Genetic Testing in Epilepsy: Improving Outcomes and Informing Gaps in Research

Abstract

While the diagnosis of epilepsy relies on the presence of seizures, it encompasses a group of phenotypically and etiologically diverse disorders in which seizures may only be one of a constellation of symptoms. There are genetic, structural, and metabolic causes, but most epilepsy syndromes have some genetic predisposition. The importance of genetics in the diagnosis and management of epilepsy has been increasingly recognized over the past 2 decades. With increased access to testing tools and new recommendations that all patients with unexplained epilepsy get genetic testing, it is becoming part of routine clinical care. Increased testing has resulted in an explosion in the number of genes and genetic changes identified and it is changing our understanding of the mechanisms of epileptogenesis. Advances in both clinical genetics and scientific discovery are expanding our potential to impact patient care as well as creating dilemmas. This brief review will highlight where we are regarding our ability to obtain a genetic diagnosis, how diagnoses impact patient care, and the next likely frontiers in diagnosis and management.

What Is the Diagnostic Yield of Genetic Testing in Practice?

There are genetic, structural, and metabolic causes for epilepsy, but most epilepsy syndromes have some genetic predisposition. ¹ The past 20 years have seen significant advances in technology related to genetic testing. Early in the 2000s, the major tools were Sanger sequencing, targeted fluorescence in situ hybridization and karyotyping. The development of chromosomal microarray (CMA) in 2005 led to the first widely available untargeted relatively high-resolution testing. ^2,3 Next generation sequencing (NGS) technology was also developed in 2005. ³ Next generation sequencing has since been used to develop clinically available gene panels, whole exome sequencing (WES) and whole genome sequencing (WGS). Using the currently available tools, the diagnostic yield of genetic testing for patients with epilepsy varies based on patient specific factors and the type of testing used, and there are relative patient specific advantages of each of these modalities.

Chromosomal microarray is the first of the modern whole genomic unbiased testing strategies to have been applied to epilepsy. Chromosomal microarray detects chromosomal copy number variants (CNVs) including deletions or duplications of large chromosomal segments down to single or partial genes. They do not show sequence changes and will miss chromosomal rearrangements if there are no associated deletions or duplications. In the absence of other neurodevelopmental disorders (NDDs), the diagnostic yield is only 3% but this increases up to 15% in populations with intellectual disability. ^{4 -8} Some advantages of CMA are that it is relatively inexpensive, has a relatively rapid turnaround time, and may detect CNVs not detected by other testing modalities. Chromosomal microarray can also detect large regions of homozygosity indicative of consanguinity as well as uniparental disomy.

The first large study applying WES to the diagnosis of epilepsy was in 2013. ⁹ Since that time, there have been many studies showing the utility and applicability of WES in epilepsies. Whole exome sequencing has the highest yield for patients with NDDs other than epilepsy. In those with epilepsy without other NDD, the yield is 9% ¹⁰ and up to 50% in some smaller studies of developmental epileptic encephalopathies (DEE). ^11,12 While the overall cost of WES is greater than many gene panels and CMA, its use in clinical practice results in a decrease in overall cost and time to diagnosis due to less unnecessary testing. ^12,13 Whole exome sequencing performed in trios with the patient and parents or other relevant family members has the highest yield of diagnosis with the least variants of unknown significance (VUS). There are limitations in the detection of deletions and duplications, so when WES is nondiagnostic, CMA may still be useful. Additionally, triplet repeat expansions are not typically detected and other modalities need to be considered if these are on the differential.

Whole genome sequencing is the most comprehensive of the testing modalities. It can detect sequence variation, CNVs, and there is the possibility of detecting triplet repeat expansions. ^14,15 Genome sequencing has the highest diagnostic yield at 48% for people with epilepsy as well as other NDD. ¹⁰ Similar to CMA and WES, WGS has the highest diagnostic yield in those patients with epilepsy accompanied by other NDD. ^10,16 It is still relatively difficult to get covered by insurance and interpretation of some findings may be more difficult due to less being known about the non-exome DNA being sequenced.

Next generation sequencing coupled with what we have learned about the genetics of epilepsy has led to the development of targeted epilepsy gene panels. These offer advantages in terms of cost and time. Because there are no secondary findings as are present in WES and WGS, there may be less genetic counseling involved. Gene content of multigene panels can vary widely between labs and there can be variability as to whether both sequence analysis and deletion duplication analysis is performed. In comparison to WES and WGS, there may be a relatively high number of VUS since results are not filtered based on phenotype and these are typically not done in trios. In addition, newly discovered genes may not be included on these panels because they are not continuously updated. Despite this, diagnostic yields can be as high as 54% in the presence of epilepsy with structural brain malformations, but overall is about 19%. ¹⁰ In neonatal and early infantile onset DEEs, the diagnostic yield can be up to 39%. ¹⁷ The age of seizure onset in general is highly predictive of diagnostic yield from gene panels. In one study, the diagnostic yield was 38% in those with seizure onset less the 3 years. ¹⁸ An additional study found that patients with seizure onset at less than 2 years of age had a 34% diagnostic yield compared to 4% for those with onset after the age of 2. ¹⁹ This suggests that gene panels may be most useful in the setting of early onset epilepsy, structural brain malformations, and in situations where genetic counseling related to incidental findings is not readily available.

Early onset DEE has the highest yield using WES trios and large gene panels with a range of 34% to 50%. ^12,20,21 As might be expected, the timing of genetic testing does not influence the diagnostic yield given the same patient factors. On the contrary, in one early study of adults with epilepsy and other NDD, the diagnostic yield was up to 47% using a combination of fragile X testing, CMA, and large gene panels with reflex to WES. Here, WES was the most effective and interestingly the highest diagnostic yield was in those with a perinatal event as the presumed cause. ²² This is consistent with the fact that the highest predictor of a genetic diagnosis in adults who received testing with a large gene panel is the age at onset with infancy being the highest, and a steep drop off between early childhood and adolescence. Similarly, the presence of intellectual disability or developmental delay was predictive of a diagnosis. ^{23 -26} Given that those with infantile onset DEE are the most likely to receive a diagnosis, it is not surprising that the most diagnosed genes in adults are like those in children who undergo genetic testing with SCN1A being the most common. ²³ Older age, seizure freedom, and institutionalization seem to be associated with the highest diagnostic gap due to lack of testing. ²⁷

What Are the Implications for Patient Care?

Despite the presumed genetic influence on the development of epilepsy, adoption of genetic testing as part of routine practice for diagnosis has been slow compared to evaluations such as imaging. This may be in part due to access and to lack of data on how it affects outcomes. For all testing modalities the question is always, how will it affect overall management? In recent years there is more evidence addressing both intangible and concrete results of genetic testing. In addition to possible changes in treatment, undergoing testing may provide patients and families with prognosis, an ability to anticipate associated comorbidities and recurrence risk. ^10,11,14 This may facilitate testing of relatives, reproductive planning, access to specific treatment services, and patient advocacy and support groups, which is often felt by families to be one of the most important benefits; and finally, an end to diagnostic odysseys. ^{10,13,14,28,29}

The impact of a genetic diagnosis on clinical management is significant. In a single cross-sectional study of consecutive patients receiving a diagnosis based on large gene panel testing, the diagnosis resulted in treatment change in 50% of patients with an overall improvement in outcomes in 75% of those in which a change was made. ²⁴ This is consistent with mixed adult and pediatric population studies in which 12% to 80% of diagnosis resulted in management changes. ^10,28 This includes treatment changes such as anti-seizure medication choice, disease specific metabolic or vitamin treatments, pathway driven medications, and gene specific trial discussions. ²⁸ Genetic testing limited to adult patients with DEE or epilepsy + other NDD indicates that 12% to 57% of diagnosis resulted in change in management. ^22,23,30 The impact on management was greatest in those with treatment refractory epilepsy. ³⁰ Importantly, some impact was documented in 100% of neurotypical individuals and 55% of those who had seizure onset after 2 years. ²⁸

Structured interviews with families of a child or adult with DEE have demonstrated that the diagnosis increased understanding and provided hope as well as a relief of guilt or self-blame associated with their loved one’s condition. They also reported increased social connection with other similarly affected families and hope. However, diagnosis often took years, and the delay was accompanied by uncertainty, anxiety, trauma, and stress with two thirds of families reporting that not having a diagnosis had influenced their decisions about having other children. ³¹ Outside of management changes, receiving a genetic diagnosis has additional personal utility and suggests benefit outside of treatment for the individual. Having a diagnosis in itself is empowering to patients and their families. ³²

What Are the New Frontiers in Genetic Testing and Application to Patient Care?

The impact of receiving a genetic diagnosis on patient management is only going to become greater as medicine and science advance. In both animal and human models, there are 2 broad categories of gene therapy research. The first is gene-specific targeted gene modulation or supplementation. There have been numerous animal studies with some DNA or RNA modulation treatments making it to human clinical trials. ^{33 -35} However, these targeted approaches require a single gene diagnosis and therefore have the potential to be very helpful for a limited and specific population. While genes such as SCN1A may represent up to 13% of diagnosed single gene causes of epilepsy, most other individual genetic changes make up less than 2% each of the total diagnoses in a population. ²³

A second line of preclinical and early clinical investigation is the use of genetic tools to modulate hyperexcitability independent of the specific change or for broad gene classes. Examples of this type of therapy include expression of molecules that lead to decreased neuronal and network activity within epileptic networks. ^33,36,37 While these types of therapies are more symptomatic treatment, advantages of these techniques include that a specific genetic cause does not have to be known so may be available to broader populations. As for specific gene therapies, cell type, region-specific, and even state specific targeting is likely going to be important in the effectiveness of these less targeted studies. ^34,37

The importance of nondiagnosis specific therapies is emphasized by the fact that despite optimal current testing modalities and patient selection, the highest yields of testing are still only 50%. ¹⁰ This may be because the patient does not have a genetic etiology for their epilepsy or because we still do not fully understand the genetic contributions to epilepsy.

Many of the studies already discussed have focused on severe, early onset, and rare epilepsies associated with other NDDs with highly penetrant genetic variants with large effect. However, genome wide association studies (GWAS) looking at more common focal and genetic generalized epilepsies have identified common variants each with small effect that are not individually large enough to stratify risk. When combined, these common variants have been used to generate polygenic risk scores (PRS) that can determine risk for generalized as well as focal epilepsies. ^38,39 Patients with familial epilepsies have been found to have higher epilepsy PRS compared to unaffected family members, sporadic epilepsy patients, and population controls regardless of the presence of a known rare variant, suggesting a roll for polygenic risk in the manifestations and penetrance of familial epilepsies. ⁴⁰ Additionally, studies of polygenic risk have been applied to predict risk of epilepsy in children with febrile seizures. ⁴¹ It is likely that continued identification of common variants will tell us more about the underlying causes of epilepsy, inheritance patterns, and inform risk as well as guide treatment in the future.

The high diagnostic yield of genetic testing in patients with malformations of cortical development indicates that there is a role for genetic testing in lesional epilepsies. Regardless of the presence of a lesion, the diagnostic yield of genetic testing for patients with focal epilepsies is 12.5% to 29%. ⁴² Genetic analysis is increasingly being included in the presurgical evaluation of patients with medication refractory focal epilepsy. ^{43 -45} Studies of the relationship of genetic diagnoses with surgical outcomes have variable results. Some data suggests that mTOR-opathies and genetic hypothalamic hamartomas have a relatively favorable outcome while synaptopathies and channelopathies have a relatively worse outcome. ⁴³ In contrast, in a cross-sectional study of WES on postoperative patients with mesial temporal lobe epilepsy, ultra-rare missense and loss of function variants were present at higher rates than in controls, but there was no difference in the proportion of subjects with variants in the group with favorable versus unfavorable outcomes. ⁴⁶ It is likely that larger studies looking at gene function in relation to outcome are needed.

In addition to polygenic risk, somatic variation is increasingly being recognized as a contributing factor to both lesional and nonlesional epilepsy, in particular due to variants in the MTOR pathway. ⁴⁵ While a criticism of the significance of somatic variant diagnosis has been that the diagnosis is too late due to testing being performed after surgery, there have been significant advances in the ability to analyze DNA from cerebral spinal fluid and from small amounts of tissue collected from depth electrodes, which may provide an opportunity to include somatic testing in the presurgical evaluation. ^47,48 Detection of somatic variation offers the possibility of determining etiology, but also of more precise mapping of epileptic networks in the future. As has been discussed, there are other benefits of a diagnosis, especially to the families.

Conclusions

The data supports that those with early onset epilepsy, epilepsy associated with other NDDs and epilepsy with congenital anomalies (particularly brain malformations) are most likely to receive a diagnosis from a genetic workup no matter when that workup is performed. However, there is the potential to impact the management of all patients with epilepsy. The National Society of Genetic Counselors has released an evidence-based practice guideline recommending that all patients with epilepsy receive a genetic workup regardless of age. ²⁹ The recommendations are for exome or genome sequencing with reflex to CMA when negative. Not all patients will receive a diagnosis despite the current standards of care and there are significant challenges related to access to testing, genetic counseling availability, provider training to understand and explain testing to their patients and interpreting results for patients with unclear phenotypes. Research into polygenic contributions to epilepsy, somatic variation, epigenetic factors, and WGS data are likely to bring the field closer to closing the diagnostic gap and more work will need to be done to close the access gap. In addition to more precise and hopefully effective management this will likely lead to improved understanding of the pathophysiology of epilepsy.