Progress in understanding the roles of kainate receptors (KARs) in synaptic integration, synaptic networks, and higher brain function has been hampered by the lack of selective pharmacological tools. We have found that UBP310 and related willardiine derivatives, previously characterized as selective GluK1 and GluK3 KAR antagonists, block postsynaptic KARs at hippocampal mossy fiber (MF) CA3 synapses while sparing AMPA and NMDA receptors. We further show that UBP310 is an antagonist of recombinant GluK2/GluK5 receptors, the major population of KARs in the brain. Postsynaptic KAR receptor blockade at MF synapses significantly reduces the sustained depolarization, which builds up during repetitive activity, and impacts on spike transmission mediated by heterosynaptic signals. In addition, KARs present in aberrant MF synapses in the epileptic hippocampus were also blocked by UBP310. Our results support a specific role for postsynaptic KARs in synaptic integration of CA3 pyramidal cells and describe a tool that will be instrumental in understanding the physiopathological role of KARs in the brain.
Kainate receptors (KARs) are ionotropic glutamate receptors composed of tetrameric assemblies of GluK1-5 subunits (Pinheiro and Mulle 2006). GluK2, GluK4, and GluK5 are thought to compose postsynaptic KARs at mossy fiber (MF) to CA3 pyramidal cell synapses (mf-CA3 synapses; Mulle et al. 1998; Contractor et al. 2003; Fernandes et al. 2009), where they contribute with small amplitude, slowly decaying currents (Castillo et al. 1997; Vignes and Collingridge 1997). Despite their small amplitude, KAR-excitatory postsynaptic currents (EPSCs) were hypothesized to be important for the integration of information because they summate temporally (Frerking and Ohliger-Frerking 2002). This hypothesis has been directly tested at excitatory synapses onto hippocampal interneurons (Goldin et al. 2007; Yang et al. 2007) and at aberrant synapses in dentate granule cells (DGCs) in temporal lobe epilepsy (TLE) (Artinian et al. 2011) but not at hippocampal mf-CA3 synapses. This is mainly due to the lack of any pharmacological tool that blocks postsynaptic KARs in CA3 pyramidal cells without affecting other ionotropic glutamate receptors. In addition to their ionotropic action to mediate KAR-EPSCs, postsynaptic KARs composed of GluK2 and GluK5 are implicated in the modulation of a slow Ca2+-activated K+ current (IsAHP; Fisahn et al. 2005; Ruiz et al. 2005), although a more recent study has failed to find a role for GluK5 in this metabotropic action of KARs (Fernandes et al. 2009). KARs are also localized presynaptically where they modulate glutamate release and greatly influence the plasticity of mf-CA3 synapses (Contractor et al. 2001; Lauri et al. 2001; Schmitz et al. 2001; Pinheiro et al. 2007).
The lack of selective pharmacological tools hampers clarifying the physiological relevance of postsynaptic KARs at mf-CA3 synapses. Recently, an intense effort was made to synthesize and characterize subunit-selective KAR antagonists, mainly against GluK1 (Jane et al. 2009). However, no selective antagonist of GluK2/GluK5 KARs yet exists, although they likely represent a major population of KARs in the brain (Wenthold et al. 1994). The implication of KAR-EPSCs at hippocampal MF synapses has only been addressed by comparing wild-type and GluK2−/− mice, which display loss of postsynaptic KARs (Mulle et al. 1998; Sachidhanandam et al. 2009). However, GluK2−/− mice are also deficient in presynaptic KARs and, consequently, display markedly decreased short-term dynamics (Contractor et al. 2001; Sachidhanandam et al. 2009). It would therefore be of great value to be able to directly test the implication of postsynaptic KARs in synaptic integration and information transfer in CA3 pyramidal cells.
KARs are also implicated in several neuropathologies. Recurrent MF synapses are a hallmark of the hippocampus from patients with TLE as well as in rodent epilepsy models with lesions in the CA3 region. KAR-EPSCs can be observed at recurrent MF synapses in the dentate gyrus in a rat model of TLE (Epsztein et al. 2005). Directly targeting the efficacy of recurrent MF synapses with pharmacological tools would be invaluable to study the debated implication of hyperexcitability associated with recurrent MFs in models of TLE.
In search for antagonists of KARs at mf-CA3 synapses, we reevaluated the efficacy of willardiine derivatives (Jane et al. 2009) on native KARs. We found that UBP310 blocks postsynaptic KARs at mf-CA3 synapses and spares AMPA and NMDA receptors. UBP310 was originally developed as a GluK1-selective antagonist (Dolman et al. 2007) and later found to also potently antagonize GluK3 (Perrais et al. 2009). We further show that UBP310 is also an antagonist of recombinant GluK2/GluK5 receptors, the major population of KARs in the brain. Using this tool, we directly investigated the implication of postsynaptic KARs in synaptic transmission, integration, and short-term plasticity at hippocampal MF synapses. Finally, we tested its potential use in inhibiting KAR-mediated EPSCs in a rodent model of TLE.
Materials and Methods
Electrophysiological Recordings from Brain Slices
Mouse parasagittal brain slices of 14- to 21-day-old C57BL/6 wild-type and GluK1 and GluK3 knockout mice were cut and stored up to 6 h in a modified artificial cerebrospinal fluid (ACSF) with partial substitution of sodium by sucrose. This solution was composed of (in mM) 87 NaCl, 25 NaHCO3, 25 glucose, 75 sucrose, 2.5 KCl, 1.25 NaH2PO4, 0.5 CaCl2, and 7 MgCl2 and was saturated with 95% O2/5% CO2. CA3 pyramidal cells were identified under infrared illumination with differential interference contrast optics and recorded in whole-cell voltage-clamp mode (3.5–4.5 MΩ electrodes, −70 mV holding potential) at room temperature. Slices were perfused with ACSF containing (in mM) 125 NaCl, 2.5 KCl, 1.25 NaH2PO4, 26 NaHCO3, 2.3 CaCl2, 1.3 MgCl2, and 25 glucose, saturated with 95% O2/5% CO2, and supplemented with bicuculline (10 μM) and CGP 55845 (1 μM). The intracellular solution contained (in mM) 140 Cs-methanesulfonate, 10 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 10 ethyleneglycol-bis(2-aminoethylether)-N,N,N′,N′-tetra acetic acid (EGTA), 2 MgCl2, 2 NaCl, and 4 Na2ATP, pH 7.3. Current-clamp recordings were made at 33 °C with a pipette solution containing (in mM) 120 K-gluconate, 20 KCl, 10 HEPES, 10 EGTA, 2 MgCl2, 1 CaCl2, 2 Na2ATP, and 0.2 Na2GTP, pH 7.3, and the Mf-EPSCs/excitatory postsynaptic potentials (EPSPs) were elicited by randomly poking the dentate hilus with a patch pipette filled with extracellular solution until synaptic responses with the characteristic features of MFs (large paired-pulse ratio and frequency facilitation) were found. The intensity of stimulation was then reduced to reach minimal stimulation conditions. The associational/commissural (A/C) pathway was stimulated with a concentric bipolar electrode placed in the stratum radiatum. Data were filtered at 2.9 KHz and digitized at 10 KHz.
Whole-cell pipettes used to record IsAHP were filled with an internal solution containing (in mM) 135 potassium methylsulfate, 5 KCl, 0.1 EGTA-Na, 10 HEPES, 2 NaCl, 5 Na2ATP, 0.4 GTP, and 10 Na-phosphocreatine. The methylsulfate anion was used to preserve IsAHP in a form similar to that observed with sharp electrode recordings (Lancaster and Adams 1986; Ruiz et al. 2005). The following drugs were used to block AMPA, NMDA, γ-aminobutyric acid (GABA)A, GABAB, mGluR5, and mGluR1 receptors, respectively: LY303070 (25 μM), DAP5 (25 μM), Bicuculline (10 μM), CGP 55845 (1 μM), MPEP (10 μM), and CPCCOEt (100 μM). Tetraethylammonium (TEA; 5 mM) was included in the mixture of antagonists to facilitate calcium spikes and reliably evoke IsAHP (Ruiz et al. 2005). The IsAHP was elicited with intrasomatic depolarizing current pulses (duration, 80 ms; amplitude, 50 mV) every 60 s, and the amplitude was measured 350–500 ms after the depolarizing voltage step to minimize the contribution of the medium AHP (Martin et al. 2001; Ruiz et al. 2005). In experiments where synaptically released glutamate was used to modulate the IsAHP through KAR activation, we delivered 3 trains of 5 stimuli at 100 Hz every second to MFs, and we observed the effect of these trains on the amplitude of IsAHP 500 ms after (Ruiz et al. 2005).
Series resistance was monitored throughout, and recordings were discarded if it varied by >20%. CA3 pyramidal cells recorded in current-clamp mode had an average resting potential of 67.7 ± 0.7 mV and a membrane resistance of 513 ± 32 MΩ; they were discarded from analysis if membrane resistance varied by 10% or more during recordings.
Electrophysiological Recordings from Recombinant Receptors
Recombinant receptors were recorded from HEK 293 cells as described previously (Pinheiro et al. 2007). Cells were cotransfected either with GluK2a and green fluorescent protein (GFP) (complementary deoxyribonucleic acid [cDNA] ratio of 2:1) or GluK2a, GluK5, and GFP (cDNA ratio of 1:3:2). One to 3 days after transfection, brightly fluorescent cells were recorded at room temperature in a solution containing (in mM) 145 NaCl, 2 KCl, 2 MgCl2, 2 CaCl2, 10 glucose, and 10 HEPES (pH 7.4, 320 mOsm). Recording pipettes (2–4 MΩ) were filled with a solution containing (in mM) 122 CsCl, 2 NaCl, 2 MgCl2, 10 EGTA, 10 HEPES, 4 Na2ATP, and 0.06 spermine (pH 7.2, 310 mOsm). Glutamate (0.1 or 1 mM) was applied for 1 or 100 ms every 20 s with a fast piezoelectric translator (Polytec PI) while cells were held at a membrane potential of −40 or −80 mV. The actual application durations were tested using open tip junction potentials (Supplementary Fig. 1). UBP310 (0.1–30 μM) was present in control and agonist lines. The data were acquired using Pulse and analyzed in Igor Pro 5.
TLE Model Experiments
An animal model of TLE was obtained using adult male Wistar rats injected with pilocarpine hydrochloride as previously described (Epsztein et al. 2005, 2010). Adult male Wistar rats (∼2 months old; Janvier Breeding Center, Le Genest-Saint-Isle, France) were injected intraperitoneally (i.p.) with pilocarpine hydrochloride (340 mg/kg dissolved in NaCl 0.9%) 30 min after a low dose of scopolamine methyl nitrate (1 mg/kg, i.p.). Approximately 60% of the rats experienced class IV/V seizures (Racine 1972). After 3 h of status epilepticus, diazepam was injected (8 mg/kg, i.p.). After a seizure-free period of several weeks, we selected for recordings and analysis only rats that experienced spontaneous seizures (13-month-old chronic epileptic rats, n = 2) since they display a strong MF sprouting (Epsztein et al. 2005). Naive age-matched rats were used as controls (13 months old, n = 3). The animals were deeply anesthetized with chloral hydrate (350 mg/kg, i.p.) and decapitated. The brain was removed rapidly, the hippocampi were dissected, and transverse 400-μm thick hippocampal slices were cut using HM650V MicroM tissue slicer in a solution containing the following (in mM): 110 choline, 2.5 KCl, 1.25 NaH2PO4, 25 NaHCO3, 7 MgCl2, 0.5 CaCl2, and 7 D-glucose (5 °C). Slices were then transferred for rest at room temperature (1 h) in oxygenated ACSF containing the following (in mM): 126 NaCl, 3.5 KCl, 1.2 NaH2PO4, 26 NaHCO3, 1.3 MgCl2, 2.0 CaCl2, and 10 D-glucose, pH 7.4. Whole-cell recordings of DGCs from chronic epileptic and control rats were obtained using the “blind” patch-clamp technique as previously described (Epsztein et al. 2005, 2010). Electrodes were filled with an internal solution containing the following (in mM): 135 gluconic acid, 135 CsOH, 10 KCl, 0.1 CaCl2, 0.2 EGTA, 10 HEPES, 2 MgATP, 0.4 NaGTP, and 0.5% biocytin, pH 7.2. Whole-cell recordings were performed in voltage-clamp mode using a Multiclamp 700B amplifier (Axon Instruments; Molecular Devices, Sunnyvale, CA). Data were filtered at 2 kHz, digitized (20 kHz) with a Digidata 1440A (Molecular Devices) to a personal computer, and acquired using Clampex 10.1 software (PClamp, Axon Instruments; Molecular Devices). Signals were analyzed offline using Clampfit 10.1 (Molecular Devices). EPSCs were evoked by bulk stimulations (0.2 Hz) with a bipolar NiCh electrode (NI-0.7F; Phymep, Paris) positioned in the inner one-third of the molecular layer of the dentate gyrus to stimulate either A/C inputs in control rats or recurrent MF inputs in epileptic rats (stimulus intensity and pulse duration were 10–45 μA and 10-45 ms, respectively). AMPAR-EPSCs were pharmacologically isolated in the presence of blockers of NMDA (40 μM D-APV) and GABAA (5 μM gabazine [SR-95531]) receptors in control rats. In epileptic rats, KAR-EPSCs were pharmacologically isolated in the presence of antagonists of AMPA receptors (30 μM GYKI53655) in addition to NMDA and GABAA receptor antagonists.
All results are presented as mean ± standard error of the mean. Comparison between groups was performed using the paired or unpaired two-tailed t-test, as appropriate, unless otherwise specified. Data plotting and statistical analyses were performed in GraphPad Prism (GraphPad Software, La Jolla, CA) or SigmaStat 3.1 (Systat Software, Richmond, CA).
UBP Willardiine Derivatives Block KAR-EPSCs at mf-CA3 Synapses
In a first set of experiments, we have studied the effects of UBP310 on native KARs at mf-CA3 synapses. Therefore, we investigated the effects of UBP310 on pharmacologically isolated KAR-EPSCs (in 50 μM D-AP5 to block NMDA receptors and 25 μM LY303070 to block AMPA receptors). We found that UBP310 dose-dependently blocked KAR-EPSCs with an IC50 of ∼250 nM (Fig. 1A,B). The effect of UBP310 could be mediated through an action on GluK1 or GluK3 KARs that are sensitive to this compound (Dolman et al. 2007; Perrais et al. 2009). Although mouse CA3 pyramidal cells express neither GluK3 nor detectable amounts of GluK1 messenger RNA after postnatal day 6 (Bureau et al. 1999), we tested UBP310 in subunit-specific knockout mice. We found that UBP310 (3 μM, which blocks KAR-EPSCs in wild-type mice) was equally effective in GluK1- and GluK3-deficient mice (Fig. 1C). We also found that UBP302 and UBP316 (2 μM) significantly inhibited isolated KAR-EPSCs with an order of potency UBP302 < UBP316 < UBP310 (Fig. 1D). Importantly, UBP310 fully spared AMPAR-EPSCs (Perrais et al. 2009) and NMDAR-EPSCs (Supplementary Fig. 2).
UBP310 Is a GluK2/GluK5 Antagonist
Our finding that UBP310 was efficient in blocking native KARs at mf-CA3 synapses, which do not contain GluK1 or GluK3, was rather surprising. However, AMPA and the GluK1-selective agonists ATPA and willardiine (from which UBP310 derives) are active on GluK2/GluK5 receptors (Alt et al. 2004) without significant activity at GluK2 alone. Because postsynaptic KARs at mf-CA3 synapses likely contain GluK2 and GluK5, we tested the ability of UBP310 to antagonize recombinant GluK2/GluK5 heteromeric receptors. In HEK 293 cells transfected with GluK2 and GluK5, glutamate-activated currents (1 mM, 100 ms) were only slightly inhibited by UBP310 (Fig. 2A). However, the kinetics of evoked currents were markedly affected with increasing concentrations of UBP310. We therefore tried to mimic the fast release and clearance of glutamate in synapses by applying short (1 ms) pulses of glutamate, as previously described (Barberis et al. 2008; see also Supplementary Fig. 1). These currents displayed slow deactivation kinetics (Fig. 2E) and UBP310 dose-dependently reduced current amplitudes with an IC50 of 1.3 μM (Fig. 2H). In cells transfected with GluK2 alone, glutamate (1 mM, 100 ms) generated large currents that were little affected in amplitude or kinetics by UBP310 (Fig. 2B). Similar results were obtained for short applications of glutamate (Fig. 2F). Consistent with the competitive nature of UBP310, its concentration-dependent antagonism was shifted to lower concentrations when only 100 μM glutamate was applied for 1 ms (Fig. 2G,H), lowering the IC50 to 0.75 μM. Furthermore, UBP310 had a significant- and concentration-dependent effect on the peak amplitude of currents activated by 100 ms pulses of 100 μM glutamate while markedly affecting the rise and decay kinetics (Fig. 2C,D). These experiments provide the first identification of a potent antagonist of recombinant GluK2/GluK5 receptors, which spares GluK2 and AMPA receptors.
Postsynaptic KARs and Spike Transmission
KARs contribute, on average, to less than 10% of the total amplitude of mf-EPSCs (e.g., Pinheiro et al. 2007), and although charge transfer is high due to the slow decay of their currents, their small amplitude raises doubts as to whether they may have a real impact on mf-CA3 synaptic transmission. Until now there was no way to selectively interfere with the function of postsynaptic KARs in CA3 pyramidal cells. We used UBP310 to directly address the role of KARs in synaptic integration. We observed that mf-EPSCs decay significantly faster in the presence of UBP310 (Fig. 3A,B) due to the loss of the slow KAR-mediated component. The effects of UBP310 are more pronounced on the EPSP waveform, and a double exponential fit reveals a significant acceleration of the slower phase of decay (Fig. 3C,D). This is somewhat expected since the fast, large-amplitude AMPAR-mediated currents will charge the membrane capacitance allowing the slow KAR-mediated component to exert a noticeable impact on membrane voltage. Our question was whether such change in mf-EPSP waveform could affect temporal summation and effectively alter mf-CA3 spike transmission. Given the effects of UBP310 on single EPSPs, blocking KARs should have the largest impact on mf-EPSP summation at frequencies around 10 Hz (Sachidhanandam et al. 2009). Indeed, UBP310 (5 μM) modified the sustained depolarization underlying the synaptic response to a train of 20 stimulations, measured as the average depolarization underneath the last 10 EPSPs (11.2 ± 1.1 mV in control and 8.9 ± 0.8 mV in UBP310; P = 0.005; Fig. 3E,F). This effect is synaptic because UPB310 did not affect cell excitability directly (Supplementary Fig. 3). However, we found that UBP310 had no significant impact on either the probability to trigger spikes (Fig. 3G) or on the latency for a first spike during the train (Supplementary Fig. 4). A plausible explanation is that spike transmission at mf-CA3 is mainly governed by the prominent presynaptic facilitation induced by repetitive trains, as mf-EPSCs can reach several hundred picoampere in the first few stimulations. The reduction in sustained depolarization during the train in UBP310 (Fig. 3E,F) may nevertheless affect spike transmission triggered by costimulation of weaker glutamatergic inputs impinging on CA3 pyramidal cells. Indeed, when a stimulation of A/C fibers was applied following a train of 5 stimulations to mf-CA3 synapses, its probability to trigger a spike was significantly reduced by UBP310 (Fig. 3H,I). Therefore, selective block of postsynaptic KARs affects the integration of diverse synaptic inputs onto CA3 pyramidal cells.
In addition to their ionotropic function at mf-CA3 synapses, KARs inhibit a Ca2+-activated K+ current (IsAHP) via a metabotropic signaling pathway, leading to increased excitability of CA3 pyramidal cells (Ruiz et al. 2005). Synaptic activation of KARs at mf-CA3 synapses leads to bimodal signaling through both ionotropic and metabotropic function (Ruiz et al. 2005). Inhibition of IsAHP is absent in GluK5−/− mice despite the preservation of KAR-EPSCs (Ruiz et al. 2005; but see Fernandes et al. 2009). We tested whether the UBP310 compound could inhibit the synaptic modulation of IsAHP by KARs. In control conditions, trains of stimulation to MFs (3 trains of 5 stimuli at 100 Hz every second) lead to a 14.1 ± 5.3% (n = 8, P < 0.0001) reversible inhibition of the IsAHP 500 ms after the last train (Fig. 4A). We repeated the protocol in the same cell in the presence of 10 μM UBP310. While KAR-EPSCs were inhibited, the inhibition of IsAHP by synaptic KARs was still present (9.3 ± 8.1% of inhibition, n = 8, P < 0.05; Fig. 4B). These results indicate that UBP310 inhibits the ionotropic function of KARs without affecting their metabotropic action.
Postsynaptic KARs and Short-term Synaptic Plasticity
Through both pharmacological and genetic approaches, it was demonstrated that presynaptic KARs contribute to the large dynamic range of mf-CA3 synapses (e.g., Contractor et al. 2001; Schmitz et al. 2001; Pinheiro et al. 2007). This view was challenged by invoking the recruitment of polysynaptic activity, much of which was attributed to the activation of postsynaptic KARs (Kwon and Castillo 2008). Because UBP310 selectively blocks postsynaptic KARs and spares other glutamate receptors and presynaptic KARs at mf-CA3 synapses (Perrais et al. 2009), it can be used to directly evaluate this possibility. Our first reasoning was that, if indeed polysynaptic activity plagued most recordings from mf-CA3 synapses, we would also observe large EPSC rise times, as reported (Kwon and Castillo 2008). However, in recordings from mouse hippocampal slices in minimal stimulation conditions, the average rise time (20–80%) of mf-EPSCs elicited at repetitive 3 Hz stimulation was only 1.11 ± 0.07 ms (Fig. 5A). Moreover, when strongly minimizing polysynaptic activity in the presence of 5 μM UBP310 to block postsynaptic KARs at mf-CA3 synapses and 250 nM NBQX to partially block AMPA receptors (74.4 ± 5.0% reduction in EPSC amplitude), the mean rise time was similar to the paired controls (1.04 ± 0.06 ms; Fig. 5A). Additionally, both the paired-pulse ratio (Fig. 5B,C) and frequency facilitation at 3 Hz (Fig. 5D,E) were unaffected by such experimental manipulation (see also Perrais et al. 2009). We also found that UBP316 had no effect on short-term synaptic plasticity at Mf-CA3 (frequency facilitation when changing the frequency of stimulation from 0.1 to 1 Hz: 809.7 ± 116.9% for ctr and 775.7 ± 98.2% for UBP316). We also recorded pharmacologically isolated KAR-EPSPs during long 25 Hz trains and found that, although it is possible to reach CA3 pyramidal cell firing through KAR activation alone, this can only occur after a high number of stimuli (>10) and is completely prevented by UBP310 (Fig. 5F). Most importantly, large presynaptic short-term facilitation was still observed when monitoring KAR-EPSCs (isolated with a low concentration of NBQX) but was significantly reduced in GluK3-deficient mice (Supplementary Fig. 5). Finally, we found that UBP310 affects neither evoked EPSPs recorded from CA3 pyramidal cells upon stimulation of the A/C pathway, consistent with the lack on KARs on these synapses, nor the frequency of spontaneous EPSCs (corresponding to all glutamatergic synaptic inputs to CA3 pyramidal cells; Supplementary Fig. 6).
UBP310 Antagonizes Postsynaptic KARs at Recurrent MF Synapses in Pilocarpine-Treated Rats
In animal models of temporal lobe epilepsies and in human patients, neuronal tissue undergoes major reorganization. This reactive plasticity is well documented in the dentate gyrus where MFs sprout (Tauck and Nadler 1985; Represa et al. 1987; Dudek and Sutula 2007) and create a powerful excitatory network between DGCs (Tauck and Nadler 1985; Dudek and Sutula 2007). Besides the axonal rewiring, recurrent MFs convert the nature of the glutamatergic transmission in the dentate gyrus because they operate via long-lasting KAR-mediated synaptic events not present in the naive condition (Epsztein et al. 2005). Here, we investigated the efficiency of UBP310 on AMPAR- and KAR-EPSCs in DGCs from control and chronic epileptic rats. First, we showed that EPSCs recorded in DGCs of control rats, with no detectable KAR-mediated component (Epsztein et al. 2005), were insensitive to bath application of 5 μM UBP310 (mean amplitude changed by −0.8 ± 8.5%, n = 11, P = 0.809; Fig. 6). By contrast, we observed that bath application of 5 μM UBP310 strongly reduced isolated KAR-EPSCs recorded in DGCs of chronic epileptic rats (mean amplitude reduced by 79.7 ± 5.6%, n = 6, P < 0.001; Fig. 6). Therefore, UBP310 is a strong inhibitor of KAR-EPSCs recorded in DGCs of chronic epileptic rats while it fully spares AMPAR-EPSCs.
This study identifies a potent antagonist of heteromeric GluK2/GluK5 and of postsynaptic KARs at mf-CA3 synapses, which spares AMPARs and NMDARs. With this tool in hand, it is now possible to dissect the functional role of postsynaptic KARs in major neuronal populations, such as CA3 pyramidal cells, in physiological and pathological conditions.
Identification of an Antagonist for a Major Population of KARs in the Brain
GluK2/GluK5 receptors are known to constitute a major population of KARs in the brain: both subunits are abundantly coexpressed in the cerebellum, neocortex, striatum, amygdala, and hippocampus, at higher levels than other KAR subunits (e.g., Bureau et al. 1999). We identified GluK2/GluK5 as a possible target for the action of UBP310 at mf-CA3 synapses by analyzing the effects of this compound in recombinantly expressed receptors. Interestingly, UBP310 acted on GluK2/GluK5 with similar potency than on native postsynaptic KARs (750 vs. 250 nM, respectively) but was much less active on homomeric GluK2. The antagonistic action of UBP310 on GluK2/K5 receptors was best observed when short, “synaptic-like,” pulses of glutamate were used, which induce slow decaying currents (Barberis et al. 2008). With longer pulses of glutamate, which induce fast receptor desensitization, UBP310 largely increased current rise and decay times without blocking receptor currents. This might explain why UBP310 was not identified in previous high-throughput screenings for antagonists (Alt et al. 2004). The sensitivity of GluK2/GluK5, but not GluK2, to UBP310 may suggest selective binding of the compound only on the GluK5 glutamate-binding site. Interestingly, the effect of UBP310 appears to be restricted to the ionotropic function of KARs, whereas the regulation of IsAHP by synaptic activation of GluK5-containing KARs was not affected. The structural basis for the differential effect of UBP310 on GluK2 and GluK2/GluK5 is unclear. The recently proposed scheme for the activation of heteromeric KARs (Mott et al. 2010), whereby high-affinity agonist binding to a nondesensitizing GluK4 (or GluK5) subunit opens the heteromeric channel, whereas low-affinity agonist binding to GluK2 desensitizes the channel complex, may apply to the mode of action of UBP310 on GluK2/GluK5. The large steady state current observed in the presence of UBP310 in response to prolonged glutamate application is consistent with the activation of GluK5 by glutamate (Mott et al. 2010; Fisher and Mott 2011). This would indicate that UBP310 preferentially binds to the glutamate-binding site on GluK2, but not on GluK5. However, this contradicts crystalographic studies of GluK1- and GluK2-binding sites in complex with UBP310, which have provided a structural explanation for the differential antagonist sensitivity (Mayer et al. 2006). However, the hypothesis that UBP310 binds to GluK2 fits with the proposal that the GluK5 subunit is responsible for the metabotropic action of KARs (Ruiz et al. 2005). Even though UBP310 antagonizes recombinant GluK2/GluK5 in a manner consistent with the block of synaptic KAR-mediated currents, it would nevertheless be interesting to test the impact of the auxiliary subunit Neto1 on the pharmacology of recombinant GluK2/GluK5 receptors. Indeed, Neto1 is associated with native KARs at Mf-CA3 synapses as well as with GluK2/GluK5 heteromers (Zhang et al. 2009; Copits et al. 2011; Straub et al. 2011). The strong activity of UBP302 and UBP316 at blocking postsynaptic KARs at mf-CA3 synapses suggests that all compounds of the UBP series could behave similarly. UBP310 was identified as a potent antagonist of GluK1 (Dolman et al. 2007) and of GluK3 (Perrais et al. 2009), and it may antagonize other heteromeric combinations of KARs apart from GluK2/GluK5, questioning whether UBP-sensitive physiological processes (e.g., Huxter et al. 2007; Wondolowski and Frerking 2009) are mediated only by GluK1-containing KARs. UBP310 may not be an antagonist for other heteromeric combinations of KARs, as for instance Gluk2/GluK3, a putative component of presynaptic KARs at Mf-CA3 synapses (Pinheiro et al. 2007).
Postsynaptic KARs and Synaptic Integration
Simulations have led to the proposal that postsynaptic KARs might have a strong influence on membrane potential and that AMPA receptors and KARs differ in their ability to encode temporal information (Frerking and Ohliger-Frerking 2002). This was directly demonstrated in hippocampal interneurons (Goldin et al. 2007; Yang et al. 2007) where GluK1 is likely a major player (Cossart et al. 1998; Frerking et al. 1998; Mulle et al. 2000). Furthermore, it was recently shown that aberrant KAR-operated synapses drastically increase the firing pattern of DGCs in TLE (Artinian et al. 2011). The role of KAR-EPSCs in synaptic integration and plasticity in CA3 pyramidal cells has, however, remained speculative due to the lack of tools to selectively interfere with their function. In addition, KARs play a major role in presynaptic short-term plasticity at mf-CA3 synapses which might be difficult to disentangle from postsynaptic KARs (see below). Here, we directly tested the physiological role of postsynaptic KARs by acutely blocking their contribution to mf-EPSPs. Consistent with the original assumptions, we found that postsynaptic KARs contribute a prolonged depolarization to mf-EPSPs. However, we found no significant effect of UBP310 on spike transfer at mf-CA3 synapses. This was rather unexpected, considering that we used frequencies that should maximize the effects of KARs on the EPSP waveform. The ionotropic action of postsynaptic KARs plays little direct role in spike transmission at mf-CA3 synapse likely because it is overcome by the powerful presynaptic short-term facilitation. However, we demonstrate that the sustained depolarization mediated by postsynaptic KARs plays a role in heterosynaptic facilitation of spike transmission through additional inputs, such as those from the A/C pathway. In addition, postsynaptic KARs may play a role in heterosynaptic long-term plasticity (Kobayashi and Poo 2004; Sachidhanandam et al. 2009).
Pre- Versus Postsynaptic KARs at Hippocampal MF Synapses
Many studies showed, through both pharmacological and genetic approaches, that presynaptic KARs significantly contribute to the large dynamic range of short-term plasticity at mf-CA3 synapses (Contractor et al. 2001; Lauri et al. 2001; Schmitz et al. 2001, 2003; Sachidhanandam et al. 2009). This view has, however, been challenged; a recent study attributed KAR-dependent presynaptic facilitation to postsynaptic KARs driving CA3 pyramidal cell spiking, hence generating polysynaptic activity that would contribute to short-term plasticity (Kwon and Castillo 2008). Although no consensus exists yet about the subunit composition of presynaptic KARs at mf-CA3 synapses, such study contradicts a large number of studies in favor of their existence (Pinheiro and Mulle 2008). We demonstrate that blocking postsynaptic KARs with UBP310 has no effect on presynaptic facilitation (see also Perrais et al. 2009) and does not affect the frequency of spontaneous EPSCs. This pharmacological argument comes in addition to the comparison of genetically modified mice, which demonstrates presynaptic deficits without postsynaptic changes (GluK3−/− mice; Pinheiro et al. 2007), or conversely the loss of postsynaptic KARs without any presynaptic impairment (Fernandes et al. 2009). Thus, our results decisively argue against the implication of postsynaptic KARs, and/or polysynaptic activity generated through their activation, in presynaptic facilitation at mf-CA3 synapses.
Overall, the identification of UBP310 as an antagonist of postsynaptic KARs and of recombinant GluK2/GluK5 likely represents a decisive step in the development of compounds to delineate the physiological roles of KARs in the regulation of synaptic networks and higher brain function. In addition, we provide evidence that UBP310 could serve to better understand the role of KARs in the mechanisms of TLE. Our data further suggest that aberrant postsynaptic KARs present at recurrent MF synapses in TLE may be a target for antagonists of GluK2/GluK5 receptors, which need to be further developed.
Centre National de la Recherche Scientifique; Fondation pour la Recherche Medicale; Conseil Régional d'Aquitaine; European Commission (EUSynapse Project, contract no. LSHM-CT-2005-019055); and Agence Nationale de la Recherche (contract KAREP).
Conflict of Interest: None declared.