Abstract
Oligodendrocytes (OLs) express functional GABAA receptors (GABAARs) that are activated by GABA released at synaptic contacts with axons or by ambient GABA in extrasynaptic domains. In both instances, the receptors’ molecular identity has not been fully defined. Furthermore, data on their structural diversity in different brain regions and information on age-dependent changes in their molecular composition are scant. This lack of knowledge has delayed access to a better understanding of the role of GABAergic signaling between neurons and OLs. Here, we used functional, and pharmacological analyses, as well as gene and protein expression of GABAAR subunits, to explore the subunit combination that could explain the receptor functional profile expressed in OLs from the neonate rat. We found that GABAAR composed of α3β2γ1 subunits mimicked the characteristics of the endogenous receptor when expressed heterologously in Xenopus laevis oocytes. Either α3 or γ1 subunit silencing by small interfering RNA transfection changed the GABA-response characteristics in oligodendrocyte precursor cells, indicating their participation in the endogenous receptor conformation. Thus, α3 subunit silencing shifted the mean EC50 for GABA from 75.1 to 46.6 µM, whereas γ1 silencing reduced the current amplitude response by 55%. We also observed that β-carbolines differentially enhance GABA responses in oligodendroglia as compared with those in neurons. These results contribute to defining the molecular and pharmacological properties of GABAARs in OLs. Additionally, the identification of β-carbolines as selective enhancers of GABAARs in OLs may help to study the role of GABAergic signaling during myelination.
Significance Statement GABAergic signaling through GABAA receptors (GABAARs) expressed in the oligodendroglial lineage contributes to the myelination control. Determining the molecular identity and the pharmacology of these receptors is essential to define their specific roles in myelination. Using GABAAR subunit expression and silencing, we identified that the GABAAR subunit combination α3β2γ1 conforms the bulk of GABAARs in oligodendrocytes from rat neonates. Furthermore, we found that these receptors have differential pharmacological properties that allow specific positive modulation by β-carbolines.
Introduction
Oligodendrocytes (OLs) and their progenitor cells (OPCs) are endowed with membrane receptors sensitive to γ-aminobutyric acid (GABA), a key neurotransmitter in the central nervous system, suggesting an important role in the communication between neurons and oligodendroglial cells. Correspondingly, specific GABAergic synaptic neuron-OPC contacts have been demonstrated in several brain regions (Lin and Bergles, 2004; Jabs et al., 2005; Kukley et al., 2008; Káradóttir et al., 2008; Tanaka et al., 2009; Orduz et al., 2015; Zonouzi et al., 2015). Volume transmission has also been proposed as an important element for axon-glial chemical communication (Vélez-Fort et al., 2010; Wake et al., 2015). The main electrical response in OLs to GABA corresponds to the activation of a pentameric receptor-channel type A [GABAA receptor (GABAAR)], allowing the efflux of Cl− ions, causing membrane depolarization and the subsequent opening of voltage-dependent Ca2+ channels (Gilbert et al., 1984; Hoppe and Kettenmann, 1989; Kirchhoff and Kettenmann, 1992). This GABA response in OLs ultimately increases the intracellular Ca2+ concentration, which is involved in numerous phenomena, including the myelination process (Cesetti et al., 2012; Cheli et al., 2015, 2016; Wake et al., 2015; Pitman and Young, 2016; Baraban et al., 2018; Krasnow et al., 2018). GABA sensitivity seems to be regulated throughout the course of OL maturation, with the lowest sensitivity observed in myelinating OLs in adulthood. Thus, the GABA response is maximal during postnatal development, a stage in which the premyelinating OLs contact with axons and initiate the myelinating program. It has been shown that GABA sensitivity is regulated by contact between the OLs and the axon through a mechanism that involves the regulation of protein subunit synthesis (Arellano et al., 2016). Importantly, GABAARs expressed in perinatal OLs and OPCs present distinctive characteristics comparable to those displayed in neurons and astrocytes (von Blankenfeld et al., 1991; Bronstein et al., 1998; Williamson et al., 1998; Arellano et al., 2016). Whether the GABAAR expressed during postnatal development has a similar molecular composition to the GABAAR expressed in OPCs at different ages or brain areas (especially in gray matter vs. white matter) remains unknown, given that the observed characteristics and molecular identity indicate diversity in composition (see, e.g., Vélez-Fort et al., 2012; Passlick et al., 2013; Balia et al., 2015).
It is fundamental to determine the GABAAR identity expressed in OLs during development, as well as its specific pharmacology, since GABAergic signaling is a prospective target for therapeutic strategies against demyelinating illness (e.g., Zonouzi et al., 2015; Shaw et al., 2018, 2019; Cisneros-Mejorado et al., 2019; Serrano-Regal et al., 2019). We proposed that the main receptor type in OLs and OPCs, from the optic nerve of P12 rats and forebrain of newborn rats, respectively, comprises a combination of the α3, β2 or β3, and γ1 or γ3 subunits (Arellano et al., 2016). This is based on the functional and pharmacological profile of the GABA response in OLs and supported by information presented by distinct groups about the expression of GABAAR subunits in the oligodendroglial lineage. Accordingly, the most abundant mRNA expressed for α subunits is for α3 and α2 subunits, whereas γ2 and β1 subunit mRNA expression is absent or low in OPC preparations (Cahoy et al., 2008; Larson et al., 2016).
A characteristic that distinguishes the oligodendroglial receptor, with respect to neuronal GABAARs, is a positive modulatory effect of β-carbolines such as N-butyl-β-carboline-3-carboxylate (β-CCB; Arellano et al., 2016; Cisneros-Mejorado et al., 2019).
In this exploratory work, oligodendroglial GABAAR subunits were cloned and heterologously expressed in accordance with the studies cited above and on the analyses of publicly available RNA expression data sets in OPCs. Then, functional, and pharmacological analyses of six possible GABAAR combinations were performed. Their characteristics were compared with those displayed by the endogenous oligodendroglial GABAAR. Priority was given to analyzing the possible binding site for β-CCB. Data showed that α3β2γ1 GABAAR closely mimicked the characteristics expressed by the endogenous receptor in OLs, including positive modulation by β-CCB that did not involve its binding to classic sites for benzodiazepines (BZDs). Moreover, knocking down the expression of either α3 or γ1 subunits using the small interfering RNA (siRNA) technique in vitro produced changes in the endogenous oligodendroglial GABA response, thus indicating that these subunits are involved in the receptor conformation. Finally, the effect of different β-carbolines, tested on both GABA responses from neurons and OLs, showed that many β-carbolines had a differential effect between these cell types, suggesting that they could be used to specifically enhance the GABAAR response expressed in OLs.
Materials and Methods
Animal Handling.
All experiments were performed by trained personnel and conducted in accordance with the Guide for the Care and Use of Laboratory Animals by the US National Institutes of Health. The animal protocols were approved by the Ethics Committees of the Instituto de Neurobiología at the Universidad Nacional Autónoma de México (UNAM) and the University of Basque Country (UPV/EHU). All possible efforts were made to comply with 3R standards and minimize animal suffering and the number of animals used.
RNA Sequencing Data Bases.
Three data sets, derived from public domain resources as described below, were used for transcriptomic analysis (Fig. 1, A and B). Bulk RNA from mouse brain OPCs expressing platelet-derived growth factor receptor alpha (PDGFRα+), a marker of OL progenitors (http://www.brainrnaseq.org; Zhang et al., 2014) were provided by Steven Sloan, who can be reached via the same website. Single cell data sets from three different regions in adult mice (P21–P90) were downloaded from the Gene Expression Omnibus (GSE75330; Marques et al., 2016) and the Allen Institute website (https://celltypes.brain-map.org/rnaseq; Tasic et al., 2016). From the first data set, a total of 81 nuclei from cortex (S1), inferred as OPCs based on the cluster analysis reported previously, and 125 nuclei from the corpus callosum were used. From the last data set, a single cell analysis of the visual cortex (V1) in adult mice (P53–P59), thus normalized RNA sequencing (RNA-Seq) from 30 NeuN− nuclei, grouped in the OPC-PDGFRα+ cluster, were used in the analysis. Detailed demographic characteristics, as well as technical white papers for data processing and quality control, can be downloaded from the same site. Confirmatory analyses of enrichment of OPC markers and lack of neuronal markers were performed for all data sets. For data analysis, gene expression level in each data set was transformed to fractional contribution (FC) to the total pool of mRNA available to produce GABAARs (Sequeira et al., 2019). Fractional contribution is defined as the percentage of expression level of each subunit gene over the sum of all 19 genes for GABAARs subunits in each mouse/cell. The means of the FC for each gene in each single cell data set (corpus callosum, S1 and V1 brain cortex) were averaged to build Fig. 1, A and B.
Diversity of GABAAR subunit expression in murine OPCs and OLs. (A) mRNA expression level for the 19 known GABAAR subunit genes analyzed in PDGFRα+ cells isolated by fluorescence-activated cell sorting from mice brains and assessed by RNA-Seq. The expression level of each GABAAR subunit is represented as its FC to the total pool of RNA available to produce GABAAR subunits. Data were obtained from http://www.brainrnaseq.org. (B) GABAAR subunit RNA expression analysis from PDGFRα+ cells from cortex (S1 and V1) and corpus callosum of adult mice. The expression level is represented as its FC to all RNA GABAAR subunits available like in (A). Data were obtained from Gene Expression Omnibus (GSE75330) and https://celltypes.brain-map.org/rnaseq. (C) The gel shows amplification by reverse transcription (RT) PCR of GABAAR subunit sequences expressed in purified OLs from rat P12 optic nerve. The expected product lengths are 266, 1311, 354, 231, and 404 and corresponded to α3, β2, β3, γ1, and γ3 subunits, respectively. bp, base pairs; MW, molecular weight marker; C(−), reaction negative control. (D) Analysis by immunocytochemistry of forebrain OPCs maintained in culture. Panels show images of the fluorescence signal for a specific antibody against the GABAAR subunit proteins (in green) as indicated in each panel, and nuclei labeling with 4′,6-diamidino-2-phenylindole dihydrochloride (blue). Bars = 50 μm.
Reverse Transcription Polymerase Chain Reaction, Cloning, and Site-Directed Mutagenesis of GABAAR Subunits from OLs.
To verify the expression of GABAAR subunits, cDNA was synthetized from total RNA extracted from fresh isolated OLs derived from optic nerves of P12 rats (see below) and used in polymerase chain reaction (PCR) assays. PCR products were amplified with the following primers: α3, forward: 5′-ACC ACA CAA ATG TGG CAC TTC-3′, reverse: 5′-AGT CAC TGC ATC TCC AAG CC-3′; β2, forward: 5′-ATG TGG AGA GTC CGG AAA AG-3′, reverse: 5′-TTT CAG TTG GGA GGC ACG TC -3′; β3, forward: 5′-CAC CCT GAT GGA ACA GTG CT-3′, reverse: 5′-ATG AGA GGA TTG TGA TCA TGA TTG-3′; γ1, forward: 5′-TAG GCG TGA GAC CCA CAG TA-3′, reverse: 5′-GCG ATT GGG CGT TGT TAT CC-3′; γ3, forward: 5′-ACC ATC AAT GCA GAG TGC CA-3′, and reverse: 5′-GCT GAG TGT GGT CAT GGT TA-3′. Glyceraldehyde-3-phosphate dehydrogenase amplification was used as a control with the oligonucleotide glyceraldehyde-3-phosphate dehydrogenase, forward, 5′-TCC CTC AAG ATT GTC AGC AA-3′, and reverse, 5′-AGA TCC ACA ACG GAT ACA TT-3′. The PCR conditions were an initial denaturation at 98°C for 30 seconds followed by 35 cycles at 98°C for 10 seconds, 55–60°C for 20 seconds, 72°C for 1 minute, and finally 72°C for 5 minutes using Phusion DNA Polymerase (Thermo Fisher Scientific, Waltham, MA). The complete coding sequences for α3, β2, β3, γ1, γ2, and γ3 subunits were amplified using the following primers: α3, forward: 5′-ATG ATA ACC ACA CAA ATG TGG C-3′, reverse: 5′-CTA CTG TTT GCG GAT CAT G-3′; β2, forward: 5′-ATG TGG AGA GTC CGG AAA-3′, reverse: 5′-TTA GTT CAC ATA GTA AAG CCA AT-3′; β3, forward: 5′-ATG TGG GGC TTT GCG GGA-3′, reverse: 5′-TCA GTT AAC ATA GTA CAG CCA GT-3′; γ1, forward: 5′-ATG GGT TCT GGG AAA GTC-3′, reverse: 5′-TTA TAA GTA TAG ATA TCC AAC CCA-3′; γ2, forward: 5′-ATG GCT GCA AAG CTG CTG-3′, and reverse: 5′-TTA AAG ATA TAG GTA TCC AAC CC-3′; γ3, forward: 5′-ATG GCT GCA AAG CTG CTG-3′, and reverse: 5′-TTA AAG ATA TAG GTA TCC AAC CC-3′. The PCR conditions were an initial denaturation at 98°C for 30 seconds followed by 35 cycles at 98°C for 10 seconds, 55–60°C for 20 seconds, 72°C for 70 seconds, and finally 72°C for 5 minutes. Each fragment obtained was cloned into pXENEX1 vector at the Ncol, BamHI, and NotI sites (e.g., Pérez-Samartín et al., 2017).
These plasmids were used as templates to make the site-directed mutagenesis of GABAAR subunits using the following primers: α3(H126R), forward: 5′-GGA CTC CAG ATA CCT TCT TCA GAA ACG G-3′, reverse: 5′-GTG AGC CAC TGA TTT TTT ACC GTT TCT GAA G-3′; α3(S294I), forward: 5′-GTT CTC ACC ATG ACC ACC TTG ATC ATC AG-3′, reverse: 5′-GGT AAA GAG TTT CTG GCA CTG ATG ATC AAG G-3′; β2(N264I), forward: 5′-CCT GAC GAT GAC CAC AAT CAT CAC CC-3′, reverse: 5′-GAG TCT CCC GGA GAT GGG TGA TGA TT-3′; γ1(S282I), forward: 5′-CGG TTT TGA CTA TGA CAA CCC TCA TCA CA-3′, reverse: 5′-GAA ACC TTA GGT AAA GAT TTT CTA GCG ATT GTG ATG AG-3′, and the Pfu polymerase (Thermo Fisher Scientific). Plasmids were linearized with the Hind III enzyme and used as templates for in vitro synthesis using the T7 mMESSAGE mMACHINE kit following the standard protocol (Ambion; Invitrogen, Grand Island, NY). cRNAs were used for heterologous expression in Xenopus laevis oocytes.
Heterologous Functional Expression of GABAAR Subunits in Xenopus Oocytes.
Ovarian follicles of X. laevis were dissected by surgery as described previously (Arellano and Miledi, 1993). Briefly, female X. laevis frogs were obtained from Xenopus I (Ann Arbor, MI). The ovarian lobules were surgically removed under sterile conditions from anesthetized frogs using 0.1% aminobenzoic acid ethyl ester (Sigma-Aldrich Co., St. Louis, MO) and rendered hypothermic. After surgery, frogs were sutured and allowed to recover from anesthesia. Frogs were maintained for 3–7 days in individual tanks until healing was complete and then housed in larger groups. No further oocytes were taken from them for at least 2 months. The follicles in stage VI (Dumont, 1972) were dissected from the lobules and maintained in normal Barth’s solution (containing in millimolar: 88 NaCl, 1 KCl, 2.4 NaHCO3, 0.33 Ca(NO3)2, 0.41 CaCl2, 0.82 MgSO4, 5 HEPES, adjusted to pH 7.4, and supplemented with 70 μg/ml gentamicin). Healthy isolated follicles were then injected with 50 nL of solution containing a mix of 2 or 3 (i.e., α3β2, α3β3, α3β2γ1, α3β2γ3, α3β3γ1, α3β3γ3, and α3β2γ2; in proportion 1:1 or 1:1:1) of the cloned GABAAR subunit cRNA sequences (5 ng per follicle) or with 50 nL H2O as a control, and cells were incubated at 18°C in normal Barth’s solution. Then, 24 hours postinjection, follicles were treated with 0.5 mg/ml collagenase type 1 A (Sigma-Aldrich Co.) in normal Ringer’s (NR) solution (containing, in millimolar, 115 NaCl, 5 KCl, 1.8 CaCl2, and 5 HEPES; pH 7.0) for 30 minutes at room temperature. Cells were then washed several times with NR solution, and the oocytes were defolliculated using fine forceps under a stereomicroscope. Oocytes were incubated at 18°C in Barth’s solution and monitored for their response to GABA and other drugs 3–6 days after cRNA injection. The electrical recording was made using the two-electrode voltage-clamp technique (Arellano and Miledi, 1993). For this, borosilicate glass micropipettes (filled with 3 M KCl) with a resistance of 1–2 MΩ were used. Currents were monitored using an Axon GenClamp 500B (Molecular Devices, San José, CA) amplifier, and signals were digitized and stored for further analysis using an analog-to-digital converter (Axon DigiData 1200; Molecular Devices) and specialized software (pClamp version 9; Molecular Devices). For electrical recordings, the oocytes were continuously superfused (10 ml/min) at room temperature with NR solution (unless otherwise stated), voltage-clamped at −60 mV and applied periodic voltage steps to −40 mV (1 second) every 40 seconds during the recording to monitor membrane conductance. All experimental groups were replicated using oocytes from at least three different frogs.
GABA and GABAAR allosteric modulators were applied through the superfusion solution to test their effects on oocytes expressing the different GABAAR subunit combinations. Dose-response curves were fitted to the following equation:
by the method of nonlinear least-squares fitting, where EC50 is the half-maximal effective concentration for GABA, nH is the slope factor (Hill coefficient), A1 and A2 are the initial and final normalized current (I) values, respectively, and [GABA] is the concentration of the neurotransmitter. For the inhibitory effect generated by Zn2+ administration was calculated the half-maximal inhibitory concentration (IC50) fitting dose-response (D-R) curves for Zn2+ with the same equation and method.
Forebrain OPCs, Optic Nerve OLs, and Neuronal Cell Culture.
Primary mixed glial cultures were prepared from newborn (P0–P2, of either sex) Sprague-Dawley rats as previously described (Arellano et al., 2016; Sánchez-Gómez et al., 2018). Briefly, forebrains were removed from the skulls, and cortices were isolated and digested by incubation (15 minutes, 37°C) in Hank’s balanced salt solution containing 2.5% trypsin and 0.4% DNAse. The cells were dissociated by passage through needles (21G and 23G), centrifuged, and resuspended in Iscove’s modified Dulbecco’s medium supplemented with 10% FBS (Hyclone; Thermo Fisher Scientific) and antibiotic/antimycotic solution (Sigma-Aldrich Co.). Cells were seeded into poly-d-lysine–coated (10 μg/ml) 75 cm2 flasks and maintained in culture at 37°C and 5% CO2. After 10–15 days in culture, the flasks were shaken (400 rpm, 2 hours, 37°C) to remove loosely adherent microglia. The remaining OPCs on top of the confluent monolayer of astrocytes were dislodged by shaking overnight at 400 rpm. The cell suspension was then filtered through a 10 μm nylon mesh and plated on bacterial grade Petri dishes for 2 hours. The nonadherent OPCs that remained in suspension were recovered and plated again on bacterial grade Petri dishes for 1 hour. The resulting enriched forebrain OPC cell suspension was centrifuged and resuspended in a chemically defined medium (OPC medium) consisting of Dulbecco’s modified Eagle’s medium supplemented with 100 μg/ml transferrin, 60 ng/ml progesterone, 40 ng/ml sodium selenite, 5 μg/ml insulin, 16 μg/ml putrescine, and 100 μg/ml bovine serum albumin. Cells were plated onto poly-d-lysine–coated 12-mm-diameter coverslips in 24-well culture dishes at a density of 5 × 103 cells per well and cultured in the presence of the mitogenic factors platelet-derived growth factor-AA (5 ng/ml) and basic fibroblast growth factor (5 ng/ml) to expand their number and prevent their differentiation (here called proliferative medium). The purity of oligodendroglial cultures was confirmed by immunostaining with the antibodies against the oligodendroglial markers PDGFRα, Olig2, and NG2 that labeled more than 80% of cells, indicating that these cultures contained mostly OPCs at this stage.
In addition to cortical OPC purification, primary cultures of OLs derived from optic nerves of 12-day-old Sprague Dawley rats of either sex were obtained as described previously (Arellano et al., 2016). Cells were seeded on 24-well plates bearing 12-mm-diameter coverslips coated with poly-d-lysine (10 μg/ml) at a density of 104 cells per well. Cells were maintained at 37°C and 5% CO2 in the differentiation medium. After 1 to 2 days in vitro, at least 97% of the cells were positive for the O4 antigen.
Primary cultures of cortical neurons were established according to a modified procedure (Larm et al., 1996). Briefly, cortical lobes of Wistar rat embryos (E18, of either sex) were isolated and digested in Hank’s balanced salt solution (Gibco; Thermo Scientific) containing 0.25% trypsin and 0.4% DNAse. After, the tissues were dissociated with needles (21G, 23G, 25G) and centrifuged. The neurons were resuspended in B27 neurobasal medium (Gibco Thermo Scientific) supplemented with 10% FBS (Hyclone; Thermo Fisher Scientific) and seeded at 105 cells per well onto poly-l-ornithine–coated (30 μg/ml) 24-well plates bearing coverslips with a 12-mm diameter. The medium was replaced by serum-free B27-supplemented neurobasal medium (with antibiotic-antimycotic and 2 mM glutamine) 24 hours later. Cultures were maintained at 37°C and 5% CO2. Rats were provided by the vivarium facilities either at the Instituto de Neurobiología at the Universidad Nacional Autónoma de México (UNAM) or at the University of Basque Country (UPV/EHU).
Whole-Cell Patch-Clamp.
For the recording of membrane currents, cells [OPCs 1–4 days in vitro (DIV), OLs 1 DIV, or neurons 12–14 DIV] plated on coverslips were transferred to a recording chamber attached to a conventional inverted microscope (IX71; Olympus, Tokyo, Japan). They were continuously superfused and maintained at room temperature (22–25°C). The standard external solution (Attali et al., 1997; Pérez-Samartín et al., 2017) contained (in millimolar) 140 NaCl, 5.4 KCl, 2 CaCl2, 1 MgCl2, and 10 (4)-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES), and it was adjusted to pH 7.3 with NaOH. Electrophysiological recording was performed using either a MultiClamp 700B or an Axopatch 200B amplifier (Molecular Devices) to establish the standard whole-cell configuration. Patch-clamp pipettes (3–5 MΩ) were filled with internal solution (Pérez-Samartín et al., 2017) containing (in millimolar) 140 KCl, 2 CaCl2, 2 MgCl2, 10 HEPES, 11 EGTA, 2 Na-ATP, and 0.2 GTP, adjusted to pH 7.3 with KOH. Currents were recorded at a holding membrane potential of −80 mV, digitized with an analog-to-digital converter (DigiData 1440; Molecular Devices), and visualized and analyzed with specialized software (pClamp v10; Molecular Devices). In most cases, peak GABA current responses generated at −80 mV by drug superfusion were used for the analysis. Agonists and other drugs were added to the external solution from stock solutions, when necessary (e.g., for β-carboline application). DMSO was no more than 0.1% in the final solution.
Immunocytochemistry.
To characterize primary cultures of OPCs and OLs, cells were immunoassayed with antibodies against oligodendroglial cell-specific markers: mouse anti-O4 (10 μg/ml; ref. MAB345; Millipore, Burlington, MA), mouse anti-PDGFRα (C-9) (1:200; ref. sc-398206; Santa Cruz Biotechnology, Dallas, TX), rabbit anti-NG2 (1:300; ref. AB5320; Millipore), and rabbit anti-Olig2 (1:500; ref. AB9610; Millipore); to detect myelin expression, the mouse anti–myelin basic protein (MBP) was used (1:1000; ref. SMI-99; Covance, Princeton, NJ); astrocytes were identified by mouse anti–glial fibrillary acidic protein (2E1) (1:200; ref. sc-33673; Santa Cruz Biotechnology). For α2, α3, γ1, and γ2 subunit immunostaining, cells were treated with the corresponding antibody (anti-GABAAα2 1:100, ref. AGA002; anti-GABAAα3 1:100, ref. AGA003; and anti-GABAAγ1 1:100, ref. AGA016; anti-GABAAγ2 1:100, ref. AGA005; from Alomone Laboratories, Jerusalem, Israel). In all cases, cells were fixed in 4% paraformaldehyde in PBS for 30 minutes at room temperature. The fixed cultures were permeabilized with 0.1% Tween-20, blocked with 5% goat serum in PBS for 30 minutes, and incubated overnight at 4°C with the antibodies diluted in PBS containing 5% goat serum and 0.1% Tween-20. Then, cells were rinsed and incubated for 2 hours at room temperature with 1:200 goat anti-mouse IgG (H + L) conjugated with fluorescein isothiocyanate (ref. 81-6511; Invitrogen) or 1:200 goat anti-rabbit IgG (H + L) conjugated with Alexa Fluor 514 (ref. A-31558; Thermo Fisher Scientific) according to the host species of the primary antibodies. After five washes with PBS, samples were stained with 4 μg/ml 4′,6-diamidino-2-phenylindole dihydrochloride (Molecular Probes, Eugene, OR) to identify cell nuclei. In all cases, the absence of nonspecific interactions of secondary antibodies was corroborated by omitting the primary antibodies. Finally, the samples were mounted on VectaShield (Vector Laboratories, Burlingame, CA), and the preparations were visualized under an LSM510 laser scanning confocal microscope or under an Apotome optical sectioning fluorescence microscope (both from Zeiss, Oberkochen, Germany).
GABAAR Subunit Silencing by siRNA Transfection.
To reduce the expression of either α3 or γ1 subunit in cultured OPCs, the interference technique using siRNAs was used. These siRNAs were commercially designed (Dharmacon Inc., Lafayette, CO) for α3 (siGENOME Rat Gabra3 [24947] siRNA-SMARTpool) and γ1 (siGENOME Rat Gabrg1 [140674] siRNA-SMARTpool) using the target sequences NM_017069.3 and NM_080586.1, respectively. OPCs were seeded at a density of 4 × 104 cells per well in 24-well plates and allowed to attach overnight. Then, 1 DIV cells were transfected with 270 ng per well of α3-siRNA or γ1-siRNA using Lipofectamine 3000 (Invitrogen) following the manufacturer’s instructions. For the control group (Control), cells were transfected in the same condition with the ON-TARGETplus Non-targeting Pool (SO-2686908G; Dharmacon Inc.). After 48–72 hours of transfection, siRNA-treated OPCs maintained in culture were used for immunocytochemistry using the methods described above or for electrophysiological monitoring. The arbitrary units of fluorescence (AUF) were estimated as optical density using ImageJ software (version 1.52i). Briefly, applying blind analysis the intensity of green was obtained from five regions of interest (ROIs) in every OPC preparation. AUF was calculated by normalizing the intensity values from each ROI against the background intensity value from each preparation, applying the following relationship: (intensity of the background – mean intensity of green measured in the ROIs)/intensity of the background (Cisneros-Mejorado et al., 2019). Then, for α3-siRNA and γ1-siRNA transfected OPCs and their corresponding control groups, the AUFs were normalized against the respective untreated groups. The GABA response elicited in transfected OPCs with either α3-siRNA or γ1-siRNA were compared with those of the corresponding control groups; OPCs were identified in the culture by their typical bipolar morphology and their rectifying current-voltage relationship, as well as the basic characteristics of the GABA response such as its low sensitivity to GABA, inhibition by Zn2+, and/or enhancement by β-CCB (see Supplemental Fig. 2 and 3).
Substances.
7-Chloro-1-methyl-5-phenyl-3H-1,4-benzodiazepin-2(1H)-one [diazepam (DZP)]; 4-ethyl-6,7-dimethoxy-9H-pyrido[3,4-b]indole-3-carboxylic acid methyl ester (DMCM); β-CCB; β-carboline-3-carboxylic acid N-methylamide; ethyl 9H-pyrido[3,4-b]indole-3-carboxylate (β-CCE); 3-(hydroxymethyl)-β-carboline (β-CHM); tert-butyl β-carboline-3-carboxylate; N-methyl-N-[3-[3-[2-thienylcarbonyl]pyrazolo[1,5-a]pyrimidin-7-yl]phenyl]acetamide (indiplon); and 8-fluoro-5,6-dihydro-5-methyl-6-oxo-4H-imidazo[1,5-a][1,4]benzodiazepine-3-carboxylic acid, ethyl ester [flumazenil (FMZ)] were all obtained from Tocris Bioscience (Bristol, United Kingdom). Salts, DMSO, GABA, and ATP were from Sigma-Aldrich Co.
Statistical Analysis.
All data were analyzed using p-CLAMP (version 10.6; Molecular Devices), R (version 3.2.3) with RStudio (Free Software Foundation Inc., Boston, MA), and GraphPad Prism (version 6.00; La Jolla, CA). Results are presented as means ± S.D., regularly; the sample size and the number of experiments were adjusted after initial data collection, and the final figures are indicated in each figure or the text. The statistical significance of differences between two data sets was tested with an unpaired two-tailed Student’s t test. For multiple comparisons, a one-way ANOVA (Tukey’s post hoc test) was used. The study was exploratory in nature and did not test a prespecified null hypothesis. The outcomes of the statistical tests cannot be interpreted as hypothesis testing but only as descriptive; thus, the indicated P values are therefore descriptive.
Results
Cloning and Expression of GABAAR Subunits in OLs.
To explore probable stoichiometries of GABAARs expressed in OPCs, we examined the expression of all 19 genes for GABAAR subunits in cells isolated by fluorescence-activated cell sorting, measured in bulk (bulk RNA-Seq) or individually (single cell RNA-Seq), from the cortex and corpus callosum, and deposited in public repositories (Zhang et al., 2014; Marques et al., 2016; Tasic et al., 2016). Only cells that were positive to PDGFRα+, or clustered within the PDGFRα+ category, were used for this analysis. The FC (Sequeira et al., 2019) of each subunit to the total pool of mRNAs available to conform GABAARs was analyzed in each data set; results are illustrated in Figs. 1, A and B. In bulk RNA-Seq, α3 was the main expressed subunit, followed by β3 and β2, whereas γ1 was the main γ subunit (Fig. 1A). In single cell RNA-Seq (Fig. 1B), a similar result was obtained. Although α subunits did not present a clear preponderance, α3 was barely the major subunit, whereas β2, β3, and γ1 were well represented in the data sets. In both cases, γ2, γ3, β1, and δ, as well as α5 and α6 subunits, were expressed at low or below the limit of detection levels. These results, together with the analysis reported previously (Larson et al., 2016), are in agreement with previous functional studies (Arellano et al., 2016) suggesting that GABAARs in OLs from neonate rats comprise a combination of α3/β2-β3/γ1-γ3 subunits. Based on this information, we amplified, cloned, and sequenced the five subunits from fresh OLs isolated from the rat optic nerve (P12) (Fig. 1C). The sequences obtained indicate that all subunits were identical to those expressed in other cell types including neurons. Thus, the α3 subunit in OL was identical to the sequence NP_058765.3 (Malherbe et al., 1990). The β2 and β3 sequences were identical to NP_037089.1 (Ymer et al., 1989) and NP_058761.1 (Ymer et al., 1989), respectively, whereas γ1 and γ3 corresponded to sequences NP_542153.1 (Ymer et al., 1990) and NP_077346.3 (Knoflach et al., 1991), respectively. Protein expression was confirmed by immunodetection in OPCs isolated from neonate forebrain for α3, α2, and γ1 subunits, but γ2 expression was not detected in the same preparations (Fig. 1D); as expected, the staining pattern for the analyzed subunits was different in cortical neuronal cells maintained in vitro (Supplemental Fig. 1).
The cRNA for each subunit was transcribed and injected into oocytes to express distinct receptor combinations as follows: α3β2, α3β3, α3β2γ1, α3β2γ3, α3β3γ1, and α3β3γ3. The functional characteristics of the receptors expressed were studied electrophysiologically to compare their behavior with that described for the endogenous GABAAR expressed in OL. First, D-R curves for GABA were built for each receptor within the concentration range of 100 nM to 1 mM (Fig. 2). All the subunit combinations expressed functional receptors, and the EC50 values estimated from the best fit for each normalized data set were as follows: α3β2, 19 ± 4.18 μM; α3β3, 10 ± 2.26 μM; α3β2γ1, 53 ± 4 μM; α3β2γ3, 21 ± 3.81 μM; α3β3γ1, 58 ± 2.26 μM; and α3β3γ3, 16 ± 2.22 μM (Fig. 2A). In all cases, the Hill coefficient (nH) was within the range of 0.82–1.33. Two main differences were observed: 1) the maximal current amplitude was obtained with the combination α3β2γ1 that reached an average peak current of 9.37 ± 5.97 μA (n = 19 oocytes), whereas combinations α3β2γ3 and α3β3γ1 had peak currents of 2.36 ± 0.88 μA (n = 13 oocytes) and 4.43 ± 2.07 μA (n = 9 oocytes), respectively (Fig. 2B); and 2) the highest EC50 values were obtained for receptors containing the γ1 subunit. These values were statistically different (P < 0.05) from those observed in receptors without the γ subunit or with γ3 (Fig. 2C).
Heterologous functional expression of oligodendroglial GABAAR subunits cloned from rat optic nerve. (A) The cRNA for GABAAR α3, β2, β3, γ1, and γ3 subunits was synthesized and injected in different combinations into Xenopus oocytes as indicated. The expressed GABA responses were monitored electrophysiologically for each combination. In this and subsequent records, GABA was applied during the times indicated by bars at the top, and oocytes were held at −60 mV. D-R curves were constructed and plotted as illustrated; each point represents the average peak response (mean ± S.D.) normalized with respect to maximal current response elicited by 1 mM GABA monitored in 10–25 oocytes from three to five frogs. (B) Comparison of the peak current (mean ± S.D.) by 1 mM GABA generated in each of the combinations studied. Each bar includes 9–19 oocytes from five donors. α3β2γ1 mean amplitude response was different to all others. (C) The graph built from data in (A) shows a comparison of the combinations containing αβγ subunits signaling the decrease in GABA sensitivity due to the presence of the γ1 subunit. The EC50 differences were statistically significant between the receptors containing γ1 vs. those containing γ3. *P < 0.001 for comparisons across the mean of −logEC50 values from each subunit combination, one-way ANOVA followed by a Tukey’s post hoc test.
Sensitivity to Zn2+.
It has consistently been shown that GABA responses in OLs are inhibited by Zn2+ (100 μM; Williamson et al., 1998; Passlick et al., 2013; Arellano et al., 2016). This observation suggests the absence of the γ2 subunit, as it confers a relative insensitivity to the cation (Hosie et al., 2003; Trudell et al., 2008; Karim et al., 2013), whereas the absence of any γ subunit renders GABAARs highly sensitive to Zn2+. However, it is not known whether γ1 or γ3 confers insensitivity as well as γ2 does, especially in combination with the α3 subunit. To achieve this information, we evaluated the effect of Zn2+ on the response elicited by GABA in oocytes expressing each of the combinations studied. Zn2+ and GABA were coapplied at 100 μM both, because this protocol inhibits more than 80% of the GABA response in receptors without γ subunits but has a weak effect on receptors with the γ2 subunit (e.g., Hosie et al., 2003; Karim et al., 2013). Figure 3 illustrates that the effect of Zn2+ on the peak of GABA response depended on the subunit combination expressed. Receptors without a γ subunit had the highest sensitivity to Zn2+ (67%–85% of inhibition), receptors with γ3 presented lower sensitivity (20%–32% of inhibition) and receptors with γ1 presented medium sensitivity to Zn2+ (within an inhibition range of 40%–50%). For comparison purposes, the current response of the receptor α3β2 containing the γ2 subunit was inhibited around 10% in similar experiments. Thus, accordingly with several other studies where the participation of γ subunits greatly reduced the sensitivity to Zn2+, this effect depended on the γ subunit type coexpressed in receptors containing α3β2 or α3β3 subunits. Among the different combinations tested in this condition, those that contained the γ1 subunit presented a medium sensitivity to Zn2+. A more detailed analysis was made to assess the potency of Zn2+ on receptors expressing this subunit.
Effect of Zn2+ on GABAARs expressed in Xenopus oocytes from oligodendroglial sequences. (A) The traces illustrate the responses to 100 μM GABA in the absence (gray trace in each case) and presence of 100 μM Zn2+ (black trace) for each receptor studied as indicated. (B) The percentage of GABA response achieved in the presence of Zn2+ in each condition was normalized against the response generated without Zn2+, then averaged and plotted in the bar graph (mean ± S.D.; 7–10 oocytes from three to five frogs). γ3 subunit conferred resistance to inhibition by Zn2+ similar to γ2 subunit, whereas presence of γ1 allowed inhibition of the GABA-response close to 50% and was statistically different to those receptors without γ subunit or containing either γ2 or γ3. One-way ANOVA (P < 0.001) followed by a post hoc Tukey test showed that all values were statistically different from each other.
The sensitivity to Zn2+ of the endogenous receptor in OPCs and OLs is about 10 μM testing the GABA response around its EC10–EC30 (Arellano et al., 2016). To compare the endogenous sensitivity with that displayed by heterologously expressed receptors, especially those containing the γ1 subunit, oocytes were injected with mRNA for α3β2γ1 or α3β3γ1 subunits, and these were compared with the receptor α3β2. Responses for each combination were generated applying 10 μM GABA alone (control condition) or together with distinct Zn2+ concentrations within the range of 0.1–1000 μM (traces in Fig. 4A). Inhibition D-R curves were built for each receptor (Fig. 4B), normalizing the peak current response in the presence of Zn2+ against the maximal control response without Zn2+. The half-maximal inhibitory concentration (IC50) values obtained were of 6.2 ± 2.68, 25.7 ± 2.68, and 69 ± 3.13 μM (five to seven oocytes for each case) for the receptors α3β2, α3β2γ1, and α3β3γ1, respectively; IC50 values were statistically different between all of them. This analysis shows that the receptor α3β2γ1 had an IC50 similar to that displayed by the endogenous receptor (Arellano et al., 2016), whereas the α3β3γ1 receptor was less sensitive, as well as the α3β2 receptors in combination with either γ2 or γ3 subunit (Fig. 4C), which had an IC50 for Zn2+ of 82.7 ± 2.69 and 75.5 ± 2.91 μM, respectively.
D-R to Zn2+ of GABAARs expressed in Xenopus oocytes from oligodendroglial sequences. (A) Top traces illustrate responses evoked by 10 μM GABA in the presence of different Zn2+ concentrations as indicated in each case for the α3β2, α3β2γ1, and α3β3γ1 receptors expressed heterologously. (B, C) The response amplitude from recordings obtained in (A) were normalized with respect to the control GABA response, without Zn2+, in the same oocytes, and the inhibition D-R curves were plotted to obtain the IC50 in each receptor. Data points represent the average (±S.D.) of five to seven oocytes (from three frogs) in each condition. Comparisons were made between the mean of the −logIC50 values of each combination of subunits. The one-way ANOVA (P < 0.001) followed by a post hoc Tukey test showed that all values were statistically different from each other.
Effect of Allosteric Modulators on the α3β2γ1 GABAAR.
A battery of allosteric modulators was analyzed on the α3β2γ1 receptor expressed in Xenopus oocytes, considering that this combination presented closer parameters of sensitivity to GABA and Zn2+ than those exhibited by receptors endogenously expressed in OLs; this combination is also in accordance with the transcriptomic analysis reported. The positive modulators DZP, indiplon, and β-CCB were analyzed by applying the respective EC10 for GABA, as a control response, and the same concentration of GABA in coapplication with each drug. Each modulator was applied alone for 40 seconds prior to coapplication with GABA (Fig. 5A). Both DZP and β-CCB enhanced the response to GABA by 198.0% ± 37.2% (11 oocytes) and 236.4% ± 40% (11 oocytes), respectively, whereas indiplon had a weak or null effect on the response (110.5% ± 3.65%, 10 oocytes) (Fig. 5B). In general, the effect of these modulators on the receptor α3β2γ1 were like those displayed by the endogenous GABAAR expressed in OLs. The interaction of β-carboline with α3β2γ1 was analyzed in greater detail (Fig. 5C). The positive effect of β-CCB (30 μM) was not blocked by FMZ (10 μM), a classic BZD site antagonist; in fact, application of FMZ together with β-CCB caused a greater response of the α3β2γ1 receptor, as illustrated in Fig. 5C. Thus, the enhancement caused by β-CCB without FMZ was of 324.1% ± 74%, whereas in the presence of FMZ was of 446.9% ± 65% (17–20 oocytes in each case). This result suggested that the β-CCB binding site did not correspond to the classic BZD binding site, an observation that was supported by similar coapplication experiments with β-CCB and GABA at α3β2γ3 and especially α3β2 receptors in which the β-carboline also caused an important increase of the response (Fig. 5D).
Pharmacological effect of allosteric modulators on the GABA response of α3β2γ1 receptors expressed in oocytes. (A) Traces illustrate the modulatory effect of drugs (10 μM; DZP, β-CCB, or indiplon) coapplied with GABA (10 μM) in each case; control response with GABA alone is shown in black. The mean effect response observed in 11 oocytes (five to six frogs) was plotted in (B) including the effect of Zn2+ in the same group of oocytes for comparison. (C) Traces illustrate the effect of 10 μM FMZ, a specific antagonist for the high affinity BZP binding site. It shows the response to 10 μM GABA alone, followed by the effect of 30 μM β-CCB in the absence and presence of FMZ that did not antagonize the enhancement caused by β-CCB. The mean effect response obtained in seven oocytes is shown in the bar graph. The increase in GABA response by β-CCB plus FMZ was statistically significant against the group without FMZ (*P < 0.001, paired two-tailed Student’s t test). (D) Effect of 3 μM β-CCB on the GABA response activated in α3β2 receptors without γ subunit, or those containing either γ1 or γ3. Mean responses (6–10 oocytes from three frogs) were statistically different for the three receptors compared with the control response elicited by GABA alone, and β-CCB enhancement in α3β2γ1 response was different to all others. *P < 0.001 one-way ANOVA followed by a Tukey’s post hoc test.
To confirm this result, the specific binding sites of high and low affinity to BZD were mutated on the α3, β2, and/or γ1 subunits. Thus, the α3β2γ1 receptor sensitivity to DZP was eliminated or strongly decreased similarly to the manner previously shown for the neuronal receptor (Benson et al., 1998; Walters et al., 2000; Middendorp et al., 2015). The results of these experiments are illustrated in Fig. 6. First, the H126R mutation in the α3 subunit eliminated the sensitivity to 1 μM DZP (Benson et al., 1998). However, positive modulation caused by β-CCB (1 μM) on the response to GABA (10 μM) was not affected, being 144.8% ± 18% in the oocytes expressing the native receptor and 202.4% ± 17.4% in the mutated receptor α3(H126R)β2γ1 (five to six oocytes from two frogs; Fig. 6A). Also, mutations α3(S294I)β2(N264I)γ1(S282I) were made to eliminate one of the low-affinity BZD sites located in the transmembrane region (Walters et al., 2000), and this receptor was expressed to explore its sensitivity to 60 μM DZP and 1 μM β-CCB. The results illustrated in Fig. 6B indicate that the potentiation of the GABA (10 μM) response produced by coapplication of 60 μM DZP was greatly reduced in the mutant receptor α3(S294I)β2(N264I)γ1(S282I) compared with the α3β2γ1 receptor response. The values were 281.1 ± 25% for the latter and 172.3% ± 10% for the mutant receptor, where persistent potentiation in this case was more likely due to DZP interaction on the high-affinity site. However, β-CCB application to the α3(S294I)β2(N264I)γ1(S282I) receptor maintained a positive modulatory effect and enhanced the response by 251.6% ± 10.5% versus the control native receptor that presented an increase of 160.1% ± 8% (10 oocytes). These data support the idea that β-CCB did not act on the classic BZD site of the GABAA α3β2γ1 receptor, but also did not act on the described transmembrane low-affinity site for BZD.
The α3β2γ1 receptor BZP binding sites and the enhancement caused by β-CCB. (A) 1 µM DZP or 3 µM β-CCB were coapplied with 10 µM GABA to analyze its effect on the high affinity site for BZP. Traces illustrate the responses obtained in oocytes that were injected with the cRNA of the native α3β2γ1 subunits (left column), and in oocytes injected with the α3(H126R)β2γ1 mutant (right column). The average responses obtained in five to six oocytes are shown in the bar graph indicating that DZP enhancement was eliminated, whereas the β-CCB enhancement was maintained in the mutated combination (line at 100% indicates control response to GABA). (B) Similar experiments in (A) were performed applying 60 µM DZP or 3 µM β-CCB coapplied with 10 µM GABA to analyze its effect on one of the low-affinity BZP sites. In the left column, the traces illustrate the responses obtained in oocytes expressing the native α3β2γ1 subunits, whereas the second group expresses the mutant α3(S294I)β2(N264I)γ1(S282I). The averages of the responses obtained in 10 oocytes (three frogs) are shown in the bar graph indicating that DZP enhancement was strongly reduced, whereas the increase generated by β-CCB was maintained in the mutated combination. All drug effects were statistically significant against the control group, with the clear exception of the mutant tested with DZP in (A) (P < 0.001 for comparisons between the control vs. drug-treated amplitude values, unpaired two-tailed Student’s t test).
Effect of β-Carbolines on the Endogenous Response to GABA in OLs and Neurons.
The effect of β-CCB acting on a different binding site to that of high-affinity for BZP on the α3β2γ1 receptor is similar to the effect observed for the endogenous oligodendroglial GABAAR. Figure 7 illustrates that the endogenous response to 10 μM GABA in cultured OLs isolated from optic nerve was enhanced by β-CCB (10 μM) but not affected by FMZ (10 μM) (n = 7). These results suggest that β-CCB did not act through the interaction with the high-affinity site for BZP in the endogenous GABAAR in OLs. In contrast, FMZ did block the potentiation of the response to GABA mediated by DZP (10 μM) in either OLs (n = 7) or cortical neurons (n = 5) (Fig. 7A). Nevertheless, in this case FMZ did not cause an extra enhancement of the GABA response when tested in conjunction with β-CCB.
β-CCB modulatory effect on the GABA response elicited in the oligodendroglial lineage. (A) GABA current response was monitored electrophysiologically in OLs (from optic nerve) or cortical neurons held at −80 mV. The enhancement mediated by either 10 μM DZP (orange bar) or 10 μM β-CCB (blue bar), in the presence and absence of 10 μM FMZ, was analyzed as summarized in the bar graph. The first black bar in each data set shows the control GABA response in the absence of drugs; the second bar of each group indicates the average of the response in the presence of a modulator; the gray bar indicates the modulator effect plus FMZ. The washing effect is in black. FMZ did not antagonize the response to β-CCB in OLs, but it was strongly effective on the potentiating DZP effect in both OLs and cortical neurons (10 μM GABA was used for OLs and 3 μM GABA for neurons). (B) D-R curves for GABA in OLs in the absence (black) and presence of 10 μM β-CCB (blue). Each data point represents the average response to GABA in eight OLs in both conditions. Data were adjusted to a D-R curve. Traces to the right illustrate the GABA current response in the presence of 10 μM β-CCB; in this and subsequent records in neural cells, drugs were applied as indicated by the top bars and cells were held at −80 mV.
Also, it was observed that in the endogenous response, β-CCB enhancement was mainly due to an increase in the sensitivity to GABA, shown in Fig. 7B, where D-R curves for GABA were built in both the absence and presence of β-CCB (10 μM). Thus, in the absence of the modulator, the EC50 was of 83.3 ± 8.6 μM (n = 5), whereas in the presence of β-CCB, the EC50 was of 18.5 ± 3.4 μM (n = 6), the EC50 differences were statistically significant (*P < 0.001 for comparisons across the mean of −logEC50 values from each group; one-way ANOVA followed by a Tukey’s post hoc test). The last EC50 value was also similar to that observed in the presence of β-CCB for the α3β2γ1 receptor heterologously expressed.
A particularly important aspect of β-carboline action is the probable differential effect on the oligodendroglial receptor, compared with the effect on receptors expressed in neurons. To obtain more information about this effect, the potency for different β-carbolines on GABA responses in both OLs and cortical neurons maintained in culture was explored. It has been shown that β-CCB distinguishes between the receptors expressed in these two cell types (Cisneros-Mejorado et al., 2019). Here we confirmed this differential effect analyzing the effect of a battery of distinct β-carbolines on the endogenous GABA response of OLs and neurons. Figure 8 illustrates that the endogenous response to GABA in OLs or neurons (EC10 in each case) was monitored in the absence (control response) or presence of one of distinct β-carbolines. The normalized response with respect to the control is indicated as an average of the effect in the tested cells. The β-carboline (10 μM) positive modulatory effect on the response to GABA (10 μM) in OLs presented the following power sequence: β-CCB > β-carboline-3-carboxylic acid N-methylamide > β-CCE > tert-butyl β-carboline-3-carboxylate > β-CHM (Fig. 8A), whereas DMCM inhibited the response. In contrast, most of the β-carbolines (same concentration) presented either a net inhibitory effect on the neuronal receptor activated by 3 μM GABA or did not have a clear effect on the amplitude of the response, as observed for β-CCB or β-CCE (Fig. 8B). Thus, it appears that distinct β-carbolines presented a differential effect on the oligodendroglial GABAAR. For β-CCB, this effect was not due to its interaction with the high-affinity binding site for BZP.
β-carbolines modulatory effect on GABA current responses from either OLs or cortical neurons. (A) The response to 10 μM GABA in the absence and presence of a β-carboline (10 μM) was tested in OLs from the optic nerve. Each drug indicated from one to six corresponds to the structures in the lower part of the Figure, and traces illustrate the current response obtained for three drugs tested as indicated. Bars are the average of the normalized response with respect to the control response without β-carboline (100% line in graph) for each of the cells registered (6–21 cells in each case). The traces in gray correspond to control current responses, whereas black traces are the GABA responses in the presence of a β-carboline. (B) For comparison, a similar analysis was performed for the GABA response elicited in cortical neurons (six to eight cells in each case) by applying 3 μM GABA and the same set of β-carbolines. Traces illustrate responses obtained for the three drugs indicated, in gray correspond to control current responses, whereas green traces are the GABA responses in the presence of a β-carboline. All drug effects were statistically significant against the control group with exception of β-CHM in OLs, as well as β-CCB and β-CCE in neurons (P < 0.001 for comparisons between the control cells vs. β-carboline treated cells, unpaired two-tailed Student’s t test). β-CCA, β-carboline-3-carboxylic acid N-methylamide; β-CCt, tert-butyl β-carboline-3-carboxylate.
Specific Silencing of Endogenous α3 or γ1 GABAAR Subunits in OLs.
Since α3 and γ1 subunits appear to define important characteristics of the endogenous oligodendroglial receptor, we performed the following experiments to directly determine their participation in the receptor conformation. For this, transfections with specific siRNA were performed for each subunit in OPCs isolated from neonate forebrain. Before performing the GABA response analysis in the transfected cells, OPCs were monitored for at least 4 days in vitro to evaluate essential characteristics such as marker expression (NG2, PDGFRα, MBP) and the electrophysiological and pharmacological GABA response profile. This was necessary because the maximal silencing effect of the respective protein had a 48-hour temporality. The result of this analysis indicated that most (85%–90%) cells maintained in culture from 2 DIV to 4 DIV in proliferating medium showed characteristics corresponding to OPCs expressing Olig2, NG2, and PDGFRα, but not glial fibrillary acidic protein or MBP, whereas at 4 DIV a significative percentage (21.6% ± 4.85%) of cells expressed O4 and presented a more complex morphology (Supplemental Fig. 2). Importantly, the GABAAR functional and pharmacological profile expressed did not change during the culture time from 2 DIV to 4 DIV. The current-voltage electrophysiological profile did not show statistically significant differences either (Supplemental Fig. 3), especially with regard to the typical rectifying behavior, due to the lack of Kir-type current activated at hyperpolarizing potentials (Pérez-Samartín et al., 2017). Thus, OPCs at 2 DIV were transfected with specific siRNA against either α3 (Fig. 9) or γ1 (Fig. 10) subunits, and the effect of this manipulation was compared with respect to cells that were transfected with a nonspecific siRNA (control), using the expression in untreated cells as reference. As illustrated in Fig. 9A, α3-siRNA transfection strongly reduced (93.7% ± 2%) the expression of the respective subunit as evidenced by the corresponding immunodetection, whereas transfection with nonspecific siRNA did not cause a notable α3 subunit expression decrease. The current amplitude generated by 1 mM GABA showed no change between these cell groups (1795.4 ± 1393 pA in control cells vs. 1539.3 ± 653 pA in α3-siRNA treated cells), although the GABA response showed three important changes. First, α3-siRNA treated cells showed an increase in the GABA response desensitization rate, as illustrated in Fig. 9B. This figure shows that control cells displayed a response desensitization fitted to an exponential decay with a mean time constant of 2109.2 ± 1321 millisecond, whereas the transfected cells presented desensitization time constants with an average of 1190 ± 253 millisecond (15–25 cells in each group; from four distinct transfections). Second, D-R curves constructed for these two experimental groups (Fig. 9C) showed that α3 subunit silencing caused an increase in the sensitivity of the expressed receptor and that the control EC50 of 75.1 ± 4.12 μM decreased to 46.6 ± 3.8 μM in α3-siRNA transfected cells (n = 10), a statistically significant difference. And third, α3-siRNA transfected cells also showed a decrease in sensitivity to β-CCB (10 μM) coapplied with GABA (10 μM) and generated a weaker increase of the GABA response to 156.4% ± 32.2% (nine cells) compared with the control effect of 227.26% ± 62.2% (10 cells), as illustrated in Fig. 9D.
α3 GABAAR subunit silencing in cells of the oligodendroglial lineage. (A) Images illustrate α3 subunit immunodetection (in green, nuclei in blue) in OPCs maintained in culture. The first panel depicts cells without treatment (untreated). The cells designated as α3-siRNA were transfected with specific interfering sequences against the subunit, whereas control cells were processed for transfection with scramble sequences. All groups correspond to cells 72 hours after the transfection process. The estimated α3 expression in AUF was normalized with respect to the untreated group as shown in the bar graph (average of seven slices from three transfected cultures). (B) Analysis of responses evoked by 1 mM GABA in control and α3-siRNA treated cells. Traces correspond to typical records in both conditions where the desensitization kinetics was adjusted to an exponential curve to estimate the time constant (τ) decay, and the graphs show the amplitude response and τ values for each group, where each data point represents the response of one cell. The mean ± S.D. is shown in gray in each case. (C) Traces illustrate GABA D-R curves built for both groups; in the graph, each data point represents the mean ± S.D. of the normalized response in 10 cells. Lines are the best fit to D-R curves and broken lines indicate the corresponding EC50. (D) Traces illustrate typical response in control (black), and α3-siRNA treated cells (red) evoked by GABA (10 μM) alone, or coapplied with 10 μM of β-CCB (blue trace). Bar graph shows the amplitude (mean ± S.D., in gray) response values for each group, where each data point represents the response in a single cell. *P < 0.001 for comparisons between the control cells vs. α3-siRNA treated cells, unpaired two-tailed Student’s t test.
γ1 GABAAR subunit silencing in cells of the oligodendroglial lineage. (A) Images illustrate γ1 subunit immunodetection (in green, nuclei in blue) in OPCs maintained in culture. The first panel shows the untreated cells. The γ1-siRNA group represents cells transfected with specific interfering sequences against the subunit. Control cells were processed for transfection with scramble sequences. (B) After 72 hours, the expression of γ1 was estimated in AUF, and the values were normalized with respect to the untreated group (average of seven slices from three transfected cultures). Also, traces illustrate responses elicited by 1 mM GABA in control and γ1-siRNA treated cells. The response amplitude for each group is represented in the graph. Each data point is the amplitude obtained in a single cell, and the mean ± S.D. is indicated in gray. (C) Traces illustrate GABA D-R curves built for both groups, where each data point represents the mean ± S.D. of the normalized response in 17–20 cells, and lines are the best fit to D-R curves. (D) Traces illustrate typical response in control (black), and γ1-siRNA treated cells (light blue) evoked by GABA (10 μM) alone, or coapplied with 10 μM of β-CCB (dark blue trace). Bar graph shows the amplitude (mean ± S.D., in gray) response values for each group, where each data point represents the response in a single cell. *P < 0.001 for comparisons between the control cells vs. γ1-siRNA treated cells, unpaired two-tailed Student’s t test.
γ1 subunit silencing in OPCs also caused changes, and although γ1 protein expression was not reduced by γ1-siRNA transfection as strongly as that designed against α3, the effect reached a 71 ± 1.6% decrease at 4 DIV as evidenced by immunodetection (Fig. 10, A and B). However, the γ1 expression decrease was accompanied by an average amplitude diminution of GABA (1 mM) peak response compared with the respective control of 55.1% ± 1.5% (1795.4 ± 1393 vs. 805.6 ± 463 pA). In this case, the corresponding EC50 values for both groups did not show a statistically significant difference (75.1 ± 4.12 μM for the control group vs. 88.9 ± 3.48 μM for γ1-siRNA transfected cells; 10–15 cells in each case, four different transfections) (Fig. 10C). In similar manner to that showed for α3-siRNA transfected cells, the γ1-siRNA transfected cells were less sensitive to β-CCB, and in the presence of the β-carboline the GABA-response was barely enhanced to 113.2% ± 16.6%, whereas the control group (same than control in Fig. 9D) was increased to 227% (Fig. 10D). These results showed that both subunits, α3 and γ1, seem to participate in the conformation of the endogenous receptor. They also suggested that the absence of α3 was compensated by another subunit, whereas γ1 subunit expression was essential to maintain the GABA response in OPCs as well as their sensitivity to β-CCB.
Discussion
The molecular mechanisms responsible for the control of myelination are only partially known, and their definition will help to identify new therapeutic targets for the treatment of various pathologic conditions. GABA signaling participates in myelination (Zonouzi et al., 2015; Arellano et al., 2016; Balia et al., 2017; Hamilton et al., 2017; Kalanjati et al., 2017; Shaw et al., 2018; Cisneros-Mejorado et al., 2019; Kalakh and Mouihate, 2019; Serrano-Regal et al., 2019); however, its specific functional role and mechanisms involved are not completely understood. To progress in these issues, this study aimed to determine the molecular identity of the main GABAAR expressed in the oligodendroglial lineage, especially in OPCs and OLs from newborn rats (Arellano et al., 2016). The results indicated that the receptor composition includes the α3 and γ1 subunits, whereas the most likely β subunit involved is β2. The identity proposed here, α3β2γ1, agrees with other studies proposing a composition α3/β2 or β3/γ1 or γ3 from a pharmacological and functional perspective (Arellano et al., 2016), and with the transcriptomic analysis performed in NG2+ cells (Larson et al., 2016), where the mentioned subunits have a high expression level. The transcriptomic analysis presented here in PDGFRα+ cells confirmed the expression of the coding sequences for various subunits, where the α2 and α3 subunits, together with β2, β3, and γ1, were all well represented. This analysis also highlighted the low or null expression of γ2, γ3, and δ subunits. Thus, coincidentally, a similar group of subunits was suggested from all the transcriptomic analyzes described and the set of subunits proposed from the functional studies, regardless of the murine species, age, and brain region. This concordance supports the relevant contribution of α3, β2, and γ1 subunits to GABAAR in cultured OLs from the neonatal brain but clearly does not exclude participation of other subunits for which solid evidence of their expression has been found in certain areas of the nervous system and in different stages of development (Passlick et al., 2013; see also Serrano-Regal et al., 2020).
GABAAR subunits protein expression was confirmed by immunodetection in cultured OLs and OPCs from the forebrain, and transcripts of α3, β2, β3, γ1, and γ3 subunits were subsequently amplified from OLs isolated and purified from rat optic nerve; the corresponding genes were cloned, and cRNA synthesized in vitro was expressed into X. laevis oocytes to analyze the functional characteristics of six possible receptors. Their characteristics were compared with those shown by the endogenous receptors reported previously (e.g., Arellano et al., 2016) to determine which combination most closely reproduced the endogenous GABA response. Results studying GABA sensitivity showed that α3β2γ1 was the most efficient in terms of amplitude response; it remains unknown if this represents an intrinsic condition for preferential assemble. However, it is important to note that substitution of β2 by β3 results in an important decrease in the amplitude of the GABA responses.
GABAARs with α3β2γ1 subunits have an EC50 of 53 μM. Substitution of β2 by β3 in this combination did not change GABA responses, whereas substitution of γ1 by γ3 caused a threefold increase in GABA sensitivity. Hence, the EC50 value for the heterologously expressed α3β2γ1 is close to that of the endogenous receptor, as all reports so far indicate an EC50 for GABA between 70 and 100 μM (e.g., Arellano et al., 2016).
On the other hand, Zn2+ sensitivity is a key parameter for this comparison because high sensitivity is an indicator of γ subunit absence. It was demonstrated that compared with γ2 and γ3 subunits, γ1 confers moderate Zn2+ sensitivity; furthermore, it was also clear that β subunit substitution also affected this parameter. The receptors containing the γ1 subunit were studied further, applying (10 μM) GABA and constructing D-R curves for Zn2+. The α3β2γ1 conformation presented an IC50 of 25 μM, which was similar to the value observed in the endogenous receptor, whereas the substitution of β2 by β3 increased threefold this parameter. Therefore, the combination containing β2 is closer to the endogenous value. Additionally, by comparing the IC50 values for receptors containing the combination α3β2 with, or without a γ subunit, once again α3β2γ1 presented the closest IC50 value to that described for the endogenous receptor.
The combination α3β2γ1 also presented an endogenous-like sensitivity for allosteric modulators, such as DZP, indiplon, and the β-carboline β-CCB. The β-CCB effect is important because it distinguishes between receptors expressed either in OLs or neurons. Given that this drug has the potential for specifically promoting GABAergic signaling in OLs in vivo (Arellano et al., 2016; Cisneros-Mejorado et al., 2019), it was critical to explore in detail the action of β-CCB on the α3β2γ1 receptor. Results indicated that β-CCB effect did not require interaction with the classic BZP binding site, since it was not antagonized by FMZ. Instead, FMZ caused a greater enhancement, suggesting, as proposed before (Sieghart, 2015), that β-CCB had a dual effect in which the classic site promoted inhibition (i.e., the inverse agonist effect) and a second site caused the positive modulatory effect. Studies have suggested that this potentiation site for β-carbolines might include residues conforming the low-affinity binding site for DZP located in the second transmembrane domains of α, β, and γ subunits (Walters et al., 2000). This is supported by results showing that the α3β2 receptor is still potentiated by β-CCB. Because subunit substitution might represent a more complex change in the final binding site configuration, we performed point mutations in α3β2γ1 to eliminate the high and low affinity DZP binding sites. Mutations eliminated DZP sensitivity as described for γ2-containing receptors (Benson et al., 1998; Walters et al., 2000); however, β-CCB was still effective as an enhancer in both mutants, indicating that its effect was not due to interaction with the proposed sites; thus, the β-CCB binding site acting as enhancer on α3β2γ1 remains unknown. The GABAAR in OLs presented similar behavior with respect to the DZP high-affinity binding site, since the β-CCB positive modulatory effect was not antagonized by FMZ. This adds evidence supporting the idea that α3β2γ1 and the endogenous receptor are similar.
The modulatory effect on the endogenous receptor by β-CCB was due to a fivefold increase in GABA sensitivity. Distinct β-carbolines presented differential potency on the oligodendroglial receptor, drugs that generated a net potentiating effect on the GABA response in OLs, such as β-CCB or β-CCE, did not have a prominent effect in neurons; thus, they can potentially be used as drugs to differentially stimulate GABAergic signaling in OLs. Lipophilicity of the β-carboline seems to be determinant for the interaction with the receptor, a highly lipophilic drug such as β-CCB had the greatest positive effect, whereas the most hydrophilic DMCM turns to be inhibitory in both cell types. This suggests that the enhancer binding site has relation with the receptor lipophilic microenvironment; thus probably the specific composition of the oligodendroglial and neuronal membranes would influence on the degree of inhibition or enhancement observed for the different drugs, a distinct lipidic microenvironment might also explain some differences observed using the heterologous expression model.
To give greater certainty about the participation of both α3 and γ1 subunits, these were specifically silenced in OPCs. Subunit silencing strongly decreased the corresponding protein expression and substantially changed the OPC GABA response. α3 subunit silencing altered GABA sensitivity of the expressed receptor, as well as the receptor desensitization kinetics and the sensitivity to β-CCB. However, its peak amplitude remained unchanged, suggesting that the α3 absence was compensated by some other subunit conforming a different receptor; α1 and α2 subunits are candidates for this because both are expressed in OPCs, which would explain the increase in GABA sensitivity and the desensitization rate. In contrast, γ1 silencing caused a clear decrease in amplitude response, suggesting that γ1 is essential for the GABAAR expression in the OPC membrane; thus, this subunit is a plausible target to be downregulated in experiments designed for the study of GABAergic signaling function during the myelination process.
Together, these results provide pharmacological (use of β-carbolines) and genetic information that can be used to generate changes in the effectiveness of the GABAAR expressed in OLs and their precursor cells in the neonatal stage, which could be crucial to modify the status of GABAergic signaling in the context of its interaction with neurons during the first peak of myelination. The transcriptomic analyzes suggest that the proposed subunits remain well represented in OPCs until the adulthood; it is therefore likely that these subunits will be involved throughout life. Given that other subunits are coexpressed and that clearly there are differences of expression in different areas, participation of some other subunits cannot yet be completely ruled out.
Thus, evidence indicates that the endogenous GABAAR expressed in oligodendroglial lineage cells, from its OPC stage to premyelinating OLs, most likely contains the α3, β2, and γ1 subunits; this contributes to the determination of the oligodendroglial receptor structure that will allow molecular studies to characterize its function in the myelination process.
Acknowledgments
We thank Jessica González Norris for editing this manuscript. We also thank Felipe Cornejo Ortiz, Dra. Nuri Aranda López, Lic. Ma. de Lourdes Lara Ayala, Ing. Elsa Nydia Hernández Ríos, M.V.Z. Martín García Servín, Alejandra Castilla León, Francisco J. Valles Valenzuela, and Ing. Ramón Martínez Olvera for their expert technical support. A.C.-M. is a researcher from Cátedras-CONACYT commissioned at Instituto de Neurobiología in the Universidad Nacional Autónoma de México (UNAM). R.P.O. is a doctoral student from Programa de Doctorado en Ciencias Biomédicas of the Universidad Nacional Autónoma de México (UNAM) and received fellowship no. 700811 from CONACYT, México.
Authorship Contributions
Participated in research design: Ordaz, Garay, Limon, Matute, Arellano.
Conducted experiments: Ordaz, Garay, Pérez-Samartín, Sánchez-Gómez, Robles-Martínez, Cisneros-Mejorado, Matute, Arellano.
Performed data analysis: Ordaz, Garay, Limon, Pérez-Samartín, Sánchez-Gómez, Cisneros-Mejorado, Matute, Arellano.
Wrote or contributed to the writing of the manuscript: Ordaz, Garay, Limon, Pérez-Samartín, Sánchez-Gómez, Robles-Martínez, Cisneros-Mejorado, Matute, Arellano.
Footnotes
- Received June 12, 2020.
- Accepted November 5, 2020.
This study was supported by grants from CONACYT-México [No. 252121] and PAPIIT UNAM-México [No. IN203519] to R.O.A.’s laboratory; National Institutes of Health National Institute of Aging [Grant R21-AG053740] and National Institute of Mental Health [Grant R21-MH113177] and the Amon G. Carter Foundation to A.L.; and CIBERNED [CB06/05/0076] and the Ministry of Economy and Competitiveness, Government of Spain [SAF2016-75292-R] and Basque Government [IT1203-19] to C.M.
Some information reported herein was presented during a poster session by Rainald Pablo Ordaz, Edith Garay, Carlos Matute, and Rogelio O. Arellano at the “XIV European Meeting on Glial Cells in Health and Disease, GLIA 2019” on 2019 July 10–13.
↵
This article has supplemental material available at molpharm.aspetjournals.org.
Abbreviations
- AUF
- arbitrary unit(s) of fluorescence
- BZD
- benzodiazepine
- β-CCB
- N-butyl β-carboline-3-carboxylate
- β-CCE
- ethyl 9H-pyrido[3,4-b]indole-3-carboxylate
- β-CHM
- 3-(hydroxymethyl)-β-carboline
- D-R
- dose-response
- DIV
- day(s) in vitro
- DMCM
- 4-ethyl-6,7-dimethoxy-9H-pyrido[3,4-b]indole-3-carboxylic acid methyl ester
- DZP
- diazepam
- FC
- fractional contribution
- FMZ
- flumazenil
- GABAAR
- GABAA receptor
- MBP
- myelin basic protein
- NR
- normal Ringer’s
- OL
- oligodendrocyte
- OPC
- oligodendrocyte precursor cell
- PCR
- polymerase chain reaction
- PDGFRα
- platelet-derived growth factor receptor-α
- ROI
- region of interest
- RNA-Seq
- RNA sequencing
- siRNA
- small interfering RNA
- Copyright © 2021 by The American Society for Pharmacology and Experimental Therapeutics
References
In this issue
GABAAR Molecular Identity in Oligodendrocytes
