Genetic mechanisms in idiopathic epilepsies


  • Yvonne G Weber MD,

    1. Department of Neurology
    Search for more papers by this author
  • Holger Lerche MD

    Corresponding author
    1. Departments of Neurology and Applied Physiology, University of Ulm, Germany.
      * Correspondence to last author at Department of Neurology, University of Ulm, Oberer Eselsberg 45, 89081 Ulm, Germany.
    Search for more papers by this author

  • See end of paper for list of abbreviations.

* Correspondence to last author at Department of Neurology, University of Ulm, Oberer Eselsberg 45, 89081 Ulm, Germany.


Idiopathic epilepsies are considered to be genetically determined. The inheritance can be monogenic and the detected mutation considered sufficient to cause the phenotype. In contrast, when the inheritance is complex, the epileptic phenotype is determined by several minor genetic defects that are much more difficult to discover. In recent years, an increasing number of mutations, mainly associated with rare monogenic idiopathic epilepsy syndromes, have been identified in genes encoding subunits of voltage- or ligand-gated ion channels. A few mutations have also been found in the frequent classical forms of idiopathic generalized epilepsies which are thought to follow a complex genetic trait, for example, in absence or juvenile myoclonic epilepsies. Functional studies characterizing the molecular defects of the mutant channels point to an important role of GABAergic synaptic inhibition in the pathophysiology of idiopathic epilepsies. As a result of genetic and functional investigations, not only will the pathophysiology of epilepsy be better understood, but newly discovered genes and pathophysiological pathways may also determine novel targets for pharmacotherapy, as has been shown for the anticonvulsant drug retigabine, which enhances the activity of neuronal KCNQ potassium channels.

With a lifetime incidence of up to 3%, epilepsy is one of the most common neurological disorders.1 It is characterized by recurring unprovoked epileptic seizures caused by synchronized electrical discharges of central neurons. Based on their origin, epileptic seizures and epileptic syndromes can be focal or generalized, whereas their underlying cause can either be symptomatic (including cortical malformations, brain tumours, or stroke) or idiopathic. Idiopathic epilepsies are assumed to be mainly genetic in origin, present with normal brain imaging, and are estimated to represent up to 47% of all epilepsies.2 Although genetic alterations can also cause symptomatic epilepsies, as in cortical malformations, this review will summarize only genetic findings in idiopathic epilepsies.

Within the past 10 years, the first gene defects have been identified that cause mostly rare monogenic idiopathic epilepsy syndromes, such as autosomal dominant nocturnal frontal lobe epilepsy or benign familial neonatal seizures (BFNS). However, only 1 to 2% of the idiopathic epilepsies seem to be monogenic, whereas most of them are believed to be polygenic, such as the most common idiopathic generalized epilepsies (IGEs), which are childhood absence epilepsy (CAE) and juvenile absence epilepsy (JAE), juvenile myoclonic epilepsy (JME), and epilepsy with generalized tonic–clonic seizures on awakening. As illustrated in Table I, the detection of genetic defects may extend the number and the classification of the idiopathic epilepsies. Most of the epilepsy-associated genes identified so far encode ion channels. Interestingly, those proteins encoded by the non-ion channel genes have been suggested to interact with ion channels. This is conceivable from a pathophysiological point of view, as ion channels provide the basis for both the electrophysiological excitability of neuronal cell membranes and the communication between neurons. Also, most anticonvulsant drugs in clinical use today modulate different types of ion channels.

Table I.   Susceptibility loci and affected genes described for various idiopathic epilepsy syndromes
DesignationAbbreviationOMIMMain clinical featuresOnsetChromosomeGeneProtein
  1. Dec, decade; unk, unknown; GTCS, generalized tonic–clonic seizures; SPS, simple partial seizures; CPS, complex partial seizures; PKC, paroxysmal kinesiogenic choreoathetosis; PED, paroxysmal exertion-induced dyskinesia; VLGR1, very large G-protein-coupled receptor 1; LGI1, leucine-rich gene, glioma-inactivated; JAE, juvenile absence epilepsy; SEP, sensory evoked potentials. Please see OMIM for further references (

Benign familial neonatal seizuresBFNS1/EBN1121200Clusters of GTCS or apnoeic spellsFirst d20q13.3KCNQ2K+ channel
Benign familial neonatal/infantile seizuresBFNIS607745Clusters of afebrile CPSFirst d/mo2q23–24.2SCN2ANa+ channel
Benign familial infantile seizuresBFIS1601764Clusters of afebrile or apnoeic spells or motor seizures or CPS/GTCS 3–9mo19q
BFIS with familial hemiplegic migraineBFIS/FHM182340BFIS, hemiplegic migraine1st/2nd dec1q21–23ATP1A2Na+/K+ATPase
Infantile convulsions and choreoathetosis syndromeICCA602066PKC, infantile convulsions1st dec16p12–q12
Rolandic epilepsy, PED, writers’crampEPRPDC608105CPS, PED1st dec16q12–p11.2
Benign epilepsy with centro-temporal spikesBECTS/Rolando117100Hemifacial partial seizures1st/2nd dec15q24
Generalized epilepsy with febrile seizures plusGEFS1604233Febrile and afebrile CPS/GTCS, absences, myoclonic seizures1st dec19q13.1SCN1BNa+ channel
GEFS26042332q24SCN1ANa+ channel
GEFS36042335q31.1–33.1GABRG2GABAA receptor
GEFS56042331p36.3GABRDGABAA receptor
Severe myoclonic epilepsy of infancy (Dravet)SMEI607208Febrile GTCS, later absences, myoclonic, SPS, CPS, stagnation in psychomotor development 1st y2q24SCN1ANa+ channel
Intractable childhood epilepsy with generalized tonic–clonic seizuresICEGTC607208Febrile GTCS, mental decline, ataxia, hypotoniaInfancy2q24SCN1ANa+ channel
Febrile seizuresFEB1121210Febrile CPS/GTCS6mo/6y8q13–21
Idiopathic generalized epilepsy (IGE)IGE1/EIG1606970Age-related occurrence of unprovoked generalized seizures (GTCS, absences and myoclonic seizures) 1st/3rd dec8q24
IGE5/EIG5/JAE6005703q26CLCN2Cl- channel
Childhood absence epilepsy (CAE)ECA1600131Absence seizures, GTCS, EEG: generalized symmetric 3Hz spike wave6–7y8q24
ECA26076815q31.1–33.1GABRG2GABAA receptor
ECA36076823q26CLCN2Cl- channel
ECA46111365q34–35GABRA1GABAA receptor
Juvenile myoclonic epilepsy (JME)EJM1254770Afebrile myoclonic jerks and GTCS on awakening
EEG: irregular generalized spikes and polyspikes
8–20y6p12–11EFHC1Ca2+ turnover
EJM46113645q34–35GABRA1GABAA receptor
EJM56019492q22–23CACNB4Ca2+ channel
Benign adult familial myoclonic epilepsyBAFME1601068Tremor, myoclonic jerks, GTCS, mild mental retardation, EEG: frontotemporal and generalized spikes, SEP: giant potentials11–40y8q24
BAFME2607876 2p11.1–q12.2
Autosomal dominant nocturnal frontal lobe epilepsyADNFLE/EFNL1600513Brief nocturnal motor seizures with hyperkinetic or tonic manifestations4–14y20q13CHRNA4Acetylcholine receptor
Lateral temporal lobe epilepsy (ADLTE)ADPEAF600512SPS, CPS including auditory symptoms8–50y10q23–26LGI1Function unk
Familial partial epilepsy with variable fociFPEVF604364Partial, nocturnal seizures with variable foci1st dec22q11–12
Generalized epilepsy/paroxysmal dyskinesiaGEPD609446Absences, paroxysmal dyskinesias1st dec10q22.3KCNMA1K+ channel

Autosomal dominant nocturnal frontal lobe epilepsy

Autosomal dominant nocturnal frontal lobe epilepsy includes frequent brief seizures occurring in childhood with hyperkinetic or tonic manifestations, typically in clusters at night. Ictal video-electroencephalographic studies have revealed partial seizures originating from the frontal lobe but also in parts of the insula, suggesting a defect of a broader network.3 The penetrance of the disease is estimated at approximately 70 to 80%. A mutation was identified in the gene CHRNA4 encoding the α4-subunit of a neuronal nicotinic acetylcholine receptor as the first ion channel mutation found in an inherited form of epilepsy.4 Altogether, five mutations in CHRNA4 and two in CHRNB2, which encodes the β2-subunit of neuronal nicotinic acetylcholine receptor, have been reported so far.5,6 Recently, another mutation in CHRNA2, encoding the neuronal nicotinic acetylcholine receptor α2-subunit, was detected.7 All these mutations reside in the pore-forming M2 transmembrane segments (Fig. 1a). Different effects on gating of heteromeric α4β2 channels leading either to a gain- or a loss-of-function were reported when most of the known mutations were functionally expressed in Xenopus oocytes or human embryonic kidney cells. The exact pathomechanism is not fully understood, but an increased acetylcholine sensitivity could be the main common gating defect of the mutations.5,6

Figure 1.

 Proposed structure of (a) nicotinic acetylcholine receptor and (b) voltage-gated potassium channels KCNQ2/3. (a) Each subunit of the pentameric receptor comprises four transmembrane regions (M1–M4, upper part of figure). M2 segments line the channel pore whereas the long extracellular amino (N)-terminal part of the α-subunit harbours the acetylcholine-binding site. Mutations found in patients with autosomal dominant nocturnal frontal lobe epilepsy are exclusively localized in the M2 segment.4–7 (b) Channel proteins share a predicted topologic arrangement of voltage-gated channels with an intracellular N terminus, six transmembrane segments (S1–S6), a pore loop (P-loop) between S5 and S6 that forms the selectivity filter of the channel, a positively charged voltage-sensing S4 segment, and a specifically large intracellular carboxy (C) terminus, which mediates the assembly of KCNQ2/3 channels into heterotetramers and other important functions such as targeting. Many benign familial neonatal seizure-causing mutations are located within this functionally important C terminus.6,15,49

Benign familial neonatal seizures

Benign familial neonatal seizures (BFNS) are characterized by clusters of seizures in the first days of life, disappearing spontaneously after weeks to months. Seizures have a partial onset, often with hemi-tonic or -clonic symptoms or with apnoeic spells, or can clinically appear as generalized. Interictal electroencephalograms (EEGs) are usually normal. The rare ictal EEGs recorded showed focal and generalized discharges. The risk of recurring seizures later in life is about 15%. Although psychomotor development is usually normal, an increasing number of cases with learning disability* have recently been described.8,9 BFNS is autosomal dominantly inherited with a penetrance of 85%. Mutations in KCNQ2 and KCNQ3 have been identified to cause it.5,6,10–12

KCNQ2 and KCNQ3 channels give rise to the M-current, a slowly activating potassium current which can be suppressed by the activation of muscarinic acetylcholine receptors.13 Co-expression of heteromeric wild-type and mutant KCNQ2/3 channels (Fig. 1b) usually revealed a reduction in the resulting potassium current of about 20 to 30%, which is apparently sufficient to cause BFNS.14 Interestingly, subtle changes in channel gating restricted to subthreshold voltages of an action potential are also sufficient to cause BFNS, proving the physiological importance of this voltage range for the action of M-channels in a human disease model.15 Although the reduction of the potassium current can explain the occurrence of epileptic seizures by a subthreshold membrane depolarization which increases neuronal firing, it is not fully understood why seizures preferentially occur in neonates. Possible explanations include: (1) an increasing axonal expression of KCNQ2 channels during maturation rendering neurons less vulnerable to an M-current reduction in adulthood;16 (2) a link to the developmental switch from GABAergic excitation to inhibition as the M-current might provide a primary inhibitory pathway in the immature brain;17 and (3) the expression of a shorter, non-functional splice variant of KCNQ2 exclusively in the fetal brain.18

Benign familial neonatal/infantile seizures (BFNIS) and benign familial infantile seizures (BFIS)

Two clinically similar but genetically distinct epilepsy syndromes are BFNIS and BFIS. Partial epileptic seizures in BFIS with or without secondary generalization occur between the age of 3 and 12 months, and in BFNIS in neonates. The rare ictal EEGs showed focal epileptic discharges in different brain regions. BFIS can be associated with other neurological disorders, such as paroxysmal dyskinesia or migraine (Table I). Different loci have been described for BFIS, but a gene responsible for it has not been identified.19-21 Mutations in the SCN2A gene encoding one of the α-subunits of voltage-gated sodium channels expressed in the mammalian brain are found in BFNIS.22 The first functional investigations revealed small gain-of-function effects of some mutations predicting increased neuronal excitability.23,24

Generalized epilepsy with febrile seizures plus (GEFS+), severe myoclonic epilepsy of infancy (SMEI; Dravet syndrome), intractable childhood epilepsy with generalized tonic–clonic seizures, and SMEI-borderline epilepsy

The term GEFS+ represents a childhood-onset autosomal dominant syndrome comprising febrile convulsions and a variety of afebrile epileptic seizure types within the same pedigree. The penetrance is about 60%. Two-thirds of affected individuals were diagnosed as having febrile seizures (FS) which are often combined with either FS persisting after the sixth year of life or with afebrile generalized tonic–clonic seizures (FS+). Additional seizure types such as absences, atonic, or myoclonic–astatic seizures have been described in other patients. Partial epilepsies occur in rare cases.

SMEI is characterized by hemi- or generalized clonic or tonic–clonic seizures in the first year of life that are often prolonged and associated with fever. During the course of the disease, patients develop afebrile generalized myoclonic, absence, or tonic–clonic seizures, but simple and complex partial seizures also occur. Cognitive deterioration appears in early childhood. In contrast to GEFS+, the syndrome is resistant to pharmacotherapy in most cases, but stiripentol seems to have a significant positive effect in patients with SMEI.25 Cranial magnetic resonance imaging in patients with SMEI found focal and generalized internal and external atrophy, which is discussed as a result of the brain encephalopathy; the rate of hippocampal sclerosis is not increased.26 Because patients with SMEI sometimes have a family history of febrile or afebrile seizures, and in some families GEFS+ and SMEI overlap, SMEI may be regarded as the most severe phenotype of the GEFS+ spectrum.27 Another similar severe epilepsy syndrome of childhood is intractable childhood epilepsy with generalized tonic–clonic seizures (ICEGTC).28 Onset and clinical course including learning disability are as in SMEI, except that myoclonic seizures do not occur. Families with some instances of ICEGTC in other family members affected by GEFS+ have been described. Therefore, we may conclude that the GEFS+ spectrum extends from simple febrile seizures to a variety of severe epilepsy syndromes of childhood such as intractable ICEGTC and SMEI, as also confirmed by genetic results described below. (For a further overview of the clinical information and references see reference 6).

In this group of diseases, the first genetic defect was discovered in a large family with GEFS+ within the SCN1B gene encoding the β1-subunit of the voltage-gated Na+ channel.29 In further studies, several groups have identified mutations in SCN1A that encode a sodium channel α-subunit in families with GEFS+,30 ICEGTC,28 and, interestingly, in patients with simple FS,31 severe myoclonic epilepsy-borderline epilepsy, and other severe epilepsies of childhood, which lack some core features of SMEI such as myoclonic or generalized tonic–clonic seizures.32 Whereas point mutations are found in families with GEFS+, most patients with SMEI carry de novo nonsense mutations predicting truncated proteins without function.33 The sodium channel blocker lamotrigine, the only drug of this class used in patients with idiopathic generalized epilepsies, deteriorates the epilepsy in patients with SMEI.34 This suggests that it is a loss-of-function sodium-channel disorder. Nevertheless, there is evidence for further genetic as well as clinical heterogeneity,35 not related to the mutations in the GABAA receptor, which will be explained below.

The expression of many SCN1A mutations in human embryonic kidney cells or Xenopus oocytes revealed both gain- and loss-of-function mechanisms. However, the loss-of-function mechanisms seem to predominate for FS and GEFS+, which is in agreement with genetic and functional studies in SMEI and ICEGTC.6 Two recently published mouse models for SCN1A, in which loss of function mutations were introduced into the endogenous mouse gene, showed spontaneous seizures and reduced sodium currents with decreased sodium-channel expression selectively in inhibitory interneurons. These results suggest that SMEI and maybe the other SCN1A-linked seizure disorders, are caused by a decreased excitability of GABAergic interneurons owing to a haploinsufficiency of SCN1A;36,37 this correlates very well with the mutations detected in GABA receptors described below.

GABA receptors were long suspected to be involved in epileptogenesis. The first mutations were identified in the γ2-subunit of GABAA receptors (GABRG2) in two families with GEFS+.38,39 One of these families presented with a typical GEFS+ phenotype,38 the other with a frequent combination of FS and absence seizures beside other GEFS+-related epilepsy syndromes.39 Three more mutations were found in other families with simple FS, GEFS+, and absence seizures.6,40 Functional expression of some of these GABRG2 mutations expressed in Xenopus oocytes or mammalian cell lines revealed a pronounced loss-of-function by altered gating or defective trafficking and reduced surface expression as a common pathogenic mechanism.6,38,39 Hence, they reduce the main mechanism for neuronal inhibition in the brain, which can explain well the occurrence of epileptic seizures.

Idiopathic generalized epilepsies

CAE, JAE, JME, and epilepsy with grand-mal seizures on awakening present the classical and most common subtypes of IGE. Absence seizures in CAE manifest typically around the sixth year of life, are of short duration, usually about 10 seconds, and typically occur in clusters of up to 100 seizures a day. In adolescence, generalized tonic–clonic seizures can occur. Absence seizures in JAE are principally similar, but less frequent and start around puberty. Myoclonic jerks are the clinical hallmark of JME, particularly of the upper extremities, which appear without loss of consciousness. They can be clinically subtle and escape clinical recognition. The disease also manifests during puberty, with seizures typically developing after awakening and being provoked by sleep deprivation. Generalized tonic–clonic seizures occur in about 75% of patients. Epilepsy with grand-mal (generalized tonic–clonic) seizures on awakening develops in adolescence. Seizures usually occur in the first hour after the patients awake, independent of the time of day. All of these clinical syndromes can overlap either within individual patients or within families and are typically associated with generalized spike-wave or poly-spike-wave discharges on EEG. Brain imaging is unremarkable.

In contrast to the monogenic syndromes discussed above, IGE follows a complex genetic inheritance, with only a few, rare, large families presenting with an apparent autosomal dominant inheritance. To identify IGE-associated genetic defects, different approaches have been applied. Linkage analyses, performed in single large families with many patients or in large samples of small, nuclear families, as well as direct candidate gene screening, enabled the identification of the first mutations in ion-channel-encoding genes associated with IGE. The results of these studies are summarized in Table I.

In addition to mutations in GABRG2 causing GEFS+ or CAE with febrile seizures, the first mutation in GABRA1, the gene encoding the α1-subunit of the GABAA receptor, was identified in a family with JME,42 and a second one later in a patient with CAE.43 As described above for all mutations in the γ2-subunit, the α1-subunit mutations also lead to a pronounced loss-of-function of the GABAA receptor when expressed in Xenopus oocytes or mammalian cells.6,42,43 In families with IGE, mutations were also found in the gene CLCN2 encoding a neuronal voltage-gated Cl channel.44,45 This channel could also play an important role in neuronal inhibition. Owing to its specific gating properties, it constitutes a chloride extrusion pathway keeping the intracellular chloride concentration at low levels, which is important for the inhibitory action of the GABAA receptor.

Perspectives and implications for therapy

Genetic techniques have been developing rapidly. Whole-genome approaches now allow the detection of common genetic variants in complex genetic syndromes. Chip technology also makes this possible. The next generation of sequencing techniques are also developing. This will enable more rapid, more effective, and cheaper genotyping, so that we expect many more genetic defects in epilepsies and other diseases to be unravelled in the near future. To translate these findings and use them for daily clinical practice is another issue. In some cases, they can explain clinical observations, such as the deterioration of myoclonic seizures in patients with SMEI upon treatment with sodium-channel blockers, and help to select the best therapy. Routine genetic testing is not yet reasonable for many of the syndromes described here, because (1) genetic defects have only been identified in a few patients with common IGE syndromes, or (2) most do not have a consequence for management and treatment. However, knowledge of the genetic defect can be relevant for diagnosis and prognosis, and may even have an impact on therapy, for example in early use of stiripentol in SMEI.

The knowledge of genetic defects and their underlying mechanisms can give rise to new therapeutic strategies in epilepsies in general. For example, defective proteins can be used as new pharmacological targets, such as KCNQ channels. Studying the pathophysiology of BFNS, a completely novel approach of anticonvulsive treatment emerged when retigabine was identified as an activator of M-currents conducted by KCNQ2 and KCNQ3 K+ channels.46 Retigabine probably binds to the activation gate of this channel,47,48 thereby shifting its voltage dependence of activation in the hyperpolarizing direction, which largely stabilizes the membrane at subthreshold voltages. This is a very potent anticonvulsive mechanism, which could be used in the future treatment of epilepsy or other disorders with a membrane depolarization and hyperexcitability of the nervous system, such as migraine, neuropathic pain, and stroke. This example shows how genetic and pathophysiological studies of epilepsy may indeed help to improve therapy.


  • *

    North American usage: mental retardation.


This work was supported by the Bundesministerium für Bildung und Forschung (BMBF-NGFN2, 01GS0478), the European Community (LSH-CT-2006-037315 EPICURE), and the Deutsche Forschungsgemeinschaft (DFG, Le1030/9-1, /10-1). HL is a Heisenberg fellow of the Deutsche Forschungsgemeinschaft.

List of abbreviations

Benign familial infantile seizure


Benign familial neonatal/infantile seizure


Benign familial neonatal seizure


Childhood absence epilepsy


Febrile seizure


Generalized epilepsy with febrile seizures plus


Intractable childhood epilepsy with generalized tonic–clonic seizures


Idiopathic generalized epilepsy


Juvenile myoclonic epilepsy


Severe myoclonic epilepsy of infancy