Genetics of inherited human epilepsies.

Major advances have recently been made in our understanding of the genetic basis of monogenic inherited epilepsies. Progress has been particularly spectacular with respect to idiopathic epilepsies, with the discovery that mutations in ion channel subunits are implicated. However, important advances have also been made in many inherited symptomatic epilepsies, for which direct molecular diagnosis is now possible, simplifying previously complex investigations, it is expected that identification of the genes implicated in familial forms of epilepsies will lead to a better understanding of the underlying pathophysiological mechanisms of these disorders and to the development of experimental models and new therapeutic strategies, in this article, we review the clinical and genetic data concerning most of the inherited human epilepsies.

Epilepsies are frequent heterogeneous disorders[1] and are caused by many factors.[2] The contribution of genetic and environmental factors varies among epileptic disorders. Genetic factors are generally thought to contribute to the etiology of 40% to 60% of human epilepsies.[2,3] Inherited epilepsies are usually classified according to whether the mode of inheritance is complex or monogenic. In epilepsies with a complex mode of inheritance, epilepsy results from the interaction between environmental factors and genetic susceptibility, whereas in monogenic epilepsies, the genetic component is prevalent, although environmental factors may contribute to phenotypic expression and could explain incomplete penetrance or variable clinical expression. Finally, in epilepsies caused by exogenous factors (the least genetically determined of the epilepsies) , genetic susceptibility could explain why only some of the individuals exposed to the same factors later develop epilepsy

Genetic studies in epilepsies are difficult to perform for several reasons. First, most epilepsies have a complex mode of inheritance and it is difficult to identify the genes involved. Nonparametric analyses in a large number of affected individuals (ie, hundreds) are necessary. However, difficulties are also encountered in genetic studies of monogenic epilepsies, particularly in the identification of large informative families with enough affected members to be useful for linkage analysis. Second, phenotype analysis can be problematic. The clinical status (ie, affected or not) of each member of the family must be determined. This involves a choice of more or less stringent electroclinical criteria to confirm the presence of the disease. The collection of reliable medical information may be difficult, especially in the first generation of affected families. Moreover, the presence of phenocopies (which are frequent for epilepsy and febrile convulsions) and possible intrafamilial phenotypic heterogeneity must be taken into account.

Despite these difficulties, major advances have been made in the genetics of epilepsy in the past 10 years. Nearly all concern epilepsies with a monogenic mode of inheritance, the least frequent of the inherited epilepsies. The progress in idiopathic epilepsies has been spectacular, with the discovery that some of them may involve mutations in ion channels, leading to the concept of “channelopathies.” However, important advances have also been made in symptomatic epilepsies, with the discovery, for example, of genes implicated in neuronal migration and various metabolic pathways.

It is expected that elucidation of the genetic basis of monogenic epilepsy will also help us understand the genetic basis of epilepsies with complex inheritance.

In this article, we review recent advances in the genetics of epilepsy, focusing on the molecular and pathophysiological aspects of some inherited epilepsies.

Idiopathic epileptic syndromes

It has long been suspected that genetic factors are prevalent in the etiology of idiopathic epilepsies. Most are characterized by a complex inheritance - idiopathic epilepsies with monogenic inheritance are rare. Those in which a locus or genes have been identified are listed in Table I. [4-46] For some of these, voltage- or ligand-gated ion channels are implicated.

Idiopathic epileptic syndromes with monogenic inheritance: the new concept of channelopathies

To date, three familial idiopathic syndromes have been found to be mediated by mutations in voltage- or lig-and-gated ion channels.

Autosomal dominant nocturnal frontal lobe epilepsy

The syndrome of autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE) was first described by Scheffer in 1994.[47,48] It is characterized clinically by the onset in infancy of frequent brief partial seizures occurring in clusters during sleep. Adult onset is less common. The motor component of seizures predominates (paroxys mal dystonic postures, thrashing, ambulation). Sometimes, the symptoms are limited to sudden awakening. Vocalizations or aura may precede the motor manifestations. Misdiagnoses are frequent, especially confusion with parasomnias (night terrors, somnambulism). Seizures usually persist in adults, but tend to be less frequent and respond to carbamazepine. Intrafamilial variations in severity are sometimes observed. Neuroimaging is normal. When ictal electroencephalography (EEG) recordings are interprétable, they show unilateral or bilateral frontal/temporal epileptic activity.

Familial studies of this rare new syndrome demonstrated autosomal dominant transmission with incomplete penetrance. One locus was found in the region 20ql3.2 by linkage analysis in a large Australian pedigree.[4] The CHRNA4 gene encoding for the alpha-4 subunit of the neuronal nicotinic acetylcholine receptor (nAChR), which has already been found in this genomic region, was a good candidate. Indeed, subsequent screening of the CHRNA4 gene in the first ADNFLE Australian family described led to identification of a mutation in this gene.[5] Other mutations of the CHRNA4 gene were subsequently detected in several families.[6-8]

nAChR receptors are heteropentameric ligand-gated ion channels. The genes for eight human nAChR subunits have been mapped. The alpha-4 subunit is expressed in all layers of the frontal cortex. The second transmembrane domain of the alpha-4 subunit is crucial to the permeability of the ion channel. Mutations of the alpha-4 subunit are thought to decrease the activity of nAChR by reducing its affinity for acetylcholine and permeability to calcium.[49,50] Neuronal nicotinic receptors are thought to be almost exclusively presynaptic, regulating the release of neurotransmitters such as glutamate. However, the mechanism by which hypoactive nAChRs cause this partial familial epilepsy is unknown.

Another ADNFLE locus has been found in the region 15q24 in one family.[9] Although this region is close to a cluster of genes encoding other nAChR subunits (CHRNA3, CHRNA5, and CHRNB4), mutations have not been found in these subunits, and the causative gene remains to be identified.

A third locus was recently identified in the pericentromeric region of chromosome l,[10] with the subsequent identification of mutations in the beta-2 subunit of nAChR (CHRNB2).[11,12] However, most ADNFLE families are not linked to CHRNB2 or CHRNA4. [51]

Bening familial neonatal convulsions

The syndrome known as benign familial neonatal convulsions (BFNC) is characterized by the occurrence of feature of unilateral or bilateral clonic, apneic, or even tonic seizures on the second or third day of life of a normal neonate. Interictal EEGs rarely show what is described as a “sharp alternating theta.” The outcome is generally favorable, although some patients will develop febrile seizures or nonfebrile seizures later in life. This familial syndrome differs in several respects from the sporadic form (benign neonatal convulsions), in which tonic seizures are never observed, the typical interictal EEG feature of a “sharp alternating theta” is more frequent, and outcome is more favorable.

BFNC was the first idiopathic epilepsy in which genetic linkage was established,[25] first to the q arm of chromosome 20, [25,26] and then to the q arm of chromosome 8.[30] Mutations in novel voltage-gated potassium channel genes KCNQ2 (region 20q)[27-29] and KCNQ3 (region 8q)[31] were found in this familial syndrome, but the existence of one or more loci is suspected. Most families are linked to KCNQ2. [31] Only one KCNQ3-linked family has been published to date. KCNQ2 and KCNQ3 are heteromeric channels with highly homologous sequences. They are predominantly expressed in all regions of brain and are functionally associated, contributing to the M current that regulates excitability of many neurons.[28,52] As demonstrated by in vitro studies, the identified mutations cause a minor loss of function of the channels.[28,29,53]

The age-dependence of this familial syndrome may result from difference in the cerebral expression of the potassium channel genes over time.[54] Interestingly mutations in KCNQ1, a voltage-gated potassium channel gene that is expressed in the heart and ear and is homologous to KCNQ2 and KCNQ3, cause two other familial syndromes: the long-QT syndrome and Jervell-Lange-Nielsen cardioauditory syndrome.[55,56]

Generalized epilepsy with febrile seizures-plus syndrome

Febrile seizures are frequent events, the genetic component of which is important. In some families, febrile seizures are associated with nonfebrile seizures, constituting the syndrome described in 1997 as generalized epilepsy with febrile seizures-plus (GEFS+).[57] In this heterogeneous familial phenotype, some affected members often have multiple febrile seizures that persist beyond the age of 6, whereas other family members have classic febrile seizures that disappear before the age of 6. Variable nonfebrile seizures are also observed. Initially, generalized seizures (tonic-clonic, myoclonic, atonic, and absence seizures) were described,[57] but hemiconvulsive, temporal, or frontal seizures were later observed in other families.[42-44,58] These afebrile seizures may begin in childhood in association with febrile seizures, after a seizure-free period, or later in life. Furthermore, not all affected members have febrile seizures. Several types of seizure can coexist in a given patient with electroclinical features that are more or less typical of generalized idiopathic epilepsies or myoclonic astatic epilepsy (Doose syndrome) , but electroclinical patterns that do not correspond to the international classification of epilepsies are also observed.[59] Some patients are intellectually disabled.[42] Outcome and response to treatment are very variable within the same family. When available, neuroimaging is normal.

GEFS+ is transmitted as an autosomal dominant trait with incomplete penetrance, and is genetically heterogeneous. The first locus was found in the region 19ql3.1, and a mutation in the SCN1B gene coding for the beta 1 subunit of the neuronal voltage-gated sodium channel was found in one family.[36] A second locus in region 2q21-q33 seems to be more frequently implicated, according to published reports in several families.[42-45] In two French families, two different mutations were identified in the SCN1A gene, which encodes for the alpha- 1 subunit of the same voltage-gated sodium channel.[46] Functional studies in Xenopus oocytes have demonstrated that mutations in the beta-1 and alpha1 subunits interfere with the functional properties of the sodium channel.

A third locus is suspected because some GEFS+ families are not linked to SCN1A or SCN1B. [36,46]

Idiopathic epilepsies with complex inheritance

Most idiopathic generalized epilepsies (including juvenile myoclonic epilepsy, juvenile absence epilepsy, childhood absence epilepsy, and epilepsy with tonic-clonic seizures on awakening) have a complex mode of inheritance. These diseases result from an interaction between genetic susceptibility (often mediated by several genes) and environmental factors. Linkage to the q arm of chromosome 8,[60,61] and the p arms of chromosomes 1[62] and 3[63] have been reported for generalized epilepsies. Because confirmatory reports in additional families have not been forthcoming, these results should be considered with caution.

Juvenile myoclonic epilepsy has been studied most extensively, with controversial findings concerning linkage to the regions 6p[64-69] and 15q14.[70]

Most febrile seizures and benign rolandic epilepsy are also thought to have complex modes of inheritance. Linkage to the q arm of chromosome 15 was suggested for benign rolandic epilepsy in one study.[71]

Inherited developmental cortical malformations (neuronal migration disorders)

These developmental disorders are an important cause of pharmacoresistant epilepsy, which is often associated with mental retardation.

Lissencephaly and double cortex syndrome

Lissencephaly is a rare disorder characterized by a reduced number of cerebral gyri due to an arrest of neuronal migration at 8 to 14 weeks of gestation.[72] The cortex is abnormally thick and the surface of the brain is smooth. Microscopically, the cortex is poorly organized with four to six primitive layers and diffuse neuronal heterotopia. Affected children have severe mental retardation, and often pharmacoresistant epilepsy and other neurological abnormalities. Various types of seizures (tonic-clonic, myoclonic, and tonic seizures, and infantile spasms) occur early in life. Lissencephaly can be isolated, as in isolated lissencephaly sequence or in hemizygous males affected with X-linked lissencephaly. However, in Miller-Dieker syndrome, lissencephaly is associated with facial dysmorphism.

In Miller-Dieker syndrome, and in around a third of patients with isolated lissencephaly sequence, a heterozygous deletion or mutation has been demonstrated in the LIS1 gene, which is located in the region 17pl3.3.[73-75] The LIS1 gene is ubiquitously expressed and encodes a noncatalytic subunit of platelet activating factor (PAF) acetylhydrolase, an enzyme that inactivates PAF.

In males affected with X-linked lissencephaly, an Xlinked dominant inherited disease, the gene involved is DCX, which encodes doublecortin and is located in the region Xq22.3-q23.[76,77] Interestingly, in females, the same mutations in the DCX gene lead to another phenotype, the double cortex syndrome, which is characterized by a laminar cerebral heterotopia.[76,77] Affected women have pharmacoresistant epilepsy, but are less mentally retarded than affected males.

More recently, rare cases of double cortex syndrome have been reported in men with mutations in the LIS1 or DCX genes.[78,79] The LISl and DCX gene products interact and interfere with dynamic properties of microtubules. The exact mechanism that underlies abnormal neuronal migration has not been elucidated.

Familial periventricular heterotopia

Periventricular heterotopia is characterized by the lining of the ventricular walls with nodules that consist of neurons that did not migrate to the cortex during brain development. X-linked periventricular heterotopia is lethal to males during the embryonic period. Affected females have epilepsy without mental retardation, associated with persistent ductus arteriosus, coagulopathies, and skeletal abnormalities. The causative gene is FLN1, which is located in the region Xq28[80] and encodes filamin 1, an actin-binding protein that interacts with other proteins of cytoskeleton.

Progressive myoclonus epilepsies

Progressive myoclonus epilepsies (PMEs) are rare disorders that have some clinical features in common, but different etiologies and variable outcomes. A specific diagnosis for some of these diseases has been possible for a long time, on the basis of characteristic stigmata detected by pathological investigation. Numerous advances in genetics now permit direct molecular diagnosis in most cases. We will focus here on the genetic bases of Unverricht-Lundborg disease and Lafora's disease. Other PMEs with their corresponding loci and genes are listed in Table II. [81-120]

Unverricht-Lundborg disease

Unverricht-Lundborg disease is an autosomal recessive PME classically with onset between 6 and 15 years of age, a slow progression, rare, late, and mild mental deterioration, and cerebellar ataxia.[121,122] However, more dramatic outcomes have been described, often precipitated by phenytoin prescription.[123] More recently, late-onset forms of the disease have been reported.[124] Both the Baltic and Mediterranean forms of the disease are caused by mutations in the cystatin B gene located in the region 21q22.3.[125,126] Rare point mutations and deletions in the coding region of the gene[81-84] lead to a loss of function of cystatin B. More frequently, expansion of a dodecamer (CGC CGC CCC GCG)n repeat in the 5' untranslated region of the gene[85-88] decreases transcription. Normal alleles contain two to three copies of the dodecamer, whereas mutant alleles contain more than 30 repeats of the dodecamer. Preliminary studies have not provided evidence of a correlation between the size of the dodecamer expansion and age at onset of the disease.[88] There are probably premutation states, since intermediate size alleles with 12 to 17 dodecamer repeats have been detected in individuals with normal phenotype who were able to transmit pathologic alleles to their offspring.[86]

The presence of these two types of mutations varies according to the geographic origin of affected families. The Baltic form of the disease is generally caused by a point mutation in one copy of the cystatin B gene and expansion of the dodecamer in the other copy or, more rarely, by point mutations in both copies of the gene.

The Mediterranean form of the disease, characterized by frequent consanguinity, results from expansion of the dodecamer on both copies of the cystatin B gene.

Cystatin B is a cystein-protease inhibitor that is thought to protect against apoptosis, but the mechanism leading to Unverricht-Lundborg disease remains to be elucidated.

Lafora's disease

Lafora's disease is an autosomal recessive PME characterized by onset between age 10 and 18, rapid neurological and cognitive decline, and fatal outcome after about 10 years of progression. Focal occipital seizures are frequent.[127] Until recently, diagnosis was established by observation of intracellular polyglucosan inclusions (Lafora bodies) on skin biopsies.[128] Direct molecular diagnosis is now possible.

Linkage analysis and homozygosity mapping localized the gene in the region 6q23-25.[129,130] The gene, identified by positional cloning,[89,90] encodes a protein tyrosine phosphatase, laforin, which is a tyrosine kinase inhibitor. Laforin is thought to be involved in glycogen metabolism. Homozygous deletions and several homozygous point mutations in the coding part of the gene have been found in affected families.[89,90] At least one other locus is probably also responsible for Lafora disease.[131,132]

Inherited neurologic disorders and chromosomal disorders with epilepsy as a part of the phenotype

Epilepsy is observed among complex neurological or extraneurological symptoms in numerous chromosomal disorders and inherited disorders affecting the central nervous system. They cannot be described in detail in this review and most are listed in Tables III and IV. [106,133-147] The frequency of epilepsy in these complex syndromes is variable.

Conclusion

Genetic studies of previously well-defined epileptic syndromes have led to the identification of causative genes in some cases, but also to the identification of new familial epileptic syndromes that are not yet included in the international classification of epilepsies and epileptic syndromes.[59] In the future, this classification will probably take into account these new familial epileptic disorders with their particular electroclinical features and prognoses.

The genetic heterogeneity of epilepsies is becoming more and more apparent. Different genes, which may or not be functionally linked, and different mutations may cause the same familial epileptic syndrome. At the same time, significant intrafamilial phenotypic heterogeneity can often be observed. This is particularly clear in the GEFS+ syndrome. One hypothesis is that the expression of the mutated genes differs among family members, causing clinical heterogeneity. Alternatively, the gene may intervene in epileptogenesis at a very general level, affecting epileptic susceptibility or modulating the epileptogenic threshold, and other genetic or environmental factors may influence the electroclinical profile of the disease in each affected subject.

There are many pathophysiological mechanisms underlying inherited epilepsies. The functional or morphological consequences of the mutations that give rise to an epileptic process are extremely variable. The discovery of dysfunction of ion channels in several idiopathic epilepsies has led to the concept of channelopathies, but abnormal neuronal migration, premature neuronal e death, metabolic disturbances, and other anomalies may also be involved.

Finally, progress in the genetics of human epilepsies has e had important consequences for clinical practice. Spell cific molecular diagnosis is now possible in symptomatic e individuals for several diseases, some of which have poor prognoses. Predictive diagnosis in presymptomatic indie viduals is also possible, although it does pose ethical problems. From a pharmacological point of view, these recent genetic discoveries should help understand the response (or resistance) of some epileptic syndromes to ri treatment and the adverse effects sometimes observed with antiepileptic drugs, and generate new antiepileptic drugs.

Table I

Genetics of idiopathic epilepsies with a monogenic mode of inheritance. AD, autosomal dominant; AR, autosomal recessive, aSeveral modes of inheritance have been described for familial febrile convulsions: polygenic inheritance seems to be prevalent; however, autosomal dominant transmission with incomplete penetrance or autosomal recessive transmission have been described for some families.

Disorder	Mode of inheritance	Chromosomal region, gene/protein
Partial idiopathic epilepsies
• Autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE)	AD	• 20q, CHRNA4[4-8]
		15q, gene?[9]
		1 (pericentromeric region),
		CHRNB2[10-12]
• Familial lateral temporal epilepsy with auditory symptoms	AD	• 10q, gene?[13-15]
• Familial mesiotemporal epilepsy	AD	• Locus?[16,17]
• Autosomal dominant partial epilepsy with variable foci	AD	• 2qter, gene?[18]
		22q, gene?[19]
• Benign familial infantile convulsions (BFIC)	AD	• 19q, gene?[20]
		Other locus?
• Infantile convulsions with paroxysmal choreoathetosis (ICCA)	AD	• 16p, gene?[21,22]
• Familial rolandic epilepsy with paroxysmal	AR	• 16p, gene (only one family
exercise-induced dystonia and writer's cramp		published)[23]
• Familial rolandic epilepsy with speech dyspraxia	AD with anticipation	• Locus? Expansion of
and mental retardation		trinucleotidic repeat is suspected
		(only one family published)[24]
Generalized idiopathic epilepsies
• Benign familial neonatal convulsion (BFNC)	AD	• 20q, EBN1, KCNQ2[28-29]
		• 8q, EBN2, KCNQ3[30,31]
		Third locus?
• Familial cortical tremor	AD	• 8q23.3-q24.1, gene?[32-34]
(or benign adult familial myoclonic epilepsy)		Other locus?
Familial febrile convulsions	Variable mode	• 8q13-q21, FEB1, gene?[35]
	of inheritancea	19p13.3, FEB2, gene?[36,38]
		11q, gene[39]
		5q14-q15, FEB4, gene?[40]
Generalized epilepsy with febrile seizures-plus (GEFS+)	AD	•19q13.1, SCN1B[41]
		2q21-q33, SCN1A[42-46]
		Third locus?

Table II

Inherited progressive myoclonus epilepsies. AD, autosomal dominant; AR, autosomal recessive, aProgressive myoclonic epilepsy may be a clinical form of the disease.

Disorder	Mode of inheritance	Locus, gene, protein
Unverricht-Lundborg disease	AR	• 21q, EMP1, cystatin B[81-88]
Lafora's desease	AR	• 6q, EMP2A, laforin[89-90]
		Other locus?
Neuronal ceroid lipofuscinoses
• Early infantile form, late infantile form,	AR	• 1p32, CLN1, lysosomal palmitoyl-
and variant of juvenile form, all with		protein thioesterase I[91,95]
cytoplasmic granular osmiophilic deposits
• Classic late infantile form	AR	• 11p15, CLN2, tripeptidyl
		peptidase I[96-98]
• Variant of late infantile form	AR	• 15q21-23, CLN6,[96] gene?
• Finnish late infantile form	AR	• 13q21-32, CLN5, new protein
		of unknown function[99]
• Turkish late infantile form[100]	AR	• Locus, CLN7? gene?
• Juvenile form	AR	• 16p12 (CLN3), novel protein
		involved in lysosomal pH
		regulation[101-103]
Adult form (Kuf' disease)[104]	AD	• Locus (CLN4)? gene?
Myoclonus epilepsy and ragged-red fibers	Maternal transmission	• Mitochondrial genome
(MERRF) syndrome		8344 tRNALys is the prevalent
		mutation[105,107]
Sialidoses	AR	•6p, α-neuraminidase[108,109]
		20q, stabilizing protein of the
		α-neuraminidase-β-
		galactosidase complex[110]
Juvenile form of Gaucher's disease	AR	• 1q, β-glucocerebrosidase[111,112]
Juvenile form of GM2 gangliosidosis	AR	• 15q23-24, β-hexosaminidase A
		α-subunit gene[113,114]
Dentatorubral-pallidoluysian atrophya	AD	• 12p; atrophin[115]
Huntington's diseasea	AD	• 4q, huntingtin[116-117]
Familial form of Alzheimer's diseasea	AD	• 14q, presenilin 1[118,119]
Mutation in neuroserpin gene a (one family published)	AD	• 3q26, neuroserpin[120]

Table III

Principal inherited disorders with epilepsy as a part of phenotype. AD, autosomal dominant; AR, autosomal recessive. *With unusual characteristics: the mutation can be passed through phenotypically normal males (norma! male carriers) and their daughters are almost never affected. In contrast, 30% of carrier females are mentally retarded.

Disorder	Mode of inheritance	Locus, gene, protein
Tuberous sclerosis	AD	• 9q34, TSC1, tuberin[133,134]
		16p13.3, TSC2, hamartin[133,135]
Type 1 neurofibromatosis	AD	• 17q11.2, NF1, neurofibromin[136]
Familial cerebral cavernomas	AD	• 7q, KR1T1 gene[137-139]
		7p
		3q
Rett's syndrome	Dominant X-linked	• Xq28, MECP2 gene[140]
Mitochondrial myophaty, encephalopathy, lactic	Maternal transmission	• Mitochondrial genome: 3243
acidosis and stroke-like episodes (MELAS)		tRNALeu is the prevalent
		mutation[106,141]
Fragile X syndrome	Dominant X-linked	• Xq27.3, FMR1 and FMR2 genes[142-144]
Some types of gangliosidosis	AR	• Variable[145-147]

Table IV

Principal chromosomal disorders associated with epilepsy.

• Trisomy 21 (Down's syndrome)

• Angelman syndrome (partial monosomy 15q11)

• Trisomy 12p

• Wolf-Hirschhorn syndrome (partial monosomy 4 p)

• Klinefelter's syndrome (XXY)

• Ring chromosome 20