Genetics of Epilepsy in Clinical Practice

Tremendous strides have been made in unraveling the genetics of epilepsy in the last 2 decades. There have been many surprises such as discovery of the importance of de novo mutations, the complexities of phenotype–genotype relationships—one syndrome having multiple genetic causes (genetic heterogeneity) and one gene being associated with different phenotypes (pleiotropy)—the very large number of genes that appear to raise risk for epilepsy, and the emerging significance of somatic (postzygotic) mutations (1–3). Moreover, the dramatically reduced cost of genetic testing, the availability of gene panels, and even whole exome or whole genome sequencing at the clinical interface mean that genetics is rapidly becoming part of routine epilepsy practice, especially, but not exclusively, in pediatric settings.

This reality can appear daunting to the clinician without specific training in genetics. One is confronted with many choices in genetic testing; the results may appear as a veritable “alphabet soup” of gene names, and the interpretation of test results may not be straightforward. While we should turn to our clinical genetic colleagues for expert assistance, it behooves all practicing epileptologists to improve their “genetic literacy” to cope with this rapidly changing landscape, which impacts everyday clinical practice. Just as every epileptologist must have expertise in reading MRI scans, although not with the sophistication of a neuroradiologist, so should we aim to have basic competency in interpreting the implications and pitfalls of genetic data. Patients and families will certainly expect this! This review outlines clinical points in epilepsy genetics for the practicing neurologist, and the companion paper deals with genomic technologies, options for testing, and results interpretation (4).

How Do I Estimate Risk to other Family Members?

It is axiomatic that precision in providing genetic risk information is dependent on precision in diagnosis. This requires both an accurate epilepsy diagnosis and a detailed family history. Tips on how to obtain optimal family history information are summarized elsewhere (1).

Epilepsies with Mendelian inheritance (e.g., autosomal dominant, autosomal recessive, X-linked) are rare, but the risk to relatives is high. Examples of dominant disorders are benign familial neonatal epilepsy (BFNE) and autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE), where 50% of first-degree relatives will have the mutated gene, and of those, most will have clinical seizures (~85% of mutation carriers for BFNE [5] and ~70% for ADNFLE [6]). Recessive epilepsies are usually quite severe disorders (e.g., Lafora disease, Unverricht-Lundborg disease), and the risk to siblings is 25%; consideration of recessive disorders should be at the forefront when the parents may be consanguineous. X-linked disorders may not be obvious unless extensive pedigrees are constructed—X-linked recessive (boys only such as ARX-related epilepsies [7]); X-linked dominant (girls affected, boys severely affected or embryonic lethal such as double cortex syndrome and periventricular nodular heterotopia) (8); and the remarkable “girls-only” disorder (where carrier males are healthy) of PCDH-19–related epilepsy (9, 10). Each has quite different counseling implications.

In most epilepsies, however, inheritance is non-Mendelian, and risks to relatives are considerably lower. The precise genetic architecture of these disorders is still uncertain—that is, the number of genes involved (few or many) and their effect size—and this is a current research challenge (1, 11, 12). In many cases, there is no known family history in first- or even second-degree relatives, and the patient may appear surprised that the cause is genetic. The clinician must explain first that there may have been mild epilepsies in relatives that they were unaware of and further inquiry may be revealing, especially among the older relatives, and second, in complex inheritance, there may not necessarily be immediate relatives with the disorder, especially if the family size is small.

For the common epilepsies, empiric risks are well established. In genetic generalized epilepsies (GGE), the risk to first-degree relatives is increased approximately 8-fold compared with the general population, meaning that first-degree relatives have about an 8% chance of being affected by age 20; the risk to second-degree relatives is much less. In focal epilepsies, the risk to first-degree relatives is approximately 2.5-fold above the general population (11, 13).

There are uncommon exceptions to these general rules, where a penetrant mutant gene segregates in a Mendelian fashion causing common epilepsy phenotypes. Examples include mutations in SLC2A1, encoding the glucose transporter GLUT1, which are found in about 1% of GGE subjects (14). For focal epilepsies, mutations in genes LGI1 and DEPDC5 are currently the most commonly identified; these cause familial lateral temporal lobe epilepsy and familial focal epilepsy with variable foci (FFEVF), respectively. Clues to these disorders include obvious family history, the presence of auditory auras in LGI1-related families (15), and heterogeneous focal epilepsy phenotypes in DEPDC5-related epilepsies (16). The spectrum of epilepsies associated with DEPDC5 mutation is rapidly expanding—small pedigrees may have exclusively nocturnal frontal lobe epilepsy or temporal lobe epilepsy thus not obviously suggesting FFEVF (16, 17). More remarkably, although MRIs are usually regarded as normal, some cases have subtle focal cortical dysplasia and occasionally more extensive focal malformations (18–20).

Why Should I Bother to Find a Genetic Cause?

Nihilistic views are sometimes expressed that finding a genetic cause does not alter management, and therefore genetic testing is unnecessary (21). If such views were ever defensible, they are certainly not now, for a number of reasons.

First, even if a condition is untreatable, patients and families nearly always want to know why the affliction has occurred. “Closure” of diagnosis is very important; it brings an end to the diagnostic odyssey, and families can stop searching for “answers” from multiple physicians or alternative medicine practitioners. This situation is particularly true when a previously healthy child develops an epileptic encephalopathy with refractory seizures and developmental regression. The parents may carry guilt, with a false belief that minor head knocks or vaccination (22) caused the disorder—identifying the true cause can provide considerable consolation and assuage unfounded blame. Even in milder familial epilepsies, subjects generally wish to know (23).

Second, a specific genetic diagnosis avoids unnecessary testing with repeated blood tests, MRIs, invasive biopsies, pre-surgical workup, and even intracranial electrodes in the vain hope that a focal lesion or some other cause might be found in an unsolved case of epileptic encephalopathy. In turn, this allows families to focus on the problem and not be distracted by searching for the cause. A genetic diagnosis may provide useful prognostic information regarding the natural history of the disorder as large case series accumulate—this has already happened with Dravet syndrome and PCDH19-related epilepsy and will occur in time for rarer epilepsies. Families often derive benefit by meeting similarly affected families, either in person or online. Family-driven groups focused on a specific disorder have become a powerful force for aggregation of accessible information, for effective public advocacy, and for encouraging and driving research. The empowerment derived by specific diagnosis, understanding the disorder and contributing to solve it cannot be underestimated (24–26).

Finally, a genetic diagnosis allows for specific counseling, as knowledge of the mutation then allows other interested family members to be tested, and much more precise information can be given, rather than broad empiric risk estimates discussed above. Where possible, this should be done in a genetic counseling environment with trained counselors and geneticists (27, 28). If there is a diagnosed familial epilepsy with a single major mutated gene, individual risk of family members can be easily determined by cascade testing, and risk to further children will also be apparent. De novo mutations provide a challenge; the risk to a further child is low, but not zero because of the possibility of parental mosaicism, which cannot easily be investigated. Counseling is also challenging when risk alleles (as opposed to genes determining Mendelian epilepsies) are detected (29); currently this is rarely a practical problem as few risk alleles are known, but this is likely to change in the near future.

The ethical, legal, and social concerns of genetic testing warrant systematic research (28,30). Concerns about the negative impact of a genetic diagnosis seem largely unfounded but need to be considered as testing becomes more widespread. Initial research supports the strong clinical impression that the impact of genetic information is largely positive; people want to know and find the data helpful (23, 31, 32).

Specific “Actionable” Genetic Findings

In addition to the general benefits of testing outlined above, there is a small but growing number of genetic diagnoses in which specific alterations in management are indicated. This may include the choice of conventional antiepileptic agents or the use of an alternative treatment.

Dravet syndrome, which is the result of mutations in SCN1A in 80% of cases, is generally easy to diagnose after a few years of life when the characteristic evolution of seizure types and developmental stagnation has occurred. There is some evidence that early aggressive therapy improves outcome, so testing should be considered early (32–34). Specific guidelines for testing are available—Dravet syndrome should be considered in infants with febrile seizures presenting around 6 months of age, especially those with prolonged and recurrent febrile seizures, hemiclonic seizures, and seizures induced by bathing (35). Recognition in adult life may lead to better seizure control and cognitive outcome (36). Sodium channel blockers such as lamotrigine and carbamazepine should be avoided, whereas valproic acid, topiramate, clobazam, and stiripentol appear to be beneficial (33, 34). Unfortunately, at present, these guidelines are based on case series evidence, with the exception of the use of stiripentol where controlled trials have shown benefit (37).

The epileptic encephalopathies associated with mutations in SCN2A and SCN8A are gradually becoming better defined (38, 39) and appear to be rarer than Dravet syndrome. Their profile of drug responsiveness may be different as the proteins encoded are localized to excitatory neurons, whereas the SC-N1A protein appears to be mainly on inhibitory interneurons. Indeed, for SCN8A encephalopathy, sodium channel blockers may be effective in some cases (39).

Loss of function mutations in the potassium channel gene KCNQ2 have been robustly associated with both BFNE and a severe epileptic encephalopathy resulting from de novo mutations. The novel drug ezogabine (retigabine) specifically targets these channels, but unfortunately toxicity to the retina, skin, and nails may limit its use, and there are little data on its effectiveness in KCNQ2 encephalopathy. However, recent experience suggests that sodium channel blockers, carbam-azepine and phenytoin, are effective in this disorder (40).

Mutations in SLC2A1, encoding the glucose transporter GLUT1, have long been known to cause a severe infantile encephalopathy with low CSF glucose that responds partially to the ketogenic diet. The spectrum of epilepsies associated with GLUT1 deficiency now includes a small proportion of GGE, especially early-onset absence epilepsy and, less commonly, other absence epilepsies as well as about 5% of patients with epilepsy with myoclonic-atonic seizures (41, 42). Some patients may have paroxysmal exercise-induced dyskinesia, increased seizures before meals, mild intellectual impairment, or cerebellar signs, which are clinical clues to the diagnosis. Many patients with mild GLUT1 deficiency respond to conventional antiepileptic drugs, but awareness of the diagnosis emphasizes the option of ketogenic diet therapy for refractory cases.

Pyridoxine-dependent epilepsy is a well-known, albeit rare, infantile epilepsy, generally diagnosed by clinical suspicion and response to pyridoxine. Mutation detection in ALDH7A1 can provide confirmation and be used to rapidly screen at-risk siblings (43). The related disorder, pyridoxamine 5′-phosphate oxidase deficiency, owing to mutations in PNPO, may only respond to pyridoxal 5′-phosphate, but the relationship of mutations to treatment response remains unclear (44).

There is now a great deal of interest in novel precision therapies based on molecular findings. Quinidine reverses the in vitro gain of function seen with KCNT1 mutations in epilepsy with migrating focal seizures of infancy, and a clinical response to quinidine has been observed (45, 46). Similarly, in a patient with a mutation in GRIN2A, encoding a glutamate receptor subunit, memantine therapy appeared effective (47). The recent discovery of the importance of mutations in DEPDC5 (16–20), which is now known to be a regulator of mTOR, raises the possibility of treatment with rapamycin analogues. All of these observations will need careful double-blind trials to establish efficacy.

Conclusion

Genetics is transforming clinical practice in epilepsy, especially in children. Useful clinical information can be provided to families, and a specific molecular diagnosis, where possible, adds greatly to the accuracy of the information. The promise of precision therapies is becoming a reality and may eventually become widely applicable.