Genetic Screening and Diagnosis in Epilepsy?
Genetic Screening and Diagnosis in Epilepsy?
For the practising neurologist seeing a patient in whom a genetic diagnosis is suspected, there may be up to four types of genetic test generally available. These are: candidate gene tests, gene panels, array comparative genomic hybridization and, in a very few centers, clinical exome sequencing. A summary of aspects of these methods is given in Table 1.
A candidate gene test may be the most direct route to a genetic diagnosis. Such testing relies on the innate Bayesian behavior of clinicians: the history suggests a diagnosis, and the informed clinician requests a specific candidate gene to be screened for pathogenic variants. All genetic testing requires the clinician to be alert to the possibility of a genetic diagnosis, and for candidate gene testing, the clinician must also have knowledge of the gene or genes that may be altered to produce the observed phenotype. There are some epilepsies with a very characteristic phenotype, which should trigger testing of a specific gene. Candidate gene testing typically uses Sanger sequencing, to which methods, such as multiplex ligation-dependent probe amplification, may be added for detecting exonic-level changes, such as exonic deletions. Dravet syndrome is amongst the best examples: with a typical history, over 80% of cases will have a pathogenic change in the gene SCN1A. Making the diagnosis can lead to treatment changes, with potential significant impact on seizure control and quality-of-life for the patient and carers. Dravet syndrome also illustrates the challenges for the clinician, especially perhaps for those, the current majority, for whom training will not have incorporated much practical genetics. In about 15% of Dravet syndrome cases, no change will be found in SCN1A; there may be mutations in several other genes causing a phenotype like Dravet syndrome, including CHD2,HCN1,GABRA1 and STXBP1: changes in more than one gene can cause a similar phenotype. Conversely, finding a mutation in SCN1A does not mean that the patient has Dravet syndrome – there are other SCN1A-related epilepsies, and some people without epilepsy or a family history of epilepsy may carry an SCN1A mutation. For clinicians, guidelines from the International League against Epilepsy's Genetics Commission should prove helpful when considering SCN1A. But, the number of genes implicated in epilepsy, complicated genotype–phenotype relationships and a relative lack of information militate against simple choices in candidate gene testing: similar guidelines do not exist for each implicated gene. Even when the clinician may seek testing of a number of candidates from an informed position, screening one gene after another is slow, costly and often impossible.
These issues drove development of gene panels, comprising a selection of implicated genes, for example, a panel of genes associated with epilepsy or epileptic encephalopathies. Technological advances have made testing of a gene set possible at a price some reimbursement systems can contemplate. The work of Lemke et al. stands out. Using next-generation methods, 265 genes were screened in each of 33 patients, with presumed disease-causing mutations rapidly identified in 16/33; another report documents a diagnostic yield of 47%. Some centers now offer clinically accredited panel sequencing in epilepsy. Panels have their limitations. They include only the selected candidate genes. Whilst gene discovery, the novel attribution of a phenotype to a gene, may not be the regular domain of the practicing clinician, nevertheless there may be a gene not yet suspected to be related to a particular type of epilepsy that is actually the cause of the phenotype in question: if that gene is not included in the panel, a genetic cause will be missed. Gene panels may also not identify microdeletions and microduplications, which have emerged as important genetic contributors to epilepsy. The cost of most commercial panels, though usually less than sequencing more than two candidate genes, is typically more than the cost of tests that provide more data.
Array comparative genomic hybridization (aCGH) is offered by many clinical genetics laboratories, and because of its higher resolution and reliability, it has replaced karyotyping as the first-line test in complex epilepsy phenotypes that do not implicate an obvious candidate gene. aCGH is indicated when the presenting epilepsy is syndromic in the sense of being associated with other features, such as facial or somatic dysmorphism, intellectual disability, autism spectrum disorder or multiple comorbidities. Microdeletions and microduplications [within the spectrum of changes called copy number variants (CNVs)] are being increasingly reported in association with complex epilepsies, and have sometimes pointed to novel candidate epilepsy genes. For example, CNVs have been reported over the last year as being especially relevant in childhood epilepsies with complex phenotypes, in Rolandic epilepsy and in genetic generalized epilepsies with intellectual disability. Genotype–phenotype correlation with long-established CNVs associated with epilepsies has become possible as increasing numbers of cases have been published. The first CNV associated with a common epilepsy, 15q13.3 microdeletion, seen in 12 patients in the classic 2009 report has now been studied in a 246-case series, 28% of whom had seizures; neuropsychiatric manifestation was common, major congenital malformations were not. Such phenotypic delineation will facilitate genetic screening and interpretation for future practice. Depending on case series and criteria for inclusion, about 12% of people with complex epilepsy might have a CNV considered relevant, making aCGH important and relevant in current practice.
The total amount of genetic information in an individual is much greater than is typically evaluated in any clinical or research setting today. Even if we ignore the epigenome, microRNAs and other RNAs, mosaicism, somatic mutation and other mechanisms (that are unlikely to be irrelevant in epilepsy), there is still a huge amount of genetic information largely untapped by most currently available clinical tests. Candidate and panel tests miss all those genes not included; aCGH does not have the resolution to detect changes below a certain size. Here, technology again will have a huge impact, as clinical evaluation of the exome and, eventually, the genome becomes available. Genome sequencing for clinical purposes must become a key outcome of the investment in human genome research, and is an aim of some nations (http://www.genomicsengland.co.uk/).
The clinical application of whole exome sequencing (WES) and whole genome sequencing (WGS) is certainly challenging. In addition to technical difficulties, interpretation of WES and WGS is likely to remain a hurdle for some time yet, requiring close collaboration between genetic scientists, bioinformaticians and, crucially, the clinicians who see the patients being sequenced. Whilst phenotyping may need to become more sophisticated, clinicians will still be needed to determine clinical relevance. Encouragingly, clinical WES has commenced and in some centers is well advanced. In epilepsy, WES has still largely been a research tool, with dramatic discoveries (, and see below). WES has demonstrated important advantages over candidate gene testing and aCGH, especially when applied to trios (proband and parents) in the search for causal de-novo mutations. WGS has proven even more powerful, for example generating a diagnostic yield of 42% in a cohort of individuals with severe intellectual disability without genetic diagnosis after WES and aCGH, in a landmark study. In time, WGS is likely to be the most useful single clinical genetic test.
Currently, there are few clinical WES studies in epilepsy, in which the focus is on finding a known cause rather than gene discovery per se. Results from clinical diagnostic exome sequencing showed that a molecular diagnosis was achieved in 35% of people with epilepsy. In the overall cohort of 500 people reported, with a variety of conditions, it was notable that trio sequencing was more productive than sequencing of a single case, and that about a quarter of positive findings were within genes characterized within the two preceding years, demonstrating the challenge faced by laboratories offering candidate gene or gene panel testing. We can anticipate, therefore, that WES and WGS are likely to soon supercede candidate and panel testing.
Genetic Screening Today
For the practising neurologist seeing a patient in whom a genetic diagnosis is suspected, there may be up to four types of genetic test generally available. These are: candidate gene tests, gene panels, array comparative genomic hybridization and, in a very few centers, clinical exome sequencing. A summary of aspects of these methods is given in Table 1.
A candidate gene test may be the most direct route to a genetic diagnosis. Such testing relies on the innate Bayesian behavior of clinicians: the history suggests a diagnosis, and the informed clinician requests a specific candidate gene to be screened for pathogenic variants. All genetic testing requires the clinician to be alert to the possibility of a genetic diagnosis, and for candidate gene testing, the clinician must also have knowledge of the gene or genes that may be altered to produce the observed phenotype. There are some epilepsies with a very characteristic phenotype, which should trigger testing of a specific gene. Candidate gene testing typically uses Sanger sequencing, to which methods, such as multiplex ligation-dependent probe amplification, may be added for detecting exonic-level changes, such as exonic deletions. Dravet syndrome is amongst the best examples: with a typical history, over 80% of cases will have a pathogenic change in the gene SCN1A. Making the diagnosis can lead to treatment changes, with potential significant impact on seizure control and quality-of-life for the patient and carers. Dravet syndrome also illustrates the challenges for the clinician, especially perhaps for those, the current majority, for whom training will not have incorporated much practical genetics. In about 15% of Dravet syndrome cases, no change will be found in SCN1A; there may be mutations in several other genes causing a phenotype like Dravet syndrome, including CHD2,HCN1,GABRA1 and STXBP1: changes in more than one gene can cause a similar phenotype. Conversely, finding a mutation in SCN1A does not mean that the patient has Dravet syndrome – there are other SCN1A-related epilepsies, and some people without epilepsy or a family history of epilepsy may carry an SCN1A mutation. For clinicians, guidelines from the International League against Epilepsy's Genetics Commission should prove helpful when considering SCN1A. But, the number of genes implicated in epilepsy, complicated genotype–phenotype relationships and a relative lack of information militate against simple choices in candidate gene testing: similar guidelines do not exist for each implicated gene. Even when the clinician may seek testing of a number of candidates from an informed position, screening one gene after another is slow, costly and often impossible.
These issues drove development of gene panels, comprising a selection of implicated genes, for example, a panel of genes associated with epilepsy or epileptic encephalopathies. Technological advances have made testing of a gene set possible at a price some reimbursement systems can contemplate. The work of Lemke et al. stands out. Using next-generation methods, 265 genes were screened in each of 33 patients, with presumed disease-causing mutations rapidly identified in 16/33; another report documents a diagnostic yield of 47%. Some centers now offer clinically accredited panel sequencing in epilepsy. Panels have their limitations. They include only the selected candidate genes. Whilst gene discovery, the novel attribution of a phenotype to a gene, may not be the regular domain of the practicing clinician, nevertheless there may be a gene not yet suspected to be related to a particular type of epilepsy that is actually the cause of the phenotype in question: if that gene is not included in the panel, a genetic cause will be missed. Gene panels may also not identify microdeletions and microduplications, which have emerged as important genetic contributors to epilepsy. The cost of most commercial panels, though usually less than sequencing more than two candidate genes, is typically more than the cost of tests that provide more data.
Array comparative genomic hybridization (aCGH) is offered by many clinical genetics laboratories, and because of its higher resolution and reliability, it has replaced karyotyping as the first-line test in complex epilepsy phenotypes that do not implicate an obvious candidate gene. aCGH is indicated when the presenting epilepsy is syndromic in the sense of being associated with other features, such as facial or somatic dysmorphism, intellectual disability, autism spectrum disorder or multiple comorbidities. Microdeletions and microduplications [within the spectrum of changes called copy number variants (CNVs)] are being increasingly reported in association with complex epilepsies, and have sometimes pointed to novel candidate epilepsy genes. For example, CNVs have been reported over the last year as being especially relevant in childhood epilepsies with complex phenotypes, in Rolandic epilepsy and in genetic generalized epilepsies with intellectual disability. Genotype–phenotype correlation with long-established CNVs associated with epilepsies has become possible as increasing numbers of cases have been published. The first CNV associated with a common epilepsy, 15q13.3 microdeletion, seen in 12 patients in the classic 2009 report has now been studied in a 246-case series, 28% of whom had seizures; neuropsychiatric manifestation was common, major congenital malformations were not. Such phenotypic delineation will facilitate genetic screening and interpretation for future practice. Depending on case series and criteria for inclusion, about 12% of people with complex epilepsy might have a CNV considered relevant, making aCGH important and relevant in current practice.
The total amount of genetic information in an individual is much greater than is typically evaluated in any clinical or research setting today. Even if we ignore the epigenome, microRNAs and other RNAs, mosaicism, somatic mutation and other mechanisms (that are unlikely to be irrelevant in epilepsy), there is still a huge amount of genetic information largely untapped by most currently available clinical tests. Candidate and panel tests miss all those genes not included; aCGH does not have the resolution to detect changes below a certain size. Here, technology again will have a huge impact, as clinical evaluation of the exome and, eventually, the genome becomes available. Genome sequencing for clinical purposes must become a key outcome of the investment in human genome research, and is an aim of some nations (http://www.genomicsengland.co.uk/).
The clinical application of whole exome sequencing (WES) and whole genome sequencing (WGS) is certainly challenging. In addition to technical difficulties, interpretation of WES and WGS is likely to remain a hurdle for some time yet, requiring close collaboration between genetic scientists, bioinformaticians and, crucially, the clinicians who see the patients being sequenced. Whilst phenotyping may need to become more sophisticated, clinicians will still be needed to determine clinical relevance. Encouragingly, clinical WES has commenced and in some centers is well advanced. In epilepsy, WES has still largely been a research tool, with dramatic discoveries (, and see below). WES has demonstrated important advantages over candidate gene testing and aCGH, especially when applied to trios (proband and parents) in the search for causal de-novo mutations. WGS has proven even more powerful, for example generating a diagnostic yield of 42% in a cohort of individuals with severe intellectual disability without genetic diagnosis after WES and aCGH, in a landmark study. In time, WGS is likely to be the most useful single clinical genetic test.
Currently, there are few clinical WES studies in epilepsy, in which the focus is on finding a known cause rather than gene discovery per se. Results from clinical diagnostic exome sequencing showed that a molecular diagnosis was achieved in 35% of people with epilepsy. In the overall cohort of 500 people reported, with a variety of conditions, it was notable that trio sequencing was more productive than sequencing of a single case, and that about a quarter of positive findings were within genes characterized within the two preceding years, demonstrating the challenge faced by laboratories offering candidate gene or gene panel testing. We can anticipate, therefore, that WES and WGS are likely to soon supercede candidate and panel testing.
Source...