Signaling Mechanism for Modulation by GLP-1 and Exendin-4 of GABA Receptors on Rat Retinal Ganglion Cells

Glucagon-like peptide-1 (GLP-1) is expressed in retinal neurons, but its role in the retina is largely unknown. Here, we demonstrated that GLP-1 or the GLP-1 receptor (GLP-1R; a G protein-coupled receptor) agonist exendin-4 suppressed γ-aminobutyric acid receptor (GABAR)-mediated currents through GLP-1Rs in isolated rat retinal ganglion cells (GCs). Pre-incubation with the stimulatory G protein (Gs) inhibitor NF 449 abolished the exendin-4 effect. The exendin-4-induced suppression was mimicked by perfusion with 8-Br-cAMP (a cAMP analog), but was eliminated by the protein kinase A (PKA) inhibitor Rp-cAMP/KT-5720. The exendin-4 effect was accompanied by an increase in [Ca2+]i of GCs through the IP3-sensitive pathway and was blocked in Ca2+-free solution. Furthermore, when the activity of calmodulin (CaM) and CaM-dependent protein kinase II (CaMKII) was inhibited, the exendin-4 effect was eliminated. Consistent with this, exendin-4 suppressed GABAR-mediated light-evoked inhibitory postsynaptic currents in GCs in rat retinal slices. These results suggest that exendin-4-induced suppression may be mediated by a distinct Gs/cAMP-PKA/IP3/Ca2+/CaM/CaMKII signaling pathway, following the activation of GLP-1Rs. Supplementary Information The online version of this article (10.1007/s12264-022-00826-9) contains supplementary material, which is available to authorized users.


Introduction
Glucagon-like-peptide-1 (GLP-1) is a metabolic hormone secreted by intestinal endocrine L-cells and stimulates insulin secretion in a glucose-dependent manner [1]. GLP-1 is also produced in the brain, particularly from preproglucagon neurons, which are distributed in the solitary tract of the brain stem, and it functions as a neuropeptide [2][3][4][5]. This peptide has been implicated in modulating neuronal cell differentiation, neurite outgrowth, and performing neuroprotective functions through the activation of GLP-1 receptors (GLP-1Rs), class B sub-group G proteincoupled receptors [2,6].
GLP-1 expression has been found in the vertebrate retina [7][8][9]. Specifically, GLP-1R immunoreactivity has been observed in the ganglion cell layer (GCL) in the human and rat retina [7,[10][11][12], and sparse staining has also been detected in the inner and outer nuclear layers of the human retina. More recently, systemic or topical administration (eye drops) of GLP-1 and GLP-1R agonists has been reported to prevent electroretinogram abnormalities and neurodegeneration of the retina in rat and mouse models of diabetes [7,10]. However, whether GLP-1 modulates retinal information processing is still largely unknown.
In the retina, ganglion cells (GCs) are the sole output neurons that integrate inhibitory signals from amacrine cells and excitatory signals from bipolar cells and transmit the processed information to higher centers [13]. GABA receptors (GABARs) and glycine receptors mediate the inhibitory signals of amacrine cells onto GCs. It has been shown that GLP-1/the GLP-1R agonist exendin-4 increases GABA A R-mediated tonic currents through GLP-1R activation in rat hippocampal CA3 pyramidal neurons [14], but no data concerning the regulation of GABARs of retinal neurons by GLP-1/GLP-1R agonists are now available. In the present work, we used whole-cell patch-clamp recording techniques to explore whether GLP-1 or exendin-4 modulates GABARs of rat retinal GCs. We first demonstrated that GLP-1 or the GLP-1R agonist exendin-4 suppressed GABAR-mediated currents (GABA currents) in isolated rat retinal GCs through GLP-1R activation. By pharmacological approaches, we further showed that a distinct G s /cAMP-protein kinase A (PKA)/inositol 1,4,5trisphosphate (IP 3 )/Ca 2? /calmodulin (CaM)/CaM-dependent protein kinase II (CaMKII) signaling pathway is responsible for the exendin-4 effect on GCs. Consistent with this, we also showed that exendin-4 suppressed GABAR-mediated the light-evoked inhibitory postsynaptic currents (L-IPSCs) of GCs via GLP-1Rs.

Materials and Methods
Animals Male Sprague-Dawley rats (15-20 days of age) were used in this study. All procedures were approved by the Animal Care and Use Committee of the Shanghai Medical College of Fudan University and were performed according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All participants signed the written informed consent.

Preparation of Isolated GCs
As previously reported [16], GCs were isolated by enzymatic digestion of retinal tissue with papain and mechanical separation. Rhodamine-labeled GCs (15-25 lm in diameter) were used for electrophysiological recording within 2-3 h after separation.

Ca 21 Imaging
We used the membrane permeability indicator Fura-2AM (Dojindo, Kumamoto, Japan) to assess changes in the intracellular Ca 2? concentration ([Ca 2? ] i ) of isolated GCs as described previously [15]. Fluorescence images were captured on an Olympus inverted microscope equipped with a digital CCD camera (Hamamatsu Photonics, Shizuoka, Japan). We used high-speed continuous scanning monochromatic light sources (Till Photonics, Grafeling, Germany) to excite at 340 nm and 380 nm. Fluorescence intensities at 340 nm and 380 nm (F 340 and F 380 ) were measured every 1-10 s, and images were acquired using C-imaging systems (Hamamatsu Photonic). The [Ca 2? ] i of the cell was proportional to the ratio of fluorescence intensity between the two images. Before an experiment, Fig. 1 Characterization of GABA-induced currents in isolated rat retinal GCs. A Representative recordings showing that the current induced in a GC by 30 lmol/L GABA is almost completely suppressed by 10 lmol/L bicuculline (BIC). The cell is clamped at V h = -60 mV and 30 lmol/L GABA is repetitively applied for 5 s at intervals of 2 min. C Representative recordings from another GC, showing that the GABA-induced current is partially attenuated by BIC, and the remaining current is completely eliminated by coapplication of 100 lmol/L CGP-35348. E Current traces of a GC showing that the GABA-induced current is partially suppressed by BIC, and the remaining current is not changed by co-application of 10 lmol/L TPMPA. G Representative recordings showing that the GABA-induced current of a GC is almost completely suppressed by 10 lmol/L gabazine. I Representative recordings of another GC showing that the GABA-induced current is partially attenuated by gabazine, and the remaining current is completely eliminated by coapplication of 100 lmol/L CGP-35348. B, D, F, H, J Bar charts showing statistical analysis of the above data. ***P \0.001, n.s., P[0.05, paired Student's t-test. The data are presented as the mean ± SEM in all figures. The data for each cell are normalized to the current amplitude of the cell in normal Ringer's (control) and then averaged. K Average current-voltage relationship of GABA A receptor-mediated currents from six GCs. Current responses for each cell at different holding potentials are normalized to the response obtained at -60 mV. Cell numbers (n) are marked inside the bars, and the cell numbers in different bars in the same subgraph are the same. It is also the case for other figures. we measured the background fluorescence level and subtracted it from the obtained data.

Statistical Analysis
Data were analyzed using Pulsefit (HEKA Elektronik, Lambrecht/Pfalz, Germany), Igor 4.0 (WaveMetrics, Lake Oswego, USA), and SigmaPlot 12.0 (Systat Software, Inc., San Jose, USA). Data are shown as the mean ± SEM. Significant differences were identified by either paired Student's t-test (for paired data) or one-way ANOVA with post hoc Tukey's test (for multiple comparisons). For all analyses, P \0.05 was considered statistically significant.

GABAR-mediated Currents in Rat GCs
The currents induced by GABA were recorded from rhodamine-labeled rat retinal GCs with relatively large somata ([15 lm in diameter) in normal Ringer's. In some GCs, the currents induced by GABA (30 lmol/L) were almost completely suppressed (7.09% ± 4.12% of control, P \0.001, n = 8) by 10 lmol/L bicuculline (BIC), an antagonist of GABA A receptors (Fig. 1A, B), which suggests that the GABA currents of these cells are exclusively mediated by GABA A receptors. However, in the remaining GCs, the GABA-induced currents were reduced to 58.76% ± 5.02% of control (P \0.001, n = 7) (Fig. 1C, D) in the presence of 10 lmol/L BIC, and the remaining currents were almost eliminated by perfusing CGP-35348 (100 lmol/L), an antagonist of GABA B receptors (6.79% ± 2.14% of control, P \0.001, n = 7) (Fig. 1D). Moreover, in the cells in which BIC only partially suppressed the GABA currents (60.09% ± 7.86% of control, P \0.001, n = 6) (Fig. 1E, F), co-application of BIC and the GABA C receptor antagonist TPMPA (10 lmol/L) did not further suppress the GABA currents (58.94% ± 5.43% of control, P \0.001,n = 6). Similar results were obtained with the application of gabazine, another GABA A receptor antagonist. That is, in some GCs, addition of 10 lmol/L gabazine almost completely suppressed the GABA currents (6.34% ± 3.79% of control, P \0.001,n = 7) (Fig. 1G, H), while in the remaining GCs, 10 lmol/L gabazine suppressed the GABA currents to 56.89% ± 7.12% of control (P \0.001,n = 6) (Fig. 1I, J), and co-application of 100 lmol/L CGP-35348 almost eliminated the currents (6.12% ± 3.05% of control, P \0.001, n = 6) (Fig. 1I, J). The current-voltage relationship of the GABA A receptor-mediated currents was linear, with a reversal potential of -3.1 ± 2.3 mV (n = 6, Fig. 1K), which was very close to the E Cl -(-4.3 mV) calculated according to the Nernst equation. These results indicated that the GABA currents of these GCs are mediated by both GABA A and GABA B receptors. All these data suggest that GCs express functional GABA A and GABA B receptors, but not GABAc receptors, and this is consistent with the results of in situ hybridization and immunohistochemistry showing that GCs express GABA A and GABA B receptors [23][24][25][26][27].

Exendin-4 Suppresses GABA Currents in GCs
The endogenous GLP-1 is highly sensitive to degradation by dipeptidyl peptidase IV (DPP-IV) [28][29][30][31][32]; for this reason, we investigated the effects of the protease-resistant long-acting GLP-1R agonist exendin-4 [28,33,34] on the GABA currents of GCs ( Fig. 2A). No detectable current in GCs was elicited by perfusion of 50 nmol/L exendin-4 (data not shown). After applying exendin-4 for *2 min, the amplitude of peak current decreased and tended to stabilize in *6 min ( Fig. 2A). The current returned to the control level following a 4-min washout. Suppression by exendin-4 of GABA currents was recorded in most of the tested GCs (16 out of 18, 88.9%). On average, the current amplitudes of these cells after 6 min incubation with 50 nmol/L exendin-4 were decreased to 69.08% ± 5.71% of control (P \0.001, n = 16, Fig. 2B). In the other two cells, exendin-4 had no effect on the GABA currents (2/18, 11.1%). When the significance test was performed on the whole data set (18 cells), exendin-4 inhibited the GABA currents to a smaller extent (76.96% ± 4.41% of control), but the difference was still statistically significant (P\0.01 vs control).
We also determined whether GLP-1 had effects similar to exendin-4. Since the maximal concentration of GLP-1 in human postprandial plasma is\40 pmol/L [1], we chose 10 pmol/L GLP-1 for extracellular perfusion. As shown in Fig. 3C and D, application of GLP-1 for 6 min significantly suppressed the GABA currents of GCs to 28.11% ± 12.50% of control (P \0.01, n = 6), and the extent of suppression was larger than that of exendin-4 (69.08% ± 5.71% of control). The currents returned to the control level following a 2-min washout. Moreover, the GLP-1induced suppression was also abolished by exendin .
Exendin-4 has been widely used in the treatment of diabetes because its plasma half-life (120 min) is much longer than GLP-1 (1.5 min) [36,37]. Therefore, we selected exendin-4 for further mechanism study. Because GLP-1R is a G-protein-coupled receptor [38], to determine whether G-protein is associated with the suppression of GABA currents by exendin-4, we added GDP-b-S, a nonhydrolyzable G-protein inhibitor, into patch pipettes. After whole-cell recording from a GC, we waited 4 min to allow GDP-b-S to diffuse throughout the cell. When the GABA currents of GCs reached a stable level at *6 min after membrane rupture (control), application of exendin-4 for 6 min failed to suppress the currents (99.35% ± 3.64% of control, P[0.05, n = 7) (Fig. 3E). Since GLP-1R can be coupled to G s or G i/o [39][40][41], we further investigated which subtype(s) of G-proteins may be associated with the exendin-4 effect. Cell suspensions were preincubated with 10 lmol/L NF 449, a G s antagonist, for at least 30 min before recording and then exposure to exendin-4 for 6 min failed to change the GABA currents of these cells (99.87% ± 4.31% of control) (P [0.05, n = 6) (Fig. 3F). Furthermore, internal infusion of 30 lmol/L mastoparan, a peptide activator of G i and G o , for 6 min did not change the GABA currents of GCs (control), and in the presence of mastoparan, applying exendin-4 for 6 min still suppressed the GABA currents to 76.85% ± 4.54% of control (P \0.01, n = 7) ( Fig. 3G). These results suggest that GLP-1R is coupled to G s in rat GCs.

cAMP-PKA Signaling Pathway Mediates Exendin-4-induced Suppression of GABA Currents
Following activation of GLP-1Rs, the main intracellular signaling pathway stimulates G s , which in turn activates adenylate cyclase, resulting in increased intracellular cAMP levels and activation of PKA [42,43]. To investigate whether this pathway is involved, we studied the effects of extracellular perfusion of 8-Br-cAMP, a Mastoparan per se has no effect on the GABA currents of GCs, and co-application of exendin-4 persists in significantly suppressing the GABA currents. The waveforms shown below the data lines are the current responses recorded at the times indicated by I, II, and III shown in E, F, and G. **P \0.01, ***P \0.001, n.s., P [0.05, one-way ANOVA with post hoc Tukey's test. membrane-permeable cAMP analog, on the GABA currents of GCs. Fig. 4A shows that GABA currents of a GC were gradually suppressed by applying 8-Br-cAMP (400 lmol/L) and reached a stable level 6 min (control) after 8-Br-cAMP application. Co-application of exendin-4 did not change the suppression of the peak currents. Average data showed that the peak current amplitude was suppressed to 64.18% ± 7.42% of control (P \ 0.01,n = 6) (Fig. 4B) by 8-Br-cAMP and applying exendin-4 did not cause further suppression (65.32% ± 3.83% of control, P\ 0.01 vs control and P[0.05 vs 8-Br-cAMP) (Fig. 4B). The action of 8-Br-cAMP was reversible. Furthermore, after intracellular application of the PKA inhibitor Rp-cAMP (50 lmol/L), exendin-4 addition no longer changed the currents (98.57% ± 5.01% of control, P [ 0.05,n = 7) (Fig. 4C, D). In addition, when another PKA inhibitor, KT-5720, was intracellularly applied to GCs, additional exendin-4 also failed to affect the GABA currents (102.86% ± 4.34% of control, P[0.05, n = 6) (Fig. 4C, D).
To investigate whether the increase of [Ca 2? ] i in the GCs is related to the PKA activation induced by exendin-4, the changes of [Ca 2? ] i were measured by Ca 2? imaging when PKA activity was inhibited by H-89. There was no change in [Ca 2? ] i in GCs, either when perfused H-89 alone (0.95 ± 0.06 vs 0.96 ± 0.07 for control, n = 12) or along with exendin-4 (0.96 ± 0.08 vs 0.96 ± 0.07 for control, n = 12) (all P [0.05) (Fig. 5C, D). The data indicate that the exendin-4-induced [Ca 2? ] i elevation may be a result of PKA activation.

PI/PC-PLC-Independent Effects of Exendin-4
While GLP-1 has been shown to regulate the phosphatidylinositol-phospholipase C (PI-PLC) or phosphatidylcholine-phospholipase C (PC-PLC) activity in mouse pancreatic b cells [56], experiments with the PI-PLC inhibitor U73122 or with the PC-PLC inhibitor D609, did not support the involvement of these two pathways in the exendin-4 effects on GCs. As shown in Fig. 7A and B, internal infusion of 10 lmol/L U73122 or 60 lmol/L D609 for 6 min did not change GABA currents of GCs (control), then addition of exendin-4 for 6 min still suppressed the GABA currents (73.28% ± 5.09% of control, P \0.01 for U73122, n = 7; 67.41% ± 5.43% of control, P \0.001 for D609, n = 8).
Taken together, the putative signaling pathway mediating the effects of exendin-4 on GABA currents of rat GCs is summarized in the schematic diagram in Fig. 8.

Exendin-4 Suppresses GABAR-mediated L-IPSCs in GCs of Rat Retinal Slices
To further verify the effect of exendin-4 on GABARs on GCs, we examined the effects of exendin-4 on GABARmediated L-IPSCs in rat retinal slice preparations. During perfusion with TTX and strychnine (see Methods for details), the GABAR-mediated L-EPSCs of GCs were recorded. As shown in Fig. S1A and S1B, bath application of 100 nmol/L exendin-4 significantly suppressed the GABAR-mediated L-IPSC to 68.38% ± 4.54% of control (P \0.001, n = 8), and the response returned to the control level after washout (91.02% ± 7.94% of control, P[0.05).
GLP-1 is a hormone released by intestinal L-cells [1]. In the brain, GLP-1 is synthesized by preproglucagon neurons, which are distributed in the nucleus of the solitary tract of the brain stem [2][3][4][5]. Real-time quantitative RT-PCR analysis has revealed that GLP-1 mRNA is present in the human retina [7]. Moreover, immunostaining of GLP-1 has been reported in neurons in the GCL of human and mouse retinas [7,8]. These data suggest that GLP-1 may be also synthesized and released from GCs and/or displaced amacrine cells in the GCL. Previous studies have shown that GLP-1 can cross the blood-brain barrier [43,61], but whether it can cross the blood-retina barrier remains unknown.

Intracellular Mechanisms Underlying Exendin-4induced Suppression of GABA Currents
GLP-1R has been reported to primarily act through G s [2,43] and has also been shown to couple with the inhibitory G i/o proteins [40]. Using selective pharmacological inhibitors, we found that G s , but not G i/o , were involved in the exendin-4-induced suppression of GABA currents (Fig. 3 E-G). Actually, studies on rat dorsal root ganglion neurons have shown that activation of G s decreases GABA-induced currents [62].
Even though the PLC/PKC signaling pathway has been shown to be involved in GLP-1-induced insulin secretion in mouse pancreatic b cells [56], this signaling pathway was unlikely involved in the effect of exendin-4 on GABA currents in GCs, due to the persistence of this effect when PI-PLC or PC-PLC was blocked (Fig. 7). In contrast, our evidence suggested that the cAMP-PKA signaling pathway mediates the exendin-4-induced inhibition in GCs. The inhibitory effect of exendin-4 on GABA current was mimicked by perfusion with the cAMP analog 8-Br-cAMP. Moreover, the exendin-4-induced inhibition of GABA currents was eliminated when PKA was inhibited by KT-5720 or Rp-cAMP (Fig. 4). These results suggest that activating G s -linked GLP-1R stimulates adenylyl cyclase and increases the cAMP level, which in turn activates PKA. cAMP and PKA are important signaling molecules responsible for GLP-1R-mediated effects, which have been shown in neurons and non-neuronal cells [63][64][65][66][67][68][69][70].
Furthermore, the exendin-4 effect on GABA currents in GCs was Ca 2? -dependent. Ca 2? imaging showed that exendin-4 significantly increased [Ca 2? ] i in GCs (Fig. 5), consistent with the results reported in cultured hippocampal neurons showing that acute application of GLP-1 induces a transient elevation of [Ca 2? ] i [58]. GLP-1 has been shown to regulate voltage-gated Ca 2? channels. For example, GLP-1 enhances L-type Ca 2? currents through activation of the cAMP-dependent PKA pathway in canine cardiomyocytes [71], whereas when rat hippocampal neurons are pre-incubated with GLP-1 for 24 h, Ca 2? currents are significantly decreased [58]. However, when we chelated extracellular Ca 2? with EGTA, exendin-4 still suppressed the GABA currents in GCs, suggesting that Ca 2? entry through voltage-gated channels was not be involved in the action of exendin-4 under our experimental conditions. In contrast, when [Ca 2? ] i was greatly reduced by internal infusion of Ca 2? -free solution, exendin-4 no longer suppressed GABA currents (Fig. 5F). Moreover, the exendin-4 effect on GABA currents was blocked when IP 3 receptors were blocked by heparin. These data indicate that the Ca 2? release from IP 3 receptor-mediated intracellular pools is associated with the exendin-4 effect on GCs. Elevated [Ca 2? ] i in rat GCs may be a result of exendin-4induced PKA activation, because our experiments demonstrated that exendin-4 no longer caused changes in [Ca 2? ] i when PKA activity was blocked by H-89 (Fig. 5C, D). Actually, PKA activation has been shown to lead to Ca 2? release from intracellular storage in heart cells and hepatocytes through the ryanodine and/or IP 3 receptor pathway [44,45,72].
Many cellular Ca 2? -stimulated signaling cascades utilize the intermediate, CaM. The binding of Ca 2? alters the conformation of CaM and increases its affinity to many CaM-binding proteins, including CaMKII [73]. In GCs, the CaM/CaMKII pathway may mediate the exendin-4 effect, because the inhibitory effect by exendin-4 on GABA currents did not occur when CaM and CaMKII were blocked by W-7 and KN-93, respectively (Fig. 6). Consistent with our results, that CaM mediates the suppression of GABA A currents induced by elevated levels of [Ca 2? ] i has been reported in turtle retinal GCs [49] and rat rod bipolar cells [52], while that Ca 2? /CaMKII mediates orexin-Ainduced inhibition of GABA A currents has been reported in HEK293 cells [54].
GABA is one of the main inhibitory neurotransmitters and is predominantly released by wide-field GABAergic amacrine cells in the inner retina. It has been reported that GABAergic amacrine cells contribute to the organization of the GC receptive field [74][75][76]. The receptive field surround of GCs may be generated by feedforward inhibition directly onto GCs [75] or presynaptically, by GABAergic feedback inhibition from amacrine cells onto bipolar cell axons [77,78]. Application of GABA suppresses both the spontaneous and light-evoked activity of all GCs and the GABA A R antagonist bicuculline potentiates the spontaneous and light-evoked activity of ON-type GCs [79], which suggests that GABA directly inhibits GCs. Light stimulation-induced GABAR-mediated responses on most GCs in this work indicated that these cells receive direct inhibitory inputs from GABAergic amacrine cells. Suppression by exendin-4 of the GABA responses of GCs suggests that exendin-4 weakens inhibitory inputs mediated by GABA, which modulates the receptive field properties of GCs. 2015AA020512), Shanghai Municipal Science and Technology Major Project (2018SHZDZX01), ZJLab, Shanghai Center for Brain Science and Brain-Inspired Technology, and Sanming Project of Medicine in Shenzhen (SZSM202011015).

Conflict of interest
The authors claim that there are no conflicts of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/.