Enhanced Tonic GABA_A Inhibition In Genetic Models Of Typical Absence Seizures

The tonic GABA_A current measured in vitro in TC neurons of the somatosensory ventrobasal thalamus of different genetic models of absence seizures is enhanced compared to the current in their respective control animals.¹⁷ This is true for a polygenic rat model, i.e., the Genetic Absence Epilepsy Rat from Strasbourg (GAERS)¹⁸ (Figure 1A1–2) and for various mice models with known spontaneous monogenic mutations, including stargazer and lethargic mice¹⁹ (Figure A3). In particular, there is a clear developmental profile of this increased GABAergic function since in GAERS up to postnatal day 16 the current is similar to that in the non-epileptic control (NEC) strain but almost doubles in amplitude at postnatal day 17 (Figure 1A2), and remains elevated well past the time of seizure onset¹⁷ (around postnatal day 30 in this strain¹⁸). In contrast, no tonic GABA_A current is detected in the GABAergic NRT neurons of GAERS and their respective non-epileptic control strain (unpublished observation), as it is indeed the case in normal Wistar rats.²⁰

Figure 1

The enhanced tonic GABA_A current in TC neurons of rat and mouse genetic models of typical absence seizures results from a malfunction in the GABA transporter GAT-1. A1. Current traces from TC neurons of postnatal day (P) 14 and 17 non-epileptic control (more...)

GAT-1 Malfunction Underlies Increased Tonic GABA_A Inhibition

The enhanced tonic GABA_A current of TC neurons in genetic absence models is not due to increased vesicular GABA release, overexpression of extrasynaptic GABA_ARs, or mis-expression of synaptic GABA_ARs, but results from a malfunction of the GABA transporter GAT-1,17 despite it being far less abundant in the thalamus than GAT-3.²¹ As shown in Figure 1B1-2 in fact, i) block of GAT-1 (by the selective blocker NO711) in GAERS and stargazer has no effect on tonic current amplitude of TC neurons, ii) block of GAT-1 in the respective non-epileptic rats and mice increases tonic current amplitude to similar values to those seen in GAERS and stargazer mice, and iii) the compensatory increase in uptake by GAT-1 that is seen following block of GAT-3 (by the selective blocker SNAP5114) in non-epileptic animals is lost in both GAERS and stargazer mice.¹⁷ A malfunction in GAT-1 also underlies the increased tonic GABA_A current in TC neurons of lethargic mice.¹⁷ These data support and enlarge previous observations that had shown a reduced GABA uptake by GAT-1²² and an increased level of extracellular GABA²³ in the ventrobasal thalamus of GAERS compared to NEC. Differently from GAERS and stargazer mice, however, GABA transport by GAT-1 is reversed in lethargic mice, with this transporter being a source of ambient GABA. Interestingly, NO711 increases tonic GABA_A current by a similar amount in dentate gyrus granule cells of GAERS and NEC,¹⁷ indicating that GAT-1 activity is not compromised in a brain area that does not participate in the generation of typical absence seizures¹⁸ and where the distribution of this transporter is primarily neuronal.²⁴ Indeed, the tonic current of dentate gyrus granule cells is not different between GAERS and NEC.¹⁷

In summary, therefore, genetic models of typical absence seizures (i.e., GAERS, stargazer, and lethargic mice) show a brain region-specific enhancement of tonic GABA_A current, which in TC neurons is due to increased extracellular GABA level resulting from a malfunction in GABA uptake by astrocytic GAT-1.

Enhanced Tonic GABA_A Inhibition In The GHB Model: Role of GABA_BRs

Systemic (or intrathalamic) injection of γ-hydroxybutyric acid (GHB) is undoubtedly the best-established pharmacological model of typical absence seizures,^25,26 and systemic administration of 4,5,6,7-tetrahydroisoxazolo[5,4-c]pyridin-3-ol) (THIP), a selective agonist at δ subunit-containing extrasynaptic GABA_ARs,27 has been shown to elicit SWDs in normal animals.²⁸ Therefore, whereas it is not surprising that THIP dose-dependently enhances the tonic GABA_A current in TC neurons of normal Wistar rats,¹⁷ one would not expect GHB, which does not bind to GABA_ARs and is believed to elicit absence seizures by activation of GABA_BRs,26 to have an effect on the tonic GABA_A current. However, GHB dose-dependently enhances the tonic GABA_A current in TC neurons of normal Wistar rats¹⁷ (Figure 2A,B) at concentrations that are similar to the brain concentrations that elicit absence seizures in vivo.²⁹ This effect of GHB is not due to some aspecific action since it is abolished by the GABA_BR antagonist CGP55845 (Figure 2B). Interestingly, application of CGP55845 alone significantly reduces the tonic GABA_A current amplitude in TC neurons of Wistar rats to 74% of control¹⁷ (Figure 2B), indicating that facilitation of extrasynaptic GABA_ARs by GABA_BRs contributes approximately one quarter of the tonic GABA_A current in normal rats. Importantly, CGP55845 also reduces the tonic current in GAERS, stargazer and lethargic mice to about 55, 65 and 57%, respectively¹⁷ (Figure 2C), suggesting that facilitation of extrasynaptic GABA_AR function by GABA_BR activation contributes up to half of the tonic current in these genetic models.

Figure 2

GABA_BRs involvement in the tonic GABA_A current of TC neurons in genetic and pharmacological models of typical absence seizures. A. Current traces form TC neurons of normal Wistar rat showing the GHB-elicited increase in tonic GABA_A current and its block (more...)

In summary, therefore, a GAT-1 malfunction in thalamic astrocytes of mouse and rat genetic models leads to an increase in ambient GABA in the sensory thalamus, which in turn elicits an enhancement in tonic GABA_A inhibition through both direct activation of extrasynaptic GABA_ARs and indirect facilitation of extrasynaptic GABA_ARs via activation of GABA_BRs.

Enhanced Tonic GABA_A Inhibition Of TC Neurons Is Necessary And Sufficient For Typical Absence Seizure Generation

Assessing the impact of enhanced tonic GABA_A current of TC neurons on the expression of absence seizures requires experiments in freely moving animals, where both the behavioural and EEG components of the seizures can be studied without bias introduced by the concomitant use of anaesthetics and/or analgesics. In fact, unrestrained and undrugged GAT-1 KO mice, which have an enhanced tonic GABA_A current in TC neurons, express ethosuximide-sensitive typical absence seizures¹⁷ (Figure 3A,F). Similarly, direct injection of the selective GAT-1 blocker NO711 into the ventrobasal thalamus by reverse microdialysis in normal Wistar rats triggers ethosuximide-sensitive typical absence seizures¹⁷ (Figure 3D,F). Moreover, in δ GABA_AR KO mice, which exhibit a markedly reduced GABA_A inhibition in TC neurons, systemic administration of GHB fails to induce absence seizures¹⁷ (Figure 3F). Similarly, the intrathalamic injection of a δ subunit-specific antisense oligodeoxynucleotide in GAERS strongly decreases both the tonic GABA_A current and spontaneous seizures 1–2 days after injection, whereas a missense oligodeoxynucleotide has no effect¹⁷ (Figure 3E). Finally, intrathalamic administration of the δ subunit-specific agonist THIP in normal Wistar rats elicits absence seizures in a concentration-dependent manner, an effect that is reversed by systemic administration of ethosuximide¹⁷ (Figure 3C,F). Taken together, these data show that enhanced tonic GABA_A inhibition in TC neurons is both necessary and sufficient for the generation of typical absence seizures. In addition, they provide a mechanistic explanation for the aggravation of absence seizures that is observed in humans and experimental models following systemic and intrathalamic administration of drugs that increase GABA levels, including tiagabine, a GABA uptake blocker, and vigabatrine, a GABA transaminase blocker.^18,30–32

Figure 3

The tonic GABA_A inhibition in TC neurons is both necessary and sufficient for typical absence seizure generation. A. Bilateral (L = left, R = right hemispheres) EEG traces from a freely moving GAT-1 knockout (KO) mouse showing spontaneous SWDs (spectrogram (more...)

Significance Of Tonic GABA_A Inhibition In Typical Absence Epilepsy

The discovery of an increased tonic GABA_A current in TC neurons represents the first abnormality that i) is common to both well-established pharmacological and genetic models of typical absence seizures, irrespective of species and known up-stream mutations, ii) is manifested both before and after the developmental onset of seizures, iii) contributes to the exquisite sensitivity of experimental absence seizures to selective GABA_BR agonists and antagonists, and iv) explains the aggravating effects of GABAergic drugs in both human and experimental absence seizures. Importantly, these results, together with the presence of powerful GABA_A IPSPs in the majority of TC neurons during absence seizures in vivo^33,34 (Figure 4A1–2) (see below), lead to another significant conclusion, i.e., that studies aiming at reproducing typical absence seizures by indiscriminatly blocking GABA_ARs of TC neurons35–38 are inherently flawed.

Figure 4

Intracellular correlates of TC and NRT neuron activity during spontaneous SWDs in vivo. A1. Intracellular recording (bottom trace) from a TC neurons in GAERS in vivo during a spontaneous SWD (top trace) shows the presence of bursts of IPSPs occurring (more...)

The finding that the increased tonic GABA_A inhibition in TC neurons of genetic models is due to a malfunction of GAT-1 shifts the emphasis from a neuronal to an astrocytic aetiology of absence epilepsy. The impaired GAT-1 activity in GAERS is not caused by a decreased expression of GAT-1 mRNA or protein levels in either thalamus or cortex.¹⁷ Also, GAT-1 cDNA from GAERS, stargazer and lethargic mice presents no genetic variants, whereas the mutations responsible for absence seizures in stargazer and lethargic mice are not present in GAERS.¹⁷ Potential abnormalities that may lead to a reduced GAT-1 function, but have not yet been tested, include the inability of this protein to reach the outer astrocytic membrane, an alteration in its subcellular location and/or abnormalities in its phosphorylation. Similalry, we are aware of no study that has investigated GAT-1 (or GAT-3) genetic variants in human CAE.

Experimental typical absence seizures can be elicited or aggravated by selective GABA_BR agonists and can be blocked by selective GABA_BR antagonists, applied either systemically or intrathalamically.^39–41 Because about 50% of the tonic GABA_A current observed in TC neurons of GAERS, stargazer and lethargic mice is abolished by a GABA_BR antagonist17 (Figure 2C), the behavioural and EEG effects of GABA_BR-selective drugs on typical absence seizures can no longer be simply explained by the ability of these drugs to affect GABA_B-mediated IPSPs and/or presynaptic GABA_BRs, but should also take into account the positive modulation by GABA_BRs of the tonic GABA_A inhibition in TC neurons.

Phasic GABA_A Inhibition In Cortex

Among the various mutations in GABA_AR subunits, those that have been identified in related or unrelated CAE patients are: α1(S326fs328X), β3(P11S), 3(S15F), β3(G32R), γ2(R43Q), γ2(R139G) and γ2(IVS6+2T•G).^12–15,42 Functional analysis of these mutant proteins expressed in heterelogous systems has shown that all of them bring about a marked decrease in GABA responses.¹² Importantly, humans carrying the γ2(R43Q) mutation have been shown to have a decreased short-interval intracortical inhibition assessed with transcranial magnetic stimulation, leading to cortical facilitation.⁴³ γ2(R43Q) mutant mice constructed by homologous recombination express spontaneous ethosuximide-sensitive absence seizures.⁴⁴ Moreover, despite the fact that the γ2 subunit has an almost ubiquitous expression and is apparently required for functional synaptic GABA_ARs, these mice exhibit a brain region-specific alteration in phasic GABA_A inhibition, since cortical layer 2/3 pyramidal cells but not TC or NRT neurons show a reduction in miniature IPSCs (mIPSCs) compared to age-matched wildtype littermates.⁴⁴ Whether the region-specificity of this malfunction is also present in humans with the γ2(R43Q) mutation, or only occurs in mice because of some species-specific promoters or silencers, is not known at present. Nevertheless, together with showing the physiological effect of a human genetic variant in a “true” physiological system as opposed to hetereologous expression systems, the key significance of this finding is that it is the first clear evidence that the downstream abnormalities of a human genetic variant in CAE can be inhibition that brain region-specific. Importantly, the cortex-specific abnormality in phasic GABA_A results from the human γ2(R43Q) mutation exquisitively complements previous experimental data showing that 1) in felines in vivo the localized block of GABA_ARs by cortical application of inhibition in the thalamus, does elicit SWDs,³³ and bicuculline, which leaves intact phasic GABA_A inhibition in TC neurons is either unaffected or 2) in rat and mouse genetic models phasic GABA_Aincreased compared to their respective non-epileptic controls^17,44–47 (see below).

On the basis of indirect evidence, a decreased cortical GABAergic inhibition has also been suggested to occur in layer 2/3 regular spiking neurons⁴⁸ as well as in layer 5 pyramidal neurons⁴⁹ of adult Wistar Albino Glaxo/Rij (WAG) rats, another well established polygenic model of typical absence seizures.⁵⁰ However, no change is observed in mIPSCs in pyramidal cell and interneurons of cortical layers 2–3 of young, pre-seizure GAERS,⁴⁵ and cortical GABAergic inhibition is intact in the feline generalized penicillin epilepsy model.⁵¹

In summary, the evidence from a human mutation indicates that a decreased phasic GABA_A inhibition in cortex characterizes typical absence seizures, though further studies are needed to clarify the extent of this decrease both in terms of different cortical regions, layers and neuronal types: this additional work might shed light on the existing discrepancies on changes in phasic GABA_A inhibition among experimental models.

Phasic GABA_A Inhibition In TC Neurons

A reduced or absent phasic GABA_AR function in the thalamus also characterizes the current view of the pathophysiology of typical absence seizures. While this may be partly true for the intra-NRT inhibition it certainly does not hold for TC neurons. Thus, both in felines that show spontaneous or cortical bicuculline-induced SWDs³³ as well as in the well-established GAERS model³⁴ the vast majority (60 and 94%, respectively) of TC neurons recorded in vivo during SWDs exhibit bursts of GABA_A IPSPs, each tightly synchronized with the EEG spike and wave complex (Figure 4A1–2). Indeed, the rise time, amplitude, frequency and decay time constant of mIPSCs and spontaneous IPSCs (sIPSCs) measured in TC neurons in vitro are not different between GAERS and NEC both prior to,⁴⁵ and after spontaneous seizure manifestation¹⁷ (i.e. postnatal days 12–25, and postnatal day 30, respectively), and paired pulse depression of evoked IPSCs is similar between the two strains⁴⁵ (Figure 5B3). As mentioned earlier, no change in TC neuron IPSCs occurs in mice carrying the human γ2(R43Q) mutation,⁴⁴ and no change in IPSC properties has been detected in TC neurons of β3 KO mice⁵² which show absence seizures as part of a much more complex neurological phenotype. Also, in lethargic, stargazer and tottering mice sIPSCs in TC neurons are similar to those in their respective littermates,¹⁷ and no difference has been found in evoked IPSCs in lethargic and tottering mice compared to control mice.⁴⁷ Finally, an increase in mIPSCs frequency, potentially leading to enhanced phasic inhibition, is observed in TC neurons of the absence seizure-prone DAB/2J mouse strain.⁴⁶

Figure 5

Phasic GABA_A inhibition in TC and NRT neurons of genetic and pharmacological models of typical absence seizures. A1. GHB reversibly decreases the amplitude of a cortical EPSC recorded in TC neurons. A2. Example of an NRT-derived IPSC (recorded in a TC (more...)

In a similar manner to the genetic models, in the best-established pharmacological model of absence seizure, i.e. the GHB model, a net increase in phasic GABA_A inhibition is observed in TC neurons of the ventrobasal thalamus.⁵³ This is because whereas GHB reversibly and dose-dependently decreases the amplitude of all sensory⁵⁴ and corticothalamic EPSCs⁵³ (Figure 5A1), the two main excitatory drives to TC neurons, at low concentrations (250 μM-1 mM) it only affects the GABA_A IPSCs in some but not all TC neurons⁵³ (Figure 5A2–3). In particular, 250 μM GHB produces an effect in <10% of tested TC neurons, whereas at 0.5-1 mM it only decreases half of the recorded IPSCs (Figure 5A4–5). Thus, at brain concentrations (250 μM - 1 mM) similar to those that elicit absence seizures in vivo GHB brings about not only the previously described GABA_BR- and K+-dependent hyperpolarization^26,55 and an increase in tonic GABA_A inhibition¹⁷ (see above) but also an imbalance between the excitatory and inhibitory drive to TC neurons, leading to a net increase in their phasic GABA_A inhibition. This, in turn, will help to impose that periodic phasic inhibition that sculptures the electrical behaviour of TC neurons during SWDs as observed in vivo in feline⁵⁶ and rat⁵⁷ models (Figure 4A1–2).

On the basis of all this evidence from both genetic and pharmacological models, as well as from mice expressing the human γ2(R43Q) mutation, it is therefore surprising that the data in support of the presence of an enhanced or unchanged phasic GABAergic function in TC neurons have been selectively dismissed over the last 15 years, such that textbooks and topical reviews^58–62 still present the pathophysiological mechanisms of TC neuron activity during typical absence seizures as the one that is observed in vitro during application of a GABA_AR antagonist.^35,36,38 It is also surprising that the in vivo evidence which is often used to support this view is the ability of bicuculline (or other GABA_AR antagonists) to generate, in ketamine-xylazine anaesthetized rats, strong field potentials at 3 Hz³⁷, that are claimed to represent SWDs even if their sensitivity to ethosuximide or other anti-absence drugs was not tested. Indeed, it is highly unlikely that these field potentials represent SWDs, since the ketamine-xylazine combination, as other general anaesthetic, are known to abolish the SWDs of any well-established in vivo model of absence epilepsy.^18,51,63

In summary, therefore, the presence of these serious issues of interpretation, together with the evidence showing an increased tonic and an unchanged/increased phasic GABA_A inhibition in TC neurons of genetic and pharmacological models, strongly support the conclusion that the block of thalamic GABA_ARs that continues to be used in many in vitro and some in vivo studies does at best elict some form of thalamic hyperexcitability of as yet unknown pathophysiological relevance, but definitely not the activity that is present during typical absence seizures.

Phasic GABA_A Inhibition In NRT Neurons

The existance of contradictory data makes it difficult at present to draw a unifying picture of the changes in the phasic GABA_A inhibition of NRT neurons that are associated with an absence seizure phenotype. Starting from inbred genetic models, mIPSCs recorded from NRT neurons of pre-seizure GAERS show a 67% higher frequency, a 25% larger amplitude and a 40% faster decay than those in age-matched NEC.⁴⁵ Moreover, paired-pulse depression of evoked GABA_A IPSCs is significantly smaller (46%) in NRT neurons of GAERS than NEC⁴⁵ (Figure 5B1–2). This smaller paired-pulse depression of IPSCs in GAERS would undoubtedly ensure a consistent hyperpolarizing drive to help promote the expression of the strong, low threshold Ca²⁺ potential-mediated bursts of action potentials that are the hallmark of NRT neuron firing at each spike and wave complex in vivo^56,57 (Figure 4B). Moreover, an increase in mIPSC frequency, potentially leading to enhanced phasic GABA_A inhibition, is observed in NRT neurons of the absence seizure-prone DAB/2J mouse strain.⁴⁶ Finally, an almost complete disappearance of the α3 subunit, that in the thalamus is selectively expressed in NRT neurons⁶⁴, has been reported in these GABAergic neurons from WAG rats,⁶⁵ suggesting a potential loss of function.

Surprisingly, however, α3 KO mice do not show spontaneous absence seizures and indeed exibit a small reduction in GHB-elicited seizures compared to wildtype littermates, a result that has been interpreted as resulting from a powerful compensatory gain in phasic GABA_A inhibition in NRT neurons.⁶⁶ On the other hand, since unknown mechanisms lead to this unexpected compensatory increase in NRT GABA_A inhibition following deletion of the α3 subunit, the lack of spontaneous absence seizures and the relative resistance to pharmacologically induced absence seizures may also be due to additional compensatory changes in phasic (or tonic?) GABA_A inhibition in cortex, or tonic inhibition in thalamus, all of which were not investigated in that study.⁶⁶ Further evidence for a pro-absence role of a decreased phasic GABA_AR function in NRT neurons has also come from the observation of an increase in high frequency discharges at 3 Hz in β3 subunit KO mice, which show a massive reduction in sIPSP frequency and amplitude in NRT, but no change in TC neurons.⁵² Since these β3 KO mice exhibit a large variety of neurological deficits and are considered a model of Angelman’s syndrome, it is difficult to unequivocally assign a causative role for this decreased intra-NRT inhibition in typical absence seizures.

In summary, whereas in inbred models and models with spontaneous mutations there is either an increase or no change in intra-NRT phasic GABA_A inhibition, data from two transgenic mice suggest that a decrease in this NRT synaptic function has a pro-absence effect. Thus, it may be tempting to conclude that abnormalities in intra-NRT phasic GABA_A inhibition are not a necessary condition for the expression of typical absence seizures, a view supported by the lack of changes in NRT IPSCs of mice carrying the human γ 2(R43Q) mutation⁴⁴. In this respect, it is worth noting that the anti-absence effect of clonazepam in humans and experimental models^67,68 may not be solely due to its ability to increase phasic GABA_A inhibition in NRT,⁶⁹ since α3 subunit-containing GABA_ARs that are the target of this benzodiazepine are also present in neocortical neurons.⁷⁰