Address correspondence and reprints requests to Dr Antonio Rodríguez-Moreno, Laboratory of Cellular Neuroscience and Plasticity, Department of Physiology, Anatomy and Cellular Biology, University Pablo de Olavide, Ctra. de Utrera, Km. 1, 41013, Sevilla, Spain. E-mail: email@example.com or Dr Talvinder S Sihra, Department of Neuroscience, Physiology and Pharmacology, University College London, Gower Street, London, WC1E 6BT, UK. E-mail: firstname.lastname@example.org
We have investigated the mechanisms underlying the facilitatory modulation mediated by kainate receptor (KAR) activation in the cortex, using isolated nerve terminals (synaptosomes) and slice preparations. In cortical nerve terminals, kainate (KA, 100 μM) produced an increase in 4-aminopyridine (4-AP)-evoked glutamate release. In thalamocortical slices, KA (1 μM) produced an increase in the amplitude of evoked excitatory post-synaptic currents (eEPSCs) at synapses established between thalamic axon terminals from the ventrobasal nucleus onto stellate neurons of L4 of the somatosensory cortex. In both, synaptosomes and slices, the effect of KA was antagonized by 6-cyano-7-nitroquinoxaline-2,3-dione, and persisted after pre-treatment with a cocktail of antagonists of other receptors whose activation could potentially have produced facilitation of release indirectly. Mechanistically, the observed effects of KA appear to be congruent in synaptosomal and slice preparations. Thus, the facilitation by KA of synaptosomal glutamate release and thalamocortical synaptic transmission were suppressed by the inhibition of protein kinase A and occluded by the stimulation of adenylyl cyclase. Dissecting this G-protein-independent regulation further in thalamocortical slices, the KAR-mediated facilitation of synaptic transmission was found to be sensitive to the block of Ca2+ permeant KARs by philanthotoxin. Intriguingly, the synaptic facilitation was abrogated by depletion of intracellular Ca2+ stores by thapsigargin, or inhibition of Ca2+-induced Ca2+-release by ryanodine. Thus, the KA-mediated modulation was contingent on both Ca2+ entry through Ca2+-permeable KARs and liberation of intracellular Ca2+ stores. Finally, sensitivity to W-7 indicated that the increased cytosolic [Ca2+] underpinning KAR-mediated regulation of synaptic transmission at thalamocortical synapses, requires downstream activation of calmodulin. We conclude that neocortical pre-synaptic KARs mediate the facilitation of glutamate release and synaptic transmission by a Ca2+-calmodulin dependent activation of an adenylyl cyclase/cAMP/protein kinase A signalling cascade, independent of G-protein involvement.
Ca2+/calmodulin involvement in pre-synaptic kainate receptors facilitation of glutamate release at thalamocortical synapses. We determined the mechanism by which kainate receptors (KARs) mediate a facilitation of glutamate release at thalamo-L4 cortical cell synapses. We find that an increase in cytosolic Ca2+ concentration, contingent on extra- and intra-cellular sources, in the pre-synaptic thalamic neuron, operates through the formation of Ca2+-calmodulin complexes to activate an AC/cAMP/PKA signalling pathway. This action of KARs may support learning and memory processes.
Glutamate receptors of the kainate-type are well established as mediators of ionotropic post-synaptic synaptic transmission. Pre-synaptically, kainate receptors (KARs) have a modulatory role in regulating neurotransmitter release. In the latter regard, KARs have been shown to have a metabotropic capacity whereby they effect the regulation of both glutamate and GABA release (see Rodríguez-Moreno and Sihra 2007a,b; Jane et al. 2009 for review). At certain excitatory glutamatergic synapses, KARs activation can actually effect biphasic modulation, whereby low agonist concentrations facilitate glutamate release, while high concentrations decrease the release of the neurotransmitter (see Rodríguez-Moreno and Sihra 2007a,b for review). Mechanistic details of how this is achieved are subject of investigation and, indeed, the subcellular location of KARs responsible for pre-synaptic modulation remains contentious.
KARs are known to be highly expressed at somatosensory synapses (Bettler and Mulle 1995; Kerchner et al. 2001; Daw et al. 2007), but the precise actions of KARs at thalamocortical projection synapses require elucidation. Thalamocortical inputs have been described to participate in short-term depression (Kidd et al. 2002), but more recently, activation of KARs at synapses between axon terminals of the ventrobasal thalamus and stellate cells in Layer 4 (L4) of the somatosensory barrel cortex, has been shown to mediate a bi-directional modulation of transmission. Thus, at ‘low’ concentrations, KA (3 μM) produced an increase in glutamate release, while higher concentrations of agonist (5–30 μM) depressed neurotransmitter release (Jouhanneau et al. 2011); the extent of KAR activation evidently determining the mode of regulation. The mechanisms that mediate these diametrically opposite actions of KA at thalamocortical synapses are unclear.
Here, we have examined the mechanism underlying the facilitatory effects of KA using two preparations. Firstly, we examined the effect of KA on glutamate release from isolated cerebral nerve terminals (synaptosomes). Utilization of isolated and purified cortical nerve terminal obviates any confounding somatodendtric (post-synaptic) effects of KA on glutamate release and allows specific establishment of the functioning of definitively nerve terminal resident KARs. Secondly, we sought to elucidate the mechanistic details in a defined thalamocortical projection more specifically, and using a more intact synaptic preparation. We utilized thalamocortical slices incorporating synapses established between axon terminals from ventrobasal thalamus and stellate cells in L4 of the somatosensory barrel cortex in adult mice. These synapses are known to be important for plasticity during development, and for somatosensory integration and network activity (see Feldman et al. 1999 for review).
Using the two complementary preparations, we first established that there are common mechanistic features involved in the facilitation of glutamate release/synaptic transmission. Thus, we found that the KARs effecting facilitation of glutamate release and synaptic transmission in synaptosomes and slices have a congruent pharmacological profile, with both showing a mandatory dependence on adenylyl cyclase (AC) and cAMP-mediated protein kinase A (PKA) activity. Extending the mechanistic dissection with thalamocortical synapses in slices, we demonstrated that the KAR-mediated facilitation of transmission is contingent on both external Ca2+ permeation into the cytosol through KARs, and downstream intracellular Ca2+-store mobilization. Finally, a major sensitivity of thalamocortical facilitation to calmodulin inhibition suggests that KARs are coupled through a Ca2+-calmodulin/AC/cAMP/PKA pathway in thalamic terminals synapsing onto stellate L4 cortical neurons.
Experiments were performed with adult (2 months old) C57Bl/6 male mice obtained from Harlan Laboratories (Spain). Experiments were carried out according to the European Union directive (609/86/EU) for the use of laboratory animals in acute experiments and were approved by the local Ethical Committee.
Preparation of synaptosomes
Synaptosomes were prepared from cerebral cortices as described previously (Sihra 1997). The final synaptosomal fraction was resuspended in HEPES-buffered incubation medium containing (mM): 140 NaCl, 5 KCl, 5 NaHCO3, 1 MgCl2·6H2O, 1.2 Na2HPO4, 10 glucose, 20 HEPES (pH 7.4). Protein concentration was then determined using a Bradford assay. Synaptosomes were centrifuged in the final wash to obtain synaptosomal pellets with 0.5 mg protein. Synaptosomal pellets were stored on ice and used within 1–2 h.
Glutamate release assay
Glutamate release was assayed by online fluorometry (Nicholls and Sihra 1986). Pelleted synaptosomes were resuspended at a protein concentration of 0.5 mg/mL in HEPES-buffered incubation medium containing 16 μM bovine serum albumin and incubated in a stirred and thermostatted cuvette at 37°C in a Perkin-Elmer LS-3B spectrofluorimeter (PerkinElmer, Waltham, MA, USA). NADP+ (1 mM), glutamate dehydrogenase (50 units/mL) and CaCl2 (1 mM) were added after 3 min. After a further 10 min of incubation, 1 mM 4-aminopyridine (4-AP) was added to stimulate glutamate release. The oxidative deamination of released glutamate, leading to the reduction of NADP+, was monitored by measuring NADPH fluorescence at excitation and emission wavelengths of 340 and 460 nm respectively. Data were accumulated at 2-s intervals. A standard of exogenous glutamate (5 nmol) was added at the end of each experiment and the fluorescence change produced by the standard addition was used to calculate the released glutamate as nanomoles glutamate per milligram synaptosomal protein. Release traces are shifted vertically to align the point of depolarization as zero release. Release values quoted in the text are levels attained at ‘steady-state’ after 4 min of depolarization (nmol/mg protein/4 min).
Thalamocortical slices (350 μm thick) were prepared according to the methodology established by Agmon and Connors (1991) and utilized in our previous studies (Rodríguez-Moreno and Paulsen 2008; Banerjee et al. 2009; Rodríguez-Moreno et al. 2011). Briefly, after decapitation, the whole brain was removed under ice-cold buffered salt solution consisting of (in mM) 124 NaCl, 2.69 KCl, 1.25 KH2PO4, 2 MgSO4, 1.8 CaCl2, 26 NaHCO3, and 10 glucose (pH 7.2, 300 mOsm), and positioned on the stage of a vibratome slicer and cut to obtain thalamocortical slices (350 μm thick), which were maintained continuously oxygenated for at least 1 h before use. All experiments were carried out at 22–25°C. During experiments, slices were continuously perfused with buffered salt solution as detailed above. Drugs were applied by switching between separate perfusion lines.
Barrels were identified under a stereomicroscope. Whole-cell patch-clamp recordings were made from stellate neurons of the layer 4 field of the somatosensory cortex. NMDA receptor-mediated (or AMPA receptor-mediated when indicated) evoked excitatory post-synaptic currents (eEPSCs) were recorded from these neurons visually identified by IR-DIC microscopy using a 40× water immersion objective. Perfusion solution contained GYKI53655 (30 μM), to block AMPA receptors, and bicuculline (20 μM), to block GABAA receptors. In experiments involving AMPA receptor-mediated currents, no GYKI53655 was used, but D-AP5 (50 μM) was included to block NMDA receptors. To evoke eEPSCs, electrical pulses were delivered to thalamic axons using a monopolar electrode placed in the fibre tract from the ventrobasal thalamus (Agmon and Connors 1991; Crair and Malenka 1995; Daw et al. 2006; Jouhanneau et al. 2011), delivering electrical pulses (constant stimulation current) at a frequency of 0.2 Hz. Responses were identified by the typical paired-pulse depression shown at the 40 ms time interval (Lee and Sherman 2008). Patch electrodes were made from borosilicate glass, and had a resistance of 4–7 MΩ when filled with (mM): 120 CsCl, 8 NaCl, 1 MgCl2, 0.2 CaCl2, 10 HEPES, 2 EGTA and 20 QX-314 (pH 7.2, 290 mOsm). Experiments were performed at +40 mV. A 40 ms paired-pulse stimulation protocol was used for paired-pulse ratio (PPR) analysis. Neurons were voltage clamped, using a Multiclamp 700B amplifier (Molecular Devices, Foster City, CA, USA). Access resistance (15–30 MΩ) was regularly monitored during recordings (calculated from the instantaneous response to a negative voltage step) and cells were rejected if it changed > 15% during the experiment. Junction potentials were not corrected. Data were filtered at 2 kHz, digitized at 5 kHz, and stored on a computer using pClamp software (Molecular Devices).
The noise-free coefficient of variation (CV) was calculated as:
Where σ2(EPSC) and σ2(noise) are the variance of the excitatory post-synaptic currents (EPSC) and baseline respectively. For each cell CVKA/CVControl was obtained. The plots comparing variation in EPSC mean amplitude (M) to the change in response variance of the EPSC amplitude were constructed as described in Bekkers and Stevens (1990) and Malinow and Tsien (1990); see Siegelbaum and Kandel (1991 for comprehensive explanation. Responses used to calculate the CV were very stable. Stability was defined as there being no change in slope over at least 10 min of the baseline (assessed comparing amplitudes and slopes of the responses every minute), with the effect of KA being evaluated at peak response. Only the cells that passed all the quality checks (including no series resistance changes, etc.) were included in the analysis. For this method to be applicable the quantal parameters are assumed to be invariable and the EPSC amplitude distribution must be described by a binomial. This means that the synaptic variance reflects the probabilistic release of neurotransmitter, that is, there is quantal variance. We could not directly test whether our data fit to a compound binomial distribution, but synaptic fluctuations were evident in most of the cells studied; depending on the time to the peak effect of KA, we generally used 200–280 responses to calculate CV (120–180 of control responses + 80–100 of responses after KA treatment). Thus, we assumed that synaptic release in our experimental situation did follow a binomial distribution. It has been established that a change in M with no change in the variance of the synaptic response denotes that only quantal content has been modified (i.e., it is a purely postsynaptic mechanism) and the experimental data are expected to be situated through a horizontal line. In contrast, if a change in the variance correlates with the change in M, it is generally suggestive of a pre-synaptic modification in release parameters.
Data were normalized taking the control as 100% of the response and presented as mean ± SEM. Signals were averaged every 12 traces. Effects of KA were measured at peak (maximum) effect compared to averaged 10 min baseline points. Significance was assessed at p < 0.05. Statistical comparisons were made using two-tailed Student's t-test for comparison of two data sets and anova for comparison of multiple data sets.
Kainate, PTx, salts and general reagents were purchased from Sigma (St. Louis, MO, USA); GYKI 53655, SYM2206, D-AP5, 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), Bicuculline, (RS)-alpha-methyl-4-carboxyphenylglycine (MCPG), (RS)-alpha-methyl-4-phosphonophenylglycine (MPPG), naloxone, 2-OH saclofen, atropine sulphate, 8-Cyclopentyl-1,3-dipropylxanthine (DPCPX), H-89, Rp-Br-cAMP, forskolin, dideoxyforskolin, IBMX, NEM, philantotoxin, ryanodin, thapsigargin, CMZ and W-7 were obtained from Tocris (Bristol, UK).
KAR activation facilitates glutamate release from cortical synaptosomes
We first established the mechanistic profile of the facilitatory actions of KA using an on-line enzymatic assay for measuring glutamate release (Nicholls and Sihra 1986) from isolated cerebrocortical nerve terminals (synaptosomes). As described previously (Perkinton and Sihra 1999), KA (100 μM) mediated a facilitation of glutamate release (Fig. 1a). Quantitation revealed that the application of 100 μM KA produced a robust and statistically significant increase in 1 mM 4-AP-evoked glutamate release (46 ± 6%, n = 12, Fig. 1b). As described previously in the hippocampus (Rodríguez-Moreno and Sihra 2004), no decrease in glutamate release was observed after KA application and, in all cases, only facilitation was seen. In these synaptosomes, metabotropic pathways mediating decrease of glutamate release following activation of a variety of inhibitory G-protein-coupled receptors have been expounded and shown to be sensitive to G-protein inhibition (Herrero et al. 1996; Wang and Sihra 2003).
In all release experiments with KA, we routinely used the non-competitive AMPA receptor antagonists GYKI53655 (30 μM) or SYM2206 (100 μM, in n = 3 control experiments assaying the effect of KA) to block any cross-activation of AMPA receptors by KA. Under these conditions, the AMPA/KA receptor blocker CNQX effectively operates as a selective KAR antagonist. Accordingly, in the presence of CNQX (100 μM), the facilitatory effect of KA was abolished (3 ± 5% increase, n = 5; Fig. 1a and b), confirming that the facilitation of glutamate release occurs by the specific activation of nerve terminal KARs, as previously suggested (Perkinton and Sihra 1999). However, there remained the possibility that the effect of KA on glutamate release from cerebral cortex synaptosomes may be indirect, given that a number of other pre-synaptic receptors may control the release of glutamate, with the neurotransmitter ligands for these receptors being released secondarily in response to KA application. To evaluate the potential participation of the most prominent of these receptors in the present experiments, we treated synaptosomes with an inhibitor cocktail which included the receptor antagonists: MCPG (1.5 mM)/MPPG (1.5 mM), naloxone (100 μM), bicuculline (20 μM), 2-OH-saclofen (150 μM), atropine sulphate (50 μM) and DPCPX (0.1 μM), to block metabotropic glutamate, opioid, GABAA, GABAB, muscarinic and adenosine receptors respectively. Under these conditions, KA effected facilitation indistinguishable from that in the absence of the inhibitor cocktail (40 ± 5%, n = 5, KA-mediated facilitation of 4-AP-evoked release in presence of the cocktail vs. 43 ± 6%, n = 12, in the absence of the cocktail; Fig. 1b). These data therefore indicated that the observed facilitation was attributable to a direct and selective action of pre-synaptic KARs on glutamate release, rather than indirect modulation through major transmitters that might secondarily activate one or more of the aforementioned receptors.
The KA-induced facilitation of glutamate release involves the cAMP/PKA cascade in cerebral cortex synaptosomes
Having confirmed the selectivity of the action of KA on glutamate release from cortical synaptosomes, we further explored the mechanism underlying the modulation. Our previous studies with hippocampal synaptosomes have pointed to the involvement of the cAMP cascade in the facilitation of glutamate release elicited by KA (Rodríguez-Moreno and Sihra 2004). We therefore analysed the effects of activating or inhibiting the AC/cAMP/PKA pathway on the KA-mediated facilitation of glutamate release from cerebrocortical synaptosomes. We first examined the effect of directly inhibiting PKA using the cell-permeable and PKA inhibitor H-89. In the presence of H-89 (100 μM), the facilitation of 4AP-evoked glutamate release mediated by the activation of KARs was prevented (4 ± 3%, n = 5, vs. 36 ± 6%, n = 5 in untreated synaptosomes, Fig. 2a and b). As H-89 at this relatively high concentration might not act selectively on PKA, to be sure the effect of H-89 reflects a PKA contingent regulation, we repeated the experiment with synaptosomes treated with the cell-permeable PKA inhibitor, Rp-Br-cAMP (100 μM). In this experimental condition, the facilitatory effect of KA on glutamate release was again prevented (2 ± 5%, n = 6, Fig. 2b). Together, these results suggest the involvement of PKA in the facilitatory action of KA.
To corroborate the involvement of AC/cAMP/PKA signalling in the KA-mediated facilitation, we next took the approach of enhancing this pathway by increasing second messenger cAMP production through direct stimulation of AC with forskolin. Forskolin (100 μM) + IBMX (50 μM, to abrogate phosphodiesterase activity) caused a 30 ± 3% (n = 5) facilitation of glutamate release (Fig. 2a and b). In the presence of forskolin and IBMX, KA (100 μM) only induced a further 10 ± 4% (n = 5) increase in the facilitation glutamate release. Clearly therefore, activation of AC by forskolin somewhat occludes the facilitatory action mediated by the KAR activation (Fig. 2a and b).
While patently not a G-protein coupled receptor, some of the modulatory effects of KARs have been attributed to metabotropic mechanisms involving G-proteins (Rodríguez-Moreno and Sihra 2007a,b). The question remained, therefore, is the AC/cAMP/PKA cascade espoused herein instigated by G-protein coupling to AC? Intriguingly, however, we observed no effect of G-protein inhibition with pertussis toxin (PTx, 35 ± 6%, n = 6) or NEM (38 ± 9%, n = 5, vs. 40 ± 7%, n = 5) on the facilitation of glutamate release mediated by KAR activation (Fig. 2c) (cf. sensitivity of inhibitory mGluR in synaptosomes to G-protein inhibition, Herrero et al. 1996).
The activation of KARs by 1 μM KA produces an increase in the amplitude of NMDA-evoked post-synaptic currents in thalamocortical slices
Having established that the facilitatory pre-synaptic KARs present in a hetereogenous population of nerve terminals (synaptosomes) is mechanistically underpinned by AC/cAMP/PKA signalling, but independently of G-protein activity, we sought to establish whether this regulation and mechanism occurred at a defined and topographically intact cortical synapse. Pursuant to the recent demonstration of the biphasic modulation of glutamatergic transmission by KA at thalamocortical synapses in juvenile mouse pups (3 μM – facilitation, 5–30 μM – depression; Jouhanneau et al. 2011), akin to that previously shown in the hippocampus (see Rodríguez-Moreno and Sihra 2007a,b; Jane et al. 2009 for review), we established the thalamocortical synapse experimental paradigm in adult animals.
In slices obtained from mature mouse cerebra, we stimulated axons in the ventrobasal thalamus and measured NMDA receptor-mediated eEPSCs in layer 4 (L4) somatosensory barrel cortex stellate neurons, by whole-cell patch clamp recordings (in the presence of 30 μM GYKI53655, to selectively block AMPA receptors and 20 μM bicuculline, to block GABAA receptors, and membrane potential held at + 40 mV). In our hands, adult thalamocortical synapses began to produce measurable facilitation of NMDA receptor-mediated eEPSCs amplitudes at 1 μM KA (46 ± 16%, n = 8, Fig. 3a and b), with 100 nM, 300 nM and 500 nM agonist having no effect (96 ± 5%, n = 6, 99 ± 6%, n = 5, and 98 ± 8%, n = 5, respectively, Fig. 3b). With 1 μM KA, synaptic facilitation was followed by a 22 ± 8% (n = 8) decrease in the eEPSCs amplitude (Fig. 3b). A similar biphasic pattern was evident with 3 μM KA (facilitation – 49 ± 15%, n = 8; depression – 30 ± 6%, n = 5), but 5 μM and 10 μM KA only produced decrease in eEPSCs amplitude (50 ± 5, n = 6 and 60 ± 7%, n = 5 respectively). To analyse the mechanisms involved in the KAR-mediated facilitation of synaptic transmission specifically, we used 1 μM KA in subsequent electrophysiological experiments. In these experiments KA 1 μM produced a change in holding current of 15 ± 6 pA in four of eight neurons recorded.
To determine whether the action of KARs responsible for the increase in eEPSCs amplitude in slices recorded from layer 4 stellate neurons, is at least mechanistically akin to that observed in synaptosomes, we compared the pharmacological profile of the regulation. First, we found that the 1 μM KA-mediated biphasic effect on the eEPSC amplitude (facilitation – 46 ± 16%, n = 8; depression – 22 ± 8%, n = 8) was abolished in the presence of 100 μM CNQX (97 ± 5%, n = 7 for the facilitation, and 98 ± 5% for the depression, Fig. 3a and c). Given that AMPA receptors were already blocked by the presence of the antagonist GYKI53655 in the slice bath, the complete antagonism by CNQX represents a specific effect on KARs. These data indicate that the facilitation (and the depression) of synaptic transmission observed at thalamocortical synapses is exclusively contingent on KAR activation, similar to the case with synaptosomes. Also pharmacologically similar to synaptosomes, KA-mediated facilitation in slices was still present (50 ± 7% facilitation of eEPSCs amplitude, n = 6) in the presence of a cocktail of neurotransmitter receptor antagonists with the same composition as that used in synaptosomal experiments (not shown). Notably, the depression that followed the facilitation of EPSCs, was also observed in the presence of the cocktail (26 ± 6%, n = 6). This again served to exclude the possibility that the KA-mediated modulation was secondary to the release of other neurotransmitters, but was rather indicative of a direct effect of KA on KARs at thalamocortical synapses.
Previous studies have attributed the pre-synaptic origin of the facilitatory action of KA at thalamocortical synapses by measuring changes in direct current during recordings (Jouhanneau et al. 2011). Here, to differentiate between the pre- or post-synaptic location of the modulation, that is, elucidate whether the KARs mediating the facilitation of glutamate release are situated and operational in the pre-synaptic thalamic neurons versus the target, post-synaptic cortical L4 cells, we used several approaches and analyses. Firstly, we measured the PPR after performing paired-pulse recordings (pair-pulse depression was observed at 40 ms pulse interval – typical of the thalamocortical synapse studied, Lee and Sherman 2008). Under control/baseline conditions, the PPR was 0.7 ± 0.05 (n = 8). Following KA treatment, the PPR rose to 1.3 ± 0.1 (n = 8; Fig. 3d), suggesting a change in release probability (Manabe et al. 1993), and thereby invoking a pre-synaptic locus of action of the KA-mediated regulation. Secondly, we examined the change in the coefficient of variation (CV) of synaptic responses versus the change in the amplitude of both NMDA- and AMPA-receptor mediated eEPSCs. We observed that the increase in the mean eEPSC amplitude was paralleled by an increase in the 1/CV2 parameter, which in most cases is suggestive of a pre-synaptic effect of KA (Fig. 3e). Thirdly, we determined the effect of KA on the proportion of synaptic failures observed. The synaptic failure rate was 50 ± 7% (n = 4) under control conditions. After KA, the rate of these failures was clearly decreased (to 21 ± 6%, n = 4, Fig. 3f); once again implicit of a pre-synaptic action of KA. Finally, we compared the effects of KA on NMDA receptor-mediated eEPSCs (in the presence of GYKI53655) and AMPA receptor-mediated eEPSCs (recorded at −70 mV, in the absence of GYKI53655, but in the presence of D-AP5 and bicuculline, to block NMDA and GABAA receptors respectively). KA produced a similar degree of increase of the NMDA receptor-mediated eEPSC (135 ± 6%, n = 5, Fig. 3g) as the AMPA receptor-mediated eEPSC (131 ± 6%, n = 10, Fig. 3g). This congruent facilitation of both NMDA and AMPA receptor-mediated eEPSC amplitudes suggests that KA impinges upstream of receptor activation, thus likely by increasing the release of glutamate. Altogether, the foregoing analyses consistently point to a pre-synaptic locus of action for KA at the thalamocortical synapse being examined. Notwithstanding, it remains to be seen in which pre-synaptic compartment, viz. terminal, axonal or somatodendritic, the respondent KARs are physically located.
KAR-mediated facilitation of synaptic transmission at thalamocortical synapses involves a cAMP-dependent cascade
Having confirmed the selectivity of the action of KA, we further explored the possibility that, mechanistically, the same second messenger system that acts to mediate effects in synaptosomes, also operates in the facilitation of eEPSCs at thalamocortical synapses recorded in slices. Firstly, to test whether PKA was involved in the facilitation of eEPSCs observed (as with glutamate release in synaptosomes), we inhibited the cAMP-dependent activation of PKA or the kinase catalytic activity itself by treating the slice with Rp-Br-cAMP or H-89 respectively. In the presence of 100 μM Rp-Br-cAMP or 2 μM H-89, the modulatory effect of 1 μM KA on the eEPSCs amplitude was abolished (99 ± 3%, n = 5 after Rp-Br-cAMP and 104 ± 5%, n = 5 after H-89, Fig. 4a and b). Rp-Br-cAMP and H-89 alone caused small decreases of the eEPSCs amplitude (9 ± 4% and 12 ± 5%, respectively, n = 5, data not shown). These data evince that, as with synaptosomes, PKA also plays an essential role in the KA-mediated modulation of thalamocortical synaptic transmission.
Next, we looked at the effect of AC activation by forskolin on the KA-mediated facilitation of eEPSCs. As the effects of forskolin are long-lasting (Tong et al. 1996), experiments were done in pre-incubated slices. The application of KA onto forskolin (30 μM) + IBMX (5 μM)-treated slices failed to affect eEPSCs amplitude (106 ± 7%, n = 6, Fig. 4b). This result therefore indicated that the prior activation of AC by forskolin also occludes the action mediated by KARs at thalamocortical synapses recorded in slices, similar in character to observations with cortical synaptosomes. Again, as done previously in synaptosomes (Wang and Sihra 2003), to ensure that the observed forskolin effect in slices was indeed because of cAMP production rather than any non-specific effect(s) of the diterpene, we performed control experiments using 1,9-dideoxyforskolin, an inactive analogue of forskolin. In the presence of dideoxyforskolin (100 μM), KA (1 μM) produced a facilitation of 40 ± 5% (to 140 ± 5% of control, n = 4) in the amplitude of the eEPSCs. This confirms that the forskolin effect on KA-mediated modulation is attributable to increases in cAMP levels produced by the stimulation of AC. Together, these results suggest that forskolin, through an AC/cAMP/PKA pathway, occludes the action of KA at thalamocortical synapses.
The foregoing data in Fig. 4 support the hypothesis that the KA-mediated facilitation of thalamocortical transmission is contingent on an AC/cAMP/PKA signalling pathway, operating pre-synaptically in the thalamocortical neuron. However, given that protein kinase C has been shown to be involved in some aspects of the KAR-mediated modulation in other slice preparations (Rodríguez-Moreno and Sihra 2007a,b for reviews), we looked at whether the effect of KA at thalamocortical synapses involved this kinase. In slices treated with calphostin C (1 μM), a highly specific inhibitor of protein kinase C, we observed no significant change in the KA-mediated facilitation (34 ± 7%, n = 5, Fig. 4b) or depression (20 ± 5%, n = 5), therefore obviating a role for the kinase in the observed modulation. To determine whether the activation of KARs to induce facilitation of glutamate release involves G-protein functioning, we studied the effect of KA in slices treated with PTx (5 μg/mL). In this experimental condition, the facilitatory effect mediated by the activation of KARs was still present (144 ± 8%, n = 5, Fig. 4b). Importantly, providing a positive control for the activity of PTx in this and foregoing experiments with synaptosomes, we found that the inhibitory effect of KA was indeed suppressed by PTx, invoking G-protein involvement of that modulation, 94 ± 8%, n = 6).
Facilitation of thalamocortical synaptic transmission/glutamate release in slices is mediated by Ca2+ permeant pre-synaptic KARs, requires mobilization of intracellular Ca2+ stores and involves Ca2+/calmodulin
The role of Ca2+ mediating facilitation at hippocampal mossy fiber-CA3 synapses following KARs activation have been subject of debate and controversy. Some studies have suggested that permeation of Ca2+ through KARs and subsequent Ca2+-induced Ca2+ release (CICR) from intracellular stores is mandatory for short-term and long-term plasticity at these synapses (Lauri et al. 2003; Scott et al. 2008). However, other studies have reported no effect of KA on cytosolic [Ca2+] (Kamiya et al. 2002), and still others have espoused a decrease in levels of the cation in response to KAR activation (Kamiya and Ozawa 1998, 2000). To test for the former possibility and the presence of Ca2+ permeant KARs at thalamocortical synapses, we examined the effect of KA on the eEPSC amplitudes in slices treated with philanthotoxin, an agent known to block unedited, Ca2+ permeable KARs (Fletcher and Lodge 1996; Scott et al. 2008). In presence of 3 μM philanthotoxin, the facilitation of synaptic transmission mediated by 1 μM KA was totally prevented (to 95 ± 8% of initial amplitude, n = 5 vs. 135 ± 5%, n = 7 observed in interleaved slices; Fig. 5a and b). This result clearly indicated that Ca2+ permeation through KARs is obligatory for the synaptic facilitation observed in thalamocortical slices.
To determine whether the aforementioned externally originating Ca2+ signal produced by KAR activation required amplification by CICR from intraterminal Ca2+-stores, we examined the effect of KA after depleting the intracellular Ca2+-stores by inhibiting the SERCA pump therein with thapsigargin. Application of thapsigargin (2 μM) indeed abolished the facilitatory effect of 1 μM KA at thalamocortical synapses, and rather produced a depression (82 ± 13%, n = 5, with thapsigargin vs. 143 ± 4%, n = 8, without thapsigargin, in interleaved slices; Fig. 5c and d). These data are indicative of an absolute requirement for intracellular Ca2+ stores in the regulation elicited by KA, but the question remained as to whether the intracellular release of Ca2+ is triggered by Ca2+. We therefore tested the ability of ryanodine, which selectively inhibits CICR (Berridge 1998), to block KAR-mediated facilitation. Ryanodine (10 μM), prevented the KAR-mediated enhancement of transmission at thalamocortical synapse (108 ± 7%, n = 6, with ryanodine vs. 138 ± 7%, n = 5, without ryanodine, in interleaved slices; Fig. 5d). Collectively, these results suggest that pre-synaptic KARs at thalamocortical synapses are Ca2+ permeable, and that Ca2+ entering the terminals through KARs triggers CICR from intraterminal Ca2+ stores.
From the foregoing results in slices, it is evident that an increase in Ca2+ via KARs is necessary for the mediation of the facilitation produced by KA. Because, in the hippocampus, the activation of a Ca2+-calmodulin complex seems to be necessary for AC activation (Andrade-Talavera et al. 2012), we tested for this possibility at thalamocortical synapses by recording eEPSCs in slices treated with the calmodulin antagonist W-7. In the presence of W-7 (25 μM), KA (1 μM)-mediated facilitation was completely prevented (88 ± 5%, n = 5, with W-7 vs. 139 ± 13%, n = 8 without W7, in interleaved slices; Fig. 6a and b). To corroborate this calmodulin dependence, we performed the experiment in the presence of an alternative calmodulin antagonist, viz. calmidazolium (CMZ, 1 μM). As with W-7, in slices treated CMZ, KA (1 μM)-mediated facilitation of glutamate release was abolished (90 ± 10%, n = 5, Fig. 6b). These results indicate that a pre-synaptic Ca2+-calmodulin complex is necessary for the synaptic regulation by KA, upstream of activation of an AC/cAMP/PKA signalling cascade.
Our results, employing electrophysiological studies in thalamocortical slices, informed by biochemical studies in isolated cortical nerve terminals (synaptosomes), show that the activation of pre-synaptic KARs at some cortical synapses produces a facilitation of synaptic transmission/glutamate release, and suggest a mechanistic coupling involving Ca2+-calmodulin/AC/cAMP/PKA activity, but independent of any G-protein activation.
The observed KA-mediated enhancement of eEPSCs at thalamocortical synapses effected was because of increased glutamate release as a similar change in NMDA receptor-mediated current (in the presence of the AMPA receptor antagonist GYKI53655), and AMPA receptor-mediated current (in the presence of the NMDA receptor antagonist AP5) was recorded. That the KA-mediated modulation displayed some common characteristics pharmacologically in synaptosomes and at a specific synapse in thalamocortical slices (i.e. AC/cAMP/PKA involvement), lends support to a mechanistic commonality, while not necessarily attributing this KA modulation to a particular nerve terminal type in the former preparation. In both, synaptosome and slice models, the facilitation by KA was blocked by CNQX, under conditions where AMPA receptors were already antagonized by GYKI53655. This supports the specific role of KARs in the regulation, particularly as a cocktail of receptor antagonists, designed to inhibit the effects of secondarily release neurotransmitters, had no effect.
In considering synaptic regulation, it remains paramount to identify the subcellular location of the receptor postulated. The strength of the synaptosome model is that, with functional post-synaptic elements eliminated in the preparation, observed pre-synaptic actions by KA must, by definition, be because of nerve terminal resident KARs. Using thalamocortical slices, although modulation could be examined at a defined synapse, to gain further insight into the locus of KAR-mediated facilitation of synaptic transmission, further analysis was warranted. Thus, to determine the pre- or post-synaptic presence and activity of KARs at thalamocortical synapses, we employed electrophysiological analysis using four different approaches. Firstly, we analysed the pair-pulse ratio of consecutive eEPSCs in thalamocortical slices. A clear increase in paired-pulse ratio of eEPSCs observed with KA application was suggestive of a change in release probability (definitively a pre-synaptic component in synaptic transmission/regulation). Secondly, we observed that the increase in the mean eEPSCs amplitude was paralleled by an increase in 1/CV2, again suggesting a pre-synaptic effect of KA.
Thus, thirdly, we determined the effect of KA on the proportion of synaptic failures. When KA was present, the proportion of failures was clearly decreased, consistent with an increase in pre-synaptic transmitter release probability, in line with the facilitation seen. Finally, similar effects of KA observed for NMDA- and AMPA receptor-mediated currents are indicative of a pre-synaptic mode of action for KARs, as no equivalence would otherwise be expected if the modulation was post-synaptic. Altogether, all four of these independent analyses confirmed and emphasized a pre-synaptic locus of action of KARs operating at thalamocortical synapses.
The question as to whether, the pre-synaptic regulation by KARs in thalamocortical neurons reflects the activity of nerve terminal/axonal or somatodendritically localized receptor, requires further deliberation. Considering the possibility that modulation mediated by putative somatodentric KARs might reflect a change in the spike threshold, we have conducted some preliminary field recording experiments to address this issue. In these experiments, consistent with Jouhanneau et al. (2011), we found no correlation between changes in fibre volley and the effect of KA on the fEPSP amplitude, and indeed, the former operated with distinct temporal characteristics in comparison to the KA-dependent synaptic modulation reported. Thus, the indications are that changes of excitability cannot explain the facilitation by KA reported herein, but rather we are moved to hypothesize the presence and operation of terminal or axonal KARs in the regulation. While, additional electrophysiological analyses could offer some elaboration of the characteristics of the thalamocortical synapse along similar lines to Jouhanneau et al. (2011), subcellular electrophysiological or pharmacological isolation of KARs proves difficult (space clamp and agonist spillover considerations respectively). Therefore, the unequivocal delineation of KAR compartmentalization prompts more direct approaches. Although technically challenging paradigms outwith the remit of the present study, to explicitly address pre-synaptic compartmentalization of KARs, two approaches warranting future consideration are: (i) The use of immunogold-based receptor localization studies; contingent on the availability of KAR antibodies with sufficient specificity and avidity. (ii) The use of targetable caged-blockers of KARs; pending development of such reagents (cf NMDA receptors studies, Rodríguez-Moreno et al. 2011). In lieu of the necessary innovations to examine KAR compartmentalization in thalamocortical neurons, the use of cortical synaptosomes does at least allow the demonstration that the KAR-mediated signalling posited does in fact operate in pre-synaptic terminals, that is, a model devoid of other functional compartments, though the shortcoming here is that a homogeneous preparation of a single terminal type is not feasible at present.
Notwithstanding the still unresolved issue about the subcellular localization of pre-synaptic thalamocortical KARs, aligning the mechanistic details of the KA-mediated regulation in synaptosomes and thalamocortical slices showed remarkable congruence. As with our previous synaptosomal studies in the hippocampus (Rodríguez-Moreno and Sihra 2004), inhibition of PKA activation using the cell-permeant cyclic nucleotide analogue Rp-Br-cAMP, led to an elimination of KA-mediated enhancement of synaptic transmission/glutamate release at thalamocortical synapses. The correspondence of mechanism was also emphasized in the current studies in that inhibition of the catalytic activity of PKA by H-89 suppressed the KA-mediated facilitation in both synaptosome and slice models of glutamate release. Likewise, direct activation of AC by forskolin (+IBMX) produced occlusion of the effects of KA in both preparations. Together, these data consistently point to an AC/cAMP/PKA signalling cascade underpinning the facilitatory regulation of glutamate release, which nonetheless remained recalcitrant to inhibition of G-proteins.
The facilitatory effects of KA on glutamate release from synaptosomes and the eEPSC in slices qualitatively display conspicuous similarity with respect to the involvement of AC/cAMP/PKA. However, the key detractor from an unequivocal suggestion of commonality of mechanism underlying the facilitation in the two models is clearly the different dose-dependencies observed. Although the undeniable difference in complexity of the two preparations likely contributes to the discrepancy, several specific reasons can be posited for the exquisite sensitivity to KA in slices compared to synaptosomes. Firstly, axonal localization of KARs has been indicated by GluK2/3 immunolabelling studies, albeit described in the hippocampus (Petralia et al. 1994). While axonal or axo-dendritic KARs would be activated by KA in a slice preparation, in isolated nerve terminals, depleted of any axonal compartment for purpose, the contribution of these receptors would be severely attenuated, if not completely eliminated. Also likely contributing to the relative sensitivity of the slices to KA compared with synaptosomes, is the enhancement of response in the former through heterosynaptic interaction of synapses through pre-synaptic, juxtasynaptic KARs (Schmitz et al. 2000). This latter phenomena, thought to contribute to frequency facilitation of glutamate responses mediated through pre-synaptic KARs in the hippocampus (Schmitz et al. 2001), would clearly not occur in the dissociated nerve terminal situation, where indeed the nature of release assay obviates any effects due to endogenously released glutamate. Secondly, the frugal possibility remains that higher concentrations of KA required to obtain facilitation in synaptosomal model reflects an uncoupling/inactivation of functional receptors because of preparative procedures. The requirement for the relatively higher concentrations of agonist in the synaptosomal preparation compared with slices is indeed a feature of studies examining KA-mediated pre-synaptic modulation using the biochemical methodology. Even in slice preparations, KA sensitivity can be seen to vary broadly depending on the synapse (autoreceptor activation in CA1 synapses in the hippocampus for example, requires 10-times higher kainate than elsewhere [Kamiya and Ozawa 1998; ]) and subunit constitution (GluK3-subunit containing receptors show 10-fold lower affinity than other non-NMDA receptors [Schiffer et al. 1997]). These foregoing observations reason against the rejection of a commonality of mechanism between the two models used here, purely on the basis of differing concentration dependencies of modulation. This is not to imply, however, that the synaptosome data obtained from a heterogenous population of terminals specifically reflect, in quantitative terms, the behaviour of thalamocortical terminals being examined in slices. In the absence of a specific marker, it is indeed not possible to quantitate how much of the synaptosome population represents thalamocortical terminals. Aside the aforementioned differences, the synaptosome model has been useful in several ways. First, it has allowed us to directly ascertain whether pre-synaptic KARs, as terminal-resident receptors, exist and modulate the release of glutamate, and determine the specific intracellular signalling pathway(s) involved in the regulation. Secondly, the consistent absence of KA-mediated depression espouses the hypothesis that pre-synaptic KARs involved in depression of glutamate release are not terminal-resident and, indeed, infer that the distinct populations of KAR mediating respective facilitation and depression of release are separate entities in different compartments.
Although the synaptosome model provided an essential basis for our studies at the outset, to obviate issues with heterogeneous readout from a mixed population of terminals obtained with this preparation, we conducted the further dissection of the details of the observed KAR-mediated facilitation in the thalamocortical slice preparation, assaying a specific/defined glutamatergic synapse. Given the lack of evidence for an upstream G-protein mediated initiation/transduction of the proposed AC/cAMP/PKA cascade involved in pre-synaptic KAR-mediated enhancement of glutamate release, we considered the potential role of Ca2+ as the instigator of the signalling at defined thalamocortical synaptic model. Potentially, KARs can effect external Ca2+ entry through, either an ionotropic activity to depolarize nerve terminals (Perkinton and Sihra 1999) and thereby activate voltage-gated Ca2+ channels or, directly, through Ca2+ permeable KARs per se (Fletcher and Lodge 1996; Scott et al. 2008). Intriguingly, in thalamocortical synapses, a blockade of the latter by philanthotoxin eliminated the KA-mediated synaptic facilitation, suggesting a strict dependence of the modulation on external Ca2+ entry via KARs. Thus, although unedited Ca2+ permeable KARs usually prevail early in neuronal development, the receptors evidently persist at thalamocortical synapses in adult brains.
Continuing the analysis of requirements for KAR-mediated regulation, we evaluated the possibility that core, but likely limited entry of Ca2+ via KARs, may be amplified by intraterminal Ca2+ store mobilization as has been described at hippocampal synapses (Lauri et al. 2003; Scott et al. 2008). Indeed, a crucial role for intraterminal Ca2+ stores was emphasized by our observations that, in the presence of thapsigargin, which effectively depletes intracellular Ca2+ stores (Irving et al. 1992), the facilitation of glutamate release mediated by KARs was abolished. Furthermore, by treating the slices with ryanodine, and thereby selectively inhibiting Ca2+-induced Ca2+-release (Berridge 1998), we demonstrated that Ca2+ entering via KARs induces Ca2+ mobilization from intraterminal Ca2+ stores.
Using two different pharmacological approaches in thalamocortical slices, we have demonstrated that Ca2+ entering via KARs and release of Ca2+ from intraterminal stores is indeed obligatory for the facilitation of glutamate release produced by KA. The question remained, how might the increase in cytosolic [Ca2+] couple to the postulated AC/cAMP/PKA signalling underlying the facilitation by KAR. A tenable possibility is that the increased cytosolic [Ca2+] activates Ca2+-dependent ACs present in thalamocortical terminals. Numerous ACs have been described, but two members of the family, viz. AC1 and AC8, are known to be activated by Ca2+-calmodulin, are prevalent in the central nervous system (see Wang and Storm 2003; Cooper 2003 for reviews), and have been shown to be essential for Ca2+-stimulated elevation of cAMP in studies with double knockouts of AC1 and AC8 (Wong et al. 1999). Our data, showing that the calmodulin antagonists W-7 and CMZ abolished the pre-synaptic KAR-mediated modulation in thalamocortial slices, support the hypothesis that formation of a Ca2+-calmodulin complex following KAR activation might stimulate AC1 and/or AC8, and thereby instigate the AC/cAMP/PKA cascade in the promotion of facilitation of glutamate release at thalamocortical synapses.
The experiments with thalamocortical slices described herein show the ventrobasal thalamus-L4 stellate synapse to be a robust model for the study of KAR-mediated modulation. Interestingly, however, we observed a facilitation of synaptic transmission/glutamate release at 1 μM KA, contrary to previous study with the same synapse, where 3 μM KA was required (Jouhanneau et al. 2011). The likely reason for this discrepancy is the different age of animals used for the experiments (we used adult animals whereas Jouhanneau et al. 2011 utilized 1 week post-natal animals). The higher concentrations of KA necessary for the activation of KARs at juvenile synapses may well reflect the low efficacy of KA at KARs lacking GluK4 and GluK5 subunits as suggested (Jouhanneau et al. 2011). Conversely, the lower effective concentrations of KA at adult thalamocortical synapses may indicate compositional plasticity of KARs, with the addition of at least some GluK4 or GluK5 subunits increasing KAR affinity for KA at mature synapses.
Although it is clear from our results that pre-synaptic KAR function certainly persists in adult animals at thalamocortical synapses, and is not restricted to the first few post-natal days as reported previously (Kidd et al. 2002), the question remains as to why, functionally, juvenile KARs may subtend relatively low affinity for KA (and glutamate) (Jouhanneau et al. 2011) compared with KARs at the same synapse in adult animals. One explanation might be that, in developing animals, KARs have an autoreceptor role, with the levels of activation determining pre-synaptic facilitation (at low [KA]) and depression (at high [KA]), and thereby presumably affecting synapse consolidation. While the exact role(s) of these KARs in adult animals remains to be elucidated, the modulation of glutamate release reported herein might well be involved in some forms of plasticity (Negrete-Díaz et al. 2007).
In our experiments, we observe that KAR activation has a biphasic effect at ventrobasal thalamus-L4 synapses as previously reported (Jouhanneau et al. 2011), inducing a depression of glutamate release at relatively high concentrations (> 1–3 μM). While the scope of the present study was to determine the mechanisms involved in the observed facilitation, and not to determine the intracellular mechanisms involved in the depression, our results indicated that the transient depression observed with higher [KA] is abolished in the presence of Rp-Br-cAMP, and is not affected by philantothoxin, Tsg, CMZ or W-7. These results suggest that the observed depression likely involves the AC/cAMP/PKA pathway as described for facilitation, indicating that these KARs may have alternative modes of facilitatory and depressive action, respectively coupling to an increase in cAMP levels (and subsequent activation of PKA), or to a decrease in cAMP levels (and subsequent decrease in activation of PKA), as has been described in the mossy fiber-CA3 synapses of the hippocampus (Negrete-Díaz et al. 2006). Our present results suggest that, while the activation of AC, to mediate a facilitation of glutamate release, involves the Ca2+-calmodulin complex, these KARs can also be negatively coupled to the AC/cAMP/PKA pathway by decreasing the activation of AC. As we reported previously for the depression of glutamate release mediated by activation of KARs at mossy-fiber-CA3 synapses, the negative coupling of the receptor to the AC/cAMP/PKA pathway mediating the depression, might involve the action of a pertussis toxin sensitive G-protein (Negrete-Díaz et al. 2006), as we indeed confirm here. Finally, the possibility remains that these diametric mechanisms are mediated by two different types of KARs. Future study will determine the exact instruments involved in the observed modulation and whether one or two different populations of KARs exist pre-synaptically at thalamocortical synapses.
In conclusion, our data show that the activation of pre-synaptic KARs by KA in thalamocortical synapses results in the facilitation of glutamate release, pharmacologically and mechanistically congruent to that observed in isolated cortical nerve terminals. We propose that the mechanism of KA-mediated pre-synaptic facilitation involves an entry of Ca2+ through Ca2+ permeable KARs, which triggers the release of Ca2+ from internal stores in thalamocortical terminals. The raised Ca2+ binds to calmodulin to form a Ca2+-calmodulin complex, which then putatively activates AC1 or AC8 to produce an increase in cAMP levels and a resultant stimulation of PKA; now established to result in an enhancement of glutamate release and hence synaptic transmission.
This study was supported by grant BFU2006-1455/BFI (from the Spanish Ministry of Education and Culture) and a grant from the ‘Eugenio Rodríguez-Pascual Foundation’ to A.R-M. P.D-F. is a recipient of a FPI fellowship from the Spanish Ministry of Science and Innovation (MICINN). The authors declare no conflicts of interest. The authors thank Dr Jasmina Jovanovic (School of Pharmacy, University College London) for her critical input. The constructive input from referees during the review of this manuscript is gratefully acknowledged.