Glucagon-like peptide-1 facilitates cerebellar parallel fiber glutamate release through PKA signaling in mice in vitro

Glucagon-like peptide-1 (GLP-1) is mainly secreted by preproglucagon neurons; it plays important roles in modulating neuronal activity and synaptic transmission through its receptors. In the present study, we investigated the effects of GLP-1 on parallel fiber–Purkinje cell (PF-PC) synaptic transmission in mouse cerebellar slices using whole-cell patch-clamp recording and pharmacology methods. In the presence of a γ-aminobutyric acid type A receptor antagonist, bath application of GLP-1 (100 nM) enhanced PF-PC synaptic transmission, with an increased amplitude of evoked excitatory postsynaptic synaptic currents (EPSCs) and a decreased paired-pulse ratio. The GLP-1-induced enhancement of evoked EPSCs was abolished by a selective GLP-1 receptor antagonist, exendin 9–39, as well as by the extracellular application of a specific protein kinase A (PKA) inhibitor, KT5720. In contrast, inhibiting postsynaptic PKA with a protein kinase inhibitor peptide-containing internal solution failed to block the GLP-1-induced enhancement of evoked EPSCs. In the presence of a mixture of gabazine (20 μM) and tetrodotoxin (1 μM), application GLP-1 significantly increased frequency, but not amplitude of miniature EPSCs via PKA signaling pathway. The GLP-1-induced increase in miniature EPSC frequency was blocked by both exendin 9–39 and KT5720. Together, our results indicate that GLP-1 receptor activation enhances glutamate release at PF-PC synapses via the PKA signaling pathway, resulting in enhanced PF-PC synaptic transmission in mice in vitro. These findings suggest that, in living animals, GLP-1 has a critical role in the modulation of cerebellar function by regulating excitatory synaptic transmission at PF-PC synapses.

www.nature.com/scientificreports/ paraventricular nucleus (PVN) 8 . GLP-1 receptors are expressed in the bed nucleus of the stria terminalis; activation of these receptors induces bidirectional membrane potential changes in neurons of this brain region 10 . Moreover, GLP-1 receptor activation reduces the conductance of voltage-dependent potassium channels, resulting in the increased spontaneous discharge frequency of olfactory mitral cells 14 . It has also been reported that the optogenetic activation of PPG neurons produces a mix of excitation and inhibition that creates a multiphasic response, shaping the firing pattern of the olfactory bulb mitral cells; this finding suggests that PPG neurons may drive the neuromodulation of olfactory output and change the synaptic map that regulates olfactory coding 15 .
The GLP-1-induced excitation of skullcap cells may also lead to changes in the excitability of the olfactory cortex and hypothalamus region by inhibiting the conductance of voltage-dependent potassium channels and enhancing the release of glutamate 16 . Furthermore, a GLP-1 receptor agonist can induce depolarization and an increased firing rate in arcuate nucleus neurons; this effect is prevented by the intracellular administration of a G-protein blocker 17 . GLP-1 can regulate neuronal activity by modulating multiple signal transduction pathways, such as the mitogen-activated protein kinases (MAPKs) ERK1/2, protein kinase A (PKA), protein kinase C (PKC), and cytosolic calcium signal pathways 11,18 . The activation of GLP-1 receptors activates adenylate cyclase, resulting in increased intracellular cyclic adenosine monophosphate (cAMP); this further modulates neuronal signal transduction and gene transcription 19 . The application of GLP-1 receptor allosteric modulators can augment cAMP signaling and increase intracellular calcium concentration 20 , as well as induce AMP-activated protein kinase (AMPK) inhibition and MAPK activation of neurons through PKA signaling pathways 21 . In addition, GLP-1 receptor activation can enhance the frequency of miniature excitatory postsynaptic currents (mEPSCs) and decrease the paired-pulse ratio (PPR) via presynaptic α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/kainate receptors in medium spiny neurons, thus suggesting that GLP-1 receptors modulate glutamate synthesis and release 22 . The application of GLP-1 agonists transiently enhances γ-aminobutyric acid (GABA) type A receptor-mediated synaptic and tonic currents in CA3 pyramidal neurons of rat brain slices 23,24 . In contrast, GLP-1 receptor blockade increases spontaneous excitatory synaptic activities, differentially modulates voltagegated and chemically gated calcium influx, and reduces pyramidal neuron overexcitation in hippocampal slices of 3 × transgenic Alzheimer's disease model mice 25 . In the arcuate nucleus, a GLP-1 analog increases mEPSC frequency and evoked EPSC amplitude in half of neurons, but also increases miniature inhibitory postsynaptic current (mIPSC) frequency and evoked IPSC amplitude in one-third of neurons, indicating that GLP-1 signaling facilitates presynaptic input in these neurons by acting on presynaptic GLP-1 receptors 17 .
GLP-1 receptors are abundantly expressed in the cerebellar cortex, including in the molecular layer, Purkinje cells (PCs), and the granular layer 26,27 . The cerebellar cortex receives many different types of afferent signalsincluding excitatory glutamatergic, inhibitory GABAergic, adrenocorticotropic, monoaminergic, and proglucagonergic projections-that are involved in the regulation of cerebellar neuronal activity 1,28 . Notably, afferents from nucleus tractus solitarius PPG neurons also project to all layers of the cerebellar cortex; however, the effect of GLP-1 on cerebellar cortical synaptic transmission has not yet been explored. We therefore studied the effects of GLP-1 on parallel fiber (PF)-PC synaptic transmission in mouse cerebellar slices.

Material and methods
Slice preparation. The experimental procedures were approved by the Animal Care and Use Committee of Yanbian University and were in accordance with the animal welfare guidelines of the U.S. National Institutes of Health. The permit number is SYXK (Ji) 2011-006. Cerebellar slices preparation has been described previously 29 . In brief, 4-6 weeks old ICR (Institute of Cancer Research) mice were deeply anaesthetized with halothane. After cutting the head of the mouse with scissors,, the whole brain was removed and placed in ice-cold artificial cerebrospinal fluid (ACSF) containing (in mM): 125NaCl, 3KCl, 1 MgSO 4 , 2 CaCl 2 , 1.25 Na 2 HPO 4 , 25 NaHCO 3 , and 10 D-glucose bubbled with 95% O 2 and 5% CO 2 (pH 7.4; 295-305 mOsm). The sagittal cerebellar slices (300 µm thick) were prepared using a Vibratome (VT 1200s, Leica, Nussloch, Germany). The slices were incubated over 1 h in a ACSF filled chamber at room temperature (24-25 °C) prior to patch-clamp recording.

Electrophysiological recordings 29 .
Whole-cell patch-clamp recordings from cerebellar PCs in slices were visualized using a 60 × water-immersion lens through a Nikon microscopy (Eclipse FN1, Nikon Corp., Tokyo, Japan). The composition of the recording electrode content is as follows (in mM): potassium gluconate 120, HEPES 10, EGTA 1, KCl 5, MgCl 2 3.5, NaCl 4, biocytin 8, Na 2 ATP 4, and Na 2 GTP 0.2 (pH 7.3 with KOH, osmolarity adjusted to 300 mOsm).Patch pipette resistances were 4-6 MΩ in the bath, with series resistances in the range of 10-20 MΩ. Membrane currents were monitored with an Axopatch 700B amplifier (Molecular Devices, Foster City, CA, USA), filtered at 5 kHz, and acquired through a Digidata 1440 series analog-to-digital interface on a personal computer using Clampex 10.4 software (Molecular devices, Foster City, CA, USA). Purkinje cells were held in voltage-clamp mode at -70 mV. Series resistance was monitored by injecting voltage pulses (10 ms, 5 mV), and only cells with stable series resistance were used for further analysis. Under the voltageclamp recording mode, gabazine (20 µM) was added to the ACSF for blocking the GABAergic inhibition 29 . PF electrical stimulation was performed by a stimulating electrode containing ACSF (0.1-0.5 MΩ). The stimulating electrode was mounted on remote-controlled manipulators(MP-385, Sutter, Novato, CA, USA), and placed in the molecular layer of the cerebellar slice. Paired-current pulses (0.2 ms, 10-100 µA; duration: 50 ms; 0.05 Hz) were used for parallel fiber stimulation. For inhibiting postsynaptic PKA in some experiments, protein kinase inhibitor-(6-22) amide (PKI; 5 µM) was added in pipette internal solution 30 . Recordings of mEPSCs were performed in the presence of a mixture of gabazine (20 μM) and tetrodotoxin (TTX; 1 μM) 31  www.nature.com/scientificreports/ Drug application. GLP-1, Exendin 9-39, TTX and gabazine were bought from Sigma-Aldrich (Shanghai, China). KT5720 and protein kinase inhibitor-(6-22) amide (PKI) were purchased from Tocris (Bristol, UK). KT5720 (1 mM) were diluted in dimethyl sulfoxide (DMSO). All the drugs were dissolved in solution and kept in frozen in aliquots, and they were applied to the cerebellar slices at 0.5 ml/min in ACSF through the peristaltic pump. When a stable whole-cell recording was formed, 5 min baseline was recorded before drugs application. In the experiments involving PKA inhibitor, KT5720 was applied at least 30 min before patch-clamp recording and continuing throughout the experiments.
Statistical analysis. Electrophysiological data were analyzed using Clampfit 10.3 software. To evaluate the pre-or postsynaptic location of the GLP-1 modulated PF-PC synaptic transmission, the paired-pulse ratio (PPR) was calculated as the second EPSC amplitude over the first EPSC amplitude [32][33][34][35] .The frequency and amplitude of mEPSCs were analyzed using Mini Analysis software (Version 6.0.3; Synaptosoft, Decatur, GA). The original traces of mEPSCs were filtered digitally at 1 kHz. Only synaptic events showing a clearly defined baseline and a peak were used for amplitude analysis. During analysis, the threshold for detection of mEPSCs was set at 4pA and the period to search an mEPSC was set at 30 ms 36 .All the parameters were maintained constant for an individual recorded cell during treatments with ACSF, drugs and recovery. All data are expressed as the mean ± S.E.M. One-way and repeated measures ANOVA followed by Tukey's post-hoc test (SPSS software) were used to determine the level of statistical significance between groups of data. P values below 0.05 were considered to indicate a statistically significant difference between experimental groups.   (Fig. 1A,B). The bath application of GLP-1 (100 nM; 5 min) increased the amplitudes of N1 and N2 (Fig. 1B,C). In the presence of GLP-1 (100 s), the normalized amplitude of N1 was significantly higher than at baseline (118.6% ± 5.9% of the baseline; P = 0.026, n = 7, Fig. 1D). Because the PPR is a widely reported indicator of a presynaptic or postsynaptic locus of the mechanism 32-35 , we further analyzed the effects of GLP-1 on the PPR. Notably, GLP-1 application significantly decreased the PPR; the normalized PPR was 88.73% ± 3.41% of the baseline (P = 0.031, n = 7, Fig. 1E). These results indicate that GLP-1 increases PF-PC synaptic transmission through GLP-1 receptor activation.

PKA activation is required for the GLP-1induced enhancement of PF-PC EPSCs.
To evaluate the role of PKA in the GLP-1-induced enhancement of PF-PC EPSCs, we used a selective PKA inhibitor, KT5720 (100 nM). For the complete inhibition of PKA, KT5720 was added to the recording chamber with artificial cerebrospinal fluid (ACSF) for 30 min before whole-cell patch-clamp recording started. Consistent with previous studies 30,37 , PKA inhibition produced a decrease in PF-PC EPSCs as well as an increase in the PPR. In the presence of KT5720, the normalized amplitude of N1 was significantly lower than that at baseline (74.2% ± 3.9% of the baseline [ACSF only]; P = 0.003, n = 6) and the normalized PPR was significantly higher (111.2% ± 2.2% of the baseline; P = 0.026, n = 6; data not shown). In the absence of PKA activity, GLP-1 application failed to increase the PF stimulation-evoked EPSCs (Fig. 3A,B); in the presence of KT5720 and GLP-1, the normalized amplitude of N1 was 100.6% ± 5.8% of the control (KT5720 alone; P = 0.76, n = 7, Fig. 3C) and the normalized www.nature.com/scientificreports/ PPR was 100.15% ± 4.02% of the control (P = 0.68, n = 7, Fig. 3D). Together, these results indicate that PKA inhibition abolishes the effects of GLP-1 on PF-PC EPSCs, suggesting that the GLP-1 enhancement of PF-PF synaptic transmission is dependent on the PKA signaling pathway.
To evaluate whether PKA antagonists act at the presynaptic terminal or the postsynaptic cell, a membraneimpermeable PKA inhibitor, PKI (5 μM), was added to the internal pipette solution 30 . With PKI in the internal solution, the bath application of GLP-1 (100 nM) increased the amplitudes of N1 and N2 (Fig. 4A,B). In the presence of GLP-1, the normalized amplitude of N1 was significantly higher than that of the baseline (117.3% ± 5.8% of the baseline; P = 0.031, n = 7, Fig. 4C) and the normalized PPR was significantly lower than that of the baseline (90.1% ± 4.01% of the baseline; P = 0.037, n = 7, Fig. 4D). These results indicate that the GLP-1-induced increase in PF-PC synaptic transmission is dependent on the presynaptic PKA signaling pathway, suggesting that GLP-1 upregulates glutamate release from PF terminals.

GLP-1 enhances mEPSCs through GLP-1 receptors and PKA signaling. The mEPSC is a widely
used indicator of presynaptic sites and can be used to observe the presynaptic mechanisms of PF-PC synaptic transmission [32][33][34][35] . To determine whether GLP-1 affects glutamatergic synaptic transmission at PF terminals, we further examined the effects of GLP-1 on mEPSCs. To record mEPSCs from PCs, we added gabazine (10 μM) and TTX (1 μM) to ACSF to block spontaneous EPSCs and GABAergic inhibitory inputs. Under the voltageclamp recording mode (Vh = − 70 mV), GLP-1 administration increased mEPSC frequency (Fig. 5A), resulting in a leftward shift of the frequency-accumulation probability curve of mEPSCs (Fig. 5B); however, there was no shift in the amplitude-accumulation probability curve (Fig. 5C). In the presence of GLP-1, the normalized frequency of mEPSCs was 112.2% ± 4.6% of the baseline (P = 0.031, n = 7, Fig. 5D) and the normalized amplitude of mEPSCs was 100.9% ± 4.0% of the baseline (P = 0.73, n = 7, Fig. 5E). These results indicate that GLP-1 significantly increases mEPSC frequency but not amplitude, suggesting that GLP-1 increases glutamate release from PF terminals.
We then used a GLP-1 antagonist to observe the effects of GLP-1 on mEPSCs in the absence of GLP-1 receptor activity. As shown in Fig. 6, mEPSC frequency and amplitude were not significantly affected by exendin 9-39 (100 nM) administration, but the effects of GLP-1 on mEPSCs were completely blocked (Fig. 6A). In the www.nature.com/scientificreports/ presence of exendin 9-39, GLP-1 administration did not lead to a significant increase in mEPSC frequency and there was no significant change in the frequency-cumulative probability curve of mEPSCs (Fig. 6B). Combined exendin 9-39 and GLP-1 administration also had no significant effects on mEPSC amplitude or the amplitudecumulative probability curve (Fig. 6C). In the presence of exendin 9-39, mEPSC frequency was 101.2% ± 4.8% of the baseline (P = 0.63, n = 7, Fig. 6D). In the presence of both exendin 9-39 and GLP-1, mEPSC frequency was 100.8% ± 4.2% of the baseline, similar to the results with exendin 9-39 alone (P = 0.72, n = 7, Fig. 6D). In the presence of exendin 9-39, mEPSC amplitude was 100.6% ± 3.5% of the baseline (P = 0.76, n = 7, Fig. 6E). In the presence of both exendin 9-39 and GLP-1, mEPSC amplitude was 100.3% ± 3.5% of the baseline, similar to the findings with exendin 9-39 alone (P = 0.78 n = 7, Fig. 6E). Together, these results indicate that GLP-1 significantly increases glutamate release from PF terminals through its receptors, resulting in significantly increased mEPSC frequency. Next, we used a specific PKA inhibitor, KT5720 (100 nM), to observe whether GLP-1 increased mEPSC frequency via the PKA signaling pathway. As shown in Fig. 7, after 30 min of KT5720 (100 nM) administration, mEPSC frequency was significantly reduced, and the frequency-cumulative probability curve of mEPSCs shifted to the right (Fig. 7A, B). In the presence of KT5720, GLP-1 did not increase mEPSC frequency and there was no significant change in the frequency-cumulative probability curve of mEPSCs (Fig. 7B). Furthermore, KT5720 application significantly decreased mEPSC amplitude and the amplitude-cumulative probability curve shifted to the left (Fig. 7C). In the presence of KT5720, GLP-1 application did not induce any further decreases in mEPSC amplitude, and the amplitude-cumulative probability curve did not change (Fig. 7C). After KT5720 was present for 30 min, the normalized frequency of mEPSCs was 74.6% ± 2.9% of the baseline (P = 0.006, n = 7, Fig. 7D). In the presence of KT5720 and GLP-1, the normalized frequency of mEPSCs was 74.7% ± 2.8% of the baseline, similar to the results with KT5720 alone (P = 0.76, n = 7, Fig. 7D). In the presence of KT5720, the normalized amplitude of mEPSCs was 65.6% ± 4.8% of the baseline (P = 0.002, n = 7, Fig. 7E). In the presence of KT5720 and GLP-1, the normalized amplitude of mEPSCs was 65.4% ± 4.6% of the baseline, similar to the findings with KT5720 alone (P = 0.74, n = 7, Fig. 7E). These results indicate that GLP-1 significantly increases glutamate release from PF terminals through the PKA signaling pathway, resulting in significantly increased mEPSC frequency in mice in vitro. Previous studies have demonstrated that GLP-1 receptor activation modulates neuronal activity and synaptic transmission in several brain areas 8,[38][39][40] . The activation of GLP-1 receptors increases glutamate release leading to an increase in the spike firing rate of hippocampal CA1 neurons in rats in vitro 38 . GLP-1 receptor expression has been observed in whole-cell recordings from the bed nucleus of the stria terminalis, hippocampus, and PVN, where GLP-1 elicits a reversible inward current or depolarization 8 . In arcuate nucleus neurons, postsynaptic GLP-1 receptor activation leads to membrane potential depolarization, which is accompanied by an increase in spontaneous action potential frequency 39,40 . However, GLP-1 receptor activation produces inhibition or excitation in distinct subpopulations of neurons in the bed nucleus of the stria terminalis, suggesting that GLP-1 affects both excitatory and inhibitory synaptic transmission in mice in vitro 10 . Our present results indicate that GLP-1 increases PF-PC EPSC amplitude, suggesting that GLP-1 enhances PF-PC transmission via the activation of its receptor. The PPR has been widely used in previous studies to investigate the presynaptic or postsynaptic location of mechanisms [32][33][34] . For example, in the lateral amygdala, kainite receptor activation inhibits the evoked EPSC amplitude but significantly increases the PPR, indicating that presynaptic kainite receptor activation inhibits glutamate release from neurons 32 . However, kainite receptor activation has also been reported to increase the evoked www.nature.com/scientificreports/ EPSC amplitude but decrease the PPR at PF-PC synapses in the cerebellar cortex, indicating that presynaptic kainite receptors upregulate glutamate release at PF-PC synapses 33,34 . In the current study, GLP-1 significantly decreased the PPR at PF-PC synapses, indicating that presynaptic GLP-1 modulates PF-PC synaptic transmission. The activation of GLP-1 receptors significantly increases the frequency of spike firing and mEPSCs in PVN gonadotropin-releasing hormone neurons in mice in vitro 41 . Consistent with previous studies 12,41,42 , the present results indicate that GLP-1 significantly increases mEPSC frequency, suggesting that GLP-1 increases glutamate release at PF-PC synapses. Notably, the effects of GLP-1 on mEPSCs were abolished by the antagonism of GLP-1 receptors in our study, indicating that GLP-1 receptor activation significantly increases glutamate release from PF terminals, resulting in a significant increase in mEPSC frequency.
GLP-1 regulates PF-PC synaptic transmission via the presynaptic PKA signaling pathway. GLP-1 regulates neuronal activity by modulating multiple signal transduction pathways, such as the ERK1/2, PKA, PKC, and cytosolic calcium signaling pathways 11,18 . The PKA signaling pathway plays a key role in regulating neurotransmitter synthesis and release, and GLP-1 receptor activation can affect the release function of glutamate and regulate the release rate of neurotransmitter vesicles through the cAMP-PKA signaling pathway 8,38 . Activated GLP-1 receptors couple to adenylate cyclase, resulting in increased intracellular cAMP and calcium concentrations; this further modulates neuronal signal transduction and gene transcription and induces AMPK inhibition and MAPK activation of neurons 21 . In the central nervous system, glutamate and GABA release are regulated by various presynaptic modulators. At cerebellar PF-PC synapses, presynaptic kainate receptor activation modulates neurotransmitter release through an adenylate cyclase-cAMP-PKA signaling pathway [33][34][35] . In the current study, PKA inhibition abolished the effects of GLP-1 on PF-PC EPSCs, suggesting that GLP-1 enhances PF-PF through activation of the presynaptic PKA signaling pathway. Our results are consistent with those of previous studies 12,42 ; for example, GLP-1 receptor activation in PVN corticotropinreleasing hormone neurons enhances AMPA receptor function and excitatory glutamate synaptic transmission via the PKA signaling pathway 42 . Moreover, GLP-1 activation induces resting membrane potential depolarization and increased spike firing frequency in ganglion cells via the cAMP pathway 12 . Importantly, we added a membrane-impermeable PKA inhibitor to the internal pipette solution in the current study, to examine whether PKA antagonists act at the presynaptic terminal or the postsynaptic cell. Blockade of postsynaptic PKA failed to prevent the GLP-1-induced enhancement of PF-PC EPSCs, thus confirming that GLP-1 increases PF-PC synaptic transmission via the presynaptic PKA signaling pathway. Moreover, GLP-1 failed to increase mEPSC www.nature.com/scientificreports/ frequency, indicating that GLP-1 increases glutamate release from PF terminals through the presynaptic PKA signaling pathway in mice in vitro. GLP-1 receptors are G-protein coupled receptors that play important roles in regulating neuronal activity and synaptic transmission via cellular signaling pathways 8,10,43 . Consistent with previous studies 43,44 , our results suggest that GLP-1 binds to its receptor and enhances adenylate cyclase activation at PF terminals; this leads to increased intracellular cAMP levels and PKA activity, thus resulting in increased PF-PC synaptic transmission. The present study provides novel insights into the regulatory function of GLP-1 in the central nervous system. Furthermore, they suggest that GLP-1 influences cerebellar PC output in living animals, thus affecting motor coordination and motor learning function.