An official website of the United States government
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Noebels JL, Avoli M, Rogawski MA, et al., editors. Jasper's Basic Mechanisms of the Epilepsies. 5th edition. New York: Oxford University Press; 2024. doi: 10.1093/med/9780197549469.003.0020
Thirty percent of children with absence seizures are pharmaco-resistant, and 60% suffer from neuropsychiatric comorbidities that often persist even after full pharmacological control of the seizures. This highlights the need for a detailed comprehension of the cellular and network mechanisms of these nonconvulsive seizures. Generally, network hyperexcitability and hypersynchrony underlying seizure ictogenesis are thought to originate from impaired inhibition or enhanced excitation. In absence seizures, there is a markedly enhanced synchrony in cortico-thalamic and cortico-basal ganglia-thalamic networks, but solid evidence from genetic animal models indicates that at the single-cell and neuronal population levels GABAergic inhibition is generally increased while excitation is mostly either unchanged or decreased. Here, recent results on intrinsic conductances and network mechanisms within cortico-thalamic and cortico-basal ganglia-thalamic circuits are highlighted that support this view.
Absence seizures are sudden and brief lapses of consciousness associated with lack of voluntary movements and spike-wave discharges (SWDs) at 2.5–4 Hz in the EEG (Crunelli and Leresche, 2002; Blumenfeld, 2005; Crunelli et al., 2020). The current failure of existing treatments to control the pharmaco-resistance (Glauser et al., 2013) and comorbidities (Masur et al., 2013; Holmes and Noebels, 2016) of absence seizures can only be addressed by a deeper understanding of the cellular and network mechanisms involved in their epileptogenesis and ictogenesis.
Seizures are defined as the transient occurrence of signs and/or symptoms due to abnormal, excessive, or synchronous neuronal activity (Scheffer et al., 2017), which in turn depends on the excitation-inhibition balance of the different neuronal populations underlying the paroxysmal activity. Thus, seizures are generally viewed as resulting either from decreased or dormant inhibition (Bekenstein and Lothman, 1993) and/or enhanced activity of excitatory neurons (Chen et al., 2003; Khosravani et al., 2004; Powell et al., 2009; Avoli, 2012), leading to an overall increase in the firing output and synchrony of the epileptic networks.
Here, we review recent evidence on intrinsic neuronal conductances (mainly T-type Ca2+ and cyclic nucleotides-gated hyperpolarization-activated, HCN, channels) and network mechanisms of absence seizures that impact the excitability and synchrony of the involved cortical and thalamic neuronal populations. The emerging picture highlights marked differences with convulsive seizures: in particular, absence seizures in genetic models show a decrease or no change in the mean ictal activity of different excitatory neuronal populations and the concomitant ictal increase in the firing of many inhibitory neuron subtypes.
Various cortico-thalamic and cortico-hippocampal networks engage in rhythmic activities serving many physiological or pathological neuronal dynamics from sensory coding (Engel et al., 2001), sleep (Steriade, 2006; Vyazovskiy et al., 2011) and spatial navigation (Buzsaki, 2011) to paroxysmal activities (Beenhakker and Huguenard, 2009; Crunelli et al., 2020). Such rhythmic activity relies on a combination of both intrinsic membrane conductances and network interactions. In particular, hippocampal CA3 (Wong and Prince, 1978) and CA1 (Jensen et al., 1996) neurons, cortical pyramidal neurons (Connors et al., 1982; Gutnick et al., 1982; McCormick et al., 1985; Gray and McCormick, 1996), and thalamocortical (TC) neurons (Llinás and Jahnsen, 1982; Hughes et al., 2002; Errington et al., 2012; Connelly et al., 2015, 2016, 2017) possess special somato-dendritic conductances that allow them to generate different rhythmic activities during normal physiological functions.
Among rhythmic activities, those generated by burst firing are ideal candidates for promoting synchrony across large neuronal assemblies as they deliver multiple presynaptic inputs to target neurons (Sakmann, 2017), with intervals that would allow the underlying network to counteract this transient activity, and thus leaving the opportunity to modulate the excitability during a short time window. In particular, the firing mode (i.e., tonic or burst firing) of TC neurons is membrane potential-dependent and includes low threshold spike (LTS)-mediated burst firing (Llinás and Jahnsen, 1982), during periods of membrane potential hyperpolarization, and tonic firing or high-threshold burst firing when the membrane potential is depolarized (Hughes et al., 2002, 2004; Lorincz et al., 2009; Molnár et al., 2021). This membrane potential-dependent shift in firing mode is also reliant on the behavioral state of the animal (Hirsch et al., 1983; Molnár et al., 2021) and thus influenced by various neuromodulatory inputs (Reimer et al., 2016).
Early in vitro work related to absence seizures showed that in thalamic slices (where GABAA receptors had been blocked) hyperexcitability dominated the activity of both TC and nucleus reticularis thalami (NRT) neurons (Von Krosigk et al., 1993; Bal et al., 1995a, 1995b; McCormick and Contreras, 2001) since both neuronal types exhibited pronounced LTS-mediated burst firing. Moreover, in anaesthetized Genetic Absence Epilepsy Rats from Strasbourg (GAERS), a well-established genetic model of absence seizures (Depaulis et al., 2016), both cortico-thalamic and NRT neurons were found hyperexcitable (Slaght et al., 2002; Polack et al., 2007; Chipaux et al., 2011; Williams et al., 2016), but a tonic hyperpolarization and/or shunting dominated the ictal dynamics of the majority of TC neurons, leading to an overall decrease in their action potential output during absence seizures (Steriade and Contreras, 1995; Pinault et al., 1998, 2006).
In contrast to the results obtained in vitro and in anesthetized animal models, electrophysiological recordings and Ca2+ imaging of different neuronal populations in cortico-thalamic networks of freely moving GAERS rats and nonanesthetized head-restrained stargazer mice, another genetic model of absence seizures (Crunelli and Leresche, 2002), have shown no evidence for hyperexcitability of principal neurons or compromised inhibition of interneurons or long-projecting GABAergic neurons. Specifically, during absence seizures GABAergic NRT neurons increase their average firing (McCafferty et al., 2018), whereas TC neurons of the somatosensory thalamus and principal neurons of the primary somatosensory cortex (McCafferty et al., 2023) decrease their total (i.e., tonic and burst) firing. Moreover, ictal activity of both principal neurons and GABAergic interneurons of the primary visual cortex remains mainly unaltered compared to interictal periods (Meyer et al., 2018). Importantly, in contrast to the results of in vitro and in anaesthetized animal models, ictal LTS-mediated burst firing constitutes a minor component of the TC neuron ictal firing in freely moving animal models and selective block of T-type Ca2+ channels (and thus LTS-mediated burst firing) in these neurons has no effect on absence seizures (Fig. 20–1) (McCafferty et al., 2018). In sharp contrast, both burst firing and the mean firing rate of NRT neurons markedly increase ictally (Fig. 20–1) (McCafferty et al., 2018), an effect that is most likely driven by the gain-of-function mutations in CaV3.2 channels that are present both in GAERS rats (Casillas-Espinosa et al., 2017) and children with absence seizures (International League Against Epilepsy Consortium on Complex Epilepsies, 2018). Notably, selective block of T-type Ca2+ channels in NRT (and cortical neurons) markedly reduces the number of absence seizures and decreases the frequency of the remaining ones (see next section), demonstrating the necessary and pivotal role of these neurons’ burst firing in the generation of these nonconvulsive seizures (Fig. 20–1) (McCafferty et al., 2018). Thus, LTS-mediated burst firing of TC neurons is not an important factor for determining the synchrony of the thalamic output during absence paroxysmal activity as previously thought. Instead, it is the cortically driven feed-forward inhibition that frames the firing of TC neurons (Fig. 20–1) (McCafferty et al., 2018), thus bringing about the strongly synchronous thalamic output to the neocortex.
Impact of T-type Ca2+ and HCN channels on absence seizures. This schematic diagram summarizes the changes in expression and/or current density of T-type Ca2+ and HCN channels in the neocortex and thalamus. The consequences of these alterations on burst (more...)
Early work on SWDs ictogenesis in vitro had suggested HCN channels as key contributors to the intrinsic excitability of thalamic and cortical neurons (Pape and McCormick, 1989; Phillips et al., 2016). The loss-of-function of dendritic HCN1 channels in layer 5 cortical pyramidal neurons of WAG/Rij rats, a genetic model of absence seizures (Coenen et al., 1992), is associated to high-frequency burst firing and to a synchronized cortical output in these epileptic animals (Fig. 20–1) (Kole et al., 2006). TC neurons in the thalamic ventrobasal complex (VB) of GAERS have increased Ih (Cain et al., 2015), and TC neurons of WAG/Rij rats show a gain of function of Ih and a negative shift in its steady-state activation (Wemhöner et al., 2015) as well as a decreased cAMP binding to HCN channels which likely promotes ictogenesis (Fig. 20–1) (Kuisle et al., 2006). Moreover, TC neurons in the lateral geniculate nucleus of WAG/Rij rats have an increased expression of HCN1, HCN2, and HCN4 (but a decrease of HCN3) and reduced cAMP sensitivity that overall give rise to an Ih with higher density, hyperpolarized voltage dependence and faster activation kinetics (Kanyshkova et al., 2012). This increased HCN function could explain the reduced LTS-mediated burst firing of TC neurons in vivo during absence seizures as the increased Ih conductance would reduce hyperpolarization (Cain et al., 2015). Moreover, in normal, nonepileptic animals VB-selective knockdown of HCN4 fails to induce absence seizures. In contrast, VB-selective HCN2 knockdown and global HCN1 (Nishitani et al., 2019) or HCN2 (Ludwig et al., 2003) knockout bring about absence seizures. Thus, thalamic HCN2 channels have an anti-absence effect (Hammelmann et al., 2019), but thalamic HCN4 channels have a pro-absence effect (Zobeiri et al., 2019). It could be speculated that in animal models of absence epilepsy, thalamic HCN4 channels might have a more substantial contribution to the total Ih of TC neurons than HCN2 channels since VB-selective knockdown of HCN2 elicits ASs (Hammelmann et al., 2019), but VB-selective knockdown of HCN4 does not (Zobeiri et al., 2019). Indeed, TC neurons of HCN4 knockout mice show a reduction of evoked intrathalamic oscillations in vitro (Zobeiri et al., 2019), but TC neurons of GAERS rats (Cain et al., 2015) and of nonepileptic mice that develop atypical ASs following a cortical infarct (Paz et al., 2013) show an increased Ih. In addition, genetic and pharmacological block of HCN channels in the VB and somatosensory cortex of various rat and mouse models markedly suppresses absence seizures (David et al., 2018; Iacone et al., 2021). Notably, systemic injection of a pan-HCN channel blocker (Ivabradine) has a dose-dependent anti-absence effect in GAERS rats and a stronger action when applied locally in the somatosensory cortex than in the VB (Iacone et al., 2021).
The consistent picture emerging from different studies in various genetic models of absence seizures indicates that the ability to generate LTS-mediated burst firing is reduced in TC neurons (through adaptive processes like an increased Ih) but is increased in NRT neurons (because of these cells’ enhanced T-type Ca2+ current density) and in neocortical layer V pyramidal neurons (because of their smaller Ih). Moreover, T-type Ca2+ channel-mediated burst firing of NRT and cortical neurons is essential for absence seizure generation, but that of TC neurons is not. Although the differences in the contribution of various HCN channel isoforms to absence seizures is still not fully understood, HCN channels have an overarching pro-absence effect, so that blockers of these channels may be considered for future anti-absence drug development.
In conjunction with intrinsic neuronal conductances, synaptic interactions are also of paramount importance in controlling network excitability. While signal propagation relies on excitation, feedforward and/or feedback inhibition is important for controlling its strength and timing. Indeed, the high temporal coincidence of excitatory and inhibitory synaptic conductances has been well documented during both sensory coding (Borg-Graham et al., 1998; Wehr and Zador, 2003, 2005; Wilent and Contreras, 2004; Haider et al., 2013) and spontaneous cortical activity (Haider et al., 2006; Okun and Lampl, 2008; Atallah and Scanziani, 2009). Moreover, a temporally restricted increase in network synchrony can be advantageous for various physiological functions (i.e., perception, memory, or attention) and constitutes the mechanism underlying various brain oscillations (Buzsaki, 2011).
Synchrony can result from hyperexcitability, but inhibition is also known to play a paramount role in generating synchrony by framing the activity of excitatory neurons in the temporal domain (Roux and Buzsáki, 2015; Cardin, 2018). For instance, during hippocampal sharp-wave ripples, that is, brief (~50 ms), high-frequency (100–200 Hz) network oscillations that have been implicated in memory replay and consolidation (Buzsáki, 2015), the permissive state of CA1 networks leads to synchronous firing among CA1 pyramidal neurons entrained by CA3 inputs. However, whereas this enhanced excitability during physiological functions is tightly controlled in the temporal domain by inhibition, a failure of inhibition can lead to epileptiform events (Karlócai et al., 2014). Indeed, reduced hippocampal inhibition (Bekenstein and Lothman, 1993) results in the cortical hyperexcitability that characterizes temporal lobe epilepsy (Badawy et al., 2015; de Curtis and Avoli, 2016; Avoli, 2019).
Thus, ictal activity is generally viewed as resulting from a shift or a loss of the excitation-inhibition balance due to increased excitation and/or reduced inhibition. In contrast to focal and generalized convulsive seizures, however, the overwhelming picture in absence seizures is that of an ictal decrease (or no change) in excitation and an ictal increase (or no change) in inhibition. In the thalamus, a prominent 50% decrease characterizes the activity of the excitatory TC neurons during absence seizures compared to interictal periods (McCafferty et al., 2018). This reduced firing results from the increased activity of both the main intrathalamic inhibitory input, originating from the GABAergic NRT neurons (McCafferty et al., 2018), and the extrathalamic inhibitory input, originating from the GABAergic neurons of the substantia nigra pars reticulata (SNr) (Fig. 20–2) (Paz et al., 2007; Arakaki et al., 2016). This increased extrathalamic inhibition comes about as a result of the rhythmic SWD-locked firing of cortico-striatal neurons that preferentially excites striatal fast-spiking neurons, which in turn inhibit the striatal output cells, that is, the GABAergic medium spiny neurons (Slaght et al., 2004; Paz et al., 2007), leading to an ictal disinhibition of the SNr GABAergic neurons (Deransart et al., 2003). Although SNr GABAergic neurons are known to predominantly target the motor thalamus, recent studies have provided solid evidence for a more widespread SNr→thalamus innervation. Specifically, using genetically targeted viral tracing and whole-brain anatomical analysis, more than 40 thalamic nuclei (including ventro-medial, ventroanterior, mediodorsal, centro-lateral/medial, and parafascicular nuclei as well as the thalamic reticular nucleus and zona incerta) were identified as targets of electrophysiologically specialized and topographically organized of SNr GABAergic neuron populations (McElvain et al., 2021). Notably, optogenetic stimulations of SNr neurons resulted in the inhibition of VB neurons via GABAergic synapses, excitation of NRT neurons, and excitation and inhibition of PO neurons (Antal et al., 2014). Morover, a recent study found that a targeted reduction in excitatory transmission from the neocortex to striatal fast-spiking interneurons could trigger SWDs (Miyamoto et al., 2019). Overall, therefore, these results indicate that the basal-ganglia system can be directly involved in the generation of SWDs.
Left: GABAergic function is highly recruited in cortex ad thalamus during absence seizures. In the neocortex, the phasic GABAergic synapses provide feedforward control of excitation (Panthi and Leitch, 2019), whereas the role of tonic inhibition (more...)
The reciprocal pace making of TC and NRT neurons (observed in in vitro studies when GABAA receptors are blocked) (Von Krosigk et al., 1993) is not a key factor in controlling the rhythmic ictal thalamic output to the cortex, since, as shown in freely moving GAERs rats, the ictal synchrony of the TC neuron population is largely determined by the cortically driven feed-forward inhibition (McCafferty et al., 2018), either directly by NRT neuron excitation and indirectly via the basal ganglia-thalamic connections (Depaulis and Charpier, 2018). Indeed, the highly precise time-locking of the TC neuron ictal firing, and thus of the thalamic output, is controlled not only by the phasic component of GABAergic inhibition but most probably also by its tonic component (Fig. 20–2) (Cope et al., 2009). In fact, a recent study (Kwak et al., 2020) has shown that an enhanced tonic GABAA inhibition not only decreases the overall firing activity of TC neurons but also reduces the spike jitter of their response to (rhythmic) synaptic inputs in normal nonepileptic animals. Though this finding needs to be directly confirmed in absence seizure models that have a constitutively higher tonic GABAA inhibition (Cope et al., 2009), it provides indirect evidence of the potential contribution made by this form of GABAergic inhibition in temporally sharpening the thalamic output to the neocortex. Ultimately, the ictal increase in the synchrony of the thalamic output may contribute to the amplitude of cortical (and EEG) SWDs, which is now recognized as a biomarker of absence seizure severity in humans (Guo et al., 2016).
Differently from the thalamus and the basal ganglia, a comprehensive and systematic analysis of the ictal firing of different excitatory and inhibitory neuronal populations that are present in different neocortical layers is still missing. Notwithstanding, the solid data that are already available demonstrate that the mean firing rate of excitatory (pyramidal) neurons in primary somatosensory and visual cortices is mainly decreased or unchanged during absence seizures compared to interictal periods (McCafferty et al., 2018; Meyer et al., 2018). Moreover, both somatostatin- and parvalbumin-containing neocortical GABAergic interneurons were reported to decrease their activity ictally (Meyer et al., 2018). The presence of different ictal firing rate changes is not surprising since in normal nonepileptic animals a decreased thalamic drive has been shown to lead to both cortical excitation (Haider et al., 2013) or inhibition (Poulet et al., 2012), but whether the former or the latter prevails during absence seizures has not been directly and systematically investigated in different cortical populations.
Nonetheless, there are indications that cortical neurons in the infragranular layers are hyperexcitable, but not necessarily hyperexcited during interictal states (Polack et al., 2007). The loss-of-function of HCN1 channels in layer 5 neocortical pyramidal neurons observed in an absence epilepsy model (Kole et al., 2006) could strengthen the functional coupling between supragranular and infragranular layers, leading to a more synchronous activity across different cortical layers and areas and ultimately to larger amplitude SWDs. Indeed, a more intense HCN channel function, as it occurs during active wakefulness because of the HCN channel upregulation by noradrenaline during this vigilance state (Phillips et al., 2016), may also represent a contributing factor to the more likely presence of absence seizures during quiet than active wakefulness (Coenen et al., 1992). Moreover, GABAergic inhibition is involved in the resonant properties of cortical networks in the 5–10 Hz frequency band (Stark et al., 2013). This property is more prominent in deep layers (David et al., 2021), suggesting that these layers can control the rhythmicity of SWDs. In contrast, layer 2/3 circuits are preferentially wired to promote the propagation of fast activities due to predominant feedforward mechanisms and oscillating properties of neurons in these upper layers (Arnal and Giraud, 2012). In addition, the largest number of VIP neurons can be found in layer 2/3 (Prönneke et al., 2015), and these neurons could easily veto the PV-mediated inhibition (Pfeffer et al., 2013), leading to hyperexcitability and increased synchrony.
Synaptic interactions involved in absence seizures result in a cortical drive of TC neurons that together with their enhanced tonic GABAA inhibition reduces the impact of sensory inputs. Moreover, these ictal network interactions favor activity in infragranular cortical layers, while inhibiting supragranular layers. Whereas it is probably trivial to try and identify one single synapse type responsible for the expression of absence seizures, there are some choke points in their underlying generating networks that await identification (Paz and Huguenard, 2015).
One key unanswered question in the absence seizure research field is the mechanism underlying the higher difference in the SWD frequency of rodent models (4–11 Hz) compared to that of humans (2.5–4 Hz). As discussed earlier, the intrinsic and synaptic mechanisms of the networks that underlie paroxysmal oscillations could likely control the SWD frequency, and there is evidence that the frequency of many physiological oscillations (e.g., sleep slow waves and theta oscillations) is higher in rodents than in humans (Mölle et al., 2009; Jacobs, 2014). Notably, our recent studies have unraveled some thalamic and cortical intrinsic properties that can impact the frequency of SWDs. Thus, decreasing NRT T-type Ca2+ channel function (McCafferty et al., 2018) or cortical Ih (Iacone et al., 2021) in rodent models reduces the frequency of their SWDs to values close to those in humans. Moreover, network resonant properties, likely due to differences in ion channel distribution (Kalmbach et al., 2018; Rich et al., 2021), as well as differences in synaptic connectivity, may explain the frequency differences. Moreover, the frontal versus parietal location of the cortical initiation network in humans versus animal models (Meeren et al., 2002; Bai et al., 2010), and the known gradient of predominant physiological frequencies in the fronto-occipital axis (Werth et al., 1997), could also account for these differences in SWD frequency.
As other seizures types, absence seizures are brought about by an increased synchrony of the underlying networks. In contrast to the ictal processes underlying convulsive seizures, however, absence seizures are mainly characterized by increased firing of different inhibitory neuronal populations with mostly unchanged or decreased activity of excitatory neurons. The intrinsic neuronal conductances and synaptic network mechanisms underlying absence seizures tend to converge toward infragranular cortical layers, exerting a driving role on thalamic nuclei which are under the influence both of intense intra- and extra-thalamic inhibition.
Work on absence seizures in our labs is supported by the Ester Floridia Neuroscience Research Foundation (grants to VC), the Hungarian Scientific Research Fund (grants NN125601 and FK123831 to MLL), the Hungarian Brain Research Program (grant KTIA_NAP_13-2-2014-0014 to MLL), and the Hungarian Ministry of Human Capacities, (grant 20391-3/2018/FEKUSTRAT to MLL). MLL is a grantee of the János Bolyai Fellowship.
The authors declare no conflict of interest.
This is an open access publication, available online and distributed under the terms of a Creative Commons Attribution-Non Commercial-No Derivatives 4.0 International licence (CC BY-NC-ND 4.0), a copy of which is available at https://creativecommons.org/licenses/by-nc-nd/4.0/. Subject to this license, all rights are reserved.
Your browsing activity is empty.
Activity recording is turned off.
See more...
