Abstract
Recent genetic research has led to the identification of a significant number of genes with a major influence on epilepsy, but the genes identified so far affect risk in a very small proportion of patients – primarily those from families consistent with Mendelian modes of inheritance. Identification of genes in the genetically complex epilepsies, affecting the vast majority of patients, is a major challenge for the next decade. In this chapter we review the types of genetic mechanisms that could be involved, study designs used to identify them, and the results of recent family and genetic studies of the complex epilepsies. Given the clinical and etiologic heterogeneity of the epilepsies, understanding the relationship of genotype to phenotype is an extremely important goal for research aimed at gene identification. We review two research designs aimed at clarifying “phenotype definition” in the epilepsies: familial aggregation studies and family concordance studies. The results of these analyses may clarify the extent to which the different clinically-defined epilepsy syndromes differ with respect to their genetic contributions, providing guidance about how best to define epilepsy subgroups likely to share susceptibility genes.
Over the last two decades, more than 20 genes with a major effect on risk for human epilepsy have been identified, providing important clues to pathogenic mechanisms and enabling some patients to discover the cause of their disorder.1 However, the genes identified so far affect risk in a very small proportion of patients – primarily those from families consistent with Mendelian modes of inheritance. Most epilepsies occur in the absence of a significant family history, and identifying and characterizing the genetic mechanisms in these “complex epilepsies” is a major challenge for the next decade.2 Here we discuss the meaning of “complex inheritance” as it applies to epilepsy, findings from current research, and approaches likely to be advantageous for gene identification in these forms of epilepsy.
EPILEPSY AS A COMPLEX DISEASE
Epilepsy is familial and genetically influenced: risk is increased two- to four-fold in the first-degree relatives of people with epilepsy of unknown cause (either idiopathic generalized epilepsy [IGE] or non-lesional focal epilepsy [NFE]);3-5 twin studies consistently show higher concordance in monozygotic than dizygotic pairs;6-9 and a large number of genes with a major effect on susceptibility has already been identified.10-12 Yet most people with epilepsy have no affected relatives. For example, in the Epilepsy Family Study of Columbia University (EFSCU),4, 5, 13 we collected family history information on 1,957 people with epilepsy ascertained from voluntary organizations without regard to their family histories. The proportion of subjects with a positive family history (≥1 first-degree relative with epilepsy) was 15% in those with IGE, and 12% in those with NFE. Most of those with a family history had just one affected relative, and very few families appeared consistent with a Mendelian model.14
Several mechanisms could account for the apparent paradox that a disorder with a genetic component usually occurs “sporadically” (i.e., in the absence of a family history). First, as with many common disorders, inheritance in the majority of cases is likely to be multifactorial, with most of the genetic influence consisting of complex disease genes – that is, genetic variants that individually have a small or modest effect, and act in concert with each other and with (as yet unspecified) environmental factors in their influence on epileptogenesis. In this model, relatives will be affected only if they inherit sufficient pathogenic variants at multiple disease loci and/or multiple nongenetic risk factors. The allele frequency distribution of these contributing loci can range from common to rare with penetrances from low to moderate. We say moderate because genes with high penetrance are considered Mendelian, and tend to lead to strong segregation patterns in families. The distribution and types of genetic variants underlying complex epilepsies are unknown, but have an important influence on optimal study designs for gene identification, as we discuss below.
While family studies provide the basis for examining genes that are inherited, some cases may also have a genetic contribution but not due to inheritance. These result from de novo germline mutations and somatic mutations occurring in critical brain regions. The importance of de novo mutations is already well documented. They are extremely important in Dravet syndrome, where more than 70% of patients have SCN1A mutations and more than 95% of these arise de novo.15, 16 Such de novo mutations most often account for a high proportion of severe early onset forms of epilepsy where affected subjects seldom reproduce (and hence mutations are not generally inherited). However, rare de novo mutations have also been identified in genes identified in Mendelian epilepsies, in some isolated cases phenotypically similar to those in the families in which the genes were found.17-22 Also, copy number variants (CNVs) have recently been associated with complex epilepsies and many of these occur de novo, although associated risks appear to be moderate rather than Mendelian.23
As is true for many complex diseases, there is no clear dividing line between Mendelian and complex epilepsy genes. Many genes likely influence risk for epilepsy, with wide variation in the magnitude of their effects. As mentioned above, genes with larger effects (higher penetrance) produce Mendelian patterns of inheritance, whereas those with smaller effects (lower penetrance) produce complex patterns. The genes identified so far have generally fallen in the high range of this continuum, which is why they were easiest to identify. However, even for many of these genes, penetrance is incomplete, suggesting that other genes or environmental factors influence their effects. Also, similar pathogenic mechanisms may be involved in Mendelian and complex epilepsies, and the discovery that most of the Mendelian genes identified so far have encoded voltage-gated or ligand-gated ion channels points to this class of genes as prime candidates for the genetically complex epilepsies.
Complexities in Phenotype Definition: Lessons from Mendelian Epilepsies
Evidence from Mendelian epilepsy syndromes provides important information about complexity in the relationship of genotype to phenotype that is likely to be relevant for non-Mendelian epilepsies. First, genetic heterogeneity is extensive, and occurs at multiple levels. At the highest level, different genetic mechanisms – Mendelian, genetically complex, or even “acquired” – can cause the same syndrome, making it very difficult, if not impossible, to classify syndromes according to genetic mechanisms in any consistent way.
For example, mutations in the leucine-rich, glioma inactivated 1 gene (LGI1) are found in approximately 33% of families with autosomal dominant partial epilepsy with auditory symptoms (ADPEAF), defined as families containing two or more individuals with temporal lobe epilepsy with ictal auditory symptoms or receptive aphasia.24 However, very few patients with the clinical features that define this syndrome come from families with autosomal dominant inheritance – the vast majority are sporadic, and de novo mutations in LGI1 are found in <2% of these sporadic cases.19, 20, 25-27 The same phenotype is observed in other patients with lesional epilepsies.28 Thus, this syndrome is due to dominant mutations in LGI1 (or other unidentified genes) in a few rare cases, and genetically complex or acquired in most.
The same principle applies to the IGEs. Recurrence risk is higher in siblings than in most other forms of epilepsy; in a recent population-based study, the risk of epilepsy to age 40 was 7.6% in the siblings of incident IGE cases, compared with 4.6% for siblings of all incident cases.29 This higher recurrence risk is consistent with a range of different underlying genetic and nongenetic models. Mutations in GABRA1 or EFHC1 have been identified in some unusual families with very high incidence of juvenile myoclonic epilepsy,30, 31 and other monogenic forms of IGE may yet be discovered. However, like other forms of epilepsy, most IGEs are presumably genetically complex, caused by a mixture of genetic and nongenetic factors. In three major twin studies that examined IGEs specifically, concordance rates in monozygotic twins ranged from 65-80%, and in dizygotic twins from 0-33%.6, 7, 9
The ILAE Commission on Classification and Terminology recently recommended that the name “Genetic Generalized Epilepsy” be used instead of IGE for these disorders.32 Although the Commission explicitly stated that use of this term “does not exclude the possibility that environmental factors (outside the individual) may contribute to the expression of disease,” we believe it is misleading to incorporate the term “genetic” into the name of a disorder in the absence of Mendelian inheritance patterns in families or molecular information about the specific genes involved. Use of this term implies IGE is exclusively genetic (which is obviously incorrect based on the evidence from monozygotic twins) and suggests that “genetic” forms of generalized epilepsy can be distinguished clinically from “nongenetic” forms.
At the second level of genetic heterogeneity, within epilepsies with clear-cut Mendelian inheritance, mutations in different genes (often encoding different subunits of the same receptor) have been found to cause the same syndrome in different families (locus heterogeneity).2, 11 At the third level, in families with a mutation in the same gene, the specific molecular change usually varies among families (allelic heterogeneity).
A second complication is variable expressivity: mutations in a single gene can produce different epilepsy phenotypes in different individuals. The best example of this is genetic epilepsy with febrile seizures plus (GEFS+), in which the phenotypes within a family with a single mutation in SCN1A can range from typical febrile seizures, febrile seizures plus (i.e., febrile seizures persisting beyond age six, or accompanied by afebrile generalized tonic seizures), IGEs, temporal lobe epilepsy, myoclonic-astatic epilepsy, or Dravet syndrome.33, 34 As with complex inheritance, this variability is likely to result from the modifying effects of other genes or environmental factors.
GENE IDENTIFICATION IN THE COMPLEX EPILEPSIES
Positional cloning
The armamentarium of methods for gene identification in complex disorders includes an array of approaches (). Almost all of the genes identified so far in the epilepsies were found using positional cloning: linkage analysis followed by sequencing of the genes in the chromosomal regions with evidence for cosegregation with disease in families. This is the optimal approach for Mendelian disorders, with a long history of success for a wide range of diseases. However, it has limited power with complex disorders, where the effect of the gene sought through linkage may be too low to be detected by this approach.35 Also, practical problems complicate the design of linkage studies in complex diseases, such as the scarcity of families containing multiple affected individuals and uncertainty about how to define the phenotype and what mode of inheritance should be assumed in the analysis.
Table 1. Methods for Gene Identification in Genetically Complex Epilepsies
Substantial efforts have focused on gene mapping in the IGEs, but few findings have been replicated, probably because of a combination of phenotypic variability, genetic heterogeneity, and low statistical power. Linkage disequilibrium mapping at two linkage regions on 6p21.3 and 18q21.1 suggested two potential susceptibility genes (BRD2, ME2) predisposing to common IGE subtypes,36, 37 but some studies have not confirmed these findings38-42 and causative mutations have not yet been identified in these genes.
Allelic Association Studies
Allelic association studies are aimed at detecting genetic variants (usually single nucleotide polymorphisms, or SNPs) that are more common in people with epilepsy than in unaffected persons from the same population. They have greater statistical power for the detection of genes with small effect35and do not require families with multiple affected individuals. While most studies have been based on a case-control design, affected individuals with parents (triads) have also been used for early onset disorders, testing for transmission disequilibrium to the affected child. In the design and interpretation of case-control studies (but not triads), the potential for confounding due to population stratification must be considered. It arises when the cases and controls in a study have different genetic ancestries, and the ancestral groups differ in their allelic distributions, so that cases and controls differ in the frequency of a SNP for reasons unrelated to the disease. While triads require no formal correction for confounding due to stratification, for case-control studies several methods have been developed to control for this potential confounding, including genomic control,43 structured association tests,44 and more recently the use of principal components analysis to define the major sources of population structure, and then adjust for them in a logistic regression analysis (for discrete outcomes) or linear regression (for continuous outcomes).45 The large number of SNPs typically genotyped in a genome-wide association (GWA) study makes these approaches very powerful for detecting and adjusting for population structure. Once population stratification has been taken into account, a significantly increased frequency of a variant in people with epilepsy would suggest that it either directly affects risk or is in linkage disequilibrium with a nearby causal variant.
Most of the association studies carried out in epilepsy so far have been relatively small, candidate gene-based studies, and few findings have been confirmed. Many of the published studies have had methodologic limitations such as small sample size, lack of control for potential population stratification, and failure to adjust for multiple statistical tests.46 One large multi-site study of common variations in 279 candidate genes failed to identify clear associations across multiple sites.47 However, the association findings are more convincing for some candidate genes, such as the calcium channel subunit gene CACNA1H.
48-50
GWA studies have yielded confirmed associations of common variants with a wide range of complex diseases,51 although in most cases the associated variants explain little of the heritability of the diseases in which they have been found.52, 53 The epilepsy field has lagged behind others in the investigation of common genetic variation. Only one GWA study has been published so far and that study, which included nearly 4,000 European patients with a wide array of non-lesional and lesional focal epilepsies, was essentially negative; none of the variants examined reached genome-wide significance.54 Power was sufficient to exclude variants with an odds ratio of 1.3 or greater, and thus the results argue against common genetic effects of this magnitude that are shared across different (and likely heterogeneous) forms of focal epilepsy in European populations. Although sample sizes of specific subgroups (e.g., non-lesional as opposed to lesional focal epilepsies, in which family studies show a stronger genetic influence4) will be smaller, leading to potentially reduced power, the effect sizes are likely to be larger in these more homogeneous subgroups, and so this type of analysis is still worthwhile to pursue. GWA studies are underway in other forms of epilepsy, such as the IGEs, and their success in identifying common variants that influence susceptibility remains to be seen.
However, studies of structural variation have made progress in identifying rare genomic variants that contribute to risk for genetically complex epilepsies.23 In several studies, approximately 1% of individuals with IGE were found to carry a rare microdeletion at chromosome 15q13.3 that had previously been identified in individuals with intellectual disability, schizophrenia, and autism.55-57 Five additional candidate microdeletions that had also been found in other neuropsychiatric disorders, at 1q21.1, 15q11.2, 16p11.2, 16p13.11, and 22q11.2, were observed collectively in 1.8% of IGE cases, with a significant excess of the 15q11.2 and 16p13.11 variants.57 A genome-wide analysis of CNVs in a wide range of epilepsies (and not only IGE) found an excess of large deletions (>100 kb) at 16p13.11 in cases with highly variable phenotypes.58 Another genome-wide study of patients with a range of idiopathic epilepsy types found deletions at 15q11.2, 15q13.3, or 16p13.11 in approximately 3% of patients.59 In studies with available family data, a substantial proportion of these microdeletions have been found to be de novo, and among those that are inherited, cosegregation with epilepsy in families is inconsistent, with many affected family members failing to carry the variant present in the proband.55-57 This lack of cosegregation with disease in families, despite very high estimated odds ratios for the variants (e.g., estimated odds ratio of 68 for the 15q13.3 microdeletion),56 is paradoxical and requires explanation. The associations of these microdeletions with a wide range of neuropsychiatric disorders, including intellectual disability, autism, schizophrenia, and epilepsy, raise intriguing questions about shared pathogenic mechanisms for these disorders.
Massively Parallel Sequencing
Although the findings from GWA studies are not yet available for most forms of epilepsy, evidence from other complex disorders suggests that the individual contribution of common variants to the overall genetic component of the epilepsies is likely to be small. On the other hand, the findings from studies of structural variation suggest that rare genetic variants play an important role in the genetic architecture of the epilepsies. The potential for GWA studies to identify rare variants (i.e. less than 2% in frequency) is limited because most are not represented, either directly or indirectly (through high linkage disequilibrium), on the panel of SNP genotypes investigated. However, rare variants can now be investigated using massively parallel sequencing approaches, so-called next generation sequencing (NGS), applied either to whole genomes or to protein-coding regions (whole exome sequencing) at costs that are not prohibitive (at least on a small scale, to date).60 Several recent studies using this approach have succeeded in identifying genes for Mendelian disorders, using very small numbers of patients or families.61 Most of these studies focused on extremely rare Mendelian disorders with high penetrance – an optimal situation for disease gene detection (just as it is for linkage analysis and positional cloning). However, application of NGS is expanding rapidly to include disorders with more complex features such as locus heterogeneity and uncertainty about phenotype definition.62, 63 We are currently conducting a study of whole genome sequencing in multiplex epilepsy families64 and another major initiative is underway to develop a collaborative NGS study in the epilepsies.
One of the challenges in this field concerns the development of study designs and statistical approaches for analysis of the large number of variants identified through NGS.65 The strategy used in our current study of multiplex epilepsy families is to sequence two affected individuals in each family, selected to be as distantly-related as possible, and “connected” by family branches that also contain affected individuals. The underlying assumption is that families containing multiple affected individuals are likely to harbor genomic variants with a strong risk-raising effect, and affected individuals in the same family are likely to carry the same pathogenic variants. Since distant relatives are unlikely to share rare variants by chance, the sharing of variants between the sequenced family members can be used as a “filter” to reduce the number of potentially causative variants for further analysis. Restriction to distant relatives connected by family branches containing other affected individuals protects against the possibility that the two relatives have different genetic causes of their epilepsy.
So far, we have completed whole genome sequencing to an average coverage of 38x in two affected individuals from each of nine families with various forms of non-acquired epilepsy (average 6.2 affected per family).64 The total number of variants averaged about 4.4 million per individual, and restriction to rare, shared, potentially functional (i.e., missense, protein truncating, or splice site-disrupting) variants reduced the number to an average of 108 per family. Our planned follow-up studies include genotyping in other family members to evaluate co-segregation with epilepsy, and case-control analyses in larger cohorts.
Co-segregation with disease is the hallmark of rare, high penetrance variants, and so multiplex families such as these are indispensable for validating any such findings. However, linkage analysis in these same pedigrees did not produce convincing evidence for any particular chromosomal region(s), and so if rare, high risk variants are involved, there must be extreme locus heterogeneity to account for the negative linkage results. Rare variants with moderate to low penetrance will be more difficult to identify and validate, as these will not segregate in families, similar to common, low penetrance variants.
PHENOTYPE DEFINITION
One of the most challenging issues for genetic research on epilepsy is the extreme clinical and etiologic heterogeneity of the disorder. The epilepsies, defined broadly as recurrent unprovoked seizures, include a wide array of different syndromes presumed to have different pathogenic mechanisms. However, the extent to which the different clinical entities also differ with respect to their genetic contributions remains unclear, and hence it is uncertain which features should be used to separate the epilepsies into subgroups likely to share susceptibility genes. This lack of clear information about the relationship of phenotype to genotype dramatically impedes efforts at gene discovery because it results in samples with uncontrolled heterogeneity and reduced statistical power.66
Here we review two types of studies that have been used to advantage to elucidate shared and distinct genetic influences on different clinically defined subsets of epilepsy. Each of these approaches can also be used to study the possibility of shared genetic susceptibility to epilepsy and other disorders, such as migraine or depression.
Familial aggregation studies
Familial aggregation studies are epidemiologic designs that examine familial risks in a population context. A sample of probands with epilepsy is ascertained and divided into subsets based on syndrome or other clinical features (age at onset, seizure type, etc.). Then in the first stage of analysis, the risk of all epilepsy is examined in the relatives of different subgroups of probands and compared with the risk in the population or in the relatives of unaffected controls. The results of this analysis provide information about the relative genetic contribution to different types of epilepsy and guidance about which probands are likely to be most informative for molecular genetic studies. They also provide extremely useful information for genetic counseling.29, 67, 68 Family studies have provided evidence for several important predictors of familial risk. First, probably the most consistent observation in the genetic epidemiology of the epilepsies is the “maternal effect”: risk for epilepsy is approximately twice as high in the offspring of affected women as in the offspring of affected men.69-71 Although this phenomenon remains unexplained, several possible causes have been excluded, such as effects on offspring epilepsy risk of intrauterine exposure to antiepileptic medications or maternal seizures, or X-linked genetic models (since the risks are similar in male and female offspring).69, 70 Second, the risk of epilepsy is increased in the relatives of probands with epilepsy of unknown cause (either IGE or NFE), but not in the relatives of probands with symptomatic epilepsy (defined as epilepsy associated with structural or metabolic insults to the CNS, such as severe head trauma, stroke, brain tumor, etc.).4 This suggest either that the genetic contributions are minimal in symptomatic epilepsies, or that the genes involved raise susceptibility only in the presence of the (uncommon) risk factors (and thus the majority of relatives, in whom the risk factors are absent, do not have increased risk). Third, the increased familial risk is greatest when the relatives are younger than 35 years, and also if the proband has onset before age 35.4 Fourth, recurrence risk patterns differ depending on relationship to the proband and the proband’s epilepsy type (IGE vs. NFE). In data from EFSCU, risks in parents and siblings were greater if the proband had IGE vs. NFE.5, 72 However, this pattern was not observed in offspring, either in EFSCU5, 72 or in a population-based study in Rochester, Minnesota.73
In the second stage of analysis, the risk for specific clinically defined subtypes of epilepsy (or other disorders) is examined in the relatives of specific subgroups of probands. If different genes influence risk for different types of epilepsy (distinct genetic effects), risk in the relatives should be increased only for the same type of epilepsy as in the proband. In contrast, if the same genes influence risk for different types of epilepsy (shared genetic effects), the increased risk in the relatives will not be restricted to the same type as in the proband. In two previous family studies, risk for NFE was significantly increased in the relatives of probands with IGE, providing evidence for shared genetic influences on IGE and NFE.72, 73
Family concordance studies
Hypotheses about shared and distinct genetic influences on different clinically defined subsets of epilepsy can also be tested using family concordance studies.74-76 As opposed to familial aggregation studies in which a systematic sampling scheme is used to ascertain probands, these studies assess the concordance of epilepsy types (syndromes, seizure types, or subsets defined by other clinical features) in sets of families containing multiple affected individuals, which are often ascertained unsystematically. The rationale for family concordance analysis is that if some of the genetic influences on different epilepsy types are distinct, families will tend to be concordant – i.e., the proportion of families in which all affected individuals have the same type of epilepsy will exceed that expected by chance. The expected proportion of concordant families is assessed using a permutation approach, taking into account the overall proportion of individuals with a given clinical feature in the dataset, the number of affected individuals in each family, and the proband’s epilepsy type.74 The results of research using this approach have provided evidence for distinct genetic influences on IGE and NFE,75 and within the IGEs, for distinct genetic influences on myoclonic and absence seizures.76-78
With regard to the shared and distinct genetic influences on IGE and NFE, the results of familial aggregation studies and family concordance studies appear to differ, but this may be an artifact of study design. In familial aggregation studies, the null hypothesis is that all genetic effects are distinct. This hypothesis is rejected (leading to the conclusion that some genetic effects are shared) when a significant increase in risk in relatives is observed for an epilepsy type different from that in the proband. In family concordance studies, the null hypothesis is exactly the opposite: that all genetic effects are shared. This hypothesis is rejected (leading to the conclusion that some genetic effects are distinct) when families show excess concordance for a specific epilepsy type. Taken together, the results of these two study designs indicate that there are both shared and distinct genetic influences on generalized and focal epilepsies, as has already been observed, for example, in families with GEFS+.
RECOMMENDATIONS FOR FUTURE STUDIES
What approaches are most promising for future studies of the genetic contributions to the complex epilepsies? First, efforts to identify, clinically characterize, and carry out positional cloning efforts in families with Mendelian epilepsies should continue. Although Mendelian epilepsy syndromes are rare, identification of the genes that cause them can provide extremely important information about basic epileptogenic mechanisms and suggest candidate genes to be investigated in complex epilepsies. GWA studies already underway should soon provide evidence about the importance of common variants in the epilepsies. Current evidence suggests that additional rare variants are likely to contribute to genetic susceptibility in epilepsy, and NGS has the greatest promise for identifying these. However, the failure to identify additional loci harboring high-risk variants through linkage analysis, in numerous multiplex families, suggests either a limited role for additional high penetrance mutations or extreme locus heterogeneity. Collaborative efforts, rigorous phenotyping, and study designs that allow analysis of subgroups predicted to share susceptibility genes are crucial to maximize the likelihood of success.
