Abstract
New technologies for mutation detection in the human genome have greatly increased our understanding of epilepsy genetics. Application of genomic technologies in the clinical setting allows for more efficient genetic diagnosis in some patients; therefore, it is important to understand the types of tests available and the types of mutations that can be detected. Making a genetic diagnosis improves overall patient care by enhancing prognosis and recurrence risk counseling and informing treatment decisions.
Major technological advances in the past decade have rapidly increased our ability to identify and characterize genetic changes across the entire human genome. The introduction of chromosome microarray platforms—including array comparative genomic hybridization (CGH) and single nucleotide polymorphism (SNP) microarrays—facilitated whole-genome studies of copy number variation and the discovery of novel microdeletion and micro-duplication syndromes. The development of next-generation sequencing technologies has revolutionized the field of medical genetics and gene discovery across many different disciplines.
Our understanding of epilepsy genetics has benefited greatly from the application of genomic technologies to identify genetic risk factors and causes of all types of epilepsy, including genetic generalized epilepsy, focal epilepsy, and epileptic encephalopathy. At an ever-increasing pace, clinical laboratories are employing the same technologies, translating research findings to diagnostic tests. Keeping up with the barrage of novel genetic findings is a significant challenge for investigators and clinicians alike, as is determining which clinical test is most appropriate to order for any given patient. This review discusses the genomic technologies most commonly employed today, the variety of genetic testing options available for patients with epilepsy, the advantages and disadvantages of each, and the interpretation of results.
Chromosome Microarrays
Copy number variants (CNVs) are deleted or duplicated stretches of chromosomal DNA and can be found throughout the genome. Large rare CNVs have long been known to cause disease; examples include the ~5 Mb deletion of 15q11-q13 that causes Angelman syndrome, the 1.5 Mb duplication that causes Charcot-Marie-Tooth type I, and the 1.5–3 Mb deletion that causes 22q11 deletion syndrome.
The ability to detect smaller CNVs in a genome-wide, largely unbiased fashion was greatly enhanced by the development of array CGH and SNP microarrays, collectively referred to as chromosome microarrays (CMA).
The first breakthrough CNV discovery in epilepsy that came from large-scale screening of affected individuals was recurrent deletion of 15q13.3 in patients with genetic generalized epilepsy (GGE) (1). Approximately 1% of individuals with GGE harbor the 1.5-Mb deletion that includes six genes, of which CHRNA7 is the most logical candidate gene for the epilepsy phenotype. Interestingly, the 15q13.3 deletion is also found in patients with intellectual disability, autism spectrum disorder, and schizophrenia, although the highest proportion of affected individuals is found in epilepsy cohorts (2–6). The odds ratio for developing epilepsy if an individual carries the deletion is very high (68, 95% CI: 29–181), but the deletion is neither necessary nor sufficient for developing GGE (2). Similarly, deletions of 15q11.2 and 16p13.11 are each found in ~1% of patients with GGE but also occasionally in patients with other forms of epilepsy (7). In addition, each is also a risk factor for other neurodevelopmental disorders. Notably, individuals with GGE and intellectual disability beyond what would be expected for their epilepsy are more likely to carry a deletion of 15q11, 15q13, or 16p13 than are individuals with GGE alone (8).
While the three deletions described above are collectively the most common, especially in GGE, other CNVs are important as well. Whole-genome screening for CNVs in various cohorts have consistently shown that ~5% of patients have a potentially pathogenic deletion or duplications. This is true for focal epilepsy (9), generalized epilepsies (10), epileptic encephalopathies (10), fever-associated epilepsy syndromes (11), and in patients with broadly defined developmental disorders and epilepsy (12). CNV testing is widely available for clinical diagnostics and should be considered early in the diagnostic process, especially for patients with severe, early-onset epilepsy or epilepsy and birth defects. Finding a chromosome abnormality that explains the clinical findings can help with prognosis and recurrence risk counseling and eliminates the need for further diagnostic testing.
Next-Generation Sequencing
Undoubtedly, the most significant recent advance in the human genetics in the past decade is the development of next-generation sequencing (NGS). This technology—also known as “massively parallel sequencing”—has largely replaced traditional Sanger sequencing in the research laboratory and in clinical diagnostics. In contrast to Sanger sequencing, in which sequencing occurs one fragment at a time, NGS allows simultaneous sequencing of millions of short fragments of DNA. An entire human genome can be sequenced in a matter of weeks for a few thousand dollars. The two most common applications of NGS are sequencing gene panels, which range from a few to several hundred genes, and whole exome sequencing, in which the exon sequence of nearly all ~20,000 human genes are sequenced. Due to technical reasons, the coverage—or percent of bases sequence well—for each gene is often better in gene panels, though a small percentage of coding sequence is missed in each test. Each of these approaches has been used in the research setting for gene discovery and in clinical labs for diagnosis.
Because all protein-coding regions of the genome are evaluated, exome sequencing is an unbiased approach to gene discovery in probands and in families. It eliminates the need for linkage data to narrow the region of the genome where the causative mutation exists in a family or for prior hypotheses about which gene or class of genes is likely to have a mutation, although such additional information can be extremely valuable for interpretation of the results. A significant advantage of exome sequencing is the ability to discover the causative mutation in previously intractable disorders caused by de novo mutation. This is achieved by exome sequencing a “trio” of the affected child and both unaffected parents to identify rare de novo variants that are present only in the affected child. Epileptic encephalopathies, which are the most severe forms of epilepsy, were once thought to be sporadic or acquired conditions; recent exome sequencing studies highlight the importance of de novo mutations in epileptic encephalopathies (13, 14). In retrospect, perhaps this should not have been surprising, given the fact that 75 to 80 percent of patients with Dravet syndrome have a de novo mutation in SCN1A. However, exome studies in large patient cohorts confirm that de novo mutations are important in epileptic encephalopathies beyond Dravet syndrome and that there is significant genetic heterogeneity. That is, mutations in many different genes can cause EE. Recently described genes for epileptic encephalopathy syndromes include KCNT1 (15, 16), SCN8A (17), TBC1D24 (18, 19), GABRA1 (20), GABRB3, ALG13, DNM1 (13, 14), and many others.
Gene panels are selected sets of genes that can be simultaneously sequenced in a patient or cohort of patients. Because fewer genes are sequenced, gene panels cost less than whole exome sequencing, which in turn allows for more patients to be sequenced. First implemented in the research setting, gene panels have been applied to large cohorts for gene discovery, leading to the recent description of causative mutations in many genes, including GRIN2A (21–23), GRIN2B (24), SLC6A1 (25), CHD2, and SYNGAP1 (26). Gene panels can also be used to validate exome findings by sequencing a large cohort of patients with similar phenotypes to identify additional mutations in candidate genes identified by exome sequencing. Proof of principle studies have shown that gene panels that include the most commonly mutated genes can be used for diagnosis, with a 10 to 50 percent diagnostic rate, depending on the panel and patient population (26–28).
NGS in Clinical Diagnostics
Gene panels are now widely applied in the clinical setting and are especially useful for disorders with significant genetic heterogeneity. The number of new epilepsy genes continues to grow, and the phenotypes associated with mutations in each gene are variable, making gene panel testing very attractive for diagnosing epilepsy patients. The number of genes included in clinical epilepsy gene panels varies greatly, from less than 20 to nearly 500. For the clinician, it is important to understand what genes or class of genes each panel includes and what the target patient population is before ordering a test. The majority of panels are designed for early onset epilepsy syndromes, and many aim to be comprehensive; however, if a specific condition or gene is suspected, the clinician should confirm that the selected panel includes the suspected gene or diagnosis. Gene panels for specific subtypes are often available, including progressive myoclonic epilepsy and familial focal epilepsy. Some panels include genes for recessive metabolic disorders and congenital disorders of glycosylation, while others highlight de novo dominant and X-linked causes. Depending on the specific phenotype, “nonepilepsy” genes panels for microcephaly syndromes, intellectual disability syndromes, or episodic ataxia may be more appropriate. The choice of company and of specific gene panel can be overwhelming. Resources to identify tests include Gene-Tests (http://www.genetests.org) and Genetic Testing Registry (https://www.ncbi.nlm.nih.gov/gtr/); consulting with a geneticist or a genetic counselor prior to ordering can be very helpful in designing an efficient and cost-effective testing strategy.
An increasing number of clinical labs offer whole exome sequencing for diagnosis. In contrast to gene panels, one need not peruse the gene list to ensure that a specific gene or diagnosis is covered. Rather, all genes are sequenced simultaneously, and all rare, potentially damaging variants are identified. Initial analysis of the data usually focuses on identifying potential mutations in known disease genes. Exome sequencing in the clinical setting broadly leads to a diagnosis is ~25% of cases overall (29). In a research study of 356 individuals with epileptic encephalopathy, trio exome sequencing identified a causative de novo mutation in ~15% of cases (14).
It is important to remember that clinical exome sequencing is not a gene discovery project—gene discovery still largely belongs in the research lab, although the lines are certainly beginning to blur. If a rare variant is identified by clinical exome sequencing in a gene that has not previously been associated with disease, it may or may not be reported as a likely causative variant or even a variant of uncertain significance. The degree to which uncertain variants are reported varies depending on the testing lab and on the patient and clinician preference for what information they would like to receive. Variants of uncertain significance can be confusing for physicians and families. Additional testing helps in some cases. For example, testing the parents may reveal that a genetic changes is de novo in a patient with epileptic encephalopathy and therefore more likely to be causative. In families with multiple affected individuals, finding the variant in all individuals with disease (and not in healthy individuals) is supportive evidence that the variant is a disease-causing mutation. Consultation with a clinical geneticist or genetic counselor should be considered when the results are uncertain. Just because a variant is de novo does not mean that it is associated with disease; even healthy individuals have an average of one de novo variant that affects protein-coding sequence. Therefore, additional functional studies or the identification of similarly affected patients with de novo mutations in the same gene are usually required to declare a new mutation or gene as causative. The Epilepsy Genetics Initiative (EGI; http://www.cureepilepsy.org/egi) aims to collect exome sequence data from patients with epilepsy and re-analyze the data on a regular basis; as patient data accumulates, novel disease-causing genes may be easier to identify. EGI and similar collaborative efforts will be essential to moving the field forward even more quickly.
Finally, for both gene panels and exomes (and even for single gene tests), availability of family DNA may be important for interpreting results. For example, it is difficult to determine the clinical importance of a rare missense change in a gene for epileptic encephalopathy that has not been previously identified in cases or controls. Testing parents helps determine whether the change is inherited and therefore likely benign, or de novo and likely causative. In familial epilepsy, the ability to determine whether a mutation segregates with disease is similarly important.
What Test Should I Send and When?
There are more options than ever for genetic testing in patients with epilepsy. While the myriad options can be daunting, some simple guidelines can help direct the testing strategy. Single gene testing should rarely be performed. However, in certain cases, testing a specific gene may be appropriate, especially for disorders in which there is a single gene that explains the vast majority of cases. Examples include patients with a clear clinical diagnosis of Dravet syndrome or Rett syndrome where testing SCN1A or MECP2, respectively, will confirm the diagnosis. Another example is GGE with paroxysmal exercise-induced dyskinesia or early onset absence epilepsy, especially when segregating in an autosomal dominant fashion, where SLC2A1 testing for GLUT1 deficiency may yield a positive result. If the single gene test is negative, however, one should move directly to CMA or gene panel testing, depending on the phenotype, rather than serially testing individual genes.
In individuals with “epilepsy-plus”—that is, epilepsy plus any other feature, including developmental delay or intellectual disability, congenital anomaly, autism spectrum disorder, and so on—consider CMA testing to detect CNVs first. This would include patients with nonspecific epileptic encephalopathy. Patients with milder epilepsy but a personal or family history of other neurodevelopmental disorders would also fall into this category; since they are at increased risk of having one of the three common microdeletions, they should also undergo CMA testing. For patients with only a mild form of epilepsy, however, CMA testing has not proven to have a high diagnostic yield. Standard karyotypes are almost never employed for diagnosis but may be used for confirmation of CMA findings. In addition, if Ring 20 syndrome is suspected (long and refractory absence seizures typically lasting 20 minutes and nocturnal tonic seizures), karyotype may be the only method to detect the abnormality.
Key Points
Chromosome microarray testing to detect CNVs should be employed early in the diagnostic process for patients with nonspecific epileptic encephalopathy and epilepsy-plus.
Gene panels are an efficient method to test for genetic mutations when multiple genes can cause the same disorder.
Whole exome sequencing is an unbiased approach to gene discovery in the research lab and genetic diagnosis in the clinic.
Having a genetic diagnosis can improve prognosis and recurrence risk counseling and inform treatment decisions.
When CMA testing is negative, the next step should be gene panel or whole exome testing. There are many considerations when deciding which to select, including phenotype (is a specific panel available?), severity (is a rapid diagnosis required for medical decision-making?), and cost (will insurance cover the test?). If a gene panel is available for the patient's phenotype, this is a reasonable next step. One advantage of gene panel testing is that, in general, the genes included are involved in the suspected disorder by definition, which can make interpretation more straightforward. However, if the gene panels available are not targeted to the phenotype, or if the phenotype is diffuse enough that the clinician is deciding between two or more large panels, exome sequencing should be considered. Serial genetic tests can quickly add up to the cost of a single exome, so considering exome sequencing early in the diagnostic odyssey can be a time- and money-saving step.
When to test depends in part on the age of the patient and the severity of the epilepsy but also on the desires of the family to use the information for family planning. Furthermore, as we move toward an era of precision or personalized medicine, early diagnosis becomes increasingly important. For infantile-onset epilepsies, especially, where early treatment may improve the medical and developmental prognosis for the patient, the more quickly a diagnosis can be made, the more effectively a treatment plan can be delineated. Determining a genetic diagnosis is important for family reasons as well. While the recurrence risk for a de novo disorder is very low, recessive conditions have a 1:4 chance of recurrence with each pregnancy. Dominant conditions have a 50% recurrence risk; although dominant conditions are often milder, some have a wide range of outcomes including devastating early onset cases. Accurate risk information can be extremely valuable for young families.
Conclusion
Our understanding of epilepsy genetics is increasing at an unprecedented rate, and clinical genetic testing options are keeping pace. Knowing how to use genetic testing efficiently for patients can improve diagnosis and inform our ability to provide accurate prognosis and recurrence risk information. For some conditions—and in the near future, for many—genetic diagnosis will guide therapy, greatly improving the lives of patients.