GABA_B Receptor Activation Protects Neurons from Apoptosis via IGF-1 Receptor Transactivation

Abstract

The G-protein-coupled receptors (GPCRs) and receptor tyrosine kinases (RTKs) play key roles in cell–cell communication. Several studies revealed important synergisms between these two types of receptors, with some of the actions of either receptor being mediated through transactivation of the other. Among the large GPCR family, GABA_B receptor is activated by the neurotransmitter GABA, and is expressed in most neurons where it mediates slow and prolonged inhibition of synaptic transmission. Here we show that this receptor is involved in the regulation of life and death decisions of cerebellar granule neurons (CGNs). We show that specific activation of GABA_B receptor can protect neurons from apoptosis through a mechanism that involves transactivation of the IGF-1 receptor (IGF-1R). Further work demonstrated that this cross talk was dependent on G_i/o-protein, PLC, cytosolic Ca²⁺, and FAK1 but independent of PKC, while IGF-1R-induced signaling involved Src kinase, PI3 kinase, and Akt activation. These results reveal a new function for this important GPCR and further highlight the importance of functional cross-talk networks between GPCRs and RTKs. Our results reveal GABA_B receptor as a potential drug target for the treatment of neurodegenerative disorders.

Introduction

GABA is a major inhibitory neurotransmitter in the vertebrate CNS (Bettler et al., 1998) that mediates fast synaptic inhibition through GABA_A (and GABA_C) ionotropic receptors as well as slow and prolonged synaptic inhibition through the metabotropic GABA_B receptor (Couve et al., 2000). The GABA_B receptor belongs to the class C G-protein-coupled receptors (GPCRs) and is composed of two subunits, GABA_B1 and GABA_B2 (Jones et al., 1998; Kaupmann et al., 1998; White et al., 1998; Kuner et al., 1999; Galvez et al., 2000). GABA_B1 contains the GABA binding site, while GABA_B2 is responsible for G_i/o-protein activation (Galvez et al., 2001; Margeta-Mitrovic et al., 2001). Upon activation of the G-protein, the Gβγ complex represses Ca²⁺ influx by inhibiting Ca²⁺ channels at presynaptic sites, suppresses neurotransmitter release (Harayama et al., 1998), and triggers the opening of K⁺ channels at the postsynaptic level (Lüscher et al., 1997; Schuler et al., 2001). Gα_i/o subunits modulate the level of cAMP by regulating adenylate cyclase activities at postsynaptic sites, and inhibition of neuronal excitability (Simonds, 1999; Billinton et al., 2001). Interestingly, recent studies suggest that in addition to a role in neuronal excitability and plasticity, GABA_B receptor may promote neuronal survival under conditions of metabolic stress or after ischemia (Dave et al., 2005; Kuramoto et al., 2007; Zhang et al., 2007). Accumulating evidences have shown that GABA_B receptor plays an important role in anxiety and depression related disorders while the different classes of antidepressants and mood stabilizers that have been shown to prevent cell death (Cryan and Kaupmann, 2005; McKernan et al., 2009). However, the mechanisms by which GABA_B receptor mediates neuroprotection remains elusive.

Insulin-like growth factor 1 (IGF-1) is essential for normal brain development (Cheng et al., 2000) and promotes neuronal survival by rescuing neurons from apoptosis (D'Mello et al., 1993). IGF-1 triggers autophosphorylation of its cognate tyrosine kinase receptor (IGF-1R) and activates the PI3 kinase/Akt signaling cascade, which in turn mediates the neuroprotective action of IGF-1 (Delcourt et al., 2007). Akt (also known as protein kinase B) is a serine/threonine kinase that functions downstream of PI3 kinase and is critical for neuronal survival (Bondy and Cheng, 2004).

In this study, by using cerebellar granule neurons (CGNs) as a cellular model of apoptosis (D'Mello et al., 1993), we further document the neuroprotective action of GABA_B receptor. Most importantly, we demonstrate that this effect of the GABA_B receptor results from functional transactivation of the IGF-1R, leading to Akt phosphorylation and survival signaling. Together, our results identify a novel cellular and molecular mechanism by which GABA_B receptor can regulate neuronal survival.

Materials and Methods

Materials.

GABA and IGF-1 from mouse were purchased from Sigma. Baclofen, CGP54626, CGP7930, FAK14, and PF573228 were purchased from Tocris Bioscience (Fisher-Bioblock). Pertussis toxin (PTX), AG1024, PP2, LY294002, wortmannin, U73122, and U73343 were purchased from Merck4Biosciences. BAPTA-AM, basal medium Eagle (BME), fetal bovine serum, and other solutions used for cell cultures were from Invitrogen. Primary antibodies including phospho-Ser⁴⁷³ Akt (193H12) antibody, Akt antibody, caspase-3 (#9662) antibody, PARP (#9542) antibody, phospho-Tyr^1135/1136 IGF-IR antibody (19H7), IGF-IR antibody (111A9), phospho-Thr^638/641 PKCα/βII (#9375) antibody, phospho-Tyr⁴¹⁶ Src kinase (#2101) antibody, Src (#2108) antibody, and CREB (#9197) antibody were purchased from Cell Signaling Technology. PKCα (sc-208), PKCβII (sc-210), and FAK (sc-557) antibody were purchased from Santa Cruz Biotechnology.

Primary cerebellar granule neuronal cultures.

One-week-old newborn KunMing mice obtained from Hubei Provincial Research Institute of Experimental Animals were decapitated, and the cerebella were dissected. The tissue was then gently triturated using fire-polished Pasteur pipettes, and the homogenate was centrifuged at 1000 rpm. The pellet was resuspended and plated in tissue culture dishes previously coated with poly-l-ornithine (Sigma). Cells were maintained in a 1:1 mixture of DMEM and F-12 nutrient (Invitrogen), supplemented with 30 mm glucose, 2 mm glutamine, 3 mm sodium bicarbonate, and 5 mm HEPES buffer, decomplemented 10% fetal calf serum, and 30 mm KCl to improve neuronal survival. Three-day-old cultures contained 1.25 × 10⁵ cells/cm².

Induction of apoptosis, TUNEL assay, and measurement of apoptotic nuclei.

After 4 d of culture, CGNs were analyzed for apoptosis. The culture cells were switched from high-K⁺ (BME, 30 mm KCl, 2 mm glutamine, and 0.5% penicillin/streptomycin; K30) to low-K⁺ medium (BME, 5 mm KCl, 2 mm glutamine, and 0.5% penicillin/streptomycin; K5). Cell death analyses were performed after 24 h of treatment.

Terminal deoxynucleotidyltransferase-mediated dUTP nick-end labeling (TUNEL) assay was performed by using the DeadEnd Fluorometric TUNEL System (Promega) according to the manufacturer's instructions. Briefly, CGNs were fixed with 4% paraformaldehyde in PBS and incubated with permeabilization buffer (0.1% Triton X-100 in PBS) for 5 min. The cells were then incubated with rTdT incubation buffer at 37°C for 1 h. Measuring green fluorescence of fluorescein-12-dUTP at 488 nm distinguished the nucleus of apoptotic cells only. Around 800–1200 cells in six to eight different fields were counted per coverslip for detecting TUNEL-positive cells under phase-contrast microscope. Cell numbers were calculated as a percentage of TUNEL-positive cells calculated from total cells.

For measuring apoptotic nuclei inside CGNs, cells were then washed twice with PBS containing 33 mm glucose, fixed with 4% paraformaldehyde in PBS for 30 min at 4°C, and incubated with 1 mg/ml Hoechst 33258 staining for 10 min. Nuclear DNA staining was examined by digital fluorescence imaging microscopy (Axiophot 2 microscope, Zeiss).

Drug treatments.

For Akt, IGF-1R, Src, and PKCα/βII activation assays, cultures were washed once with Ca²⁺-free HEPES-buffered solution (HBS) (containing 10 mm HEPES, pH 7.4, 140 mm NaCl, 4 mm KCl, 2 mm MgSO₄, and 1 mm KH₂PO₄) and preincubated at 37°C with or without indicated inhibitors dissolved in HBS for 60 min. For PTX treatment, the cultures were pretreated 14–16 h with PTX (200 ng/ml) or not. Cells were then stimulated at the indicated time by incubating with GABA, baclofen, CGP7930, or IGF-1 prepared in fresh HBS. At the end of the treatment, the cells were washed quickly with ice-cold Ca²⁺-free PBS pH 7.4, and 200 μl of lysis buffer was added to the cells and placed immediately on ice.

For caspase-3 and PARP assays, the cultures were incubated at different times in K5 media containing the indicated drugs. The cell monolayer was immediately scraped into ice-cold lysis buffer after treatment with K5 media, and the cell lysate was used for Western blotting detection. Drugs were dissolved in HBS with or without dimethyl sulfoxide (DMSO) or alcohol. Whenever DMSO or alcohol were used, HBS containing the same concentration of DMSO and/or alcohol were used as the control vehicle. All immunoblots were performed in at least three independent experiments.

RNAi transfection in MEF cells and primary cerebellar granule neurons.

For IGF-R RNAi knockdown experiment in MEF cells, mammalian shRNA constructs were designed as described previously (Dong et al., 2007). First, we design the target site of small hairpin RNAs (shRNA) for IGF-1R from 3476 to 3494 cDNA (NM_000875.2). The forward DNA template was as follows: gatccccgaagatttcacagtcaacttcctgtcagattgactgtgaaatcttcggtttttg, and the reverse DNA template was as follows: aattcaaaaaccgaagatttcacagtcaatctgacaggaagttgactgtgaaatcttcggg. Synthesized shRNA template oligonucleotides were phosphorylated, annealed, and then ligated into linearized pSIH-H1-copGFP shRNA vector (System Biosciences) digested with EcoRI/BamHI. Murine embryonic fibroblasts (MEFs) were first transfected with shRNA plasmid by Lipofectamine 2000 reagent (Invitrogen) and then with GABA_B1-HA and GABA_B2-Flag plasmid after 24 h; cells were then treated with drugs. For IGF-1R RNAi knockdown experiments in CGNs, primary cerebellar granule neurons were transfected with IGF-IRα/β siRNA (sc-35638), or control siRNA-A (sc-37007) following Santa Cruz siRNA transfection protocol in vitro for 2 d. Cells were then treated with drugs. For PKCα, PKCβII, and FAK RNAi knockdown experiments in MEF cells, MEF cells were transfected following Santa Cruz siRNA transfection protocol in vitro with PKCα (sc-208), PKCβII (sc-210), and FAK siRNA (sc-35353) or control siRNA-A (sc-37007) for 2 d and then transfected with GABA_B1-HA and GABA_B2-Flag plasmid for 24 h; cells were then treated with drugs.

Western blot analysis.

Cell lysates were sonicated, and protein concentrations were determined using Bradford reagent (Bio-Rad Laboratories). Equal amounts of protein (20 μg) were resolved by SDS-PAGE on 8–12% gels. Proteins were transferred to nitrocellulose membranes (Millipore) and blocked in blocking buffer (5% nonfat dry milk in TBS and 0.1% Tween 20) for 1 h at room temperature. The blots were then incubated with primary antibodies at the relevant dilution (Cell Signaling Technology) overnight at 4°C, and with horseradish peroxidase-linked secondary antibodies (1:20,000; Cell Signaling Technology) for 2 h. Immunoblots were detected by using enhanced chemiluminescence reagents (Pierce) and visualized by using x-ray film. The density of immunoreactive bands was measured by using NIH Image software, and all bands were normalized to percentages of control values.

Immunoprecipitation.

Proteins from CGNs were prepared under weakly denaturing conditions to permit the protein and protein interaction. Cultured cells were scraped into a microtube containing ice-cold 1× cell lysis buffer (20 mm Tris, pH 7.5, 150 mm NaCl, 1 mm EDTA, 1 mm EGTA, 1% Triton X-100, 2.5 mm sodium pyrophosphate, 1 mm β-glycerophosphate, 1 mm Na₃VO₄, and 1 μg/ml leupeptin) and homogenized by sonication. The homogenate was centrifuged at 12,000 rpm for 5 min at 4°C. The supernatant was then transferred to a new tube and incubated with GABA_B1 antibody (sc-14006, Santa Cruz Biotechnology) or normal rabbit IgG (#2729, Cell Signaling Technology) overnight at 4°C with gentle rocking. Fifty percent protein A-agarose/Sepharose beads (50 μl of 50% bead slurry, Millipore) were added to the sample and incubated to allow for binding with primary antibody for 2 h at 4°C. Tubes were centrifuged for 30 s at 4°C. The pellet was washed five times with 500 μl of 1× cell lysis buffer and kept on ice during washes. Pellets were resuspended with 3× SDS sample buffer, vortexed and boiled for 5 min, and then centrifuged for 5 min at 12,000 rpm. The boiled samples were then loaded onto SDS-PAGE gel for Western blotting.

cAMP production assay.

Cultures were washed once with HBS and preincubated 2 h with HBS at 37°C after 3–4 d cultured in vitro and then stimulated 30 min with different concentrations of baclofen or forskolin dissolved in stimulation buffer. The assay steps were performed according to protocol of LANCE cAMP 384 Kit (PerkinElmer).

Statistical analysis.

Data are presented as means ± SEM of at least three independent experiments. Statistical analysis was performed by Student's t test. Values with p < 0.05 were considered statistically significant.

Results

GABA_B receptor protects cerebellar granule neurons from low-potassium-induced apoptosis

CGN survival during development involves neuronal excitation, a situation mimicked in cultures by incubating in high-potassium-containing media (30 mm, K30). By reducing the potassium concentration to 5 mm (K5), the cells show rapid neuronal cell death; thus, CGNs have been widely used as a model to study neuronal apoptosis (D'Mello et al., 1993). By measuring the number of apoptotic CGNs by TUNEL assay, we found that the selective agonist of GABA_B receptor, baclofen, at 30 μm or higher concentrations significantly decreased the number of apoptotic CGNs after potassium deprivation (Fig. 1A, top left). Nuclear DNA staining with Hoechst 33258 confirmed that baclofen decreased the number of apoptotic CGNs (Fig. 1A, top right). Caspase-3 activity is commonly associated with apoptotic processes, and was found to be largely increased in CGNs after transfer to K5 media. Poly(ADP-ribose) polymerase-1 (PARP-1), which can be proteolytically cleaved by caspase-3 at the DEVD site to generate a 85 kDa and a 24 kDa fragment, is recognized as a known caspase-3 substrate (Du et al., 1997). Consistent with the anti-apoptotic effects of the GABA_B receptor, baclofen decreased caspase-3 activation and PARP-1 activity (Fig. 1A, bottom; supplemental Fig. 4A,B, available at www.jneurosci.org as supplemental material) but have no effect on CREB protein expression, used here as loading control (Tu et al., 2007). This protective effect resulted from the activation of the GABA_B receptor and was reversed by the GABA_B receptor antagonist CGP54626 (Fig. 1B). The endogenous agonist of the GABA_B receptor, GABA, significantly decreased the number of apoptotic CGNs, and this effect was not blocked by pretreatment with the GABA_A and GABA_C receptor antagonists, bicuculline and TPMPA, respectively (Fig. 1C). The positive allosteric modulator of GABA_B receptor, CGP7930, also significantly decreased the number of apoptotic CGNs (Fig. 1D, left) and caspase-3 activity induced by potassium deprivation (Fig. 1D, right; supplemental Fig. 4A, available at www.jneurosci.org as supplemental material). Together, these results demonstrate that specific activation of GABA_B receptor, but not GABA_A or GABA_C receptors, protects CGNs from low-potassium-induced apoptosis.

Figure 1.

GABA_B receptor activation protect neurons from potassium-deprivation-induced apoptosis. A, Dose–response of baclofen on apoptosis when CGNs were transferred from high (30 mm K⁺, K30) to low (5 mm K⁺, K5) potassium concentrations for 24 h. Apoptotic cells were measured by TUNEL assay (left) and observed by Hoechst 33258 staining (right). Time course study of the effect of baclofen at 300 μm on caspase-3 activation and PARP activity were shown in the bottom panel and CREB protein expression was used as loading control. B, C, Effect of GABA_A [bicuculline (10 μm)] and GABA_C [TPMPA (10 μm)] receptor antagonists on apoptotic CGNs after GABA (100 μm) stimulation. For all TUNEL assay or Hoechst 33258 staining, cells were treated with drugs in K5 media for 18–20 h before fixing and staining. For caspase-3 activation, cells were treated with drugs in K5 media for 0.1–12 h. D, Dose–response of the GABA_B receptor-positive allosteric modulator CGP7930 on apoptotic CGNs (left) and the effects of CGP7930 at 30 μm on caspase-3 activation (right). **p < 0.01, ***p < 0.001, versus K30 conditions. ⁺⁺⁺p < 0.001; ns, no significance, versus in the absence of antagonist. The blots shown are representative of three separate experiments.

GABA_B receptor-mediated neuroprotection is via G_i/o-protein and PI3 kinase but not intracellular cAMP levels

To elucidate the downstream signaling pathways involved in GABA_B receptor-induced neuroprotection, we analyzed the importance of the G_i/o-protein, which is coupled to the GABA_B receptor in neurons (Mannoury la Cour et al., 2008), and PI3 kinase, which plays an important role in cell survival (Downward, 2004). Under low-potassium-induced apoptosis conditions, the G_i/o inhibitor, PTX, reversed the protective effect of baclofen/CGP7930 on cell survival (Fig. 2A, left). In agreement with this, PTX suppressed the inhibitory action of baclofen and CGP7930 on caspase-3 degradation at low potassium concentrations (Fig. 2A, right; supplemental Fig. 4C, available at www.jneurosci.org as supplemental material). Interestingly, two specific inhibitors of PI3 kinase, LY294002 (Fig. 2B, left) and wortmannin (data not shown), abolished the neuroprotective effect of baclofen and CGP7930. LY294002 also reversed the effect of baclofen and CGP7930 on caspase-3 activity in CGNs (Fig. 2B, right; supplemental Fig. 4D, available at www.jneurosci.org as supplemental material). Apoptosis of CGNs induced by low potassium has been known to be prevented either by elevated intracellular cAMP levels (D'Mello et al., 1993; Galli et al., 1995). Although coupled to G_i/o type of G-proteins, the GABA_B receptor could well increase cAMP formation through an indirect pathway, then leading to inhibition of apoptosis. This is unlikely since no increase in cAMP production could be detected after baclofen stimulation of CGNs, whereas a nice increase of cAMP could be detected upon adenylate cyclase activation with forskolin (Fig. 2C).

Figure 2.

The neuroprotective effect of GABA_B receptor is dependent on G_i/o-protein and PI3 kinase but independent of intracellular cAMP levels. A, B, Effect of PTX (200 ng/ml) (A) and LY294002 (20 μm) (B) on the protective effect of baclofen (300 μm) or CGP7930 (50 μm) on apoptotic CGNs (left) and caspase-3 activation (right). For TUNEL assay, cells were treated with drugs in K5 media for 18–20 h. For caspase-3 activation, cells were treated with drugs in K5 media for 1–12 h. ***p < 0.001, versus K30 conditions. ⁺⁺⁺p < 0.001, versus the absence of PTX or LY294002. The blots shown are representative of three separate experiments. C, GABA_B receptor activation does not induce cAMP production in CGNs. Dose–response effect of baclofen or adenylyl cyclase activator forskolin on cAMP production.

These results demonstrate that G_i/o-protein and PI3 kinase but not intracellular cAMP are involved in GABA_B receptor-mediated neuroprotection of CGNs from potassium deprivation.

GABA_B receptor activates PI3 kinase/Akt via G_i/o-protein, PLC, and Src kinases

The GABA_B receptor can activate Akt, a downstream PI3 kinase effector that is important for cell survival (Schlessinger, 2000), and we explore whether this signaling pathway is involved in the neuroprotective effect of this receptor. In CGNs, GABA caused a rapid and transient increase in Akt phosphorylation on Ser⁴⁷³ residue with no changes in total Akt protein expression levels (Fig. 3A). Similar results were obtained with baclofen or CGP7930 (Fig. 3A), and both baclofen and CGP7930 induced Akt phosphorylation in a dosage-dependent manner (Fig. 3B). CGP54626 can antagonize baclofen's effect on Akt phosphorylation (Fig. 3C), while CGP7930-induced Akt phosphorylation was not antagonized by CGP54626, consistent with its action at a different site on the GABA_B receptor (Binet et al., 2004). In addition, GABA-mediated Akt phosphorylation could not be blocked by pretreatment with bicuculline or TPMPA (Fig. 3D), and is in agreement with the specific involvement of GABA_B receptor in Akt activation.

Figure 3.

GABA_B receptor activation can increase Akt phosphorylation. A, Time course of Akt activation in CGNs after incubation with GABA (100 μm), baclofen (100 μm), or CGP7930 (50 μm). Representative immunoblots were used to quantify phosphorylated Akt levels (pAkt). *p < 0.05, **p < 0.01, ***p < 0.001, versus basal level for baclofen dose–response. B, Response of baclofen or CGP7930-induced Akt phosphorylation after 10 min incubation. Immunoblots were used to quantify pAkt levels. *p < 0.05, **p < 0.01, versus basal levels. C, D, Cells were treated with CGP54626, bicuculline, or TPMPA at 10 μm for 20 min before the stimulation of baclofen/GABA at 100 μm (10 min). Akt phosphorylation was measured by quantifying pAkt from immunoblots. ***p < 0.001, versus basal levels. ⁺⁺⁺p < 0.001; ns, nonsignificance, versus treated with baclofen, GABA, or CGP7930 in the absence of antagonist. The blots shown are representative of three separate experiments.

We then further analyzed the signaling pathways that lead the GABA_B receptor to activate Akt. PTX treatment completely inhibited GABA_B receptor-induced Akt activation, demonstrating the involvement of G_i/o-protein in the GABA_B receptor-mediated pathway (Fig. 4A). In various cell types, G_i/o-coupled receptors have been shown to activate the Src family kinases (Satoh et al., 1992; Luttrell et al., 1996; Thodeti et al., 2000). PP2, a Src-family kinase inhibitor, completely abolished Akt phosphorylation induced by baclofen and CGP7930 (Fig. 4B). G_i/o-coupled receptors have been shown to enhance phospholipase C (PLC) activity and activate protein kinase C (PKC) through Gβγ subunits (Blank et al., 1992; Selbie and Hill, 1998). Accordingly, we showed that pretreating CGNs with U73122, an inhibitor of PLC but not its inactive analog U73343, completely abolished Akt phosphorylation (Fig. 4C). We showed that baclofen can induce the phosphorylation of PKCα/βII on Thr^638/641 residues, an effect that was blocked by U73122 (Fig. 4E). We verified that PI3 kinase is involved in GABA_B receptor-mediated Akt phosphorylation by showing that LY294002 can completely block this effect (Fig. 4D).

Figure 4.

G_i/o-protein, Src family kinase, PLC, and PI3 kinase are required for GABA_B receptor-mediated Akt phosphorylation. A–D, Cells were pretreated with PTX (200 ng/ml), PP2 (5 μm), U73122 (5 μm), U73343 (5 μm), or LY294002 (10 μm) for 60 min, and Akt phosphorylation was induced by 100 μm baclofen (or 50 μm CGP7930) for 10 min. pAkt levels were quantified by immunoblotting. E, Time course of the PKCα/βII phosphorylation in CGNs after incubation with baclofen at 100 μm. Cells were pretreated with U73122 at 5 μm for 60 min before baclofen stimulation (100 μm). To quantify pAkt or pPKCα/βII, immunoblots from separate experiments (mean ± SEM; n = 3–5) were analyzed, and a representative immunoblot is shown. For all results, *p < 0.05, **p < 0.01, and ***p < 0.001, versus basal levels. ⁺⁺p < 0.01, ⁺⁺⁺p < 0.001; ns, nonsignificance, versus treated with baclofen or CGP7930 in the absence of inhibitors.

Together, these results demonstrate that G_i/o-protein, Src-family kinase, PLC, and PI3 kinase are required for GABA_B receptor-induced Akt phosphorylation, and may be required for the neuroprotective effect of GABA_B receptor in CGNs.

IGF-1R is required for GABA_B receptor-mediated Akt phosphorylation and neuroprotection

It has been shown that apoptosis induced by potassium deprivation in CGNs can be inhibited by IGF-1 (D'Mello et al., 1993). In our work, we confirmed that IGF-1 can induce a sustain phosphorylation on IGF-1R at Tyr^1135/1136 residues and Akt at Ser⁴⁷³ residue (Fig. 5A). This GABA_B receptor-induced Akt activation was blocked by pretreatment with two different classes of IGF-IR inhibitors, AG1024 (Fig. 5B) and NDGA (7H) (supplemental Fig. 1A, available at www.jneurosci.org as supplemental material) (Blecha et al., 2007).

Figure 5.

IGF-1R is required for GABA_B receptor-mediated neuroprotection. A, Time course of IGF-1 (10 ng/ml)-induced IGF-1R and Akt phosphorylation in CGNs. B, Effect of pretreatment with AG1024 at 0.1 μm for 60 min on Akt phosphorylation, followed by induction by baclofen (100 μm) or CGP7930 (50 μm) for 10 min. To quantify pAkt or pIGF-1R, immunoblots from separate experiments (mean ± SEM; n = 3–5) were analyzed. For all results, *p < 0.05, **p < 0.01, and ***p < 0.001, versus the basal levels, and ⁺⁺⁺p < 0.001, versus treated with baclofen or CGP7930 in the absence of AG1024. C, Effect of AG1024 (0.1 μm) on apoptotic CGNs after baclofen (100 μm) stimulation. For all TUNEL assay (left) or Hoechst 33258 staining (right), cells were treated with drugs in K5 media for 18–20 h before fixing and staining. ***p < 0.001; ns, no significance, versus K5 conditions. The blots shown are representative of three separate experiments. D, Effect of AG1024 (0.1 μm) on caspase-3 activity in CGNs induced by GABA (100 μm), baclofen (100 μm), or CGP7930 (50 μm). The blots shown are representative of three separate experiments. For caspase-3 activation, cells were treated with drugs in K5 media for 2–12 h. E, Effect of shRNA of IGF-1R (IGF-R3467) on Akt phosphorylation induced by baclofen (100 μm) for 5 min in MEF cells cotransfected GABA_B1 and GABA_B2. Expression of IGF-R3467 decreased MEF cells endogenous IGF-1R expression but not Src kinase expression (left). The blots shown are representative of three separate experiments.

Several lines of evidence confirmed that IGF-1R is necessary for GABA_B receptor-mediated neuroprotection. First, AG1024 reversed the decrease of caspase-3 activity induced by GABA, baclofen, or CGP7930 (Fig. 5C; supplemental Fig. 1E, available at www.jneurosci.org as supplemental material). Second, AG1024 suppressed the effect of baclofen on the protection of CGNs from apoptosis (Fig. 5D). Third, transfection with shRNA of IGF-1R into MEF cells cotransfected with GABA_B1 and GABA_B2 could inhibit IGF-1R expression, which in turn inhibited baclofen-induced Akt phosphorylation (Fig. 5E). Finally, transfection with IGF-1R siRNA into CGNs could also inhibit baclofen-induced Akt phosphorylation by reducing endogenous IGF-1R expression in CGNs (supplemental Fig. 1B, available at www.jneurosci.org as supplemental material). Furthermore, we verified that CGP54626, a GABA_B antagonist, failed to modify the effect of IGF-1 on cell signaling (supplemental Fig. 1C, available at www.jneurosci.org as supplemental material) and CGN neuroprotection from apoptosis (supplemental Fig. 1D, available at www.jneurosci.org as supplemental material).

GABA_B receptor activation induces IGF-1R transactivation

It has been reported that the GPCR angiotensin II type 1 receptors (AT1Rs) can transactivate IGF-1R to mediate downstream signaling from AT1Rs to PI3 kinase (Zahradka et al., 2004). We first showed that baclofen or CGP7930 can cause a transient increase of IGF-1R phosphorylation without altering IGF-1R protein expression levels (Fig. 6A), and that this effect was abolished by inhibition with CGP54626 and PTX (Fig. 6B).

Figure 6.

Activation of GABA_B receptor transactivated IGF-1R. A, Time course of baclofen- (100 μm) and CGP7930- (50 μm) induced IGF-1R phosphorylation in CGNs. B, Effects of pretreatment with PTX (200 ng/ml) or CGP54626 (10 μm) on IGF-1R phosphorylation induced by baclofen (100 μm) or CGP7930 (50 μm) for 10 min, and representative immunoblots. C, Western blot analysis of IGF-1R coimmunoprecipitated from CGNs with GABA_B1 antibody. The membrane was immunoblotted by anti-IGF-1R antibodies. The total lysate (input) was analyzed. D, Effect of the neutralizing antibody anti-IGF-1R (αIR3; 2 μg/ml) on baclofen- (100 μm) or IGF-1 (10 ng/ml) induced IGF-1R phosphorylation for 10 min. For all IGF-IR phosphorylation results, representative immunoblots are shown under the quantified data of pIGF-IR analyzed from separate experiments (mean ± SEM; n = 3–5). For all results, *p < 0.05, **p < 0.01, and ***p < 0.001, versus the basal levels, and ⁺p < 0.05, ⁺⁺p < 0.01, ⁺⁺⁺p < 0.001, and ns, nonsignificance, versus treatment with baclofen or CGP7930 in the absence of CGP54626, PTX, IGF-1, or mAb αIR3.

Functional cross talk between two receptors could be due to their association in the same signaling platform or through indirect interactions (Rives et al., 2009). We found that IGF-1R could be coimmunoprecipitated with GABA_B1 from CGN lysates (Fig. 6C), indicating that IGF-1R and GABA_B1 are part of the same signaling complex (Fig. 6C).

Two different mechanisms have been proposed for GPCR-induced transactivation of receptor tyrosine kinases: ligand-dependent or ligand-independent mechanisms (Shah and Catt, 2004; Delcourt et al., 2007). The αIR3 mouse monoclonal antibody against IGF-1R specifically recognizes the extracellular α-subunit of IGF-1R and inhibits its ligand-mediated effects (Kienlen Campard et al., 1997). We found that αIR3 had no effect on baclofen-induced IGF-1R phosphorylation (Fig. 6D), whereas it decreased IGF-1-induced IGF-1R phosphorylation.

Together, these results demonstrate that GABA_B receptor activation of G_i/o-protein can induce ligand-independent IGF-1R transactivation in CGNs.

GABA_B receptor-induced IGF-1R transactivation requires PLC through Ca²⁺-dependent FAK1 pathway

The selective PLC inhibitor U73122, but not its inactive analog U73343, abolished baclofen-induced IGF-1R phosphorylation (Fig. 7A). However, PP2 failed to block baclofen-stimulated IGF-1R phosphorylation, whereas AG1024 completely inhibited baclofen-induced IGF-1R phosphorylation (Fig. 7B), indicating that Src is not involved in this pathway. Our data show that PLC is involved in GABA_B receptor-induced IGF-1R phosphorylation and that Src kinases are acting downstream of IGF-1R. Note that we detected two bands using the IGF-IR antibody (111A9) from Cell Signaling Technology (Fig. 7A, right). We think that the fragment with a lower molecular weight corresponds to a degradation product of IGF-1R, as it can also be found in the Cell Signaling Technology product information provided for IGF-IR antibody (111A9).

Figure 7.

PLC downstream effectors FAK1 but not Src kinase is involved in GABA_B receptor-mediated IGF-1R transactivation. A, Cells were pretreated with U73122 (5 μm) or U73343 (5 μm) for 60 min and induced by baclofen at 100 μm for 10 min. IGF-1R phosphorylation was measured by immunoblotting. B, PP2 (5 μm) was used to pretreat cells for 60 min or AG1024 (0.1 μm) before baclofen (100 μm) for 10 min. IGF-1R phosphorylation was detected by immunoblotting. C, Time course of baclofen (100 μm)-induced Src kinase phosphorylation in CGNs. D, Effect of pretreatment with U73122 (5 μm) or AG1024 (0.1 μm) on Src kinase phosphorylation induced by baclofen (100 μm) for 10 min. E, Effect of intracellular Ca²⁺ and FAK1 on baclofen-induced Akt phosphorylation for 10 min. Cells were pretreated with cell membrane-permeable Ca²⁺ chelator BAPTA-AM (10 μm) for 30 min and FAK1 inhibitors [FAK14 (50 μm) or PF573228 (10 μm)] for 10 min before the CGNs were stimulated with baclofen (100 μm) for 10 min, respectively. Akt phosphorylation was measured by immunoblotting. For all results, representative immunoblots are shown under the quantified data of pIGF-IR, pSrc, or pAkt analyzed from separate experiments (mean ± SEM; n = 3–5). *p < 0.05, **p < 0.01, and ***p < 0.001, versus the basal levels, and ⁺⁺p < 0.01, ⁺⁺⁺p < 0.001, ns, no significance, versus treated with baclofen in the absence of inhibitors. F, MEF cells cotransfected with GABA_B1 and GABA_B2 were pretreated with control RNAi or RNAi of FAK1 for 48 h and then stimulated with baclofen for 5 min, respectively. IGF-1R and Akt phosphorylation were measured by immunoblotting. The blots shown are representative of three separate experiments.

Further experiments showed that baclofen can induce Src kinase phosphorylation at Tyr⁴¹⁶ residue in a transient manner (Fig. 7C), an effect that was blocked by both U73122 and AG1024 (Fig. 7D), but not by U73343 (supplemental Fig. 2A, available at www.jneurosci.org as supplemental material). However, LY294002 did not inhibit baclofen-induced phosphorylation of either IGF-1R or Src (supplemental Fig. 2B, available at www.jneurosci.org as supplemental material), suggesting that Src kinase can act upstream of PI3 kinase/Akt.

Finally, we examined the possible involvement of PLC downstream effectors such as PKC, intracellular Ca²⁺, and focal adhesion kinase (FAK1) on GABA_B receptor-induced IGF-1R phosphorylation. PKC inhibitor (Gö6983 or Gö6976) failed to inhibit baclofen-induced Akt phosphorylation, whereas it inhibited baclofen-induced CREB phosphorylation in CGNs (supplemental Fig. 3A,B, available at www.jneurosci.org as supplemental material). Furthermore, RNAi-mediated knockdown experiments of PKCα and PKCβII failed to block baclofen-induced Akt phosphorylation in MEF cells coexpressing GABA_B1 and GABA_B2 (supplemental Fig. 3C, available at www.jneurosci.org as supplemental material). These data suggest that GABA_B receptor-induced IGF-1R/Akt signaling does not need PKC activation. Group 1 metabotropic glutamate receptors (mGluR1 and mGluR5) that stimulate PLC can lead to both mobilization of intracellular Ca²⁺ and activation of PKC. However, mGluR1- and mGluR5-induced tyrosine phosphorylation of the downstream FAK1 results from the Ca²⁺/CaM signaling and not from PKC activation (Shinohara et al., 2001). Here we showed that pretreatment of CGNs with either BAPTA-AM (cell membrane-permeable Ca²⁺ chelator) or FAK1 inhibitors (FAK14 or PF573228) inhibited baclofen-induced Akt phosphorylation (Fig. 7E). Furthermore, we showed that RNAi knockdown for FAK1 could block baclofen-induced phosphorylation of both IGF-1R and Akt in MEF cells coexpressing GABA_B1 and GABA_B2 (Fig. 7F). These indicated that GABA_B receptor induced IGF-1R transactivation through PLC/Ca²⁺-dependent FAK1 pathway, which in turn activated Akt.

Together, these results demonstrate that GABA_B receptor-induced IGF-IR transactivation is mediated by PLC/FAK1 pathway via an intracellular Ca²⁺-dependent but PKC-independent mechanism and that GABA_B receptor can induce Src kinase phosphorylation via IGF-1R transactivation (Fig. 8).

Figure 8.

Schematic representation of the signaling pathway mediated by GABA_B receptor in CGN neuroprotection. Activation of GABA_B receptor leads to the release of Gβγ subunits from G_i/o-protein to activate PLC/Ca²⁺-dependent FAK1 pathway, which in turn transactivates IGF-IR to induce Akt phosphorylation through the Src family kinases.

Discussion

GABA_B receptor has previously been shown to promote neuronal survival, but the mechanism of such neuroprotection remains elusive. In the present study, we elucidated the mechanism by which GABA_B receptor induced anti-apoptotic effect in CGNs, a model system in which apoptosis can be induced by potassium deprivation. We demonstrated that GABA_B receptor-induced neuroprotection is mediated by Akt phosphorylation through a signaling pathway that involves IGF-1R transactivation. We show that the mechanism of GABA_B receptor-induced Akt activation is through G_i/o-protein and PLC/Ca²⁺-dependent FAK1, which leads to IGF-1R transactivation and Src family kinase and PI3 kinase/Akt phosphorylation.

Apoptosis of CGNs induced by low potassium is known to be prevented either by elevated intracellular cAMP levels or by IGF-1 (D'Mello et al., 1993; Galli et al., 1995). In accordance with this, activation of pituitary adenylate cyclase-activating polypeptide (PACAP) receptor, which is known to couple to G_s, produces the anti-apoptotic effect in CGNs by increasing intracellular cAMP levels (Kienlen Campard et al., 1997; Delcourt et al., 2007). In our present work, we demonstrated that GABA can protect CGNs from apoptosis, and this effect is mediated by GABA_B receptor coupling to G_i/o-protein and not by the GABA_A and GABA_C receptors. Interestingly, we found that this anti-apoptotic effect mediated by GABA_B receptor may not require changes in intracellular cAMP levels. Works have shown that GABA_B receptor from rat brain membranes can couple to both G_o- and G_i-proteins, resulting in an inhibition of adenylate cyclase activity and cAMP production, but not to G_q and G_s (Nishikawa et al., 1997; Odagaki et al., 2000; Odagaki and Koyama, 2001). Previous studies suggested that due to the low expression of G_i in comparison to G_o, GABA_B receptor may be preferentially coupled to G_o in neurons (Gierschik et al., 1986). PAM effects such as CGP7930 that act on the GABA_B2 have been shown to enhance efficacy of coupling to G_o over G_i (Mannoury la Cour et al., 2008). Finally, unlike G_i, most functions of G_o can be interpreted through the actions of a common pool of Gβγ (Ghil et al., 2006). Together, our data suggest that the anti-apoptotic effect of GABA_B receptor is mediated by Gβγ subunits released from G_i/o-protein and IGF-1R signaling pathway rather than by an increase in cAMP levels.

Earlier reports have shown that IGF-1 acts as a cell survival growth factor, protecting CGNs from apoptosis through the PI3 kinase/Akt pathway (Linseman et al., 2002; Subramaniam et al., 2005). Our data demonstrated that GABA_B receptor can transactivate IGF-1R. Recent studies in different cell types have established the potential involvement of RTKs in transducing the growth-promoting signals of GPCRs (Shah and Catt, 2004). To date, a few GPCRs in the CNS have been investigated in terms of their ability to transactivate RTKs in neurons, although much is known for peripheral GPCRs transactivating epidermal growth factor receptor (EGFR), neurotrophin receptor (Trk), platelet-derived growth factor receptor (PDGFR), and fibroblast growth factor receptor (FGFR) (Peavy et al., 2001; Shah and Catt, 2004). From these studies, two different mechanisms of RTK transactivation by GPCRs have been reported. One is ligand-dependent such as gonadotropin-releasing hormone (GnRH) receptors-EGFR transactivation resulting from the induction of metalloprotease-dependent ectodomain shedding from heparin-binding EGF (Peavy et al., 2001; Shah et al., 2003; Shah and Catt, 2004). Alternatively, the receptor can promote the release of the RTK agonist as observed for the μ-opioid receptors-FGFR-1 transactivation (Belcheva et al., 2002). The second model is ligand independent and results from a direct phosphorylation of the transactivated RTKs by a GPCR downstream tyrosine kinase. This has been well illustrated by the adenosine- and PACAP-induced Trk receptor transactivation (Lee and Chao, 2001; Lee et al., 2002) and dopamine D₄ receptor-mediated PDGFR transactivation (Heldin and Westermark, 1999). Of major importance, whereas the first mechanism leads to the RTK activation in cells surrounding the activated GPCR, leading to a diffusion of the effect, the second mechanism is exclusively mediated by intracellular events such that RTK transactivation is limited to the cells expressing the activated GPCR.

Here, we show that GABA_B receptor-induced IGF-1R transactivation is independent of endogenous IGF-1, and likely results mainly from an intracellular pathway, such that only CGNs activated by GABA_B receptor agonists are expected to benefit from the neuroprotective effect. Our coimmunoprecipitation experiments show that GABA_B receptor and IGF-1R can be part of the same signaling complex; however, there is no evidence suggesting whether these two receptors form direct interactions, nor that these may be required for the observed cross talk. This will be the subject of future work. Our results suggest that GABA_B receptor may transactivate IGF-1R through a ligand-independent mechanism that may involve downstream GABA_B receptor signaling.

To date, there are limited numbers of reports demonstrating IGF-1R transactivation by GPCRs. The best characterized example is angiotensin II type 1 (AT1) receptor, which was shown to activate PI3 kinase through IGF-1R transactivation in smooth muscle cells (Zahradka et al., 2004). However, the mechanism involved is not exactly the same as those reported here. In smooth muscle cells, AT1 receptor couples to both G_i/o- and G_q-proteins and mediates IGF-1R phosphorylation through Src activation (Zahradka et al., 2004). In CGNs, we show that GABA_B receptor-mediated IGF-1R phosphorylation is independent of Src family kinases activation, but Src is involved in the downstream effect of IGF-1R, in agreement with the IGF-1R action on these neurons (Delcourt et al., 2007).

Given the relative abundance of G_i/o-protein in neurons, this G-protein is a likely source of Gβγ subunits for PLC activation (Blank et al., 1992). Accordingly, a number of GPCR agonists that activate G_i/o, including LPA and thrombin, can activate PLC through Gβγ (Rozengurt, 2007), then leading to a rapid increase in the intracellular concentration of Ca²⁺ (Berridge et al., 2000), and activation of protein kinase C (PKC) through DAG (Dempsey et al., 2000). We show here that activation of GABA_B receptor-induced PKCα/βII phosphorylation in CGNs, in agreement with other studies (Thompson and Gahwiler, 1992; Tosetti et al., 2004; Pontier et al., 2006), consistent with GABA_B receptor-mediated PLC activity through Gβγ subunits. However, our data show that PKC activation is not required for GABA_B receptor-mediated IGF-1R transactivation and Akt phosphorylation. This is consistent with the absence of IGF-1R phosphorylation by PKC, in contrast to what has been reported for EGFR (Hunter et al., 1984).

Although PKC activation does not appear to be involved, GABA_B receptor-mediated increase in intracellular Ca²⁺ concentration (New et al., 2006; Mizuta et al., 2008) plays a major role in IGF-1R transactivation. Recently, a large body of evidence suggested that a rapid increase of the tyrosine phosphorylation for a nonreceptor tyrosine kinase, FAK, is a common response to multiple GPCR agonists (Rozengurt, 2002). Interestingly, whereas tyrosine phosphorylation of FAK induced by the neuropeptide bombesin is not mediated by either of the two signaling arms initiated by PLC, namely Ca²⁺ and PKC (Sinnett-Smith et al., 1993), activation of mGluR1/5 induces FAK tyrosine phosphorylation through PLC-modulated Ca²⁺/calmodulin signaling but independent of PLC-induced PKC activation (Shinohara et al., 2001). Indeed, Ca²⁺ transient signals mediated by GPCRs were reported to induce rapid phosphorylation of FAK at serine residues (Ser⁹¹⁰, Ser⁸⁴³) through Ca²⁺/calmodulin-dependent protein kinase II (CaMKII), indicating that FAK is a point of integration of tyrosine and serine phosphorylation (Fan et al., 2005; Jacamo et al., 2007; Jiang et al., 2007). Our results showed that Ca²⁺ signaling and FAK1 are involved in GABA_B receptor-IGF-1R transactivation, suggesting that GABA_B receptor-mediated Ca²⁺ signaling may induce phosphorylation of FAK at Ser⁸⁴³, which in turn induces FAK tyrosine kinase activity and downstream signaling. Accordingly, FAK was reported to be involved in phosphorylation of IGF-1R and may influence its downstream signaling such as Akt and ERK and receptor stability through a direct interaction (Liu et al., 2008; Watanabe et al., 2008; Andersson et al., 2009; Zheng et al., 2009). Together, we demonstrated that one GPCR could transactivate IGF-1R through Gβγ/PLC/Ca²⁺-dependent FAK1 pathway. Whether and how GABA_B-mediated Gβγ subunits/PLC/Ca²⁺ signaling/FAK Ser⁸⁴³ phosphorylation signaling could participate in IGF-1R transactivation through phosphorylation of FAK at tyrosine residue remains to be established.

Our results may have important significance in cerebellum development and function, as GABA_B receptor has been shown to actively participate in the regulation of life and death decisions in CGNs during cerebellum development. GABA was reported to be acting as a trophic factor that can influence proliferation, migration, differentiation, synapse maturation, and cell death in immature neurons in the cerebellum (Eilers et al., 2001) through the GABA_A receptors (Owens and Kriegstein, 2002; Ben-Ari et al., 2007). The developmental switch from GABA-mediated excitation to inhibition occurs after birth (Ganguly et al., 2001; Ben-Ari et al., 2007). In mature neurons, the inhibitory action of GABA is balanced by the excitatory effect of glutamate (Foster and Kemp, 2006). Overall, GABA_B receptor-mediated IGF-1R pathways may be important in protecting neurons from apoptosis after GABA_A receptor turn inhibitory in CGNs from postnatal mice. The anti-apoptotic effects of the GABA_B receptor could be used to balance the effect of glutamate attenuating the survival properties of IGF-1 in hippocampal neurons, by decreasing IGF-1R signaling and responsiveness (Zheng and Quirion, 2009). Further studies will be required to elucidate the neuroprotective effects of GABA_B receptor, and to verify whether GABA_B receptor plays a similar role in other regions of the brain where this receptor is strongly expressed, such as the cortex and hippocampus.