More than sixty years after the discovery that gamma-amino-butyric acid (GABA) is a major inhibitory transmitter in the brain, this article reviews its history, with special emphasis on the significant involvement of Canadian scientists. It is now well-established that GABA is the main inhibitory neurotransmitter in the adult forebrain.Reference Martin and Olsen 1 Once released from interneuronal terminals, it activates pre- and postsynaptic GABA receptors, which are categorized into three types: A, B and C.Reference Martin and Olsen 1 , Reference Farrant and Kaila 2 GABAA receptors activate ionotropic anionic channels while GABAB receptors are metabotropic, acting through second messengers. In addition, presynaptic GABAB receptors control transmitter release from excitatory and inhibitory terminals whereas such function remains controversial for GABAA receptors.Reference Draguhn, Axmacher and Kolbaev 3 GABAC receptors, which activate ionotropic channels, are presumably confined to the retina in the adult,Reference Enz, Brandstätter and Hartveit 4 though inhibitory functions in the adult hippocampus have been reported.Reference Alakuijala, Alakuijala and Pasternack 5
GABAA receptors in various regions of the brain have different subunit compositions with specific functional and pharmacological characteristics.Reference Barnard, Skolnick and Olsen 6 , Reference Olsen and Sieghart 7 Moreover, GABAA receptors are located both synaptically (low affinity) and extra-synaptically (high affinity), the latter being activated by spillover of synaptically released GABA; these two categories of receptors are believed to mediate phasic and tonic inhibition, respectively.Reference Semyanov, Walker and Kullmann 8 - Reference Yeung, Canning and Zhu 10 Our review is directed to the early steps that established the function of GABA as an inhibitory neurotransmitterReference Krnjević 11 as well as on more recent evidence pointing to GABAA receptor signaling as a powerful mechanism underlying neuronal network synchronization both during developmentReference Ben-Ari, Gaiarsa and Tyzio 12 and in the adult brain.Reference Avoli and de Curtis 13 Indeed, several studies have identified a paradoxical synchronizing role played by GABAA receptors in cortical structures. Along this road we also discuss the ability of several clinically relevant drugs to modulate GABAergic function and the involvement of this neurotransmitter in neurological disorders.
Establishing GABA as the Major Inhibitory Neurotransmitter
The standard criteria for identification of a neurotransmitter are: (i) its presence in presynaptic neurons; (ii) its release from these presynaptic terminals following activation of these neurons; (iii) its ability to mimic the synaptic response when exogenously applied; and (iv) the ability of agonists or antagonists to enhance or block, respectively, both the response induced by presynaptic terminal activation and that evoked by application of the candidate substance.Reference Obata 14 , Reference Werman 15 It took nearly two decades for GABA to comply with these criteria. Almost simultaneously, in November 1950, three independent studies reported the presence of large amounts of GABA in amphibian and mammalian brains.Reference Awapara, Lanua and Fuerst 16 - Reference Udenfried 18 However, while GABA is readily obtained by alpha-decarboxylation of glutamic acid - the most prominent amino-acid in brain - as an omega amino-acid it cannot be used for the synthesis of proteins and peptides; its function, therefore, remained a matter of speculation for several years.
The first clarifying event came about when the Austrian zoologist Ernst Florey happened to test mammalian brain extracts on the crayfish abdominal stretch receptors.Reference Florey 19 Powerful suppression of the spontaneous firing of the receptor neuron led Florey to conclude that the brain extracts contained a very effective inhibitory agent, which he appropriately named Factor I. To identify the chemical nature of Factor I, he came to Montreal, where there followed several major developments. Working in KAC Elliott’s neurochemistry laboratory at the Montreal Neurological Institute, Hugh McLennan and Florey used the ‘cortical cup’ technique pioneered by Elliott and JasperReference Elliott and Jasper 20 to show that Factor I was released from the cat’s brain in situ Reference Florey and McLennan 21 and that topical applications of Factor I could depress transmission both in sympathetic ganglia and in the lumbar spinal cord.Reference Florey and McLennan 21 , Reference Florey and McLennan 22 Subsequently, after an exhaustive search for active brain constituents, Elliott’s group concluded that GABA accounted for most of Factor I inhibitory action on the crayfish receptor neuron.Reference Bazemore, Elliott and Florey 23 , Reference Bazemore, Elliott and Florey 24 In support of a functional role for GABA came several studies by Herbert Jasper: first, showing depressant effects of GABA on spontaneous spindle waves and evoked potentials obtained with electroencephalogram (EEG) recordings from the cerebral cortex;Reference Iwama and Jasper 25 and then that the release of GABA from the cortex consistently varied with electrographic activity.Reference Jasper, Khan and Elliott 26 , Reference Jasper and Koyama 27
In other respects, further progress was anything but straightforward. In Canada, two of the leaders in the field, Florey and McLennan, found reason to believe that Factor I and GABA were not identical.Reference Florey and McLennan 28 In Australia, following up on these authors’ earlier report demonstrating that Factor I inhibits transmission in the spinal cord, Curtis et al. applied GABA by iontophoresis from micropipettes to a variety of neurons in the feline spinal cord.Reference Curtis, Phillis and Watkins 29 GABA (and some related amino acids) consistently depressed neuronal activity; however, for a number of reasons – including the apparent lack of an inhibitory post-synaptic potential (IPSP)-like hyperpolarizing action and resistance to antagonism by strychnine, known as an effective blocker of spinal inhibition - they concluded that the effects of GABA action were non-specific. A few years later, however, a similar iontophoretic approach, testing single units in the neocortex and cerebellum, led Krnjević and Phillis to propose that GABA was most likely to be the inhibitory transmitter in the brain.Reference Krnjević and Phillis 30 This view was soon supported by the demonstration that (i) the GABA content of individual synaptosomes was sufficient to inhibit cortical cell firingReference Krnjević and Whittaker 31 and (ii) stimulation of the neocortexReference Jasper and Koyama 27 or cerebellumReference Obata and Takeda 32 caused GABA release. Compelling evidence for GABA’s role as inhibitory transmitter came with the demonstration that its effects, when iontophoretically applied to corticalReference Krnjević and Schwartz 33 , Reference Dreifuss, Kelly and Krnjević 34 cuneateReference Galindo, Krnjević and Schwartz 35 and Deiters neuronsReference Obata, Ito, Ochi and Sato 36 , had characteristics (e.g., reversal potentials) similar to those of the IPSPs generated by these cells.
As further mentioned in the following sections, many convulsants act by blocking inhibitory transmission. Indeed, one of the best known, picrotoxin, had proved to be an effective blocker of Factor I and GABA’s inhibitory action on the stretch receptor neuronReference Elliott and Florey 37 , Reference Van der Kloot and Robbins 38 as well as on mammalian brain stem neurons.Reference Galindo 39 , Reference Obata and Highstein 40 Another Canadian, RH Manske - working at the National Research Council Laboratories in Ottawa - had previously discovered bicuculline, the more direct antagonist of GABA receptors, during his systematic investigations of plant alkaloids. In 1933, he identified an unknown compound in the root bulbs of the common Canadian spring flower, Dutchman’s breeches (Dicentra cucularia), which he named bicuculline.Reference Manske 41 Testing samples of bicuculline provided by Manske, pharmacologists at the University of Toronto shortly after found that bicuculline was a powerful convulsant when administered to frogs and rabbits.Reference Welch and Henderson 42 Nothing further happened with bicuculline until 1970 when, after testing various known convulsants, Curtis et al. reported that bicuculline not only blocked inhibition in the central nervous system (CNS) but was a very effective GABA antagonist.Reference Curtis, Duggan and Felix 43 Thus, by the early 1970s, GABA had become widely accepted as the principal inhibitory neurotransmitter, especially in the brain.Reference Nistri and Constanti 44
GABA Receptors as Targets for Drugs
Having established the role of GABA as the main inhibitory transmitter at central synapses, researchers tried to identify the interactions occurring between this receptor and neurotropic drugs or endogenous substances. In the 1970s, several studies revealed that benzodiazepines could enhance GABAergic inhibition in the CNS,Reference Costa, Guidotti and Mao 45 - Reference Choi, Farb and Fischbach 47 and led to the identification of a benzodiazepine site in the GABA receptor.Reference Braestrup and Squires 48 , Reference Möhler and Okada 49 At that time, it was still unclear whether GABA receptors consisted of structurally and functionally different subtypes: the metabotropic GABAB receptor was not discovered before the end of that decade.Reference Bowery, Hill and Hudson 50 Nonetheless, electrophysiological, biochemical and behavioral data clearly indicated that benzodiazepine ligands modulated ionotropic GABAergic function. Subsequent molecular and pharmacological studies demonstrated an allosteric ‘benzodiazepine site’ on most α subunit-containing GABAA receptors.Reference Olsen 51
Over the last four decades, additional agents have been shown to modulate GABAA receptor function and it appears that these actions depend on separate binding sites.Reference Olsen 51 These include, barbiturates (which depress neuronal activity and, at higher doses, may have a direct action on the GABAA receptor),Reference Nicoll 52 - Reference Rho, Donevan and Rogawski 57 steroid metabolitesReference Majewska, Harrison and Schwartz 58 and neurosteroids,Reference Majewska 59 , Reference Belelli and Lambert 60 ethanolReference Nestoros 61 - Reference Wallner, Hanchar and Olsen 63 as well as propofol.Reference Krasowski, Nishikawa and Nikolaeva 64
Decreasing GABA Function Leads to Epileptiform Activity and Seizures
In 1954, two independent groups reported seizures in infants fed with a formula that was accidentally deficient in pyridoxine (also known as vitamin B6); pyridoxine is the coenzyme for the synthesis of GABA from glutamic acid via the enzyme glutamic acid decarboxylase.Reference Coursin 65 , Reference Molony and Parmelee 66 Further studies showed that GABA prevented seizures and that drugs interfering with GABA synthesis could induce convulsions.Reference Hawkins and Sarett 67 , Reference Benassi and Bertolotti 68 This evidence was later corroborated by experiments aimed at identifying the cellular and pharmacological mechanisms underlying epileptiform synchronization. First, it was shown that drugs capable of inducing focal interictal-like discharges preparations in vivo Reference Dichter and Spencer 69 , Reference Ayala, Dichter and Gumnit 70 and in vitro Reference Schwartzkroin and Prince 71 were GABAA receptor antagonists, and that the occurrence of this type of epileptiform activity was characterized by reduction in recurrent and/or feed-forward inhibition.Reference Schwartzkroin and Prince 72 , Reference Dingledine and Gjerstad 73 Second, it was reported that the onset of electrographic seizures induced by repetitive activation of hippocampal inputs are associated with rapid fading of IPSPs.Reference Ben-Ari, Krnjević and Reinhardt 74 Shortly later, Kostopoulos et al. found that cortical recurrent inhibition, which was maintained during an EEG pattern of generalized spike-wave discharge, became markedly reduced before the onset of electrographic tonic-clonic seizures.Reference Kostopoulos, Avoli and Gloor 75 Therefore, early in 1980s, weakening of GABAA receptor signaling was regarded by many neuroscientists as the main cause of seizure activity.
Role of GABA in Synchronizing Adult Neuronal Networks
Early intracellular recordings from thalamic relay cells in vivo highlighted the potential role of presumptive GABAergic IPSPs in causing what was known as rebound excitation, and ultimately leading to low frequency (7-14 Hz) thalamocortical oscillations such as those associated with sleep spindles.Reference Andersen and Andersson 76 , Reference Purpura 77 A major advance in understanding the mechanism of EEG spindles and the switch between waking and sleep came with the demonstration by Steriade’s group at Laval University (Québec) that spindle generation depends on the activity of GABAergic inhibitory neurons in the reticular thalamic nucleus.Reference Steriade, Deschênes and Domich 78 , Reference Steriade, Domich and Oakson 79 Along with the experiments performed in vitro by Llinas,Reference Llinás 80 evidence obtained by several research groups during the last three decades have shown the importance of GABA receptor-mediated IPSPs for hyperpolarizing thalamocortical neurons at membrane potential levels that inactivate low-threshold Ca2+ conductances and activate the Ih current.Reference McCormick and Pape 81 - Reference Crunelli, David and Leresche 84
Depolarizing effects of GABA were identified early in the spinal cord;Reference Eccles, Schmidt and Willis 85 , Reference De Groat, Lalley and Saum 86 but GABAA receptor-activation was later observed in hippocampal pyramidal cells, most often following activation of receptors located on the dendrites.Reference Andersen, Dingledine and Gjerstad 87 , Reference Alger and Nicoll 88 These depolarizing effects were interpreted to reflect a higher concentration of Cl- in the dendrites than in the soma of principal cells, where hyperpolarizations continued to be recorded.Reference Misgeld, Deisz and Dodt 89 Subsequent experiments have shown that GABAA receptor-mediated depolarizations in cortical neurons can also be generated by HCO3 −, which also passes through the anionic channels that are opened during inhibitionReference Kelly, Krnjević and Morris 90 , Reference Bormann, Hamill and Sakmann 91 but with an equilibrium potential more positive than for Cl−.Reference Grover, Lambert and Schwartzkroin 92 - Reference Staley, Soldo and Proctor 94 The depolarizing effect of HCO3 − and K+ efflux and resulting gamma-frequency oscillations are sharply enhanced in the young hippocampus when carbonic anhydrase increases cytoplasmic HCO3 −.Reference Rivera, Voipio and Kaila 95 In addition, these GABAA receptor-mediated, HCO3 −-dependent depolarizations can also activate voltage-gated Ca2+ channels.Reference Autere, Lamsa and Kaila 96
As mentioned in the previous section, studies published between the 1950s and 1980s demonstrated that reducing GABAA receptor signaling can lead to seizures in vivo and interictal-like activity in vitro. This view was challenged by reports that epileptiform synchronization occurs during pharmacological conditions that do not interfere with GABAA receptor function and, in some cases, they appeared to enhance it.Reference Avoli and de Curtis 13 Specifically, epileptiform discharges were observed during application of Mg2+ free-medium in concomitance with the preservation of IPSPs.Reference Mody, Lambert and Heinemann 97 , Reference Tancredi, Hwa and Zona 98 In addition, it was shown that cortical networks made hyperexcitable by reducing the function of specific K+ channels could generate synchronous, propagating neuronal activity, even when ionotropic glutamatergic transmission was blocked.Reference Perreault and Avoli 99 - Reference Dickson and Alonso 102 It was also found that under these pharmacological conditions, each synchronous event was associated with elevations in extracellular potassium ion concentration ([K+]).Reference Avoli, Barbarosie and Lucke 103 - Reference Morris, Obrocea and Avoli 105 Interestingly, shortly before, Mary Morris at the University of Ottawa had found in the hippocampal slice preparation that selective activation of GABAA receptors caused increases in extracellular [K+].Reference Barolet and Morris 106 This finding was later elucidated in Kai Kaila’s laboratory; specifically, it was shown that excessive activation of GABAA receptors leads to accumulation of Cl- inside the postsynaptic cells and to a subsequent increase in K-Cl co-transporter (KCC2) activity that makes K+ and Cl- move to the extracellular space.Reference Viitanen, Ruusuvuori and Kaila 107
GABA Receptor Signaling and High Frequency Oscillations
Synchronization of cortical interneurons also plays a critical role in the generation of fast EEG oscillations in the mature CNS; these include beta-gamma (at 20–80 Hz) and high frequency oscillations (>80 Hz, so called ripples).Reference Buzsáki, Horváth and Urioste 108 Both beta-gamma rhythms and ripples - which are recorded from several cortical structures including those of the limbic system -Reference Murthy and Fetz 109 - Reference Chrobak and Buzsáki 112 are implicated in higher brain processes such as attention, sensorimotor integration, consciousness, learning and memory.Reference Girardeau, Benchenane and Wiener 113 , Reference Montgomery and Buzsáki 114 In vivo studies have shown that ripples represent population IPSPs generated by principal neurons entrained by synchronously active interneuronal networks.Reference Buzsáki, Horváth and Urioste 108 , Reference Chrobak and Buzsáki 112 , Reference Ylinen, Bragin and Nádasdy 115
Similar fast oscillations are reproduced in vitro by bath applying the cholinergic agonist carbachol, high-K+, kainic acid, or metabotropic glutamate receptor agonistsReference Whittington, Traub and Jefferys 116 - Reference Dickson, Biella and de Curtis 119 as well as by electrical tetanic stimulation.Reference Whittington, Stanford and Colling 120 , Reference Bracci, Vreugdenhil and Hack 121 These studies showed that fast activities reflect the synchronization of inhibitory GABAergic networks,Reference Whittington and Traub 122 , Reference Mann and Paulsen 123 with or without the contribution of excitatory glutamatergic networks and gap junctions.Reference Bartos, Vida and Frotscher 124 - Reference Tamás, Buhl and Lörincz 126 The hypothesis that gamma oscillations reflect interactions within interneuron networks is also supported by computer modeling.Reference Wang and Buzsáki 127 , Reference Traub 128
GABA and Brain Maturation
GABAA receptor-mediated depolarizations in the developing CNS were first reported in rabbit hippocampus.Reference Mueller, Chesnut and Schwartzkroin 129 However, depolarizing GABAA receptor-dependent depolarizations were clearly established in immature brain tissue when Ben-Ari et al.Reference Ben-Ari, Cherubini and Corradetti 130 found giant GABAA receptor-mediated depolarizing potentials that were spontaneously generated by rat CA3 pyramidal cells during the first 12 days of postnatal life. It was later shown that, at least in rodents, these GABAergic potentials are consistently recorded during the first post-natal week and that, as the brain matures, they gradually change to hyperpolarization,Reference Cherubini, Gaiarsa and Ben-Ari 131 , Reference Zhang, Spigelman and Carlen 132 which is due to maturation of the outward Cl- transporter KCC2.Reference Rivera, Voipio and Kaila 95 , Reference Rivera, Voipio and Payne 133 In line with this view, synaptic response generated by glycine, which are also Cl- mediated, change from depolarizing to hyperpolarizing during ontogeny.
GABAergic excitation in the immature CNS may play a role in the growth, differentiation, maturation and preservation of neurons as well as in establishing their synaptic connectivity during development.Reference Ben-Ari, Gaiarsa and Tyzio 12 These processes depend on Ca2+ influx that is presumably the main signaling system causing oscillations of cytoplasmic calcium ion concentration and activation of several Ca2+-binding proteins. Since glutamatergic AMPA receptor-type pathways are not operative in the immature brain, GABAA receptor-mediated depolarizing currents in concert with those caused by activation of N-methyl-D-aspartate (NMDA) receptors, which are present quite early on, may cause significant Ca2+ influx.Reference Ben-Ari, Gaiarsa and Tyzio 12
GABA and Tonic Inhibition
Activation of high affinity GABAA receptors, localized extra- or perisynaptically, also causes tonic inhibition; these receptors have distinct, subunit compositions that include the δ α4, α5 and α6 subunits.Reference Semyanov, Walker and Kullmann 8 - Reference Yeung, Canning and Zhu 10 , Reference Nusser, Sieghart and Somogyi 134 - Reference Glykys, Mann and Mody 136 This tonic (‘always on’) current appears to be activated by taurine in the mouse ventrobasal thalamus, thus reducing the excitability of thalamocortical relay neurons.Reference Jia, Yue and Chandra 137 Neurosteroids have been proposed to modulate preferentially tonic rather than phasic inhibition.Reference Belelli and Lambert 60 , Reference Mihalek, Banerjee and Korpi 138 - Reference Lambert, Cooper and Simmons 140 Indeed, neuroactive steroids may play a role in catamenial epilepsy and in temporal lobe epilepsy, as suggested by their ability to delay the establishment of this chronic condition following pilocarpine-induced status epilepticus in rodents.Reference Biagini, Panuccio and Avoli 141
Over the last decade tonic currents have also been proposed to be involved in the potentiating effects of ethanol on GABAA receptor-mediated inhibition.Reference Mody, Glykys and Wei 142 For instance, in the hippocampus, low concentrations of ethanol selectively augment the tonic inhibitory currents mediated by δ subunit-containing GABAA receptors.Reference Wei, Faria and Mody 143 In addition, it has been reported that chronic intermittent ethanol treatment causes a switch of its actions in the hippocampus from extrasynaptic to synaptic GABAA receptors.Reference Liang, Zhang and Cagetti 144
GABA and Neurological Disorders
Dysfunction of GABAA receptor signaling - such as loss or rearrangement of inhibitory interneurons, changes in subunit composition, intracellular ionic homeostasis, etc… - has been documented in several neurological conditions including epileptic disordersReference Olsen, Bureau and Houser 145 - Reference Avoli, Louvel and Pumain 147 and mental retardation conditions such as the Fragile XReference Curia, Papouin and Séguéla 148 - Reference Conde, Palomar and Lama 150 and Down syndromes.Reference Fernandez, Morishita and Zuniga 151 - Reference Deidda, Parrini and Naskar 153 Because a relative excess of GABAergic inhibition may be a factor in conditions associated with memory loss such as aging,Reference Rissman and Mobley 152 the findings that both synaptic plasticity and long term memory can be rescued by agents reducing GABA-mediated inhibitionReference Izquierdo and Medina 154 - Reference Zhu, Huang and Kalikulov 157 indicate potential lines of treatment.
To review all the studies published on these topics is beyond the purpose of this review. However, one should pay attention to the evolving concept of the role played by GABA under some specific pathological conditions such as the generation of seizures in epileptic patients and animal models. For instance, and contrary to expectations from evidence obtained in the 1970s by employing acute models of focal epileptiform discharges,Reference Ayala, Dichter and Gumnit 70 the onset of seizures recorded from epileptic patients undergoing presurgical electrophysiological investigations is characterized by marked reduction of unit firing.Reference Truccolo, Donoghue and Hochberg 158 , Reference Schevon, Weiss and McKhann 159 In addition, recent data obtained from animal models of focal epilepsy, indicate that seizure onset is accompanied by increased activity of inhibitory interneurons that may, in fact, silence principal neurons.Reference Grasse, Karunakaran and Moxon 160 - Reference Toyoda, Fujita and Thamattoor 162 This paradoxical role of inhibition in initiating seizures is in line with evidence obtained from several in vitro and in vivo acute studies that have demonstrated the participation of GABAergic, often depolarizing, currents in both the initiation and maintenance of prolonged periods of epileptiform synchronization.Reference Avoli and de Curtis 13 , Reference Velazquez and Carlen 163 - Reference Fujiwara-Tsukamoto, Isomura and Imanishi 167 The exact mechanisms by which GABAA receptor signaling facilitates seizure activity are still under scrutiny but evidence obtained in vitro from animal preparationsReference Avoli and de Curtis 13 , Reference Avoli, Barbarosie and Lucke 103 , Reference Avoli, Louvel and Kurcewicz 104 and slices of the human dysplastic cortexReference Avoli, Louvel and Pumain 147 point at increases in extracellular [K+] caused by GABA release from interneurons at the onset of seizure activity. It is indeed well established that elevating extracellular [K+] increases excitability and the occurrence of epileptiform discharges both in vivo and in vitro.Reference Zuckermann and Glaser 168 - Reference Traynelis and Dingledine 171 Evidence that GABAA receptor signaling can promote seizures may explain the disappointingly limited efficacy of antiepileptic drugs designed to potentiate GABA receptor-mediated inhibition.
Concluding Remarks
GABA has not ceased to surprise us. Throughout the brain, it is present in, and released from a variety of inhibitory neurons with different characteristics, and it acts on an even greater variety of receptors. Indeed, such major brain structures as the cerebellar cortex and the striatum consist of mostly GABAergic neurons on which depends efficient locomotion. However, albeit well-established, GABA’s inhibitory function can differ greatly according to the type and composition of the targeted receptors and is, moreover, astonishingly malleable. Not only is it susceptible to modulation by a rich variety of drugs - many of which are in extensive clinical use – but, in addition, is characterized by a quite unexpected plasticity, arising from its membrane action, the activation of anionic channels; whether the action is mainly inhibitory or excitatory depends on the predominance of the relative contributions of chloride and bicarbonate ions, and of course the transmembrane Cl- gradient, determined either locally or generally by the direction of net Cl- transport; all these factors are indeed subject to genetic, developmental and pathological changes. These characteristics help to explain why GABA dysfunction is manifested in so many different ways by neurological patients.
Disclosures
Massimo Avoli and Krešimir Krnjević do not have anything to disclose.