|Home | About | Journals | Submit | Contact Us | Français|
This review is a summary of a talk presented at the 2015 American Epilepsy Society Annual Meeting. Its purposes are 1) to review developments in epilepsy genetics, 2) to discuss which groups of patients with epilepsy might benefit from genetic testing, and 3) to present a rational approach to genetic testing in epilepsy in the rapidly evolving era of genomic medicine.
In the past decade, we have experienced a genomic revolution in medicine. Now faced with single nucleotide variants and copy number abnormalities that might explain our patients' epilepsies, we must think carefully about which populations of patients with epilepsy might most benefit from genetic testing, what type of testing makes sense for them, and how best to educate patients and families about the possible outcomes and limitations of genetic testing. Finally, as genetic testing becomes more available, to most effectively manage the growing complexity of epilepsy genetics, we need to consider emphasizing genetic education for neurologists during training and partnering with geneticists and genetic counselors with expertise in epilepsy.
Genetics has had a long-recognized role in epilepsy, particularly familial epilepsy and epilepsy of unknown cause that was formerly called “idiopathic epilepsy” (1). Since the sequencing of the human genome in 2001 (2, 3), we have experienced a paradigm shift in many fields of medicine, including epilepsy. The early discoveries of genes responsible for familial epilepsies—for example, SCN1A, CHRNA4, and LGI1—provided proof of principle that substantiated many decades of hypotheses that patients with familial epilepsy have a genetic etiology (4–6). The notion that some sporadic cases of epilepsy had a genetic basis was initially theoretical, until associations were established, for example, between SCN1A and severe myoclonic epilepsy of infancy (Dravet syndrome) (7). Over the past 15 years, the list of genes associated with epilepsy has expanded dramatically as modern sequencing technologies and analysis capacities have accelerated both research and clinical genetic analysis (8, 9). Though genome-wide association studies did not change the landscape of epilepsy genetics dramatically (10), a large number of studies implicating new dominant and recessive epilepsies—strengthened by genetic evidence and/or functional data—rapidly unfolded in the past decade (9). A key development is that de novo pathogenic variants in epilepsy-associated genes are now recognized to have a major role in the etiology of epileptic encephalopathies, in the form of single nucleotide variants (11–13) and copy number variants (14–18).
A brief word on terminology: The term mutation refers to an event resulting in a change in the expected DNA sequence. In the medical literature, “mutation” has conventionally referred to such changes for which there was strong evidence for pathogenicity. More recently, the term “variant” has been suggested as a preferred term by the American College of Medical Genetics, with various levels of evidence for pathogenicity that can be attributed to each variant; these are based on the strengths of association between a given gene and a given phenotype (gene-level evidence) and based on amino acid conservation across species and gene paralogs, family segregation data, and presence or absence in clinical and research databases or population databases (variant-level evidence; e.g., Exome Aggregation Consortium, www.exac.broadinstitute.org) (19).
The primary types of genetic variants associated with epilepsy, as with other diseases, are single nucleotide variants (1 base pair), small insertions and deletions that may or may not result in a shift in the reading frame of the gene, and structural variation in the form of microdeletions/microduplications or chromosomal monosomy or trisomy. Chromosomal rearrangements are another category of genetic variation that may result in small or large losses or gains in copy number at their breakpoints.
Mutational events give rise to DNA variation on a regular basis, sometimes leading to cancer and developmental disease and other times occurring without known direct consequence (20–24). Ancestral variants, arising several generations ago, may be tolerated in the heterozygous state such that the individual's cells all have one reference allele, or “normal” copy of the gene, and one alternate or variant allele, or “abnormal” copy of the gene. These variants may be passed down from generation to generation without causing symptoms or any discernable phenotype (e.g., overt seizures or EEG abnormality). Two copies of the same variant allele may be present in the same individual when the population allele frequency is sufficiently high that chance mating occurs between two carriers (e.g., the delta F508 allele associated with cystic fibrosis) or in the case of parental consanguinity, which occurs in many parts of the world. Two different variant alleles in the same gene may occur and result in a compound heterozygous condition. When an individual is found to have two alleles in “trans” configuration, having inherited one from each parent (or one inherited and the other de novo but affecting both alleles of the gene), compound heterozygous recessive disease results if both variant alleles affect normal protein function.
De novo variants are those that are present and detectable in an individual and apparently absent in his or her parents when DNA is assayed, traditionally from leukocytes. Both the inherited heterozygous variant allele in a gene associated with a recessive condition and the variant that is pathogenic in the heterozygous state (dominantly inherited or de novo) occurred at some point as a truly de novo, or new, mutation, likely during meiosis during the formation of a gamete (oocyte or spermatocyte) (25–27). In such a situation, the variant is present from the point when the individual was a zygote and onward and is thus detectable in blood.
Mutational events also occur after the stage of the zygote, during mitosis (21, 25, 28). These events have classically been associated with the development of cancer; when they occur early in the course of embryonic development, and do not result in the death of the involved daughter cells, they may present as a developmental lesion (20–24, 29). A role for postzygotic mutation, also referred to as somatic mutation, has been postulated for epilepsy (30). While not yet demonstrated for nonlesional focal epilepsy, this mechanism has been demonstrated to explain focal cortical dysplasia, hemimegalencephaly, and neurocutaneous syndromes with megalencephaly and prominent brain malformations (31–41). The extent to which focal malformations and, by extension, focal epilepsy can be explained by variants that arose in the postzygotic embryo remains to be determined.
Genetic discoveries in the research arena were soon followed by the availability of genetic testing in the clinical arena, initially with single gene sequencing and chromosomal microarray analysis (CMA) and later with gene panels for sequencing of a few genes at a time. Now, larger panels of genes can be assayed, with both sequencing that will identify single nucleotide variants and indels as well as deletion/duplication testing that can identify copy number variants at a resolution of ~20 kB. Exome sequencing is clinically available yet costly. In general, for epilepsy diagnoses, the more it costs the less likely it is to be covered by insurance, which presents a hurdle for many families and their physicians who are seeking a genetic diagnosis (42).
Who might benefit from genetic testing? And why should we embark on the process of seeking approval and coverage for genetic testing in the realm of epilepsy, when we are also managing medications, side effects, and developmental and behavioral concerns in ever more complex healthcare systems? Genetic testing can be time-consuming, particularly when done with proper pretest counseling about the range of effects that may ensue and their meaning (see below). Follow-up of results, parental testing in some cases, and return of positive results often demand additional visits and can be logistically challenging. Why then pursue genetic testing in the epilepsy clinic? The two main reasons for seeking a genetic diagnosis in epilepsy are 1) diagnostic certainty, which may aid in prognosis, and 2) potential impact on treatment (see examples in Table 1). So, it would make sense to focus testing at this time on those in whom there is a high likelihood of a diagnostic finding and on those whose refractory epilepsy may be influenced by a precise genetic diagnosis that can guide treatment. In most cases, we cannot answer the question of whether treatment will be affected until after we perform genetic testing and have the results. For patients with refractory epilepsy, particularly infants and young children, if there is any chance of that occurring, parents and physicians alike are naturally compelled to undertake all possible testing that could in any way ameliorate the clinical situation. This is the case regardless of the likelihood of the availability of a rational therapy based on the diagnosis. In this group, we seek an end to a “diagnostic odyssey” for families and a removal of a sense of blame that many of them report prior to a conclusive molecular diagnosis (42).
The highest yield group of patients for genetic testing to date can be summarized as “epilepsy plus,” consisting of epilepsy with accompanying dysmorphic features (though a specific genetic syndrome may not be evident early on), intellectual disability, autism, and cognitive regression (12, 13, 43–45). In these groups, the yield of CMA is about 5% (14) and sequencing via panel testing or exome sequencing, 20 to 50 percent (44, 46–48). This group is also highly likely to experience refractory epilepsy. There are some conditions, the inherited metabolic epilepsies, that are treatable and simply should not be missed; these include pyridoxine-dependent epilepsy caused by recessive variants in ALDH7A1 (49). There is a growing list of genes that suggest a modification in treatment (e.g., SCN1A—avoid phenytoin and lamotrigine, in general though not always; SCN2A and SCN8A—high-dose sodium channel-affecting agents such as phenytoin may be effective) (42). There are others for which observations across many centers may have a role in suggesting treatment (e.g., PCDH19 ) but for which prospective trials are still needed to know if there is a clear specific response to particular treatments. For some genes, there is a hope of “precision” medicine, or treatment based on the biology of the genetic dysfunction. This includes KCNT1 (51), GRIN2A (52), and DEPDC5 and other mTORopathies, including focal cortical dysplasia (53). Two major caveats should be mentioned here: 1) the presence of any variant in an epilepsy-associated gene does not automatically mean that the variant is pathogenic, nor does it inform us as to the type of functional change that may be present (e.g., gain vs loss of function), and 2) while case reports are helpful in establishing proof of principle and pointing to potential treatments, we need as a community to come together to conduct clinical trials with clear, uniform outcomes to study the effects of rational, target medications for each rare genetic epilepsy (54). Both careful vetting of variants and consultation with colleagues with expertise in specific genes are often needed to aid in the most accurate determination of whether or not to pursue gene-based treatment.
There is not a single algorithm for genetic testing in epilepsy, and there are pros and cons to each strategy. The first step is to consider the clinical diagnosis. Is the patient dysmorphic? If so, a CMA to assess for 20 kB or larger deletions and duplications is warranted, as such findings are likely to be missed with exome sequencing and possible gene-panel testing, depending on the panel used. Is there an epilepsy syndrome associated with a gene or list of genes, such as Dravet syndrome with SCN1A or Rett syndrome with MECP2? If so, complete sequencing with full coverage of every exon (not always a guarantee with exome sequencing) and deletion/duplication coverage of one or more genes should be a priority. If the CMA and initial panel suggested by a specific phenotype (e.g., infantile epilepsy panel, progressive myoclonic epilepsy panel, or Angelman-like panel) are not revealing of a genetic diagnosis, but a genetic cause is highly suspected based on the phenotype, exome sequencing is a logical next step in the evaluation. In some cases, where several phenotypes are present and waiting for results of successive panels would prove inefficient, exome sequencing (+/− CMA) may be pursued early on at some centers. The incremental yield of exome sequencing after a negative evaluation with CMA and panels is not well documented yet, but there are numerous case examples where exome sequencing has provided a diagnosis that was not made with CMA or panel testing. These include examples in which the relevant gene was not included on a panel because it had not yet been associated with epilepsy when the test was run and examples when mosaicism is present and exome sequencing analysis allows identification of variants where other techniques may not.
With all the new and evolving information about genes important for epilepsy, variant interpretation, functional characterization of epilepsy genes, and emerging clinical trials for gene-specified populations of patients, it is important for clinicians to inform their patients of what they are getting themselves into. Genetic test results may be positive—with unequivocal findings in genes strongly associated with epilepsy of variants that have been previously reported as associated with disease and/or functionally characterized. They may be negative, with no explanatory results discovered, and it should be noted that this does not rule out that the cause is genetic—it may be, and we in 2017 (or the current year) are not yet able to elucidate it with current knowledge and technologies. They may be equivocal with variants of uncertain significance in disease-relevant genes; this situation may require parental testing, particularly for severe cases in which the presence of a de novo variant would be most compelling, and may require an ongoing discussion that incorporates incrementally accrued knowledge over time. Families should be told what the next steps might be in each scenario—possible treatment considerations and referral to gene-based support networks if positive, or research enrollment if equivocal or negative (e.g., the Epilepsy Genetics Initiative supported by the National Institutes for Neurological Disease and Stroke and Citizens United for Research in Epilepsy).
A 14-year-old young woman with juvenile myoclonic epilepsy (JME) and a family history of generalized epilepsy (mother) and febrile seizures (maternal aunt) comes for consultation to consider a possible genetic diagnosis.
If she had a mutation in a gene always associated with a benign course, we could provide reassurance that she has a nonprogressive epilepsy. Suppose, for example, we were faced with school decline in a patient like this, which can be multifactorial and needs to be addressed at many levels. In that case, confirmation of a nonprogressive condition may not be required but may be helpful. In a classic case of JME in an otherwise well patient, with normal EEG background, genetic testing may not add to the clinical picture in 2017. There may well be a role, however, for pharmacogenetics that will help with treatment choice and side effect minimization in the future. What can we tell her about the next generation—will she pass on a genetic trait such that some of her children will have epilepsy? As the genetics of the more common epilepsies is worked out, in part through large collaborative sequencing efforts (55), it may be possible in the future to help with risk stratification for siblings and for successive generations. Currently, data suggest a risk to family members of about 5% (56), but without specific genes to evaluate, this remains a generic prediction.
A 6-month-old boy with new-onset infantile spasms with hypsarrhythmia and normal MRI is referred for genetic testing as part of his evaluation.
If he was found to have a pathogenic variant in an “epilepsy gene,” repeated testing for possible metabolic disease (with lumbar puncture, etc.) and repeated imaging, as often occurs in patients with infantile spasms who then go on to have focal seizures, could be stopped.
Would there be an impact on treatment? As discussed above, there is a growing list of genes associated with modifications in treatment as well as rational approaches to treatment that will hopefully coalesce into multi-center clinical trials in the coming years.
Genetics plays a major role in epilepsy, particularly in patients with refractory epilepsy. The well-informed neurologist can triage who most needs and could benefit from genetic testing, choose the appropriate testing, and explain the findings in the context of the ever-important clinical scenario. What is needed is a thoughtful genetic approach incorporated into the overall clinical evaluation of a patient with epilepsy. The ability to access and incorporate new information about types of testing, specific genes, and possible treatments can augment the success of this approach. Depending on the practice setting, a neurologist can address all of these issues and stay current by partnering with a colleague in genetics or by seeking epilepsy genetics expertise from physicians and genetic counselors in a specialized epilepsy genetics program.
Editor's Note: Authors have a Conflict of Interest disclosure which is posted under the Supplemental Materials link.