Gene Expression, Genetics, and Genomics in Epilepsy: Some Answers, More Questions


Address correspondence and reprint requests to Dr. Peter B. Crino, Department of Neurology, University of Pennsylvania, 3 West Gates Bldg., 3400 Spruce St., Philadelphia, PA 19104. E-mail:


Summary:  The rapid technical progress made in molecular genetics has provided new strategies to study the molecular pathogenesis of human epilepsy. In particular, the abilities to assay the expression of many thousands of genes simultaneously with cDNA or oligonucleotide arrays and to rapidly screen thousands of DNA basepairs permits exciting insights into how human epilepsy may result from alterations in gene transcription and sequence. These approaches can show how monogenic and even complex genetic disorders lead to network alterations and seizures. Most recently, investigation of single nucleotide polymorphisms (SNPs) has shown that even subtle alterations in gene sequence across the genome can raise or lower seizure threshold. Clearly, there is a complex interplay between gene expression, genetics, and genomics which ultimately leads to seizure onset and epilepsy. Identifying the contribution that each plays in epileptogenesis may help define new therapeutic targets.

Over the past two decades, there has been a revolution in understanding the molecular pathogenesis of human epilepsy. In fact, the mere observation that epilepsy can result from genetic and more recently, genomic, influences, has been a major advance. We now know that no fewer than eleven epilepsy syndromes are monogenic disorders resulting from mutations in known ion channel genes (Meisler et al., 2001; Avanzini and Franceschetti, 2003; Noebels, 2003). Moreover, there are numerous autosomal or X-linked genetic disorders in which epilepsy figures prominently as a phenotypic feature (Crino et al., 2002; Barkovich et al., 2005). In many of these disorders, there are associated malformations of cortical development while for syndromes resulting from ion channel gene mutations, the brain cytoarchitecture is presumably intact.

While the discovery of clear links between altered gene sequence and seizures are reported with increasing frequency, the molecular mechanisms that underlie epilepsy pathogenesis remain a challenge to define. For example, we now know that mutations in a single gene, i.e., SCN1A (Meisler et al., 2001; Avanzini and Franceschetti, 2003; Noebels, 2003) can lead to distinct epilepsy syndromes, i.e., GEFs plus (Scheffer et al., 2005) or SMEI (Mulley et al., 2003), (so-called variable expressivity) and in some cases, mutations in multiple genes, i.e., TSC1 or TSC2, Lis-1, lead to a common clinical phenotype such as infantile spasms (so-called locus heterogeneity). Thus, other factors may account for the phenotypic distinctions between individual epilepsy syndromes. Moreover, epilepsy in the majority of individuals is a sporadic disorder associated either with environmental precipitants, i.e., brain tumors, stroke, or trauma or from unidentified causes. Are there genetic influences for these epilepsies as well? Of pivotal importance are two additional molecular arenas in which genes play into epileptogenesis. First, there is mounting evidence that altered gene expression is associated with recurrent seizures and second, recent studies have shown that gene sequence polymorphisms, i.e., common sequence variations in common gene alleles found throughout the population, may be associated with a lowered threshold for seizures.

This review will address three molecular approaches to investigate epilepsy pathogenesis including studies of gene expression, gene mutations, and polymorphic alterations in gene sequence. Each approach provides new insights into epilepsy but at the same time leaves more questions unanswered. Perhaps the most compelling question is how alterations in gene expression or sequence (mutation or polymorphisms) actually lead to seizure onset, that is, by a direct effect on encoded protein function, via an effect on a downstream cascade, or both?


There are numerous animals systems to study epilepsy and epileptogensis (for reviews, see Yang and Frankel, 2004; Stafstrom and Sutula, 2005; Pitkänen and McIntosh, 2006; Cortez et al., 2006). The analysis of gene expression, gene mutations, and genetic backdrop can also be assayed in animal model systems. Some models involve pharmacological treatment of an animal with compounds such as kainate or pilocarpine to induce seizures while others result from electrical stimulation or traumatic injury to the brain. These models have been used to study cell proliferation, cell death, changes in receptor pharmacology, and cellular excitability. Perhaps most compelling are the models that result from gene knockout strategies. The most common genes that lead to experimental epilepsy models are ion channel genes including sodium, potassium, and calcium channel genes. These models are particularly relevant to human epilepsy since several human epilepsy syndromes also result from single ion channel gene mutations. Several important questions arise however regarding the mechanism in which these gene defects or knockouts lead to epilepsy. First, each knockout tends to aim for a maximal inactivation by gene deletion. However, in human channelopathies or other genetic disorders associated with epilepsy, there may be a broad ranges of gene mutations from deletions that completely remove a gene to subtle missense mutations associated with amino acid sequence alterations which only diminish but not fully inactivate gene function (Mulley et al., 2005). Thus, from a molecular perspective animal model systems do not necessarily fully recapitulate the range of molecular events in human epilepsy syndrome that lead to seizures. Second, it is clear that loss of function in a particular ion channel leads to altered membrane potential and excitability. While the loss of gene function is likely the primary cause for seizures in these animals, the effects of this single gene defect and subsequent seizures lead to changes in expression of a variety of other genes and proteins. Thus, it is often difficult to decipher the differential contributions of transcriptional or translational alterations in make to seizure onset and recurrent seizures. Finally, the genetic background of the animal models is generally homogeneous whereas the background of humans is not. Thus, a pivotal question is to what extent a heterogeneous genomic background, i.e., SNP patterns, found in humans, may contribute to the full phenotype of any given epilepsy syndrome (see in the section, Genomics). For example, a gene mutation that leads to seizures in a human might not cause seizures in an experimental animal because the particular genomic backdrop of the human also contributes to seizure susceptibility. Conversely, a gene mutation in mice that results in seizures may have no phenotype in humans because other differences in gene sequence can compensate for or counteract these effects. These considerations have obvious relevance to the generation of animal epilepsy models as well as the use of these models for antiepileptic drug development.


It is now virtually axiomatic that seizures can cause alterations in gene expression (Morgan et al., 1987; Gall et al., 1991; Newton et al., 2003). Altered gene expression is a measure of the differences in relative abundance of cellular mRNAs that code for proteins. The effects of seizures on gene expression have been defined in both in vitro and in vivo systems by numerous approaches including in situ hybridization, Northern analysis, RT-PCR, and gene array assays. Early studies used in situ hybridization to define the expression of individual mRNAs (Gall et al., 1991) whereas current technologies use gene arrays to define differential expression of thousands of genes at once (Elliot et al., 2003). The first study using gene array technologies as a strategy to assay coordinate changes in brain gene expression in a human epilepsy syndrome was performed in resected tissue from tuberous sclerosis complex (TSC) patients (Crino et al., 1996) and subsequent analysis of a variety of human epilepsy subtypes such as temporal lobe epilepsy (Brooks-Kayal et al., 1998; Becker et al., 2003), focal cortical dysplasia (Crino et al., 2001; Kim et al., 2003; Baybis et al., 2004), and neocortical epilepsy (Rakhade et al., 2005) have provided new insights into transcriptional alterations associated with epilepsy. In fact, expression of genes across broad families such as growth, angiogenesis, and transcription factors, cell-signaling molecules, cytoskeletal elements, neurotransmitter receptor subunits and uptake sites as well as ion channels is altered following acute or chronic seizures (Becker et al., 2003; Elliot et al., 2003; Lukasiuk et al., 2003). Clearly, in experimental animals seizures can induce almost instantaneous changes in gene transcription as has been shown for the immediate early genes c-fos and c-jun as well as more prolonged changes in gene transcription over more prolonged time periods. In addition, gene expression changes have been reported in the pilocarpine seizure model following status epilepticus (Elliot et al., 2003) or following electrical stimulation models (Lukasiuk et al., 2003) suggesting that early transcriptional alterations may be the first steps that lead to seizure onset.

From a mechanistic perspective, it is tacitly assumed that altered mRNA levels predict similar changes in functional protein expression that disrupt normal cell function, although this relationship is not universal. Indeed, seizure-induced changes in mRNA expression are associated with long-lasting changes in protein expression that have functional relevance in terms of cellular architecture, network reorganization, cell proliferation, cell death, and perhaps most interesting, fostering recurrent seizures. Presumably, alterations in, for example, ion channel or neurotransmitter receptor subunits can lead to changes in neuronal excitability that may be linked with seizure recurrence and that enhanced expression of cell death pathway genes is associated with apoptosis in the hippocampus. Altered gene expression also highlights differences in gene transcription, that is, how genomic DNA is converted to mRNA. This process is governed by a complex assortment of transcription factors, repressors, and enhancers that regulate mRNA expression. Thus, if the levels of a particular mRNA are increased following seizures, there may be numerous mechanisms by which the increase occurs. In either scenario, the “holy grail” of gene expression analysis studies in epilepsy has been to define either a single gene or a panel of genes that are the first steps in the epileptogenesis cascade. The therapeutic implications of these essential genes are numerous.

An important consideration is that altered gene expression may actually antedate seizure onset. i.e., that a particular transcriptional pattern leads to seizures. How these changes are initiated and how they may lead to seizures are pivotal, yet sadly, unanswered questions. It is conceivable that only a limited number of genes may be required for seizure initiation. For example recent data suggests that even in divergent epilepsy types, there may be a common “signature” of genes exhibiting altered expression (Rakhade et al., 2005).

Finally, an alternative hypothesis is that altered mRNA levels following seizures represent epiphenomenal changes that are associative rather than predictive or mechanistic or that mRNA expression changes reflect the effects of a larger network disruption. This is unlikely for several reasons. First, in monogenic epilepsy syndromes resulting from loss of function or deletion mutations, altered levels of the mutant or deleted gene transcript have been reported suggesting that transcript level of expression is significant and biologically relevant. Second, changes in gene expression in monogenic epilepsy syndromes likely represent downstream effects of the mutant gene and thus are very likely to be part of a cascade of events leading to seizure onset. Third, corroborative changes in gene expression have been reported in mouse models of human epilepsies and human epilepsy tissue. For example, in TSC changes in several genes such as nestin, brain lipid binding protein, and vimentin, have been identified in the Tsc1 conditional knockout mouse and in human tuber specimens (Ess et al., 2004).

A new development over the past decade has been the implementation of technologies that permit analysis of gene expression in single cell types (Crino et al., 1996). Microdissection of individual cells provides a strategy to greatly enhance the ability to define gene expression changes in single cells types and remove contamination of expression results by heterogeneous cell populations. For example, while analysis of mRNA expression in whole hippocampi can provide a broad overview of differential gene expression in a seizure paradigm, the highly heterogeneous cell population in the hippocampus may lead to artifactual and erroneous results. For example, suppose expression of a gene of interest is reduced in a particular cell types, i.e., pyramidal cells but is increased in a distinct cell type, i.e., granule cells. By analyzing a tissue homogenate, no net change in expression of this candidate gene may be detected. By using single cell microdissection, the actual transcriptional alterations in each cell type can be defined and a more accurate biological picture can be viewed. This approach has led to major advance in understanding the cellular pathology associated with focal cortical dysplasia (Baybis et al., 2004) and temporal lobe epilepsy (Brooks-Kayal et al., 1998).

An interesting area that has not been investigated in epilepsy research is subcellular localization of mRNAs in neurons (Eberwine et al., 2002). It is now widely accepted that mRNAs are present in neuronal dendrites and that transport of these mRNAs is tightly regulated during development and dendrite outgrowth. Local translation of these transcripts into protein occurs at great distance form the neuronal cytoplasm and serves to locally regulate the impact of synaptic input. In fact, local protein synthesis can be modulated by many of the same signals that figure prominently in long-term potentiation and enhanced excitability (Job and Eberwine, 2001). Several glutamate receptor subunit mRNAs have been identified in dendrites and clearly their local translation can be modulated by dynamic alterations in the synaptic milieu, such as calcium blockade. Thus, it seems quite likely that dendritic gene expression may be altered in human epilepsy and that this may lead to altered network properties and seizure initiation or propagation (Tongiorgi et al., 2006). One interesting though unproven hypothesis is that particular gene mutations or sequence polymorphisms might alter a sequence necessary for mRNA trafficking into dendrites and thus the actually effect of the mutation would be a disruption of mRNA targeting to neuronal dendrites.

A recently identified and fascinating group of candidate molecules that has been implicated in cancer biology are so-called micro-RNAs or miRNAs (Massirer and Pasquinelli, 2006). These noncoding transcripts are approximately 19–24 basepairs in size and have been implicated in mRNA degradation or translational inhibition. An increasing number of miRNAs have been identified in neurons and several have been implicated in a variety of brain disorders, including brain tumors. miRNAs are promoter-driven transcripts with particular chromosomal loci throughout the genome and thus are susceptible to sequence changes (polymorphisms or mutations) like coding genes. It seems plausible that these molecules may play at least some in role in epilepsy, perhaps as modulators of gene expression or protein translation.

While altered gene expression can be defined in human epilepsy and animal seizure models, in single cells or in single dendrites, there are numerous and critically important practical questions regarding the relevance of these data to human epilepsy. In particular, to what extent are reported changes in gene expression consistent across animal or patient samples, experimental conditions, seizure paradigms, and cell types and how we can reconcile any disparities to develop a cohesive hypothesis regarding transcriptional alterations and epileptogenesis. For example, while increased c-fos expression has been reported in several experimental seizure and epilepsy models, a similar consistency for many other genes in these models has not been established. How do we reconcile disparate gene expression changes in human epilepsy and an animal seizure models? If the expression of a set of genes is up-regulated in an animal model of a particular epilepsy syndrome, is it necessary to show that expression of these same genes are enhanced in human tissue specimens to make them biologically plausible? Conversely, if changes in gene expression are defined in human tissue specimens, to what extent do these changes need to be corroborated in animal systems to prove their biological credibility or therapeutic relevance? Of course, the many vagaries of human tissue research can complicate interpretation of gene expression data, yet it seems illogical that a mouse or fly system would better model human disease than studying the actual disease.

Further complicating gene expression studies are the experimental methods used to define altered expression. Thus, for example, how do we reconcile studies that define expression of a single mRNA by in situ hybridization with a study that implements oligonucleotide array analyses of several thousand genes? Indeed, are disparities in approaches helpful to understand and model the full range of seizure effects on gene expression or do they merely add unnecessary complexity? Finally, how do changes in gene expression identified using kainate, pilocarpine, or electrical stimulation seizure models help us understand the molecular pathogenesis of seizures induced by traumatic brain injury or brain tumors?

One starting place to address these issues will be to critically evaluate all reported changes in gene expression, how they were defined, and in what system they were identified. A second approach will be to look for gene expression changes that occur in multiple epilepsy syndromes as a strategy to define common transcriptional alterations for all epilepsy subtypes. Conceivably, each modeled seizure type may exhibit a unique profile of gene expression. To prove this, comparative analysis of data sets from numerous models systems will need to be analyzed or conversely, pooled analysis of multiple models may yield a common pattern of differential gene expression.


A complete review of human epilepsy genetics is beyond the scope of any single review. Epilepsy often clusters in families and in fact there is a two- to fourfold increase risk of epilepsy in patients with a first degree relative with epilepsy (Annegers et al., 1982). Interestingly, only a small number of epilepsy patients can be placed into a pedigree suggestive of Mendelian inheritance pattern (Ottman et al., 1996a, 1996b). Evidence suggests that having a primary relative with epilepsy (onset before age 35) is associated with an increased risk of developing epilepsy, and thus other genetic paradigms must be considered, i.e., modifier gene effects or epigenetic control of gene expression in the genome.

Although epilepsy is a complex and heterogeneous disorder, it is now quite clear that a number of human epilepsy syndromes, for example, generalized epilepsy with febrile seizures plus, SMEI, and benign neonatal convulsions, can result from mutations in single genes (Avanzini et al., 2007). In these disorders, a deleterious change in gene sequence (mutation) presumably leads to loss of encoded protein function (loss of function mutation). Occasionally, a mutation leads to new protein function, e.g., a so-called gain of function mutation. As a consequence of the gene mutation, seizures or epilepsy are in effect a defining phenotypic feature. The search for gene loci or causative gene mutations is underway for several generalized epilepsy syndromes including childhood absence and juvenile myoclonic epilepsy. In contrast, there are numerous neurological disorders that result from an identified gene mutation in which epilepsy is but one common feature such as Angelman's syndrome, TSC, Unverricht-Lundborg syndrome, and Miller-Dieker lissencephaly syndrome. In these disorders, there are other associated phenotypic features as well as epilepsy, and in many of these conditions, structural alterations in the brain likely contribute to epileptogenesis. In all analyses aimed at defining gene mutations associated with epilepsy, comprehensive definition of each phenotype is essential since subtle phenotypic differences may actually belie distinct gene mutations. For example, the often subtle, yet differing anterior to posterior pathological gradients of lissencephaly syndromes reflects distinct mutations in either the DCX or LIS-1 genes (Dobyns et al., 1999). These distinctions are especially important in generating targeted knockout models for these types of disorders since neither the DCX or Lis-1 knockout mice exhibit defects in cortical cytoarchitecture that fully model human lissencephaly.

Even with the detection of single gene mutations, a pivotal issue is whether epileptogenesis occurs solely as a consequence of an alteration in gene function, i.e., changes in excitability following a sodium channel or GABA receptor subunit mutation, or whether a cascade of downstream events initiated by altered function of the mutant protein is necessary for seizure initiation. Expressing a mutant channel in vitro can have deleterious effects in individual cells on spontaneous bursting or even response to proconvulsant agents but in the brain additional changes in network properties are surely a consequence. In fact, alterations in expression of numerous genes and their cognate proteins as a downstream consequence of the gene mutation may be critically important in to the epileptogenic process. In this scenario, while the altered gene sequence may be necessary to initiate the cascade, it is actually the cascade itself that is the pathological process. Investigators must be wary of viewing the effects of a gene mutation in isolation and must also consider how a gene mutation can affect related proteins and the overall network.

Alternatively, for most epilepsy phenotypes, there is little if any evidence for inheritance patterns. For example, most localization-related epilepsies typically appear as sporadic disorders in patients with no affected family members. Thus, the only plausible “genetic” mechanism is a de novo gene mutation. From an experimental perspective, these patients provide an enormous challenge because there are few insights into pathogenic mechanisms or candidate genes for most types of sporadic seizures, no family pedigree can be drawn, and a genome wide search is a daunting prospect. Perhaps even more challenging is that brain structure in many sporadic epilepsies is intact and thus morphological analysis may not be helpful. However, for some sporadic epilepsy, there are common histopathological features, such as gliosis, loss of particular cell types, changes in brain structure, or the appearance of abnormal cells, that point toward a pathogenic mechanism or pathway. For example, the pathological features of mesial temporal sclerosis such as neuronal death, gliosis, and aberrant axon sprouting supports the hypothesis that whatever the inciting event may be in these individuals, the net effects on brain structure are consistent. The presence of cytomegalic cells in balloon cell cortical dysplasia suggests a consistent pathogenic event. In fact, from a pathophysiological perspective, epilepsy per se may be the least important phenotypic feature in these disorders.

One fascinating possibility is that sporadic epilepsies result from somatic mutations occurring in brain progenitor cells, which are not present in the germline DNA (Fig. 1; Crino, 2005; Weiss, 2005). Thus, while sporadic epilepsies might not follow Mendelian inheritance patterns, they are nonetheless the result of a genetic or mutational event. In this scenario, a deleterious gene mutation might occur during the many rounds of cellular mitosis occurring throughout the course of brain development. In the case of epilepsy associated with focal brain malformations, a mutation could affect the migration or maturation of daughter cells derived from these progenitors and lead to a focal malformation of the cortex associated with seizures. However, somatic mutations could account for seizures even in the absence of altered brain structure. For example, a somatic mutation in an ion channel or neurotransmitter receptor subunit might not lead to altered cortical lamination, but nonetheless might disrupt network properties in a focal brain area. Indeed, there are many cases of so-called “nonlesional” neocortical epilepsy in which no alterations in cytoarchitecture are detected in the resected brain tissue. Future studies to sequence genomic DNA from resected brain tissue specimens may aid in identifying somatic mutations that are confined to small regions of the cortex. This hypothesis could be extended to all brain regions so that, for example, a mutation in a neurotransmitter receptor subunit, or uptake site might occur in a subset of thalamic progenitor cells at a particular developmental epoch, and result in epilepsy. The recent development of conditional knockout mouse strains which target gene knockout to a particular brain region, developmental epoch, or even single cell types will likely open new doors to understanding these highly complex, spatiotemporally regulated mutational mechanisms. A corollary hypothesis is that an accumulation of mutations in a variety of distinct genes could lead to neuronal hyperexcitability. Thus, somatic mutations in multiple genes during brain development might lead to seizure onset. Perhaps even more exciting might be the possibility that, in select brain regions, such as the dentate gyrus in which progenitor cell turnover is constitutive and even increased in response to seizures, there is an accumulation of gene sequence alterations that fosters deleterious changes in hippocampal network properties, and leads to recurrent seizures. Indeed, recurrent seizures might even promote somatic gene mutations in astrocytes or in the small pool of dividing cells in the dentate gyrus.

Figure 1.

Somatic mutations as a cause for sporadic epilepsy. Left, a focal brain area (red) identified as a seizure focus associated with a particular clinical semiology. Histological analysis of the tissue (middle, right) may reveal normal (A) or abnormal cytoarchitecture (B). At the molecular level, analysis of DNA may reveal a gene mutation (bottom arrow) that accounts for either seizure onset, or seizure onset plus altered cytoarchitecture in this region.

Modifier genes

An important consideration for understanding variable penetrance and expressivity of select epilepsy phenotypes is the presence of modifier genes that act synergistically or antagonistically with known epilepsy susceptibility loci (Durner et al., 2001; for review, see Mulley et al., 2003). These sites in the genome, while not directly responsible for seizures, can affect the impact of gene mutations or sequence alterations that culminate in seizures. Modifier genes may affect epilepsy phenotype in idiopathic generalized epilepsies but also conceivably in localization related epilepsies. In mouse epilepsy models, recent studies have demonstrated that strain backgrounds can determine seizure onset in the setting of an engineered gene defect. For example, mutations in the Scn 2a (Q54) gene lead to an epilepsy phenotype. Congenic C57BL/6J.Q54 mice exhibit decreased incidence of spontaneous seizures, delayed seizure onset, and longer survival in comparison with C57BL/6J × SJL/JF(1)Q 54 mice (Bergren et al., 2005). This observation indicates that strain type carries dominant modifier genes at one or more loci that could determine the severity of the epilepsy phenotype.

In humans, variable expressivity among family members is a common feature of inherited epilepsy syndrome of most types, suggesting that genetic modifiers may influence the clinical manifestation of epilepsy. Documented examples of such susceptibility genes include a polymorphic allele of the T-type calcium channel CACNA1H (Khosravani et al., 2004) and a rare variant of the GABA(A) receptor delta subunit gene (Dibbens et al., 2004), in which genetic variation is associated with experimental alterations in ion channel properties and possibly seizure susceptibility. These loci are not commonly involved in complex epilepsy suggesting the likelihood of considerable underlying polygenic heterogeneity.


While epilepsy may result from gene mutations, an alternative hypothesis is that epigenetic mechanisms lead to altered gene function. Epigenetic effects are defined as altered transcription of a select gene as a consequence of modifications, typically methylation or alkylation, of upstream transcriptional regulatory sequences. Epigenetics provides a new and largely unexplored area of research in epilepsy. It is widely known in cancer research that sequences within gene promoter regions known as CpG islands, which contain concatameric repeats of cytosine and guanine residues, can be methylated or alkylated (Feinberg, 2005). Addition of methyl or alkyl groups to these sequences upstream of a particular gene can transcriptionally silence the expression of that gene. Rett syndrome is associated with epilepsy and the causative gene mutation is the MeCP2 gene whose function serves to methylate a wide variety of genes (Segawa and Nomura, 2005). Excessive methylation along 15q11.13 is associated with Angelman syndrome and seizures (Lossie et al., 2001) Thus, altered expression of genes associated with epileptogenesis in other syndromes may be controlled not solely by mutations but by transcriptional silencing through DNA methylation or alkylation. Interestingly, valproic acid (VPA), a drug that has been used for decades in the treatment of epilepsy and as a mood stabilizer, triggers replication-independent active demethylation of DNA. Thus, VPA can potentially reverse DNA methylation patterns and erase stable methylation imprints on DNA in nondividing cells (Detich et al., 2003).


Genomics refers to the analysis of gene sequence alterations that may predict clinical outcome or phenotype but that do not represent deleterious mutations per se. In particular, single nucleotide polymorphisms (SNPs) represent single basepair changes that are present in more than 1% of the general population. Nonsynonomous SNPs may result in amino acid sequence changes whereas synonomous SNPs do not alter amino acid sequence. Nonsynonomous SNPs can lead to subtle or overt changes in protein structure that may have functional consequences in terms of excitability or seizure susceptibility. For example, a recent study showed that a SNP in the KCNJ10 potassium channel gene confers increased seizure susceptibility in a variety of epilepsy subtypes (Buono et al., 2004; Ferraro et al., 2004). A common SNP has been reported in the T-type Ca2+ channel gene CACNA1H associated with childhood absence seizures (Vitko et al., 2005). The compelling aspect of SNP research is the growing body of evidence documenting that SNPs relevant to epilepsy occur throughout the genome and thus that subtle alterations in basepair sequence in a set number of candidate genes can culminate in an epilepsy phenotype (Fig. 2). This so-called “common variant-common allele” phenomenon suggests that certain types of epilepsy result from an inherent predisposition or predilection based on the unique or additive effects of SNPs within a single or multiple genes (Ottman, 2001; Ottman, 2005). Thus, epilepsy may reflect a SNP dose effect in which seizure susceptibility is directly related to the number and functional significance of relevant SNPs throughout the individual genome. Genomic influences may account for the majority of patients with no obvious structural or inherited etiology for their seizures, so-called cryptogenic epilepsy. In these individuals, the presence of common allelic variants leads to epilepsy. Allelic association studies to detect genetic variants in candidate genes (as well as in a genome wide search) that occur more frequently in individuals with epilepsy than in unaffected individuals are currently underway to investigate how SNPs are associated with increased epilepsy risk.

Figure 2.

Dose effect for SNPs as susceptibility loci. In the absence of any seizure susceptibility SNPs (left), seizures are absent. With increasing numbers of SNPs either within one gene, a subset of genes, or across the genome (middle and right), there is an increasing likelihood of seizures. The effects of SNPs on seizure susceptibility may be viewed in isolation (right) or in combination with structural alterations in the brain (middle).

SNP variations may also account for some forms of symptomatic epilepsy. This notion has practical importance since it may account for the incidence of seizures in some, but not all, patients with tumors, stroke, or trauma. For example, even in the setting of brain injury, cortical malformation, or tumor, only a proportion of patients will develop epilepsy. It is conceivable that these individuals develop seizures not just because of injury to the cortex but rather because of combinatorial effects of a brain lesion plus a particular set of predispositional SNPs. In fact, a more expansive hypothesis could postulate that all epilepsies to some extent reflect a SNP dosage effect conferring high, moderate, or low seizure susceptibility either alone or in combination with structural, metabolic, or pharmacological alterations in the brain. Even in the case of monogenic epilepsy syndromes, variable phenotypic expressivity might reflect the effect of a single gene defect superimposed on a particular set of predispositional SNPs.

A relationship between gene expression changes in the setting of functional or silent SNPs has not yet been investigated. Thus, it seems plausible that changes in gene sequence that lead to alterations in network properties and culminate in seizures are likely to induce consequent changes in gene expression. As in all seizure types, transcriptional changes likely lead to changes in protein expression that have functional consequences for seizure initiation and propagation. One final point is that SNPs may also serve important roles as biomarkers for select epilepsy syndromes akin to apoE in Alzheimer's disease rather than causal events. A reliable biomarker for epilepsy would be useful as a means to predict seizure onset, prognosis, and potentially response to therapy.

Conclusions and future directions

The advances in molecular genetics of the past decade have led to major advances in our understanding of human epilepsy. A critical future direction will be to define the complex interplay between single gene mutations, alterations in gene expression, genetic backdrop and epileptogenesis in human patients. Likely, these dynamic molecular variables account for the phenotypic variability of observed for syndromes associated with a single gene defect. Defining the role that functional gene sequence polymorphisms play in seizure susceptibility is of paramount importance since these alterations may make major contributions to both sporadic and inherited epilepsies. Finally, a critical analysis of gene expression alterations across animal models and human syndromes by functional or hierarchical clustering may aid in identifying a subset of genes directly related to seizure onset. Of course, the long-term goal of molecular genetic analysis of human and experimental epilepsies is to define better tools for diagnosis, counseling, and treatment.


Acknowledgment:  This work was supported by NS045877. The author thanks Tom Ferraro, Ph.D., for comments on the manuscript.