Wrestling with epilepsy; potential roles for kainate receptor SUMOylation in regulating neuronal excitability

Epilepsy is defined as the tendency to have recurrent seizures caused by sudden bursts of synchronized electrical activity in the brain. In the UK, approximately one in 20 people will have a single seizure at some point in their life, whereas around one in 130 people have epilepsy [101]. Epilepsy continues to be a serious problem even though neurologists are now equipped with an unprecedented number of antiepileptic drugs. In fact, approximately 30% of patients are resistant to currently available therapies and experience uncontrolled seizures. The introduction of several new antiepileptic drugs has not changed this unfavorable outcome since most of these ‘new‘ drugs act on a limited number of ‘old‘ targets. Indeed, the majority of currently available antiepileptic drugs either inhibit voltage-gated sodium/calcium channels or promote inhibition mediated by GABA_A receptors [1]. This narrow scope of biological activity translates not only into limited efficacy but also considerable untoward effects, hampering successful epilepsy therapy.

Excitatory ionotropic glutamate receptors (GluRs) have been considered as potential targets of antiepileptic drugs for a long time [2]. GluRs are integral membrane proteins that mediate the majority of excitatory neurotransmission in the brain. They are classified into three main families, named after their selective agonists: N-methyl-d-aspartate receptors (NMDARs), amino-3-hydroxy-5-methyl-4-iso-xazole propionic acid receptors (AMPARs) and kainate receptors (KARs). Synaptic release of glutamate activates these receptors and permits sodium and calcium influx resulting in excitatory postsynaptic potentials. Under strong activation this leads to the generation of action potentials and excitation of the postsynaptic cell [3]. Therefore, they play key roles in mechanisms responsible for seizure generation, and antagonists of various GluRs have powerful anticonvulsant properties in animal models of epilepsy [2]. However, translation of these promising results into clinical benefit has proven problematic. NMDAR antagonists lack clinical efficacy and produce serious side effects, including neurotoxicity. Clinical trials with AMPAR antagonists also failed to deliver robust efficacy at doses devoid of neurological side effects [4]. The observed low safety margin could be explained at least in part by the ubiquitous presence of both NMDARs and AMPARs in virtually every neuron of the brain. Thus, inhibition of these receptors ultimately leads to interference with important physiological processes resulting in central side effects, such as dizziness, sedation, psychosis or memory problems [5]. In this respect, KARs that are developmentally regulated and have somewhat restricted distribution in the adult brain [6], may appear as more attractive targets for antiepileptic drug discovery since limbic structures such as the hippocampus and amygdala, which play a pivotal role in temporal lobe epilepsy, are particularly rich in KARs [7].

In common with other GluRs, KARs are tetrameric structures assembled from combinations of five different subunits: GluR5, -6, -7, KA1 and -2. GluR5–7 can function as homomeric and heteromeric receptors, while KA1 and -2 play an auxiliary role and can associate with any of the GluR5–7 subunits [7]. The targeting of KAR subunits to different cell compartments and neuronal populations confers specific functions. KARs in presynaptic terminals regulate neurotransmitter release, in axons and somatic compartments they control cell excitability and at postsynaptic densities they mediate synaptic neurotransmission [8–12]. In addition, KAR function can be regulated by insertion or removal at the plasma membrane and they are tightly regulated during development [6,13,14] and synaptic plasticity [13,15]. Age-dependent changes in KAR gene expression are particularly marked [6], and KARs have been proposed as key players in synapse formation and maturation [16–18].

Kainic acid, the prototypic agonist of KARs, produces robust seizures in rodents, which are associated with pathology that closely resembles several features of human temporal lobe epilepsy. These seizures originate in the hippocampus and spread to other limbic structures causing the characteristic pattern of neuronal loss that is highly pronounced in the CA3 hippocampal region [19]. Although kainic acid also activates AMPARs it has been suggested that its proepileptic actions in this region are due to activation of KARs located on hippocampal mossy fiber synapses. Indeed, CA3 hippocampal neurons of GluR6-deficent mice are much less sensitive to both in vitro application of kainate and seizures induced by its systemic injection [20]. Numerous long-term morphological and functional changes in the brains of animals that sustained kainic acid seizures have also been described. Perhaps the most widely described phenomenon is neuronal reorganization, including aberrant mossy fiber sprouting and loss of interneurons [21–23]. Interestingly, following a latent period of a few weeks, recurrent and spontaneous seizures develop with increased frequency in these animals [24]. Thus, kainate-induced seizures can be very useful in understanding the mechanisms responsible for epileptogenesis and human temporal lobe epilepsy.

Since activation of KARs appears to play an important role in seizure generation, a reasonable hypothesis is that KAR antagonists should have strong anticonvulsant properties. However, appropriate KAR-selective compounds are not yet available and KAR- and subunit-selective pharmacology continues to present a major challenge. Initial compounds, such as NBQX, distinguish between NMDARs and non-NMDARs, but are poorly selective between KARs and AMPARs. More recent compounds with better selectivity for KARs, particularly those containing the GluR5 subunit, are becoming available [25,26]. However, compounds that selectively inhibit KARs containing the GluR6 subunit, which is largely responsible for the convulsant effects of kainic acid, are still unavailable.

In addition to chemistry and pharmacological approaches, an alternative strategy to modulate KAR activity is to regulate the number of receptors available for activation. The transport and targeting of KARs and the rates, extent and location of KAR surface expression (collectively termed trafficking) are regulated by interactions with intracellular binding proteins. KARs constantly cycle in and out of the plasma membrane by both activity-dependent and constitutive exo- and endocytosis [27] and, like other glutamate receptor subtypes [28], they are likely to undergo lateral diffusion. The rates, extent and location of KAR surface expression are regulated by interactions with intracellular binding proteins [29–31] that act in concert to control KAR availability and function [13,15].

Direct [29–31] and indirect [32] KAR-interacting proteins have been identified. For example, proteins such as syntenin, PICK1, GRIP and PSD95 bind to a specialized protein-interaction motif termed a PDZ domain present at the extreme C-termini of several KAR subunits. Blocking GRIP or PICK1 binding with peptide antagonists rapidly decreases KAR-mediated synaptic transmission, suggesting that both interactions are required to maintain KAR synaptic function [30]. Thus, specific protein interactions may regulate the insertion and removal of surface KARs in response to different forms of activity, but our current understanding of the specific mechanisms underlying constitutive and activity-dependent trafficking is too limited to formulate specific therapeutic intervention strategies.

Another avenue for investigation is the regulation of KARs by post-translational modification. Protein post-translational modification involves the covalent attachment of chemical groups, lipids, sugars or other proteins and is a crucial factor in the spatial and temporal regulation of protein function. SUMOylation is the reversible covalent attachment of a member of the small ubiquitin-like modifier (SUMO) family of proteins to lysine residues in a target protein. SUMO is conjugated to its target by an enzyme pathway analogous to the ubiquitin pathway. SUMOylation requires the coordinated actions of E1 (SUMO-activating enzyme), E2 (SUMO-conjugating enzyme, Ubc9) and, for most, but not all substrates, E3 (SUMO ligase, e.g., PIAS3), and deSUMOylation is mediated via the SENP family of isopeptidases [33–35].

Neuronal SUMOylation of nuclear proteins appears to be involved in numerous neurodegenerative conditions [36,37]. Thus, there is currently intense interest in SUMOylation; more than 100 SUMOylated proteins have been reported and several hundred potential SUMOylation substrates have been identified by proteome studies [38,39]. These are mainly nuclear-localized proteins but also include extranuclear signaling proteins, synapse scaffolding proteins, presynaptic vesicle proteins, voltage-gated channels and endocytosis-related proteins. Among these there are a large number of proteins that could be involved in epilepsy. For example, SUMO conjugation has been reported to modulate the function of two neuronal potassium channels, K2P1 and Kv1.5, implicating SUMOylation in neuronal excitability. Taken with the fact that KARs are regulated by SUMO (see below), this suggests that SUMOylation could be a general mechanism for regulating ion-channel function.

Recently, it has been demonstrated that the KAR subunit GluR6, alone among AMPAR and KAR subunits, is subject to the protein-based post-translational modification SUMOylation [40]. GluR6 in neurons is SUMOylated in vivo but exhibits low levels of SUMOylation under resting conditions. However, it is rapidly SUMOylated in response to a glutamate or kainate challenge. Interestingly, although both kainate and NMDA application evoke robust internalization of GluR6, the fate of the internalized receptors differs [27] and GluR6 is not SUMOylated following NMDA treatment. Reducing GluR6 SUMOylation using SENP-1 prevented kainate – but not NMDA-evoked KAR endocytosis. Furthermore, a mutated non-SUMOylatable form of GluR6 prevented kainate-evoked KAR endocytosis [40].

Consistent with these imaging and cell biology results, electrophysiological recordings in hippocampal slices demonstrated that KAR-mediated excitatory postsynaptic currents are decreased by infusion of active SUMO-1 from the recording pipette and enhanced by infusion of SENP-1 [40]. These experiments confirm that SUMOylation of GluR6 regulates KAR endocytosis and modifies synaptic transmission, revealing a previously unsuspected role for SUMO in the regulation of normal synaptic function.

SUMO regulation of KARs may also be important in regulating the response of neurons to glutamate release under pathological conditions such as epilepsy and ischemia. Interestingly, it has been shown that SUMO-1 mRNA levels are dramatically raised in ischemic cells [41], although what happens to SUMO mRNA levels in epilepsy is unknown. Nonetheless, it is known that GluR6 is chronically downregulated by excitotoxic stress via a c-fos-dependent mechanism [42], and one attractive possibility is that upregulation of SUMO levels leads to a downregulation in KARs, thereby reducing neuronal excitability.

Thus, it is possible that increasing the amount of GluR6 SUMOylation could reduce neuronal activity. Hence GluR6 SUMOylation presents a new potential target for drug development that could one day be used as a way to treat epilepsy by preventing over-excitation.

The study of SUMOylation in neurons is at an early stage but we anticipate that, analogous to better-characterized post-translational modifications, such as ubiquitination and phosphorylation, SUMOylation will have far-reaching implications for understanding normal and pathological synaptic function.

Financial & competing interests disclosure

Work in Jeremy Henley‘s lab relating to this review is funded by the MRC, the Wellcome Trust and EU grants GRIPPANT (contract number 005320) and ENI-NET (contract number 019063). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed

No writing assistance was utilized in the production of this manuscript.