JLE

Epileptic Disorders

MENU

GAT-1 (rs2697153) and GAT-3 (rs2272400) polymorphisms are associated with febrile seizures and temporal lobe epilepsy Volume 22, numéro 2, April 2020

Temporal lobe epilepsy (TLE) is one of the most prevalent types of focal epilepsy (Téllez-Zenteno and Hernández-Ronquillo, 2012) and occurs usually in late childhood to early adulthood. TLE seizures are frequently drug-resistant (Engel, 2001). They are classified according to semiology (Fisher et al., 2017) or pathology (Wyler et al., 1992). Hippocampal sclerosis (HS) is the most prevalent pathological substrate (Jay et al., 1993) and is associated with a relatively uniform clinical presentation.

The exact aetiology remains controversial; current hypotheses focus on certain susceptibility genes (Stögmann et al., 2002) and a failure of inhibitory neurotransmission and/or potentiation of excitatory neurotransmission. Two mechanisms that occur within the hippocampus are hallmarks of TLE pathophysiology: hilar mossy cell loss leading to decreased inhibition of granular cells of the dentate gyrus (Sloviter, 1994), and granular cell sprouting which increases excitatory feedback on the dentate gyrus (Babb, Brown, et al., 1984; Babb, Lieb, et al., 1984; Adzhubei et al., 2010; Schmeiser et al., 2017). As inhibition is mainly effectuated through γ-aminobutyric acid (GABA), this neurotransmitter is likely to play a pivotal role in epilepsy (Treiman, 2001; Mody and Pearce, 2004; Guazzi and Striano, 2019).

After release into the synaptic cleft, GABA activity is terminated by reuptake into glial cells and GABA-ergic neurons via a family of electrogenic, sodium- and chloride-dependent transporters (Kaila et al., 1992; Cammack, Rakhilin and Schwartz, 1994; Petroff, 2002). During this voltage-dependent transport, two sodium ions and one chloride ion are exchanged for one GABA molecule (Radian and Kanner, 1983). After reuptake in the neuron, GABA is re-utilized as transmitter located in synaptic vesicles. After reuptake in the astroglial cell, GABA is metabolized (Waagepetersen, Sonnewald and Schousboe, 2003). Molecular cloning has revealed four distinct GABA transporters, termed GABA transporter-1 to -3 (GAT-1 to -3) and betaine/GABA transporter (BGT-1). In the hippocampus, extracellular GABA is taken up primarily by both presynaptic neuronal GAT-1 and astrocytic GAT-3. Studies in TLE patients have shown changes in extracellular GABA concentrations, related to alterations in GAT quantity and/or function (During and Spencer, 1993; During, Ryder and Spencer, 1995; Williamson, Telfeian and Spencer, 1995; Mathern et al., 1999; Hoogland et al., 2004). A reduced number of GATs and/or GAT dysfunction is consequently suspected in the hippocampus of TLE patients. Recent data have confirmed previous observations that GAT expression is spatially reduced, in an isoform-specific manner, in HS (Schijns et al., 2015). Even though it is clear that GABA-ergic neurotransmission is disturbed, it is unclear why. Regarding epileptogenesis, only two risk factors have been clearly identified: traumatic brain injury (TBI) and febrile seizures (FS). It is not clear, however, why a minority of patients develop epilepsy after TBI or FS, and why a significant part of the TLE-FS+ subgroup shows HS.

Perhaps a specific genetic background predisposes to the development of TLE, while another genetic background is protective. As variants in the GAT genes could play such a role, we have investigated the frequency of single-nucleotide polymorphisms (SNPs) in genes encoding GAT-1 (SLC6A1 gene, chromosome 3) and GAT-3 (SLC6A11 gene, chromosome 3) in a group of drug-resistant TLE patients with and without a background of FS and TBI, and in healthy controls. Regarding GAT-1, the SNP with dbSNP identifier rs2697153 (A>G allele, 5’ flanking region, exon 1) has been associated with increased panic attacks in a group of anxiety disorder patients (Thoeringer et al., 2009). This SNP is located in the 5’ flanking region of SLC6A1 within only 10 kb of exon 1, potentially including the putative promoter region, providing functional and structural relevance to this specific SNP (Thoeringer et al., 2009). Another study (Carvill et al., 2015) has identified six SCL6A1 (GAT-1) gene mutations in seven individuals with myoclonic-atonic seizures. The alterations in these patients most probably led to loss of GAT-1 function with consequently decreased clearance of synaptic GABA, provoking both phasic and tonic inhibition, leading to enhanced spike-wave discharges and, hypothetically, seizures. Fifteen other SNPs in the SLC6A1 gene were identified, but this study focused on non-clinical samples and therefore the detected variants could not be related to certain diseases (Hirunsatit et al., 2007). Johannesen et al. recently found four recurring missense variants, suggesting possible mutational “hot spots” within SLC6A1. They studied patients with myoclonic atonic epilepsy (MAE) syndrome and found that the clinical hallmark of these gene variants is cognitive impairment, expressed with different levels of intellectual disability. Most of these patients also had epilepsy since childhood. This study, together with the study of Carvill et al., suggests that SLC6A1 mutation-positive patients have a combination of intellectual disability, language delay and epilepsy, most frequently associated with MAE syndrome (Johannesen et al., 2018).

Regarding GAT-3, the SNP with dbSNP identifier rs2272400 (GAT-3 c.1572 C>T, exon 12) has been associated with antiepileptic drug resistance in a Korean study (Kim et al., 2011). Structurally, the SNP rs2272400 is located in the coding region of GAT-3 on exon 12, however, this does not lead to amino acid changes. It has become increasingly clear that “silent” SNPs may influence promoter activity (and thereby gene expression) or may lead to synthesis of proteins with the same amino acid sequence, but different structural and functional properties (Kim et al., 2011). No other SNPs in the SLC6A11 gene have thus far been identified. We therefore performed an association study for both SNPs in a cohort of drug-resistant TLE patients and healthy controls.

Materials and methods

Study population

All TLE patients (n=138) were diagnosed with drug-resistant TLE (Kwan et al., 2010). They were subjected to resective brain surgery (anterior temporal lobectomy and amygdalahippocampectomy) after an extensive preoperative work-up including video-EEG, 3T-MRI and neuropsychological examination. Temporal lobe neocortical and hippocampal specimens were collected during surgery. Upon resection, specimens were immediately frozen on dry ice and stored at -80̊C until molecular analysis. All TLE patients gave their written informed consent to use their resected tissue samples for medical research. The research was conducted in The Netherlands only. According to Dutch legislation, it is not obligatory to obtain permission from the local ethics committee for medical research with patient data and tissue samples. At the time of tissue collection for this study, ethical approval was deemed unnecessary not only because of this legislation but also because the tissue collection did not determine the course of surgery or treatment, either before or during follow-up, and the investigated samples would otherwise have been discarded. All tissue samples and patient data were anonymized for members of the experimental research team.

Based on routine histopathological evaluation, samples consisted of both HS and non-HS (tumour and vascular malformation) cases. HS samples were graded as mild HS (Grade 1-2) or severe HS (Grade 3-4) (Wyler et al., 1992). Controls (n=94) consisted of buccal swabs or blood samples collected from a pool of anonymized healthy subjects without a (familial) history of neurological or psychiatric disease. The control group was matched with the study population for age, gender and ethnicity.

SNP genotyping

In both patients and controls, the alleles of the SNPs in the GAT-1 gene (SLC6A1), SNP rs2697153, and the GAT-3 gene (SLC6A11), SNP rs2272400, were analysed. Genomic DNA was extracted from 20-25 mg frozen neocortical tissue (patients) and blood samples/buccal swabs (controls) using a QIAamp DNA Mini Kit (Qiagen, Hilden, Germany). The DNA samples were subsequently submitted to PCR, followed by allele-specific restriction fragment length polymorphism (RFLP) analysis, as follows: PCR reactions for GAT-1 and GAT-3 were carried out in a volume of 50 μL containing 200 ng genomic DNA, 0.5 μM of each primer (GAT-1 forward and reverse primer: 5’TCAATTGGGCACGAGGGTAG3’ and 5’CCTTAGGATGTCAAAGGGCCA3’, respectively [Thoeringer et al., 2009]; GAT-3 forward and reverse primer: 5’GGATCACCTTCCGCCTTTCT3’ and 5’CGTGGGTTGGAGGGTAGATG3’, respectively [Dong-Uk et al., 2011]), 0.2 mM dNTPs (Invitrogen), 1 U fast start Taq polymerase (Roche Applied Science, The Netherlands), and 5 μL PCR buffer (10x) with 1.5 mM MgCl2. Next, SNP-dependent fragments were generated by digestion of PCR products with allele-specific restriction endonucleases according to the manufacturer's recommendations (supplementary table 1 for lengths of bp fragments). Finally, fragments were resolved by gel electrophoresis on a 3.5% agarose gel containing Gelstar and visualized by UV light using a Biorad Geldoc 2000 system with Quantity One software (supplementary figure 1A, B).

Statistical analysis

Statistical analysis was performed using the Statistical Package for Social Sciences 12.0 software. Differences in genotype between patient groups and controls were analysed by the two-tailed Fisher exact test and the Chi-Square-Test. A p value <0.05 was considered significant.

Results

Demographic characteristics of the patient cohort are presented in table 1.

The patient group included 138 neocortical samples, of which we were able to genotype 115 for GAT-1 (tables 2, 4) and 129 for GAT-3 (tables 3, 5).

The control group consisted of 94 samples, from which we were able to genotype 75 samples for GAT-1 (table 2) and 86 samples for GAT-3 (table 3).

Regarding the three GAT-1 genotypes, GA was most common (48% in epilepsy patients and 56% in controls), while the GG genotype was least common (12% in epilepsy patients and 21% in controls) (table 2). The AA genotype was significantly more common in the epilepsy group than in controls (40 vs 23%; p=0.03).

Regarding the three GAT-3 genotypes, CC was most common (96% in epilepsy patients and 100% in controls), while the TT genotype was not present at all. In the control group, the CT genotype was not present either, though it was found in 4% of epilepsy cases (table 3) (p=0.16). Furthermore, the C allele was present in 98% of epilepsy cases and in all controls, indicating that the T allele was only found in 2% of epilepsy cases and not at all in controls.

Within the patient group, data on FS history was available for 136 patients (tables 4, 5). Of these, 31/136 patients (23%) had a history of FS (FS+). In this FS+ group, we were able to determine the GAT-1 genotype in 25 and the GAT-3 genotype in 28 patients (tables 4, 5). In the FS- group, we were able to determine the GAT-1 genotype in 87, and the GAT-3 genotype in 100 patients (tables 4, 5). The GAT-3 CT genotype was significantly more common in the FS+ group (14%) than in the FS- group (1%; p=0.01). Both the patient group and the control group were in Hardy-Weinberg equilibrium.

Within the epilepsy group, data was available regarding TBI history in 133 patients (tables 6, 7). Of these, 13 had a history of TBI (10%), of whom the GAT-1 genotype was determined in 11 patients, and GAT-3 genotype in 13. Of the 120 patients without a history of TBI, the GAT-1 genotype was determined in 88 patients, and GAT-3 genotype in 111. The GAT-1 GG genotype was present in 18% of the TBI+ group and in 10% of the TBI- group. The GAT-3 CT genotype was not present in the TBI+ group at all, but was present in 4% of the TBI- group. The GAT-1 and GAT-3 genotypes did not significantly differ between TBI+ and TBI- patients (tables 6, 7).

Discussion

In this association study, we analysed the rs2697153 SNP in the GAT-1 gene (SLC6A1) and the rs2272400 SNP in the GAT-3 gene (SLC6A11) in a well-circumscribed cohort of patients with drug-resistant TLE and matched healthy controls. The SNPs were selected based on relevant recent literature (Thoeringer et al., 2009; Kim et al., 2011; Carvill et al., 2015; Johannesen et al., 2018), however, the clinical significance of both SNPs is not yet reported in major databases (dbSNP, gnomAD and ExAC).

We observed that the AA GAT-1 genotype was significantly more frequent in patients than in controls. The obtained frequencies in the control group for both SNPs were comparable with data from the above-mentioned major databases (dbSNP, gnomAD) corresponding to the European population (dbSNP: rs2697153 G=0.4085, A=0.5915 and rs2272400 C=0.99257, T=0.00743).

A potential role for GAT-1 in epilepsy is supported by Carvill et al. (2015), who found six mutations in the GAT-1 gene of myoclonic-atonic seizure patients. These mutations probably cause a decrease or loss of GABA transport activity (Ben-Yona and Kanner, 2013). Decreased GABA transport results in increased extracellular GABA levels and can paradoxically, as described in GAT-1 knockout mice (Jensen et al., 2003), provoke seizures and hypersynchronous epileptiform neuronal activity. Elevated synaptic GABA concentrations can overstimulate extra synaptic GABAA and GABAB receptors and enhance phasic and tonic inhibition, which consequently can be associated with spontaneous spike-wave discharges (Hosford, Wang and Cao, 1997). Cope et al. have shown that spontaneous discharges, typical of absence seizures, have been recorded in GAT-1 knockout mice and that GABAA-mediated tonic inhibition was increased (Cope et al., 2009). Other studies have shown that prolonged GABAB receptor activation stimulates low-voltage activated Ca2+ channels, which can cause recurrent excitation within the thalamocortical system through excessive Na+ spikes, associated with epilepsy (Han, Cortez and Snead, 2012).

A second finding of this study was that the CT genotype and T allele of GAT-3 SNP rs2272400 were more common in TLE patients than controls. The TLE group was subdivided into groups with and without epilepsy risk factors, FS and TBI. This sub-analysis revealed that the FS+ group contained more CT-genotyped patients (14%) than the FS- group (1%; p<0.05). This GAT-3 SNP is a so-called synonymous SNP on exon 12, indicating that there is no amino acid change in the GAT-3 protein. It remains to be elucidated whether this SNP can result in alternative mRNA or a change in turnover or capacity of GAT-3 protein, or is merely an indicator of a co-segregating functional SNP, as described for other silent SNPs (Komar, 2007; Shastry, 2009).

Regarding the function of GAT-3, astrocytic GAT-3 is capable of both uptake of GABA from the synaptic cleft (Wu, Wang and Richerson, 2003; Schousboe et al., 2004) and transport reversal, i.e. release of GABA into the synaptic cleft. These mechanisms may change under pathological conditions (Raiteri et al., 2002). Dysfunction of GAT-3 can be due to reduced transporter reversal, resulting in reduced extracellular GABA levels, thereby contributing to a prolongation of the ictal state (Kinney, 2005). The significance and possible pathological role of the GAT-3 T allele in TLE with a history of FS has not been revealed yet. A possible hypothesis is that (a rapid increase of) high temperature induces a progressive dysfunction of the already altered GABA transporter, leading to further lowering of the synaptic GABA level, lower inhibitory tonus, and consequently a lower seizure threshold. A similar mechanism has already been described for FS and GABAA receptor subunit mutations (Kang, Shen and Macdonald, 2006). Thus, the exact pathophysiological mechanism of this SNP in TLE and FS remains to be clarified. Though this study demonstrates a significant association between both SNPs and TLE, a limitation is due to the fact that not all ethnic groups were represented in this study. Therefore, we suggest validating this study in another cohort, preferably including different genetic backgrounds.

Conclusion

The results of this study demonstrate that the AA genotype of the GAT-1 transporter SNP (rs2697153) is significantly more common in epilepsy patients than in controls. The second finding is that the CT genotype, and as a consequence, the T allele, of the GAT-3 transporter SNP (rs2272400, GAT-3 c.1572 C>T) are significantly more frequent in TLE patients, especially the FS+ group, compared with controls. These data suggest that the susceptibility to develop TLE is associated with SNPs in genes encoding GAT-1 and -3. In fact, GAT-3 c1572T may be a contributing factor in patients with TLE and FS. The exact pathophysiological mechanism remains to be elucidated. Generalizability of our findings would require validation in cohorts with different ethnic backgrounds, and the establishment of an international biobank containing genetic material from more epilepsy patients would be very helpful to achieve this.

Highlights

  • Susceptibility to develop TLE is associated with SNPs in GAT-1 and -3 genes
  • AA genotype of GAT-1 SNP rs2697153 is significantly more frequent in TLE patients
  • CT genotype of GAT-3 SNP rs2272400 is significantly more frequent in TLE patients
  • GAT-3 SNP rs2272400 may be a contributing factor in patients with TLE and febrile seizures

Supplementary data

Supplementary materials are available on the www.epilepticdisorders.com website.

Disclosures

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


* This work has previously been presented as a scientific poster at the XXIIIrd Congress of the ESSFN 2018 - Edinburgh, Scotland (26/09/2018-29/09/2018) and at the Congress of the European Association of Neurosurgical Societies (EANS) 2018 – Brussels, Belgium (21/10/2018-25/10/2018).