GABA_ARs are integral membrane ion channels—permeable to Cl⁻ and HCO₃⁻ anions—composed of five subunits that mediate the major form of fast inhibitory neurotransmission in the CNS (Olsen and Sieghart, 2008; Doyon et al., 2016). There are, at least, 19 distinct GABA_AR subunit genes, which include 6 α (α1-α6), 3 β (β1-β3), 3 γ (γ1-γ3), 3 ρ (ρ1-ρ3), and 1 gene of the respective δ, ε, θ, and π subunits (Sieghart and Savić, 2018). This diversity results in different homomeric or heteromeric subunit combinations that may have specific locations in the CNS, particular pharmacology, and, consequently, distinctive functional characteristics (Vogt, 2015). The subunit profile that forms GABA_ARs depends on several factors including brain region, cell type, developmental stage, and physiological or pathophysiological conditions (Levitan et al., 1988; Seeburg et al., 1990; Waldvogel and Faull, 2015). Currently, 11 GABA_AR subtypes with different subunit combinations have been identified, being most of them heteromeric receptors formed by αxβxγx or αxβxδ (where x represents any subtype of a given subunit; Figure 1A), whereas others are homomeric receptors formed by ρ subunits (Barnard et al., 1998).

Activation of GABA_ARs is relevant for the modulation of myelinating cell physiology (Magnaghi, 2007; Vélez-Fort et al., 2012), however, the specific subunit composition of GABA_ARs expressed in these cells remains unknown. Electrophysiological recordings in OPCs and OLs reveal differences between the response of the GABA_AR expressed in these cells and those expressed in neurons and astrocytes (Table 1), suggesting the presence of a novel GABA_AR subtype with unique stoichiometry in the oligodendroglial lineage (von Blankenfeld et al., 1991; Williamson et al., 1998; Vélez-Fort et al., 2012; Arellano et al., 2016).

von Blankenfeld et al. (1991) suggested that GABA_ARs in murine OPCs and OLs carry a γ subunit required to form the benzodiazepine binding site, as they observed potentiation of the GABA response in these cells with classic benzodiazepines. Moreover, the inverse agonist β-carboline methyl 4-ethyl-6,7-dimethoxy-9H-β-carboline-3-carboxylate (DMCM) reduced the GABA-induced current responses in oligodendroglial cells, unlike what happens in astrocytes. Contrary to that, Williamson et al. (1998) reported no influence of flunitrazepam or DMCM in the response elicited by GABA in rat-derived OPCs, indicating an absence of the γ subunit in the GABA_ARs expressed by these cells. This observation was supported by the inhibitory effect of Zn²⁺, which is characteristic of receptors that lack the γ2 subunit. RT-PCR analyses conducted in the same study did not find expression of γ2, α1, α6, and δ subunit mRNAs in OPCs. Although amplification of other subunits was demonstrated, the results were interpreted with caution since the preparation was 85% pure for OPCs and the presence of GABA_ARs from other cell types could not be excluded. In a third study conducted by Bronstein et al. (1998), the GABA response of an immortalized murine glial cell line that expresses mature myelin proteins was insensitive to diazepam and sensitive to Zn²⁺, reinforcing the idea of γ-subunit absence.

Later, Passlick et al. (2013) reported the expression of two types of GABA_ARs in hippocampal NG2 cells from juvenile mice by functional and pharmacological analyses and single-cell RT-PCR. NG2 cells from this brain area express, on the one hand, postsynaptic GABA_ARs comprised of a combination of α1, α2, β3, γ1, and γ2 subunits and, on the other hand, they have extrasynaptic GABA_ARs mostly lacking the γ2 subunit (Figure 1B).

Our studies of GABA_AR responses conducted in cultured immature OLs from the rat forebrain and mature OLs from the optic nerve showed that these cells are diazepam-sensitive, suggesting once more the presence of a γ subunit (Arellano et al., 2016). This positive modulation by benzodiazepines was observed when using low GABA concentrations (≤EC₃₀), which may explain the discrepancies with previous studies. Concerning the specific subtype of γ subunit, two observations indicate that γ2 may not contribute to oligodendroglial GABA_ARs: (1) Zn²⁺ blocks GABA responses; and (2) indiplon, a positive allosteric modulator acting on γ2 subunit-containing receptors, does not modulate GABA currents. Therefore, these receptors likely contain either γ1 or γ3 subunit (Arellano et al., 2016).

Regarding β subunits, potentiation of the GABA response by loreclezole suggests the presence of β2 or β3 subunits, since β1 subunit-containing receptors are insensitive to this drug (Arellano et al., 2016). Finally, concerning α subunits, α3 is the most likely candidate because it forms receptors with low sensitivity to GABA (Karim et al., 2013), as it is the case of OLs (EC₅₀ between 70 and 100 μM).

Together, these pharmacological studies suggest that the composition of the GABA_AR expressed in rat-derived oligodendroglial cells is a combination of the α3, β2 or β3, and γ1 or γ3 subunits (Figure 1B). Also, we confirmed the expression of the α3 subunit by immunocytochemistry in cultured OLs (Arellano et al., 2016). These observations are supported by previous functional genomic analyses performed in OLs (Cahoy et al., 2008). Moreover, RNA sequencing (RNA-Seq) transcriptional analyses of purified NG2 cells obtained from P17 mice revealed that the α3 subunit is the most expressed α-subtype, while among β subunits, β3 and β2 are much more abundant than β1. Lastly, γ1 is expressed at much higher levels than γ2 and γ3 (Larson et al., 2016).

An important factor for the diversity of GABA_ARs expressed and the subunits involved in their conformation will undoubtedly be the species. For example, regarding humans, a recent study of the GABA_AR-subunit expression in OPCs isolated from the middle temporal gyrus of healthy adults—based on the single-nucleus RNA-Seq analysis by Hodge et al. (2019)—showed that OPCs from this brain area express high mRNA levels of α3, all β subunits, γ2 and, interestingly, the ε subunit (Figure 2). These mRNAs, if translated and incorporated into functional receptors, would increase the variety of potential configurations and may have important functional and pharmacological consequences (Jones and Henderson, 2007; Bollan et al., 2008; Belujon et al., 2009).

GABA_A-type receptors are relevant to SC physiology (Magnaghi et al., 2001, 2006). However, their pharmacological and functional properties, as well as their molecular identity, are not clear. GABA_AR subunit composition in rat-derived cultured SCs includes α2 and α3, as well as the three β subunits, while mRNAs of α1 and γ2 subunits have been found at much lower levels (Magnaghi et al., 2006; Figure 1C). Moreover, the presence of α2, α3, and β3 proteins were confirmed by immunocytochemistry. Alpha-2 and β3 subunits are also expressed in SC-like adult stem cells derived from bone marrow or adipose tissue, as their levels are upregulated following SC differentiation in vitro (Faroni et al., 2012). Also, GABA_AR stimulation with muscimol increases the proliferation rate of SCs, meaning that GABAergic signaling has an important role in these cells (Magnaghi et al., 2006). However, despite these important findings, there is still little knowledge about the specific composition of GABA_ARs expressed in differentiated SCs.

GABA_BRs are G-protein coupled receptors (GPCRs) responsible for the slower and prolonged GABA-mediated inhibitory transmission. They were first described pharmacologically as bicuculline-insensitive metabotropic receptors that were activated by the GABA analog baclofen (Bowery and Hudson, 1979; Hill and Bowery, 1981). Functional GABA_BRs are heterodimers constituted by two receptor subunits, GABA_B1 and GABA_B2, that cooperate to perform signal activation (Kaupmann et al., 1998; Kuner et al., 1999). GABA_B1 is responsible for ligand binding, while GABA_B2 contains binding sites for allosteric modulators (Galvez et al., 2001; Binet et al., 2004), couples with G_i/o-protein, and is necessary for trafficking the heterodimer to the cell surface, where the receptor becomes active (Calver et al., 2000; Couve et al., 2000). Among the effector elements involved in GABA_BR signaling pathways in neurons are voltage-gated Ca²⁺ channels (VGCC), inwardly-rectifying potassium channels (Kir) and adenylyl cyclase (AC; Bowery et al., 2002; Bettler et al., 2004; Figure 3). However, the specific coupling of GABA_BRs to the molecular effector may differ depending on the cell type and region analyzed (Booker et al., 2018).

Myelinating cells express both subunits of GABA_BRs, which are negatively-coupled to AC (Magnaghi et al., 2004; Luyt et al., 2007). However, their functional characteristics are not as well known as in the case of neurons. Regarding oligodendroglial cells, we recently confirmed by immunocytochemistry and RT-qPCR the expression of GABA_B1 and GABA_B2 subunits in OPCs and OLs from the rat cerebral cortex and in OLs from the optic nerve (Serrano-Regal et al., 2020). We also performed calcium imaging assays and electrophysiological recordings in these cells and observed that baclofen does not modify their response to KCl 50 mM, calcium influx, and Kir currents. These results strongly suggest that GABA_BRs in oligodendroglial cells are not coupled to Ca²⁺ and Kir channels in the same manner as in other cell types (Serrano-Regal et al., 2020). Likewise, GABA_BRs from CA1 somatostatin interneurons, unlike pyramidal neurons, are not coupled to the canonical Kir3 signaling cascade (Booker et al., 2018), which suggests a functional diversity of downstream effectors depending on cell type and location, and warrants the need of exploring these features in OLs as well as in SCs.

Charles et al. (2003) did not find any colocalization of GABA_B1 subunit and MBP expression in myelinating OLs in the white matter of the rat spinal cord and suggested that GABA_BR expression in developing OLs decreases during differentiation. Following this, Luyt et al. (2007) observed downregulation of the GABA_B1 subunit in mature OLs from the mouse periventricular white matter, while the expression of GABA_B2 remained constant. As they reported changes in GABA_B1/GABA_B2 ratios in mature OLs, they raised the possibility that one subunit alone or in combination with another protein could make GABA_BR functional in different cell types (Calver et al., 2000; Luyt et al., 2007). In contrast, our recent results show that cultured oligodendroglial cells from the rat forebrain and mature OLs from the optic nerve express GABA_B1 and GABA_B2 subunits at different stages of maturation, as well as mature OLs from the juvenile and adult rodent corpus callosum in vivo (Serrano-Regal et al., 2020). These discrepancies can be explained by the different regions analyzed, as GABA_B1/GABA_B2 expression exhibits important regional variations (Luyt et al., 2007). Since oligodendroglial cells are extremely diversified, the different cells targeted in these studies may correspond to distinct oligodendroglial and OPC subpopulations (Marques et al., 2016; Spitzer et al., 2019; Marisca et al., 2020).

SCs also express different isoforms of the GABA_BR, such as -1a, -1b, -1c, and -2 (Magnaghi et al., 2004, 2006).

Downregulation of GABA_BR expression occurs in pre- and non-myelinating SCs (Corell et al., 2015). Neurosteroids modulate the expression of GABA_BR subunits in cultured SCs and, as GABA_BR is downregulated with age in the PNS, the GABA synthesized in the adult sciatic nerve acts through the ionotropic GABA_AR, present both in neurons and SCs (Magnaghi et al., 2006). Interestingly, the conditional knockout of the GABA_B1 subunit in SCs changes the expression of GABA_AR subunits α3, α4, β1, and δ (Faroni et al., 2019), suggesting that GABA_BRs in these cells regulate somehow the expression of GABA_ARs and/or their subunits.

Overall, more detailed analyses such as single-cell RNA-seq would help to better figure out the expression of GABA_B1 and GABA_B2 subunits along the oligodendroglial and SC lineages.

Contrary to mature neurons, activation of GABA_ARs in OPCs leads to depolarization and an increase in cytosolic Ca²⁺ levels (Kirchhoff and Kettenmann, 1992), resulting from Ca²⁺ influx through activated VGCC (Paez and Lyons, 2020). Thus, a rise in Ca²⁺ in the cytosol may regulate OPC proliferation, migration and maturation and, consequently, OL (re)myelination (Cheli et al., 2016; Santiago-González et al., 2017; Baraban et al., 2018; Krasnow et al., 2018; Marisca et al., 2020).

As oligodendroglial cells constitute a highly dynamic and heterogeneous population, the expression of GABA_ARs and their different subunits changes as these cells progress along their lineage, as occurs with the expression of certain ion-channels (Spitzer et al., 2019), and these changes can affect their intercellular relationship and differentiation. In line with this, we observed that oligodendroglial GABA_AR expression in vitro is dependent on the close interaction between axons and OLs, as OLs cultured alone lose GABA responses with differentiation (Arellano et al., 2016). Moreover, the presence of the γ2 subunit, which is associated with a possible role in neuron-OPC synapse formation, decreases with age along with the density of GABAergic synaptic contacts in cortical NG2 cells of mature mice (Balia et al., 2015). Thus, NG2 cells may switch the expression of GABA_ARs from synaptic (with γ2 subunit) to extrasynaptic (without γ2 subunit) during development (Vélez-Fort et al., 2010). Surprisingly, genetic inactivation of oligodendroglial γ2 does not affect OPC proliferation and differentiation, while it causes progressive and specific depletion of the OPC pool that lacks γ2-mediated synaptic activity without affecting the oligodendrocyte production (Balia et al., 2017). These observations indicate that GABAergic communication in cortical OPCs through γ2-containing GABA_ARs does not play a role in oligodendrogenesis but rather modulates OPC maintenance.

GABAergic signaling regulates OPC population and OPC differentiation and myelination in the cerebellar white matter in vivo (Zonouzi et al., 2015). In turn, hypoxia causes a strong downregulation of the GABAergic synaptic input from local interneurons to OPCs (NG2 cells) as well as an increase in the proliferation of these cells and a delay in their maturation, which limits myelination. These effects are mimicked in control animals when blocking GABA_ARs with their antagonist bicuculline. However, they are reverted when applying tiagabine, a selective inhibitor of the GABA transporter GAT-1 that increases GABA availability in the extracellular space. Treatment with tiagabine results in a decrease of NG2 cell proliferation and an increase of myelinating OLs, reverting the hypomyelinating effect caused by perinatal hypoxia (Zonouzi et al., 2015). These findings strongly suggest that GABAergic signaling (either neuronal activity-dependent or independent) influences OPC development and differentiation and, therefore, it may help to develop novel therapies to improve OPC differentiation into damaged brain areas.

GABAergic signaling through GABA_ARs may also be relevant for the stronger remyelination that occurs following focal demyelination in the corpus callosum of late pregnant rats compared to virgin and postpartum ones (Kalakh and Mouihate, 2019). This pregnancy-associated promyelinating effect was lost when either the GABA_AR was blocked or when 5α-reductase, the rate-limiting enzyme for the endogenous GABA_AR activator ALLO, was inhibited (Kalakh and Mouihate, 2019). Moreover, N-butyl-β-carboline-3-carboxylate (β-CCB), a selective drug activating preferentially oligodendroglial GABA_ARs, promotes remyelination in a model of gliotoxin-induced demyelination in the rat cerebellar caudal peduncle as assessed using magnetic resonance imaging (MRI) together with myelin staining (Cisneros-Mejorado et al., 2020). Together, these results strongly suggest that GABA_AR-mediated signaling promotes myelination and remyelination in OLs either directly or indirectly. However, at odds with these data, activation of GABA_ARs by endogenous GABA in cortical organotypic cultures reduces the number of oligodendroglial cells and myelination whereas enlarges internode length, influencing the velocity of the nerve impulse propagation (Hamilton et al., 2017). These effects may be due to GABA released by glial cells. However, we could not assess this idea as gabazine treatment of OPC cultures had no significant effect on myelin protein production (Serrano-Regal et al., 2020).

Neurosteroid therapy is another pharmacological approach to modulate GABA_AR activity in the nervous system, as they act as allosteric modulators of these receptors in the nanomolar concentration range (Lambert et al., 2009). Indeed, progestin ALLO increases myelin basic protein (MBP) production in rat-derived cerebellar organotypic slices, an effect that requires GABA_ARs (Ghoumari et al., 2003). This observation points to these receptors as mediators of myelination. The effects of neurosteroids are of special interest in postnatal development, as they may help to prevent neurodevelopmental disorders associated with preterm birth (Shaw et al., 2019). ALLO, which is mainly synthesized in the placenta, has an important role during nervous system development. In premature neonates, ALLO concentration decreases abruptly and this decrease is associated, in part, with hypomyelination (Shaw et al., 2015). Consequently, experimental administration of the ALLO analog ganaxolone as replacement therapy in guinea pig-preterm neonates showed positive effects on myelination, through its interaction with GABA_ARs. Therefore, neurosteroid replacement could be a good therapeutic option to improve myelination in this condition (Shaw et al., 2019).

Neurosteroids may also enhance GABA_AR function in SCs. Thus, ALLO acting via GABA_A receptor can influence peripheral myelin protein 22 (PMP22) synthesis (Magnaghi et al., 2006). Moreover, ALLO modulates SC morphology, motility, and myelination in SC/dorsal root ganglia neuron (DRG) co-cultures via the Src/focal adhesion kinase (FAK) pathway, a signaling cascade that involves GABA_ARs and relies on actin rearrangements (Melfi et al., 2017). Therefore, neurosteroids represent a promising molecular approach for the treatment of peripheric pathologies. Together, the studies discussed in this section connect GABA_AR signaling with OPC/SC differentiation and/or myelination using pharmacological approaches. However, the results observed cannot be solely attributed to the action of GABA on GABA_ARs. Although pharmacology may be a good strategy to enhance OPC/SC differentiation and myelination in pathological conditions, potential side-effects must also be considered. To minimize them, it would be of great interest to use more specific drugs acting on myelinating cell GABA_ARs and/or to use genetic approaches to specifically target the different GABA_AR subunits expressed in these cells.

An early study suggests that GABA_BRs may be relevant for OPC development as baclofen increases migration and proliferation in cultured OPCs derived from periventricular white matter (Luyt et al., 2007). However, more recent studies could not confirm those findings as baclofen did not change OPC proliferation or total OPC number in dissociated and organotypic cultures derived from the cortex (Hamilton et al., 2017; Serrano-Regal et al., 2020). This discrepancy could reflect the different brain areas studied. Thus, different subpopulations of OPCs may exist in white and gray matter that behave differently or have different responses to baclofen, as proposed by Luyt et al. (2007). In contrast, baclofen reduced cell proliferation of SC cultures (Magnaghi et al., 2004). However, in dissociated developing DRG primary cultures, in which SCs proliferate spontaneously in vitro, baclofen did not affect (Corell et al., 2015).

On the other hand, GABA and baclofen modulate OPC differentiation, as well as the myelination capacity of mature OLs cultured with DRG neurons, pointing out GABA_BRs as relevant modulators of OL maturation and myelination (Serrano-Regal et al., 2020). Consistent with these observations, GABA_BRs also regulate SC differentiation both in the myelinating and non-myelinating phenotypes. Thus, forskolin-induced SC differentiation in vitro correlates with a redistribution of GABA_B1 and GABA_B2 subunits of GABA_BRs. Indeed, in the cytoskeleton rearrangement that takes place during differentiation, GABA_BRs colocalize with f-actin on the SC elongated processes (Procacci et al., 2013).

Apart from being essential for SC commitment to a non-myelinating phenotype during development, GABA_BRs are key modulators of neuronal-SC interactions regarding myelination, as GABA_B1 receptor total null mice showed altered levels of PMP22 and myelin protein zero (P0) as well as thinner myelin sheaths. These mice also presented fiber alterations, which causes changes in pain behavior, gait abnormalities, and motor coordination disturbances (Magnaghi et al., 2008). Together, these findings suggest a role for GABA_BRs in the control of SC myelination. Moreover, both GABA_BR subunits in addition to GABA and GAD_65/67 were found at the node of Ranvier in a sub-population of myelinated sensory fibers (Corell et al., 2015). Surprisingly, GABA_BR expression is upregulated in SCs of injured nerves, which may be interpreted as an adaptive response for stimulating the neighboring axons to re-grow distally to the injury (Corell et al., 2015).

Finally, conditional deletion of the GABA_B1 subunit in SCs altered their proliferation, migration, and myelination capacities, as well as reduced neurite length of co-cultured DRGs (Faroni et al., 2019). Furthermore, molecular and transcriptomic changes were also observed both in SCs and DRGs derived from mice lacking GABA_B1 subunit in SCs (P0-GABA-B1^fl/fl). Interestingly, the expression of some GABA_AR subunits by SCs and DRGs was also altered, indicating a possible role of GABA_BRs in regulating the expression of GABA_ARs in these cells (Faroni et al., 2019). Similar studies using conditional deletion of GABA_BR subunits in oligodendroglia will help to understand the role of these receptors and their impact on myelination and pain and motor behavior.