Synapsins (Syn I, Syn II, and Syn III) are a family of synaptic vesicle phosphoproteins regulating synaptic transmission and plasticity. SYN1/2 genes have been identified as major epilepsy susceptibility genes in humans and synapsin I/II/III triple knockout (TKO) mice are epileptic. However, excitatory and inhibitory synaptic transmission and short-term plasticity have never been analyzed in intact neuronal circuits of TKO mice. To clarify the generation and expression of the epileptic phenotype, we performed patch-clamp recordings in the CA1 region of acute hippocampal slices from 1-month-old presymptomatic and 6-month-old epileptic TKO mice and age-matched controls. We found a strong imbalance between basal glutamatergic and γ-aminobutyric acid (GABA)ergic transmission with increased evoked excitatory postsynaptic current and impaired evoked inhibitory postsynaptic current amplitude. This imbalance was accompanied by a parallel derangement of short-term plasticity paradigms, with enhanced facilitation of glutamatergic transmission in the presymptomatic phase and milder depression of inhibitory synapses in the symptomatic phase. Interestingly, a lower tonic GABAA current due to the impaired GABA release is responsible for the more depolarized resting potential found in TKO CA1 neurons, which makes them more susceptible to fire. All these changes preceded the appearance of epilepsy, indicating that the distinct changes in excitatory and inhibitory transmission due to the absence of Syns initiate the epileptogenic process.
The dissection of the genetic basis of epilepsy has been so far very difficult due to the variable phenotypes and numerous genes involved (Noebels 2003). Animal models offer important clues to the elucidation of the pathogenetic role of inherited or de novo mutations affecting neuron-specific genes. In addition to mutations affecting membrane excitability, a large number of potential epilepsy genes involved in neural development, synaptogenesis and synaptic transmission have been uncovered (Noebels 2003; Steinlein 2004; Cavalleri et al. 2007). Although limited data exist in human, the large protein families operating in synaptic vesicle (SV) trafficking and neurotransmitter release include several potential epileptic genes.
Notwithstanding the large number of SV genes inactivated in animal models, only few mutants exhibit an epileptic phenotype, namely knockout (KO) mice for members of the synapsin (Syn) and SV2 families of SV proteins (Noebels 2003). Interestingly, a form of familial syndromic epilepsy characterized by a nonsense mutation in the SYN1 gene present in all affected family members was initially reported (Garcia et al. 2004). Very recently, an additional nonsense and several missense mutations in the SYN1 gene associated with epilepsy and/or autism were discovered and found to be implicated in synaptic dysfunctions (Fassio et al. 2011).
Synapsin I, II, and III are major SV-specific phosphoproteins implicated in neural development, synaptic transmission and plasticity (Cesca et al. 2010). In mature neurons, Syns control SV trafficking between the reserve pool (RP) and the readily releasable pool (RRP) in a phosphorylation-dependent fashion (Chi et al. 2001, 2003; Menegon et al. 2006; Messa et al. 2010; for review, see Cesca et al. 2010). Moreover, Syns play a role in the final postdocking steps of exocytosis, including SV priming and fusion (Hilfiker et al. 1998, 2005; Sun et al. 2006; Baldelli et al. 2007; Chiappalone et al. 2009).
Mice lacking Syn I, Syn II, Syn I/II, and Syn I/II/III (but not Syn III; Feng et al. 2002) are normal at birth, but, at the expression peak of Syn I/II in vivo (2–3 months; Lohmann et al. 1978; Bogen et al. 2009), they start showing epileptic seizures whose severity increases with age and is proportional to the number of inactivated SYN genes (Li et al. 1995; Rosahl et al. 1995; Gitler et al. 2004; Etholm and Heggelund 2009; Boido et al. 2010; Ketzef et al. 2011). Indeed, acute slices from Syn I/II/III triple KO (TKO) mice display intense interictal and ictal activities evoked by 4-aminopyridine, which precede the appearance of the epileptic phenotype and become more severe with age (Boido et al. 2010).
To better clarify the role of Syns in neurotransmission and epileptogenesis, we investigated excitatory and inhibitory transmission in acute slices of 1-month-old presymptomatic and 6-month-old symptomatic TKO mice. The results indicate that Syns are involved in the fine balance between excitatory and inhibitory synapses and that their deletion triggers a strong imbalance in excitatory and inhibitory transmission, which precedes the appearance of epilepsy. Such imbalance is associated with a deficit in tonic γ-aminobutyric acid (GABA) current and an increased excitability of CA1 pyramidal neurons.
Materials and Methods
Homozygous Syn TKO mice (Gitler et al. 2004) were kindly provided by Drs Hung-Teh Kao (Brown University, Providence, RI) and Paul Greengard (The Rockefeller University, New York, NY). TKO mice were rederived on a C57BL/6J background (Charles River, Calco, Italy), obtaining single and multiple Syn KO strains up to the Syn TKO and matched wild-type (WT) mice. One-month-old (young) and 6-month-old (adult) TKO mice of either sex together with age- and sex-matched littermates were used. All experiments were carried out in accordance with the guidelines established by the European Communities Council (Directive of 24 November 1986) and approved by the National Council on Health and Animal Care (authorization ID 227, prot. 4127 25 March 2008).
Horizontal corticohippocampal slices were prepared as shown by Bischofberger et al. (2006). Mice were anaesthetized with halothane (Sigma-Aldrich, Milan, Italy) and decapitated; the brain was quickly removed and immersed in an ice-cold “cutting” solution composed of (in millimolars) 125 NaCl, 25 NaHCO3, 25 glucose, 2.5 KCl, 1.25 NaH2PO4, 1 CaCl2, 2 MgCl2, 0.4 ascorbic acid, 2 NaPyruvate, 3 myo-inositol, and saturated with 95% O2 and 5% CO2. Horizontal corticohippocampal slices (250 μm thick) were cut using an HM 650 vibratome (Microm International GmbH, Walldorf, Germany) in ice-cold oxygenated cutting solution. Slices were first incubated in cutting solution at 35 °C for 30 min and then transferred to a “submerged” recording chamber in a “standard recording” solution composed of (in millimolars) 125 NaCl, 25 NaHCO3, 25 glucose, 2.5 KCl, 1.25 NaH2PO4, 2 CaCl2, and 1 MgCl2. This solution was constantly oxygenated, maintained at 33 °C, and superfused at a rate of 1.5 mL/min.
Bicuculline methiodide (BMI; 30 μM) and D-2-amino-5-phosphonopentanoic acid (APV; 100 μM) were added to the standard recording solution to record excitatory neurotransmission, whereas APV (100 μM) and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; 10 μM) were applied to discern inhibitory neurotransmission. To evaluate the tonic current, BMI was applied to neurons that were recorded in the presence of CNQX/APV and (2S)-3-[[(1S)-1-(3,4-dichlorophenyl) ethyl] amino-2-hydroxypropyl] (phenylmethyl) phosphinic acid (CGP 55845, 5 μM). All drugs were obtained from Tocris Bioscience (Bristol, UK).
Whole-cell current-clamp and voltage-clamp recordings from CA1 pyramidal neurons of hippocampal slices were performed with glass pipettes (∼3.8 to 5 MΩ) pulled from borosilicate glass (Kimble Glass Inc., Vineland, NJ) and filled with the following intracellular solution (in millimolars): 126 Kgluconate, 4 NaCl, 1 MgSO4, 0.02 CaCl2, 0.1 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid, 15 glucose, 5 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, 3 Adenosine-5′-triphosphate, 0.1 Guanosine-5′-triphosphate. In voltage-clamp experiments, lidocaine (QX314; 10 mM; Sigma-Aldrich) was added to the intracellular solution. Pyramidal neurons of CA1 region of hippocampus were visualized with a ×40 water immersion objective and an infrared camera. Voltage-clamp recordings were conducted at the holding potential of −70 mV for excitatory neurotransmission and −40 mV for inhibitory neurotransmission. Electrophysiological data were acquired using the MultiClamp 700B amplifier and the pClamp 9.2 software (Axon Instruments, Molecular Devices, Sunnyvale, CA). Data were acquired at 10 kHz and filtered at 2 kHz; series resistance was monitored throughout the experiment and whenever it changed more than 15%, the recording was not included in the analysis. Evoked excitatory postsynaptic currents (eEPSCs) were evoked in pyramidal neurons by electrical stimulation of Schaffer collaterals in the stratum radiatum (see inset of Fig. 1A). Evoked inhibitory postsynaptic currents (eIPSCs) were evoked by electrical stimulation of an area between the stratum radiatum and the stratum pyramidale (see inset of Fig. 1B), which was most effective in eliciting inhibitory currents in pyramidal neurons (Freund and Buzsaki 1996). Stimulation was performed by administering test pulses of 0.25 ms duration at 0.1 Hz, using a bipolar tungsten electrode. For the results shown in Figure 1, stimulus intensity was varied from the lowest intensity able to evoke ePSCs to an intensity eliciting the maximal ePSC amplitude. For the remaining data presented in the paper, the stimulation was adjusted at an intensity able to evoke 2/3 of maximal ePSC amplitude (except for the experiments of cumulative amplitude analysis, see below). Spontaneous and miniature EPSCs and IPSCs were also recorded.
Postsynaptic Current Analysis
The amplitude, decay, and rise time of eEPSCs and eIPSCs were calculated by using the Clampfit 9.0 application of pClamp 9.2, MiniAnalysis (Synaptosoft, GA) and a homemade software developed in the R-CRAN environment (http://www.R-project.org). The eIPSC decays were fitted with a biexponential curve (A1 * exp(−t/τ1) + A2 * exp(−t/τ2)) and the weighted decay constant (τW = (A1 * τ1 + A2 * τ2)/(A1 + A2)) was later computed.
Estimation of Pr by Cumulative Amplitude Analysis
The release probability Pr was estimated applying the cumulative amplitude analysis, previously used on cultured neurons (Rosenmund and Stevens 1996; Baldelli et al. 2005, 2007), Calix of Held synapses (Schneggenburger et al. 2002), and hippocampal slices (Wesseling and Lo 2002). High-frequency stimulation (2 s at 40 and 20 Hz, for excitatory and inhibitory synapses, respectively) was applied to presynaptic fibers with a bipolar tungsten electrode. The analysis assumes that the depression during the steady state induced by the train is limited by a constant recycling of SVs and that equilibrium is present between released and recycled SVs. The extracellular stimulation lasted 0.25 ms and its intensity was adjusted as the minimal intensity needed to elicit successful synaptic transmission in the 100% of the stimuli during low frequency stimulation. The few cells displaying a short-lived facilitation phase before severe depression were excluded from the analysis. Cumulative amplitudes of PSCs during the train were normalized to the mean amplitude of the first responses recorded in neurons from young WT mice and linearly fitted. The number of data points to include in the fit of the steady state phase was evaluated by calculating, for each cell, the best linear fit which included the maximal number of data points starting from the last data point. According to this procedure, the Y-intercept yielded a nondimensional estimation of the changes in the size of the ensemble RRP (i.e., the total RRP of the synapses activated by the stimulus) relative to young WT and the ratio of the first PSC evoked by the stimulation train to the size of cumulative RRP yielded an estimation of Pr. While the relative RRP changes can be potentially affected by differences in the number of activated fibers, the estimation of Pr is substantially independent of changes in the number of stimulated fibers. The constant amplitude of the electrical artifacts associated with each stimulus indicates that the bipolar electrodes maintained their efficacy in eliciting action potentials (APs) over the course of each stimulation train. Moreover, the very low PSC amplitude fluctuations observed during long-lasting stimulation at 0.1 Hz (less than 10%) indicate that, within the same slice, the number of stimulated fibers remained constant for the entire duration of the stimulation protocol.
Paired stimuli were delivered at increasing interstimulus intervals (ISIs; 10–500 and 25–2000 ms for excitatory and inhibitory synapses, respectively). Paired-pulse ratio (PPR) was calculated as the ratio of the amplitude of the second to the first response (I2/I1). The current traces used for the analysis were averages of 12 consecutive responses repeated every 10 s.
The post-tetanic response was measured 10 s after the end of a high-frequency train (40 Hz for 2 s for excitatory synapses and at 20 Hz for 2 s for inhibitory synapses) by lowering the stimulation to 0.1 Hz. Each post-tetanic response was normalized to the mean baseline value obtained before the application of the train stimulation and was then averaged with others.
Synaptic Depression and Recovery
For each experiment, the baseline was calculated by averaging the amplitude of ePSCs obtained in response to stimulation at 0.1 Hz for 5 min. Then, high-frequency repetitive stimulation protocols were applied (20 Hz for 20 s and 10 Hz for 30 s for excitatory and inhibitory synapses, respectively) at the end of which the stimulation frequency was returned to 0.1 Hz to evaluate the rate and extent of recovery from synaptic depression. Each cell was normalized to the mean baseline value and averaged with other cells.
Tonic Current Analysis
To evaluate the tonic GABA current, BMI was applied in voltage-clamp configuration at a holding potential of −50 mV in the presence of APV/CNQX and CGP 55845. The maximal amplitude of the tonic current was estimated by applying BMI in the presence of saturating concentrations of exogenous GABA (5 μM). The current shift was calculated as the mean current during 1 min of recording, starting 4 min after BMI application. The real membrane potential was also monitored in the current-clamp configuration before and after the application of BMI and used to calculate the percentage of depolarization induced by the application of BMI and the mean membrane potential reached after the blockade of tonic current.
Current-clamp recordings were used to estimate the resting membrane potential. Firing activity was evoked by injecting into the neurons increasing current steps (30 pA) from a membrane potential kept constantly at −70 mV by injection of a holding current. The voltage threshold needed to induce regular firing or bursting activity was determined from the potential reached during the current injection step that led to the spiking activity. Bursts were defined as multispike responses to current injection followed by a silent period (number of spikes per burst ≥ 2; interspike interval ≤ 15 ms). The mean intraburst frequency was calculated as the ratio of the number of APs in the burst to the duration of the same burst.
We adopted the mechanistic mathematical model of neurotransmitter release developed by Sun et al. (2005) that explores the contributions of RRP size, Pr, and Ca2+-dependent depression or facilitation of synaptic release to different forms of short-term plasticity. According to this model, the average release probability P(t) (that is determined experimentally in the cumulative analysis), expressed in terms of the probability of releasing one SV (α) and the number of SVs (n) constituting the RRP, can be written as
Young WT and TKO mice were deeply anesthetized with urethan and intracardially perfused with 4% paraformaldehyde in phosphate-buffered saline (PBS). Brains were removed from the skull and postfixed overnight in the same fixative, equilibrated in 30% sucrose, and fast frozen by immersion in isopentane precooled at −80 °C. Fifty-micrometer cryosections were made using an HM 450 sliding microtome coupled to a fast freezing unit KS 34 (Microm International GmbH). Floating sections were blocked in 5% horse serum, 0.3% Triton X100 in PBS (blocking buffer) for 2 h at room temperature (RT), incubated with primary antibodies for 48 h at 4 °C, and with fluorescently conjugated secondary antibodies for 3 h at RT under constant agitation. Both primary and secondary antibodies were diluted in blocking buffer. Slices were then incubated in 3.3 mg/mL Hoechst-33342 (Sigma) in PBS for 10 min at RT, washed in PBS, and mounted in Mowiol 4-88 (Dako, Milan, Italy). The following primary antibodies were used: monoclonal antineuronal class III β-tubulin (MMS-435P, Covance, NJ), polyclonal anti-GABAA receptor subunit α5 (ab10098, Abcam, UK), polyclonal antibody anti-GABAA receptor subunit δ (kind gift of Dr Sieghart, Center for Brain Research, Medical University Wien, Austria). Fluorescently conjugated secondary antibodies were from Molecular Probes (Invitrogen).
Statistical analysis was performed with Origin 7.0 (OriginLab Corporation, Northampton, MA), Sigmastat 3.5 (Systat Software Inc., Chicago, IL) and SPSS 8.0 (SPSS, Chicago, IL) using the following tests: analysis of variance (ANOVA) for repeated measures, two-way ANOVA, Bonferroni's or least significant difference multiple comparison tests, or unpaired Student's t-test. Data were expressed as mean values ± standard error of the mean (SEM). Figures were edited with Corel Draw 12 (Corel Corporation, Ottawa, Canada).
A Strong Imbalance between Basal Glutamatergic and GABAergic Transmission Exists in the Hippocampus of TKO Mice and Precedes the Appearance of Epilepsy
The evoked release of glutamate and GABA was investigated in acute hippocampal slices from TKO mice at 2 distinct ages, namely before the appearance of the first signs of epilepsy (1-month-old TKO mice) and after an overt epileptic phenotype had occurred (6-month-old TKO mice) and compared with age-matched WT mice. To evaluate release, afferent fibers were stimulated at increasing intensities, and the PSC amplitude was plotted against the stimulation intensity (Fig. 1A–D). Strikingly, opposite changes in the ePSC amplitude were observed at excitatory and inhibitory synapses. Excitatory synapses from presymptomatic TKO mice showed significantly higher amplitudes, particularly at high stimulation intensities, than excitatory synapses from age-matched WT mice (Fig. 1A). On the contrary, the input–output relationship at inhibitory synapses demonstrated that eIPSCs in CA1 pyramidal cells from presymptomatic TKO mice have significantly lower amplitude than eIPSCs recorded in age-matched control neurons, particularly at high intensity of stimulation (Fig. 1B). Such imbalance of excitatory and inhibitory transmission persisted in the adult age, accompanying the development of epilepsy (Fig. 1C and Fig. 1D for glutamatergic and GABAergic synapses, respectively). These data indicate that an increase in glutamatergic and a decrease in GABAergic neurotransmission are present well before the appearance of the epileptic phenotype in the TKO hippocampus and can be at the basis of the epileptogenetic process.
The analysis of the kinetics of ePSCs also revealed changes, which were consistent with the imbalance in PSC amplitudes. In fact, while the rise time was not altered, a slower decay was observed in eEPSCs of CA1 pyramidal neurons from presymptomatic TKO mice, which persisted in the adult age, indicating a further increase in the charge transfer in TKO excitatory synapses (Fig. 1E,G). On the contrary, the analysis of eIPSC kinetics revealed no significant differences in the rise times at both ages, while the weighted decay time constant (τW = [A1 * τ1 + A2 * τ2]/[A1 + A2], see Materials and Methods) was significantly shorter in symptomatic TKO neurons (Fig. 1F,H), further contributing to the impairment in GABAergic transmission. On the other hand, a shorter τW may also reflect a lower GABA release in response to a single AP (Petrini et al. 2011).
The Imbalance between Glutamatergic and GABAergic Transmission Is due to Opposite Changes in the Ensemble RRP Size
Syn I, Syn II, Syn I/II, and Syn I/II/III KO mice showed a reduced number of SVs in both excitatory and inhibitory presynaptic terminals (Rosahl et al. 1995; Gitler et al. 2004; Siksou et al. 2007), and functional analysis of RRP in cultured neurons of Syn I KO and TKO identified a decreased size of inhibitory RRP in the absence of changes in the average release probability Pr (Gitler et al. 2004; Baldelli et al. 2007; Chiappalone et al. 2009). Moreover, a higher excitatory RRP size was found in Syn I KO glutamatergic synapses (Chiappalone et al. 2009). We estimated Pr in young and adult WT and TKO slices by applying the cumulative amplitude analysis to presymptomatic and symptomatic excitatory and inhibitory synapses of acute brain slices. To this aim, a high-frequency stimulation protocol sufficient to fully empty the RRP (2 s at 40 and 20 Hz for excitatory and inhibitory neurons, respectively) was applied to evoke PSCs (Fig. 2A,B and Fig. 2F,G for excitatory and inhibitory neurons, respectively).
The cumulative amplitude analysis, performed on the responses of excitatory synapses, revealed that excitatory Pr was not significantly affected among the experimental groups (Fig. 2D). No differences were also observed in the amplitude of miniature EPSCs (9.33 ± 0.25 and 9.57 ± 0.20 pA for young WT and TKO neurons, respectively; 9.21 ± 0.19 and 9.05 ± 0.24 pA for adult WT and TKO neurons, respectively), indicating that quantal size was not affected in the various experimental groups. On the other hand, neurons from presymptomatic TKO mice showed an increase in the excitatory ensemble RRP with respect to age-matched WT (Fig. 2E), an effect that increased with age and whose magnitude roughly corresponded to the relative increase in eEPSC amplitude (Fig. 2C).
Cumulative analysis performed at inhibitory synapses revealed that no significant changes were observed in the inhibitory Pr (Fig. 2I), in the absence of changes in the amplitude of miniature IPSCs (9.95 ± 0.21 and 9.22 ± 0.54 pA for young WT and TKO neurons, respectively; 8.34 ± 0.34 and 8.97 ± 0.47 pA for adult WT and TKO neurons, respectively). However, inhibitory synapses from presymptomatic TKO mice exhibited opposite changes with respect to the glutamatergic transmission, with a significant decrease in the ensemble RRP, which persisted in adult epileptic mice (Fig. 2L) and paralleled the decrease in eIPSC amplitude (Fig. 2H).
A Similar Imbalance in Paired-Pulse Responses at Glutamatergic and GABAergic Synapses Exists in the Hippocampus of Presymptomatic, but Not Epileptic, TKO Mice
Short-term plasticity paradigms, such as facilitation, depression, or potentiation, profoundly affect network activity and filtering properties (Abbott and Regehr 2004). Thus, the response to paired stimuli administered at increasing ISIs was assessed in both excitatory and inhibitory synapses of young and adult TKO and WT mice (Fig. 3). Previous studies identified an increased paired-pulse facilitation (PPF) of excitatory transmission in the CA1 region of Syn I KO mice (but not in Syn II or Syn III KO mice; Rosahl et al. 1993, 1995; Feng et al. 2002), while patch-clamp studies found no differences in PPR of inhibitory synapses between WT and Syn I KO neurons (Baldelli et al. 2007; Chiappalone et al. 2009).
Analysis of PPF in excitatory synapses of presymptomatic TKO mice revealed a higher facilitation in the ISI range 10–250 ms (Fig. 3A), in agreement with previous extracellular recordings in Syn I mice (Rosahl et al. 1993, 1995). However, this enhanced facilitation was lost in adult mice, and the PPF profiles of adult TKO and WT excitatory synapses became superimposable (Fig. 3C) because of a significant age-dependent decrease of PPF in symptomatic TKO mice (two-way ANOVA and Bonferroni's multiple comparison test, P < 0.05 at 50 ms ISI). Thus, PPF in excitatory synapses of adult TKO mice resembles that reported for Syn II KO mice, where no differences in PPF were detected (Rosahl et al. 1995).
We then tested PPR at inhibitory synapses of hippocampal CA1 neurons from presymptomatic and symptomatic TKO mice by administering pairs of stimuli at increasing ISIs (25–2000 ms; Fig. 3B,D). The analysis of the PPR of eIPSCs in presymptomatic TKO mice revealed a significant increase of paired-pulse depression (PPD) at short ISI (25 and 50 ms; Fig. 3B) and no differences at longer ISIs. On the opposite, PPRs of eIPSCs in symptomatic TKO mice displayed a clear-cut switch of PPD behavior (Fig. 3D), with inhibitory TKO synapses displaying a significantly lower PPD over the whole ISI tested. This result was contributed by opposite age-dependent changes of PPR in WT and TKO mice, namely a PPR decrease in adult WT mice and a concomitant PPR increase in adult TKO mice (two-way ANOVA and Bonferroni's multiple comparison test, P < 0.05 at 50 ms ISI). This result indicates the occurrence of epilepsy-dependent changes that strengthen inhibitory transmission over aging and resemble the reduction of PPD found in other models of epilepsy (Wu and Leung 1997; Merlo et al. 2007; Inaba et al. 2009).
The Genotype- and Age-Dependent Changes in PPR Are Attributable to Changes in RRP and Ca2+ Sensitivity of Release
Changes in PPR are generally attributed to changes in the probability of release in response to the first of the paired stimuli. However, the changes in PPR detected at glutamatergic and GABAergic synapses of TKO mice (Fig. 3) were not associated with significant changes in release probability as analyzed by cumulative amplitude analysis (see Fig. 2). In order to account for the PPR changes, we performed a modeling analysis based on the mechanistic mathematical model of neurotransmitter release developed by Sun et al. (2005), which reproduces Ca2+-dependent depression or facilitation of release (Fig. 4A) and implements the experimental finding that the PPR not only is directly affected by the release probability but also depends on the RRP size. The latter dependence plays a critical role in synapses with small RRP as the hippocampal ones. In neurons from presymptomatic TKO mice, the results of our model indicated that a change in RRP size in the same range of that determined experimentally, rather than a change in Pr, can account for the higher facilitation at excitatory synapses (Fig. 4B and Table 1). However, the complete loss of PPF observed at the same synapses in adult TKO mice could not be easily explained on the same basis, since the changes in quantal parameters observed in presymptomatic mice were still present and even enhanced. To account for this age-dependent effect, the modeling approach predicted that an impaired Ca2+ sensitivity of TKO terminals upon development of epilepsy could underlie the lack of facilitation (Fig. 4D). The entry of Ca2+ participating in facilitation was δF = 0.004 and the remaining parameters were as in the original model (Sun et al. 2005): KF = 4 (mM), kmax = 0.03 (ms−1), k0 = 0.002 (ms−1), τD = 50 (ms), KD = 2 (mM), and R = 0.0001 (ms−1).
|WT young||TKO young||WT adult||TKO adult|
|k0||6.10−4||—||3.6 × 10−4||—|
|WT young||TKO young||WT adult||TKO adult|
|k0||6.10−4||—||3.6 × 10−4||—|
Note: The model includes an additional parameter (t-delay) respect to Sun et al. (2005) to account for the absence of PPF at low ISIs (see also Fig. 4A). The optimal parameters were in the same range of the original model except for the facilitation time constant of adult inhibitory synapses (see Results). The parameters that were kept identical during the simultaneous fit in WT and TKO data are indicated by the minus symbol (−). RRPs were normalized to the corresponding RRP value of young WT mice (35 and 50 SVs for excitatory and inhibitory synapses, respectively) to facilitate the comparison with Figure 2E,L.
Modeling of PPR at GABAergic synapses based on the quantal parameters determined in presymptomatic and symptomatic TKO mice also allowed to fit the experimental results of PPD found in young and adult WT and TKO mice. Cumulative amplitude analysis showed a decrease in RRP size, which persisted in adult animals, and was not accompanied by significant changes in the Pr. The inhibitory presymptomatic PPR was fitted by letting some of the parameters of the original model (Sun et al. 2005) vary. The model fitted well the observed differences in the RRP and the similar Pr values of WT and TKO inhibitory synapses, except for the initial attenuation of PPD (Fig. 4C and Table 1). We adopted the same hypothesis used for excitatory transmission in adult mice (see above) in modeling PPR from adult inhibitory synapses. To explain the higher PPR of TKO synapses with respect to WT synapses, we hypothesized an increased Ca2+ sensitivity. The effective difference in the PPR profile observed on the whole ISI range tested (25–2000 ms) was well reproduced, including the long-lasting effect of the facilitation mechanism, with a facilitation time constant set to 1500 ms (Fig. 4E and Table 1). In our model, we also took into account the intrinsic postsynaptic dynamics by introducing GABA receptor desensitization. Although the fitting of inhibitory PPRs did improve at short ISIs, the main conclusions of the modeling did not change (not shown).
Post-tetanic Potentiation in Glutamatergic Synapses Is Lost in Both Presymptomatic and Epileptic TKO Mice
Post-tetanic potentiation (PTP) is a well-known paradigm of short-term plasticity, which lasts for 30 s to several min following a tetanic stimulation and is predominantly generated by a Ca2+ accumulation within the presynaptic terminal (Xu-Friedman and Regehr 2004). It was previously shown that Syn II and Syn I/II KO mice, but not Syn I KO mice, display an impairment of PTP in Schaffer-to-CA1 excitatory synapses (Rosahl et al. 1995), while no data are available for Syn III and Syn I/II/III KO mice. To evaluate PTP at excitatory synapses, a high-frequency train (2 s at 40 Hz) was applied to the Schaffer collateral pathway. The responses were analyzed 10 s after the end of the train, upon return of the stimulation to 0.1 Hz. In WT mice, the excitatory PTP was present in both young and adult groups, whereas it was completely absent in both presymptomatic and symptomatic TKO neurons (Fig. 5A,C). On the other hand, inhibitory synapses of young TKO and WT mice did not exhibit significant changes in the eIPSC amplitudes after the high-frequency train (2 s at 20 Hz; Fig. 5B), and this lack of post-tetanic response was maintained in slices from adult mice of both genotypes (Fig. 5D).
The Age-Dependent Increase in Synaptic Depression of Excitatory Synapses Is Accelerated by Lack of Syns
To evaluate synaptic depression in glutamatergic synapses, the Schaffer collateral pathway was stimulated with a train of 400 stimuli delivered at 20 Hz. In line with previous results (Gitler et al. 2004), pyramidal neurons from presymptomatic TKO mice showed a transient facilitation followed by an intense and sustained depression, while age-matched WT synapses responded to the stimulation with a long-lasting facilitation (Fig 6A). In adulthood, the facilitation of WT mice became transient and was followed by a mild depression phase, whereas in adult TKO mice, the facilitation phase vanished and was substituted by a severe and profound depression (Fig 6C). These results confirm a key involvement of Syns in the response to high-frequency stimulation and in the presynaptic mechanisms counteracting synaptic depression. When the stimulation frequency was lowered to 0.1 Hz, excitatory synapses from symptomatic TKO mice showed a slower recovery from depression than age-matched WT synapses, while the recovery rate from depression did not significantly differ between presymptomatic and symptomatic TKO mice (Fig. 6A,C). Both the severity of depression and the slow recovery are likely attributable to the combination of impairment in SV recruitment to the RRP and decrease in the RP size previously demonstrated by electron microscopy in TKO terminals (Gitler et al. 2004; Siksou et al. 2007).
Previous studies on synaptic depression in inhibitory synapses from Syn KO mice reported a larger synaptic depression in primary neurons from Syn I KO mice (Baldelli et al. 2007) but no significant effects on the depression kinetics in TKO autaptic neurons (Gitler et al. 2004). To evaluate synaptic depression at inhibitory synapses of the CA1 region, the area between stratum radiatum and stratum pyramidale was stimulated with a train of 300 stimuli delivered at a frequency of 10 Hz. In young animals, no significant differences were detected between WT and TKO synapses (Fig. 6B), while in inhibitory synapses of adult TKO mice, depression was attenuated at the beginning of the train (first 10 s) but subsequently reached the same steady state level observed in age-matched WT mice (Fig. 6D). The early attenuation of depression is consistent with the appearance of a lower PPD in the same synapses of symptomatic TKO mice (see Fig. 3D). When the stimulation frequency was returned to 0.1 Hz, inhibitory synapses from both young and adult TKO mice displayed a recovery from depression closely similar to that observed in age-matched WT mice (Fig. 6B,D).
CA1 Pyramidal Neurons from TKO Mice Display an Age-Dependent Depolarization of the Membrane Potential due to a Markedly Decreased Tonic GABAA Current
To investigate the intrinsic excitability of pyramidal neurons, we analyzed the membrane potential of patched CA1 pyramidal cells in slices from young and adult WT and TKO mice. Presymptomatic TKO pyramidal neurons showed a significantly more depolarized membrane potential than that of age-matched WT neurons, a difference that became more pronounced with age (Fig. 7A).
Since TKO mice suffer from an impaired GABAergic transmission, we tested whether the depolarization was due to a reduced tonic GABA current. Thus, the GABAA receptor blocker BMI was applied in voltage-clamp configuration at a holding potential of −50 mV (Fig. 7B). Indeed, the amplitude of the tonic current in neurons from presymptomatic TKO mice was significantly smaller than in age-matched WT, and this difference persisted with age (Fig. 7C). Then, the membrane potential reached by the cell after the application of BMI was measured (Fig. 7D). Interestingly, in the absence of tonic current, the membrane potentials of TKO and WT neurons were not significantly different, underpinning the causal link between the downregulation of the tonic current and the depolarized state of CA1 pyramidal neurons of TKO mice. We next studied whether the impairment of the tonic current in presymptomatic TKO mice was due to the impaired GABA release and spillover or to a change in the density/activity of extrasynaptic GABAA receptors containing α5 or δ subunit (Scimemi et al. 2005; Glykys et al. 2008). When the maximal amplitude of the tonic current in CA1 pyramidal neurons was evaluated by application of BMI in the presence of saturating concentrations of exogenous GABA, no significant changes were observed in presymptomatic TKO mice with respect to age-matched WT neurons (Fig. 7E), indicating that the impaired GABA release observed in TKO slices is responsible for the decrease in the tonic current. Moreover, no apparent changes in α5 and δ subunit immunoreactivities were observed in the CA1 region of presymptomatic TKO mice as compared with age-matched WT mice (Fig. 7F,G).
Firing Is Increased in Bursting and Regular-Spiking CA1 Pyramidal Neurons in the Hippocampus of TKO Mice
To investigate whether the decreased tonic current and the depolarized state of the resting potential of CA1 pyramidal neurons affect their intrinsic discharge behavior, current-clamp experiments were conducted to elicit APs in the patched CA1 pyramidal cells in slices from young presymptomatic WT and TKO mice (Fig. 8). The injected current was increased to reach the AP threshold. Distinct firing behaviors (bursting vs. regular-firing neurons) were observed in the pyramidal cell population of WT mice (21 bursting neurons of 57, corresponding to 36.8%). The incidence of bursting cells in TKO slices was not significantly different from that of WT mice (14 bursting neurons of 43 corresponding to 32.5%; chi-square test, P > 0.5).
In the regular-spiking neuronal population (Fig. 8A), TKO neurons displayed a significantly increased input resistance (Fig. 8B), consistent with the decreased tonic current, and responded to increasing steps of injected current with a significantly higher firing frequency with respect to WT neurons (Fig. 8C). In the bursting population (Fig. 8D), the mean membrane threshold to elicit bursting activity was more hyperpolarized (Fig. 8E) and the mean frequency of APs within the burst was significantly higher (Fig. 8F) in TKO than in WT mice. Taken together, the data indicate that the complex imbalance in basal transmission and short-term plasticity of excitatory and inhibitory synapses impinging on CA1 pyramidal neurons, together with the decrease in the GABA tonic current, trigger a state of hyperexcitability, which could play a central role in the development of epilepsy.
Abnormalities in presynaptic function affecting neurotransmitter release and short-term plasticity have been identified at the basis of epilepsy and cognitive disorders, such as autism spectrum disorders (Ramocki and Zoghbi 2008; Sudhof 2008; Chao et al. 2010). One family of such presynaptic proteins, whose deletion or loss-of-function mutations generate an epileptic phenotype in mouse and man, are the synapsins (Garcia et al. 2004; Cavalleri et al. 2007; Fassio et al. 2011). Mice lacking Syn I, Syn II, Syn I/II, or Syn I/II/III are all prone to epileptic seizures starting at an age of 2–3 months (Rosahl et al. 1995; Gitler et al. 2004; Ketzef et al. 2011), a period of intense synapse maturation and refinement in which Syn I/II has already reached high expression levels (Lohmann et al. 1978; Bogen et al. 2009). A similar onset during childhood or adolescence is present in the epileptic patients affected by SYN1 mutations (Garcia et al. 2004).
Deletion of SYN Genes Alters the E/I Balance in Evoked Synaptic Transmission
Previous work revealed that eIPSCs were depressed in Syn I KO and TKO autaptic neurons, as well as in experimental models of temporal lobe epilepsy (Gitler et al. 2004; Baldelli et al. 2007; see, e.g., Hirsch et al. 1999). On the other hand, eEPSCs were found to be enhanced in Syn I KO or not affected in TKO neurons (Gitler et al. 2004; Chiappalone et al. 2009). Acute slices from TKO mice recapitulated these in vitro findings and the amplitude of evoked PSCs was decreased in GABAergic synapses and increased in glutamatergic synapses. These effects were present well before the appearance of the first seizures and were associated with a slower EPSC kinetics, which further contributed to the excitation/inhibition imbalance. Although the extent of inhibition is known to control postsynaptic excitability, both the increase in basal excitatory transmission and the decrease in basal inhibitory transmission seem to be present before the appearance of the epileptic phenotype, as previously observed in Syn I KO autaptic excitatory neurons (Chiappalone et al. 2009). Interestingly, the IPSC kinetics was not altered in the presymptomatic phase but was significantly shortened after the appearance of the epileptic phenotype, possibly representing a consequence of epileptogenesis and contributing to a further impairment of the charge transfer in inhibitory synapses.
While no changes in the average release probability and in the quantal size were found between the 2 genotypes, we found that opposite changes in the size of excitatory and inhibitory ensemble RRP fully accounted for the opposite changes in PSC amplitudes observed in inhibitory and excitatory synapses. As a higher excitability was observed in TKO excitatory neurons, it cannot be excluded that a larger number of fibers is recruited by stimulation and thus contributes to the relative increase in the excitatory ensemble RRP. However, previous results demonstrated that a similar increase in the strength of glutamatergic transmission in autaptic Syn I KO neurons was associated with a significant increase in the RRP size (Chiappalone et al. 2009). On the other hand, the impairment in the ensemble RRP size of GABAergic synapses appears to play a causal role in decreasing the strength of inhibitory transmission in TKO mice. Such RRP decrease is consistent with the decrease in the size of both RP and RRP in TKO terminals as shown by electron microscopy (Gitler et al. 2004; Siksou et al. 2007) and the decreased RRP size demonstrated by cumulative analysis in Syn I KO inhibitory neurons (Baldelli et al. 2007; Chiappalone et al. 2009).
Deletion of SYN Genes Causes Complex Impairments in Short-term Plasticity
Analysis of PPF showed an increased facilitation of excitatory transmission in presymptomatic TKO neurons, as previously observed in Syn I−/− mice (Rosahl et al. 1993, 1995). In the absence of genotype-specific changes in Pr, this is in apparent contrast with the notion that the extent of PPF is inversely related to the initial release probability (Xu-Friedman and Regehr 2004). However, the modeling approach reproducing Ca2+-dependent depression/facilitation of synaptic release indicated that the relative variations observed in the ensemble RRP size may be responsible for changes in facilitation, supporting the idea that the changes in the ensemble RRP size were not due to differences in the number of activated fibers. The same model also predicted that the normalization of PPF in adult TKO synapses, in spite of the persistent RRP increase, could be ascribed to a lower Ca2+dependency of release, which can arise from defects in signal transduction pathways impinging on presynaptic Ca2+ channels.
The more intense PPD at short ISIs of inhibitory synapses from pre-symptomatic TKO mice is also consistent, as predicted by the model, with a decreased RRP at the same Pr level. On the other hand, the complete switch of PPD changes in adult animals, in which inhibitory TKO synapses displayed a milder depression over a wide ISI range, could be explained by a higher Ca2+ sensitivity of release. A similar PPD reduction observed after hippocampal kindling was associated with a reduced GABA autoinhibition caused by a downregulation of presynaptic GABAB receptors (Buhl et al. 1996; Wu and Leung 1997). The decreased facilitation of excitatory transmission and the lower depression of inhibitory transmission found in symptomatic TKO mice can have an adaptive compensatory role against the buildup of hyperexcitability that accompanies epileptogenesis, as proposed for other animal models of epilepsy (Wu and Leung 1997; Merlo et al. 2007; Inaba et al. 2009).
Glutamatergic synapses from presymptomatic TKO mice showed no PTP, a very pronounced synaptic depression preceded by a transient facilitation phase and an impaired recovery from depression, supporting the role of Syns in the mobilization of SVs during repetitive stimulation. Age significantly worsened the response of both TKO and WT to sustained high frequency stimulation and also adult WT mice started to show a mild depression, suggesting that an age-dependent switch between facilitation and depression in excitatory transmission is more precocious in TKO mice with respect to age-matched controls. In accordance with TKO autapses (Gitler et al. 2004), inhibitory synapses from presymptomatic TKO mice showed no genotype-dependent differences in depression. Consistent with the milder PPD discussed above, the same synapses in adult mice displayed a transient phase of milder depression that can represent a compensatory response of inhibitory systems during epileptogenesis.
The alterations in short-term plasticity responses can contribute to the initial E/I imbalance and, in a later stage, be the expression of adaptive responses which can either reinforce inhibition or play a direct role in triggering epileptic paroxysms by changing the filtering properties of the neuronal network toward bursting activities (Avoli et al. 2002; Abbott and Regehr 2004). A reinforcement of inhibition upon epileptogenesis, accompanied by an upregulation of GAD67, has recently been reported in TKO mice (Ketzef et al. 2011).
A Decreased Tonic Inhibition Associated with the E/I Imbalance Triggers Hyperexcitability of TKO Pyramidal Neurons
Our results first demonstrated that the membrane potential of TKO CA1 pyramidal neurons was more depolarized than that of WT pyramidal cells and that this alteration became more severe during epileptogenesis. This effect on the membrane potential was due, at least in part, to a significant decrease in the tonic GABAA current, which was apparent before epileptogenesis. The similar maximal tonic current, evaluated by BMI application in the presence of saturating exogenous GABA, and the comparable expression of α5/δ subunit–containing GABAA receptors indicate that the impaired tonic current is brought about by the decreased GABA spillover consequent to the decreased phasic synaptic release of GABA. A decreased tonic inhibition, besides inducing a depolarization of the resting potential of target cells, may also enlarge the temporal and spatial window for signal integration, thus increasing the firing probability (Farrant and Nusser 2005). Interestingly, a decrease in the tonic current is characteristic of certain animal models of epilepsy, where it directly contributes to network hyperexcitability (Spigelman et al. 2002; Glykys and Mody 2006; Curia et al. 2009).
Modification in intrinsic excitability could contribute to brain hyperexcitability during epileptogenesis (Sanabria et al. 2001; Avoli et al. 2002; Becker et al. 2008). In addition to the depolarized state and increased input resistance of CA1 pyramidal neurons due to the decreased tonic current, bursting TKO pyramidal neurons showed a lower threshold for burst generation and a higher intraburst frequency, while regular-spiking neurons displayed an increased firing rate in response to current injection. These results suggest that the E/I imbalance can modify not only synaptic plasticity but also the membrane potential due to impaired tonic inhibition. Under these conditions, bursting TKO neurons can act as a pacemaker of spontaneous interictal events and can recruit normally firing CA1 pyramidal neurons into the synchronized discharges (Sanabria et al. 2001; Becker et al. 2008; Huberfeld et al. 2011).
Epilepsy is a network phenomenon in whose pathogenesis changes in excitability, inhibitory tone, or E/I balance can play a major role. A constitutive imbalance in excitatory and inhibitory transmission can be considered as the initial cause of hyperexcitability in TKO mice, triggering the secondary plasticity effects that characterize the process of epileptogenesis. Since the extent of inhibition provides stability to neuronal networks, such imbalance may force neuronal circuits into a state of basal heightened excitability, which facilitates the spontaneous or stimulus-evoked onset of seizures, not only in mice bearing Syn gene deletion, but also in man with loss-of-function mutations in SYN genes.
This study was supported by research grants from the Italian Ministry of University and Research (PRIN to F.B. and P.B.), the Italian Ministry of Health Progetto Giovani (to P.B.), and the Compagnia di San Paolo, Torino (to F.B. and P.B.). The support of Telethon-Italy (GGP09134 to F.B. and F.V. and GGP09066 to P.B.) is also acknowledged.
We thank Drs Hung-Teh Kao (Brown University, Providence, RI) and Paul Greengard (The Rockefeller University, New York, NY) for providing us with the TKO mutant mouse strain, Werner Sieghart (Center for Brain Research, Medical University Wien, Austria) for the kind gift of the anti-δ subunit antibody, Silvia Casagrande (Department of Experimental Medicine, University of Genova, Italy) for help in breeding the WT and TKO colonies, and Andrea Arrigoni (The Italian Institute of Technology, Genova, Italy) for help in immunohistochemistry experiments. Conflict of Interest: None declared.