Prolonged Epileptiform Discharges Induced by Altered Group I Metabotropic Glutamate Receptor-Mediated Synaptic Responses in Hippocampal Slices of a Fragile X Mouse Model

Introduction

Fragile X syndrome (FXS) is the most common hereditary form of mental retardation. The syndrome is caused by a trinucleotide repeat (CGG) expansion that results in epigenetic silencing of the gene FMR1 and prevents the expression of the encoded protein, the fragile X mental retardation protein (FMRP) (O'Donnell and Warren, 2002). FMRP is an RNA-binding protein and may function as a negative regulator of protein synthesis at the translation level (Caudy et al., 2002; Ishizuka et al., 2002; Jin et al., 2004; Aschrafi et al., 2005; Qin et al., 2005) (but see Todd et al., 2003).

In the CNS, synaptically activated mRNA translation is reported to be involved in several forms of synaptic plasticity (Steward and Schuman, 2003; Kelleher et al., 2004; Klann and Dever, 2004). In particular, translational processes appear to underlie group I metabotropic glutamate receptor (mGluR)-mediated synaptic plasticity (Huber et al., 2000; Raymond et al., 2000; Karachot et al., 2001; Hou and Klann, 2004; Shin et al., 2004; Aschrafi et al., 2005). FMRP may be coupled to group I mGluR activities (Weiler et al., 1997) and may mediate synaptic plasticity via its action on mRNA translation. In a fragile X mouse model, in which FMRP is functionally absent, the mRNA translation-dependent long-term depression induced by group I mGluR stimulation is enhanced (Huber et al., 2002).

Epileptogenesis is another form of plasticity elicited by group I mGluR activation. In hippocampal slices, transient application of the group I mGluR agonist (S)-3,5-dihydroxyphenylglycine (DHPG) elicited persistent prolonged (>2 s) synchronized discharges of the CA3 population (Taylor et al., 1995; Zhao et al., 2004). These prolonged synchronized discharges resemble ictal discharges in epilepsy and have been used as a model to study epileptogenesis (Wong et al., 2004). Three properties of the group I mGluR-dependent ictal-like discharges make them particularly interesting as an epileptogenesis model: (1) their induction is dependent on mRNA translation (Merlin et al., 1998); (2) once induced by the agonist, they are maintained for hours even after the agonist washout (Merlin and Wong, 1997); (3) expression of the prolonged synchronized discharges is reversibly suppressed by group I mGluR antagonists (Merlin and Wong, 1997). The usefulness of the model has been limited in that epileptogenic activity can be elicited only by an agonist, DHPG, and not via synaptic activation of group I mGluRs (Lee et al., 2002).

Here, we report that, in an FXS mouse model, group I mGluR-mediated epileptogenic responses, normally activatable only by an exogenous agonist, were elicited by synaptic activities. The unleashing of a synaptic mechanism for group I mGluR-mediated epileptogenesis may also be operational in vivo, because fragile X mice have enhanced audiogenic seizure susceptibility (Musumeci et al., 2000; Chen and Toth, 2001; Yan et al., 2004) and this form of seizure can be suppressed by an mGluR5 antagonist (Yan et al., 2005). Our finding is also relevant to an understanding of the FXS, because ∼25% of fragile X patients suffer from epilepsy during development (Musumeci et al., 1999; Sabaratnam et al., 2001).

Materials and Methods

Animals. Male F₁ hybrid mice were obtained from crosses of male C57BL/fmr1-tm1Cgr mice and FVB/fmr1-tm1Cgr female mice (C57 × FVB) to produce Fmr1 knock-out (ko) mutant mice (Yan et al., 2004). Control animals were wild-type C57 × FVB. Three- to 4-week-old animals were used.

Slice preparation. Transverse hippocampal slices (400 μm thick) were prepared as described previously (Bianchi and Wong, 1995; Lee et al., 2002) and placed on the nylon mesh of an interface recording chamber (Fine Science Tools, Vancouver, British Columbia, Canada). The artificial CSF (ACSF) contained the following (in mm): 157 Na⁺, 136 Cl^–, 5 K⁺, 1.6 Mg²⁺, 2Ca²⁺, 26HCO^–₃, and 11 d-glucose. Slices were perfused with ACSF continuously bubbled with 95% O₂/5% CO₂ to maintain the pH near 7.4. The temperature was maintained at 34–36°C.

Electrophysiological recordings. In all experiments, the slices were allowed to recover from the isolation procedure for at least 1.5 h before the beginning of the recordings. Intracellular recordings were performed using an Axoclamp 2A amplifier (Molecular Devices, Palo Alto, CA). Electrodes were pulled with thin-walled glass tubing (World Precision Instruments, Sarasota, FL) and had resistances of 30–50 MΩ when filled with potassium acetate (2 m). Some extracellular recordings were obtained with 3–8 MΩ electrodes filled with ACSF. Voltage signals were digitized and stored in an Intel Pentium-based computer using a Digidata 1322A converter controlled by pClamp 8 software (Molecular Devices).

Pharmacological agents. Baseline epileptiform activities for experiments in wild-type and Fmr1 ko mice hippocampal slices were elicited by continuous bath perfusion of the GABA_A antagonist bicuculline (50 μm). In another group of wild-type slices, the group I mGluR agonist DHPG (50 μm) was used to induce synchronized activities. DHPG, the mGluR5-selective antagonist 2-methyl-6-(phenylethynyl)-pyridine hydrochloride (MPEP) (50 μm), the mGluR1-selective antagonist (S)-(+)-α-amino-4-carboxy-2-methylbenzeneacetic acid (LY367385) (50 μm), the translation inhibitors anisomycin and cycloheximide (15–20 and 60 μm, respectively), and transcription inhibitor actinomycin D (25 μm) were obtained from Tocris Cookson (Ellisville, MO). 2-(2-Amino-3-methyoxyphenyl)-4 H-1-benzopyran-4-one (PD98059) (50 μm) was from Calbiochem (La Jolla, CA). Bicuculline and the other chemicals were purchased from Sigma (St. Louis, MO). All agents were stored in stock solutions at –80°C for no more than 1–2 weeks and diluted into ACSF at the indicated final concentrations at the time of the experiments.

Data analysis. CA3 pyramidal cells included in this study had stable resting membrane potentials of less than –60 mV, overshooting action potentials, and input resistance of >30 MΩ. In some cases, hyperpolarizing DC was injected into the cells to prevent intrinsic firing and identify network activities. The durations of individual synchronized discharges were measured from the beginning of the first action potential to the repolarization of the last action potential of the discharge at membrane potentials kept within a few millivolts throughout the experiment. In the frequency histogram plots, the durations of all synchronized discharges recorded in a 6 min period for each slice in the various experimental conditions were included. The histograms were fitted with first- or second-order Gaussian equations. Based on the distribution of the synchronized discharge durations shown in Figure 1 Dbi, “short” and “prolonged” discharges refer to events shorter and longer, respectively, than 1.5 s. Clampfit 8 (Axon Instruments) and Sigma Plot (SPSS, Chicago, IL) software were used for all data analysis. Average data were expressed as mean ± SEM. Student's t test was used for statistical comparisons, and differences were considered significant when p < 0.05.

Results

Bath application of bicuculline (50 μm) to a hippocampal slice prepared from a wild-type mouse first induced occasional synchronized discharges with a mean duration of 0.573 ± 0.026 s (short synchronized discharges; n = 20) in the CA3 area. Within 10 min, the short synchronized discharges began to occur regularly and rhythmically (frequency range, 0.042–0.123 Hz; n = 25 discharges) (Fig. 1Aa). Overall, the frequency and duration of the short synchronized discharges remained stable for at least 2 h in all preparations examined (n = 10) (Fig. 1A,C,Da). The average duration of short synchronized discharges recorded ∼2 h after their initial appearance was 0.625 ± 0.021 s (n = 30 discharges) (Fig. 1Ab), a value not significantly different from that recorded when they first appeared (p = 0.122; Student's t test).

Figure 1B shows records obtained from hippocampal slices prepared from Fmr1 ko mice. Addition of bicuculline (50 μm) elicited an initial period of short rhythmic synchronized discharges with average duration of 0.627 ± 0.021 s, not significantly different from that recorded in the wild-type slice. In contrast to the events recorded from wild-type slices, in all of the Fmr1 ko preparations, a variable initial period of 5–40 min was followed by sporadic prolongation of the short synchronized discharges interspersed between the regular rhythmic short synchronized discharges. Prolonged discharges were considered to be extended from short discharges for two reasons: (1) they occurred at the time point at which a short synchronized discharge would have occurred as predicted by the ongoing rhythmicity, and (2) the initial action potential pattern of a prolonged discharge resembled that of the short synchronized discharge (Fig. 1B, compare a, b), suggesting that the prolonged discharge was triggered through a mechanism initiated by the short synchronized discharge. Over time, the frequency and duration of the prolonged discharges increased (Fig. 1C), and, by 90 min of bicuculline perfusion, the prolonged discharges occurred at regular rhythms (frequency range, 0.011–0.021 Hz; n = 6) and became the dominant spontaneous event (Fig. 1Bb, inset; C). Figure 1, Bb and C, shows that prolonged synchronized discharges at ∼2 h of perfusion with bicuculline were significantly longer (6.347 ± 0.021 s; n = 6 discharges) than the short synchronized discharges (p << 0.001). Often, even when the prolonged discharges became rhythmic, short-duration discharges were still observed and they coexisted with the prolonged discharges (Fig. 1Bb, inset; C). Population data from 21 Fmr1 ko mice indicate that, after >80 min of bicuculline application, in all preparations short and prolonged synchronized discharges are distributed into two relatively distinct groups (Fig. 1Dbi), and the longest discharges are significantly prolonged compared with those recorded at the beginning of their appearance (Fig. 1Dbii).

Prolongation of the bicuculline-induced short synchronized discharges is an event unique to the Fmr1 ko preparation. In the wild-type preparation, prolongation of the short synchronized discharges induced by bicuculline was observed only after addition of the group I mGluR agonist DHPG (25–50 μm). The DHPG-induced prolonged synchronized discharges have two distinct properties: (1) the extended discharges are sustained by activation of group I mGluRs (Merlin and Wong, 1997), and (2) the induction of the events is protein synthesis dependent (Merlin et al., 1998). It is possible that the bicuculline-induced prolonged synchronized discharges in Fmr1 ko slices are also dependent on the group I mGluR-mediated mechanism that is now triggered by synaptically released glutamate, without the need for added DHPG. If this were the case, the prolonged discharges elicited by bicuculline in the Fmr1 ko mice should be pharmacologically similar to those elicited in the wild-type slices after DHPG stimulation (i.e., suppressed by group I mGluR antagonists), and their induction should be sensitive to translation inhibitors. We tested these two possibilities.

The effectiveness of MPEP, a blocker of the group I mGluR subtype mGluR5, on synchronized discharges elicited by bicuculline was examined. In the wild-type preparations, addition of MPEP (50 μm) did not significantly affect the duration of rhythmic short synchronized discharges observed after 80 min of bicuculline treatment (bicuculline, 80 min: 0.478 ± 0.014 s; n = 7; bicuculline plus MPEP, 60 min: 0.454 ± 0.011 s; n = 10; p = 0.203) (Fig. 2Aa,Ab,C). Figure 2, B and C, shows results obtained from hippocampal slices prepared from an Fmr1 ko mouse. Rhythmic short synchronized discharges initially elicited by bicuculline developed into mainly prolonged synchronized discharges by ∼80 min of maintained bicuculline perfusion (duration, 4.230 ± 0.111 s; n = 7) (Fig. 2Ba,C). Within 30 min after MPEP perfusion, the prolonged discharges shortened significantly (bicuculline plus MPEP, 60 min: 0.534 ± 0.012 s; n = 6; p << 0.001) and the spontaneous activities reverted back to rhythmic short synchronized discharges (Fig. 2Bb,C). Summary data from slices prepared from seven Fmr1 ko animals show that most slices exhibited rhythmic patterns of short and prolonged synchronized discharges after 80 min of bicuculline perfusion (Fig. 2Dbi). Addition of MPEP after the 80 min time point suppressed the prolonged synchronized discharges, leaving behind rhythmic short synchronized events at an accelerated frequency (Fig. 2Bb,Db). In wild-type preparations, MPEP did not significantly affect the duration of short synchronized discharges recorded after 80 min of bicuculline treatment (Fig. 2Da).

LY367385 (50 μm), an antagonist of the other group I mGluR subtype mGluR1, was also effective in suppressing bicuculline-induced prolonged synchronized discharges in Fmr1 ko preparations. In four experiments, the five longest discharges in a 6 min period after 80–120 min of bicuculline perfusion were 6.107 ± 0.735, 3.599 ± 0.054, 5.907 ± 0.171, and 5.855 ± 0.246 s. After 60–90 min of LY367385 application, the five longest discharges were 0.440 ± 0.027, 1.056 ± 0.051, 0.675 ± 0.032, and 4.191 ± 0.248 s, respectively (p < 0.01 in all cases).

Previous data show that, in hippocampal slices prepared from guinea pigs, agents that inhibit mRNA translation such as anisomycin and cycloheximide prevented the activation of prolonged synchronized discharges by DHPG (Merlin et al., 1998). We reevaluated the action of these inhibitors on wild-type mouse preparations.

Addition of DHPG to hippocampal slices elicited a phase of short rhythmic synchronized discharges, which gradually evolved into a mixed pattern of short and prolonged discharges. After ∼40 min of DHPG perfusion, the spontaneous synchronized activity settled into a rhythmic pattern of primarily prolonged discharges with interspersed short discharges (Fig. 3Aa). This activity was ongoing relatively unaltered for up to 4 h (the longest period recorded). The average duration of the prolonged discharges (4.559 ± 0.114 s; n = 17) (Fig. 3Ad; C, top panel) was similar to that of bicuculline-induced prolonged discharges in Fmr1 ko slices (Fig. 1Bb,Dbi). Pretreatment of hippocampal slices with 20 μm anisomycin for 1 h before the addition of DHPG prevented the appearance of prolonged discharges in two of six slices. In the remaining preparations treated with anisomycin, short discharges evolved into prolonged ones, but rhythmic prolonged discharges appeared only transiently for a period of 10–30 min, and by 60 min of DHPG in none of the slices was synchronized activity observed (Fig. 3Ab,B). This was unlikely attributable to non-specific effects on synchronizing mechanisms of the CA3 network, because bicuculline addition to anisomycin-treated slices did elicit short synchronized discharges (Fig. 3Ae). Pretreatment of slices with cycloheximide (60 μm; n = 3) produced similar suppressive effects on the elicitation of DHPG-mediated prolonged discharges (Fig. 3B). Thus, inhibitors of mRNA translation antagonized the generation of DHPG-induced prolonged discharges. In contrast, pretreatment of slices with actinomycin D (25 μm), a transcription inhibitor, did not affect prolonged discharge activities elicited by DHPG (Fig. 3Ac,Af) (duration, 4.309 ± 0.071 s; n = 16) (Fig. 3C, bottom panel). The results suggest that active translation is required for the generation of DHPG-activated prolonged discharges.

Prolonged discharges observed in bicuculline-treated Fmr1 ko hippocampal slices were also suppressed by anisomycin pretreatment (Fig. 4B,D). Rhythmic short synchronized discharges were the only activity pattern recorded under this condition. Thus, in the presence of anisomycin, bicuculline elicited similar responses in Fmr1 ko preparations as those elicited by bicuculline in wild-type preparations. Additional experiments showed that anisomycin pretreatment did not affect the time course of short synchronized discharges in the wild-type preparation (Fig. 4A,D). The results suggest that the generation of short synchronized discharges in the wild-type and Fmr1 ko bicuculline-treated preparations is protein synthesis independent. Generation of bicuculline-mediated prolonged discharges, however, is dependent on a mRNA translation, a process that is activatable only in the Fmr1 ko preparations.

Previous studies show that, in wild-type slices, the induction of prolonged synchronized discharges by DHPG requires activation of the extracellular signal-regulated kinase 1/2 (ERK1/2) (Zhao et al., 2004). We examined whether ERK1/2 is also involved in the activation of bicuculline-induced prolonged synchronized discharges in the Fmr1 ko hippocampal slices. Inhibition of ERK1/2 was attempted with pretreatment of the slices with the mitogen-activated protein kinase kinase (MEK) inhibitor PD98059 (50 μm). In these slices, only short synchronized discharges were observed in bicuculline for periods of up to 2 h (Fig. 4C,D).

Discussion

The results show that prolonged synchronized discharges evoked by bicuculline in Fmr1 ko hippocampal slices share similar properties with the prolonged synchronized discharges activated by the group I mGluR agonist DHPG in wild-type slices. The data suggest that, in the Fmr1 ko preparation, synaptically released glutamate is effective in inducing group I mGluR-mediated prolonged synchronized discharges.

In the wild-type preparation, group I mGluRs are activated by DHPG. The activation of the receptors initiates the signaling cascade and, via protein synthesis, induces the prolonged synchronized discharges. Signaling through this cascade is probably the limiting step in the period required for the expression of the fully extended prolonged synchronized discharges. In Fmr1 ko preparations, group I mGluRs are activated by synaptic glutamate synchronously released in the presence of bicuculline. Activation of group I mGluRs in the ko was effective in initiating the signaling cascade for the protein synthesis-dependent induction of the prolonged synchronized discharges. Prolonged synchronized discharges appeared after a latent period of 5–40 min after the appearance of the short synchronized discharges. This period of time may again reflect translation processes necessary for the induction of prolonged synchronized discharges. Thus, the prolonged synchronized discharges induced by DHPG in wild-type slices and those induced by bicuculline in Fmr1 ko slices appear to share a common mechanism dependent on the group I mGluR-activated protein translation process.

In hippocampal neurons of wild-type preparations, group I mGluRs are activated synaptically during evoked responses (Nakamura et al., 1999) and during bicuculline-induced short synchronized discharges (Lee et al., 2002). However, despite such synaptic activation, group I mGluR-mediated prolonged discharges are not elicited in bicuculline. Lee et al. (2002) showed that, in the wild-type preparation, synaptically activated group I mGluR responses mainly result from the direct actions of the receptor-coupled G-proteins and do not involve phospholipase C β1 (PLCβ1) activation and the signaling cascade downstream to PLCβ1 activation. Group I mGluR-activated mRNA translation is probably ERK1/2 dependent (Fig. 4C,D) and requires signaling downstream to PLCβ1 activation (Zhao et al., 2004). Thus, synaptic stimulation of group I mGluRs may not be effective in activating mRNA translation and the resulting prolonged synchronized discharges in the wild-type hippocampal neurons. The unleashing of this process in the Fmr1 ko preparation suggests that synaptically induced prolonged synchronized discharges and the associated mRNA translation are repressed by FMRP (Fig. 4E). A similar role of FMPR in downregulating a group I mGluR-induced, translation-dependent long-term depression has led to the hypothesis that exaggerated group I mGluR function is a fundamental abnormality in fragile X syndrome (Bear et al., 2004). Our results provide an additional example to this general concept and extend the hypothesis to include epileptogenesis as a possible outcome of an exaggerated group I mGluR function.

Epilepsy is a syndrome suffered by ∼25% of fragile X patients during development (Musumeci et al., 1999; Sabaratnam et al., 2001). Recent studies show that the susceptibility to audiogenic seizures is enhanced in Fmr1 ko mice (Musumeci et al., 2000; Chen and Toth, 2001; Yan et al., 2004). In addition, these seizures are effectively suppressed by the mGluR5 antagonist MPEP (Yan et al., 2005). Our data suggest that the expression of prolonged synchronized discharges requires the contribution of both mGluR1 and mGluR5. Antagonism of either receptor subtype is effective in preventing the expression of prolonged synchronized discharges.

The epileptogenic property of synaptic group I mGluRs in the FXS mouse model is uncovered in our studies by using bicuculline to block GABA_A receptor-mediated inhibition. It is possible that enhanced group I mGluR-mediated synaptic responses can also be recruited by intense neuronal activity in the Fmr1 ko preparation to elicit epileptiform activities, particularly in view of the finding that GABA_A receptor-mediated inhibition is down-regulated in Fmr1 ko mice (El Idrissi et al., 2005). As noted above, enhanced synaptic group I mGluR responses may indeed be operational to sustain increased audiogenic seizure susceptibility in the FXS mouse model (Yan et al., 2005).

In summary, our data are consistent with the emerging hypothesis that exaggerated group I mGluR translational responses contribute to synaptic abnormalities in fragile X syndrome (Bear et al., 2004). Together with available data, our findings suggest that FMRP regulation of synaptically induced group I mGluR-mediated mRNA translation is involved in the normal neuronal signaling and that the absence of such regulation, as in fragile X syndrome, could cause brain pathologies including epileptogenesis.