Article Text

Download PDFPDF
Genomic testing in neurology
  1. Vani Jain,
  2. Rachel Irving,
  3. Angharad Williams
  1. All Wales Medical Genomics Service, University Hospital of Wales Healthcare NHS Trust, Cardiff, UK
  1. Correspondence to Dr Vani Jain, All Wales Medical Genomics Service, University Hospital of Wales Healthcare NHS Trust, Cardiff CF14 4XW, UK; vani.jain{at}wales.nhs.uk

Abstract

Genomic testing has been available for neurological conditions for decades. However, in recent years, there has been a significant change in its availability, range and cost, as well as improvements in the technology and knowledge that underpin how the genome is interrogated. Neurologists can encounter a wide range of genetic conditions, and so their understanding of genomic testing is fundamental to modern clinical practice.

  • genetics
  • neurogenetics

Data availability statement

Data sharing not applicable as no datasets generated and/or analysed for this study.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Thousands of genetic conditions involve the neurological system. The UK rare disease framework highlighted that people with rare diseases often experience a lengthy diagnostic delay before a diagnosis is reached.1 The ‘mainstreaming’ of genomic testing across all specialities has increased awareness of genomic tests as part of the diagnostic armament. With this comes the hope that people with rare diseases will have more coordinated care and better access to research and emerging therapies. This article aims to help neurologists increase their understanding of genomic investigations and to provide general guidance on how to proceed after a result is issued.

What is a variant?

A variant is an alteration in the genomic architecture. The term ‘variant’ describes different types of genetic alteration (figure 1) and it has now largely replaced the term ‘mutation’ in common parlance; it means much the same as the word ‘mutant’ but is less likely to be misused as a term of abuse. Comparing a person’s genomic sequencing data with a reference genome can highlight variants that differ (ie, can find the ‘abnormal’ by comparing it with the ‘normal’). The current reference genome, known as GRCh38, is an amalgamation of genomic sequences from a few people. It does not claim to represent an ‘ideal’ human genome and it is well recognised to have its limitations, one of these being that it cannot reflect the natural variation across ethnicities.

Figure 1

Genomic variants can affect chromosomes, nuclear genes and mitochondrial DNA. The molecular confirmation of a genetic diagnosis may require tests that interrogate the genomic material at different levels. Genes account for ~2% of the whole genome. The remaining ~98% is classed as non-coding as it is not translated into protein. Genes are made up of exons (coding sequence) and introns. Although both are transcribed into mRNA, the intronic sequence is spliced out, resulting in a mature mRNA sequence that is translated into a protein.

Variants may be disease related but they are more often completely benign, so that evidence is needed to conclude that a variant is a cause of disease. Although not discussed here, there are also epigenetic changes to the genome that modify gene function through mechanisms that do not alter the DNA sequence.

Genomic variation is necessary for our survival as a species. Although each of our genetic codes is broadly similar, we each have millions of variants that make us unique individuals. Most variants are not ‘disease causing’. Many may minutely increase or decrease disease susceptibility, contributing to ‘polygenic’ conditions. The focus of genomic testing in the clinical setting, as will be discussed here, is on identifying genomic variants associated with monogenic disorders.

Protein-coding genes are transcribed into mRNA and then translated into proteins. For each gene, this process depends on a complex set of mechanisms and a variant may affect any part of this. For example, as well as altering the amino acid sequence of a protein, variants may affect the length or quantity of mRNA/protein that is produced. For other variants, the quantity of protein product may be adequate but it may not be trafficked to the correct part of the cell, or it may be unable to bind to a downstream protein.

There are an estimated 21 000 human genes, many of which have yet to be linked to disease.2 However, over 5000 genetic disorders are recognised and every year new gene/disease associations are ‘discovered’ through a multitude of research groups. It is important to be aware that genomic testing in the clinical and research setting can vary significantly. Clinical laboratories usually report variants in genes where there is a known association with a condition, as these results will probably have the greatest clinical utility. Although research projects can do the same, their aim may be to discover new genes or associations, the results of which may not be available for many years.

Types of genomic test

Genetic conditions can be highly variable in their onset, their features and their progression. Unless there is a pathognomonic sign on examination or a classical clinical/radiological picture, the differential diagnosis can be wide. In addition, many conditions show ‘locus heterogeneity’; variants in different genes can cause the same clinical features (table 1). Therefore, although single-gene testing may be available for some conditions, most genetic testing is broader in order to detect many different conditions.

Table 1

Useful terminology

No test is perfect and each requires its own skill set to make the best use of it. Consequently, depending on the clinical presentation, different types of genomic test may be needed either concurrently or sequentially.3 Discussions with the local genomics laboratory, clinical genetics service or, if available, a neurogenetics multidisciplinary team (MDT) meeting can help with deciding on the best approach. For adult patients who had genetic testing when they were younger, newer genomic tests may be available if a diagnosis is yet to be reached.

Next-generation sequencing (NGS)

‘Sequencing’ of DNA means reading the genetic chemical code letter by letter. Since the completion of the human genome project 20 years ago, sequencing technologies have advanced significantly. Newer, massively parallel sequencing methods are collectively known as ‘NGS’. As the cost has come down, NGS has largely replaced the low throughput testing methods that were previously used, which had involved time-consuming sequencing techniques that made sequential testing of large numbers of genes impractical.

NGS is primarily used to identify small sequence variants. It may also detect mosaic variants (variants that are present in some cells but not others). In addition to small variants in coding and non-coding regions, it can detect larger ‘copy number variants’ (CNVs) involving single exons up to multiple genes, structural rearrangements of chromosomes, unstable repeat sequences and mitochondrial DNA variants. However, NGS is not guaranteed to identify these and other genomic tests may be needed.

Just as different MR imaging sequences can highlight different features within the central nervous system, so NGS testing has its different approaches. Figure 2 illustrates that sequencing can involve looking at a single gene, many genes, all genes or the whole genome. However, even if whole-genome sequencing (WGS) has been performed, this does not necessarily mean that all the whole-genome data has been analysed. A panel of genes of interest (eg, adult-onset dystonia) may be analysed from the WGS data but only variants in that ‘virtual panel’ will be reported. The use of WGS allows a higher quality of sequencing data to be obtained and creates the possibility of data reanalysis for ‘test-negative’ patients in the future as panels expand over time.

Figure 2

Next-generation sequencing (NGS) techniques can be applied in different ways so as to capture the region of interest. This may range from the whole genome down to a single gene. Mitochondrial DNA (mtDNA) sequencing can be requested as a standalone test or variants may be detected through whole-genome sequencing.

Although in many cases, the affected person’s DNA may be the only sample used for NGS testing, there can be value also in obtaining samples from both parents and performing a ‘trio’ analysis, if this is possible. This allows comparison of the ‘child’s’ sequence with their parents. For those with complex phenotypes, especially those involving multiple body systems, trio WGS is likely to be a more efficient testing approach than requesting multiple panels.

Detecting CNVs

A CNV is a dosage imbalance due to loss or gain of genetic material.4 A CNV disorder occurs due to deletion or duplication of an exon, multiple exons, a gene or multiple genes. Examples of genes where CNVs are a well-recognised mechanism are PMP22 (where a whole gene duplication causes Charcot–Marie–Tooth disease type 1A) and DMD (Duchenne and Becker muscular dystrophies, where deletions or duplications of exons within the gene are well-recognised causes of both conditions).

Microarray analysis is used to detect larger CNVs across chromosomes. Microarray technologies include array comparative genomic hybridisation (array CGH) and single-nucleotide polymorphism array (SNP array). Both technologies use ‘hybridisation’ (complementary bonding of single strand DNA molecules) to detect dosage imbalance.

  • An array CGH compares the hybridisation of patient DNA to oligonucleotide probes (small molecules of DNA) distributed across the genome, with that of a reference DNA.

  • An SNP array measures and quantifies hybridisation of patient DNA using DNA probes reflecting genomic regions where there are differences between individuals at a single base pair (SNP).

Each centre usually provides one or the other, although SNP array is becoming used more commonly. Although such testing may not be first line or used at all for some neurological presentations, it is often requested where the phenotype includes intellectual disability, neuropsychiatric features, seizures and/or a distinctive facial appearance. Large CNVs can involve multiple genes, which may cause a combination of clinical features suggesting a particular syndrome.

Genomic testing for unstable repeat expansion disorders

Special consideration needs to be taken for those conditions that are caused by unstable repeat expansions, many of which are neurological, such as Huntington’s disease.5 Repeat sequences are difficult to size accurately using standard NGS. Other test methods such as triplet repeat primed PCR may be needed, although newer NGS techniques are now able to detect these types of variants.6 When requesting an NGS test, it is advisable to check whether such genes of interest are being included or whether separate testing is needed.

Genomic testing for mitochondrial disorders

Mitochondrial disorders are highly variable in their clinical presentation and display significant genetic heterogeneity. Their complexity led to the establishment in the UK of the National Health Service (NHS) Rare Mitochondrial Disorders Service (https://mitochondrialdisease.nhs.uk). Mitochondrial disorders may be due to variants in nuclear genes that follow a Mendelian pattern of inheritance, or due to variants in the mitochondrial DNA that show a matrilineal pattern of inheritance, but which can of course still affect the sons of affected women, although not the children of the sons. When considering a mitochondrial condition, a diagnosis may be identified on a gene panel or through whole-exome/genome sequencing. However, the diagnosis may require testing that specifically sequences the mitochondrial genome7 and depending on the presentation, may require additional samples, such as urine or muscle tissue.

Choosing a test

Introducing genomic testing early in the diagnostic pathway may prevent the need for other investigations. But care should be taken to consider the clinical phenotype, and the results of other investigations that could guide genomic testing and subsequent diagnosis. The family history should be explored as this may suggest a pattern of inheritance if there are other ‘affected’ family members. The wider medical history and presence of non-neurological signs should also be taken into account. Lynch et al provide a useful example of how clinical, biochemical, radiological and genomic information can be brought together in the diagnosis of adult-onset leukodystrophies.8

There is a wide range of genomic tests available for neurological presentations. In England, genomic testing within the NHS is provided in line with the NHS Genomic Test Directory (https://www.england.nhs.uk/publication/national-genomic-test-directories/). This groups tests by specialty/body system and for each test provides eligibility criteria and testing methods. The genes tested under each indication are listed on Panel App (https://panelapp.genomicsengland.co.uk/). Neurological tests form a sizeable proportion of the tests available within the directory.

For some patients, it is clear which test to request. For others, with previous ‘normal’ results from genomic investigations in the family, or a wide differential diagnosis, or a complex clinical picture, the testing route may be less clear. Increasingly, regional and national multidisciplinary meetings and clinics are used to help formulate investigation and management plans.9 10 Where there is uncertainty before testing, it is advisable to discuss the case within a neurogenetics meeting in order to garner opinion on possible testing strategies. The direction taken will depend not only on clinical presentation and family history but also on the patient’s wishes and family structure. Complex results, particularly where there is an unexpected variant, can also benefit from wider discussion so as to plan additional investigations or decide on how to give a difficult result.

DNA storage

When a genomic test is requested, any remaining DNA is usually stored automatically. However, DNA storage without any testing may be an option in various situations; for example, for patients who are worried about the results of a diagnostic genomic test and they do not want to have results while they are alive. For those with a rapid disease progression, DNA storage is advisable. DNA can be stored for many years and, if of suitable quality, could be tested years later for the benefit of family members.

Consent for genomic testing

Although genetic counselling has traditionally been seen as the remit of clinical genetics services, it is a necessary part of any diagnostic genomic test. The possible results of a genomic test will differ depending on the exact test. Figure 3 lists some key discussion points to be broached during the consent process. There can be unrealistic expectations about what genomic testing can deliver. It is a common misconception that a genetic cause has been fully ‘ruled out’ if no causative variant is identified. This may be true in some circumstances, such as in Huntington’s disease where a specific type of variant in a specific part of the HTT gene causes disease. However, if a patient with a personal and family history of ataxia has a series of ‘normal’ genomic tests and there is no obvious non-genetic explanation for this, a genetic cause is still highly likely.

Figure 3

Discussion points during consent for genomic tests.

Where a genomic test could potentially diagnose numerous different conditions, it is not always feasible to discuss each of these in detail with the patient. For gene panels, clinicians can give a general description of the type of conditions being tested for, whereas for broader tests (eg, analysing chromosomes or full exome/genome sequencing) they should discuss the reasoning behind why this approach is being taken as well as the possibility of incidental findings. An incidental finding refers to an unexpected finding, unrelated to the primary clinical indication for which testing was performed. The broader the test, the greater the chance of identifying significant incidental variants.11 12 Even with phenotype-specific panel testing, a confirmed diagnosis could be associated with a range of other neurological and non-neurological features. For any genomic test, it is imperative to discuss the possibility of variants of uncertain significance. This can be complex to explain to patients but highlighting the possibility of their identification can save much confusion in the future.

Genomic results can have wide-ranging implications. Some patients have a lived experience of a condition in their family, while for others, the suggestion of a genetic diagnosis could be a shock. As well as discussing the reason for the test and the possible results, issues may be raised regarding the impact on family members, personal relationships, mental health and long-term care, or other concerns about employment, driving, family planning, insurance and confidentiality.13 Moreover, even if a clinical diagnosis is highly likely, some patients may not want it confirmed with a genomic test. The question of ‘what difference will a genetic result make?’ is frequently asked in a genetics clinic and should be tackled with honesty, giving patients and families the time to decide how they may wish to proceed. The Joint Committee on Genomics in Medicine report, ‘Consent and confidentiality in genomic medicine’, uses case studies to provide essential guidance for all healthcare professionals on a range of scenarios that can arise.14

Predictive testing for people who are asymptomatic but at risk of a familial condition, such as Huntington’s disease, is undertaken by clinical genetics services. Usually over a series of appointments, patients are counselled about the condition which involves discussing their risk of inheriting the familial variant, what the results may (or may not) tell them and the potential reproductive options, among other things. However, neurologists may be asked to test individuals with non-specific symptoms where the parent, grandparent or more distant relative is affected with an autosomal dominant condition. If it is unclear whether a test will be diagnostic or predictive, then we strongly recommend discussing this with clinical genetics and/or a neurogenetics MDT. In the scenario where an individual is at 25% risk (ie, where the grandparent but not the parent is clinically affected) identifying the familial variant will automatically reveal the genetic status of the parent. Although this is not an entirely unfamiliar situation within clinical genetics, no two situations are alike and careful counselling is required.

Requesting genomic testing

Most genomic tests can be performed using a blood sample taken in an EDTA tube; however, a sample of lithium heparin is required for a karyotype. Different genomic tests require a certain quantity and quality of DNA. Testing also needs to have been validated for the sample type. For example, the laboratory may perform a test using blood but be unable to provide a reliable result using DNA from saliva. For cases where RNA testing is required, a PAXgene RNA tube is needed and can be requested from the local genomics laboratory.

For many centres, samples and requests will go through a local genomics laboratory who forward the request and sample to another laboratory if needed. Many laboratories have specialist expertise in particular areas and are available for advice if needed.

The information provided by the clinician is highly valuable to the laboratory and should include detailed phenotypic information, age of onset, progression, imaging and family history. Genomic clinical scientists may request further information or suggest alternative testing strategies. If the clinician has a differential diagnosis in mind, this should definitely be mentioned as it may prompt closer analysis of certain genes in the event that no obvious variants are identified.

What is variant interpretation?

We each have millions of variants that differ from the reference genome. The identification of a variant, even in a gene that looks to fit with the patient’s clinical features, is not automatically deemed to be ‘disease causing’. Following quality control and filtering of sequencing variants, the remaining variants need to be interpreted individually by a genomic clinical scientist (figure 4). If trio analysis were performed, the data from the parents will have aided in the filtering of sequence variants. Variant interpretation takes time as it requires working through a standardised process (figure 5). The laboratory may request further phenotypic information from the clinician or suggest discussion at a local MDT before releasing the official result.15

Figure 4

The sequencing data from the patient may be from a single gene, a gene panel or from whole exome/genome sequencing. With each stage of data filtering, the number of variants lessens.

Figure 5

Variant interpretation is a process that involves answering a series of questions. Laboratories use databases and programmes to answer some of the questions but it is heavily reliant on the skills of genomic clinical scientists. This process needs to be repeated for each variant separately.

Receiving results with no confirmed diagnosis

In addition to the actual result, genomic test reports can include helpful information particularly where there is no clear diagnosis. The report should include the testing method and types of variants that were and were not detectable. Other information may include additional tests that could be arranged, such as parental testing, advice on specialist clinics or specialists with whom the case could be discussed. Clinical genetics teams are also available, either through referral, MDTs or through an on-call service if the query is more urgent.

Table 2 provides questions to be considered if testing does not identify any significant genomic variants. For situations where a variant of uncertain significance has been reported, further work and review of the variant in the future may be needed (figure 6).

Figure 6

Reclassification of a variant of uncertain significance (VUS) means that additional evidence has been found that clarifies the variant as being B/LB or P/LP. Only P/LP variants are considered diagnostic. As such, P/LP variants can be used for predictive/carrier testing in unaffected family members and reproductive options may be available. The initial VUS test report may give suggestions on what type of additional evidence could be used. If there is additional information available, the clinician can request reinterpretation of the variant but there is no guarantee that reclassification will happen.

Table 2

Case 1: dentatorubral pallidoluysian atrophy (DRPLA)

A 20-year-old man, J, had been under the care of neurology and clinical genetics for many years. He had intellectual disability from early childhood. At age 8, he developed generalised tonic–clonic and absence seizures. His behaviour became more challenging in his teens. Neurological examination in childhood had been normal and multiple MR scans of the brain up to age 18 had shown no clear abnormalities. Cerebrospinal fluid (CSF) examination and metabolic investigations were normal. Genomic testing over the years had included testing for fragile X syndrome, a karyotype, an array CGH and a gene panel for early-onset epilepsy. In addition, he had trio whole-exome sequencing through the research project known as 'Deciphering Developmental Disorders'. No diagnosis had been identified.

His seizures became increasingly difficult to control. He required a wheelchair due to worsening balance and he was losing weight. His mother discovered that a distant relative had died of DRPLA and she asked if her son could have the same condition (figure 7).

Figure 7

Pedigree showing the family history for case 1. Black arrow=proband. DRPLA, dentatorubral pallidoluysian atrophy.

DRPLA is a very rare autosomal dominant neurodegenerative condition with similarities to Huntington’s disease.16 When the condition presents in childhood, clinical features include ataxia, intellectual disability and seizures (usually progressive myoclonic epilepsy). It is caused by a trinucleotide repeat expansion in the ATN1 gene. J’s history had not prompted consideration of DRPLA before, as it is so rare, his seizures and history were not typical, and the family history was previously unknown.

Appropriate genomic testing had been performed for J’s initial presentation, but none of the tests requested at the time would have detected an unstable repeat condition such as DRPLA. A diagnostic ATN1 test was arranged, which confirmed a ‘pathogenic’ expansion with 68 CAG repeats in one copy of the gene. J’s diagnosis also confirms that his sister has a 50% risk of inheriting the ATN1 pathogenic expansion and she can come forward for predictive genetic testing.

Case 2: frontotemporal dementia (FTD)

A 72-year-old woman, M, saw a neurologist at a specialist dementia clinic and received a clinical diagnosis of FTD. She also had a family history of dementia (figure 8). C9orf72 gene testing was arranged with no pathogenic variant identified. This was followed by a gene panel for neurodegenerative conditions. Testing identified a variant in the presenilin-1 (PSEN1) gene that was classed as a variant of uncertain significance. The variant was a deletion and insertion in an early part of the coding region. It was predicted by bioinformatic software to be ‘deleterious’ in nature and was not present in any control population databases. This, along with the fact that PSEN1 variants can be associated with behavioural variant FTD as well as Alzheimer’s disease, suggested it could be the cause of M’s diagnosis of FTD. However, the variant had not been described in any clinical databases or in the medical literature, which made its significance uncertain. The report concluded that a diagnosis ‘must rest on clinical grounds alone’, but familial segregation analysis may be appropriate.

Figure 8

Pedigree showing the family history for case 2. Black arrow=proband. FTD, frontotemporal dementia.

The presence or absence of Alzheimer’s disease biomarkers in the CSF could provide additional information but the family did not feel it appropriate to put M through an invasive procedure, and amyloid PET scanning was not possible locally. M’s sister was offered a neurology appointment at the specialist clinic where she could have more detailed phenotyping and be tested for the PSEN1 variant. The family were informed that finding the variant in a second affected relative may still be insufficient to clarify the significance of the variant.

There were no further genomic tests that could be suggested. Her family members, including her children, cannot be offered predictive genetic testing while the result remains a variant of uncertain significance, as testing them would not give clarity on their risk. This also means that reproductive options, such as prenatal diagnosis and preimplantation genetic diagnosis, are not possible. In time, if other affected individuals are found to have the same (or a similar) PSEN1 variant, whether they be in the family or in other families, this may lead to reclassification of the variant as either benign or pathogenic.

Case 1 highlights that review of the clinical picture and change in family history can provide crucial information that leads to diagnosis. Despite multiple ‘normal’ genomic results, a genetic cause was still highly likely. Case 2 highlights the frustration of a result that feels like it explains the diagnosis but where the evidence of the variant being (likely) pathogenic is not strong enough to establish this or to change its classification. As families make life-changing decisions based on results, a high level of evidence is required for variants to be classed as likely or definitely pathogenic.

Next steps after genetic diagnosis

Confirmation of a genetic diagnosis allows more detailed information on the condition to be provided and for families to be directed towards support groups, treatment options and research opportunities. If the diagnosis suggests a multisystem condition, patients may need referral to other specialists. Many people with rare conditions become experts themselves and find that they are required to explain their diagnosis to other healthcare professionals.

A diagnosis, whether confirmed or suspected, frequently leads to questions about implications for other family members or reproductive options. Referral to a clinical genetics service can be made for discussion of the genomic result, risk to family and counselling about reproductive options such as prenatal testing and preimplantation genetic diagnosis.17 18

Conclusion

Genomic testing is broadening our understanding of well-known conditions and identifying new ones. Genomics, integrated with other ‘omics’ technologies, is transforming our understanding of disease pathogenesis.19 Despite technical advances, clinical skill and judgement remain just as important as always.

At present, treatment is a possibility for a minority of those with a genetic diagnosis; however, recent progress in the development of therapies for conditions such as spinal muscular atrophy, Duchenne muscular dystrophy and Huntington’s disease provide hope for the future. Many patients will choose to undergo genomic testing for diagnostic clarity but others will not, or will wish to delay a decision and to reflect first. Open conversations and shared decision-making are essential if genomic testing is to benefit our patients and their families.

Further reading

Key points

  • The clinical phenotype is key to determining the most appropriate genomic test. It is important to consider the type of genomic testing method before requesting a test to ensure likely diagnoses are covered.

  • Discussion with clinical geneticists, genetic counsellors and genomic clinical scientists or within a neurogenetics multidisciplinary team may help to decide on testing pathways and how to relay information to patients.

  • Informed consent is essential to ensure that the reasons for testing and possible outcomes are understood.

  • A ‘normal’ genomic test result does not necessarily exclude a genetic diagnosis. If a confirmed genetic diagnosis is not reached, consider further steps that could be taken.

Data availability statement

Data sharing not applicable as no datasets generated and/or analysed for this study.

Ethics statements

Patient consent for publication

Ethics approval

Not applicable.

Acknowledgments

We would like to thank the patients and families who have contributed to this paper.

References

Footnotes

  • Contributors VJ wrote and revised the article. RI and AW reviewed and revised the article.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Commissioned. Externally peer reviewed by Ed Newman, Glasgow, UK, and Helen Grote, London, UK.