25:870--873. Brucato FH, Mott DD, Lewis DV, Swartzwelder HS. 1995. GABAB receptors modulate synaptically-evoked responses in the rat dentate gyrus, in vivo. Brain Res. 677:326--332.

Page 10 of 11 Neuroligin-2 and Dentate Gyrus Excitability

d

Jedlicka et al.

Chih B, Engelman H, Scheiffele P. 2005. Control of excitatory and inhibitory synapse formation by neuroligins. Science. 307:1324--1328. Chubykin AA, Atasoy D, Etherton MR, Brose N, Kavalali ET, Gibson JR, Sudhof TC. 2007. Activity-dependent validation of excitatory versus inhibitory synapses by neuroligin-1 versus neuroligin-2. Neuron. 54:919--931. Cline H. 2005. Synaptogenesis: a balancing act between excitation and inhibition. Curr Biol. 15:R203--R205. Cooke SF, Wu JQ, Plattner F, Errington M, Rowan M, Peters M, Hirano A, Bradshaw KD, Anwyl R, Bliss TVP, et al. 2006. Autophosphorylation of alpha CaMKII is not a general requirement for NMDA receptordependent LTP in the adult mouse. J Physiol. 574:805--818. Coulter DA, Carlson GC. 2007. Functional regulation of the dentate gyrus by GABA-mediated inhibition. Prog Brain Res. 163:235--243. Craig AM, Kang Y. 2007. Neurexin-neuroligin signaling in synapse development. Curr Opin Neurobiol. 17:43--52. Davies CH, Starkey SJ, Pozza MF, Collingridge GL. 1991. GABA autoreceptors regulate the induction of LTP. Nature. 349:609--611. Davison AP, Morse TM, Migliore M, Shepherd GM, Hines ML. 2004. Semi-automated population of an online database of neuronal models (ModelDB) with citation information, using PubMed for validation. Neuroinformatics. 2:327--332. Dong N, Qi J, Chen G. 2007. Molecular reconstitution of functional GABAergic synapses with expression of neuroligin-2 and GABAA receptors. Mol Cell Neurosci. 35:14--23. Feng J, Schroer R, Yan J, Song W, Yang C, Bockholt A, Cook EH, Jr., Skinner C, Schwartz CE, Sommer SS. 2006. High frequency of neurexin 1beta signal peptide structural variants in patients with autism. Neurosci Lett. 409:10--13. Fletcher EL, Koulen P, Wassle H. 1998. GABAA and GABAC receptors on mammalian rod bipolar cells. J Comp Neurol. 396:351--365. Freund TF, Katona I. 2007. Perisomatic inhibition. Neuron. 56:33--42. Fritschy JM. 2008. Epilepsy, E/I balance and GABA(A) receptor plasticity. Front Mol Neurosci. 1:5. Fu Z, Vicini S. 2009. Neuroligin-2 accelerates GABAergic synapse maturation in cerebellar granule cells. Mol Cell Neurosci. 42:45--55. Garber K. 2007. Autism’s cause may reside in abnormalities at the synapse. Science. 317:190--191. Gibson JR, Huber KM, Su¨dhof TC. 2009. Neuroligin-2 deletion selectively decreases inhibitory synaptic transmission originating from fast-spiking but not from somatostatin-positive interneurons. J Neurosci. 29:13883--13897. Graf ER, Zhang X, Jin SX, Linhoff MW, Craig AM. 2004. Neurexins induce differentiation of GABA and glutamate postsynaptic specializations via neuroligins. Cell. 119:1013--1026. Heine M, Thoumine O, Mondin M, Tessier B, Giannone G, Choquet D. 2008. Activity-independent and subunit-specific recruitment of functional AMPA receptors at neurexin/neuroligin contacts. Proc Natl Acad Sci U S A. 105:20947--20952. Hines ML, Carnevale NT. 1997. The NEURON simulation environment. Neural Comput. 9:1179--1209. Hines ML, Morse T, Migliore M, Carnevale NT, Shepherd GM. 2004. ModelDB: a database to support computational neuroscience. J Comput Neurosci. 17:7--11. Hines RM, Wu L, Hines DJ, Steenland H, Mansour S, Dahlhaus R, Singaraja RR, Cao X, Sammler E, Hormuzdi SG, et al. 2008. Synaptic imbalance, stereotypies, and impaired social interactions in mice with altered neuroligin 2 expression. J Neurosci. 28:6055--6067. Hoon M, Bauer G, Fritschy JM, Moser T, Falkenburger BH, Varoqueaux F. 2009. Neuroligin 2 controls the maturation of GABAergic synapses and information processing in the retina. J Neurosci. 29:8039--8050. Jamain S, Quach H, Betancur C, Rastam M, Colineaux C, Gillberg IC, Soderstrom H, Giros B, Leboyer M, Gillberg C, et al. 2003. Mutations of the X-linked genes encoding neuroligins NLGN3 and NLGN4 are associated with autism. Nat Genet. 34:27--29. Jedlicka P, Deller T, Schwarzacher SW. 2010. Computational modeling of GABA(A) receptor-mediated paired-pulse inhibition in the dentate gyrus. J Comput Neurosci. doi:10.1007/s10827-010-0214-y. Jedlicka P, Papadopoulos T, Deller T, Betz H, Schwarzacher SW. 2009. Increased network excitability and impaired induction of long-term

Downloaded from cercor.oxfordjournals.org at Universitaetsbibliothek Johann Christian Senckenberg on December 14, 2010

review, see Persico and Bourgeron 2006; Garber 2007; Su¨dhof 2008). Furthermore, mutations in the NL2-binding partner neurexin 1 have also been associated with autism (Feng et al. 2006; Szatmari et al. 2007). Although no signs of disturbed social behavior have been found in NL2 KO mice so far, they display increased anxiety which is often observed in autism-spectrum disorders (Blundell et al. 2009). In addition, NL2deficient mice exhibit decreased inhibition which has been suggested to underlie susceptibility to seizures seen in autistic individuals (Rubenstein and Merzenich 2003). In summary, our work provides insights into NL-dependent mechanisms of synapse stabilization and thus may help to link synaptic dysfunctions to cognitive diseases.

Pouille FR, Watkinson O, Scanziani M, Trevelyan A. 2008. Differential control of pyramidal neurons’ dynamic range by somatic and dendritic inhibition. Neuroscience 2008, Washington, Poster 136.11/F8. Poulopoulos A, Aramuni G, Meyer G, Soykan T, Hoon M, Papadopoulos T, Zhang M, Paarmann I, Fuchs C, Harvey K, et al. Forthcoming 2009. Neuroligin 2 drives postsynaptic assembly at inhibitory synapses through gephyrin and collybistin. Neuron. 63:628--642. Prange O, Wong TP, Gerrow K, Wang YT, El-Husseini A. 2004. A balance between excitatory and inhibitory synapses is controlled by PSD-95 and neuroligin. Proc Natl Acad Sci U S A. 101: 13915--13920. Rubenstein JL, Merzenich MM. 2003. Model of autism: increased ratio of excitation/inhibition in key neural systems. Genes Brain Behav. 2:255--267. Santhakumar V, Aradi I, Soltesz I. 2005. Role of mossy fiber sprouting and mossy cell loss in hyperexcitability: a network model of the dentate gyrus incorporating cell types and axonal topography. J Neurophysiol. 93:437--453. Schweizer C, Balsiger S, Bluethmann H, Mansuy IM, Fritschy JM, Mohler H, Lu¨scher B. 2003. The gamma 2 subunit of GABA(A) receptors is required for maintenance of receptors at mature synapses. Mol Cell Neurosci. 24:442--450. Sloviter RS. 1991. Feedforward and feedback inhibition of hippocampal principal cell activity evoked by perforant path stimulation: GABAmediated mechanisms that regulate excitability in vivo. Hippocampus. 1:31--40. Su¨dhof TC. 2008. Neuroligins and neurexins link synaptic function to cognitive disease. Nature. 455:903--911. Szatmari P, Paterson AD, Zwaigenbaum L, et al. 2007. Mapping autism risk loci using genetic linkage and chromosomal rearrangements. Nat Genet. 39:319--328. Thomas MJ, Mameli M, Carta M, Valenzuela CF, Li PK, Partridge LD. 2005. Neurosteroid paradoxical enhancement of paired-pulse inhibition through paired-pulse facilitation of inhibitory circuits in dentate granule cells. Neuropharmacology. 48:584--596. Varoqueaux F, Aramuni G, Rawson RL, Mohrmann R, Missler M, Gottmann K, Zhang W, Sudhof TC, Brose N. 2006. Neuroligins determine synapse maturation and function. Neuron. 51:741--754. Varoqueaux F, Jamain S, Brose N. 2004. Neuroligin 2 is exclusively localized to inhibitory synapses. Eur J Cell Biol. 83:449--456. Wigstro¨m H, Gustafsson B. 1983. Facilitated induction of hippocampal long-lasting potentiation during blockade of inhibition. Nature. 301:603--604. Winkels R, Jedlicka P, Weise FK, Schultz C, Deller T, Schwarzacher SW. 2009. Reduced excitability in the dentate gyrus network of betaIVspectrin mutant mice in vivo. Hippocampus. 19:677--686. Ylisaukko-oja T, Rehnstrom K, Auranen M, Vanhala R, Alen R, Kempas E, Ellonen P, Turunen JA, Makkonen I, Riikonen R, et al. 2005. Analysis of four neuroligin genes as candidates for autism. Eur J Hum Genet. 13:1285--1292. Zappone CA, Sloviter RS. 2004. Translamellar disinhibition in the rat hippocampal dentate gyrus after seizure-induced degeneration of vulnerable hilar neurons. J Neurosci. 24:853--864. Zucker RS, Regehr WG. 2002. Short-term synaptic plasticity. Annu Rev Physiol. 64:355--405.

Cerebral Cortex Page 11 of 11

Downloaded from cercor.oxfordjournals.org at Universitaetsbibliothek Johann Christian Senckenberg on December 14, 2010

potentiation in the dentate gyrus of collybistin-deficient mice in vivo. Mol Cell Neurosci. 41:94--100. Jedlicka P, Schwarzacher SW, Winkels R, Kienzler F, Frotscher M, Bramham CR, Schultz C, Bas OC, Deller T. 2009. Impairment of in vivo theta-burst long-term potentiation and network excitability in the dentate gyrus of synaptopodin-deficient mice lacking the spine apparatus and the cisternal organelle. Hippocampus. 19: 130--140. Jones MW, Errington ML, French PJ, Fine A, Bliss TVP, Garel S, Charnay P, Bozon B, Laroche S, Davis S. 2001. A requirement for the immediate early gene Zif268 in the expression of late LTP and longterm memories. Nat Neurosc. 4:289--296. Kraushaar U, Jonas P. 2000. Efficacy and stability of quantal GABA release at a hippocampal interneuron-principal neuron synapse. J Neurosci. 20:5594--5607. Lambert NA, Wilson WA. 1994. Temporally distinct mechanisms of usedependent depression at inhibitory synapses in the rat hippocampus in vitro. J Neurophysiol. 72:121--130. Laumonnier F, Bonnet-Brilhault F, Gomot M, Blanc R, David A, Moizard MP, Raynaud M, Ronce N, Lemonnier E, Calvas P, et al. 2004. X-linked mental retardation and autism are associated with a mutation in the NLGN4 gene, a member of the neuroligin family. Am J Hum Genet. 74:552--557. Lawson-Yuen A, Saldivar JS, Sommer S, Picker J. 2008. Familial deletion within NLGN4 associated with autism and Tourette syndrome. Eur J Hum Genet. 16:614--618. Levinson JN, Chery N, Huang K, Wong TP, Gerrow K, Kang R, Prange O, Wang YT, El-Husseini A. 2005. Neuroligins mediate excitatory and inhibitory synapse formation: involvement of PSD-95 and neurexin1beta in neuroligin-induced synaptic specificity. J Biol Chem. 280:17312--17319. Levinson JN, El-Husseini A. 2005. Building excitatory and inhibitory synapses: balancing neuroligin partnerships. Neuron. 48:171--174. Lise´ MF, El-Husseini A. 2006. The neuroligin and neurexin families: from structure to function at the synapse. Cell Mol Life Sci. 63: 1833--1849. Lomo T. 2009. Excitability changes within transverse lamellae of dentate granule cells and their longitudinal spread following orthodromic or antidromic activation. Hippocampus. 19:633--648. Moser EI. 1996. Altered inhibition of dentate granule cells during spatial learning in an exploration task. J Neurosci. 16:1247--1259. Papadopoulos T, Korte M, Eulenburg V, Kubota H, Retiounskaia M, Harvey RJ, Harvey K, O’Sullivan GA, Laube B, Hulsmann S, et al. 2007. Impaired GABAergic transmission and altered hippocampal synaptic plasticity in collybistin-deficient mice. EMBO J. 26:3888--3899. Patrizi A, Scelfo B, Viltono L, Briatore F, Fukaya M, Watanabe M, Strata P, Varoqueaux F, Brose N, Fritschy JM, et al. 2008. Synapse formation and clustering of neuroligin-2 in the absence of GABAA receptors. Proc Natl Acad Sci U S A. 105:13151--13156. Paulsen O, Moser EI. 1998. A model of hippocampal memory encoding and retrieval: GABAergic control of synaptic plasticity. Trends Neurosci. 21:273--278. Persico AM, Bourgeron T. 2006. Searching for ways out of the autism maze: genetic, epigenetic and environmental clues. Trends Neurosci. 29:349--358. Polleux F, Lauder JM. 2004. Toward a developmental neurobiology of autism. Ment Retard Dev Disabil Res Rev. 10:303--317.