JLE

Epileptic Disorders

MENU

Update on the genetics of the epilepsy-aphasia spectrum and role of GRIN2A mutations Volume 21, supplement 1, June 2019

Figures


  • Figure 1

Tables

The so-called “idiopathic” childhood focal epilepsies (IFE) correspond to a broad spectrum of childhood epilepsy syndromes with specific age-dependent onset and typical EEG features such as multifocal spikes or spike-waves of various topographies. Childhood epilepsy with centro-temporal spikes (ECTS), also known as Rolandic epilepsy (RE), is the most frequent IFE. The relationships between RE and various comorbid manifestations and conditions, such as migraine, cognitive and behavioral issues, or reading impairment, have increasingly been recognized. Furthermore, the association with transient or permanent speech and/or language impairment has long been reported; hence the identification of the genetic syndrome of RE with verbal dyspraxia (Scheffer et al., 1995). Epileptic Encephalopathy related to Status Epilepticus during slow Sleep (ESES) and Landau-Kleffner syndrome (LKS) -also known as ’acquired’ epileptic aphasia - are two closely related epileptic encephalopathies (EEs) that represent more severe and less frequent forms of the IFE continuum. Indeed, all those syndromes are now considered different clinical expressions of the same pathological spectrum (Rudolf et al., 2009). They all share the association of usually infrequent seizures with paroxysmal EEG discharges activated during drowsiness and sleep, sometimes fulfilling, in a subset of the patients, the criteria of status epilepticus during slow-wave sleep (SES), with more or less severe acquired neuropsychological and behavioral deficits.

General considerations about the genetic architecture of childhood focal epilepsies

A more modern view, that takes into account the recent advances in the genetic origin of various types of epilepsies, has recently challenged the classical distinction between idiopathic and symptomatic epilepsies (Berg et al., 2010). Indeed, based on some of the many possible examples, it has somehow unexpectedly demonstrated that various types of EEs, such as the Dravet syndrome or the Ohtahara syndrome, can have simple, monogenic causes (Depienne et al., 2012; Allen et al., 2013). Conversely, epilepsies of genetic origin (formerly considered as ’idiopathic’) can be associated with comorbid neurological conditions (e.g. migraine, behavioral or cognitive issues) or with structural lesions (e.g. cortical dysplasia). In the IFE, the possible existence of behavioral and cognitive issues, for instance, inherently challenged the use of the ’benign’ term. It had long been assumed that in contrast to generalized epilepsies, most focal epilepsies are caused by lesions, infections, tumors, etc. and are hardly under genetic influence. Twin studies and familial concurrences then indicated that focal epilepsies could also be sustained by genetic factors (Ryan, 1995). As an example, the mapping and the subsequent identification of the first ’monogenic epilepsy’ gene (CHRNA4) encoding a nicotinic acetylcholine receptor subunit was obtained for autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE) (Steinlein et al., 1995). Since then, several genes responsible for rare types of monogenic focal epilepsies have been identified (Berkovic et al., 2006).

Familial aggregation has long been recognized in RE (Neubauer et al., 1998). Relatives of RE patients display higher risk of epilepsy (notably RE, LKS or ESES) than control individuals (De Tiege et al., 2006; Vears et al., 2012; Dimassi et al., 2014). Most RE, however, do not show a simple inheritance. In contrast to RE, the genetic influence in LKS and ESES has long remained controversial (Landau and Kleffner, 1957; Rudolf et al., 2009) and a role for autoimmunity has even been hypothesized (Connolly et al., 1999; Nieuwenhuis et al., 2006). Recent advances in molecular cytogenetics and in next-generation DNA sequencing have dramatically helped in solving this issue. Consistent with the existence of genomic defects (copy number variations) that may have possible pathophysiological influence in numerous human disorders including the epilepsies (Helbig et al., 2009; Mefford et al., 2010), the screening of a series of 61 patients with LKS or ESES led to an overall picture with highly heterogeneous genomic architecture (Lesca et al., 2012). A large number of potentially pathogenic alterations corresponded to genomic regions or genes (e.g. encoding cell adhesion proteins) that were either associated with the spectrum of autism disorders, or involved in speech or language impairment - which was of interest given the well-known association of LKS and ESES with autism-like manifestations (e.g. regression, disturbance of social interactions, perseveration) and with language disorders.

The role of GRIN2A mutations in RE/ESES/LKS

The study by Lesca and colleagues led to detection of several de novo genomic alterations including deletions of the NMDA glutamate receptor (NMDAR) subunit gene GRIN2A (Lesca et al., 2012). Besides the obvious crucial function of NMDARs in the brain (Burnashev & Szepetowski, 2015), a few GRIN2A defects had previously been reported in patients with severe neurodevelopmental disorders (Reutlinger et al., 2010; Endele et al., 2010). Since then, the crucial and direct causal role of de novo or inherited GRIN2A mutations (microdeletions, splice-site, nonsense and missense mutations) in LKS, in ESES, and in RE with verbal dyspraxia has been demonstrated in three parallel studies (Carvill et al., 2013; Lemke et al., 2013; Lesca et al., 2013). The mutations were found in all the different domains of the corresponding GluN2A (formerly known as NR2A) subunit, making it difficult to draw any clear genotype-phenotype correlation (Burnashev and Szepetowski, 2015; von Stülpnagel et al., 2017). However, de novo mutations of GRIN2A might cluster in and around the ligand-binding sites and the transmembrane domains (Strehlow et al., 2015).

The existence of a simple, unifying pathophysiological mechanism remains elusive: whereas the existence of microdeletions and nonsense mutations indicated loss-of-function (LOF) effects, some missense mutations seemed to lead to gain-of-function (GOF), at least in vitro, and some may even have multiple effects (Yuan et al., 2014; Swanger et al., 2016; Sibarov et al., 2017). Interestingly, a recent study has indicated that some GRIN2A-associated EEs caused by gain-of-function mutations might be treatable by the NMDA receptor blocker, memantine (Pierson et al., 2014). This study confers hope for more targeted therapeutic strategies in GRIN2A-related seizure disorders caused by GOF mutations. Nevertheless, an alteration in the Glun2B-to-GluN2A developmental switch, as suggested by altered kinetics of the mutant NMDARs (Carvill et al., 2013; Lesca et al., 2013), and a dysfunction of the thalamocortical network are likely. As a matter of fact, Grin2a knockout mice one month old, exhibit a series of transient brain microstructural alterations that involve the thalamus and the neocortex, and also display rare spontaneous epileptiform discharges in the third postnatal week (Salmi et al., 2018). A case report suggested that the epileptic discharges could have a triggering role in the speech deterioration observed in children carrying a deleterious variant of GRIN2A (Sculier et al., 2017).

Based on the initial studies, it has been estimated that up to 20% of patients with RE, LKS or ESES have a mutation in GRIN2A. A recent exome-wide study of the mutational burden in patients with typical and atypical RE showed that GRIN2A was the only gene with a significantly enriched burden associated with deleterious and LOF mutations (Bobbili et al., 2018). However, interestingly, the statistical significance of this burden disappeared after excluding atypical RE patients, indicating that GRIN2A is more likely to cause phenotypes within the severe end of the disease spectrum (Bobbili et al., 2018).

Other rare monogenic causes of ESES

Besides GRIN2A, other rare monogenic causes, particularly for ESES, have been reported (table 1). ESES has been reported in patients with LOF mutations in KCNA2 encoding the potassium channel subunit Kv1.2. The patients usually present with febrile and multiple afebrile, often focal seizure types, multifocal epileptiform discharges strongly activated by sleep, mild to moderate intellectual disability, delayed speech development, and sometimes ataxia (Syrbe et al., 2015; Masnada et al., 2017). Interestingly, mutations in another potassium channel gene, KCNB1, have also been detected in patients with developmental and epileptic encephalopathies and an ESES-like pattern on the EEG (figure 1) (de Kovel et al., 2017; Marini et al., 2017).

Furthermore, several genetic and genomic defects of the CNKSR2 gene, encoding an adaptor protein of the postsynaptic density, have been identified in patients with clinical features reminiscent of the epilepsy-aphasia spectrum (Lesca et al., 2012; Vaags et al., 2014; Damiano et al., 2017). Last but not least, recent publications also suggest that ESES may be a frequent and underestimated feature in patients with Christianson syndrome, a severe neurodevelopmental disorder due to mutations in SLC9A6, encoding the endosomal solute carrier (Na K/H) exchanger 6 (NHE6) (Zanni et al., 2014; Mathieu et al., 2018).

Genetic susceptibility factors

16p11.2 microduplications encompassing PRRT2 have not been identified in patients with RE or atypical RE (Dimassi et al., 2014; Reinthaler et al., 2014). PRRT2 has been associated with a wide range of paroxysmal neurological disorders including infantile convulsions, paroxysmal dyskinesia (mostly kinesigenic), or hemiplegic migraine (Cloarec et al., 2012; Lee et al., 2012), thus duplication of this gene may also confer risk for the development of RE.

Other potential genetic risk factors for RE have been suggested. For example, a dominant-negative rare variant previously identified in the SRPX2 (Sushi-Repeat Protein, X-linked 2) protein (Roll et al., 2006) co-segregated with a p.Ala716Thr GRIN2A missense mutation in most affected members of a family with RE, verbal dyspraxia, and intellectual disability (Lesca et al., 2013). The GRIN2A mutation likely plays a key role in this family, although it had no obvious effect on a subset of NMDAR properties in vitro (Sibarov et al., 2017). On the other hand, SRPX2 causes transcriptional down-regulation of FOXP2 (Roll et al., 2010), which is associated with verbal dyspraxia (Lai et al., 2001). Furthermore, knockdown of Srpx2 in utero had dramatic consequences on neuronal migration in the developing rat cerebral cortex and led to postnatal epileptiform activity that was prevented by maternal administration of a tubulin deacetylase inhibitor (Salmi et al., 2013). Moreover, it was shown in mice that Srpx2 also influences synaptogenesis and ultrasonic vocalization (Sia et al., 2013). Overall, rare SRPX2 variants might correspond to genetic risk factors for various neurodevelopmental disorders; as a matter of fact, one splice-site SRPX2 mutation was reported in a patient with autism (Lim et al., 2013), and the p.N327S rare variant was detected in a novel patient with LKS (Reinthaler et al., 2014).

Furthermore, ELP4 has been proposed as a susceptibility gene for RE (Strug et al., 2009), however, this has never been confirmed. Mutations that might influence RE were also identified in the paralogous RBFOX1 and RBFOX3 neuronal splicing regulator genes (Lal et al., 2013), in the mammalian target of rapamycin regulator gene DEPDC5 (Lal et al., 2014), and recently also in the GABRG2 gene (Reinthaler et al., 2015).

Conclusion

In conclusion, several genes have been associated with LKS/ESES/RE, however, mutations in GRIN2A represent by far the most common genetic cause identified so far. As GRIN2A mutations unambiguously participate in the epilepsy-aphasia spectrum, whether and how genetic modifiers and environmental factors together with GRIN2A mutations contribute to the huge phenotypic variability, seen in the affected patients and families, emerges as an important question that surely deserves future investigation. Identification of other genetic and non-genetic factors and the ongoing studies with corresponding animal models, such as the Grin2a KO murine model, will help our understanding of the pathophysiology of this fascinating group of disorders situated at the crossroads between epileptic, cognitive, behavioral, and speech and language disorders.

Acknowledgements and disclosures

This work was supported by ANR (Agence Nationale de la Recherche) grant ‘EPILAND’ (ANR-2010-BLAN-1405 01) with EuroBiomed label, by the National PHRC grant 2010 N̊ 03-08, by the European Union Seventh Framework Programme FP7/2007-2013 under the project DESIRE (grant agreement n̊602531) and by INSERM (Institut National de la Santé et de la Recherche Médicale).

None of the authors have any conflict of interest to declare.